Clinical Dilemmas In: Diabetes

14.
9 mm 187 x 235 mm
Clinical Dilemmas in
Clinical Dilemmas in Diabetes

Diabetes
Edited by
Adrian Vella MD is the Earl and Annette R. McDonough Professor of Medicine, and the Research
Diabetes
Chair of the Division of Diabetes, Endocrinology & Metabolism in the Department of Medicine,
Mayo Clinic College of Medicine, USA.
Clinical Dilemmas in Diabetes answers the clinical questions commonly encountered when
diagnosing, treating, and managing patients with diabetes and its associated complications.
Designed to support informed, evidence-based care, this authoritative clinical guide includes
contributions from leading endocrinologists and diabetes researchers that discuss a diverse range
of recent developments. Concise and focused chapters cover prediabetes, diabetes diagnosis, initial
Second Edition
evaluation and management, disease complications, and cardiovascular disease and risk factors.
Edited by Adrian Vella
Now in its second edition, Clinical Dilemmas in Diabetes contains extensively reviewed and
revised information throughout. New and updated chapters examine prediction, diagnosis, and
management of early Type 1 diabetes, ophthalmic complications, screening asymptomatic patients
for cardiovascular disease, new agents for treatment of dyslipidemia, closed loop systems in Type 1
diabetes, upper gastrointestinal manifestations, managing hyperglycemia in critically ill patients,
and more. Edited by Dr. Vella at the Mayo Clinic, this highly practical resource:
Vella
• Encourages evidence-based clinical decision-making, rather than algorithm-based approaches
• P rovides clear guidance on common problematic areas, especially in cases where conflicts in
treatment for the disease and the complications occur
• E mphasizes the importance of translating the results of clinical trials to individual care and
management of diabetes
• C
ontains effective learning and revision tools, including learning points, chapter introductions
Edition
Second
and summaries, tables, figures, charts, and full references
Part of the popular Clinical Dilemmas series, Clinical Dilemmas in Diabetes is a must-have guide
for anyone involved in the treatment of patients with diabetes, particularly endocrinologists,
diabetes specialists and consultants, cardiologists, residents, fellows, specialist nurses, and general
practitioners with an interest in diabetes.
Cover Design: Wiley

Cover Images: © dboystudio/Shutterstock, SCIEPRO/SCIENCE
PHOTO LIBRARY/Getty Images, MIRIAM MASLO/Science Source,
DR P. MARAZZI/Science Source, JUAN GAERTNER/SCIENCE
PHOTO LIBRARY/Getty Images, Scott Camazine/Getty Images
www.wiley.com
Diabetes
Diabetes
Second Edition
EDITED BY
Adrian Vella MD FRCP(Edin.)

Division of Diabetes, Endocrinology & Metabolism
Earl and Annette R. McDonough Professor
Mayo Clinic
Rochester, MN, USA
This edition first published 2022
© 2022 John Wiley & Sons Ltd
Edition History
Blackwell Publishing Ltd (1e, 2011)
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or
otherwise, except as permitted by law. Advice on how to obtain permission to reuse material
from this title is available at http://www.wiley.com/go/permissions.
The right of Adrian Vella to be identified as the author of the editorial material in this work has
been asserted in accordance with law.
Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley
products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some
content that appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty
The contents of this work are intended to further general scientific research, understanding, and
discussion only and are not intended and should not be relied upon as recommending or
promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In
view of ongoing research, equipment modifications, changes in governmental regulations, and
the constant flow of information relating to the use of medicines, equipment, and devices, the
reader is urged to review and evaluate the information provided in the package insert or
instructions for each medicine, equipment, or device for, among other things, any changes in the
instructions or indication of usage and for added warnings and precautions. While the publisher
and authors have used their best efforts in preparing this work, they make no representations or
warranties with respect to the accuracy or completeness of the contents of this work and
specifically disclaim all warranties, including without limitation any implied warranties of
merchantability or fitness for a particular purpose. No warranty may be created or extended by
sales representatives, written sales materials or promotional statements for this work. The fact
that an organization, website, or product is referred to in this work as a citation and/or potential
source of further information does not mean that the publisher and authors endorse the
information or services the organization, website, or product may provide or recommendations it
may make. This work is sold with the understanding that the publisher is not engaged in
rendering professional services. The advice and strategies contained herein may not be suitable
for your situation. You should consult with a specialist where appropriate. Further, readers should
be aware that websites listed in this work may have changed or disappeared between when this
work was written and when it is read. Neither the publisher nor authors shall be liable for any loss
of profit or any other commercial damages, including but not limited to special, incidental,
consequential, or other damages.
Library of Congress Cataloging-in-Publication Data
Names: Vella, Adrian, editor.
Title: Clinical dilemmas in diabetes / edited by Adrian Vella.
Description: Second edition. | Hoboken, NJ : Wiley-Blackwell, 2022. |
Includes bibliographical references and index.
Identifiers: LCCN 2021025568 (print) | LCCN 2021025569 (ebook) | ISBN
9781119603160 (paperback) | ISBN 9781119603177 (adobe pdf) | ISBN
9781119603184 (epub)
Subjects: MESH: Diabetes Mellitus–therapy | Diabetes Mellitus–diagnosis |
Evidence-Based Medicine–methods
Classification: LCC RC660.4 (print) | LCC RC660.4 (ebook) | NLM WK 810 |
DDC 616.4/62–dc23
LC record available at https://lccn.loc.gov/2021025568
LC ebook record available at https://lccn.loc.gov/2021025569
Cover Design: Wiley
Cover Images: © dboystudio/Shutterstock, SCIEPRO/SCIENCE PHOTO LIBRARY/Getty
Images, MIRIAM MASLO/Science Source, DR P. MARAZZI/Science Source, JUAN
GAERTNER/SCIENCE PHOTO LIBRARY/Getty Images, Scott Camazine/Getty Images
Set in 8.75/12pt Minion by Straive, Pondicherry, India
10 9 8 7 6 5 4 3 2 1
Contents
List of Contributors, vii 8 Does HbA1c Remain the Most Important Therapeutic
Target in Outpatient Management of Diabetes? 101
Preface, x Kristen Gonzales and Steve A. Smith
9 Technology Issues: Continuous Glucose Monitoring,
Insulin Pumps, and Closed Loop Control for Patients
Part I Prediabetes and the Diagnosis with Diabetes 115
of Diabetes Ravinder Jeet Kaur, Shafaq R. Rizvi and Yogish C. Kudva
1 “Is Prediabetes a Risk Factor or Is It a Disease?” 3 10 Optimizing Diet in Patients with Diabetes 124
Jacob Kohlenberg and Adrian Vella Meera Shah
2 Early Diagnosis of Type 1 Diabetes – Useful or a Pyrrhic 11 Are Insulin Sensitizers Useful Additions to Insulin
Victory? 17 Therapy? 133
Paolo Pozzilli and Silvia Pieralice John W. Richard, III and Philip Raskin
12 Incretin‐Based Therapy for the Management of Type 2
3 Reclassifying or Declassifying Diabetes? Can Clinical
Diabetes 146
Characteristics Guide Classification and
Treatment? 37 Kristin Gonzales and Adrian Vella
Adrian Vella
4 How Should Secondary Causes of Diabetes
Part III Diagnosis and Management
Be Excluded? 45
Tomás P. Griffin, Aonghus O’Loughlin, and Sean F. of Cardiovascular Risk Factors
Dinneen and Cardiovascular Disease
5 How to Screen Appropriately for Monogenic Diabetes 68 13 Screening Patients with Prediabetes and Diabetes
Adrian Vella for Cardiovascular Disease 163
Sabreen Ahmed, Saritha Tirumalasetty and Vivian Fonseca
14 Choosing Medications for Type 2 Diabetes – What
Weighting Should Be Given to Cardiovascular Risk
Reduction? 174
Part II Initial Evaluation and Management
Adrian Vella
of Diabetes
15 Choosing Medications for Weight Loss in Type 2 Diabetes
6 Managing Gestational Diabetes During and After Mellitus 187
Pregnancy 75 Shubhada Jagasia, Chase Dean Hendrickson and
Aoife M. Egan and Fidelma P. Dunne Alexander J. Williams
7 What Is the Role of Self‐Monitoring in Diabetes? Is There 16 Are Statins the Optimal Therapy for Cardiovascular Risk
a Role for Postprandial Glucose Monitoring? How Does in Patients with Diabetes? What Newer Agents Are There
Continuous Glucose Monitoring Integrate into Clinical for the Treatment for Dyslipidemia in Diabetes? Are
Practice? 87 Triglycerides an Important Risk Factor for Diabetes? 198
Rami Almokayyad and Robert Cuddihy Recie Davern and Timothy O’Brien
v
vi Contents
17 New Agents for Treatment of Dyslipidemia 211 21 Diagnosis and Management of Ophthalmic

Vinaya Simha Complications of Diabetes 252
18 The Role of Bariatric Surgery in Obese Patients
Andrew J. Barkmeier and Konstantin Astafurov
with Diabetes: Primary or Rescue Therapy? 220 22 Upper Gastrointestinal Manifestations of Diabetes 269
Praveena Gandikota and Blandine Laferrère Michael Camilleri
19 Treatment Strategies in Patients with Diabetes Mellitus
and Ischemic Heart Disease 228 Index 280
Madeline K. Mahowald, Chiam Leker Locker, Mandeep
Singh and Robert L. Frye
Part IV Management of Disease

Complications
20 Diabetes Management in Patients with Critical and
Non-Critical Illness 241
Kalpana Muthusamy, Kristin Grdinovac and
John M. Miles
List of Contributors
Sabreen Ahmed Robert Cuddihy Aoife M. Egan

Section of Endocrinology International Diabetes Center Division of Endocrinology, Diabetes,
Tulane University School of Medicine World Health Organization Collaborating Metabolism and nutrition
New Orleans, LA Center for Diabetes Education, Mayo Clinic
USA Translation and Computer Rochester, MN
Technology USA
Rami Almokayyad Minneapolis, MN
Division of Endocrinology, Department USA
Vivian Fonseca
of Medicine Section of Endocrinology
University of Minnesota Medical School Recie Davern Tulane University School of Medicine
University of Minnesota Department of Medicine and New Orleans, LA
Minneapolis, MN Endocrinology/Diabetes Mellitus USA
USA University College Hospital Galway and
National University of Ireland Robert L. Frye
Konstantin Astafurov Galway Department of Cardiovascular Disease
NJRetina/Rutgers Robert Wood Johnson Ireland Mayo Clinic
Medical School Rochester, MN
New Brunswick, NJ USA
Sean F. Dinneen
USA
Discipline of Medicine
NUI Galway Praveena Gandikota
Andrew J. Barkmeier Galway Endocrine, Diabetes and Nutrition Division
Department of Ophthalmology Ireland Department of Medicine
Mayo Clinic and St Luke’s Roosevelt Hospital
Rochester, MN Centre for Diabetes, Endocrinology and New York, NY
USA Metabolism USA
Galway University Hospitals
Galway
Michael Camilleri Kristen Gonzales
Ireland
Clinical Enteric Neuroscience Translational Department of Internal Medicine
and Epidemiological Research Program, Division of Endocrinology, Diabetes, and
Division of Gastroenterology and Fidelma P. Dunne Metabolism
Hepatology Galway Diabetes Research Centre University of New Mexico Health Sciences
Mayo Clinic National University of Ireland Center
Rochester, MN Galway Albuquerque, NM
USA Ireland USA
vii
viii List of Contributors
Kristin Grdinovac Blandine Laferrère Silvia Pieralice

University of Kansas Division of Endocrinology, Diabetes Department of Endocrinology & Diabetes
Kansas City, KS and Nutrition Obesity Research Center Campus Bio-Medico University of Rome
USA Department of Medicine Rome
St Luke’s Roosevelt Hospital Center Italy
Columbia University College of Physicians
Tomás P. Griffin
and Surgeons Paolo Pozzilli
Discipline of Medicine
New York, NY Department of Endocrinology & Diabetes
NUI Galway
USA Campus Bio-Medico University of Rome
Galway
Rome
Ireland
Italy
and Chiam Leker Locker
Centre for Diabetes, Endocrinology and Department of Cardiovascular Disease
Philip Raskin
Metabolism Mayo Clinic
Clifton and Betsy Robinson Chair in
Galway University Hospitals Rochester, MN
Biomedical Research
Galway USA
University of Texas Southwestern Medical
Ireland
Center at Dallas
Madeline K. Mahowald Dallas, TX
Chase Dean Hendrickson, Department of Cardiovascular Disease USA
Division of Diabetes, Endocrinology, and Mayo Clinic
Metabolism Rochester, MN John W. Richard III
Vanderbilt University Medical Center USA Division of Endocrinology, Diabetes,
Nashville, TN Nutrition and Metabolism
USA John M. Miles University of Texas Southwestern Medical
University of Kansas Center at Dallas
Kansas City, KS Dallas, TX
Shubhada Jagasia USA
Division of Diabetes, Endocrinology and USA
Metabolism Kalpana Muthusamy
Vanderbilt University Medical Center Shafaq R. Rizvi
Division of Endocrinology
Nashville, TN Division of Endocrinology, Diabetes,
Olmsted Medical Center
USA Metabolism and Nutrition
Rochester, MN
Mayo Clinic
USA
Rochester, MN
Ravinder Jeet Kaur USA
Division of Endocrinology, Diabetes,
Timothy O’Brien
Metabolism and Nutrition
Department of Medicine and Meera Shah
Mayo Clinic
Endocrinology/Diabetes Mellitus Division of Endocrinology, Diabetes,
Rochester, MN
University College Hospital Galway and Metabolism, and Nutrition
USA
National University of Ireland Mayo Clinic
Galway Rochester, MN
Jacob Kohlenberg Ireland USA
Division of Endocrinology, Diabetes,
Metabolism, and Nutrition Aonghus O’Loughlin Vinaya Simha
Mayo Clinic Discipline of Medicine Division of Endocrinology, Diabetes,
Rochester, MN NUI Galway Metabolism and nutrition
USA Galway Mayo Clinic
Ireland Rochester, MN
Yogish C. Kudva and USA
Division of Endocrinology, Diabetes, Centre for Diabetes, Endocrinology and
Metabolism and Nutrition Metabolism Mandeep Singh
Mayo Clinic Galway University Hospitals Department of Cardiovascular Disease
Rochester, MN Galway Mayo Clinic
USA Ireland Rochester, MN
USA
List of Contributors ix
Steve A. Smith Adrian Vella Alexander J. Williams

Division of Endocrinology, Diabetes, Division of Endocrinology, Diabetes, Division of Diabetes, Endocrinology, and
Metabolism, and Nutrition Metabolism, and Nutrition Metabolism
Mayo Clinic Mayo Clinic Vanderbilt University Medical Center
Rochester, MN Rochester, MN Nashville, TN
USA USA USA
Saritha Tirumalasetty
Section of Endocrinology
Tulane University School of
Medicine
New Orleans, LA
USA
Preface
While a week may be a long time in politics, the decade that round hole of an algorithm or guideline. There has been
has elapsed since the original publication of Clinical an expansion in the areas covered, reflecting my initial
Dilemmas in Diabetes could be considered to be a very desire for a book that reflects the growing facets of meta-
(very) long time given the developments in clinical practice bolic care.
that occurred. New therapeutics for the treatment of dys- I would like to express my gratitude to Dr. Robert A.
lipidemia and diabetes have appeared. Progress in the Rizza MD, who served as a co-editor for the first edition of
development of an artificial pancreas and in sensor tech- the book and as always serves as a sounding board for
nology has improved the care of type 1 diabetes and some many of my ideas. None of this would have been possible
diabetes medications may directly alter cardiovascular out- without him. Of course, the other important characters in
comes. However, the original motivations behind the this saga would be my parents—especially my father, who
book—translating trial data into clinical practice—remain has now retired from General Practice— but whose
very relevant. The astute clinician will need to balance risks approach to problems I have always tried to emulate. Last,
and benefits, convenience and cost with changing clinical but not least, I am very grateful to my wife Elsa, who hung
needs, comorbidities, and patients’ social structures. on gamely while I edited proofs of the first edition during
As before, the topics covered deserve discussion and our honeymoon, and to our children Katie and Lucy, who
debate to avoid making every square peg fit into the are always along for the ride.
x
PART I
Prediabetes and the Diagnosis
of Diabetes
1 “Is Prediabetes a Risk Factor or Is It
a Disease?”
Jacob Kohlenberg1 and Adrian Vella2
1
Fellow and Instructor in Medicine, Division of Endocrinology, Diabetes, Metabolism, and Nutrition, Mayo Clinic,
Rochester, MN, USA
2
Professor of Medicine, Division of Endocrinology, Diabetes, Metabolism, and Nutrition, Mayo Clinic, Rochester,
MN, USA
LE A R NI NG P OI NT S between 140–199 mg/dL [2]. Unlike the ADA, the WHO

does not include HbA1c as a diagnostic criterion for
⚫⚫ Prediabetes is a heterogeneous condition with variable
prediabetes.
risk of progression to type 2 diabetes The definitions of both prediabetes and DM have
⚫⚫ In addition to diabetes risk, it is associated with an evolved in recent decades. The WHO first defined the
increased risk of vascular disease. “borderline state” in 1965 as a 2‐h PG during a 50 or 100 g
To date, lifestyle modification is the single most
⚫⚫
OGTT between 110–129 mg/dL [3]. The ADA has long
important tool for altering the natural history of
prediabetes and progression to type 2 diabetes.
recognized IGT, and its definition has undergone little
change since its inception. First adopted by the ADA in
1997 and WHO in 1999, the term IFG was originally
defined as FPG 110–125 mg/dL [4]. However, in 2003, the
ADA revised the criteria for IFG to 100–125 mg/dL based
What is prediabetes?
on data from multiple studies showing that the risk of DM
Prediabetes is defined as an elevated fasting plasma glu increases markedly at a FPG concentration > 100 mg/dL
cose (FPG), and/or an elevated 2‐hour plasma glucose [5]. In 2010, the ADA added HbA1c as a diagnostic crite
(2‐h PG) during a 75‐gram (g) oral glucose tolerance test rion for prediabetes because the relationship between
(OGTT), and/or an elevated Hemoglobin A1c (HbA1c), HbA1c and the risk of retinopathy was similar to corre
without meeting diagnostic criteria for overt diabetes sponding FPG and 2‐h PG thresholds [6].
mellitus (DM) [1]. The 2020 American Diabetes
Association (ADA) Guidelines define prediabetes as
Rationale for the diagnostic criteria
impaired fasting glucose (IFG) with a FPG of 100–125
for diabetes mellitus and prediabetes
mg/dL, and/or impaired glucose tolerance (IGT) with a
2‐h PG during a 75‐g OGTT of 140–199 mg/dL, and/or a Individuals with prediabetes have abnormal glucose regu
HbA1c of 5.7–6.4% [1]. In contrast to the ADA, the 2016 lation and increased risk for developing DM type 2 (DM2)
World Health Organization (WHO) Guidelines define and its complications [6]. The diagnostic thresholds for
intermediate hyperglycemia as IFG between 110–125 mg/ DM are based on [1] the bimodal distribution of FPG and
dL and/or IGT with a 2‐h PG during a 75‐g OGTT 2‐h PG during an OGTT and [2] the glycemic thresholds
Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

© 2022 John Wiley & Sons Ltd. Published 2022 by John Wiley & Sons Ltd.
3
4 Prediabetes and the Diagnosis of Diabetes
Bimodal distribution of two-hour glucose levels

30
25
20
Percent
15
10
0
70 100 130 160 190 220 250 280 310 340 370 400 430
Two-hour plasma glucose mg/dL
FIG 1.1 Histogram with superimposed composite and component curves to describe the distribution of two‐hour plasma glucose
levels following an oral glucose load. Glucose concentrations and frequencies were arbitrarily chosen to illustrate a bimodal
distribution. The bimodal glucose distribution can be used to separate individuals into two groups, those with normoglycemia and
those with hyperglycemia. Intermediate glucose concentrations between normoglycemia and hyperglycemia helped define the
diagnostic thresholds for prediabetes.
for the development of microvascular complications, spe Epidemiology of prediabetes

cifically retinopathy (Figure 1.1). The bimodal distribution
The 2020 National Diabetes Statistics Report published by
and glycemic thresholds for the development of microvas
the Centers for Disease Control and Prevention estimated
cular complications have been demonstrated in many pop
the prevalence of prediabetes (defined by 2020 ADA crite
ulations including the Pima, Nauruans, South Africans,
Americans, Chinese, and Egyptians [7–15]. The bimodal ria) to be 38.0% (95% confidence interval (CI) 35.2–40.8)
distribution of glucose has been used to separate individu among adults in the United States (U.S.) [1, 16]. Overall,
als into two groups: those with normoglycemia and those the prevalence of prediabetes has not changed significantly
with hyperglycemia. IFG and IGT were defined as interme from 2005–2016. However, the number of U.S. adults who
diates between normoglycemia and hyperglycemia. The are aware that they have prediabetes increased from 6.5%
diagnostic thresholds for FPG, 2‐h PG during a 75‐g (95% CI 5.3–7.9) in 2005–2008 to 13.3% (95% CI 11.0–16.0)
OGTT, and HbA1c are relatively concordant in dis in 2013–2016. Further, the prevalence of prediabetes
criminating between the two components of a bimodal increases with age: 29.1% (95% CI 25.2–33.3) of adults 18–44
frequency distribution and their associations with years of age; 46.3% (95% CI 43.5–49.1) of adults 45–64 years
microvascular complications [4]. It is important to of age; and 51.0% (95% CI 46.5–55.5) of adults ≥ 65 years of
note that defining a lower limit of an intermediate cat age. Prediabetes is also more common in men, whose preva
egory of FPG, 2‐h PG during a 75‐g OGTT, and HbA1c lence is 42.3% (95% CI 38.1–46.5), compared to women,
is somewhat arbitrary because the risk of developing whose prevalence is 33.7% (95% CI 30.7–36.8). The preva
DM is a continuum that extends into the normoglyce lence of prediabetes is similar among racial/ethnic groups
mic range [6]. and among individuals of different education levels [16].
“Is Prediabetes a Risk Factor or Is It a Disease?” 5
Pathogenesis of impaired fasting glucose Using human pancreatic tissue obtained at autopsy,
and impaired glucose tolerance obese individuals with IFG have a 40% deficit in relative
beta‐cell volume compared to obese individuals with NFG
In epidemiologic studies, isolated IGT consistently has a
[30]. Additionally, the diabetes‐associated genetic varia
higher prevalence than isolated IFG [17]. The prevalence
tion in TCF7L2 is associated with impaired insulin secre
of IFG and IGT increases with age [17]. In adults less than
tion [31–33]. The inability of insulin secretion to
55 years of age, IGT is more common in women and IFG is
compensate for a decline in insulin action results in
more common in men [17]. This suggests that these two
hyperglycemia.
states have different pathophysiologic mechanisms.
It is also essential to consider the role of hepatic extrac
Genetics and lifestyle influence the pathogenesis of DM
tion in the pathophysiology of prediabetes and DM [26].
[18, 19]. Although more than 400 genetic signals have been
The peripheral insulin concentration reflects portal insulin
identified as influencing risk for DM2, single polymor
secretion, hepatic extraction, distribution, and degradation
phisms add only small degrees of risk [20]. Polymorphisms
[34]. Hepatic extraction of secreted insulin has a direct
in the Transcription factor 7‐like 2 (TCF7L2) locus have
relationship with disposition index. Reduced hepatic
the largest‐known effect on risk for DM [20, 21]. Compared
extraction in prediabetes and DM may be a compensatory
with non‐carriers, heterozygous and homozygous carriers
measure to increase circulating insulin. Interestingly, insu
of the at‐risk TCF7L2 variants have relative risks of devel
lin pulse characteristics influence hepatic extraction of
oping DM of 1.45 and 2.41, respectively [20, 21].
insulin, and the diabetes‐associated genetic variation in
The relationship between genetics and lifestyle on the
TCF7L2 is associated with abnormal insulin pulse charac
pathogenesis of DM was emphasized by a study comparing
teristics [33].
the prevalence of DM2 in the Pima population in Mexico
Furthermore, lack of glucagon suppression contributes
versus the Pima population in the U.S. [22]. The prevalence
to hyperglycemia in subjects with impaired insulin secre
of DM2 in the Pima population in Mexico was 6.9%, com
tion [24]. The diabetes‐associated genetic variation in
pared to 38% in the Pima population in the U.S. The preva
TCF7L2 is associated with impaired glucagon suppression
lence of obesity in the Pima population in Mexico was
[31, 35]. In a longitudinal study, defects in α‐cell function
significantly lower than that in the Pima population in the
with elevated fasting glucagon concentrations were associ
U.S. By comparison, the latter group also had significantly
ated with a subsequent decline in β‐cell function [36].
lower physical activity levels.
Overall, islet cell function is the primary regulator of
Screening recommendations
glucose metabolism, but multiple additional factors con
for prediabetes
tribute. Postprandial and fasting glucose concentrations
are determined by insulin secretion, hepatic extraction, The 2020 ADA Guidelines recommend screening asymp
insulin action, glucagon suppression, glucose effectiveness, tomatic overweight or obese adults for DM or prediabetes
and the rate of gastric emptying (Figure 1.2) [23–29]. if at least one of the following risk factors is present: first‐
Studies using the minimal model to quantitate insulin degree relative with DM; high‐risk race/ethnicity (e.g.
secretion and insulin action demonstrated that both indi African American, Latino, Native American, Asian
ces are lower in subjects with NFG/IGT and IFG/IGT than American, Pacific Islander); history of cardiovascular dis
in subjects with NFG/NGT [25, 26]. However, there was no ease; hypertension; high‐density lipoprotein cholesterol <
significant difference in insulin secretion and insulin 35 mg/dL; triglyceride level > 250 mg/dL; presence of poly
action between subjects with IFG/NGT and those with cystic ovary syndrome; physical inactivity; or other clinical
NFG/NGT. This implies that insulin secretion declines in considerations associated with reduced insulin action such
concert with insulin action across the spectrum of predia as acanthosis nigricans [1]. The 2020 ADA Guidelines fur
betes; however, subjects with isolated IFG may have an ther recommend that patients with prediabetes be tested
altered glucose threshold for insulin secretion without yearly and that women with a history of gestational diabetes
other intrinsic defects in β‐cell function. be tested lifelong at least every three years. For all other
Peripheral Systematic appearance of portal

tissue insulin/peripheral insulin action
Glucose
Stomach
effectiveness Liver &
splanchnic
Glucose tissues Hepatic insulin
Absorption effectiveness extraction/hepatic
Glucose insulin action
EGP Portal insulin secretion
Glucagon
suppression
Glucagon
suppression
Pancreas
FIG 1.2 Postprandial and fasting glucose concentrations are determined by insulin secretion, hepatic insulin extraction, insulin
action, glucagon suppression, glucose effectiveness, endogenous glucose production, and the rate of gastric emptying [23–29].
EGP = Endogenous Glucose Production.
patients, screening should begin at age 45, and if results are prediabetes and DM [39]. An analysis of NHANES data
normal, should be repeated at least every three years. from 2005–2010 examined adults without self‐reported
DM at baseline with available measurements for HbA1c,
Reproducibility, sensitivity, and specificity FPG, and 2‐h PG [40]. Prediabetes and DM were defined
of glycemic measurements and the role according to the current ADA Guidelines [1]. Using HbA1c
of oral glucose tolerance testing in clinical thresholds of ≥ 6.5% for DM and ≥ 5.7% for prediabetes
practice resulted in low sensitivity (24.9% and 35.4%, respectively)
A study that analyzed data from the Second Examination and high specificity in identifying patients diagnosed with
of the Third National Health and Nutrition Examination FPG and 2‐h PG. When FPG and HbA1c were used
Survey (NHANES) found that the within‐person coeffi together for diagnosis, the false‐negative rate was 45.7% for
cient of variation was 16.7% (95% CI 15.0–18.3) for 2‐h DM and 9.2% for prediabetes.
PG, 5.7% (95% CI 5.3–6.1) for FPG, and only 3.6% (95% CI Similar findings were reported in the Insulin Resistance
3.2–4.0) for HbA1c [37]. A study of non‐diabetic adults Atherosclerosis Study (IRAS), which examined 855 sub
reported a reproducibility rate of 65.6% for a 75‐g OGTT jects and defined DM and prediabetes in accordance with
repeated twice over a six‐week period [38]. the 2020 ADA Guidelines (Tables 1.1 and 1.2) [1, 41].
HbA1c, FPG, and 2‐h PG during a 75‐g OGTT have HbA1c ≥ 6.5%, FPG ≥ 126 mg/dL, and 2‐h PG ≥ 200 mg/dL
different sensitivities and specificities in the diagnosis of identified 32.3%, 44.8%, and 86.8% of individuals with
TABLE 1.1 The sensitivity of various indices of glycemia used that the OGTT continues to have a role in clinical practice.
to diagnose type 2 diabetes mellitus in the Insulin Resistance An individual eating three daily meals will be in a post
Atherosclerosis Study. Diabetes mellitus was defined according
prandial state for 6–9 hours per day [42]. Therefore, it is
to the 2020 American Diabetes Association criteria. This data
illustrates that omitting an oral glucose tolerance test leads logical that the knowledge gained from a standardized glu
to under‐diagnosis of diabetes mellitus, even when both fasting cose load is clinically informative. The 2020 ADA
plasma glucose and hemoglobin A1c are measured concur- Guidelines state that FPG, 2‐h PG during a 75‐g OGTT,
rently. HbA1c = Hemoglobin A1c. FPG = Fasting Plasma and HbA1c are equally appropriate to test for prediabetes
Glucose. 2‐h PG = 2‐hour Plasma Glucose [41].
and DM [1]. Many clinicians screen with both a FPG and
Sensitivity HbA1c. However, the omission of an OGTT leads to a sig
nificant under‐diagnosis of both DM and prediabetes.
HbA1c ≥ 6.5% 32.3%
FPG ≥ 126 mg/dL 44.8%
2‐h PG ≥ 200 mg/dL 86.8% Risk of progression from prediabetes
HbA1c ≥ 6.5% and/or FPG ≥ 126 mg/dL 52.2% to diabetes mellitus
FPG ≥ 126 mg/dL and/or 2‐h PG ≥ 200 mg/dL 97.1%
The transition from prediabetes to DM2 is variable and
influenced by heredity and lifestyle [18, 19, 43]. Multiple
TABLE 1.2 The sensitivity of various indices of glycemia used variables have been used to estimate the risk of progression
to diagnose prediabetes in the Insulin Resistance Atherosclerosis to DM2, including glycemic indices, anthropometric data,
Study. Prediabetes was defined according to the 2020 comorbidities, metabolomics, and genetic variants.
American Diabetes Association criteria. This data illustrates that
Numerous prediction models have been created based on
omitting an oral glucose tolerance test leads to under‐diagnosis
of prediabetes, even when both fasting plasma glucose
these variables [43–51].
and hemoglobin A1c are measured concurrently. HbA1c = Baseline FPG is a significant predictor of an individual’s
Hemoglobin A1c. FPG = Fasting Plasma Glucose. 2‐h PG = 2‐hour risk for developing DM [43]. In nondiabetic adults residing
Plasma Glucose [41]. in Minnesota, baseline FPG levels < 100 mg/dL, 100–109
mg/dL, and 110–125 mg/dL were associated with a 7%,
Sensitivity
19%, and 39% risk, respectively, with progression to DM
HbA1c 5.7–6.4% 23.6% over a median of 9 years. A discrete gradient of risk for
FPG 100–125 mg/dL 69.1% progression to DM was also observed among subjects with
2‐h PG 140–199 mg/dL 59.5%
a baseline FPG < 100 mg/dL.
HbA1c 5.7–6.4% and/or FPG 100–125 mg/dL 75.6%
FPG 100–125 mg/dL and/or 2‐h PG 140–199 95.8%
During a 75‐g OGTT, 30‐minute, 60‐minute, and 120‐
mg/dL minute PG concentrations were all significant predictors
for future risk of DM [49, 52]. One‐hour plasma glucose
during a 75‐g OGTT has been shown to be a better pre
DM, respectively. The combination of HbA1c ≥ 6.5% and/ dictor of future DM than FPG, HbA1c, and PG at 30 min
or FPG ≥ 126 mg/dL detected 52.2% of all individuals with ute and 120 minutes during a 75‐g OGTT [50]. In one
DM. HbA1c between 5.7–6.4%, FPG between 100–125 study, the hazard ratio for the development of DM was 9.5
mg/dL, and 2‐h PG between 140–199 mg/dL identified (7.90–11.43) for those with combined IFG‐IGT at base
23.6%, 69.1%, and 59.5% of all non‐diabetic subjects with line, 4.5 (4.03–5.02) for those with isolated IGT at base
prediabetes. The combination of HbA1c 5.7–6.4% and/or line, and 3.98 (3.16–5.02) for those with isolated IFG at
FPG 100–125 mg/dL detected 75.6% of all non‐diabetic baseline [45].
subjects with prediabetes. Baseline HbA1c is also a strong predictor for risk of
Despite the limitations of the OGTT – including lower future DM, and the risk of incident DM significantly
reproducibility and reduced patient convenience compared increased across the HbA1c range of 5.0–6.5% [46]. A sys
to FPG and HbA1c – the aforementioned studies suggest tematic review found that the 5‐year incidence of DM was
25–50%, 9–25%, and 5–9% among those with a baseline Pooled data analysis of nine studies from five countries
HbA1c of 6.0–6.5%, 5.5–6.0%, and 5.0–5.5%, respectively. examined glycemic thresholds for diabetes‐specific retin
Numerous studies have shown that demographic data, opathy (defined as moderate or more severe retinopathy)
anthropometric data, and comorbidities are associated in 44 623 subjects [57]. A curvilinear relationship was
with progression to DM. Factors positively associated with found to exist for FPG and HbA1c when diabetic retinopa
progression to DM include: family history of diabetes, for thy was plotted against continuous glycemic measures.
mer or active smoking, higher BMI, abdominal obesity, Diabetes‐specific retinopathy began to increase from a
increased waist circumference, hypertension, elevated tri FPG of 6.0–6.4 mmol/L and from a HbA1c of 6.0–6.4%.
glycerides, low HDL cholesterol, and elevated high sensi Based on vigintile distributions, glycemic thresholds for
tivity C‐reactive protein [45, 53, 54]. Increasing age has diabetes‐specific retinopathy were observed over the range
also been shown to be a predictor of DM in some studies. of 6.4–6.8 mmol/L for FPG, 9.8–10.6 mmol/L for 2‐h PG,
Metabolomics profiles have been investigated as a tool and 6.3–6.7% for HbA1C. Compared with the first vigin
to estimate the risk of developing DM. Numerous metabo tile, the odds ratios for diabetes‐specific retinopathy for the
lites have been found to be positively and negatively associ above vigintile distributions were 2.5 (95% CI 1.2–5.2) for
ated with progression to DM [51]. More than 400 distinct FPG, 10.1 (95% CI 1.3–79.4) for 2 h PG, and 4.5 (95% CI
genetic signals that affect the risk of developing DM2 have 1.4–15.2) for HbA1c.
been identified [20, 44]. Polygenic scores have been used to A major limitation of many studies that examine the
estimate the combined genetic risk for the development prevalence of retinopathy is the diagnostic criteria used to
of DM. define diabetes‐specific retinopathy [57]. The use of an
overly broad definition for retinopathy leads to the inclu
sion of subjects with mild retinopathy, which may have eti
Microvascular complications associated
ologies other than hyperglycemia. Overall, the rate of
with prediabetes
retinopathy increases with the degree and duration of
Retinopathy hyperglycemia [56, 57]. Subjects with prediabetes have a
In a large cohort of Pima individuals, the frequency of higher prevalence of retinopathy than subjects with NFG/
retinopathy – defined as the presence of at least one hem NGT, although the prevalence remains relatively low. In
orrhage, one microaneurysm, or proliferative retinopa many populations, the glycemic threshold for retinopathy
thy – was directly related to baseline FPG and 2‐h PG [55]. occurs in the prediabetic range of dysglycemia [57, 58].
Beginning at a baseline FPG threshold of 6.0–6.4 mmol/L
and 2‐h PG threshold of 9.0–10.6 mmol/L, there was a sig Nephropathy
nificant increase in period prevalence (10‐year interval) of Analysis of NHANES data from 1999–2006 revealed that
retinopathy. Additional data from the Pima individuals chronic kidney disease (CKD), defined as either reduced
illustrated that beyond a HbA1c threshold of 6.2% there kidney function or elevated albuminuria (urinary albu
was a significant increase in the prevalence of retinopathy min‐creatinine ratio ≥ 30 mg/g), was present in 17.1% of
[12]. Similar glycemic thresholds for an increase in the prediabetics compared to 11.8% of individuals with nor
prevalence of retinopathy were observed in a cross‐sec moglycemia [59]. A multiethnic study found that 16.1% of
tional study of Egyptians and data from the Third National subjects with IGT had microalbuminuria, compared to 4%
Health and Nutrition Examination Survey [13, 14]. of subjects with NGT [60]. After adjusting for multiple
Additionally, in the Diabetes Prevention Program (DPP), variables, glycemic status was found to be the most signifi
diabetic retinopathy was detected in 7.9% of the subjects cant determinant of urinary albumin concentration. A sys
with IGT and in 12.6% of the subjects who developed DM tematic review and meta‐analysis including 9 studies with
[56]. The prevalence of retinopathy is significantly higher 185 452 subjects reported that prediabetes was modestly
in subjects with DM, but retinopathy can develop in sub associated with an increased risk of CKD [61]. After adjust
jects with prediabetic range dysglycemia. ing for established risk factors, the relative risk of CKD was
1.11 (95% CI 1.02–1.21) for subjects with FPG between developing CVD. It is possible that much of this risk is due
6.1–6.9 mmol/L. to the increased risk of ultimately progressing to DM.
The prevalence and five‐year incidence of nephropathy Approximately 25% of first myocardial infarctions (MIs)
increases as FPG, 2‐h PG, and HbA1c rise [12, 58]. Of note, are unrecognized, which are predictive of future major car
the association of glycemia with nephropathy is weaker diovascular events including death [69]. In a multi‐ethnic
than the association between glycemia and retinopathy. population‐based cohort study adjusted for cardiovascular
When plotting prevalence of microalbuminuria against risk factors, it was shown that subjects with IFG have a
FPG, 2‐h PG, and HbA1c, there is a visible inflection point higher prevalence of unrecognized MIs than those with
and subsequent increase in microalbuminuria prevalence NFG, with an odds ratio of 1.60 (95% CI 1.01–2.48).
beyond a FPG of 5.5 mmol/L, 2‐h PG of 5.5 mmol/L (and In the Whitehall Study, after 5–10 years of follow‐up,
again at 9.3 mmol/L), and HbA1c of 5.8%. In summary, survival by baseline glucose tolerance status diverged
multiple studies suggest that prediabetic‐range hyperglyce among the groups, and median survival differed by approx
mia is associated with higher rates of nephropathy. imately 4 years between the normoglycemic and glucose
intolerant groups [70]. Overall, all‐cause mortality, cardio
Neuropathy vascular mortality, and respiratory mortality were higher
Although neural dysfunction is associated with hypergly among participants with glucose intolerance. The hazard of
cemia, clinicians should be mindful that neurologic deficits coronary mortality rose beginning at a 2‐h PG of 83 mg/dL;
can be attributable to non‐glycemic causes in individuals however, the graded relationship diminished after adjust
with prediabetes and DM [62]. Subjects with isolated IGT ing for multiple variables including baseline CVD.
were shown to have subclinical neural dysfunction that was A systematic review and meta‐analysis incorporated 53
generally asymptomatic and characterized by small‐fiber prospective studies with 1 611 339 subjects who were fol
neuropathy and mild impairment of cardiovascular auto lowed for a median of 9.5 years for cardiovascular and
nomic function [63]. Erectile dysfunction has also been mortality outcomes [68]. IFG and IGT diagnostic criteria
shown to be independently associated with IFG, with an were in accordance with 2020 ADA guidelines [1].
odds ratio of 1.26 (95% confidence interval 1.08–1.46) Compared to individuals with normoglycemia, those with
[64]. Prediabetic‐range hyperglycemia is also associated IGT or IFG had an increased risk of composite CVD (rela
with chronic idiopathic axonal polyneuropathy (CIAP) tive risk (RR) 1.13 for IFG and 1.30 for IGT), coronary
[65]. In a study of 100 subjects with CIAP, 36 individuals heart disease (RR 1.10 for IFG and 1.20 for IGT), stroke
were found to have IFG, 3 had FPG ≥ 126 mg/dL, 38 had (RR 1.06 for IFG and 1.20 for IGT), and all‐cause mortality
IGT, and 24 had 2‐h PG ≥ 200 mg/dL. Overall, the preva (RR 1.13 for IFG and 1.32 for IGT). One limitation of this
lence of dysglycemia in this cohort was approximately 2‐ systematic review and meta‐analysis is that some of the
fold higher than in an age‐matched general population included studies did not adjust for progression to DM dur
group. ing the follow‐up period.
A separate meta‐analysis examined 102 prospective
studies with 698 782 subjects and showed that FPG was
Macrovascular complications
modestly and non‐linearly associated with vascular dis
and mortality associated
ease, with hazard ratios for coronary heart disease of 1.11
with prediabetes
for FPG of 5.6–6.09 mmol/L and 1.17 for FPG of 6.1–6.99
There is an increased prevalence of cardiovascular disease mmol/L [71]. Another meta‐analysis that included 97 pro
(CVD) in individuals with prediabetes, but this relation spective studies with 820 900 subjects calculated hazard
ship is confounded by common‐risk factors present in ratios for cause‐specific death according to baseline FPG
CVD and prediabetes [66–68]. However, after accounting [72]. After adjusting for multiple variables and excluding
for non‐glycemic cardiovascular risk factors, both IFG and subjects with known CVD at baseline, FPG was found to be
IGT are still associated with a modestly increased risk of nonlinearly related to risk of death. Compared with
subjects with NFG, subjects with IFG had hazard ratios of included nephropathy, retinopathy, and neuropathy – did
1.13 for cancer deaths, 1.17 for vascular deaths, and 1.12 not differ among the 3 treatment groups. However, for
for non‐cancer and non‐vascular deaths. women, intensive lifestyle intervention significantly
reduced aggregate microvascular disease at 15 years com
pared to metformin and compared to placebo. Additionally,
Management of prediabetes
for those subjects with a baseline BMI ≥ 35 kg/m2, the RR
The goals of prediabetes management include preserving for the development of aggregate microvascular disease
β‐cell function, delaying or preventing the onset of DM, was significantly lower in the intensive lifestyle interven
delaying or preventing the developing of microvascular tion group compared to the placebo group. Among partici
and macrovascular complications associated with hyper pants whose most recent HbA1c was ≥ 6.5%, the intensive
glycemia, and reducing the cost of diabetes care. lifestyle intervention group showed statistically significant
reductions in the aggregate microvascular outcome, retin
Diabetes prevention program opathy, and neuropathy compared with placebo and
The Diabetes Prevention Program (DPP) enrolled 3234 metformin.
nondiabetic adult subjects at 27 centers in the U.S. [18]. Other benefits of intensive lifestyle changes were seen in
Eligibility criteria included FPG of 95–125 mg/dL and a DPP subjects [76]. From baseline to year three after rand
2‐h PG during a 75‐g OGTT 140–199 mg/dL. Subjects omization, hypertension increased in the placebo and met
were assigned to one of three groups: (1) intensive lifestyle formin groups but decreased in the intensive lifestyle
modification (goal ≥ 7% weight loss of initial body weight group. From baseline to year three, dyslipidemia pro
and ≥ 150 minutes of moderate intensity physical activity/ gressed in all three groups but the progression was less in
week); (2) metformin 850 mg twice daily plus standard the intensive lifestyle group compared to the metformin
lifestyle recommendations; or (3) placebo plus standard and placebo groups. After a mean follow‐up of 3.2 years in
lifestyle recommendations. Standard lifestyle recommen the DPP, there were significant improvements in quality of
dations were provided in writing and through annual brief life measures for the intensive lifestyle group, but not for
individual sessions [73]. In contrast, the intensive lifestyle the other two groups [77]. From a payer perspective, 10
modification provided comprehensive instruction in a years after randomization in DPP, intensive lifestyle
structured 16‐lesson curriculum. changes were cost‐effective, and metformin was marginally
The original DPP results were published after an aver cost‐saving compared to placebo [78].
age follow‐up of 2.8 years. The estimated cumulative inci
dence of DM at three years was significantly different Da Qing study
among all groups: 28.9% in the placebo group; 21.7% in the In 1986, a population‐based survey identified subjects in
metformin group; and 14.4% in the intensive lifestyle Da Qing, China with IGT [79]. These subjects were then
group. Weight loss was the main predictor of reduced DM randomized into four groups: control group, diet only,
incidence, with a hazard ratio per five‐kilogram (kg) exercise only, and diet plus exercise. At six years post rand
weight loss of 0.42 (95% CI 0.35–0.51). Further, for every omization, the mean rate of DM was significantly higher in
one kg of weight loss, there was a 16% reduction in risk of the control group at 66%, compared to 47% in the diet
progression to DM [74]. group, 45% in the exercise group, and 44% in the diet plus
Following randomization, DPPOS followed subjects for exercise group.
15 years and metformin continued to be provided to the The original Da Qing participants were followed for up
group originally assigned to it [75]. Over 15 years of fol to 30 years after randomization to assess the effects of
low‐up, the cumulative incidence of DM was 62% in the intervention of DM incidence, microvascular and macro
placebo group, 56% in the metformin group, and 55% in vascular complications, and mortality [80]. Active inter
the intensive lifestyle group. At the end of DPPOS, the vention occurred for the first six years after randomization
aggregate prevalence of microvascular outcomes – which until 1992, after which subjects were informed of the study
results and asked to continue with normal medical care. No to DM, including pioglitazone, rosiglitazone, and troglita
specific interventions were offered after the initial six years, zone [84–86]. Additionally, combined hormone replace
and the three intervention groups were combined into one ment therapy in post‐menopausal women, glipizide,
group for analysis purposes. valsartan, orlistat, and acarbose also significantly reduced
Over the 30‐year follow‐up period, the intervention the progression to DM compared to placebo [87–92].
group had a median delay in DM onset by four years (NNT Interestingly, during the Study to Prevent Non‐Insulin
10) compared to the control group and a significantly lower Dependent Diabetes Mellitus (STOP‐NIDDM), acarbose
cumulative incidence of DM onset (HR 0.61) [80]. At 30 was associated with a 49% relative risk reduction in cardio
years, there were 26% fewer CVD events in the interven vascular events compared to placebo with a HR 0.51 (95%
tion group compared to the control group. The difference CI 0.28–0.95). However, the methodology of STOP‐
between the two groups continued to increase over time. NIDDM has been heavily criticized and the validity of the
At 30 years, the incidence of retinopathy was 40% lower results has been questioned [93]. Many pharmacologic
in the intervention group than in the control group, and therapies have also been studied and shown to be ineffec
incidence of nephropathy and neuropathy were numeri tive in preventing progression to DM, including vitamin D,
cally lower in the intervention group but not significantly nateglinide, glimepiride, and ramipril [94–97].
different [80]. The median delay of composite microvascu
lar disease outcome was 5.2 years in the intervention group Bariatric surgery for the prevention or delay
(NNT 10). Cardiovascular and all‐cause mortality were of diabetes mellitus
also significantly lower in the intervention group (25.6% A Swedish study enrolled a large cohort of adult nondia
and 35.2%, respectively) than in the control group (45.5% and betic subjects who had already chosen to undergo various
56.3%, respectively). The median delay in CVD death and bariatric surgeries including banding (19%), vertical
all‐cause mortality in the intervention group were 7.3 years banded gastroplasty (69%), or gastric bypass (12%), and
and 4.8 years, respectively, with NNT of 10 for both matched them with a nonrandomized control group [98].
outcomes. At the time of enrollment, 17.2% of the control group had
IFG, compared to 16% of the surgical group. After a median
Pharmacologic therapy for the prevention or follow‐up time of 10 years, bariatric surgery compared
delay of diabetes mellitus with usual care reduced progression to DM by 87% among
A systematic review that included 20 randomized con subjects with IFG. Another study reported that 98.6% of
trolled trials examined the efficacy of metformin for the subjects with prediabetic range dysglycemia had normal
prevention or delay of DM [81]. The overall conclusion was FPG following gastric bypass [99].
that for at‐risk subjects, metformin compared with placebo
reduced or delayed the risk of progression to DM. The inci Exercise and diet for the prevention or delay
dence of DM was not significantly different when compar of diabetes mellitus
ing metformin plus intensive diet and exercise and identical A systematic review and meta‐analysis included 28 studies
intensive diet and exercise alone. with 1 261 991 subjects and investigated the role of physical
In addition to metformin, other medications have activity in reducing progression to DM [100]. Overall, for
proven efficacy in reducing progression to DM. Once‐daily those who achieved 11.25 metabolic equivalent of task
subcutaneous liraglutide 3.0 mg as adjunct therapy to life (MET) hours/week (equivalent to 150 minutes/week of
style modifications reduced progression to DM compared moderate activity) there was a risk reduction of 26% for
to placebo after 160 weeks [82]. In adults with elevated DM. Further risk reductions occurred at greater MET
CVD risk and prediabetes or newly established DM, a hours/week. The Health Professionals Follow‐up Study
once‐daily insulin glargine injection reduced progression showed that engaging in weight training or aerobic exercise
to DM compared to placebo [83]. Several thiazolidinedi for at least 150 minutes/week was independently associated
ones also have proven efficacy in reducing the progression with a lower risk of DM of 34% and 52%, respectively
[101]. The greatest reduction in DM risk was seen in men misclassification and under‐diagnosis of DM and predia
who engaged in both aerobic exercise and weight training betes. The following question then arises: are there clinical
for at least 150 minutes/week, with a risk reduction of 59%. consequences to misclassifying diabetes as only prediabe
The Prevención con Dieta Mediterránea (PREDIMED) tes or prediabetes as normal glucose metabolism?
study enrolled older adult subjects at high risk for but with We conclude that the answer is yes. There is a clear asso
out baseline CVD and randomized participants to either a ciation of increasing microvascular complications with ris
Mediterranean diet supplemented with extra‐virgin olive ing glucose concentrations, and patients with prediabetes
oil, a Mediterranean diet supplemented with mixed nuts, have an increased prevalence of retinopathy, nephropathy,
or a control diet based on general advice to reduce fat and neuropathy compared to individuals with normal glu
intake [102]. Compared to the control group, the cose metabolism [13, 14, 55, 57, 59–61, 63, 104–106].
Mediterranean diet supplemented with extra‐virgin olive There is also an increased prevalence of CVD, cardiovascu
oil group had significantly lower progression to DM with a lar mortality, and all‐cause mortality in patients with pre
HR of 0.60 (95% CI 0.54–0.85). The beneficial effect seen diabetes compared to those with normal glucose
in this study was thought to be due to dietary composition metabolism [66–68, 70].
itself and not attributed to calorie restriction, increased Therefore, recognition and treatment of prediabetes is
physical activity, or weight loss. essential to minimize morbidity and mortality. Intensive
lifestyle changes emphasizing weight loss and physical
American diabetes association 2020 guideline activity have been shown to prevent or delay progression to
recommendations for the management DM, potentially decrease microvascular complications,
of prediabetes potentially reduce macrovascular complications, improve
The 2020 ADA guidelines recommend that patients with comorbidities, improve quality of life, reduce medical
prediabetes participate in an intensive behavioral lifestyle costs, and decrease mortality [18, 75–80]. In addition to
intervention program modeled on the DPP to achieve and lifestyle changes, numerous medications are also effective
maintain 7% weight loss and to achieve ≥ 150 minutes/ in preventing or delaying progression to DM including
week of moderate intensity physical activity (such as brisk metformin, glipizide, liraglutide, insulin glargine, several
walking) [103]. Additionally, the ADA recommends that a thiazolidinediones, orlistat, acarbose, valsartan, and estro
variety of eating patterns are acceptable for patients with gen/progestin [18, 82–85, 87–92, 107].
prediabetes including the Mediterranean diet and a low‐ The totality of available evidence suggests that prediabe
calorie, low‐fat diet. Regarding pharmacologic therapy, the tes is not only a risk factor, it is in fact a disease. Perhaps the
2020 ADA guidelines state that metformin for prevention term “prediabetes” should be changed to “early diabetes”
of DM should be considered in patients with prediabetes, and managed as such clinically.
especially those with BMI ≥ 35 kg/m2, those aged < 60
years, and women with prior GDM.
References
1. American Diabetes Association, 2. Classification and
Conclusion diagnosis of diabetes: standards of medical care in
diabetes – 2020. Diabetes Care. 2020;43(Suppl 1):S14–S31.
In summary, patients with prediabetes have abnormal glu
2. World Health Organization, Global Report on Diabetes. 2016.
cose regulation mediated by impaired insulin secretion and
3. Diabetes Mellitus. Report of a WHO expert committee.
reduced insulin action, as well as an increased risk of pro
World Health Organ Tech Rep Ser. 1965;310:1–44.
gression to DM [18, 25, 26]. Many clinicians screen for dia 4. Report of the Expert Committee on the Diagnosis and
betes and prediabetes using the combination of FPG and Classification of Diabetes Mellitus. Diabetes Care.
HbA1c. However, even when used together, the combined 1997;20(7):1183–1197.
sensitivity for diagnosing DM and prediabetes is poor [41]. 5. Genuth S et al. Follow‐up report on the diagnosis of diabetes
Therefore, omission of a 75‐g OGTT results in both mellitus. Diabetes Care. 2003;26(11):3160–3167.
6. American Diabetes Association, Standards of medical care 21. Grant SF et al. Variant of transcription factor 7‐like
in diabetes – 2010. Diabetes Care. 2010;33(Suppl 1):S11–61. 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet.
7. Bennett PH et al. Epidemiologic studies of diabetes in the 2006;38(3):320–323.
Pima Indians. Recent Prog Horm Res. 1976;32:333–376. 22. Schulz LO et al. Effects of traditional and western environ
8. Zimmet P and Whitehouse S. Bimodality of fasting and ments on prevalence of type 2 diabetes in Pima Indians in
two‐hour glucose tolerance distributions in a Micronesian Mexico and the U.S. Diabetes Care. 2006;29(8):1866–1871.
population. Diabetes. 1978;27(8):793–800. 23. Bergman RN, Phillips LS and Cobelli C. Physiologic evalua
9. Rosenthal M et al. Evidence of bimodality of two hour tion of factors controlling glucose tolerance in man: meas
plasma glucose concentrations in Mexican Americans: urement of insulin sensitivity and beta‐cell glucose sensitivity
results from the San Antonio Heart study. J Chronic Dis. from the response to intravenous glucose. J Clin Invest.
1985;38(1):5–16. 1981;68(6):1456–1467.
10. Omar MA et al. South African Indians show a high preva 24. Shah P et al. Lack of suppression of glucagon contributes to
lence of NIDDM and bimodality in plasma glucose distribu postprandial hyperglycemia in subjects with type 2 diabetes
tion patterns. Diabetes Care. 1994;17(1):70–73. mellitus. J Clin Endocrinol Metab. 2000;85(11):4053–4059.
11. Fan J et al. Bimodality of 2‐h plasma glucose distributions in 25. Bock G et al. Pathogenesis of pre‐diabetes: mechanisms of
whites: the Rancho Bernardo study. Diabetes Care. 2005;28(6): fasting and postprandial hyperglycemia in people with
1451–1456. impaired fasting glucose and/or impaired glucose tolerance.
12. McCance DR et al. Comparison of tests for glycated haemo Diabetes. 2006;55(12):3536–3549.
globin and fasting and two hour plasma glucose concentra 26. Sathananthan A et al. A concerted decline in insulin secre
tions as diagnostic methods for diabetes. BMJ. 1994;308(6940): tion and action occurs across the spectrum of fasting and
1323–1328. postchallenge glucose concentrations. Clin Endocrinol (Oxf).
13. Davidson MB et al. Relationship between fasting plasma 2012;76(2):212–219.
glucose and glycosylated hemoglobin: potential for false‐ 27. Basu A et al. Impaired basal glucose effectiveness in NIDDM:
positive diagnoses of type 2 diabetes using new diagnostic contribution of defects in glucose disappearance and pro
criteria. JAMA. 1999;281(13):1203–1210. duction, measured using an optimized minimal model inde
14. Engelgau MM et al. Comparison of fasting and 2‐hour glu pendent protocol. Diabetes. 1997;46(3):421–432.
cose and HbA1c levels for diagnosing diabetes. Diagnostic 28. Camilleri M. Peripheral mechanisms in appetite regulation.
criteria and performance revisited. Diabetes Care. 1997;20(5): Gastroenterology. 2015;148(6):1219–1233.
785–791. 29. Camilleri M and Shin A. Novel and validated approaches for
15. Zhang R et al. The association of retinopathy and plasma gastric emptying scintigraphy in patients with suspected
glucose and HbA1c: a validation of diabetes diagnostic crite gastroparesis. Dig Dis Sci. 2013;58(7):1813–1815.
ria in a Chinese population. J Diabetes Res. 2016;2016: 30. Butler AE et al. Beta‐cell deficit and increased beta‐cell
4034129. apoptosis in humans with type 2 diabetes. Diabetes. 2003;
16. Centers for Disease Control and Prevention. National 52(1):102–110.
Diabetes Statistics Report, 2020. 2020. 31. Shah M et al. TCF7L2 Genotype and alpha‐Cell Function in
17. Abdul‐Ghani MA, Tripathy D and DeFronzo RA. Humans Without Diabetes. Diabetes. 2016;65(2):371–380.
Contributions of beta‐cell dysfunction and insulin resist 32. Adams JD and Vella A. What can diabetes – associated genetic
ance to the pathogenesis of impaired glucose tolerance and variation in TCF7L2 teach us about the pathogenesis of Type 2
impaired fasting glucose. Diabetes Care. 2006;29(5): Diabetes? Metab Syndr Relat Disord. 2018;16(8):383–389.
1130–1139. 33. Laurenti MC et al. Diabetes‐associated genetic variation in
18. Knowler WC et al. Reduction in the incidence of type 2 dia TCF7L2 alters pulsatile insulin secretion in humans. JCI
betes with lifestyle intervention or metformin. N Engl J Med. Insight. 2020;5(7).
2002;346(6):393–403. 34. Meier JJ, Veldhuis JD, and Butler PC. Pulsatile insulin secretion
19. Talmud PJ et al. Sixty‐five common genetic variants and pre dictates systemic insulin delivery by regulating hepatic insulin
diction of type 2 diabetes. Diabetes. 2015;64(5):1830–1840. extraction in humans. Diabetes. 2005;54(6):1649–1656.
20. Mahajan A et al. Fine‐mapping type 2 diabetes loci to single‐ 35. Smushkin G et al. Diabetes‐associated common genetic varia
variant resolution using high‐density imputation and islet‐ tion and its association with GLP‐1 concentrations and response
specific epigenome maps. Nat Genet. 2018;50(11):1505–1513. to exogenous GLP‐1. Diabetes. 2012;61(5):1082–1089.
36. Adams JD et al. Fasting glucagon concentrations are associ 50. Peddinti G et al. 1‐hour post‐OGTT glucose improves the
ated with longitudinal decline of beta‐cell function in non‐ early prediction of type 2 diabetes by clinical and metabolic
diabetic humans. Metabolism. 2020;105:154175. markers. J Clin Endocrinol Metab. 2019;104(4):1131–1140.
37. Selvin E et al. Short‐term variability in measures of glycemia 51. Peddinti G et al. Early metabolic markers identify potential
and implications for the classification of diabetes. Arch Intern targets for the prevention of type 2 diabetes. Diabetologia.
Med. 2007;167(14):1545–1551. 2017;60(9):1740–1750.
38. Ko GT et al. The reproducibility and usefulness of the oral 52. Priya M et al. 1‐hour venous plasma glucose and incident
glucose tolerance test in screening for diabetes and other prediabetes and diabetes in Asian Indians. Diabetes Technol
cardiovascular risk factors. Ann Clin Biochem. 1998;35 Ther. 2013;15(6):497–502.
(Pt 1):62–67. 53. Wu S et al. Transition from pre‐diabetes to diabetes and pre
39. Meijnikman AS et al. Not performing an OGTT results in dictors of risk in Mexican‐Americans. Diabetes Metab Syndr
significant underdiagnosis of (pre)diabetes in a high risk Obes. 2017;10:491–503.
adult Caucasian population. Int J Obes (Lond). 2017;41(11): 54. Abdul‐Ghani MA et al. What is the best predictor of future
1615–1620. type 2 diabetes? Diabetes Care. 2007;30(6):1544–1548.
40. Guo F, Moellering DR, and Garvey WT. Use of HbA1c for 55. Gabir MM et al. Plasma glucose and prediction of microvas
diagnoses of diabetes and prediabetes: comparison with cular disease and mortality: evaluation of 1997 American
diagnoses based on fasting and 2‐hr glucose values and Diabetes Association and 1999 World Health Organization
effects of gender, race, and age. Metab Syndr Relat Disord. criteria for diagnosis of diabetes. Diabetes Care.
2014;12(5):258–268. 2000;23(8):1113–1118.
41. Lorenzo C et al. A1C between 5.7 and 6.4% as a marker for 56. Diabetes Prevention Program Research Group. The preva
identifying pre‐diabetes, insulin sensitivity and secretion, lence of retinopathy in impaired glucose tolerance and
and cardiovascular risk factors: the Insulin Resistance recent‐onset diabetes in the Diabetes Prevention Program.
Atherosclerosis Study (IRAS). Diabetes Care. 2010;33(9): Diabet Med. 2007;24(2):137–144.
2104–2109. 57. Colagiuri S et al. Glycemic thresholds for diabetes‐specific
42. Tuomilehto J. Point: a glucose tolerance test is important for retinopathy: implications for diagnostic criteria for diabetes.
clinical practice. Diabetes Care. 2002;25(10):1880–1882. Diabetes Care. 2011;34(1):145–150.
43. Dinneen SF et al. Effects of changing diagnostic criteria on 58. Tapp RJ et al. Diagnostic thresholds for diabetes: the associa
the risk of developing diabetes. Diabetes Care. 1998;21(9): tion of retinopathy and albuminuria with glycaemia.
1408–1413. Diabetes Res Clin Pract. 2006;73(3):315–321.
44. Udler MS et al. Genetic risk scores for diabetes diagnosis and 59. Plantinga LC et al. Prevalence of chronic kidney disease in
precision medicine. Endocr Rev. 2019;40(6):1500–1520. US adults with undiagnosed diabetes or prediabetes. Clin J
45. Han SJ et al. Incidence and predictors of type 2 diabetes Am Soc Nephrol. 2010;5(4):673–682.
among Koreans: a 12‐year follow up of the Korean Genome 60. Metcalf PA et al. Microalbuminuria in a middle‐aged work
and Epidemiology Study. Diabetes Res Clin Pract. 2017;123: force. Effect of hyperglycemia and ethnicity. Diabetes Care.
173–180. 1993;16(11):1485–1493.
46. Zhang X et al. A1C level and future risk of diabetes: a sys 61. Echouffo‐Tcheugui JB et al. Association between prediabe
tematic review. Diabetes Care. 2010;33(7):1665–1673. tes and risk of chronic kidney disease: a systematic review
47. Hulman A et al. Glucose patterns during an oral glucose tol and meta‐analysis. Diabet Med. 2016;33(12):1615–1624.
erance test and associations with future diabetes, cardiovas 62. Dyck PJ et al. The prevalence by staged severity of various
cular disease and all‐cause mortality rate. Diabetologia. types of diabetic neuropathy, retinopathy, and nephropathy
2018;61(1):101–107. in a population‐based cohort: the Rochester Diabetic
48. Abdul‐Ghani MA et al. One‐hour plasma glucose concen Neuropathy Study. Neurology. 1993;43(4):817–824.
tration and the metabolic syndrome identify subjects at high 63. Putz Z et al. Noninvasive evaluation of neural impairment in
risk for future type 2 diabetes. Diabetes Care. subjects with impaired glucose tolerance. Diabetes Care.
2008;31(8):1650–1655. 2009;32(1):181–183.
49. Abdul‐Ghani MA et al. Fasting versus postload plasma glucose 64. Grover SA et al. The prevalence of erectile dysfunction in the
concentration and the risk for future type 2 diabetes: results primary care setting: importance of risk factors for diabetes
from the Botnia Study. Diabetes Care. 2009;32(2):281–286. and vascular disease. Arch Intern Med. 2006;166(2):213–219.
65. Hoffman‐Snyder C et al. Value of the oral glucose tolerance 79. Pan XR et al. Effects of diet and exercise in preventing
test in the evaluation of chronic idiopathic axonal polyneu NIDDM in people with impaired glucose tolerance. The Da
ropathy. Arch Neurol. 2006;63(8):1075–1079. Qing IGT and Diabetes Study. Diabetes Care. 1997;
66. Nathan DM et al. Impaired fasting glucose and impaired 20(4):537–544.
glucose tolerance: implications for care. Diabetes Care. 2007; 80. Gong Q et al. Morbidity and mortality after lifestyle inter
30(3):753–759. vention for people with impaired glucose tolerance: 30‐year
67. Barr EL et al. Risk of cardiovascular and all‐cause mortality results of the Da Qing Diabetes Prevention Outcome Study.
in individuals with diabetes mellitus, impaired fasting glu Lancet Diabetes Endocrinol. 2019;7(6):452–461.
cose, and impaired glucose tolerance: the Australian 81. Madsen KS et al. Metformin for prevention or delay of type 2
Diabetes, Obesity, and Lifestyle Study (AusDiab). Circulation. diabetes mellitus and its associated complications in persons
2007;116(2):151–157. at increased risk for the development of type 2 diabetes mel
68. Huang Y et al. Association between prediabetes and risk of litus. Cochrane Database Syst Rev. 2019;12:CD008558.
cardiovascular disease and all cause mortality: systematic 82. le Roux CW et al. 3 years of liraglutide versus placebo for
review and meta‐analysis. BMJ. 2016;355:i5953. type 2 diabetes risk reduction and weight management in
69. Stacey RB et al. Prediabetes and the association with unrec individuals with prediabetes: a randomised, double‐blind
ognized myocardial infarction in the multi‐ethnic study of trial. Lancet. 2017;389(10077):1399–1409.
atherosclerosis. Am Heart J. 2015;170(5):923–928. 83. ORIGIN Trial Investigators. Basal insulin and cardiovascu
70. Brunner EJ et al. Relation between blood glucose and coro lar and other outcomes in dysglycemia. N Engl J Med.
nary mortality over 33 years in the Whitehall Study. Diabetes 2012;367(4):319–328.
Care. 2006;29(1):26–31. 84. DeFronzo RA et al. Pioglitazone for diabetes prevention in
71. Emerging Risk Factors Collaboration. Diabetes mellitus, impaired glucose tolerance. N Engl J Med. 2011;364(12):
fasting blood glucose concentration, and risk of vascular dis 1104–1115.
ease: a collaborative meta‐analysis of 102 prospective stud 85. The DREAM (Diabetes REduction Assessment with ramipril
ies. Lancet. 2010;375(9733):2215–2222. and rosiglitazone Medication) Trial Investigators. Effect of
72. Rao Kondapally Seshasai S et al. Diabetes mellitus, fasting rosiglitazone on the frequency of diabetes in patients with
glucose, and risk of cause‐specific death. N Engl J Med. impaired glucose tolerance or impaired fasting glucose: a ran
2011;364(9):829–841. domised controlled trial. Lancet. 2006;368(9541):1096–1105.
73. Diabetes Prevention Program Research Group. The Diabetes 86. Knowler WC et al. Prevention of type 2 diabetes with trogl
Prevention Program (DPP): description of lifestyle interven itazone in the Diabetes Prevention Program. Diabetes.
tion. Diabetes Care. 2002;25(12):2165–2171. 2005;54(4):1150–1156.
74. Hamman RF et al. Effect of weight loss with lifestyle interven 87. Torgerson JS et al. XENical in the prevention of diabetes in
tion on risk of diabetes. Diabetes Care. 2006;29(9):2102–2107. obese subjects (XENDOS) study: a randomized study of orl
75. Diabetes Prevention Program Research Group. Long‐term istat as an adjunct to lifestyle changes for the prevention of
effects of lifestyle intervention or metformin on diabetes type 2 diabetes in obese patients. Diabetes Care. 2004;27(1):
development and microvascular complications over 15‐year 155–161.
follow‐up: the Diabetes Prevention Program Outcomes 88. Chiasson JL et al. Acarbose for prevention of type 2 diabetes
Study. Lancet Diabetes Endocrinol. 2015;3(11):866–875. mellitus: the STOP‐NIDDM randomised trial. Lancet.
76. Ratner R et al. Impact of intensive lifestyle and metformin 2002;359(9323):2072–2077.
therapy on cardiovascular disease risk factors in the diabetes 89. Group NS et al. Effect of valsartan on the incidence of diabe
prevention program. Diabetes Care. 2005;28(4):888–894. tes and cardiovascular events. N Engl J Med. 2010;362(16):
77. Florez H et al. Impact of lifestyle intervention and met 1477–1490.
formin on health‐related quality of life: the diabetes preven 90. Eriksson JG et al. Long‐term beneficial effects of glipizide
tion program randomized trial. J Gen Intern Med. treatment on glucose tolerance in subjects with impaired
2012;27(12):1594–1601. glucose tolerance. J Intern Med. 2006;259(6):553–560.
78. Diabetes Prevention Program Research Group. The 10‐year 91. Kanaya AM et al. Glycemic effects of postmenopausal hor
cost‐effectiveness of lifestyle intervention or metformin for mone therapy: the Heart and Estrogen/progestin Replacement
diabetes prevention: an intent‐to‐treat analysis of the DPP/ Study. A randomized, double‐blind, placebo‐controlled trial.
DPPOS. Diabetes Care. 2012;35(4):723–730. Ann Intern Med. 2003;138(1):1–9.
92. Margolis KL et al. Effect of oestrogen plus progestin on the 100. Smith AD et al. Physical activity and incident type 2 diabe
incidence of diabetes in postmenopausal women: results tes mellitus: a systematic review and dose‐response meta‐
from the Women’s Health Initiative Hormone Trial. Dia analysis of prospective cohort studies. Diabetologia. 2016;
betologia. 2004;47(7):1175–1187. 59(12):2527–2545.
93. Sawicki PT and Kaiser T. Response to Chiasson et al.: 101. Grontved A et al. A prospective study of weight training
Acarbose for the prevention of Type 2 diabetes, hypertension and risk of type 2 diabetes mellitus in men. Arch Intern
and cardiovascular disease in subjects with impaired glucose Med. 2012;172(17):1306–1312.
tolerance: facts and interpretations concerning the critical 102. Salas‐Salvado J et al. Prevention of diabetes with Medi
analysis of the STOP‐NIDDM Trial data. Diabetologia. terranean diets: a subgroup analysis of a randomized trial.
2004;47(6):976–977. Ann Intern Med. 2014;160(1):1–10.
94. Lindblad U et al. Can sulphonylurea addition to lifestyle 103. American Diabetes Association. 3. Prevention or delay of
changes help to delay diabetes development in subjects type 2 diabetes: standards of medical care in diabe
with impaired fasting glucose? The Nepi ANti tes – 2020. Diabetes Care. 2020;43(Suppl 1):S32–S36.
diabetes StudY (NANSY). Diabetes Obes Metab. 2011;13(2): 104. Hu W et al. Association of elevated glycosylated hemoglobin
185–188. A1c with hyperfiltration in a middle‐aged and elderly Chinese
95. Pittas AG et al. Vitamin D supplementation and prevention population with prediabetes or newly diagnosed diabetes: a
of type 2 diabetes. N Engl J Med. 2019;381(6):520–530. cross‐sectional study. BMC Endocr Disord. 2015;15:47.
96. NAVIGATOR Study Group et al. Effect of nateglinide on the 105. Kannan MA et al. Prevalence of neuropathy in patients
incidence of diabetes and cardiovascular events. N Engl J with impaired glucose tolerance using various electrophys
Med. 2010;362(16):1463–1476. iological tests. Neurol India. 2014;62(6):656–661.
97. The DREAM Trial Investigators. Effect of ramipril on the 106. Wu JS et al. Epidemiological evidence of altered cardiac
incidence of diabetes. N Engl J Med. 2006;355(15): autonomic function in subjects with impaired glucose tol
1551–1562. erance but not isolated impaired fasting glucose. J Clin
98. Carlsson LM et al. Bariatric surgery and prevention of type 2 Endocrinol Metab. 2007;92(10):3885–3889.
diabetes in Swedish obese subjects. N Engl J Med. 107. Holman RR et al. Effects of acarbose on cardiovascular and
2012;367(8):695–704. diabetes outcomes in patients with coronary heart disease and
99. Pories WJ et al. Is type II diabetes mellitus (NIDDM) a impaired glucose tolerance (ACE): a randomised, double‐
surgical disease? Ann Surg. 1992;215(6):633–642; discus blind, placebo‐controlled trial. Lancet Diabetes Endocrinol.
sion 643. 2017;5(11):877–886.
2 Early Diagnosis of Type 1
Diabetes – Useful or a Pyrrhic Victory?
Paolo Pozzilli and Silvia Pieralice
Department of Endocrinology & Diabetes, Campus Bio-Medico University of Rome, Rome, Italy
T1D have been shown in countries having both high and

LE A R NI NG P OI NT S
low prevalence figures, with an indication of a steeper
●● Making an early diagnosis of type 1 diabetes can
increase in some of the low-prevalence countries. T1D
permit intervention with agents that could modify the accounts for about 10% of all cases of diabetes, occurs most
course of disease. commonly in people of European descent and affects
●● Numerous clinical trials have attempted to change the 2 million people in Europe and North America [1]. The
course of disease or prevent its onset. Of these, early lowest incidence has been found in Asia and Oceania, the
immune intervention holds the most promise.
highest in Europe.
As far as its pathogenesis is concerned, T1D results from
the autoimmune destruction of pancreatic insulin-
Introduction secreting β-cells. Genetic, metabolic, and environmental
factors act together to precipitate the onset of the disease.
Type 1 diabetes (T1D) is one of the most widespread
The excess mortality associated with diabetes-related com-
chronic disease occurring in children and young adults. In
plications and the increasing prevalence of the disease
2001, people with diabetes (all types included) were
among the youth, emphasize the importance of novel ther-
177 million worldwide, in 2010 285 million, in 2019
apeutic strategies to prevent or slow down the autoimmune
approximately 463 million and it is predicted that some
process.
700 million people worldwide will live with diabetes in
2045 [1]. Pathogenesis of T1D: An Update in View
As the prevalence of diabetes continues to rise world- of Defining Preventive Tools
wide, disease-related morbidity and mortality are emerg-
ing as major healthcare problems. Epidemiologic data There are three main categories of factors involved in the
show that diabetes-related long-term complications begin pathogenesis of T1D. These are genetic, immunological,
early in the natural history of the disease [2]–[4]. These and environmental factors (Figure 2.1).
findings indicate that early identification and management Like other organ-specific autoimmune diseases, T1D
of individuals at increased risk of T1D has the potential to shows specific human leukocyte antigen (HLA) associa-
reduce both the clinical onset of the disease and its related tions. The HLA complex on chromosome 6 comprises the
complications. first gene shown to be associated with the disease and it is
The global incidence of T1D in children and adolescents considered to contribute to almost half of the familial basis
is rising with an estimated overall annual increase of of T1D. Two combinations of HLA haplotypes are of par-
approximately 3%. These recent epidemiologic trends in ticular importance. They are DR4-DQ8 and DR3-DQ2,

17
Immune Environmental
dysregulation triggers and
regulators
IAA
GADA ICA512A ICA
Interactions
Loss of first phase
between
β-cell mass
insulin response
genes
(IVGTT)
imparting Variable insulitis
susceptibility β-cell sensitivity Glucose intolerance
and resistance to injury Absence of C-peptide
Overt diabetes
Pre-diabetes
Time
FIG 2.1 Pathogenesis and natural history of Type 1 diabetes. Atkinson MA and Eisenbarth GS Type 1 diabetes: new perspectives on
disease pathogenesis and treatment. Lancet 2001;358(9277):221–229.
which are present in 90% of children with T1D [5]. A third e mphasizes the potential importance of current therapeu-
haplotype, DR15-DQ6, is found in less than 1% of children tic strategies targeting this interaction [10].
with T1D, compared to more than 20% of the general pop- Genetic studies have highlighted the importance of
ulation, and it is considered to be protective. The genotype large, well-characterized populations in the identification
combining the two susceptibility haplotypes (DR4-DQ8/ of susceptibility genes for T1D. Recruitment of increas-
DR3-DQ2) contributes to the greatest risk of the disease ingly large populations of patients with T1D and their
and it appears frequently in children with an earlier onset. families is required to provide statistically powerful cohorts
First-degree relatives of these children are themselves at in which to identify other disease-associated genes. Some
greater risk of T1D compared to those of children in whom genes have a relatively minor individual impact on suscep-
the disease develops later. tibility to disease but could nevertheless provide more clues
Candidate gene studies also identified the insulin gene to future preventive therapies.
on chromosome 11 as another important genetic suscepti- The presence of autoantibodies to β-cells is the hallmark
bility factor, contributing 10% of the genetic susceptibility of T1D (Figure 2.1). Abnormal activation of the T-cell-
to T1D [6]. Similarly, an allele of the gene acting as a nega- mediated immune system in susceptible individuals leads
tive regulator of T-cell activation, cytotoxic T lymphocyte to an inflammatory response within the islets as well as to a
antigen 4 (CTLA-4), found on chromosome 2q33, is con- humoral response with production of antibodies to β-cell
sidered to be another susceptibility gene for T1D [7]. A antigens. Islet-cell antibodies (ICA) were the first ones
variant of PTPN22, the gene encoding Lymphoid described, followed by more specific autoantibodies to
Phosphatase (LYP), which is a suppressor of T-cell activa- insulin (IAA), glutamic acid decarboxylase (GAD), the pro-
tion, has been deemed as another susceptibility gene [8]. tein tyrosine phosphatase (IA-2), and most recently Zinc
Similarly, variation in IL2RA which encodes the α-chain of Transporter 8 (ZnT8A), all of which can be easily detected
the IL-2 receptor is also associated with T1D [9]. The by sensitive radioimmunoassay to identify subjects at risk
observation that these susceptibility genes for T1D all play of developing T1D [11]. These autoantibodies are common
important roles in antigen presentation to T- cells, in both childhood and adult onset T1D, with many subjects
Early Diagnosis of Type 1 Diabetes – Useful or a Pyrrhic Victory? 19
being positive for multiple autoantibodies. The type of tions, and severe systemic infections in newborn infants.
immune response is age-dependent, but seroconversion to Most infections, however, are subclinical or manifest with
multiple autoantibody positivity usually occurs tightly clus- mild respiratory symptoms. The primary replication of the
tered in time and is associated with genetic risk. virus occurs in the lymphoid tissues of the pharynx and
The presence of one or more type of antibodies can pre- small intestine, and during the following viremic phase the
cede the clinical onset of T1D by years or even decades. virus can spread to various organs including the β-cells.
These autoantibodies are usually persistent, although a Theoretically, Enterovirus could cause β-cell damage by
small group of individuals may revert to being seronegative two main mechanisms. They may infect β-cells and destroy
without progressing to clinical diabetes. The presence and them directly or they may induce an autoimmune response
persistence of positivity to multiple antibodies increases against β-cells. Direct virus-induced damage has been sup-
the likelihood of progression to clinical disease. ported by studies showing that Enterovirus are present in β-
On that note, recent evidence shows that antibodies spe- cells in patients who have died from severe systemic
cific to oxidative post- translationally modified insulin Enterovirus infection and that the islet-cells of these patients
(oxPTM-INS) are present in the majority of newly diag- are damaged. Enterovirus can also infect and damage β-cells
nosed individuals with T1D being significantly more abun- in vitro and induce the expression of interferon-alpha and
dant than autoantibodies to native insulin (NT-INS) [12]. HLA-class I molecules in β-cells thus mimicking the situa-
Furthermore, subsequent analysis found that oxPTM-INS tion observed in the pancreas of patients affected by T1D.
auto-reactivity is present before diabetes diagnosis in over The first reports connecting Enterovirus infections to T1D
90% of individuals, suggesting a potential role for oxPTM- were published more than 30 years ago, showing that the
INS-Ab as a predictive biomarker of T1D [13]. seasonal variation in the onset of T1D follows that of
The progressive reduction of insulin-secretory reserve Enterovirus infections. At the same time antibodies against
leads primarily to the loss of the first phase insulin secre- Coxsackievirus B serotypes were found to be more frequent
tion in response to an intravenous glucose tolerance test, in patients with newly diagnosed T1D than in control sub-
and therefore to a state of absolute insulin deficiency. jects [14]. Enterovirus have also been isolated from patients
Regarding the role of environmental factors, it should be with newly diagnosed T1D. In one case report Coxsackievirus
underlined that the increase in incidence of T1D is too B4 was isolated from the pancreas of a child who had died
rapid to be caused by alterations in the genetic background from diabetic ketoacidosis, and this virus caused diabetes
and is likely to be the result of environmental changes. when transferred to a susceptible mouse strain. The β-cells
of diabetic patients also express interferon-alpha, a cytokine
What Are the Environmental Factors Triggering that is induced during viral infections, suggesting the pres-
Type 1 Diabetes? ence of some virus in the β-cells. Prospective studies are par-
Certain viral infections may play a role in the pathogenesis ticularly valuable in the evaluation of viral triggers because
of human T1D. Congenital rubella is the classical example they cover all stages of the β-cell damaging process.
of virus-induced diabetes in human beings, but effective Enterovirus are not the only viruses that have been con-
immunization programs have eliminated congenital rubella nected to the pathogenesis of T1D. Mumps, measles, cyto-
in most Western countries. Currently, the main candidate megalovirus, and retroviruses also have been found to be
for a viral trigger of human diabetes is members of the associated with T1D, but the evidence is less convincing
group of Enterovirus [14]. They are small non-enveloped than that for Enterovirus.
RNA viruses, which belong to the Picornavirus family.
They consist of more than 60 different serotypes, with the The Role of Cow’s Milk
Polioviruses being their best- known representatives. There is evidence that cow’s milk proteins can act as trig-
Enterovirus infections are frequent among children and gers for the autoimmune process of β-cell destruction
adolescents causing aseptic meningitis, myocarditis, rash, based on studies indicating bottle feeding as a triggering
hand-food-and-mouth disease, paralysis, respiratory infec- factor for an autoimmune response to β-cell [15].
There are several arguments for the milk hypothesis in immune response towards β-cells. The TRIGR trial (see
T1D, including the following: section on primary prevention of T1D) investigated
whether the administration of a cow’s milk hydrolysate
●● Epidemiological studies show increased risk for T1D if
could prevent or delay the onset of T1D.
the breast-feeding period is short and cow’s milk is
introduced before 3–4 months of age.
The Role of Vitamin D Deficiency
●● Skim milk powder can be "diabetogenic" in diabetes-
Several epidemiological studies have described an intrigu-
prone BB rats.
ing correlation between geographical latitude and the inci-
●● Patients with T1D have increased levels of antibodies
dence of T1D and an inverse correlation between monthly
against cow’s milk constituents.
hours of sunshine and the incidence of diabetes. A seasonal
●● Milk albumin and β-casein have some structural similarity
pattern of disease onset has also been described for T1D,
to the islet autoantigen ICA69 and GLUT-2, respectively.
once again suggesting an inverse correlation between sun-
A number of hypotheses have been postulated to explain light and the disease [16]. Vitamin D is an obvious candi-
the pathogenic role of cow’s milk. One of the most convinc- date as a mediator of this sunshine effect.
ing ones is that immature gut mucosa allows the passage of Dietary vitamin D supplementation is often recom-
high molecular weight, potentially antigenic proteins mended in pregnant women and in children to prevent
which share some molecular mimicry with pancreatic β- vitamin D deficiency. Cod liver oil taken during the first
cells. Among diabetogenic proteins in cow’s milk, β-casein, year of life reportedly reduced the risk of childhood-onset
β-lactoglobulin, and albumin have been implicated as T1D and a multicenter case-control study also showed an
sources of potential antigens. association between vitamin D supplementation in infancy
Casein represents the major protein in cow’s milk. and a decreased risk of T1D. Two further meta-analyses of
Human and bovine β-casein are approximately 70% homol- retrospective studies showed that the risk of T1D was lower
ogous and 30% identical. There are several reasons why it is in children who were supplemented with calcitriol com-
thought that β-casein is a good candidate to explain the pared with those who were not supplemented [17].
observed association between cow’s milk consumption and Nonetheless, it remains to be determined whether these
T1D: (1) it has several structural differences from the observations are the result of supplementation of vitamin
homologous human protein; (2) casein is probably the milk D to supraphysiological levels or are simply the result of the
fraction promoting diabetes in the NOD mouse, since a prevention of vitamin D deficiency. On that note, other
protein-free diet prevents the disease while a diet contain- clinical studies reported no effect of vitamin D supplemen-
ing casein as the sole source of protein produces diabetes in tation on β-cell function in recent-onset T1D [18]. In
the same animals; (3) several sequence homologies exist summary, despite continuing interest in vitamin D supple-
between bovine β-casein and β-cell autoantigens; (4) spe- mentation as a potential intervention to prevent islet auto-
cific cellular and humoral immune responses toward immunity and T1D, there is still little supporting evidence
bovine β-casein are detectable in most T1D patients at the from prospective birth cohort studies.
time of diagnosis, highly suggestive that this protein may
participate in the immune events triggering the disease; (5)
Prediction of T1D as the Basis for Disease
casein hydrolysate was shown to be non-diabetogenic in
Prevention
the BB rat and NOD mouse models, therefore it was
thought that this dietary intervention might be beneficial in There are different approaches for the identification of indi-
humans as well for disease prevention. viduals at risk for T1D. These approaches are based on family
The rationale behind the use of cow’s milk hydrolysate history of T1D, genetic disease markers, autoimmune mark-
for primary prevention of T1D is based on several epide- ers, or metabolic markers of T1D (Figure 2.2). These alterna-
miological and in vitro studies, indicating that intact cow’s tives may also be combined in various ways to improve the
milk, if given before three months of age, may induce an predictive characteristics of the screening strategy. The
Diabetes 100
Beta cell mass (%)

80
60
40
20
years
5 10 15 20
Secondary prevention
Genetic analysis
Tertiary prevention
Primary prevention
FIG 2.2 Strategies to preserve β-cell mass in T1D. Modified from Reimann M, Bonifacio E, Solimena M et al. An update on
preventive and regenerative therapies in diabetes mellitus. Pharmacol Ther 121(3): 317–331, 2009.
importance of understanding the natural history of immune activity directed against β-cells. While direct evaluation of
mediated pre-diabetes lies in the development of prevention T-cell activity might be preferable, antibody determina-
strategies. Several randomized clinical intervention trials tions are generally used for screening because these assays
have been concluded and the next generation of such trials are more robust. Antibody titers are often used in combi-
will rely upon improved and simplified identification of indi- nation with an assessment of the genetic susceptibility, pri-
viduals who are at high risk of progression to T1D. This is marily evaluated by HLA typing. In the near future, testing
essential to ensure that trials have sufficient statistical power for oxPTM-INS-Ab may help to identify children progress-
to detect a given effect of the intervention within the time ing to overt diabetes.
available for the study. Such understanding is also needed to Interventions are generally designed to delay or prevent
avoid exposing those who will not develop T1D to the risk of T1D by impacting some phases of the immune pathogen-
adverse effects of the intervention. esis of the disease. As discussed below, current trials are
attempting to modify the course of disease progress at
many points along the presumed pathogenic pathway.
Prevention of T1D: Current Status
Most prevention trials include only relatives of T1D
Although the process by which pancreatic β-cells are patients, a group in which risk prediction strategies are
destroyed is not well understood, several risk factors and most established. Trials in genetically at-risk infants evalu-
immune-related markers are known to accurately identify ate whether avoiding one of the putative environmental
first-degree relatives of patients with T1D who may develop triggers for T1D can delay or prevent its onset.
the disease. Since we now can predict the development of
T1D, investigators have begun to explore the use of inter- Primary Prevention
vention therapy to halt or even prevent β-cell destruction Primary prevention identifies and attempts to protect indi-
in such individuals. The autoimmune pathogenesis of T1D viduals at risk from developing T1D. It can therefore reduce
determines the efforts to prevent it. Susceptible individuals both the need for diabetes care and the need to treat
are identified by searching for evidence of autoimmune diabetes-related complications (Figure 2.2).
T1D is relatively easy to prevent in animal models of the TABLE 2.1 Prevention in T1D
disease and an array of therapies is effective. However, the
Study
mechanism of prevention is usually poorly defined, and
there is a lack of surrogate assays of the immune response to Primary prevention DAISY:
define which therapies are likely to prevent diabetes in cow’s milk intake and IA development
humans. Inability to define surrogate assays probably results TRIGR:
from a fine balance of the immune system, so that even with Casein hydrolysate vs cow’s milk formula
inbred strains of animals, only a subset progress to diabetes, Babydiet:
Delayed introduction of dietary gluten
and thus relatively small changes in immune function may
TEDDY:
prevent disease. These observations have led to the hypoth-
Timed introduction of gluten-containing
esis that identifying children at a very high genetic risk for cereals
diabetes, prior to development of measurable β-cell autoim- DIPP:
munity, and treating them at that point may be a more effec- Intranasal insulin
tive means of diabetes prevention. Studies for the primary Secondary ENDIT: Nicotinamide
prevention of T1D, i.e., prior to the expression of islet prevention DPT-1: Insulin/oral insulin
autoantibodies, are currently being designed and imple- DIPP: Intranasal insulin
INIT: Intranasal Insulin Trial
mented. These studies target young children at a very high
Tertiary prevention Cyclosporine
genetic risk for T1D and propose treatments that are very
Nicotinamide
safe. These studies require large-scale screening to identify Vitamin D
high-risk subjects and a follow-up over a long period of Insulin
time to observe the outcome of anti-islet autoimmunity as a Anti-CD3
surrogate marker for the disease and onset of hyperglycemia Anti-CD20
Anti TNF alfa
as the main end point (Table 2.1).
CTLA4-Ig
As mentioned above, various nutritional components GAD- Alum
have been suggested to modulate the risk of T1D. Among Anti IL-1
them, several epidemiological and in vitro studies indicat- Anti IL-6
ing that intact cow’s milk, if given before 3 months of age, Anti IL-8
CXCR1/2 inhibitor
may induce an immune response towards β-cells.
In a prospective study called DAISY, which followed
children at increased T1D risk for IA and T1D develop-
ment, cow’s milk protein intake was associated with genetic disease susceptibility (diabetogenic HLA alleles and
increased IA risk in children with low/moderate risk first-degree relatives with T1D) and whether the use of a
HLA-DR genotypes [hazard ratio (HR): 1.41, 95% confi- cow’s milk hydrolysate could protect from the disease. The
dence interval (CI): 1.08–1.84], but not in children with recruitment was carried out over a 5-year period in nine
high risk HLA-DR genotypes. Furthermore, cow’s milk European countries, six major centers in the USA, 12 cent-
protein intake was associated with progression to T1D ers in Canada, and three centers in Australia. Due to statis-
(HR: 1.59, CI: 1.13–2.25) in children with IA [15]. tical considerations, the frequency of the high-risk HLA
TRIGR is a large international randomized double-blind genotype, consent and drop-out rates, the trial required
intervention study intended to provide information on the initial access to 8000 pregnancies, which ultimately yielded
incidence of predictive islet- cell autoantibodies vs the 5156 infants necessary for randomization. 1081 were rand-
actual occurrence of clinical diabetes in two treatment omized to be weaned to the extensively hydrolyzed casein
groups [19]. The aim of this trial was to investigate whether formula and 1078 to a conventional adapted cow’s milk for-
early exposure to complex dietary proteins of cow’s milk mula supplemented with 20% of the casein hydrolysate. The
could increase the risk of T1D in new-born infants with participants were observed for a median of 11.5 years.
The strength of this trial was the larger number of par- followed 8676 children with increased genetic risk of T1D
ticipants, which provides substantially greater statistical in the U.S., Finland, Germany, and Sweden, authors
power in a more heterogeneous study population compared observed that later introduction of gluten-containing cere-
with previous studies and, therefore, provides a more defin- als was associated with increased risk of any IA using Cox
itive answer to whether weaning to an extensively hydro- regression models [22]. In this study, the HRs for every
lyzed formula might be protective of diabetes. This trial 1-month delay in gluten introduction were 1.05 (95% CI
showed that among infants at risk for T1D, weaning to a 1.01, 1.10; P = 0.02) and 1.08 (95% CI 1.00, 1.16; P = 0.04),
hydrolyzed formula compared with a conventional formula respectively. The risk of IA associated with introducing
did not reduce the cumulative incidence of T1D. Thus, the gluten before 4 months of age was lower (HR 0.68; 95% CI
absolute risk of T1D was 8.4% among participants rand- 0.47, 0.99), whereas the risk of IA associated with introduc-
omized to the casein hydrolysate vs 7.6% among those ran- ing it after 9 months of age was higher (HR 1.57; 95% CI
domized to the conventional formula; the hazard ratio for 1.07, 2.31) than introduction between 4 and 9 months of
T1D adjusted for HLA risk group, duration of breastfeed- age. Interestingly, another sub- analysis performed on
ing, duration of study formula consumption, sex, and TEDDY participants shows that administration of probiot-
region while treating study center as a random effect was ics during the first 27 days of life reduced the risk for a first-
1.1 (95% CI, 0.8 to 1.5; P = 0.46). The median age at diagno- appearing β- cell autoantibody in children with the
sis of type 1 diabetes was similar in the 2 groups (6.0 years HLA-DQ2/8 genotype [23]. However, once again, authors
[Q1–Q3, 3.1–8.9] vs 5.8 years [Q1–Q3, 2.6–9.1]) [19]. acknowledged the need to confirm these results in further
In conclusion, studies investigating cow’s milk as an studies before any recommendation of dietary intake or
environmental factor show inconsistent results and do not probiotics use is made.
currently support the need to revise the dietary recommen- Also noteworthy are some observational studies sug-
dations for infants at risk for T1D. gesting a protective role of vitamin D and long-chain n-3
Another outstanding question concerns the association fatty acid against T1D by modulating the immune system.
between timing of gluten introduction and IA develop- Nonetheless, clinical trials have been inconclusive so
ment. The BABYDIET study was an intervention study far [14, 24].
aimed to determine whether primary intervention through Finally, the Diabetes Prediction and Prevention Project
delayed introduction of dietary gluten was feasible and (DIPP Study) (Clinical trial NCT03269084; www.
could reduce the incidence of IA in high-risk first degree clinicaltrials.gov) was a longitudinal study on T1D predic-
relatives of patients with T1D [20]. The study was based on tion and prevention carried out in the university hospitals
the premise that introduction of foods containing gluten or of Turku, Tampere, and Oulu (Finland). The aim of the
cereal before the age of 3 months was associated with an study was to investigate longitudinally the dietary factors
increased risk of IA in childhood. New-born children were in relation to the development of diabetic autoantibodies
eligible if they were younger than 3 months, were offspring and clinical T1D. The diet of children was followed up by a
or siblings of patients with T1D and had HLA genotypes structured questionnaire and by 3-day dietary records at
that confer a high T1D risk. In this study, authors did not various ages. A food frequency questionnaire was applied
find a benefit in delaying gluten exposure with respect to for studying the dietary intake of pregnant and lactating
autoimmunity associated with diabetes and celiac disease mothers.
at 3 years of age. The aims of this project were: (1) to identify infants at
The follow-up findings of the BABYDIET study con- increased genetic risk for T1D from the general popula-
firm previous results, failing to demonstrate that an inter- tion at birth; (2) to monitor such children for the
vention based on delayed gluten introduction will reduce appearance of diabetes- associated autoantibodies, to
the risk of developing autoimmunity related to T1D [21]. identify those at high risk to develop clinical disease and
Conversely, in The Environmental Determinants of to characterize the natural course of T1D; (3) to identify
Diabetes in the Young (TEDDY) study, which prospectively the environmental factors inducing the seroconversion
to autoantibody- positivity in children at increased European Nicotinamide Diabetes Intervention

genetic risk; and (4) to evaluate whether it is possible to Trial (ENDIT)
delay or prevent progression to clinical T1D by daily The ENDIT study, conducted predominantly in Europe,
administration of intranasal insulin. Whereas points examined whether nicotinamide could lead to a reduction
1–4 have been fulfilled and useful information has been in the rate of progression to T1D in at risk relatives of T1D
obtained, the trial with intranasal insulin started soon probands. Over 40,000 first- degree relatives aged 5–40
after detection of autoantibodies, and did not show any years were screened in centers in Europe and North
beneficial effect of this treatment in preventing or delay- America. The study was designed to recruit at least 422
ing the disease [25]. subjects with ICA titers ≥ 20 JDF units to be randomized to
In conclusion, since the failure of ENDIT and DPT1 tri- either a nicotinamide-or a placebo-treated group. With an
als (see secondary prevention) in preventing the onset of expected rate of progression to diabetes of 40% in the pla-
T1D in subjects who are β- cell autoantibody positive, cebo arm, the proposed 5-year observation period should
interest has switched to prevention trials starting before have allowed a 90% power to observe a 35% reduction in
islet-cell autoimmunity has developed. These primary pre- the incidence of disease [26, 27].
vention trials of T1D offer an exciting view of how our Nicotinamide treatment at the doses used did not show
knowledge of the pathogenesis of this disease can lead to any significant effect on the primary outcome – progres-
the possibility of intervening at birth. There is still a long sion to T1D. A total of 159 participants developed the dis-
way to go; however, the rationale is sound and the pros- ease within 5 years of randomization to treatment, 82
pects seem good. (30%) in the active treatment group and 77 (28%) in the
placebo group. The unadjusted Cox proportional hazard
Secondary Prevention estimate showed no difference between the placebo and
Secondary prevention of T1D aims to reduce the inci- nicotinamide groups on an intention to treat basis. Nor was
dence of the disease by stopping progression of β-cell any difference was found between groups after adjustment
destruction in individuals with signs of such a process. for age at baseline, glucose concentrations at 2-h glucose in
This approach may represent a viable alternative to an the OGTT, and number of islet autoantibodies. The pro-
actual cure by blocking the autoimmune response while portion of relatives who developed diabetes within 5 years
β-cell mass is still functional. For this reason, in the last was almost identical in those treated with nicotinamide
few decades, prevention studies have started to focus on and those treated with placebo, and there was no sugges-
subjects at risk of developing the disease, in which the tion of a treatment effect in any of the subgroups defined
autoimmune process has begun but the β-cell mass is still by well established markers of additional risk.
preserved. The most-used approach at this stage of T1D is A useful message of this trial has been that large-scale
as an antigen-specific therapy, based on insulin adminis- collaborations were essential to move things forward and
tration to establish tolerance. Generally, mechanistic that the place for single-center trials was limited.
long-term effects of these treatments rely on effector
T-cell depletion with expansion of T regulatory cells (T- DPT-1 Trials
reg) (Figure 2.2). The Diabetes Prevention Trial -Type 1 (DPT-1) consisted
Three large multicenter trials of diabetes prevention in of two clinical trials that sought to delay or prevent T1D.
autoantibody-positive relatives have been completed. The Nine medical centers and more than 350 clinics in the
European Nicotinamide Diabetes Intervention Trial United States and Canada took part in the two trials of the
(ENDIT), the Diabetes Prevention Trial-Type 1 (DPT-1) DPT-1 [28].
and the Type 1 Diabetes Prediction and Prevention Study Individuals who were eligible for testing were identified as
(DIPP). However, no significant benefits of these treat- follows: age 3 to 45 years, with a brother or sister, child or
ments on the prevention of clinical onset of T1D were parent with T1D; age 3 to 20 years, with a cousin, uncle or
found (Table 2.1). aunt, nephew or niece, grandparent, or half-sibling with T1D.
Those who met these criteria had ICA antibodies measured. response (FPIR) to intravenous glucose, assessed oral glu-
To be eligible, a subject had to be positive for ICAs. cose tolerance (OGT), and determined presence or absence
Animal research and small studies indicated that small, of HLADQA1* 0102/DQB1*0602 (a protective haplotype
regular doses of insulin could prevent or delay T1D in sub- that excluded subjects from participation).
jects at risk. One DPT-1 trial tested whether low-dose insu- The study was a double-masked, placebo-controlled,
lin injections could prevent or delay the development of randomized clinical trial, in which participants were
T1D in people at high risk for developing T1D within 5 assigned to receive capsules of either oral insulin, 7.5 mg of
years. recombinant human insulin crystals (Eli Lilly, Indianapolis,
First-degree relatives, 3 to 45 years of age, and second- IN), or matched placebo. Subjects consumed the capsule
degree relatives, 3 to 20 years of age, of patients with T1D (insulin or placebo) as a single daily dose before breakfast
were screened for islet-cell antibodies. Those with an islet- each day, either by taking the capsule or, if the subject could
cell antibody titer of 10 JDF units or higher were offered not swallow capsules, sprinkling its contents in juice or on
staging evaluations. food.
Subjects identified as having a high risk of T1D were eli- In the primary analysis of relatives selected and rand-
gible for random assignment to the experimental interven- omized in DPT-1, oral insulin did not delay or prevent
tion (parenteral insulin therapy) or to a control group that development of diabetes. There was greater variability in
underwent close observation. the IAA assay for values 39–79 nU/ml than for values
The results demonstrated that insulin, in small doses, ≥80 nU/ml, particularly in confirmation of a positive result
can indeed be administered safely to persons who are at (98.7% overall confirmation for values ≥80 nU/ml com-
risk for T1D. The increase in presumed and definite hypo- pared with 70.6% for values 39–79 nU/ml). This prompted
glycemia among the subjects in the intervention group did comparison of the rate of evolution of diabetes by entry
not adversely affect cognitive function. IAA level. The cohort with confirmed IAA ≥80 nU/ml (the
In high-risk relatives of patients with diabetes, the original entry IAA criterion) progressed to diabetes at a
insulin regimen did not delay or prevent the development faster rate than those subjects who did not have confirmed
of T1D [27]. There are several potential explanations for IAA ≥80 nU/ml. In addition, those with confirmed IAA
the lack of effect observed so far. One is that the interven- ≥ 80 nU/ml had other risk characteristics that suggested
tion took place too late in the disease process to slow more rapid evolution to diabetes, including younger age,
down the progression of disease. Studies conducted ear- greater likelihood of having other antibodies, and greater
lier in the disease process may be more successful. loss of β-cell function [28].
Moreover, the low dose insulin used in the trial may have The effect of intervention in each of these two sub-
failed to achieve such an effect on β-cells, but the dose groups was further evaluated.
was limited by the risk of hypoglycemia. With a different The group with confirmed IAA ≥ 80 nU/ml showed a
dosing scheme or a different regimen, insulin or insulin- beneficial effect of oral insulin, whereas the group who did
like peptides might alter the course of development of not have confirmed IAA ≥ 80 nU/ml showed a trend sug-
diabetes. gesting a detrimental effect of oral insulin [28]. This group
The other study was an oral insulin trial that sought to also had a much lower overall rate of development of
prevent T1D in subjects with a moderate risk for develop- diabetes.
ing diabetes [28]. Furthermore, the rate of progression seemed to increase
First-degree (ages 3–45 years) and second-degree (ages when oral insulin therapy was stopped, suggesting that the
3–20 years) relatives of patients with T1D were screened therapy was probably effective but required ongoing
for ICAs. Those with ICA titer ≥10 JDF units were invited administration [29]. This observation has prompted a
to undergo staging evaluations. larger and justified follow-up study with oral insulin to con-
Staging confirmed ICA positivity, measured insulin firm these preliminary studies (Clinical trial NCT00419562;
autoantibody (IAA) status, assessed first- phase insulin www.clinicaltrials.gov).
In conclusion, neither low-dose insulin injections in The best results in this field were obtained 30 years ago
subjects at high risk for developing T1D nor insulin cap- with the use of cyclosporine, subsequently abandoned
sules taken orally by those at moderate risk for T1D were because of transient benefits and undesired adverse
successful at preventing or delaying the disease. effects [33].
In the following years none of the several treatments
Type 1 Diabetes Prediction and Prevention that have been proposed has obtained appreciable results
Study (DIPP) but for nicotinamide [34] (Table 2.1).
The DIPP study was a randomized double- blind trial Over the last few decades, there has been growing inter-
investigating whether nasal insulin could reduce the inci- est in vitamin D and its active metabolites in relation to
dence of T1D in children with HLA genotypes and autoan- T1D and its immune pathogenesis. Vitamin D metabolites
tibodies conferring increased risk of disease [25]. Daily have been shown to exert several immunomodulatory
doses of intranasal insulin were administered; however, effects and 1,25- dihydroxyvitamin D3 [1,25- (OH)2D3]
after 1.8 years of observation, no differences were found in can either prevent or suppress autoimmune encephalomy-
the rate of progression to T1D. elitis, inflammatory bowel disease, and other autoimmune
Similar results have been obtained in the Intranasal diseases. Based on this rationale, several interventional and
Insulin Trial (INIT I). This pilot study, based in Australia randomized controlled trials evaluated the role for vitamin
and New Zealand, treated autoantibody-positive subjects D in the treatment of T1D, with mixed results.
with intranasal insulin, showing that intranasal insulin did Previous data in humans have demonstrated that reduc-
not prevent T1D onset. However, investigators found that tion in vitamin D supplementation is associated with a
intranasal insulin administration induced immune changes higher risk of the disease, whereas its supplementation is
consistent with mucosal tolerance to insulin, justifying a associated with a decreased frequency of T1D [35]. Other
formal trial to determine if intranasal insulin is immuno- authors observed a significant increase in T-reg cells in
therapeutic and retards progression to clinical diabe- T1D patients supplemented with cholecalciferol at differ-
tes [30]. The INIT II study is still ongoing and will expand ent dosages [36, 37]. Similar effects were reported by
the number of enrolled subjects. Thus, clinical trials evalu- Treiber et al., who administrated 70 IU/Kg of cholecalcif-
ating insulin administration for disease prevention have erol daily for 12 months, demonstrating not only the
demonstrated to date limited success in preventing the dis- enhancement of the T-reg cells, but also an increased T-reg
ease’s progression. cell suppressive capacity among the supplemented
group [38].
Tertiary Prevention Different trials showed a better preservation of residual
Tertiary prevention is aimed at delaying or preventing the pancreatic β-cell function in T1D patients supplemented
development of complications in subjects who already have with different forms of vitamin D (cholecalciferol, alfacal-
T1D. A landmark trial investigating patients with T1D cidiol or calcitriol), as proven by the significantly higher
showed that good glycemic control [31] as well as low gly- level of fasting C-peptide and/or lower needed daily insulin
cemic variability [32] can reduce the likelihood of micro- dose observed in supplemented groups [37, 39]. However,
vascular complications leading to blindness or kidney there are also studies that indicate no significant role for
disease, but the trend toward a decrease in macrovascular vitamin D in the treatment of T1D. The IMDIAB XI trial
disease was not statistically significant. Diabetes education was an open-label randomized trial designed to determine
of health care professionals and those affected by diabetes whether supplementation with the active form of vitamin
plays a key role in the tertiary prevention of the disease. D (calcitriol) at diagnosis of T1D could improve parame-
Tertiary prevention is identified by the maintenance of the ters of glycemic control [40]. The secretion of C-peptide as
residual β-cell function present at disease onset and can be an index of residual pancreatic β-cell function was the pri-
realized by immune suppression or immune modulation mary end point, with HbA1c and insulin requirement as
since the time of clinical diagnosis of T1D (Figure 2.2). secondary end points. The aim of this study was to
investigate whether supplementation with the active form therapy than the untreated patients. Those who received
of vitamin D (calcitriol) in subjects with recent-onset T1D, the antibody treatment also had better HbA1c levels. The
protects residual pancreatic β-cell function and improves anti-CD3 was designed to act on the immune system’s
glycemic control (HbA1c and insulin requirement). In this T-cells in a more specific manner than previous attempts at
open-label randomized trial, 70 subjects with recent-onset immune intervention in early diabetes.
T1D, mean age 13.6 ± 7.6 years, were randomized to calci- Most recently, a phase 2, randomized, placebo-
triol (0.25 microg on alternate days) or nicotinamide controlled, double-blind trial showed that a single 14-day
(25 mg/kg daily) and were followed up for 1 year. Intensive course of teplizumab (an Fc receptor–nonbinding anti-
insulin therapy was implemented with three daily injec- CD3 monoclonal antibody) significantly slowed progres-
tions of regular insulin + NPH insulin at bedtime. No sig- sion to clinical T1D in high-risk, nondiabetic relatives of
nificant differences were observed between calcitriol and patients with diabetes who had at least two autoantibod-
nicotinamide groups in respect of baseline/stimulated ies. At the conclusion of the trial, the percentage of
C-peptide or HbA1c 1 year after diagnosis, but the insulin diabetes-free persons in the teplizumab group (57%) was
dose at 3 and 6 months was significantly reduced in the double that in the placebo group (28%). Median delay in
calcitriol group. In conclusion, at the dosage used, calcitriol the diagnosis of diabetes was 2 years. However, the cohort
had a modest effect on residual pancreatic β-cell function was relatively small, and the estimated power limited.
and only temporarily reduced the insulin dose [40]. These Furthermore, authors did not assess the potential devel-
results were confirmed later by the same group (IMDIAB opment of antibodies to teplizumab, which would be a
XIII Trial), who found no effect of calcitriol in protecting concern [44].
β-cell function in subjects with recent-onset T1D and high In the DEFEND-1 and DEFEND-2 phase III trials, anti-
C- peptide at diagnosis followed- up for 2 years [41]. CD3 antibody Otelixizumab revealed a narrow therapeutic
Currently, a pilot study (POSEIDON) is investigating the window. At a low dose, there was no preservation of β-
safety and efficacy of a regimen that combines Omega-3 mass [45]. This was indeed observed at 18 months, but
fatty acids and cholecalciferol in subjects at T1D onset with significant adverse effects. Otelixizumab is a chimeric
(NCT03406897). Thus, omega-3 fatty acids have effects on monoclonal antibody that targets the CD3/T-cell receptor,
several immunotypes and are known to increase T-reg cell which is genetically modified by removing the glycosyla-
differentiation. The recruitment is still open. tion site in the Fc domain, thus affecting binding of com-
plement or Fc receptors. This reduces secondary reactions
Immune Intervention Therapies at Diagnosis of T1D due to cytokine release. Otelixizumab downregulates path-
Other strategies for prevention of β-cells damage with ogenic T-cells and upregulates T-reg cells, thus halting the
immune intervention at onset of the disease are based on autoimmune process in T1D.
immunotolerance (monoclonal antibodies, antigen-based Only one trial has focused on β-cell function in T1D
treatments, pro-inflammatory cytokine-based treatments) after treatment with Rituximab, which produces B cell
(Figure 2.3, Tables 2.2 and 2.3). depletion [46]. Although T1D is a T-cell mediated auto-
In the last decades, experience obtained with the use of immune disorder, B lymphocytes play a pathogenetic
the antiCD3 monoclonal antibody in two studies (one in the role by acting as antigen-presenting cells (APC) and
USA and the other in Europe) has revitalized the interest in modulating the islet environment. In recent-onset T1D
these types of intervention [42, 43]. The drug, a modified subjects, administration of Rituximab reduced HbA1c
form of anti-CD3 antibody that minimizes first-dose side levels and exogenous insulin demand due to the preser-
effects, was studied by comparing 12 subjects aged 7 to vation of C-peptide levels over 1 year. Interestingly, after
30 who were treated with the antibody to an equal number 2 years’ follow-up, Rituximab delayed the decrease in
of patients in a control group who did not receive the drug. C-peptide levels but did not influence insulin dose, sug-
One year after treatment with anti- CD3, the treated gesting B cell deletion is not sufficient to restore β-cell
patients produced more insulin and needed less insulin tolerance.
Thymocyte Depletion + Immunomobilization
Delay & AbATE– ATG + G-CSF –2016
Teplizumab ATG + G-CSF + Rapamycin + Islet Transplant – 2018
Alemtuzumab + Anakinra + Etanercept + Liraglutide +
DEFEND 1 & DEFEND 2– Plerixafor – 2021
1-alpha\ ATG + G-CSF + Low dose IL-2 + Etanercept +
otelixizumab Exenatide – 2021
hydroxyvitamin D3 Cytokines
DILT1D–IL-2
Anti-IL-6 (Siltuximab) – 2017
1alpha, 25- REPAIR- Anti-IL-6 Receptor (Tocilizumab) – 2018
Autologous Canakinumab Autologous Anti-IL12/IL-23 (Ustekinumab) – 2017
dihydroxyvitamin D3 T1D–
Hematopoietic CD3+CD4+CD25+CD127– Anti-IL12/IL-23 (Ustekinumab) + INGAP – 2017
Stem Cell AIDA– Sitagliptin + Anti-IL21 (NNC0114–0006) + Liraglutide - 2019
polyclonal Treg cell
Autologous Transplant Anakinra Lansoprazole Aldesleukin (IL-2) – 2016
therapy
Non-myeloablative ChAglyCD3 IL-2–2017
Hematopoietic Stem Low dose rhIL-2–2018
Anakinra
Cell Transplant with Low dose Proleukin (IL-2) – 2023
Protégé– Anti-TNFα (Golimumab) – 2019 + 2021
ATG + Teplizumab Autologous Autologous
Teplizumab Bone marrow T cells
Cyclophosphamide Umbilical Cord IMCY-0098–2018
Conditioning Stem Cells Blood Otelixizumab – 2018
Exenatide +/– START – ATG
Teplizumab – 2022
Daclizumab DIATOR– Wharton′s Tregs
Mycophenolate Atorvastatin Autologous Cord
jelly-derived ATG + G-CSF CLBSO3 (Autologous Ex Vivo Expanded Polyclonal
mofetil + Blood + oral
Mesenchymal Regulatory T-cells) – 2020
Daclizumab BCG Docosahexaenoic] CD4+CD127lo/-CD25+ Polyclonal Tregs + IL-2 – 2021
Stem cells G-CSF
Vaccination acid + Vitamin D Umbilical Cord Blood Tregs + Liraglitude – 2019
Umbilical Cord Blood Tregs – 2019
Tregs + anti-CD20 antibody – N/A
T cells Costimulation
+/– T&B T cell- Stem Hygiene Stem Increti Thymocytes +/– Abatacept (CTLA4-lg) – 2018
Stem Cells Vitamin D T cells IL-1 T cells UCB IL-1 Treg
Incretin cells APC cells Hypothesis cells n& H2+ Immunomobilizatio
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Ongoing
Global Immune T cell co- Innate Vitamin Innate

B cells Cytokines Suppression UCB DCs Vitamin D Treg +/– β cell Immunity Stem cells Immunity Innate Immunity
stimulation D
Alpha-1 Proteinase Inhibitor – 2017
PGD2 receptor antagonist MK-1092 – 2018
Autologous Immunoregulatory Densritic Cells – 2019
Stem Cells
Allogeneic mesenchymal Stem Cells – 2017
Aplha-1 Aplha-1 Mesenchymal Stem Cells – 2017
Autologous antitrypsin
Tolerogenic antitrypsin Allogeneic Adipose Mesenchymal Stem Cells with
ENBREL Bone Marrow Mononuclear Cells - 2018
(Etanercept) Dendritic Stem Cell Educator Therapy – 2019
cells TIDAL – Alefacept aLD-SCs / CB-MSCs + G-CSF – 2019
Cholecalciferol Chemotherapeutics
Oral IFN-α
Imatinib – 2019
Abatacept Neuropeptide
Rapamycin Substance P – 2017
+ IL-2 CXCR1/2 Inhibitor
Autologous Immunoablation (ATG + Cyclophosphamide) + Ladarixin – 2018
Umbilical Incretin
DF-IL2– Autologous Hematopoietic Stem Cell Transplant Sitagliptin – 2017
Cord Blood IL-2
Transfusion + G-CSF Exenatide – 2017
Saxagliptin – 2017
Albiglutide – 2017
Rituximab Umbilical Cord Mesenchymal Stromal Cell + Vildagliptin – 2017
Cholecalciferol Autologous Bone marrow Mononuclear Cell Liraglitude – 2018 + N/A
Stem Cell Transplant Rapamycin + Vildagliptin – 2018
Calcitriol 𝛃 cell Stress
Alpha1-Antitrypsin – 2017
Adipose-derived Insulin-secreting Mesenchymal Tauroursodeoxycholic Acid (TUDCA) – 2018
Cyclosporin + Verapamil – 2019
Methotrexate Stromal Cells + Bone marrow-derived
Alpha difluoromethylornithine (DFMO) – 2019
Hematopoietic Stem Cell Infusion
Vitamin D
Ergocalciferol – 2020
Stem Cell Educator Therapy
Environmental Factors
Prebiotic Fiber – 2017
Autologous Mesenchymal Stromal Cells Probiotics – 2018
GNbAC1 – 2018
Enterovirus vaccination – 2022
BCG vaccination – 2023
FIG 2.3 T1D clinical research over the past 10 year. Atkinson MA, Roep BO, Posgai A et al. The challenge of modulating β-cell autoimmunity in type 1 diabetes.
Lancet Diabetes Endocrinol 2019;7(1):52–64.
TABLE 2.2 Current TrialNet studies. (Source: www.trialnet.net)
Study Status Description
Teplizumab Prevention Study Completed This study tested the drug teplizumab to see if it could delay
or prevent progression of early stage T1D (stage 2) and
prevent clinical diagnosis (stage 3).
Oral Insulin Prevention Study Completed This study tested the drug oral insulin to see if it can delay or
prevent T1D (stage 1) from progressing to stage 2 and
ultimately prevent clinical diagnosis (stage 3).
Hydroxychloroquine (HCQ) Currently Enrolling This study tests the drug hydroxychloroquine (HCQ) to see if
it can delay or prevent early stage T1D (stage 1) from
progressing to abnormal glucose tolerance (stage 2) and
ultimately prevent clinical diagnosis (stage 3).
ATG/GCSF New Onset Study Completed This study was designed to build on prior findings of a pilot
study suggesting thymoglobulin (ATG) combined with
pegylated granulocyte colony stimulating factor (GCSF)
preserved insulin production for more than 1 year after
treatment in people who had type 1 diabetes for 4 months
to 2 years.
Abatacept Prevention Study Currently Not Enrolling This study tests the drug abatacept to see if it can delay or
prevent progression of early stage T1D (stage 1 or stage 2),
and ultimately prevent clinical diagnosis (stage 3).
Pathway to Prevention Currently Enrolling This study screens and observes relatives of people with type
1 diabetes to learn more about how the disease occurs
The use of immunosuppressive drugs has also been pro- fully established. It is however clear that GAD65 is one of
posed as a possible approach to decelerate the evolution of the most important targets when the immune system
the disease. attacks the insulin producing β-cells in autoimmune diabe-
In a multicenter, double- masked, randomized con- tes. Thus, treatment with rhGAD65 is thought to induce
trolled trial, Abatacept (CTLA4-Ig) has been administered tolerance to GAD65, thereby intervening in the autoim-
on days 1, 14, 28, and then every 28 days through a 30 min mune attack and preserving the capacity to produce insulin
intravenous infusion at a dose of 10 mg/kg [47]. Results in patients with autoimmune diabetes.
from this study showed that Abatacept was more efficient Although ongoing studies are investigating whether
compared to placebo in preserving the β-cell mass, as evi- rhGAD65 can preserve β-cell function in recently diag-
denced by stimulated C-peptide secretion. However, the nosed individuals with T1D, trials carried out to date do
effect diminished with time; therefore, further investiga- not demonstrate a significant reduction in the loss of stim-
tion will be necessary in order to unravel whether the ben- ulated C-peptide in recently diagnosed children and young
eficial effect persists after cessation of infusions. Indeed, adults (10–20 years) with T1D over a 15-month period [48].
Abatacept is a potential candidate to be used in tertiary GAD65 was also targeted in NOD mice in order to
prevention trials, and is a candidate for use in combination reduce the number of GAD65-specific T effector cells [49],
therapies for recent-onset T1D patients. achieving normoglycemia in 70% of NOD mice. Based on
GAD65 (the 65 kDa isoform of glutamic acid decarbox- this findings, antigen-based immunotherapy therapy with
ylase) is a human enzyme that has an important role in the subcutaneous GAD-alum have been tested to slow down
nervous system and in several nervous system diseases, e.g. the course of loss of insulin in patients with recently diag-
Parkinson’s disease and chronic pain. nosed T1D. Recently, the prevention trial DIAPREV-IT has
GAD65 is also found in the insulin producing β-cells of showed that GAD-Alum as a subcutaneous prime and
the pancreas, although its function at this site is not yet boost injection was safe in prediabetic young children but
TABLE 2.3 Immunotherapy trials that are awaiting, or currently recruiting, participants. (Source: www.clinicaltrials.gov)
ClinicalTrials.gov
identifier Title Intervention Primary outcome
NCT02307695 The Effect of Saxagliptin on Glucose Fluctuation and Saxagliptin MAGE at 24 weeks
Immune Regulation in Patients with Type 1 Diabetes
NCT01559025 Evaluation of Vildagliptin (Galvus®) as add-on to Vildagliptin MMTT C-peptide at
Insulin in Residual β-cell Function and Inflammatory 3, 6, 9 and
Markers in New-onset Type 1 Diabetes Mellitus 12 month
NCT02442544 Effect of Prebiotic Fibre on Gut Microbiota, Intestinal Prebiotic 1:1 HbA1C at 3 months
Permeability and Glycaemic Control in Children with oligofructose:inulin
Type 1 Diabetes: A Pilot Randomized, Double Blind,
Placebo Controlled Study
NCT02820558 A Phase I Study of Safety and Pharmacological Activity Substance P Safety at
of Substance P in the Reversal of Recent-Onset Type 1 20–27 days
Diabetes
NCT02505893 A Monocentric, Open-label Pilot Study to Assess the ATG + pG-CSF + rapamycin Safety and MMTT
Safety and Efficacy of Minimal Islet Transplantation in + human pancreatic islet C-peptide at
Patients with New-onset Type 1 Diabetes 12 months
NCT02940418 Use of Stem Cells in Diabetes Mellitus Type 1 AD-MSCs + BM-MNCs Safety at 6 months
NCT02293837 EXTEND Tocilizumab MMTT C-peptide at
12 months
NCT02617654 A Randomized, Double-blinded Placebo-controlled, Liraglutide MMTT C-peptide at
Paralleled Designed, Investigator Sponsored Study of 12 months
the Effect of the GLP-1 Receptor Agonist Liraglutide
on Β-cell Function in C-peptide Positive Type 1
Diabetic Patients
NCT03170544 A Single Ascending Dose Clinical Trial to Study the MK-1092 Safety and maximal
Safety, Tolerability, Pharmacokinetics, and glucose infusion
Pharmacodynamics of MK-1092 in Healthy Subjects rate at 33 days
and in Subjects With Type 1 Diabetes Mellitus
NCT02814838 A Phase 2, Multicentre, Randomized, Double-blind, Ladarixin MMTT C-peptide at
Placebo-controlled Study in Patients with New-onset 13±1 weeks
Type 1 Diabetes
NCT02411253 DIABIL-2 rhIL-2 MMTT C-peptide at
12 months
NCT02803892 MONORAPA Rapamycin /+ Vildagliptin MMTT C-peptide at
4±1, 12±2 weeks
NCT02218619 Clinical Investigation of Efficacy of Tauroursodeoxycholic Acid MMTT C-peptide at
Tauroursodeoxycholic Acid (TUDCA) to Enhance (TUDCA) 6, 12, and
Pancreatic Β-cell Survival in Type 1 Diabetes by 18 months
Reducing Endoplasmic Reticulum Stress
NCT03272269 A Phase I Placebo-controlled, Double-blind, Dose IMCY-0098 Safety at 24 weeks
Escalation Clinical Trial to Evaluate the Safety and
Immune Responses of Imcyse’s IMCY-0098 in Patients
With Recent Onset Type 1 Diabetes
NCT03032354 Effect of Lactobacillus Rhamnosus GG and Probiotics MMTT C-peptide at
Bifidobacterium Lactis BB 12 on Beta-cell Function in 6 and 12 months
Children With Newly Diagnosed Type 1 Diabetes -a
Randomized Controlled Trial
NCT02644759 Transplantation of Autologous Stem Cells for the aLD-SCs/CB-MSCs + G-CSF Insulin requirement
Treatment of Type 1 Diabetes Mellitus at 1 month
TABLE 2.3 (Continued )
ClinicalTrials.gov
identifier Title Intervention Primary outcome
NCT02354911 A Randomized, Double-Blind, Placebo-Controlled, Immunoregulatory Dendritic MMTT C-peptide at

Cross-Over Study of the Safety and Efficacy of Cells 12 and 24 months
Autologous Immunoregulatory Dendritic Cells in
Patients With Type 1 Diabetes
NCT02624804 A Pilot Study of the Therapeutic Potential of Stem Cell Stem Cell Educator Therapy Safety at
Educator Therapy in Type 1 Diabetes 12 months
NCT02846545 T1GER Golimumab MMTT C-peptide at
12 months
NCT03011021 Safety and Efficacy of Umbilical Cord Blood UCB -Tregs + Liraglutide Safety at
Regulatory T Cells Plus Liraglutide on Autoimmune 24 months
Diabetes
NCT02932826 Safety Study and Therapeutic Effects of Umbilical UCB - Tregs Safety at
Cord Blood Treg Cells on Autoimmune Diabetes 24 months
NCT02384889 Targeting Polyamines Using DFMO in Persons with Difluoromethylornithine Safety with dose
Type 1 Diabetes: A Randomized, Double-Masked, escalation at
Placebo-Controlled Phase I Study to Evaluate the 6 months
Safety, Tolerability, and Initial Pharmacodynamics of
Multiple Ascending Doses
NCT03046927 Vitamin D and Residual Beta-Cell Function in Type 1 Ergocalciferol MMTT C-peptide at
Diabetes 12 months
NCT01773707 CTLA4-Ig (Abatacept)for Prevention of Abnormal Abatacept Abnormal glucose
Glucose Tolerance and Diabetes in Relatives At -Risk tolerance
for Type 1
NCT03298542 A Phase 1b Study to Evaluate SIMPONI (Golimumab) Golimumab Safety at 26, 52,
Therapy in Children, Adolescents and Young Adults and 78 weeks
With Pre-Symptomatic Type 1 Diabetes
NCT03182426 Autologous Hematopoietic Stem Cell Mobilization Alemtuzumab + anakinra + Safety and MMTT
(Plerixafor) and Immunologic Reset in New Onset Type etanercept + liraglutide + C-peptide at 3, 6,
1 Diabetes Mellitus plerixafor 9, 12, 18, and
24 months
NCT02772679 TILT Tregs + IL-2 Safety and Treg
proportion at 3
years
NCT02804165 Gene-virus Interactions Implicated in Type 1 Diabetes Enterovirus vaccination T1D diagnosis
NCT03243058 A Randomized, Double Blind, Phase I/II Trial of Proleukin (IL-2) MMTT C-peptide at
Low-Dose Interlekin-2 Immunotherapy in Established 12 months
Type 1 Diabetes
NCT02081326 Repeat BCG Vaccinations for the Treatment of BCG HbA1C at 1,2,3,4
Established Type 1 Diabetes and 5 years
EudraCT: TregVac2.0 Tregs + anti-CD20 antibody (Phase II Trial
2014-004319-35 assessing expanded
polytTregs in
patients with recent
onset Type 1
diabetes mellitus)
EudraCT Incretin-based therapy in non-symptomatic, early Liraglutide MMTT C-peptide at
2014-004760-37 diagnosed Type 1 Diabetics 3, 6, 9, and
12 months
did not affect progression to T1D [50]. The beneficial for trial involved 76 patients with new-onset T1D, randomly
prevention of T1D of GAD-alum should be tested in future (2:1) assigned to receive either Ladarixin treatment
prevention studies. (400 mg b.i.d. for 3 cycles of 14 days on/14 days off – treat-
Pro-inflammatory cytokine-based treatments have ment group) or placebo (control group). Although results
proven to be safe and effective for treatment of various indicated no statistically significant differences in stimu-
autoimmune diseases. Thus, inhibition of expression of lated C-peptide at weeks 13 and 26, investigators noted
those molecules can induce important changes in pancre- 76.6% of patients receiving Ladarixin had an HbA1c below
atic β- cells induce important changes in pancreatic 7% and a daily insulin requirement of less than 0.50 IU/kg
β-cells [51]. compared to just 45.8% of patients receiving placebo.
The aim of using the Anti-Interleukin-1 in newly diag- Furthermore, in a prespecified subgroup analysis of
nosed T1D subjects is to test the feasibility, safety/tolerabil- patients with fasting C-peptide below the median value of
ity and potential efficacy of anti-IL-1 therapy in maintaining the trial population at baseline, MMTT AUC of C-peptide
or enhancing β-cell function in people with new onset trended at week 13 and reached statistical significance at
T1D. Anti-IL-1 administration for rheumatoid arthritis week 26.
has been proven to be well tolerated in patients [52, 53].
IL-1 is also involved in T1D progression by activating Incretin-based Therapies
T-helper cells and improving the number of circulating Recent knowledge regarding the heterogeneity in the
memory T-cells [54]. The active substance is interleukin-1 extent of the β-cell impairment and pancreatic lesions as
receptor antagonist, a blocker of an immune-signal mole- well as the differences in circulating T-cell and autoanti-
cule named interleukin- 1. Two randomized placebo- body immune signatures underscore the potential applica-
controlled trials aimed to assess whether canakinumab, a tions for incretin treatments, which improve capacity for
human monoclonal anti-interleukin-1 antibody, or anak- insulin production by residual β-cells and suppress gluca-
inra, a human interleukin-1 receptor antagonist, improved gon secretion, as well as the need for therapeutics to reduce
β-cell function in recent-onset T1D, but their effectiveness β-cell stress co-administered with immunomodulatory
was not demonstrated [54, 55]. therapy to reverse autoimmunity in symptomatic T1D.
More recently, the ongoing clinical trial EXTEND Hence, therapies once considered only applicable to those
(Clinical trial NCT02293837; www.clinicaltrials.gov) is with T2D may be of potential benefit for those with T1D.
currently examining whether the blockade of IL-6 signal- In this regard, ongoing T1D immunotherapy trials are
ing through tocilizumab, an anti-IL-6 receptor antibody, investigating the potential benefits on β-cell function in
can induce a protection of β-cell function in T1D patients C-peptide positive early diagnosed T1D patients.
(ages 6 to 17 years) (Table 2.3). In a post hoc analysis of pooled data from five rand-
Interleukin-8 appears to be another important mediator omized, placebo-controlled studies, the dipeptidyl pepti-
in the progression of T1D. Circulating levels of IL-8 are dase 4 (DPP- 4) inhibitor saxagliptin improved β-cell
elevated in children with T1D compared to non-diabetic function as assessed by HOMA2 of β-cell function and
controls. Furthermore, levels of IL-8 correlate with glyce- post-prandial C-peptide from baseline in patients with
mic control, higher level being associated to poorer glucose Latent Autoimmune Diabetes in the Adult (LADA) [56].
control. As a result, the modulation or inhibition of IL8 Another small study found that sitagliptin, as an add-on
activity may be a valid target for the development of novel treatment to insulin, had a beneficial effect on C-peptide
treatments aimed to control the progression of T1D. decline compared with insulin alone. Moreover, a recent
A multicenter, randomized, double- blind, placebo- trial evaluated the effect of saxagliptin in combination with
controlled phase 2 trial of CXCR1/2 IL- 8 inhibitor vitamin D3 in subjects with LADA with promising
(Ladarixin) has just presented its results at the American results [57]. As far as T1D is concerned, the effect of saxa-
Diabetes Association’s (ADA) 80th Scientific Sessions gliptin on immune regulation have been investigated in a
(Clinical trial NCT02814838; www.clinicaltrials.gov). The phase IV trial (NCT02307695). Similarly, a phase III
r andomized controlled trial is evaluating the action of vild- maintain stable glucose control. Early diagnosis of T1D is
agliptin in the prevention of progressive β-cell dysfunction crucial if we want to restore and to save β-cell mass.
in patients with newly diagnose of T1D (NCT01559025). Different trials using antigen- specific or nonspecific
In another ongoing study, researchers are testing the effi- interventions have shown some benefit in modulation of
cacy of 4 weeks rapamycin treatment and 4 weeks rapamy- the autoimmune process and in preventing the loss of insu-
cin treatment plus 3 months vildagliptin treatment versus lin secretion in the short term after early diagnosis of T1D.
placebo in increasing endogenous insulin production and Unfortunately, there are still limitations to current strat-
correcting glycemic lability (NCT02803892). Glucagon- egies, including a lack of suitable markers to predict and
like peptide (GLP-1) analogues have been tested in large- monitor the success of interventions, uncertainty about the
scale clinical trial to prove their various benefits for β-cell long-term adverse effects or the duration of treatment
and glucolipid metabolism in T2D and obesity patients. effect and the feasibility of restoration of β-cell mass.
This has led to concerns regarding the potential applica- Moreover, we should remember that T1D is a heterogene-
tions in T1D patients. A small phase II randomized con- ous disease with an age at onset spanning from childhood
trolled trial (NCT02617654) is currently investigating the to adult age.
effect of 52 weeks of treatment with liraglutide 1.8 mg/day, Ideally, the interventions would be specific for T1D, free
compared to placebo, on stimulated C-peptide concentra- of adverse effects, and effective prior to disease onset, with
tions in patients with long-standing type 1 diabetes and long-term and clinically meaningful improvements over
residual insulin production (primary outcome: MMTT standard therapies. The success of these approaches will
C-peptide at 12 months). eventually be evaluated by their impact on glycemic con-
Other studies on preservation of β-cell function using trol as this is the definitive determinant of long-term out-
incretin-based therapies are currently active but not come of the disease.
recruiting participants (NCT02443155; NCT02127047). In conclusion we can affirm that early diagnosis of T1D
Results from these studies are warranted to prove that is very valuable and is not a Pyrrhic victory. Hopefully a
incretin-based therapy might preserve C-peptide secretion better comparison would be with the encounter of David
(Table 2.3). and Goliath, where early immune intervention dramati-
cally changes the clinical course of T1D.
Conclusions
Today, one of the therapeutic goals in T1D is the preserva- Acknowledgments
tion of the residual C-peptide secretion that is detected in a We would like to thank Juvenile Diabetes Research
significant percentage of patients at diagnosis and which Foundation (JDRF), National Institute of Health (NIH)
potentially may influence the clinical course of the Consortia, Centro Internazionale Studi Diabete (CISD),
disease. Diabete e Metabolismo (DEM) Foundation and University
Several studies have been demonstrated that residual Campus Bio-Medico that support clinical research on T1D
C-peptide secretion, after T1D diagnosis, depends on in our University Hospital.
genetic factors, the patient’s age at the diabetes diagnosis,
the number of anti- islet antibodies, and the residual
C-peptide secretion. In the same way, intensive insulin References
therapy and immunomodulators drugs may be useful in
1. Nam Han Cho. IDF Diabetes Atlas, 9th edn. Brussels: 2019.
this direction. 2017.
The ultimate goal of any therapeutic intervention is to 2. de Ferranti SD, de Boer IH, Fonseca V et al. Type 1 diabetes
prevent or reverse T1D by abrogation of pathogenic auto- mellitus and cardiovascular disease: a scientific statement
reactivity and by preservation or restoration of the β-cell from the American Heart Association Andamerican
mass and function to physiologically sufficient levels to Diabetes Association. Circulation, 2014;130:1110–1130.
3. Dabelea D, Stafford JM, Mayer-Davis EJ et al. Association of analysis of observational cohort and randomised interven-
type 1 diabetes vs type 2 diabetes diagnosed during child- tion studies. BMJ. 2014;348:g1903.
hood and adolescence with complications during teenage 18. Napoli N, Strollo R, Pitocco D et al. Effect of calcitriol on
years and young adulthood. JAMA. 2017;317:825–835. bone turnover and osteocalcin in recent-onset type 1 diabe-
4. Hamman RF, Bell RA, Dabelea D et al. The SEARCH for tes. PLoS One. 2013;8:e56488.
diabetes in youth study: Rationale, findings, and future 19. Knip M, Akerblom HK, Al Taji E et al. Effect of hydrolyzed
directions. Diabetes Care. 2014;37:3336–3344. infant formula vs conventional formula on risk of type 1 dia-
5. Cerolsaletti K, Hao W, and Greenbaum CJ. Genetics coming betes the TRIGR randomized clinical trial. JAMA.
of age in type 1 diabetes. Diabetes Care. 2019;42:189–191. 2018;319:38–48.
6. Lee HS and Hwang JS. Genetic aspects of type 1 diabetes. 20. Hummel S, Pflüger M, Hummel M et al. Primary dietary
Annals of Pediatric Endocrinology and Metabolism. intervention study to reduce the risk of islet autoimmunity
2019;24:143–148. in children at increased risk for type 1 diabetes. Diabetes
7. Forgetta V, Manousaki D, Istomine R et al. Rare genetic vari- Care. 2011;34:1301–1305.
ants of large effect influence risk of type 1 diabetes. Diabetes. 21. Beyerlein A, Chmiel R, Hummel S et al. Timing of gluten
2020;69:784–795. introduction and islet autoimmunity in young children:
8. Haider MZ, Rasoul MA, Al-Mahdi M et al. Association of Updated results from the Babydiet study. Diabetes Care.
protein tyrosine phosphatase non- receptor type 22 gene 2014;37:e194–195.
functional variant c1858t, HLA- dq/dr genotypes and 22. Uusitalo U, Lee HS, Andren Aronsson C et al. Early infant
autoantibodies with susceptibility to type-1 diabetes mellitus diet and islet autoimmunity in the TEDDY study. Diabetes
in Kuwaiti Arabs. PLoS One. 2018;13:e0198652. Care. 2018;41:522–530.
9. Vella A, Cooper JD, Lowe CE et al. Localization of a type 1 23. Uusitalo U, Liu X, Yang J et al. Association of early exposure
diabetes locus in the IL2RA/CD25 region by use of tag of probiotics and islet autoimmunity in the TEDDY study.
single-nucleotide polymorphisms. Am J Hum Genet JAMA Pediatr. 2016;170:20–28.
2005;76(5):773–779. 24. Sørensen IM, Joner G, Jenum PA et al. Serum long chain n-3
10. Todd JA, Walker NM, Cooper JD et al. Robust associations fatty acids (EPA and DHA) in the pregnant mother are inde-
of four new chromosome regions from genome-wide analy- pendent of risk of type 1 diabetes in the offspring. Diabetes.
ses of type 1 diabetes. Nat Genet 2007;39(7):857–864. Metab. Res. Rev. 2012;28:431–438.
11. Regnell SE and Lernmark Å. Early prediction of autoim- 25. Näntö-Salonen K, Kupila A, Simell S et al. Nasal insulin to
mune (type 1) diabetes. Diabetologia. 2017;60:1370–1381. prevent type 1 diabetes in children with HLA genotypes and
12. Strollo R, Vinci C, Arshad MH et al. Antibodies to post- autoantibodies conferring increased risk of disease: a
translationally modified insulin in type 1 diabetes. double- blind, randomised controlled trial. Lancet.
Diabetologia. 2015;58:2851–2860. 2008;372:1746–1755.
13. Strollo R, Vinci C, Napoli N et al. Antibodies to post- 26. Bingley PJ and Gale EAM. Progression to type 1 diabetes in
translationally modified insulin as a novel biomarker for islet cell antibody- positive relatives in the European
prediction of type 1 diabetes in children. Diabetologia. Nicotinamide Diabetes Intervention Trial: The role of addi-
2017;60:1467–1474. tional immune, genetic and metabolic markers of risk.
14. Rewers M and Ludvigsson J. Environmental risk factors for Diabetologia. 2006;49:881–890.
type 1 diabetes. The Lancet. 2016;387:2340–2348. 27. Beik P, Ciesielska M, Kucza M et al. Prevention of type 1
15. Lamb MM, Miller M, Seifert JA et al. The effect of childhood diabetes: Past experiences and future opportunities. J. Clin.
cow’s milk intake and HLA-DR genotype on risk of islet auto- Med. 2020;9.
immunity and type 1 diabetes: The Diabetes Autoimmunity 28. Skyler JS, Krischer JP, Wolfsdorf J. Effects of oral insulin in
Study in the Young. Pediatr. Diabetes. 2015;16:31–38. relatives of patients with type 1 diabetes: The diabetes pre-
16. Schramm S, Lahner H, Jöckel KH et al. Impact of season and vention trial-type 1. Diabetes Care. 2005;28:1068–1076.
different vitamin D thresholds on prevalence of vitamin D 29. Vehik K, Cuthbertson D, Ruhlig H et al. Long-term outcome of indi-
deficiency in epidemiological cohorts – a note of caution. viduals treated with oral insulin: Diabetes Prevention Trial-Type 1
Endocrine. 2017;56:658–666. (DPT-1) oral insulin trial. Diabetes Care. 2011;34:1585–1590.
17. Chowdhury R, Kunutsor S, Vitezova A et al. Vitamin D and 30. Ryhänen SJ, Harkonen T, Siljander H et al. Impact of intra-
risk of cause specific death: Systematic review and meta- nasal insulin on insulin antibody affinity and isotypes in
young children with HLA-conferred susceptibility to type 1 43. Keymeulen B, Vandemeulebroucke E, Ziegler AG et al.
diabetes. Diabetes Care. 2011;34:1383–1388. Insulin needs after CD3-antibody therapy in new-onset type
31. Gubitosi-Klug RA, Lachin JM, Backlund JYC et al. Intensive 1 diabetes. N. Engl. J. Med. 2005;352:2598–2608.
diabetes treatment and cardiovascular outcomes in type 1 44. Herold KC, Bundy BN, Long SA et al. An anti-CD3 anti-
diabetes: The DCCT/EDIC study 30- year follow- up. body, teplizumab, in relatives at risk for type 1 diabetes. N.
Diabetes Care. 2016;39(5):686–693. Engl. J. Med. 2019;381:603–613.
32. Lachin JM, Bebu I, Bergenstal RM et al. Association of glyce- 45. Aronson R, Gottlieb PA, Christiansen JS et al. Low-dose ote-
mic variability in type 1 diabetes with progression of micro- lixizumab anti- CD3 monoclonal antibody DEFEND- 1
vascular outcomes in the diabetes control and complications study: Results of the randomized phase III study in recent-
trial. Diabetes Care. 2017;40:777–783. onset human type 1 diabetes. Diabetes Care 2014;37:
33. Bone RN and Evans-Molina C. Combination immunotherapy 2746–2754.
for type 1 diabetes. Current Diabetes Reports. 2017;17:50 . 46. Pescovitz MD, Greenbaum CJ, Bundy B et al. B-lymphocyte
34. Atkinson MA, Von Herrath M, Powers AC, and Clare- depletion with rituximab and β- cell function: Two- year
Salzler M. Current concepts on the pathogenesis of type 1 results. Diabetes Care. 2014;37:453–459.
diabetes-considerations for attempts to prevent and reverse 47. Orban T, Bundy B, Becker DJ et al. Co-stimulation modula-
the disease. Diabetes Care. 2015;38:979–988. tion with abatacept in patients with recent-onset type 1 dia-
35. Rak K and Bronkowska M. Immunomodulatory effect of betes: A randomised double-masked controlled trial. Lancet.
vitamin D and its potential role in the prevention and treat- 2011;378:412–419.
ment of type 1 diabetes mellitus: A narrative review. 48. Ludvigsson J, Krisky D, Casas R et al. GAD65 antigen ther-
Molecules. 2018;24:24. apy in recently diagnosed type 1 diabetes mellitus. N. Engl. J.
36. Gabbay MAL, Sato MN, Finazzo C et al. Effect of cholecal- Med. 2012;366:433–442.
ciferol as adjunctive therapy with insulin on protective 49. Ludvigsson J. Autoantigen treatment in type 1 diabetes:
immunologic profile and decline of residual β-cell function Unsolved questions on how to select autoantigen and
in new-onset type 1 diabetes mellitus. Arch. Pediatr. Adolesc. administration route. International Journal of Molecular
Med. 2012;166:601–607. Sciences. 2020;21.
37. Bogdanou D, Penna- Martinez M, Filmann N et al. 50. Elding Larsson H, Lundgren M, Jonsdottir B et al. Safety and
T-lymphocyte and glycemic status after vitamin D treatment efficacy of autoantigen- specific therapy with 2 doses of
in type 1 diabetes: A randomized controlled trial with alum-formulated glutamate decarboxylase in children with
sequential crossover. Diabetes. Metab. Res. Rev. 2017;33. multiple islet autoantibodies and risk for type 1 diabetes: A
38. Treiber G, Prietl B, Frohlich-Reiterer E et al. Cholecalciferol randomized clinical trial. Pediatr. Diabetes. 2018;19:
supplementation improves suppressive capacity of regula- 410–419.
tory T-cells in young patients with new-onset type 1 diabetes 51. Lopes M, Kutlu B, Miani M et al. Temporal profiling of
mellitus: A randomized clinical trial. Clin. Immunol. cytokine-induced genes in pancreatic β- cells by meta-
2015;161:217–224. analysis and network inference. Genomics. 2014;103:
39. Ataie-Jafari A, Loke SC, Rahmat AB et al. A randomized 264–275.
placebo-controlled trial of alphacalcidol on the preservation 52. Sota J, Vitale A, Insalaco A et al. Safety profile of the inter-
of beta cell function in children with recent onset type 1 dia- leukin-1 inhibitors anakinra and canakinumab in real-life
betes. Clin. Nutr. 2013;32:911–917. clinical practice: a nationwide multicenter retrospective
40. Pitocco D, Crino A, Di Stasio E et al. The effects of calcitriol observational study. Clin. Rheumatol. 2018;37:
and nicotinamide on residual pancreatic β-cell function in 2233–2240.
patients with recent-onset Type 1 diabetes (IMDIAB XI). 53. Cabello-Olmo M, Araña M, Radichev I et al. New insights
Diabet. Med. 2006;23:920–923. into immunotherapy strategies for treating autoimmune
41. Bizzarri C, Pitocco D, Napoli N et al. No protective effect of diabetes. International Journal of Molecular Sciences.
calcitriol on β-cell function in recent-onset type 1 diabetes: 2019;20.
The IMDIAB XIII trial. Diabetes Care. 2010;33:1962–1963. 54. Moran A, Bundy B, Becker DJ et al. Interleukin-1 antago-
42. Herold KC, Hagopian W, Auger JA et al. Anti-CD3 mono- nism in type 1 diabetes of recent onset: Two multicentre,
clonal antibody in new-onset type 1 diabetes mellitus. N. randomised, double-blind, placebo-controlled trials. Lancet.
Engl. J. Med. 2002;346:1692–1698. 2013;381:1905–1915.
55. Bundy BN and Krischer JP. A quantitative measure of treat- 57. Zhang Z, Yan X, Wu C et al. Adding vitamin D3 to the dipep-
ment response in recent-onset type 1 diabetes. Endocrinol. tidyl peptidase-4 inhibitor saxagliptin has the potential to pro-
Diabetes Metab. 2020;3:e00143. tect β-cell function in LADA patients: A 1-year pilot study.
56. Buzzetti R, Pozzilli P, Frederich R et al. Saxagliptin improves Diabetes. Metab. Res. Rev. 2020;36:e3298.
glycaemic control and C-peptide secretion in latent autoim-
mune diabetes in adults (LADA). Diabetes. Metab. Res. Rev.
2016;32:289–296.
3 Reclassifying or Declassifying Diabetes?
Can Clinical Characteristics Guide
Classification and Treatment?
Adrian Vella
Professor of Medicine, Division of Endocrinology, Diabetes, Metabolism, and Nutrition, Mayo Clinic, Rochester, MN, USA
LE A R NI NG P OI NT S of diabetes would require documenting an immune infil

trate within islets [2]. This is extremely invasive, has a sig
●● Type 2 diabetes is differentiated from type 1 diabetes
nificant risk of serious complications and for these reasons
by what it is not i.e. a lack of evidence of immune‐ is almost never undertaken (nor should it).
mediated destruction of insulin‐secreting cells. This Several authors have argued persuasively that if the ulti
oversimplification may miss some of the heterogeneity mate goal of diabetes management is to achieve glycemic
present in type 2 diabetes.
control safely and effectively, then the underlying diagnosis
●● Current methods of differentiating type 1 from type 2
diabetes have significant limitations and lack sensitivity
may be less important [1]. In the absence of therapeutic
and specificity. choices this may be less important but given the prolifera
●● Glucose‐ and lipo‐toxicity can adversely affect insulin tion of therapeutic classes in diabetes pharmacotherapy a
secretion in the short term although this state of affairs greater understanding of the pathophysiologic abnormali
is not necessarily permanent.
ties resulting in hyperglycemia may help optimize therapy.
●● Attention to a possible diagnosis of monogenic
diabetes and achieving glycemic control safely and
This is perhaps best illustrated by the demonstration that
effectively are more likely to be clinically relevant than several forms of monogenic diabetes respond to sulfonylu
current attempts to sub‐classify type 2 diabetes. reas and do not necessarily need insulin therapy [3]. Early
detection and treatment of hemochromatosis may prevent
loss of pancreatic islet function and diabetes associated
with generalized lipodystrophy responding to leptin ther
The pathophysiology of hyperglycemia –
apy serve as other examples.
type 1 vs type 2 diabetes
How does hyperglycemia arise? In the postprandial state
It is important to appreciate that the classification of type 1 glucose concentrations are the net result of stimulation of
and type 2 diabetes was originally somewhat arbitrary and insulin secretion, suppression of glucagon secretion, the
has now evolved into a “positive” diagnosis of type 1 diabe ability of insulin (insulin action) and glucose (glucose
tes – based on the presence of autoantibodies – and a diag effectiveness) to suppress endogenous glucose production
nosis of type 2 diabetes “by exclusion” i.e. no evidence of and stimulate glucose uptake by the tissues, and the rate of
islet autoimmunity. However, this approach has significant gastric emptying. In type 1 diabetes, all defects are ulti
limitations driven in part by the nature of the tests used mately secondary to insulin deficiency whereas in type 2
and by the heterogeneity of type 2 diabetes [1]. Of course, diabetes the relative contribution of these parameters is
the ultimate demonstration of an immune‐mediated cause more variable [4, 5].

37
Defective and/or delayed insulin secretion is the hall system regulates satiety and, to a lesser extent, caloric
mark of all forms of diabetes and the presence of hypergly intake, there has been no clear association with predisposi
cemia implies that the degree of insulin secretion is tion to diabetes. For example, most of the benefits of bari
inadequate for the prevailing insulin action [6]. Increasing atric surgery occur through changes in caloric intake and
secretory demands may overwhelm the ability of β‐cells to weight loss [21] and are likely independent of the rate of
assemble insulin leading to protein misfolding and a cas gastric emptying – outcomes after sleeve gastrectomy and
cade of mechanisms, including the unfolded protein Roux‐en‐Y‐Gastric Bypass (RYGB) are broadly compara
response, that result in dysfunction and if unchecked the ble – despite a far higher rate of gastric (pouch) emptying
death of the β‐cell [7, 8]. Common genetic variation asso after RYGB [22].
ciated with type 2 diabetes has helped to identify multiple
pathways associated with the synthesis and secretion of
Genetic and environmental influences
insulin. It is important to note, however, that the genetic
on disease presentation
predisposition to fasting hyperglycemia differs from that
predisposing to glucose intolerance [9, 10]. This is con Having somewhat downplayed the notion that there are
cordant with the observation that impaired fasting glucose two discrete forms of diabetes, one must acknowledge that
can occur independently of impaired glucose tolerance and the extremes of the disease spectrum are fairly easy to rec
vice‐versa in prediabetes [11]. ognize and seem to differ in their genetic predisposition.
Abnormalities of glucagon suppression are observed in Both type 1 and type 2 exhibit genetic predisposition to the
both type 1 and type 2 diabetes and were initially attributed disease but the environmental contribution to type 2 dia
to insulin deficiency within the islet [12]. However, insulin betes is much more prominent [23].
restraint of α‐cell secretion may not be as important as pre Common genetic variation in more than 200 loci has
viously thought [13, 14]. Certainly there are other parac been associated with type 2 diabetes. The risk conferred by
rine regulators of glucagon secretion [15]. Abnormal most of these variants is small – indeed knowledge of
glucagon secretion arises early in prediabetes and occurs genetic variation at the 18 loci with greatest effect on dis
independently of defects in insulin secretion [14]. ease risk did not appreciably alter the performance of a pre
Underlining its importance in the pathogenesis of diabetes diction model utilizing anthropometric information and
is the observation that people with diabetes‐associated family history [24]. Although later models incorporating
genetic variation exhibit defects of α‐cell function [16, 17]. genetic information from additional loci improved their
The factors altering the ability of insulin and of glucose predictive performance, especially in younger adults, it
itself to suppress endogenous glucose production (and remains apparent that genetic information is unhelpful in
release into the circulation) as well as stimulate uptake are predicting type 2 diabetes risk at an individual level [25].
less well understood although weight, adiposity and physi In contrast, genetic variation in the human leukocyte
cal activity all influence these parameters. Genetic predis antigen (HLA) confers more than half of the genetic risk of
position to defects in insulin action, for example, is less type 1 diabetes. HLA binds and processes antigen‐presen
well characterized. Certain syndromes such as polycystic tation to the immune system. A handful of other loci
ovarian syndrome are associated with diabetes through involved in immune response pathways confer significant
effects on insulin resistance [18, 19]. additional risk. Other variants (~50) also contribute
The final variable affecting glycemic control is upper smaller effects. Most of the loci associated with type 1 dia
gastrointestinal function. The stomach and proximal small betes alter immune regulation [26–28].
bowel function in unison to dampen fluctuations in the The environmental events, if any, that trigger the cas
rate of appearance of ingested calories into the duodenum cade of immune‐mediated destruction in type 1 diabetes
and jejunum after meals. Multiple neural and hormonal are not well defined. Often hyperglycemia is detected for
inputs regulate gastric volume and wall tension (allowing the first time in the context of a febrile illness, but it is well
accommodation of ingested food), pyloric tone and the accepted that significant islet destruction has occurred by
rate of gastric emptying [20]. Although this integrated the time hyperglycemia develops. In contrast, excess caloric
Reclassifying or Declassifying Diabetes? 39
intake and physical inactivity are associated with type 2 situations similar to everyday life e.g. an oral glucose toler
diabetes. However, this is not uniform, so that of all patients ance test or a mixed meal tolerance test may be better at
coming to bariatric surgery in the United States (BMI > detecting subtle defects in insulin secretion as compared to
35 kg/m2) only a third have diabetes. The reasons for this supraphysiologic stimuli. The latter use stimuli intended to
heterogeneity are unclear at present [29]. Another interest produce submaximal insulin secretion e.g. glucagon or
ing observation is the association of shift work and circa arginine stimulation tests [36–38]. Similarly, an intrave
dian misalignment with type 2 diabetes – this may be nous bolus of glucose that can change peripheral concen
mediated through effects on insulin synthesis and secre trations of glucose very rapidly (as part of an intravenous
tion. Genetic variation in the melatonin receptor glucose tolerance test – IVGTT) is also a supraphysiologic
(MTNR1B) – a key part of circadian signaling – is associ stimulus to β‐cell secretion. Intravenous glucose chal
ated with type 2 diabetes [30, 31]. lenges produce a characteristic biphasic pattern of insu
lin secretion in healthy subjects that is not observed in
people with early type 1 diabetes or in people with type 2
Measuring insulin secretion
diabetes [39].
Insulin is secreted by the β‐cell in response to various stim While isolated measurement of insulin secretion in
uli. However, peripheral insulin concentrations may be an response to a standardized stimulus might provide a quali
imperfect measure of insulin secretion because they repre tative measurement of β‐cell function, truly quantitative
sent the net sum of two opposing processes. The first is measures require expressing insulin secretion as a function
actual insulin secretion into the portal vein while the sec of insulin action. Such indices generated through mathe
ond is hepatic extraction of insulin that occurs across the matical modelling can predict risk of progression to diabe
liver as insulin reaches the systemic circulation. It is uncer tes or quantify response to an intervention. However, at
tain whether hepatic extraction is an active or a passive present they have limited utility on an individual basis due
process, but it is likely proportional to the magnitude of to lack of sufficient normative data [6]. Nevertheless,
insulin secretion [32] and declines as β‐cell function absent or near absent C‐peptide concentrations in the fast
declines [33]. ing state or in response to stimuli may help demonstrate an
C‐peptide arises from the post‐translational processing absence of endogenous insulin secretion.
of insulin as preproinsulin which folds upon itself to form
specific disulfide bonds, resulting in a dimeric structure
Autoantibodies
after cleavage of the connecting (C‐)peptide. This peptide
is secreted in a 1:1 ratio with insulin but does not undergo Several autoantibodies can be used in clinical practice to
hepatic extraction. Therefore, in theory, C‐peptide concen document the presence of islet autoimmunity. These are
trations serve as a better measure of insulin secretion than antibodies directed against Glutamic Acid Decarboxylase
do insulin concentrations themselves. However, C‐peptide (GAD‐65), insulin, tyrosine phosphatases, insulinoma‐
which is cleared by the kidney (and therefore cannot be associated protein 2 (IA‐2 and IA‐2β), and the zinc trans
used reliably in renal dysfunction or failure) has a half‐life porter (ZnT8). The greater the number of autoantibodies
of ~30 minutes and therefore accumulates in the circula that are positive and the higher the titer of these antibodies,
tion compared to insulin (half‐life of < 5minutes). the more likely that islet autoimmunity is present [40].
Deconvolution of insulin secretion rates from C‐peptide However, people with a clinical presentation compatible
concentrations requires knowledge of the kinetics under with type 1 diabetes are often antibody negative. The oppo
ling C‐peptide clearance. This can be estimated from site sometimes applies. Indeed, in a cohort of Belgian
anthropometric criteria thanks to the work of Van Cauter patients with type 1 diabetes presenting before age 40, 24%
et al. so that insulin secretion can be measured accurately did not have autoantibodies [41]. In the Botnia study, 4.4%
in humans with intact renal function [34, 35]. of healthy individuals had GAD antibodies [42].
The other factor when assessing insulin secretion is the Latent autoimmune diabetes of adults (LADA) likely
nature of the stimulus – a physiological stimulus mimicking represents a “forme fruste” of type 1 diabetes where a lower
genetic burden of predisposition results in a later age at glycemic control seems to prolong the persistence of
presentation, decreased amount and degree of autoanti endogenous insulin secretion in affected patients.
body positivity, and a longer persistence of endogenous
insulin secretion. Patients with LADA tend to have a lower
Cystic fibrosis‐related diabetes
BMI than patients with type 2 diabetes. Of note, antibodies
against insulin and IA‐2 are rarely present in late‐onset dia Diabetes is the most common extra‐pulmonary complica
betes. Islet autoantibodies are most useful in screening tion of cystic fibrosis and rises with age so that 50% of
family members of patients with type 1 diabetes for risk of affected patients over age 30 have diabetes [51]. There is a
developing the disease [43]. recognized abnormality in insulin secretion which is likely
explained by the actions of the Cystic Fibrosis
Transmembrane Regulator (CFTR) on β‐cell electrophysi
“Atypical” diabetes, glucose toxicity
ology. Drugs that improve CFTR function also have salu
and the honeymoon period
tary effects on insulin secretion [52]. Exocrine pancreatic
Atypical diabetes or “ketosis‐prone type 2 diabetes” were insufficiency may decrease intraluminal release of nutri
monikers used to describe obese minority (typically ents that stimulate incretin hormone secretion and delay
African‐American) patients presenting with diabetic gastric emptying thereby contributing to postprandial
ketoacidosis as their first manifestation of diabetes [44]. hyperglycemia. Endocrine dysfunction is also strongly cor
Typically, after acute metabolic control is achieved, such related with destruction of the exocrine pancreas.
patients are able to maintain glycemic control over long Nutritional deficiency and also the systemic inflamma
periods of time without insulin (often with diet alone). The tion associated with cystic fibrosis likely impair insulin
concept of glucose toxicity where hyperglycemia per se action which is also a feature of Cystic Fibrosis‐related
impairs insulin secretion and action has been put forward Diabetes. Macrovascular complications are uncommon
to explain these observations [45]. Indeed, during recov and microvascular complications occur at lower rates than
ery, these indices do not differ from those of age‐ and they do in “classical” type 1 and type 2 diabetes [53].
weight‐matched controls [46]. Similar short‐term expo
sure to lipotoxicity also has little effect on β‐cell
Novel approaches to reclassifying
function [47].
diabetes
Short‐term intensive insulin therapy in a cohort of
Chinese patients with a first presentation of diabetes pro Ahlqvist et al. undertook a data‐driven cluster analysis in
duced a “remission” of diabetes with preservation of insu patients with newly diagnosed diabetes from a Swedish
lin secretion at one year when compared to conventionally cohort. They utilized six variables measured at the time of
treated patients [48]. Diabetes remission, independent of diagnosis – age at diagnosis, BMI, HbA1c, GAD antibodies,
weight‐loss, is also reported after bariatric surgery but – at and homoeostatic model assessment (HOMA) estimates of
least in the short‐term is likely explained by caloric restric β‐cell function and insulin resistance. The investigators
tion and β‐cell “rest” [49]. This is in keeping with the con identified five clusters based on these variables – one clus
cept that insulin over‐production may produce ter was antibody positive, and relatively lean with poor
endoplasmic reticulum stress and if unchecked lead to β‐ metabolic control. The second cluster differed from the
cell death [50]. first in that antibodies were absent, while the third cluster
These factors probably explain the resumption of had an elevated BMI and insulin resistance. Cluster 4 was
endogenous insulin secretion and decreased or absent also obese but defects in insulin action as measured by
exogenous insulin requirements during the honeymoon HOMA were not as severe as in the 3rd cluster. Finally,
period in type 1 diabetes. The term refers to a variable patients in the 5th cluster were older but, like cluster 4, had
interval after presentation with hyperglycemia and subse less severe metabolic abnormalities. These clusters were
quent restoration of metabolic and glycemic control. Some subsequently validated in 3 independent cohorts. Over sub
patients have a prolonged remission and indeed tight sequent follow‐up, the incidence of diabetic ketoacidosis
and of non‐alcoholic fatty liver disease were highest in the should help improve the care of patients with unclear clas
first and third cluster, respectively. This latter cluster also sification and with associated comorbidities.
had the highest risk of chronic kidney disease [54].
HOMA measures have significant limitations and poor Bibliography
predictive value for progression to diabetes [55] –
1. Gale EA. Declassifying diabetes. Diabetologia.
although in fairness the investigators used C‐peptide
2006;49:1989–1995.
rather than insulin to estimate β‐cell function. Whether
2. Kaestner KH, Powers AC, Naji A, Consortium H, Atkinson
this will help guide therapeutic interventions and improve MA. NIH Initiative to Improve Understanding of the
outcomes in individual patients remains to be Pancreas, Islet, and Autoimmunity in Type 1 Diabetes:
ascertained. The Human Pancreas Analysis Program (HPAP). Diabetes.
Another approach to characterizing different sub‐types 2019;68:1394–1402.
of diabetes is based on common genetic variation and its 3. Hattersley AT, Pearson ER. Minireview: pharmacogenetics
influence on quantitative traits such as glucose tolerance in and beyond: the interaction of therapeutic response,
response to a standardized challenge. For example, Dimas beta‐cell physiology, and genetics in diabetes.
et al. identified 5 clusters of genetic risk loci that altered Endocrinology. 2006;147:2657–2663.
glycemic traits [56]. In this study examining data from ~58 4. Smushkin G, Vella A. What is type 2 diabetes? Medicine
(Baltimore). 2010;38:597–601.
000 non‐diabetic subjects, one cluster altered insulin sensi
5. Vella A, Camilleri M, Rizza RA. The gastrointestinal tract
tivity. A second cluster altered fasting glucose while a third
and glucose tolerance. Curr Opin Clin Nutr Metab Care.
cluster altered the ratio of proinsulin to insulin in the fast 2004;7:479–484.
ing state. Another cluster was primarily characterized by 6. Cobelli C, Dalla Man C, Toffolo G, Basu R, Vella A, Rizza
defects in post‐challenge insulin secretion. The final clus R. The oral minimal model method. Diabetes.
ter of risk loci did not alter glycemic traits. While this exer 2014;63:1203–1213.
cise has certainly helped understand the effect of genotype 7. Haataja L, Manickam N, Soliman A, Tsai B, Liu M, Arvan P.
on phenotype, the effect size of each risk allele on a given Disulfide mispairing during proinsulin folding in the
phenotypic trait is so small as be unhelpful in terms of pre endoplasmic reticulum. Diabetes. 2016;.
dicting individual clinical behavior. 8. Sun J, Cui J, He Q, Chen Z, Arvan P, Liu M. Proinsulin
Pharmacogenomics has also held promise as a way of misfolding and endoplasmic reticulum stress during the
development and progression of diabetes. Mol Aspects Med
classifying diabetes. This hope was reinforced by the discov
2015;42:105–118.
ery that the target for thiazolidinediones and sulfonylureas
9. Smushkin G, Vella A. Genetics of type 2 diabetes. Curr
are risk loci for type 2 diabetes. Unfortunately, their effect Opin Clin Nutr Metab Care. 2010;13:471–477.
on response to therapy is difficult to discern at an individ 10. Saxena R, Hivert MF, Langenberg C, Tanaka T, Pankow JS,
ual level and genotype at PPARG and KCNJ11 should not Vollenweider P, Lyssenko V, Bouatia‐Naji N, Dupuis J,
alter therapeutic choices in type 2 diabetes [57, 58]. Jackson AU, Kao WH, Li M, Glazer NL, Manning AK, Luan
J, Stringham HM, Prokopenko I, Johnson T, Grarup N,
Boesgaard TW, Lecoeur C, Shrader P, O’Connell J, Ingelsson
Conclusions E, Couper DJ, Rice K, Song K, Andreasen CH, Dina C,
Multiple roads lead to the pathophysiologic defects that Kottgen A, Le Bacquer O, Pattou F, Taneera J, Steinthorsdottir
cause hyperglycemia. While autoimmune diabetes can be V, Rybin D, Ardlie K, Sampson M, Qi L, van Hoek M,
Weedon MN, Aulchenko YS, Voight BF, Grallert H, Balkau
characterized quite readily in many patients, the remaining
B, Bergman RN, Bielinski SJ, Bonnefond A, Bonnycastle LL,
patients are far more heterogeneous. Efforts to reinvent the
Borch‐Johnsen K, Bottcher Y, Brunner E, Buchanan TA,
classification, at least to date, have not appreciably changed, Bumpstead SJ, Cavalcanti‐Proenca C, Charpentier G, Chen
or guided, the management of individual patients – the YD, Chines PS, Collins FS, Cornelis M, G JC, Delplanque J,
notable exception being monogenic forms of diabetes. Doney A, Egan JM, Erdos MR, Firmann M, Forouhi NG,
Nevertheless, appreciating the underlying defects as well as Fox CS, Goodarzi MO, Graessler J, Hingorani A, Isomaa B,
the limitations of autoantibody and β‐cell function testing Jorgensen T, Kivimaki M, Kovacs P, Krohn K, Kumari M,
Lauritzen T, Levy‐Marchal C, Mayor V, McAteer JB, Meyre 18. Corbould A, Kim YB, Youngren JF, Pender C, Kahn BB, Lee
D, Mitchell BD, Mohlke KL, Morken MA, Narisu N, Palmer A, Dunaif A. Insulin resistance in the skeletal muscle of
CN, Pakyz R, Pascoe L, Payne F, Pearson D, Rathmann W, women with PCOS involves intrinsic and acquired defects in
Sandbaek A, Sayer AA, Scott LJ, Sharp SJ, Sijbrands E, insulin signaling. American Journal of Physiology Endocrinology
Singleton A, Siscovick DS, Smith NL, Sparso T, Swift AJ, and Metabolism. 2005;288:E1047–1054.
Syddall H, Thorleifsson G, Tonjes A, Tuomi T, Tuomilehto J, 19. Urbanek M, Legro RS, Driscoll D, Strauss JF, 3rd, Dunaif A,
Valle TT, Waeber G, Walley A, Waterworth DM, Zeggini E, Spielman RS. Searching for the polycystic ovary syndrome
Zhao JH, Illig T, Wichmann HE, Wilson JF, van Duijn C, Hu genes. J Pediatr Endocrinol Metab. 2000;13 Suppl 5:1311–1313.
FB, Morris AD, Frayling TM, Hattersley AT, Thorsteinsdottir 20. Vella A, Camilleri M. The gastrointestinal tract as an inte
U, Stefansson K, Nilsson P, Syvanen AC, Shuldiner AR, grator of mechanical and hormonal response to nutrient
Walker M, Bornstein SR, Schwarz P, Williams GH, Nathan ingestion. Diabetes. 2017;66:2729–2737.
DM, Kuusisto J, Laakso M, Cooper C, Marmot M, Ferrucci 21. Ikramuddin S, Korner J, Lee WJ, Thomas AJ, Connett JE,
L, Mooser V, Stumvoll M, Loos RJ, Altshuler D, Psaty BM, Bantle JP, Leslie DB, Wang Q, Inabnet WB, 3rd, Jeffery RW,
Rotter JI, Boerwinkle E, Hansen T, Pedersen O, Florez JC, Chong K, Chuang LM, Jensen MD, Vella A, Ahmed L, Belani
McCarthy MI, Boehnke M, Barroso I, Sladek R, Froguel P, K, Billington CJ. Lifestyle intervention and medical manage
Meigs JB, Groop L, Wareham NJ, Watanabe RM. Genetic vari ment with vs without Roux‐en‐Y Gastric Bypass and control
ation in GIPR influences the glucose and insulin responses to of hemoglobin A1c, LDL cholesterol, and systolic blood
an oral glucose challenge. Nat Genet. 2010;42:142–148. pressure at 5 years in the Diabetes Surgery Study. JAMA.
11. Bock G, Dalla Man C, Campioni M, Chittilapilly E, Basu R, 2018;319:266–278.
Toffolo G, Cobelli C, Rizza R. Pathogenesis of pre‐diabetes: 22. Schauer PR, Bhatt DL, Kirwan JP, Wolski K, Brethauer SA,
mechanisms of fasting and postprandial hyperglycemia in Navaneethan SD, Aminian A, Pothier CE, Kim ES, Nissen
people with impaired fasting glucose and/or impaired glu SE, Kashyap SR. Bariatric surgery versus intensive medical
cose tolerance. Diabetes. 2006;55:3536–3549. therapy for diabetes – 3‐year outcomes. The N Engl J Med.
12. Unger RH, Orci L. The essential role of glucagon in the 2014;370:2002–2013.
pathogenesis of diabetes mellitus. Lancet. 1975;1:14–16. 23. Florez JC, Hirschhorn J, Altshuler D. The inherited basis of
13. Faerch K, Vistisen D, Pacini G, Torekov SS, Johansen NB, diabetes mellitus: implications for the genetic analysis of com
Witte DR, Jonsson A, Pedersen O, Hansen T, Lauritzen T, plex traits. Annu Rev Genomics Hum Genet. 2003;4:257–291.
Jorgensen ME, Ahren B, Holst JJ. Insulin resistance is 24. Meigs JB, Shrader P, Sullivan LM, McAteer JB, Fox CS,
accompanied by increased fasting glucagon and delayed Dupuis J, Manning AK, Florez JC, Wilson PW, D’Agostino
glucagon suppression in individuals with normal and RB, Sr., Cupples LA. Genotype score in addition to common
impaired glucose regulation. Diabetes. 2016;65:3473–3481. risk factors for prediction of type 2 diabetes. N Engl J Med.
14. Sharma A, Varghese RT, Shah M, Man CD, Cobelli C, Rizza 2008;359:2208–2219.
RA, Bailey KR, Vella A. Impaired insulin action is associated 25. Khera AV, Chaffin M, Aragam KG, Haas ME, Roselli C, Choi
with increased glucagon concentrations in non‐diabetic SH, Natarajan P, Lander ES, Lubitz SA, Ellinor PT, Kathiresan
humans. J Clin Endocrinol Metab. 2018;103:314–319. S. Genome‐wide polygenic scores for common diseases
15. Ashcroft FM, Rorsman P. K(ATP) channels and islet hor identify individuals with risk equivalent to monogenic
mone secretion: new insights and controversies. Nature mutations. Nat Genet. 2018;50:1219–1224.
Reviews. 2013;9:660–669. 26. Lowe CE, Cooper JD, Brusko T, Walker NM, Smyth DJ,
16. Karanth S, Adams JD, Serrano MLA, Quittner‐Strom EB, Bailey R, Bourget K, Plagnol V, Field S, Atkinson M, Clayton
Simcox J, Villanueva CJ, Ozcan L, Holland WL, Yost HJ, DG, Wicker LS, Todd JA. Large‐scale genetic fine mapping
Vella A, Schlegel A. A hepatocyte FOXN3‐alpha cell gluca and genotype‐phenotype associations implicate polymor
gon axis regulates fasting glucose. Cell Rep. 2018;24: phism in the IL2RA region in type 1 diabetes. Nat Genet.
312–319. 2007;39:1074–1082.
17. Shah M, Varghese RT, Miles JM, Piccinini F, Dalla Man C, 27. Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K,
Cobelli C, Bailey KR, Rizza RA, Vella A. TCF7L2 Genotype Plagnol V, Bailey R, Nejentsev S, Field SF, Payne F, Lowe CE,
and alpha‐cell function in humans without diabetes. Szeszko JS, Hafler JP, Zeitels L, Yang JH, Vella A, Nutland S,
Diabetes. 2016;65:371–380. Stevens HE, Schuilenburg H, Coleman G, Maisuria M,
Meadows W, Smink LJ, Healy B, Burren OS, Lam AA, Performance of individually measured vs population‐based
Ovington NR, Allen J, Adlem E, Leung HT, Wallace C, C‐peptide kinetics to assess beta‐cell function in the pres
Howson JM, Guja C, Ionescu‐Tirgoviste C, Simmonds MJ, ence and absence of acute insulin resistance. Diabetes Obes
Heward JM, Gough SC, Dunger DB, Wicker LS, Clayton Metab. 2018;20:549–555.
DG. Robust associations of four new chromosome regions 35. Van Cauter E, Mestrez F, Sturis J, Polonsky KS. Estimation of
from genome‐wide analyses of type 1 diabetes. Nat Genet. insulin secretion rates from C‐peptide levels. Comparison of
2007;39:857–864. individual and standard kinetic parameters for C‐peptide
28. Smyth DJ, Howson JM, Payne F, Maier LM, Bailey R, clearance. Diabetes. 1992;41:368–377.
Holland K, Lowe CE, Cooper JD, Hulme JS, Vella A, 36. Robertson RP, Bogachus LD, Oseid E, Parazzoli S, Patti ME,
Dahlman I, Lam AC, Nutland S, Walker NM, Twells RC, Rickels MR, Schuetz C, Dunn T, Pruett T, Balamurugan AN,
Todd JA. Analysis of polymorphisms in 16 genes in type 1 Sutherland DE, Beilman G, Bellin MD. Assessment of beta‐
diabetes that have been associated with other immune‐ cell mass and alpha‐ and beta‐cell survival and function by
mediated diseases. BMC Med Genet. 2006;7:20. arginine stimulation in human autologous islet recipients.
29. Ikramuddin S, Billington CJ, Lee WJ, Bantle JP, Thomas AJ, Diabetes. 2015;64:565–572.
Connett JE, Leslie DB, Inabnet WB, 3rd, Jeffery RW, Chong 37. Robertson RP, Raymond RH, Lee DS, Calle RA, Ghosh A,
K, Chuang LM, Sarr MG, Jensen MD, Vella A, Ahmed L, Savage PJ, Shankar SS, Vassileva MT, Weir GC, Fryburg DA,
Belani K, Schone JL, Olofson AE, Bainbridge HA, Laqua PS, Beta cell project team of the foundation for the NIHBC.
Wang Q, Korner J. Roux‐en‐Y gastric bypass for diabetes Arginine is preferred to glucagon for stimulation testing of
(the Diabetes Surgery Study): 2‐year outcomes of a 5‐year, beta‐cell function. Am J Physiol Endocrinol Metab.
randomised, controlled trial. Lancet Diabetes Endocrinol. 2014;307:E720–727.
2015;3:413–422. 38. Shankar SS, Vella A, Raymond RH, Staten MA, Calle RA,
30. Sharma A, Laurenti MC, Dalla Man C, Varghese RT, Cobelli Bergman RN, Cao C, Chen D, Cobelli C, Dalla Man C, Deeg
C, Rizza RA, Matveyenko A, Vella A. Glucose metabolism M, Dong JQ, Lee DS, Polidori D, Robertson RP, Ruetten H,
during rotational shift‐work in healthcare workers. Stefanovski D, Vassileva MT, Weir GC, Fryburg DA,
Diabetologia. 2017;60:1483–1490. Foundation for the National Institutes of Health beta‐Cell
31. Bouatia‐Naji N, Bonnefond A, Cavalcanti‐Proenca C, Project T. Standardized mixed‐meal tolerance and arginine
Sparso T, Holmkvist J, Marchand M, Delplanque J, Lobbens stimulation tests provide reproducible and complementary
S, Rocheleau G, Durand E, De Graeve F, Chevre JC, Borch‐ measures of beta‐cell function: results from the foundation
Johnsen K, Hartikainen AL, Ruokonen A, Tichet J, Marre M, for the National Institutes of Health Biomarkers Consortium
Weill J, Heude B, Tauber M, Lemaire K, Schuit F, Elliott P, Investigative Series. Diabetes Care. 2016;39:1602–1613.
Jorgensen T, Charpentier G, Hadjadj S, Cauchi S, Vaxillaire 39. Bergman RN. Lilly lecture 1989. Toward physiological
M, Sladek R, Visvikis‐Siest S, Balkau B, Levy‐Marchal C, understanding of glucose tolerance. Minimal‐model
Pattou F, Meyre D, Blakemore AI, Jarvelin MR, Walley AJ, approach. Diabetes. 1989;38:1512–1527.
Hansen T, Dina C, Pedersen O, Froguel P. A variant near 40. Skyler JS, Bakris GL, Bonifacio E, Darsow T, Eckel RH,
MTNR1B is associated with increased fasting plasma glu Groop L, Groop PH, Handelsman Y, Insel RA, Mathieu C,
cose levels and type 2 diabetes risk. Nat Genet. 2009;41: McElvaine AT, Palmer JP, Pugliese A, Schatz DA, Sosenko
89–94. JM, Wilding JP, Ratner RE. Differentiation of diabetes by
32. Meier JJ, Veldhuis JD, Butler PC. Pulsatile insulin secretion pathophysiolog, natural history, and prognosis. Diabetes.
dictates systemic insulin delivery by regulating hepatic insu 2017;66:241–255.
lin extraction in humans. Diabetes. 2005;54:1649–1656. 41. Decochez K, Tits J, Coolens JL, Van Gaal L, Krzentowski G,
33. Sathananthan A, Man CD, Zinsmeister AR, Camilleri M, Winnock F, Anckaert E, Weets I, Pipeleers DG, Gorus FK.
Rodeheffer RJ, Toffolo G, Cobelli C, Rizza RA, Vella A. A High frequency of persisting or increasing islet‐specific
concerted decline in insulin secretion and action occurs autoantibody levels after diagnosis of type 1 diabetes pre
across the spectrum of fasting and postchallenge glucose senting before 40 years of age. The Belgian Diabetes Registry.
concentrations. Clin Endocrinol (Oxf). 2012;76:212–219. Diabetes Care. 2000;23:838–844.
34. Varghese RT, Dalla Man C, Laurenti MC, Piccinini F, Sharma 42. Tuomi T, Carlsson A, Li H, Isomaa B, Miettinen A, Nilsson
A, Shah M, Bailey KR, Rizza RA, Cobelli C, Vella A. A, Nissen M, Ehrnstrom BO, Forsen B, Snickars B, Lahti K,
Forsblom C, Saloranta C, Taskinen MR, Groop LC. Clinical 53. Moheet A, Moran A. CF‐related diabetes: containing the
and genetic characteristics of type 2 diabetes with and with metabolic miscreant of cystic fibrosis. Pediatr Pulmonol
out GAD antibodies. Diabetes. 1999;48:150–157. 2017;52:S37–S43.
43. Hummel M, Bonifacio E, Schmid S, Walter M, Knopff A, 54. Ahlqvist E, Storm P, Karajamaki A, Martinell M, Dorkhan
Ziegler AG. Brief communication: early appearance of islet M, Carlsson A, Vikman P, Prasad RB, Aly DM, Almgren P,
autoantibodies predicts childhood type 1 diabetes in offspring Wessman Y, Shaat N, Spegel P, Mulder H, Lindholm E,
of diabetic parents. Ann Intern Med. 2004;140:882–886. Melander O, Hansson O, Malmqvist U, Lernmark A, Lahti
44. Umpierrez GE, Smiley D, Kitabchi AE. Narrative review: K, Forsen T, Tuomi T, Rosengren AH, Groop L. Novel sub
ketosis‐prone type 2 diabetes mellitus. Annals of Internal groups of adult‐onset diabetes and their association with
Medicine. 2006;144:350–357. outcomes: a data‐driven cluster analysis of six variables.
45. Poitout V, Robertson RP. Glucolipotoxicity: fuel excess and Lancet Diabetes Endocrinol. 2018;6:361–369.
beta‐cell dysfunction. Endocr Rev. 2008;29:351–366. 55. Xiang AH, Watanabe RM, Buchanan TA. HOMA and
46. Gosmanov AR, Smiley D, Robalino G, Siqueira JM, Peng L, Matsuda indices of insulin sensitivity: poor correlation with
Kitabchi AE, Umpierrez GE. Effects of intravenous glucose minimal model‐based estimates of insulin sensitivity in lon
load on insulin secretion in patients with ketosis‐prone dia gitudinal settings. Diabetologia. 2014;57:334–338.
betes during near‐normoglycemia remission. Diabetes Care. 56. Dimas AS, Lagou V, Barker A, Knowles JW, Magi R, Hivert
2010;33:854–860. MF, Benazzo A, Rybin D, Jackson AU, Stringham HM, Song
47. Umpierrez GE, Smiley D, Robalino G, Peng L, Gosmanov C, Fischer‐Rosinsky A, Boesgaard TW, Grarup N, Abbasi
AR, Kitabchi AE. Lack of lipotoxicity effect on {beta}‐cell FA, Assimes TL, Hao K, Yang X, Lecoeur C, Barroso I,
dysfunction in ketosis‐prone type 2 diabetes. Diabetes Care. Bonnycastle LL, Bottcher Y, Bumpstead S, Chines PS, Erdos
2010;33:626–631. MR, Graessler J, Kovacs P, Morken MA, Narisu N, Payne F,
48. Weng J, Li Y, Xu W, Shi L, Zhang Q, Zhu D, Hu Y, Zhou Z, Stancakova A, Swift AJ, Tonjes A, Bornstein SR, Cauchi S,
Yan X, Tian H, Ran X, Luo Z, Xian J, Yan L, Li F, Zeng L, Froguel P, Meyre D, Schwarz PE, Haring HU, Smith U,
Chen Y, Yang L, Yan S, Liu J, Li M, Fu Z, Cheng H. Effect of Boehnke M, Bergman RN, Collins FS, Mohlke KL,
intensive insulin therapy on beta‐cell function and glycae Tuomilehto J, Quertemous T, Lind L, Hansen T, Pedersen O,
mic control in patients with newly diagnosed type 2 diabe Walker M, Pfeiffer AF, Spranger J, Stumvoll M, Meigs JB,
tes: a multicentre randomised parallel‐group trial. Lancet. Wareham NJ, Kuusisto J, Laakso M, Langenberg C, Dupuis J,
2008;371:1753–1760. Watanabe RM, Florez JC, Ingelsson E, McCarthy MI,
49. Laedtke T, Kjems L, Porksen N, Schmitz O, Veldhuis J, Kao Prokopenko I, Investigators M. Impact of type 2 diabetes
PC, Butler PC. Overnight inhibition of insulin secretion susceptibility variants on quantitative glycemic traits
restores pulsatility and proinsulin/insulin ratio in type 2 dia reveals mechanistic heterogeneity. Diabetes. 2014;63:
betes. Am J Physiol Endocrinol Metab. 2000;279:E520–528. 2158–2171.
50. Arunagiri A, Haataja L, Cunningham CN, Shrestha N, Tsai 57. Florez JC, Jablonski KA, Kahn SE, Franks PW, Dabelea D,
B, Qi L, Liu M, Arvan P. Misfolded proinsulin in the endo Hamman RF, Knowler WC, Nathan DM, Altshuler D. Type 2
plasmic reticulum during development of beta cell failure in diabetes‐associated missense polymorphisms KCNJ11 E23K
diabetes. Ann N Y Acad Sci. 2018;. and ABCC8 A1369S influence progression to diabetes and
51. Moran A, Dunitz J, Nathan B, Saeed A, Holme B, Thomas W. response to interventions in the Diabetes Prevention
Cystic fibrosis‐related diabetes: current trends in pre Program. Diabetes. 2007;56:531–536.
valence, incidence, and mortality. Diabetes Care. 2009;32: 58. Florez JC, Jablonski KA, Sun MW, Bayley N, Kahn SE,
1626–1631. Shamoon H, Hamman RF, Knowler WC, Nathan DM,
52. Bellin MD, Laguna T, Leschyshyn J, Regelmann W, Dunitz J, Altshuler D. Effects of the type 2 diabetes‐associated PPARG
Billings J, Moran A. Insulin secretion improves in cystic P12A polymorphism on progression to diabetes and
fibrosis following ivacaftor correction of CFTR. a small pilot response to troglitazone. J Clin Endocrinol Metab.
study. Pediatr Diabetes. 2013;14:417–421. 2007;92:1502–1509.
4 How Should Secondary Causes
of Diabetes Be Excluded?
Tomás P. Griffin, Aonghus O’Loughlin, and Sean F. Dinneen
Discipline of Medicine, NUI Galway, Galway, Ireland and Centre for Diabetes, Endocrinology and Metabolism, Galway
University Hospitals, Galway, Ireland
describe a variety of forms of diabetes for which a cause

was clearly identified. Problems arose with this approach
●● Current diabetes clinical practice guidelines do not
because many patients with “non‐insulin‐dependent”
address the issue of population‐wide screening for diabetes ended up being treated with insulin and having
specific types of diabetes due to secondary causes. the very cumbersome label of “insulin‐treated non‐insu
●● Case finding (or opportunistic screening) should be lin‐dependent” diabetes. Current American Diabetes
undertaken among patients with diabetes and clinical Association (ADA) Guidelines recommend categorizing
features of the condition under consideration (on a
case by case basis).
diabetes into four groups: Type 1 diabetes (autoimmune,
●● Serum ferritin and transferrin saturation should be insulin deficient), Type 2 diabetes (insulin resistance,
used to assess iron stores when screening for progressive loss of insulin secretory β‐cell function), ges
hereditary haemochromatosis. tational diabetes (diagnosis of diabetes in an individual
●● When screening for Cushing’s syndrome the following who had no known diagnosis of diabetes prior to preg
tests should be considered two 24‐hour urinary free
cortisol, a 1 mg overnight dexamethasone suppression
nancy), and specific types of diabetes due to other
test, or two late‐night salivary cortisol. causes [3] (Table 4.1). Currently, approximately 463 mil
●● Annual oral glucose tolerance testing is recommended lion adults worldwide aged between 20 and 79 years have
in adolescents and adults with cystic fibrosis. diabetes (9.3% of individuals in this age bracket) [4]. This
●● Baseline measurement of height, weight (with is expected to rise to 578.4 million by 2030 and 700.2 mil
calculation of body mass index), measurement of FPG
and fasting lipid profile should be undertaken before
lion by 2045 [4]. Approximately 90% have type 2 diabetes,
commencing atypical antipsychotic drugs. 5–10% have type 1 diabetes and the remainder are spe
cific types of diabetes due to other causes [5].
In this chapter, we explore the specific types of diabetes
due to other causes (often referred to as secondary diabe
tes). Recognition of these specific types of diabetes is
Introduction
important as diagnosis impacts on clinical care decisions.
In the early 1990s the terminology used to classify diabetes Correct diagnosis reduces the risk of complications such as
mellitus was based mainly on the approach used to treat the diabetic ketoacidosis (DKA), hyperosmolar hyperglycemic
condition; “insulin dependent” or “non‐insulin‐dependent” syndrome (HHS), or the unnecessary initiation of insulin
diabetes. The term “other specific types” was used to therapy (which can have a profound impact on quality of

45
TABLE 4.1 Types of diabetes.
Type of diabetes
Type 1 Diabetes Mellitus Immune‐mediated Endocrinopathies Acromegaly
Idiopathic Cushing’s syndrome

Type 2 Diabetes Mellitus Hyperthyroidism
Gestational Diabetes Mellitus Glucagonoma
Specific types of diabetes due to Phaeochromocytoma
other causes
Genetic defects of β‐cell function MODY3 (HNF‐1α) Somatostatinoma
MODY1 (HNF‐4α) Hyperaldosteronism
MODY2 (GCK) Others
MODY5 (HNF‐1β) Drug or chemical‐induced Immune checkpoint
inhibitors
MODY4 (IPF‐1) Glucocorticoids
MODY6 (ND1) Thyroid hormone
Other rarer forms of MODY Diazoxide
Transient neonatal diabetes Statin
Permanent neonatal diabetes Β‐adrenergic agonists
Mitochondrial diabetes Others
Other Infections Congenital rubella
Genetic defects in insulin action Type A insulin resistance Cytomegalovirus
Leprechaunism Other
Rabson–Mendenhall syndrome Other genetic syndromes Down’s syndrome
Lipoatrophic diabetes Klinefelter syndrome
Others Turner syndrome
Diseases of the exocrine pancreas Pancreatitis Wolfram syndrome
Trauma/pancreatectomy Friedreich ataxia
Neoplasia Myotonic dystrophy
Cystic fibrosis Other
Haemochromatosis
Other
life) particularly those with some types of monogenic dia r ecognition can have a profound effect (therapeutic or oth
betes. There is a lack of consensus among clinicians on erwise) on an individual patient and/or their family
when and how to test for other causes of diabetes. As well members.
as traditional forms of secondary diabetes, we will discuss Specific types of diabetes due to other causes can be
newer genetic syndromes that have been elucidated in identified in many ways. In some cases, patients present
recent years and new therapies (immune checkpoint inhib with a new diagnosis of diabetes and the challenge is to find
itors (ICIs)) which can contribute to the onset of diabetes. the underlying cause. To delve beyond the most common
Although these conditions may not be curable their diagnoses of type 1 and type 2 diabetes can be challenging
How Should Secondary Causes of Diabetes Be Excluded? 47
for a physician due to the vast number of other causes and c omprehensive [6]. They include questions relating to the
the fact that some are exceedingly rare. On the other hand, condition (the condition should be an important public
patients can present with a diagnosis which is known to be health problem, does the natural history of the condition
associated with (cystic fibrosis (CF)) or cause diabetes include progression from latent to declared disease?), the
(post‐pancreatectomy) where the challenge for the physi test (is the test simple, safe, validated and precise? Are there
cian is to diagnose or screen for diabetes. It is worth con appropriate cut‐off levels for the test?), the intervention (is
trasting the challenge of searching for secondary forms of there evidence that early treatment in those who have been
hypertension with that of searching for other causes of dia screened leads to better outcomes than for those who
betes. In a hypertensive cohort, there are certain “red flags” receive usual care?) and the screening program itself (is
that most clinicians are alert to which may indicate an there randomized controlled trial evidence that the screen
underlying secondary cause. These include a young age at ing program itself is cost effective and reduces morbidity or
onset of hypertension, difficult to control blood pressure mortality?) [6]. At present, only 5 of 39 adult conditions
(e.g. requiring ≥ 3 antihypertensive drugs), the presence of have been considered by the UK National Screening
certain clinical (e.g. paroxysmal symptoms) or biochemical Committee (NSC) undergo widespread screening (breast,
(e.g. hypokalemia) markers at first presentation. Unfortuna bowel and cervical cancer, diabetic retinopathy, and
tely, these “red flags” do not easily translate into the diabe abdominal aortic aneurysm [7]). Diabetes itself has not
tes setting. Onset of diabetes at a young age is not unusual been recommended for population‐wide screening by the
and is likely to be classified as type 1 diabetes if the patient UK NSC, but the United States Preventive Services Task
is lean or as type 2 if the patient is overweight. The latter is Force (USPSTF) recommends screening for “abnormal
increasingly recognized as a feature of the obesity epidemic blood glucose as part of cardiovascular risk assessment in
and is especially common among ethnic minority groups. adults aged 40–70 years who are overweight or obese” [8].
Similarly, diabetes that is difficult to control with tablets Although the exercise of searching for other forms of dia
would not typically suggest an underlying cause but rather betes is not, strictly speaking, screening (which is searching
it would lead to the earlier introduction of insulin. The for disease in a healthy population) it is a very similar pro
“failure” of oral anti‐hyperglycemic agents is not unusual cess. Table 4.2 summarizes the main secondary forms of
(albeit due to poor compliance or relative insulin defi diabetes mellitus and screening methods used to detect
ciency) and does not represent a “red flag” for secondary them. Prior to discussing the identification of other forms
causes. It is in the clinical domain that the similarities of diabetes, it is worth reminding ourselves of the current
between approaches to identifying secondary hypertension ADA criteria for diagnosis of diabetes [3]:
and secondary diabetes hold up. Indeed, several endo
●● Fasting (no caloric intake for ≥ 8 h) plasma glucose ≥
crinopathies such as Cushing’s syndrome or acromegaly
7.0 mmol/L.
can lead to both diabetes and hypertension. Because these
●● 2‐h plasma glucose ≥ 11.1 mmol/L during an oral glu
conditions are so rare, clinical acumen is the most impor
cose tolerance test (OGTT) performed in accordance
tant factor in identifying other causes of diabetes. A poten
with WHO criteria.
tial portent for the onset of diabetes is the commencement
●● HbA1C ≥ 48 mmol/mol (carried out in a laboratory
of new therapies such as glucocorticoids or ICIs. In this
using a method that is NGSP certified and standardized
chapter we will address the question “should we be under
to the Diabetes Control and Complications Trial
taking more systematic case finding among our patients
(DCCT) assay).
with diabetes?”
●● In a patient with classic symptoms of hyperglycemia or
Screening for a disease at population level is a complex
hyperglycemic crisis, a random plasma glucose ≥
and expensive endeavor. Public Health England (PHE) for
11.1 mmol/L.
example provide guidance on conditions for which screen
ing programs should or should not be developed. The In the absence of classical clinical symptoms such as
criteria by which the PHE appraises the viability, effec hyperglycemic crisis, classic symptoms of hyperglyce
tiveness and appropriateness of a screening program are mia or random plasma glucose ≥ 11.1 mmol/L diagnosis
TABLE 4.2 Overview of secondary diabetes due to pancreatic disease and hormone excess.
Evidence for benefit Screening for DM

Screening for condition from population‐wide among individuals
Condition among people with DM screening with the condition
Pancreatic Hereditary Serum ferritin ++ OGTT/FG

disease haemochromatosis Transferrin saturation
Genotyping
CFRD N/A + Annual OGTT
Pancreatic cancer Abdominal CT scan +/‐ EUS − OGTT
Hormone Cushing’s 24hour UFC − OGTT

excess 1 mg DST
Late night salivary cortisol
Acromegaly IGF‐1 +/‐ OGTT with growth − OGTT
hormone (70)
Phaeochromocytoma Plasma metanephrines sampled − OGTT
following 30 minutes supine
rest (142‐144)
EUS: Endoscopic Ultrasound; DST: Dexamethasone Suppression Test; OGTT: Oral Glucose Tolerance Test; IGF‐1: Insulin‐like Growth
Factor; FG: Fasting Glucose
Evidence for benefit from population screening: ++ Very Strong Evidence and improvement of dysglycemia with treatment; +
Strong Evidence; − Little Evidence
requires two abnormal test results from the same sample increased and unregulated absorption of iron by entero
or in two separate test samples [3]. cytes in the gastrointestinal tract due a deficiency of hepci
din [9]. HH is typically inherited as an autosomal recessive
trait due to mutations such as C282Y, H63D and S65C in
Hereditary haemochromatosis
the HFE gene. Affected individuals may develop excessive
Hereditary haemochromatosis (HH) is a genetic disorder deposition of iron in the parenchymal cells of certain
characterized by iron overload [9]. It is associated with organs including liver, pancreas, skin and heart. Clinical
Clini c a l Vi g n e tt e :
A 43‐year‐old Caucasian male presented with a 3‐month history of polydipsia, polyuria, 4 kg unintentional weight loss, fatigue,
and lethargy. Random blood glucose was 16 mmol/L. He was diagnosed with haemochromatosis 12 years prior, at which time
his serum ferritin was > 4000 ng/mL and hepatic fibrosis was found on liver biopsy. Regular phlebotomy was initiated but
attendance was erratic in recent years. Medication included methotrexate and folic acid for haemochromatosis‐related arthritis.
The patient drank 24 units of alcohol per week. Normal libido and erectile function with no cardiac symptoms were noted.
On examination he had a bronzed appearance and no stigmata of chronic liver disease. Body mass index was 26 kg/m2
and blood pressure was 135/87 mmHg. Cardiovascular, respiratory, and gastrointestinal system examinations were normal.
Ankles were swollen over the lateral malleoli with scars from previous arthroscopies. Normal peripheral pulses and no loss of
vibration sensation were noted. Biochemical investigations revealed a HbA1c of 133 mmol/mol, normal liver chemistries,
normal alpha‐fetoprotein, serum ferritin 1340 ng/mL, transferrin saturation 97% and normal serum testosterone and
gonadotrophins. He was instructed in self‐monitoring of blood glucose and reported home readings that were consistently
> 15 mmol/L. A multiple daily injection regimen with insulin glargine and aspart was commenced. By attending for regular
venesection, he maintained his serum ferritin < 50 ng/mL
manifestations include deranged liver enzymes, hepatic some of the end‐organ damage particularly if phlebotomy is
fibrosis/cirrhosis, cardiomyopathy, hyperglycemia, diabe initiated early in the course of the disease. Referred to as
tes mellitus and a bronzed appearance on the skin. Joint “bronze diabetes”, the prevalence of diabetes in patients with
pains, fatigue, reduced libido, and erectile dysfunction HH is estimated at between 13–23% [15].
are other potential problems. This vignette illustrates a Diabetes in HH results from β‐cell dysfunction (iron
typical patient who developed diabetes secondary to deposition and hepcidin expression in pancreatic islets)
haemochromatosis. and decreased insulin secretion [16]. In patients with
C282Y HH, there is an association between the diagnosis
Natural history of haemochromatosis of diabetes and a higher fibrosis stage (independent of
Historically, HH often presented with end‐organ damage gender, alcohol, steatosis and hepatic iron overload) [15].
due to increased levels of iron deposition in tissue. Although Interestingly, diabetes itself does not improve with serial
the genetic defect is present from birth HH rarely presents phlebotomy [17]. There is a lack of evidence on the effect
< 40 years [10]. In women, signs and symptoms present later of phlebotomy on preventing the onset of diabetes [17].
as menstruation can reduce iron accumulation prior to men
opause. Factors such as excessive alcohol consumption lead The case for screening
to earlier presentation. Prior to the identification of the HFE HH appears to be a perfect condition for screening: insidi
gene, confirmation of the diagnosis was undertaken by doc ous onset over many years, good biochemical and genetic
umenting excessive iron stores on biochemistry and measur tests available for case detection, a safe treatment associ
ing hepatic or bone marrow iron levels by quantitative ated with benefit. Despite these considerations,
methods. Initial evaluation of suspected cases currently USPSTF [13, 18], UK NSC [19], European Association for
involves laboratory measurement of serum iron, ferritin and the Study of the Liver (EASL) [20] and American
total iron binding capacity and estimation of the transferrin Association for the Study of Liver Diseases [21] do not rec
saturation. Due to ease of access to these laboratory tests, ommend population‐wide screening for HH. The USPSTF
patients are typically asymptomatic at presentation. based this decision on:
Confirmation of the diagnosis involves genetic testing for
1 Full‐blown clinical disease is rare in the general
the commonest mutations (C282Y, H63D, and often S65C).
population.
The frequency of homozygosity for the C282Y mutation
2 Disease penetrance is low among those with a high‐risk
(commonest mutation) is 4.4 per 1000 in Caucasian popula
genotype.
tions in the US. It is less common in Hispanic (0.27 per
3 Lack of evidence that early phlebotomy in the screen‐
1000), African American (0.14 per 1000) and Asian
detected patient provides additional benefit over phle
American (< 0.001 per 1000) populations [11]. Not everyone
botomy in clinically detected patients.
homozygous for the C282Y mutation develops the full‐
4 Potential for harm (anxiety, denial of life insurance, dis
blown clinical syndrome [12]. Among those with C282Y
crimination through health insurance) through labeling
homozygosity 38–50% develop iron overload and only
of homozygous individuals as “diseased” or “high risk”
10–33% develop HH associated morbidity [13].
despite the possibility that they may never develop clini
Historically, untreated HH was a cause of premature mor
cal disease [22].
bidity and mortality (due to liver cirrhosis, diabetes, conges
tive cardiac failure. . .). With the advent of phlebotomy, The UK NSC currently recommends further research to
outcomes depend primarily on degree of iron overload and determine which of those with a genetic susceptibility for
sustained aggressive treatment if indicated. In patients with HH with normal iron stores and which of those with
diagnosed HH (C282Y homozygotes), only those with a genetic susceptibility for HH with iron overload but no evi
serum ferritin> 2000 ug/L had increased mortality (primar dence of end organ damage will eventually develop the full‐
ily as a result of liver disease) [14]. The goal of phlebotomy is blown syndrome. Screening of family members (such as
to achieve a serum ferritin level < 50 ng/ml. This may reverse first‐degree relatives) of patients with a diagnosis of HH is
recommended [23]. Studies have shown that often well as patients with symptoms or manifestations linked to
homozygous relatives have biochemical and clinical evi the clinical disorder. A modified ACG algorithm is repro
dence of disease [24, 25]. duced in Figure 4.1.
The evidence for screening for HH among individuals
with pre‐existing diabetes is lacking. There are conflicting
Cystic fibrosis‐related diabetes
reports as to whether there is an association between dia
betes and a genetic susceptibility to HH. This is due to the Cystic fibrosis (CF) is the most common life shortening
fact that diabetes is prevalent in the general population and autosomal recessive disorder seen in Caucasian popula
genetic susceptibility to HH does not always lead to iron tions. The prevalence of CF among those with Northern
accumulation [17]. In our practice (Ireland has the highest European ancestry is 1 in 2500–3000 live births with a car
genotype prevalence reported), we measure iron stores in rier frequency of 1 in 25 [30]. It results from a defect in the
patients with newly diagnosed diabetes. This is not a popu CF gene which encodes a protein called cystic fibrosis
lation‐wide approach to screening (not all patients with transmembrane regulator (CFTR). The genetic defect leads
newly diagnosed diabetes will be referred to hospital). It is to disordered ion transport which in turn affects the func
common for iron stores to be measured in patients with tion of sweat glands, the exocrine pancreas, the gastrointes
diabetes and another potential manifestation of HH (such tinal and respiratory tracts. The classical clinical picture is
as abnormal liver function tests or a bronzed appearance recurrent pulmonary infections, exocrine pancreatic insuf
on the skin). There is no current formal recommendation ficiency, malnutrition, and premature death. In the past
to screen for HH in patients with newly diagnosed diabe 2–3 decades, improvements in the pulmonary, antimicro
tes [17] and further research is required. When measuring bial, and nutritional management means that patients with
iron stores, it is important to remember that serum ferritin CF are now living into their 3rd, 4th, or 5th decades. Cystic
may be elevated as an acute phase reactant (especially in fibrosis‐related diabetes (CFRD) is the commonest comor
patients with newly diagnosed diabetes) and may not bidity (20% adolescents; 40–50% adults) [31].
reflect iron overload. A combination of serum ferritin and
transferrin saturation is the preferred method of biochemi Natural history of CFRD
cal screening. The primary pathogenetic defect in CFRD is insulin defi
Screening HH patients for diabetes is complicated; ciency resulting from pancreatic exocrine damage and
HbA1c is likely to be underestimated following venesec pancreatic fibrosis. Glucagon deficiency also occurs and
tion. In those undergoing active venesection, we recom may explain why ketoacidosis is seldom seen in CFRD. It is
mend the use of fasting plasma glucose (FPG)/OGTT to uncommon for CFRD to appear in childhood. Varying
diagnosis diabetes. Considering the prevalence of diabetes degrees of insulin insufficiency occur as a result of acute
among those with HH, it is sensible that all patients with and chronic illness [32, 33]. A single‐center North
HH are screened for diabetes at time of diagnosis. American study that undertakes annual OGTT in its
Thereafter, provided they have good HH control, the crite patients with CF reported CFRD rates of 2% in children,
ria for testing for diabetes or prediabetes in asymptomatic 19% in adolescents, 40% in individuals in their 20s, and
adults outlined by the ADA could be adhered to [28]. 45–50% in those aged ≥ 30 years [31]. In the 30–39 years
So where does all of this leave the practicing diabetes age group, women with CFRD outnumbered men but typi
clinician? In the absence of guidance from the ADA, the cally there was no gender difference in prevalence. CFRD is
Clinical Practice Guideline of the American College of associated with reduced survival and poorer prognosis in
Gastroenterology (ACG) provides sensible guidance on females compared to males [33].
screening for HH in routine clinical practice [29]. It
emphasizes the measurement of serum iron, transferrin A comparison with other forms of diabetes
saturation and serum ferritin as a first step and targets rela Although CFRD shares many similarities with Type 1 and
tives of individuals known to have the genetic disorder, as Type 2 diabetes, it represents a distinct form of diabetes.
Target Population
Adult 1st degree

Step 1 symptomatic asymptomatic relative of HH
Fasting transferrin saturation

and serum ferritin
TS <45% and TS >45% and ferritin

normal ferritin elevated
No further iron Genotype

Step 2
evaluation
C282Y/C282Y
Compound heterozygote
C282Y/H63D
Heterozygote C282Y or Age < 40 years Age > 40 years
Ferritin<1000 and/or Ferritin
non-C282Y
and Normal >1000
or Elevated
ALT/AST ALT/AST
Exclude other liver or
Step 3 hematologic disease
Therapeutic Liver biopsy for HIC
± Liver biopsy Phlebotomy and histopathology
FIG 4.1 Proposed algorithm for screening and treatment of haemochromatosis in patients with diabetes [23, 29]. TS: Transferrin
Saturation; HIC: Hepatic Iron Concentration.
The presence of CF does not out rule the possibility of type type 2 diabetes, this is not the case in CFRD. Instead a more
1 diabetes rather than CFRD. DKA is uncommon in liberal individualized diet is recommended and the glu
patients with CFRD and the presence of DKA should cose‐lowering therapy (usually insulin) is adjusted to
prompt a diabetes autoantibody screen for type 1 diabe achieve near normal plasma glucose levels [32]. Current
tes [32]. Table 4.3 compares features of CFRD with the dogma is that patients with CFRD should be treated with
2 major forms of diabetes seen in the population at large. insulin therapy and that oral hypoglycemic agents may not
Although insulin deficiency is the primary defect in be as effective in optimizing nutritional and metabolic out
patients with CFRD, insulin resistance is becoming more comes [32]. As well as achieving and maintaining good
common as median age of survival increases. This is espe glycemic control the anabolic effects of insulin are also
cially true during periods of intercurrent illness (e.g. res believed to be advantageous to patients with CFRD. While
piratory infections, steroid therapy). Autoimmunity does the primary concerns are the pulmonary and nutritional
not appear to be a factor in the pathogenesis of CFRD. Low consequences [32], similar to patients with type 1 or type 2
body mass index is associated with premature mortality in diabetes, patients should undergo annual monitoring for
CF [34] and this can lead to conflict in the dietary manage the microvascular complications of diabetes. This should
ment of patients with CFRD. While calorie restriction is begin 5 years after initial diagnosis or, if the date of diagno
usually recommended in the management of patients with sis is not known, when fasting hyperglycemia was first
TABLE 4.3 Clinical features of T1DM, T2DM and CFRD. Adapted from [49].
T1DM T2DM CFRD
BMI Usually normal Usually increased Usually decreased

Presentation Acute Sub‐clinical Sub‐clinical
Age of onset Childhood and Adulthood Increasing incidence with age
adolescence
Autoantibodies Yes No Probably No
β cell function Severely reduced/absent Reduced Severely reduced
Insulin action Usually normal Severely impaired Somewhat impaired*
Treatment Insulin Diet, glucose lowering Insulin and nutrition
medication including insulin
Microvascular Yes Yes Reduced
complications
Macrovascular Yes Yes No
complications
Cause of death Cardiovascular disease Cardiovascular disease Pulmonary disease
and nephropathy
BMI: body mass index; CFRD: Cystic fibrosis‐related diabetes; T1DM: Type 1 Diabetes Mellitus; T2DM: Type 2 Diabetes Mellitus
*
Insulin sensitivity reduced during acute illness
observed [32]. Microvascular complications do not typi and within an individual patient. For example, at times of
cally occur within the first 5 years of diagnosis but compli severe pulmonary infection or during use of corticosteroid
cations such as microalbuminuria have been observed in therapy, glucose intolerance can develop but may not be
4–21% [35–37] and diabetic retinopathy in 10–27% [35– permanent. CFRD is part of a spectrum of glucose toler
38]. Macrovascular complications are infrequent and while ance in patients with CF:
most patients with diabetes die as a result of macrovascular
●● Normal glucose tolerance (FPG < 5.6 mmol/l; 2‐hour
disease, patients with CFRD are most likely to succumb to
OGTT glucose < 7.8 mmol/l)
pulmonary complications of their CF [39].
●● Indeterminate glucose tolerance (INDET) (FPG <
5.6 mmol/l; 1‐hour OGTT glucose ≥ 11.1 mmol/l; 2‐
The case for screening
hour OGTT glucose < 7.8 mmol/l)
Patients with CFRD often have no clear symptoms of dia
●● Impaired glucose tolerance (FPG < 5.6 mmol/l; 2‐hour
betes. Early diagnosis and management of CFRD is essen
OGTT glucose 7.8–11.1 mmol/l)
tial as the presence of CFRD is associated with severity of
●● CFRD FH− (FPG < 7.0 mmol/l; 2‐hour OGTT glucose ≥
lung disease [40], more difficulty in maintaining weight
11.1 mmol/l)
and increased mortality. Screening for CFRD is not a diag
●● CFRD FH+ (FPG ≥ 7.0 mmol/l; 2‐hour OGTT glucose ≥
nostic exercise but rather an exercise in case finding, initi
11.1 mmol/l)
ating earlier treatment and (ideally) achieving improved
●● Impaired fasting glucose (FPG 5.6–6.9 mmol/l; 2‐hour
outcomes. Patients with a new diagnosis of diabetes are not
OGTT glucose N/A) [42].
screened for CF; it is extremely unlikely that patients with
CF develop diabetes as the first manifestation of their dis The ADA and the CF foundation recommend the 2‐
ease. Case finding of patients with CF in a general diabetes hour 75 g OGTT as the screening test of choice [32]. This
clinic will not be discussed. Instead, we will examine the should be performed when the patient is well (≥ 6 weeks
case for screening patients with CF for diabetes. Few after an acute exacerbation) [32] and fasting for ≥ 8 hours
patients with CF have truly normal glucose status [41]. having consumed at least 600 kcal or 150 g of carbohydrate
Glucose tolerance status varies across individuals with CF per day for the preceding 3 days [41]. OGTTs should be
performed on 2 separate days in the absence of classical Pancreatic‐cancer‐associated diabetes

symptoms of hyperglycemia such as polyuria and polydip
Diabetes is a risk factor for pancreatic cancer [50]. Up to
sia. An abnormal FPG or elevated HbA1c can be used to
80% of patients with a new diagnosis of pancreatic cancer
confirm the diagnosis, but if these are normal the OGTT
have either hyperglycemia or diabetes [51]. Pancreatic can
must be repeated [41]. Annual screening for CFRD is rec
cer can present as a new diagnosis of diabetes [52]. These
ommended from 10 years in all patients without a known
associations between pancreatic cancer and diabetes have
diagnosis of CFRD [3, 32]. HbA1c is not recommended as
long been recognized but whether the pancreatic cancer
a screening tool as it is not sensitive enough [32].
causes diabetes or vice versa has been difficult to ascer
During acute illness or if requiring glucocorticoid ther
tain [53]. The association is likely bi‐directional. Diabetes
apy, glucose levels should be monitored both fasting and 2
due to pancreatic cancer is likely mediated through
hours postprandially for at least the first 48 hours of treat
humoral factors produced by the tumor, insulin resistance
ment. CFRD can manifest during initiation of continuous
or due to destruction of β‐cells [54]. Interestingly, in a
enteral feeling or gastrostomy tube feeding (diagnosis
group of patients with pancreatic cancer who underwent
based on mid or immediate post feeding values of ≥
pancreatico‐duodenectomy, 17/30 (57%) of new‐onset dia
11.1 mmol/l confirmed on two separate nights) and it is
betes resolved after tumor resection [54]. It is worth
important to check for CFRD at these times.
remembering that in patients who do not have diabetes
Insulin therapy is a safe and effective therapy in reversing
pre‐operatively there is a risk of developing de‐novo
chronic weight loss in patients with CF without fasting
diabetes post pancreatectomy. Among patients undergoing
hyperglycemia [44]. Initiation of insulin glargine is associ
distal pancreatectomy between 14–39% developed a de‐
ated with decreased progression of lung disease in patients
novo diagnosis of diabetes post‐operatively [55].
with CF and fasting hyperglycemia [45]. In patients with
Diabetes is associated with an almost 2‐fold increased
CFRD, insulin improved nutritional status, temporarily
risk of pancreatic cancer [56, 57]. Approximately 1% of
improved lung function [46] and improved protein
patients with diabetes diagnosed at ≥ 50 years of age will
synthesis [47]. Although current clinical practice guidelines
have a diagnosis of pancreatic cancer within the first 3
recommend insulin and suggest that oral hypoglycemics
years of meeting the diagnostic criteria for diabetes [58].
may not be the optimal therapy, a systematic review has sug
Diabetes itself is prevalent in the age range where pancre
gested that there is no clear difference between the two [43].
atic cancer is most common [59]. As the 5‐year survival of
Randomized controlled trials are required to evaluate this.
pancreatic cancer is so poor [60], new‐onset diabetes may
More intensive monitoring of blood glucose through
be a marker of early stage cancer [51] and could poten
use of continuous glucose sensors should be considered if
tially lead to earlier recognition with a greater chance of
hyperglycemic symptoms develop or at times of acute pul
cure. In the diabetes literature it has been suggested that
monary infection. It is unlikely that a clinical trial compar
age ≥ 50 years, a lean body habitus and lack of a family
ing treatment versus no treatment of screen detected CFRD
history of type 2 diabetes [61] are clinical indicators of the
will be undertaken. In the absence of such a trial the next
potential for underlying pancreatic cancer. However, this
best level of evidence comes from well described case
clinical impression has not been borne out by careful
series. One such case series was recently reported from the
comparison of the characteristics of patients with pancre
University of Minnesota, a unit that has experience in the
atic‐cancer‐associated diabetes and type 2 diabetes. A
management of CFRD. The report highlights a temporal
group at the Mayo Clinic has shown that a similar per
trend over the past 3 decades towards improved outcomes
centage of patients in both groups are overweight and
and greater longevity in patients with CFRD [31]. The
obese and have a positive family history for diabetes. In
authors attribute this to universal screening (which they
fact, obesity itself is a risk factor for pancreatic cancer [62];
undertake from age 6), earlier recognition of CFRD and
while effective treatment of diabetes is associated with a
more aggressive treatment with insulin and nutritional
reduced risk [53].
strategies.
For new onset diabetes mellitus to be useful as a poten people [68]; PPGL annual incidence 0.8/100 000 person
tial indicator of underlying pancreatic cancer, it will need years [69]). Screening could be associated with high rates
to be combined with a tumor biomarker. As yet, no such of false positive test results with resultant increased distress
biomarker has been identified. Adrenomedullin has been and cost from unnecessary testing. Instead case finding
associated with increased aggressiveness of pancreatic can should be undertaken when additional clinical features
cer [63] and may contribute to pancreatic associated diabe consistent with the endocrinopathy are present.
tes [64]. Studies are ongoing to determine its potential as a In the case of acromegaly, these would include enlarge
marker of pancreatic cancer in patients with new‐onset ment of the hands and feet, characteristic changes in facial
diabetes [65]. The presence of symptoms such as jaundice, appearance, headache, visual disturbance, and hyperhidro
abdominal pain, anorexia, and weight loss is problematic sis. These changes are often subtle and take many years to
because by the time the patient develops these symptoms manifest, they may not be apparent to the patient or even to
the tumor is almost always unresectable. Until a suitable family members. A comparison of the patient’s current
biomarker is identified, widespread use of abdominal appearance with that obtained from old photographs and
imaging among older patients with new onset diabetes is rings or shoes getting too small can be informative. The
not recommended as a method of searching for occult pan Acromegaly: Endocrine Society Clinical Practice Guideline
creatic cancer. The USPSTF does not recommend screen recommends measuring IGF‐1 in patients without the typ
ing for pancreatic cancer among asymptomatic patients in ical manifestations of acromegaly but who have several
the general population. This advice also applies to individ linked conditions such as obstructive sleep apnea syn
uals who have risk factors such as new‐onset diabetes and drome, type 2 diabetes, hypertension, hyperhidrosis, carpal
pre‐existing diabetes [66]. tunnel syndrome or severe arthritis [70]. All patients who
present with acromegaly should be evaluated for diabetes.
Diabetes is found in 11–38% of patients with acromegaly
Diabetes due to hormone excess
– excess growth hormone leads to insulin resistance [70].
Several disorders resulting in hormone excess can cause The classical features of PPGL include paroxysmal
diabetes including Cushing’s syndrome (glucocorticoid hypertension, palpitations, headaches, pallor, perspiration,
excess), acromegaly (growth hormone excess), hyperaldo and pain. Table 4.3 includes initial testing that can be
steronism (aldosterone excess), hyperthyroidism, pheo undertaken if the clinical suspicion of growth hormone or
chromocytoma/ paraganglioma (PPGL) (catecholamine catecholamine excess is sufficiently high to justify a case
excess), glucagonoma and somatostatinoma. Diabetes can finding exercise. One caveat with acromegaly screening is
be a feature of autoimmune polyglandular syndrome (more that poorly controlled diabetes is associated with elevated
common in type 2 than type 1) and POEMS syndrome levels of growth hormone [71] and normal or low
(polyneuropathy, organomegaly, endocrinopathy, mono IGF‐1 [72].
clonal gammopathy, skin changes). The benefit of early Cushing’s syndrome merits mention because it is the
identification and management of these conditions in a commonest endocrine disorder associated with diabetes.
patient with concomitant diabetes is that treatment of the This is due to the widespread use of exogenous glucocorti
endocrinopathy may lead to cure or amelioration of the coids (iatrogenic Cushing’s syndrome) in the treatment of a
glucose intolerance. Furthermore, many of these endo variety of inflammatory and malignant diseases. A detailed
crinopathies are associated with other metabolic complica medication history should highlight the use of exogenous
tions such as hypertension [67] which can lead to worsening steroids. Glucocorticoids may be contained in topical, over
of the micro‐ and macro‐ vascular complications of diabe the counter or herbal preparations or may be taken sur
tes. In the case of acromegaly and pheochromocytoma rou reptitiously. Furthermore, use of some steroids (such as
tine screening among individuals with diabetes is not medroxyprogesterone acetate [73]) that are not primarily
justified since both conditions are extremely rare (acro glucocorticoids has been associated with Cushing’s syn
megaly annual incidence range from 0.2–1.1 cases/100 000 drome. The mechanism whereby glucocorticoid excess
causes diabetes is multifactorial involving increased insulin ing, proximal myopathy, and striae. The tests recom
resistance and beta‐cell dysfunction. If Cushing’s syndrome mended by the Endocrine Society are listed in Table 4.3.
from exogenous glucocorticoid use is excluded, then The choice of which test(s) to use depends on several fac
(endogenous) Cushing’s syndrome is rare (3 new cases of tors including concomitant medication use, local labora
endogenous Cushing’s syndrome per million per year [74]). tory expertise as well as the index of clinical suspicion for
However, among populations of patients with diabetes who severe Cushing’s. Diagnostic criteria that suggest Cushing’s
have undergone screening for Cushing’s syndrome preva syndrome are two or more 24‐hour urinary free cortisol
lence as high as 3–5% has been reported [75, 76]. In a results above the upper limit of normal for the assay, a posi
cross‐sectional prospective study of patients attending 24 tive overnight 1‐mg dexamethasone suppression test (i.e.
diabetes clinics across Italy Terzolo et al. found no evidence an 08:00 cortisol > 50 nmol/L after administration of dexa
to support screening patients with type 2 diabetes for methasone 1 mg at midnight) or two or more late‐night
Cushing’s syndrome. However, this group concluded that salivary cortisol values above 4 nmol/L. Some authors
Cushing’s syndrome was more common than previously advocate using higher cut‐off levels as a way of reducing
thought (5/813 had a definitive diagnosis of Cushing’s syn the number of false positives. Current best evidence sug
drome). These results support case‐finding in those gests that screening for Cushing’s syndrome should take
patients with poorly controlled diabetes and hypertension place on a case by case basis with signs such as easy bruis
despite best medical therapy [77]. Budyal et al. conducted a ing, facial plethora, proximal myopathy, and striae seen as
single‐center study to determine if patients with type 2 dia “red flags”. All patients with Cushing’s syndrome should be
betes should be screened for subclinical Cushing’s syn regularly screened for diabetes.
drome (unregulated cortisol production without the
classical signs and symptoms of Cushing’s syndrome).
Post‐transplantation diabetes mellitus
They found that out of a cohort of 993 participants no one
(PTDM)
had subclinical Cushing’s and thus screening should only
take place if mandated by clinical suspicion [78]. Some Post‐transplantation diabetes mellitus (PTDM) refers to
authors have previously advocated screening all patients the recognition of diabetes in post‐transplant recipients.
with newly diagnosed diabetes (in a hospital setting) [75] This finding does not depend on the timing of the diagno
while others have suggested that screening for Cushing’s sis or whether it was present but undiagnosed prior to
should be limited to patients with poorly controlled diabe transplantation. New‐onset diabetes mellitus after trans
tes. While diabetes is prevalent among patients with plantation (NODAT) is the term used to describe those
Cushing’s syndrome, Cushing’s syndrome remains rela patients who prior to transplantation did not have diabetes
tively rare among patients with diabetes. Recent Endocrine and who develop new‐onset diabetes post transplantation.
Society Guidelines support testing for Cushing’s syndrome NODAT does not include patients who likely had diabetes
following a detailed drug history to exclude exogenous pre‐transplant but were not diagnosed. Other terminology
steroids among groups of patients with certain “overlap used in this area includes “transplant associated hypergly
disorders”, i.e. conditions which although common in the cemia”. There are wide variations in the reported preva
community can also be a feature of occult Cushing’s syn lence of PTDM due in part to varying definitions of PTDM,
drome [79]. These include osteoporosis, hypertension, follow‐up strategies over time, type of transplant, immuno
polycystic ovary syndrome and diabetes mellitus. The suppressive regimen used and donor and recipient charac
guidelines state that Cushing’s is more likely when these teristics [80]. Hyperglycemia can occur early in the
conditions occur at a young age. Another way of increasing post‐operative period (when the dose of immunosuppres
the pre‐test probability of Cushing’s is to limit screening to sant therapy is high). This can regress to normoglycemia or
those patients with diabetes and one or more of the clinical result in persistent long‐term hyperglycemia [81]. PTDM
features associated with tissue glucocorticoid excess. These has been reported in 4–25% of renal, 2.5–25% of liver,
include (but are not limited to) facial plethora, easy bruis 4–40% of heart [82] and 25–40% of lung transplants [83].
Risk factors include increasing age, ethnicity, pre‐trans annually thereafter [95]. The OGTT is the diagnostic test
plant overweight or obesity, post‐transplant weight gain, of choice [94]. HbA1c can be difficult to interpret in the
adult polycystic kidney disease, a strong family history of immediate post‐transplant period especially in those with
diabetes, previous glucose intolerance (should be assessed renal disease as HbA1c readings are impacted by blood
in all patients prior to transplantation) and a history of loss, blood transfusion and iron replacement [81].
cytomegalovirus or hepatitis C infection [84]. Hepatitis C In the absence of randomized control trial evidence,
infection itself is associated with an increased risk of type 2 management of PTDM is like that of type 2 diabetes –
diabetes mellitus [85]. Eradication of hepatitis C infection involving structured diabetes care and the use of oral anti‐
with direct‐acting antiviral agents has been shown to hyperglycemic medication and insulin. As for patients with
improve diabetes control [86]. Many immunosuppressant type 2 diabetes, drug‐specific and patient factors need to be
drugs including calcineurin inhibitors (tacrolimus > cyclo considered when selecting appropriate therapy including
sporin) as well as corticosteroids are diabetogenic. The lat the degree of dysglycemia, risk of hypoglycemia, weight
ter causes diabetes predominantly through increasing change, cardiovascular and renal effects, cost, and route of
insulin resistance, enhancing gluconeogenesis and lipolysis administration. The ADA advise that irrespective of the
and decreasing glucose uptake in skeletal muscle in patients risk of NODAT, the immunosuppressive therapy associated
whose beta‐cell response is compromised [81]. Calcineurin with the best outcomes and graft survival for the patient
inhibitors likely contribute to diabetes through a combina should be used [94]. In addition, patients post transplanta
tion of β‐cell apoptosis, impaired insulin secretion and tion may have renal or hepatic impairment, the severity of
insulin resistance although further study is required to which may be a contraindication or require dose adjust
fully understand the mechanism [87]. ment for certain medications (such as biguanides) and
consideration should be given to the interaction of immu
Clinical features nosuppressants and oral anti‐hyperglycemic agents [96].
Weight gain is a common occurrence after transplanta Modification of the immunosuppressant regimen can be
tion [81] and the risk of developing PTDM increases by a considered in certain cases (under appropriate supervision
factor of 1.4 for every 10 kg increase in weight over and guidance, for example changing from tacrolimus to
60 kg [88]. Contributors to weight gain are likely multifac ciclosporin or mycophenolate mofetil plus azathio
torial and include steroid therapy and fewer dietary restric prine [97]) must be balanced with the risk of graft rejec
tions (for example no longer requiring a strict renal tion. Furthermore, the benefits of reducing corticosteroid
diet) [81]. Taken with the increasing insulin resistance dosage and reducing or substituting calcineurin inhibitors
associated with the immunosuppressant medication, the have not been proven fully in clinical trials. A screening
clinical features of PTDM resemble those of type 2 program should involve pre‐ and post‐transplant testing of
diabetes. glucose tolerance. A knowledge of pre‐transplant risk fac
tors is essential to characterize the subsequent risk of
The case for screening PTDM.
PTDM predicts graft failure/function [89], mortality [90,
91], infection and micro‐[92]/macro‐vascular [93] compli
Diabetes associated with atypical
cations of diabetes. Early recognition of PTDM with the
antipsychotic drug use
potential to prevent the associated morbidity and mortality
has been advocated. The ADA advise that patients should Antipsychotic drugs are licensed for use in schizophrenia
be screened when stable on immunosuppressive therapy and related psychotic disorders. Irrespective of antipsy
and no active infection [94]. There are no formal recom chotic therapy, schizophrenia itself is associated with an
mendations regarding the exact frequency of screening for increased risk of developing diabetes, the cause of which is
PTDM. Some authors recommend measurement of FPG multifactorial including both environmental and biological
weekly for the first 4 weeks, at 3, 5 and 12 months and factors such as sedentary lifestyle and poor diet [98] as well
as disease and treatment specific effects [99]. First‐genera obesity than age and sex matched controls [104], a risk fac
tion antipsychotics were useful for alleviating the positive tor for the onset of type 2 diabetes [105]. Patients with
symptoms of psychosis such as hallucinations and delu schizophrenia who are prescribed second‐generation
sions. They are less efficacious at managing the negative antipsychotics compared to first generation antipsychotics
symptoms of psychosis which include withdrawal, poverty have a relative risk of having diabetes of 1.32 (95%CI 1.15–
of speech and cognitive problems. The introduction of sec 1.51) [106]. A systematic review indicated that there is an
ond‐generation (or atypical) antipsychotics was greeted increased risk of type 2 diabetes among children and youth
with enthusiasm as they were efficacious against both posi (aged 6–24 years) prescribed antipsychotic drugs that
tive and negative psychotic symptoms. These agents have increased with cumulative dose [107].
less debilitating extrapyramidal side‐effects than the first‐ Although high‐quality cohort studies have been diffi
generation agents but have a stronger diabetogenic cult to undertake, it is generally accepted that these disor
effect [100]. Second‐generation antipsychotics including ders are associated with an increased risk of obesity,
aripiprazole, asenapine, clozapine, iloperidone, olanzapine, dyslipidemia, diabetes, and resultant increased cardiovas
quetiapine, and risperidone have transformed the lives of cular morbidity. Because of the effect of the underlying dis
many patients with schizophrenia and have enabled many ease, it is difficult to tease out the exact contribution of an
individuals to live in the community as opposed to being additional drug effect on top of this. Clinical experience
managed as inpatients in psychiatric institutions. These does suggest that many of the atypical antipsychotics are
medications have become popular with psychiatrists and associated with weight gain and metabolic abnormalities.
general practitioners and are now prescribed for many psy Between 15–72% of patients on antipsychotics have weight
chiatric conditions other than those for which they are gain during the acute or maintenance phase of treatment of
licensed (attention‐deficit hyperactivity disorder, depres schizophrenia and affective disorders [99]. Atypical antip
sion, eating disorders, insomnia, obsessive compulsive dis sychotics except for ziprasidone also appear to be associ
order. . .). Among nursing home residents for example (in ated with weight gain (clozapine followed by olanzapine
the 2004 National Nursing Home Survey), 86% of second‐ and quetiapine have the highest weight gain) [108]. Use of
generation antipsychotics prescribed were for off‐label these agents has been associated with profound metabolic
indications [101]. Side‐effects include agranulocytosis disturbance including DKA. In some case reports presenta
associated with clozapine (very rare; regular monitoring of tion with DKA is followed by subsequent recovery of β‐cell
the patient’s full blood count is mandatory) to the much function akin to the so‐called “Flatbush” or “ketosis‐prone
commoner metabolic disorders which we will discuss in type 2 diabetes” [109]. More commonly the metabolic
detail. derangement is less profound and associated with weight
gain and dyslipidemia. As well as causing new diabetes
Natural history among individuals not previously recognized as having the
Psychiatric disorders for which atypical antipsychotics are disease these drugs can also lead to worsening of pre‐exist
prescribed are a risk factor for or a complication of diabe ing diabetes and a need for intensification of anti‐hypergly
tes [102]. The underlying disorders for which the atypical cemic therapy. Co‐ordinated shared care plans should be
antipsychotic drugs are used (schizophrenia, other psy established by the psychiatry team and diabetes centers to
chotic disorders, major depression, etc.) are often associ optimize patient care [99].
ated with a sedentary lifestyle and poor dietary habits (less
fruit and vegetables consumption [103]). In patients with The case for screening
severe mental illness, modification of risk factors associ In consultation with the European Association for the
ated with diabetes such as dietary and exercise interven Study of Diabetes (EASD) and the European Society of
tions can be difficult due to low mood, impaired judgement, Cardiology (ESC), the European Psychiatric Association
and/or disordered thinking. Drug‐naïve and drug‐free (EPA) published a position statement outlining how
patients with schizophrenia are more likely to have central patients with serious mental illness have an increased risk
of cardiovascular disease and diabetes [99]. Patients with Diabetes associated with HIV infection
severe mental illnesses have worse physical health, and its treatment
decreased life expectancy and a 2‐ to 3‐fold increased mor
With the advent of safe and effective antiretroviral therapy
tality rate compared to the general population [99]. There
(ART) over the past 40 years human immunodeficiency
is a gradation of metabolic risk across drugs in the
virus (HIV) infection in the developed world has seen a
class [110] with clozapine and olanzapine carrying a higher
transformation from a disease associated with opportunis
risk of metabolic side‐effects than quetiapine and risperi
tic infection, AIDS and premature death to a chronic dis
done [110]. For example a metanalysis found that clozap
ease associated with good long‐term survival. Improved
ine (4.45 kg) and olanzapine (4.15 kg) had greater risk of
life expectancy of patients with HIV and ART, has seen the
weight gain than quetiapine and risperidone (2.1 kg) while
development of chronic disease complications in particular
drugs like aripiprazole, amisulpride, and ziprasidone had
cardiometabolic diseases such as diabetes and increased
limited effect on weight [111]. Causes for this weight gain
cardiovascular event rates [115]. In the Multicenter AIDS
are multifactorial and poorly understood (increased appe
cohort study, the prevalence of diabetes was 14% among
tite and modified energy expenditure) [99]. A recent
411 HIV‐infected men using ART (4.7 cases per 100‐per
review concluded that it is unclear if antipsychotic medica son years) compared to 5% among 711 HIV seronegative
tion exacerbates a predisposition to new onset of type 2 men (1.4 cases per 100‐person years) [116]. Major reported
diabetes or whether there is a direct effect [113]. risk factors for the development of dysglycemia in those
The EPA position statement called for increased coop with HIV include family history of diabetes, increasing age,
eration and shared cared among the relevant healthcare hepatitis C infection, elevated BMI, fat redistribution and
professionals to improve care for those who have a serious medications such as indinavir and lopinavir/ritona
mental illness [99]. Management is often complicated by vir [117]. The long‐term complications of diabetes can be
lack of insight into the underlying conditions (both psychi accentuated among patients with diabetes and HIV [117].
atric and medical) and the patients’ inability to follow a Similar to atypical antipsychotics, clinicians prescribing
treatment program. All patients at initial presentation these drugs need to be aware that the risk of metabolic
should have a detailed medical history (including past derangement varies between the different agents [118]. In
medical history, family history, smoking, exercise, dietary). those with dysglycemia, it is important to consider if alter
Clinical examination should include recording of height native ARTs with a less adverse glycemic profile could be
and weight (and calculation of body mass index), waist cir used while still effectively and safely managing HIV. That
cumference and blood pressure as well as biochemical said considerable controversy exists on the reported links
measurement of fasting glucose, HbA1c and lipid levels. between ART and diabetes [119–121], which may be due
Although these are measured routinely in diabetes clinics to variable standards/methods of screening for diabetes, to
around the world they are not undertaken routinely in heterogeneity between studies used in systematic
Departments of Psychiatry. However, if we are to stem the reviews and meta‐analyses and to lack of long‐term
rise in metabolic disorders that would appear to be hap follow‐up [120].
pening with increasing use of these drugs then this is an For patients with HIV, the ADA recommends screening
important first step. The ADA recommends annual for diabetes and prediabetes with a fasting glucose test
screening for patients who are prescribed atypical antipsy before starting antiretroviral therapy, at the time of switch
chotic medications for prediabetes or diabetes in accord ing antiretroviral therapy, and 3–6 months after starting or
ance with ADA guidelines [28]. In addition, among switching antiretroviral therapy. If initial screening results
treatment naïve patients with serious mental illness who are normal, fasting glucose should be checked annu
do not have diabetes at baseline the EPA recommends ally [28]. HbA1c is not an appropriate screening tool in
repeat fasting glucose and lipids at week 6 and 12 post ini patients with HIV as it underestimates the degree of dys
tiation of therapy [99]. glycemia by 1.6±0.2 mmol/L [122]. Kim et al. found that
HbA1c is particularly impacted in those on nucleotide addition to FPG should be used to monitor patients with
analog reverse transcriptase inhibitors particularly patients known type 1 or 2 diabetes on ICIs.
with macrocytosis or on abacavir [122].
Monogenic forms of diabetes

Diabetes secondary to immunotherapy
The development of immunotherapy as a treatment for a C ase V ign et t e:
A 23‐year‐old Caucasian male was admitted to hospital
variety of cancers has revolutionized clinical outcomes
with acute urinary retention and renal failure. He
(improved prognosis) for these conditions [134]. complained of progressive thirst, polyuria and nocturia
CTLA‐4 inhibitors such as Ipilimumab, PD‐1 inhibitors over the last 18 months. His past history was remarkable
such as Pembrolizumab and Nivolumab and PD‐L1 inhibi for diabetes mellitus diagnosed 10 years previously. This
tors such as Atezolizumab, Avelumab, and Durvalumab in was managed with insulin aspart and insulin glargine. He
had poor eyesight since childhood which was attributed to
addition to combination therapies (CTLA‐4 and PD‐1 inhi
Leber’s optic atrophy. Family history: mother – type 2
bition) such as Ipilimumab/Nivolumab are approved for diabetes mellitus (on insulin therapy), gestational diabetes;
treatment in a broad spectrum of conditions from mela father – hypercholesterolemia; grandmother – diabetes
noma to renal cell cancer to Hodgkin’s lymphoma [134]. mellitus, died aged 73 years; three sisters alive and well
The number of cancers which have been shown to benefit with no history of diabetes.
On examination he was very hard of hearing. Blood
from immunotherapy are increasing [135].
pressure was 170/95 mmHg. Body mass index was
While other endocrinopathies such as hypothyroidism, elevated at 29.8 kg/m2. He had markedly decreased visual
hyperthyroidism, and hypophysitis are more common, dia acuity. Dilated fundoscopy revealed pale optic disks and
betes secondary to ICIs represents a rare but important scattered retinal hemorrhages. There was no evidence of
complication of immunotherapy; the first presentation can peripheral neuropathy. The remainder of the physical
examination was normal. An ultrasound revealed bilateral
be with life‐threatening DKA [136]. Insulin‐dependent
hydronephrosis. His acute kidney injury resolved with
diabetes occurs in up to 1% of patients who receive either a insertion of a suprapubic catheter. An audiogram revealed
PD‐1 or PD‐L1 inhibitors [137]. Reports suggest that time bilateral sensorineural deafness. A water deprivation test
to onset of diabetes from the first dose of the ICI varies confirmed cranial diabetes insipidus.
from 5–790 days (median 116 days) [138]. The number of Genetic testing revealed a complex heterozygote state
with c.506G and c.1611_1624del14 mutations in the
doses before the onset of diabetes varies from 1–24
WFS1 gene on chromosome 4 confirming the clinical
(median 3) [138]. impression of Wolfram’s or DIDMOAD (Diabetes Insipidus,
FPG should be measured before commencing ICIs and Diabetes Mellitus, Optic Atrophy and Deafness) syndrome.
at the start of each course of treatment for the first Both parents were confirmed to be carriers and the family
6 months, every two courses for the next 6 months and as received genetic counselling on the inheritance of this
autosomal recessive disorder.
clinical signs and symptoms dictate [139]. The Society for
Endocrinology recommends the measurement of FPG in
all unwell patients who are on ICIs [135, 140]. Furthermore,
One of the main justifications for identifying specific
they advise the management of DKA and HHS in accord
types of diabetes due to other causes is the potential for
ance with local guidelines and measurement of autoanti
cure through treatment of the underlying condition.
bodies and C‐peptide [135] and ICIs should be held until
However, at present this does not apply to monogenic
the patient has recovered [141]. Patients on ICIs often
forms of diabetes. Instead justification comes from:
require high‐dose steroids which in itself will make the
patient prone to developing steroid‐induced diabetes [135]. 1 Potential to adapt the treatment regimen to the type
HbA1c is not suitable for diagnosing diabetes in the acute of monogenic diabetes (a step towards precision
setting, due to the often‐rapid onset of diabetes but in medicine).
2 Ability to provide an individual patient and their family patient and allows patients to be treated with a more cost‐
with a more accurate prognosis (genetic counselling). effective therapy.
3 Potential to advance our understanding of diabetes in This phenomenon (of stopping insulin after many years
general through elucidation of rare molecular and cel of use) has also been reported in patients with permanent
lular defects causing diabetes [124]. neonatal diabetes due to mutations in the Kir6.2 subunit of
the ATP sensitive potassium channels in the beta‐cell
Monogenic diabetes represents a diverse group of con (KCNJ11 mutation) (inherited spontaneously (80%) or
ditions with varied presentations and underlying genetic autosomal dominant) [125, 132]. Cellular mechanism
causes [125]. In the case described above, the family was relates to the mutated channels remaining open resulting
not aware of the progressive nature of this rare condition in the beta‐cell membrane remaining in a hyperpolarized
(1 in 770 000 in the UK [126]) Wolfram syndrome with state thereby preventing insulin from being secreted.
almost inevitable death before age 50 (median age of death Sulfonylureas leads to closure of the channel thus enabling
39 (range 25–49) years) [126]; most families appreciate the insulin secretion. Diabetes occurring before 6 months of
additional knowledge gained through genetic testing and age is typically (80–85% [133]) due to monogenic causes,
genetic counselling. which can occur transiently or permanently. Again, early
Examples where identification of monogenic forms of identification is essential as some forms of neonatal diabe
diabetes result in a better fit between the diabetes syn tes can be treated effectively with sulphonylureas while
drome and the treatment regimen come from both the others require intensive insulin strategies.
Maturity Onset Diabetes of the Young (MODY) and neo Patients with other monogenic forms of diabetes are
natal diabetes syndromes. MODY is inherited in an autoso often misdiagnosed as having type 1 diabetes (if childhood
mal dominant pattern, is often underdiagnosed and onset) or type 2 diabetes (if adult onset). These monogenic
accounts for > 1% of all cases of diabetes [123]. Several dif forms include that associated with the m.3243A>G muta
ferent gene defects have been recognized in MODY. MODY tion in mitochondrial DNA as well as lipodystrophic forms
3, due to a defect in the hepatocyte nuclear factor 1‐alpha of diabetes due to mutations in the LMNA gene causing
(HNF1α) gene, is the commonest form of MODY (30–65% familial partial lipodystrophy. The phenotypic characteris
of MODY patients). Other common forms of MODY tics that should alert the clinician to the possibility of a
include MODY 2 (due to GCK mutation; 30–50% of monogenic form of diabetes include (1) young age of onset,
MODY patients), MODY 1 (due to HNF4α mutation; (2) family history demonstrating either a pattern of autoso
5–10% of MODY patients) and MODY 5 (due to HNF1β mal dominant inheritance (e.g. MODY kindreds) or of
mutation; < 5% of MODY patients) [127]. Several clinical maternal transmission (e.g. mitochondrial diabetes), (3)
clues may suggest a diagnosis of MODY and the mutation association with deafness (e.g. mitochondrial diabetes), or
contributing to the diagnosis for example lower renal (4) neurological or neuromuscular disorders (e.g. Frederich’s
threshold for glucose (HNF1α) [128], large birthweight ataxia).
(HNF4α) [129], transient or persistent neonatal hypogly The ADA [3] recommends screening for monogenic
cemia (HNF4α) [130], and renal disease (HNF‐1β) [131]. diabetes syndromes in the following circumstances:
Of import, patients with MODY 2 often require no treat
1 Diagnosis of diabetes < 6 months (neonatal diabetes).
ment (stable fasting hyperglycemia), while patients with
2 Diagnosis of diabetes in children and early adulthood
MODY 3 and MODY 1 can be managed effectively with
with a pattern of inheritance suggestive of Autosomal
sulphonylureas as a first‐line agent. Because diabetes due
Dominance (MODY).
to MODY can be diagnosed in childhood or adolescence,
many of these patients are started on insulin on the In addition, the ADA recommends consultation with a
assumption that they have type 1 diabetes. The ability to center that has expertise in the management of monogenic
recognize previously undiagnosed MODY and potentially diabetes to interpret the mutations, optimize patient care,
remove the need for lifelong insulin is a huge benefit to the and facilitate genetic counselling [3].
References 16. Barton JC, Acton RT. Diabetes in HFE Hemochromatosis. J

Diabetes Res. 2017;2017:9826930. DOI: 10.1155/2017/982
1. Report of the Expert Committee on the Diagnosis and 6930.
Classification of Diabetes Mellitus. Diabetes Care. 17. Utzschneider KM, Kowdley KV. Hereditary hemochromato
1997;20:1183–1197. sis and diabetes mellitus: implications for clinical practice. Nat
2. Gale EA. Declassifying diabetes. Diabetologia. Rev Endocrinol. 2010;6:26–33. DOI: 10.1038/nrendo.2009.241.
2006;49:1989–1995. DOI: 10.1007/s00125‐006‐0348‐7. 18. Screening for hemochromatosis: recommendation statement.
3. Classification and diagnosis of diabetes: standards of Ann Intern Med. 2006;145:204–208. DOI: 145/3/204 [pii].
medical care in diabetes – 2020. Diabetes Care. 19. UK National Screening Committee. Screening for Haemo
2020;43:S14–S31. DOI: 10.2337/dc20‐S002. chromatosis. 2015.
4. International Diabetes Federation. IDF Diabetes Atlas, 9th 20. European Association For The Study Of The L. EASL clini
edn. Brussels, Belgium, 2019. Available at www.diabetes cal practice guidelines for HFE hemochromatosis. J Hepatol.
atlas.org. 2010;53:3–22. DOI: 10.1016/j.jhep.2010.03.001.
5. McCulloch DK. Classification of Diabetes Mellitus and 21. Bacon BR, Adams PC, Kowdley KV, Powell LW, Tavill AS,
Genetic Diabetic Syndromes. Waltham, MA. UpToDate. American Association for the Study of Liver D. Diagnosis and
2020. management of hemochromatosis: 2011 practice guideline by
6. Public Health England. Guidance. Criteria for Appraising the American Association for the Study of Liver Diseases.
the Viability, Effectiveness and Appropriateness of a Hepatology. 2011;54:328–343. DOI: 10.1002/hep.24330.
Screening Programme. 2015. 22. Force USPST. Screening for hemochromatosis: recommen
7. UK National Screening Committee (UK NSC). Current UK dation statement. Ann Intern Med. 2006;145:204–208. DOI:
NSC recommendations. 10.7326/0003‐4819‐145‐3‐200608010‐00008.
8. U.S. Preventive Services Task Force. Abnormal Blood 23. Kowdley KV, Brown KE, Ahn J, Sundaram V. ACG Clinical
Glucose and Type 2 Diabetes Mellitus: Screening. 2015. Guideline: hereditary hemochromatosis. Am J Gastro
9. Brissot P, Pietrangelo A, Adams PC, de Graaff B, McLaren enterol. 2019;114:1202–1218. DOI: 10.14309/ajg.00000000000
CE, Loreal O. Haemochromatosis. Nat Rev Dis Primers. 00315.
2018;4:18016. DOI: 10.1038/nrdp.2018.16. 24. Bulaj ZJ, Ajioka RS, Phillips JD, LaSalle BA, Jorde LB, Griffen
10. Crownover BK, Covey CJ. Hereditary hemochromatosis. LM et al. Disease‐related conditions in relatives of patients
Am Fam Physician. 2013;87:183–190. with hemochromatosis. N Engl J Med. 2000;343:1529–1535.
11. Adams PC, Reboussin DM, Barton JC, McLaren CE, Eckfeldt DOI: 10.1056/NEJM200011233432104.
JH, McLaren GD et al. Hemochromatosis and iron‐overload 25. Jacobs EM, Hendriks JC, Marx JJ, van Deursen CT,
screening in a racially diverse population. N Engl J Med. Kreeftenberg HG, de Vries RA et al. Morbidity and mortality
2005;352:1769–1778. DOI: 10.1056/NEJMoa041534. in first‐degree relatives of C282Y homozygous probands
12. Rossi E, Jeffrey GP. Clinical penetrance of C282Y homozy with clinically detected haemochromatosis compared with
gous HFE haemochromatosis. Clin Biochem Rev. 2004;25: the general population: the HEmochromatosis FAmily Study
183–190. (HEFAS). Neth J Med. 2007;65:425–433.
13. Whitlock EP, Garlitz BA, Harris EL, Beil TL, Smith PR. 26. O’Brien T, Barrett B, Murray DM, Dinneen S, O’Sullivan DJ.
Screening for hereditary hemochromatosis: a systematic Usefulness of biochemical screening of diabetic patients for
review for the U.S. Preventive Services Task Force. Ann hemochromatosis. Diabetes Care. 1990;13:532–534.
Intern Med. 2006;145:209–223. DOI: 145/3/209 [pii]. 27. Pilling LC, Tamosauskaite J, Jones G, Wood AR, Jones L,
14. Bardou‐Jacquet E, Morcet J, Manet G, Laine F, Perrin M, Kuo CL et al. Common conditions associated with heredi
Jouanolle AM et al. Decreased cardiovascular and extrahe tary haemochromatosis genetic variants: cohort study in UK
patic cancer‐related mortality in treated patients with mild Biobank. BMJ. 2019;364:k5222. DOI: 10.1136/bmj.k5222.
HFE hemochromatosis. J Hepatol. 2015;62:682–689. DOI: 28. American Diabetes A. 2. Classification and diagnosis of dia
10.1016/j.jhep.2014.10.025. betes: standards of medical care in diabetes – 2020. Diabetes
15. Wood MJ, Powell LW, Dixon JL, Ramm GA. Clinical cofac Care. 2020;43:S14–S31. DOI: 10.2337/dc20‐S002.
tors and hepatic fibrosis in hereditary hemochromatosis: the 29. Tavill AS. Diagnosis and management of hemochromatosis.
role of diabetes mellitus. Hepatology. 2012;56:904–911. DOI: Hepatology. 2001;33:1321–1328. DOI: S0270913901325302
10.1002/hep.25720. [pii] 10.1053/jhep.2001.24783.
30. Ioannou L, McClaren BJ, Massie J, Lewis S, Metcalfe SA, 41. Moran A, Becker D, Casella SJ, Gottlieb PA, Kirkman MS,
Forrest L et al. Population‐based carrier screening for cystic Marshall BC et al. Epidemiology, pathophysiology, and
fibrosis: a systematic review of 23 years of research. Genet prognostic implications of cystic fibrosis‐related diabetes: a
Med. 2014;16:207–216. DOI: 10.1038/gim.2013.125. technical review. Diabetes Care. 2010;33:2677–2683. DOI:
31. Moran A, Dunitz J, Nathan B, Saeed A, Holme B, Thomas W. 10.2337/dc10–1279.
Cystic fibrosis‐related diabetes: current trends in prevalence, 42. Ode KL, Frohnert B, Laguna T, Phillips J, Holme B,
incidence, and mortality. Diabetes Care. 2009;32:1626–1631. Regelmann W et al. Oral glucose tolerance testing in chil
DOI: dc09‐0586 [pii] 10.2337/dc09‐0586. dren with cystic fibrosis. Pediatr Diabetes. 2010;11:487–492.
32. Moran A, Brunzell C, Cohen RC, Katz M, Marshall BC, DOI: 10.1111/j.1399‐5448.2009.00632.x.
Onady G et al. Clinical care guidelines for cystic fibrosis‐ 43. Onady GM, Stolfi A. Insulin and oral agents for managing
related diabetes: a position statement of the American cystic fibrosis‐related diabetes. Cochrane Database Syst Rev.
Diabetes Association and a clinical practice guideline of the 2016;4:CD004730. DOI: 10.1002/14651858.CD004730.pub4.
Cystic Fibrosis Foundation, endorsed by the Pediatric 44. Moran A, Pekow P, Grover P, Zorn M, Slovis B, Pilewski J
Endocrine Society. Diabetes Care. 2010;33:2697–2708. DOI: et al. Insulin therapy to improve BMI in cystic fibrosis‐
10.2337/dc10‐1768. related diabetes without fasting hyperglycemia: results of the
33. Milla CE, Billings J, Moran A. Diabetes is associated with cystic fibrosis related diabetes therapy trial. Diabetes Care.
dramatically decreased survival in female but not male sub 2009;32:1783–1788. DOI: 10.2337/dc09‐0585.
jects with cystic fibrosis. Diabetes Care. 2005;28:2141–2144. 45. Mozzillo E, Franzese A, Valerio G, Sepe A, De Simone I,
DOI: 10.2337/diacare.28.9.2141. Mazzarella G et al. One‐year glargine treatment can improve
34. Chamnan P, Shine BS, Haworth CS, Bilton D, Adler AI. the course of lung disease in children and adolescents with
Diabetes as a determinant of mortality in cystic fibrosis. cystic fibrosis and early glucose derangements. Pediatr
Diabetes Care. 2010;33:311–316. DOI: 10.2337/dc09‐1215. Diabetes. 2009;10:162–167. DOI: 10.1111/j.1399‐5448.2008.
35. Schwarzenberg SJ, Thomas W, Olsen TW, Grover T, Walk D, 00451.x.
Milla C et al. Microvascular complications in cystic fibrosis‐ 46. Mohan K, Israel KL, Miller H, Grainger R, Ledson MJ,
related diabetes. Diabetes Care. 2007;30:1056–1061. DOI: Walshaw MJ. Long‐term effect of insulin treatment in cystic
10.2337/dc06‐1576. fibrosis‐related diabetes. Respiration. 2008;76:181–186.
36. Andersen HU, Lanng S, Pressler T, Laugesen CS, Mathiesen DOI: 10.1159/000110206.
ER. Cystic fibrosis‐related diabetes: the presence of micro 47. Rafii M, Chapman K, Stewart C, Kelly E, Hanna A, Wilson
vascular diabetes complications. Diabetes Care. 2006;29: DC et al. Changes in response to insulin and the effects of
2660–2663. DOI: 10.2337/dc06‐0654. varying glucose tolerance on whole‐body protein metabo
37. van den Berg JM, Morton AM, Kok SW, Pijl H, Conway SP, lism in patients with cystic fibrosis. Am J Clin Nutr.
Heijerman HG. Microvascular complications in patients 2005;81:421–426. DOI: 10.1093/ajcn.81.2.421.
with cystic fibrosis‐related diabetes (CFRD). J Cyst Fibros. 48. Moran A, Hardin D, Rodman D, Allen HF, Beall RJ, Borowitz
2008;7:515–519. DOI: 10.1016/j.jcf.2008.05.008. D et al. Diagnosis, screening and management of cystic
38. Yung B, Landers A, Mathalone B, Gyi KM, Hodson ME. fibrosis related diabetes mellitus: a consensus conference
Diabetic retinopathy in adult patients with cystic fibrosis‐ report. Diabetes Res Clin Pract. 1999;45:61–73. DOI:
related diabetes. Respir Med. 1998;92:871–872. DOI: 10.1016/ 10.1016/s0168‐8227(99)00058‐3.
s0954‐6111(98)90390‐0. 49. O’Riordan SM, Robinson PD, Donaghue KC, Moran A,
39. Litvin M, Nwachukwu S. Cystic Fibrosis Related Diabetes: a Consensus ICP. Management of cystic fibrosis‐related diabetes.
Unique Challenge in Diabetes Care. Mo Med. 2016;113: Pediatr Diabetes. 2008;9:338–344. DOI: 10.1111/j.1399‐5448
384–389. .2008.00437.x.
40. Koch C, Rainisio M, Madessani U, Harms HK, Hodson ME, 50. Tan J, You Y, Guo F, Xu J, Dai H, Bie P. Association of ele
Mastella G et al. Presence of cystic fibrosis‐related diabetes vated risk of pancreatic cancer in diabetic patients: a system
mellitus is tightly linked to poor lung function in patients atic review and meta‐analysis. Oncol Lett. 2017;13:1247–1255.
with cystic fibrosis: data from the European Epidemiologic DOI: 10.3892/ol.2017.5586.
Registry of Cystic Fibrosis. Pediatr Pulmonol. 2001;32: 51. Pannala R, Basu A, Petersen GM, Chari ST. New‐onset dia
343–350. DOI: 10.1002/ppul.1142. betes: a potential clue to the early diagnosis of pancreatic
cancer. Lancet Oncol. 2009;10:88–95. DOI: 10.1016/S1470‐ ADMR. Cancer Res. 2007;67:2666–2675. DOI: 10.1158/0008‐
2045(08)70337‐1. 5472.CAN‐06‐3362.
52. Patel R, Ede J, Collins J, Willens D. Pancreatic cancer pre 64. Aggarwal G, Ramachandran V, Javeed N, Arumugam T,
senting as new‐onset diabetes. Case Rep Oncol. 2014;7: Dutta S, Klee GG et al. Adrenomedullin is up‐regulated in
171–174. DOI: 10.1159/000360812. patients with pancreatic cancer and causes insulin resistance
53. Magruder JT, Elahi D, Andersen DK. Diabetes and pancre in beta cells and mice. Gastroenterology. 2012;143:1510–
atic cancer: chicken or egg? Pancreas. 2011;40:339–351. 1517 e1511. DOI: 10.1053/j.gastro.2012.08.044.
DOI: 10.1097/MPA.0b013e318209e05d. 65. Antolino L, Rocca M, Todde F, Catarinozzi E, Aurello P,
54. Pannala R, Leirness JB, Bamlet WR, Basu A, Petersen GM, Bollanti L et al. Can pancreatic cancer be detected by adre
Chari ST. Prevalence and clinical profile of pancreatic can nomedullin in patients with new‐onset diabetes? The
cer‐associated diabetes mellitus. Gastroenterology. 2008;134: PaCANOD cohort study protocol. Tumori. 2018;104:312–
981–987. DOI: 10.1053/j.gastro.2008.01.039. 314. DOI: 10.5301/tj.5000693.
55. De Bruijn KM, van Eijck CH. New‐onset diabetes after distal 66. Force USPST, Owens DK, Davidson KW, Krist AH, Barry
pancreatectomy: a systematic review. Ann Surg. 2015;261: MJ, Cabana M et al. Screening for pancreatic cancer: US
854–861. DOI: 10.1097/SLA.0000000000000819. Preventive Services Task Force Reaffirmation Recommen
56. Ben Q, Xu M, Ning X, Liu J, Hong S, Huang W et al. Diabetes dation Statement. JAMA. 2019;322:438–444. DOI: 10.1001/
mellitus and risk of pancreatic cancer: a meta‐analysis of jama.2019.10232.
cohort studies. Eur J Cancer. 2011;47:1928–1937. DOI: 67. O’Shea PM, Griffin TP, Fitzgibbon M. Hypertension: the
10.1016/j.ejca.2011.03.003. role of biochemistry in the diagnosis and management. Clin
57. Li D, Tang H, Hassan MM, Holly EA, Bracci PM, Silverman Chim Acta. 2017;465:131–143. DOI: 10.1016/j.cca.2016.
DT. Diabetes and risk of pancreatic cancer: a pooled analysis 12.014.
of three large case‐control studies. Cancer Causes Control. 68. Lavrentaki A, Paluzzi A, Wass JAH, Karavitaki N.
2011;22:189–197. DOI: 10.1007/s10552‐010‐9686‐3. Epidemiology of acromegaly: review of population studies.
58. Chari ST, Leibson CL, Rabe KG, Ransom J, de Andrade M, Pituitary. 2017;20:4–9. DOI: 10.1007/s11102‐016‐0754‐x.
Petersen GM. Probability of pancreatic cancer following dia 69. Beard CM, Sheps SG, Kurland LT, Carney JA, Lie JT.
betes: a population‐based study. Gastroenterology. 2005;129: Occurrence of pheochromocytoma in Rochester, Minnesota,
504–511. DOI: 10.1016/j.gastro.2005.05.007. 1950 through 1979. Mayo Clin Proc. 1983;58:802–804.
59. Fernandez‐del Castillo C. Clinical manifestations, diagnosis, 70. Katznelson L, Laws ER, Jr., Melmed S, Molitch ME, Murad
and staging of exocrine pancreatic cancer. Waltham, MA. MH, Utz A et al. Acromegaly: an endocrine society clinical
UpToDate. 2020. practice guideline. J Clin Endocrinol Metab. 2014;99:3933–
60. McGuigan A, Kelly P, Turkington RC, Jones C, Coleman 3951. DOI: 10.1210/jc.2014‐2700.
HG, McCain RS. Pancreatic cancer: a review of clinical diag 71. Vigneri R, Squatrito S, Pezzino V, Filetti S, Branca S, Polosa
nosis, epidemiology, treatment and outcomes. World J P. Growth hormone levels in diabetes. Correlation with the
Gastroenterol. 2018;24:4846–4861. DOI: 10.3748/wjg.v24. clinical control of the disease. Diabetes. 1976;25:167–172.
i43.4846. DOI: 10.2337/diab.25.3.167.
61. Noy A, Bilezikian JP. Clinical review 63: diabetes and pan 72. Lim DJ, Kwon HS, Cho JH, Kim SH, Choi YH, Yoon KH
creatic cancer: clues to the early diagnosis of pancreatic et al. Acromegaly associated with type 2 diabetes showing
malignancy. J Clin Endocrinol Metab. 1994;79:1223–1231. normal IGF‐1 levels under poorly controlled glycemia.
DOI: 10.1210/jcem.79.5.7962312. Endocr J. 2007;54:537–541. DOI: 10.1507/endocrj.k06‐083.
62. Xu M, Jung X, Hines OJ, Eibl G, Chen Y. Obesity and pancre 73. Siminoski K, Goss P, Drucker DJ. The Cushing Syndrome
atic cancer: overview of epidemiology and potential preven induced by medroxyprogesterone acetate. Annals of Internal
tion by weight loss. Pancreas. 2018;47:158–162. DOI: Medicine. 1989;111:758–760. DOI: 10.7326/0003‐4819‐111‐
10.1097/MPA.0000000000000974. 9‐758.
63. Ramachandran V, Arumugam T, Hwang RF, Greenson JK, 74. Wengander S, Trimpou P, Papakokkinou E, Ragnarsson O.
Simeone DM, Logsdon CD. Adrenomedullin is expressed in The incidence of endogenous Cushing’s syndrome in the
pancreatic cancer and stimulates cell proliferation and inva modern era. Clin Endocrinol (Oxf). 2019;91:263–270. DOI:
sion in an autocrine manner via the adrenomedullin receptor, 10.1111/cen.14014.
75. Reimondo G, Pia A, Allasino B, Tassone F, Bovio S, Borretta 86. Hum J, Jou JH, Green PK, Berry K, Lundblad J, Hettinger BD
G et al. Screening of Cushing’s syndrome in adult patients et al. Improvement in glycemic control of type 2 diabetes
with newly diagnosed diabetes mellitus. Clin Endocrinol after successful treatment of hepatitis C virus. Diabetes Care.
(Oxf). 2007;67:225–229. DOI: 10.1111/j.1365‐2265.2007. 2017;40:1173–1180. DOI: 10.2337/dc17‐0485.
02865.x. 87. Chakkera HA, Mandarino LJ. Calcineurin inhibition and
76. Boscaro M, Arnaldi G. Approach to the patient with possible new‐onset diabetes mellitus after transplantation. Trans
Cushing’s syndrome. J Clin Endocrinol Metab. 2009;94:3121– plantation. 2013;95:647–652. DOI: 10.1097/TP.0b013e31826
3131. DOI: 94/9/3121 [pii] 10.1210/jc.2009‐0612. e592e.
77. Terzolo M, Reimondo G, Chiodini I, Castello R, Giordano R, 88. Cosio FG, Pesavento TE, Kim S, Osei K, Henry M, Ferguson
Ciccarelli E et al. Screening of Cushing’s syndrome in outpa RM. Patient survival after renal transplantation: IV. Impact
tients with type 2 diabetes: results of a prospective multicen of post‐transplant diabetes. Kidney Int. 2002;62:1440–1446.
tric study in Italy. J Clin Endocrinol Metab. 2012;97: DOI: kid582 [pii] 10.1111/j.1523–1755.2002.kid582.x.
3467–3475. DOI: 10.1210/jc.2012‐1323. 89. Elmagd MM, Bakr MA, Metwally AH, Wahab AM. Clinico
78. Budyal S, Jadhav SS, Kasaliwal R, Patt H, Khare S, Shivane V epidemiologic study of posttransplant diabetes after living‐
et al. Is it worthwhile to screen patients with type 2 diabetes donor renal transplant. Exp Clin Transplant. 2008;6:42–47.
mellitus for subclinical Cushing’s syndrome? Endocrine 90. Kasiske BL, Snyder JJ, Gilbertson D, Matas AJ. Diabetes mel
Connections. 2015;4:242–248. DOI: 10.1530/EC‐15‐0078. litus after kidney transplantation in the United States. Am J
79. Nieman L, Biller B, Findling J, Newell‐Price J, Savage M, Transplant. 2003;3:178–185. DOI: 10.1034/j.1600‐6143.2003.
Stewart P et al. The diagnosis of Cushing’s syndrome: an 00010.x.
Endocrine Society Clinical Practice Guideline. Eur J 91. Miles AM, Sumrani N, Horowitz R, Homel P, Maursky V,
Endocrinol. 2009. DOI: EJE‐09‐0695 [pii] 10.1530/EJE‐09 Markell MS et al. Diabetes mellitus after renal transplanta
‐0695. tion: as deleterious as non‐transplant‐associated diabetes?
80. Kuypers DR, Claes K, Bammens B, Evenepoel P, Transplantation. 1998;65:380–384. DOI: 10.1097/00007890‐
Vanrenterghem Y. Early clinical assessment of glucose 199802150‐00014.
metabolism in renal allograft recipients: diagnosis and pre 92. Londero TM, Giaretta LS, Farenzena LP, Manfro RC, Canani
diction of post‐transplant diabetes mellitus (PTDM). LH, Lavinsky D et al. Microvascular complications of post
Nephrol Dial Transplant. 2008;23:2033–2042. DOI: 10.1093/ transplant diabetes mellitus in kidney transplant recipients:
ndt/gfm875. a longitudinal study. J Clin Endocrinol Metab. 2019;104:557–
81. Chowdhury TA. Post‐transplant diabetes mellitus. Clin 567. DOI: 10.1210/jc.2018‐01521.
Med (Lond). 2019;19:392–395. DOI: 10.7861/clinmed. 93. Sharif A, Baboolal K. Complications associated with new‐
2019‐0195. onset diabetes after kidney transplantation. Nat Rev Nephrol.
82. Munshi VN, Saghafian S, Cook CB, Werner KT, Chakkera 2011;8:34–42. DOI: 10.1038/nrneph.2011.174.
HA. Comparison of post‐transplantation diabetes mellitus 94. American Diabetes A. 2. Classification and diagnosis of dia
incidence and risk factors between kidney and liver trans betes: standards of medical care in diabetes – 2019. Diabetes
plantation patients. PLoS One. 2020;15:e0226873. DOI: Care. 2019;42:S13–S28. DOI: 10.2337/dc19‐S002.
10.1371/journal.pone.0226873. 95. Bodziak KA, Hricik DE. New‐onset diabetes mellitus after
83. Hackman KL, Snell GI, Bach LA. Prevalence and predictors solid organ transplantation. Transpl Int. 2009;22:519–530.
of diabetes after lung transplantation: a prospective, longitu DOI: TRI800 [pii] 10.1111/j.1432‐2277.2008.00800.x.
dinal study. Diabetes Care. 2014;37:2919–2925. DOI: 10.2337/ 96. Vanhove T, Remijsen Q, Kuypers D, Gillard P. Drug‐drug
dc14‐0663. interactions between immunosuppressants and antidiabetic
84. Rodrigo E, Fernandez‐Fresnedo G, Valero R, Ruiz JC, Pinera drugs in the treatment of post‐transplant diabetes mellitus.
C, Palomar R et al. New‐onset diabetes after kidney trans Transplantation Reviews (Orlando, Fla). 2017;31:69–77. DOI: 10.
plantation: risk factors. J Am Soc Nephrol. 2006;17:S291–S295. 1016/j.trre.2016.09.001.
DOI: 10.1681/ASN.2006080929. 97. Oberholzer J, Thielke J, Hatipoglu B, Testa G, Sankary HN,
85. Lecube A, Hernandez C, Genesca J, Simo R. Glucose abnor Benedetti E. Immediate conversion from tacrolimus to
malities in patients with hepatitis C virus infection: epidemi cyclosporine in the treatment of posttransplantation diabetes
ology and pathogenesis. Diabetes Care. 2006;29:1140–1149. mellitus. Transplant Proc. 2005;37:999–1000. DOI: 10.1016/
DOI: 10.2337/diacare.2951140. j.transproceed.2004.12.085.
98. Rouillon F, Sorbara F. Schizophrenia and diabetes: epide 109. Hui Fang CE, Rafey MF, Cunningham A, Dinneen SF,
miological data. European Psychiatry. 2005;20:S345–S348. Finucane FM. Risperidone‐induced type 2 diabetes pre
DOI: https://doi.org/10.1016/S0924‐9338(05)80189‐0. senting with diabetic ketoacidosis. Endocrinol Diabetes
99. De Hert M, Dekker JM, Wood D, Kahl KG, Holt RI, Moller Metab Case Rep. 2018;May 10. DOI: 10.1530/EDM‐18‐0031.
HJ. Cardiovascular disease and diabetes in people with 110. Consensus development conference on antipsychotic
severe mental illness position statement from the European drugs and obesity and diabetes. Diabetes Care. 2004;27:
Psychiatric Association (EPA), supported by the European w596–601.
Association for the Study of Diabetes (EASD) and the 111. Citrome L. Risk‐benefit analysis of available treatments for
European Society of Cardiology (ESC). Eur Psychiatry. schizophrenia. Psychiatric Times. 2007;24:27–30.
2009;24:412–424. DOI: 10.1016/j.eurpsy.2009.01.005. 112. Ader M, Kim SP, Catalano KJ, Ionut V, Hucking K, Richey
100. Cohen D, Stolk RP, Grobbee DE, Gispen‐de Wied CC. JM et al. Metabolic dysregulation with atypical antipsy
Hyperglycemia and diabetes in patients with schizophrenia chotics occurs in the absence of underlying disease: a pla
or schizoaffective disorders. Diabetes Care. 2006;29:786–791. cebo‐controlled study of olanzapine and risperidone in
DOI: 10.2337/diacare.29.04.06.dc05‐1261. dogs. Diabetes. 2005;54:862–871. DOI: 54/3/862 [pii].
101. Kamble P, Sherer J, Chen H, Aparasu R. Off‐label use of 113. Whicher CA, Price HC, Holt RIG. Mechanisms in endocri
second‐generation antipsychotic agents among elderly nology: antipsychotic medication and type 2 diabetes and
nursing home residents. Psychiatric Services. 2010;61:130– impaired glucose regulation. Eur J Endocrinol.
136. DOI: 10.1176/ps.2010.61.2.130. 2018;178:R245–R258. DOI: 10.1530/EJE‐18‐0022.
102. Llorente MD, Urrutia V. Diabetes, Psychiatric disorders, 114. Churchward S OS, Olotu VO, Thalitaya MD. Setting stand
and the metabolic effects of antipsychotic medications. ards for physical health monitoring in patients on antipsy
Clinical Diabetes. 2006;24:18–24. DOI: 10.2337/diaclin. chotics. Psychiatric Bulletin. 2009;33:451–454.
24.1.18. 115. Morse CG, Kovacs JA. Metabolic and skeletal complica
103. McCreadie R, Macdonald E, Blacklock C, Tilak‐Singh D, tions of HIV infection: the price of success. JAMA.
Wiles D, Halliday J et al. Dietary intake of schizophrenic 2006;296:844–854. DOI: 296/7/844 [pii] 10.1001/
patients in Nithsdale, Scotland: case‐control study. BMJ. jama.296.7.844.
1998;317:784–785. DOI: 10.1136/bmj.317.7161.784. 116. Brown TT, Cole SR, Li X, Kingsley LA, Palella FJ, Riddler
104. Thakore JH, Mann JN, Vlahos I, Martin A, Reznek R. SA et al. Antiretroviral therapy and the prevalence and
Increased visceral fat distribution in drug‐naive and drug‐ incidence of diabetes mellitus in the multicenter AIDS
free patients with schizophrenia. Int J Obes Relat Metab cohort study. Arch Intern Med. 2005;165:1179–1184. DOI:
Disord. 2002;26:137–141. DOI: 10.1038/sj.ijo.0801840. 10.1001/archinte.165.10.1179.
105. Freemantle N, Holmes J, Hockey A, Kumar S. How strong 117. Wohl DA, McComsey G, Tebas P, Brown TT, Glesby MJ,
is the association between abdominal obesity and the inci Reeds D et al. Current concepts in the diagnosis and man
dence of type 2 diabetes? International Journal of Clinical agement of metabolic complications of HIV infection and
Practice. 2008;62:1391–1396. DOI: 10.1111/j.1742–1241. its therapy. Clin Infect Dis. 2006;43:645–653. DOI: 10.1086/
2008.01805.x. 507333.
106. Smith M, Hopkins D, Peveler RC, Holt RI, Woodward M, 118. De Wit S, Sabin CA, Weber R, Worm SW, Reiss P, Cazanave
Ismail K. First‐ v. second‐generation antipsychotics and risk C et al. Incidence and risk factors for new‐onset diabetes in
for diabetes in schizophrenia: systematic review and meta‐ HIV‐infected patients: the Data Collection on Adverse
analysis. Br J Psychiatry. 2008;192:406–411. DOI: 10.1192/ Events of Anti‐HIV Drugs (D:A:D) study. Diabetes Care.
bjp.bp.107.037184. 2008;31:1224–1229. DOI: dc07‐2013 [pii] 10.2337/dc07
107. Bobo WV, Cooper WO, Stein CM, Olfson M, Graham D, ‐2013.
Daugherty J et al. Antipsychotics and the risk of type 2 dia 119. Nduka CU, Stranges S, Kimani PK, Sarki AM, Uthman
betes mellitus in children and youth. JAMA Psychiatry. OA. Is there sufficient evidence for a causal association
2013;70:1067–1075. DOI: 10.1001/jamapsychiatry.2013.2053. between antiretroviral therapy and diabetes in HIV‐
108. Taylor DM, McAskill R. Atypical antipsychotics and infected patients? A meta‐analysis. Diabetes Metab Res
weight gain – a systematic review. Acta Psychiatrica Rev. 2017;33. DOI: 10.1002/dmrr.2902.
Scandinavica. 2000;101:416–432. DOI: 10.1034/j.1600‐0447 120. Echecopar‐Sabogal J, D’Angelo‐Piaggio L, Chanamé‐Baca
.2000.101006416.x. DM, Ugarte‐Gil C. Association between the use of protease
inhibitors in highly active antiretroviral therapy and inci linked to mutations in the hepatocyte nuclear factor‐1alpha
dence of diabetes mellitus and/or metabolic syndrome in (MODY3) gene. Diabetes Care. 2005;28:2774–2776. DOI:
HIV‐infected patients: a systematic review and meta‐analysis. 10.2337/diacare.28.11.2774.
Int J STD AIDS. 2018;29:443–452. DOI: 10.1177/0956462417 132. Hattersley A, Bruining J, Shield J, Njolstad P, Donaghue
732226. KC. The diagnosis and management of monogenic diabetes
121. Prioreschi A, Munthali RJ, Soepnel L, Goldstein JA, in children and adolescents. Pediatr Diabetes. 2009;10
Micklesfield LK, Aronoff DM et al. Incidence and preva Suppl 12:33–42. DOI: PDI571 [pii] 10.1111/j.1399‐5448.2009.
lence of type 2 diabetes mellitus with HIV infection in 00571.x.
Africa: a systematic review and meta‐analysis. BMJ Open. 133. De Franco E, Flanagan SE, Houghton JA, Lango Allen H,
2017;7:e013953. DOI: 10.1136/bmjopen‐2016‐013953. Mackay DJ, Temple IK et al. The effect of early, comprehen
122. Kim PS, Woods C, Georgoff P, Crum D, Rosenberg A, sive genomic testing on clinical care in neonatal diabetes:
Smith M et al. A1C underestimates glycemia in HIV infec an international cohort study. Lancet. 2015;386:957–963.
tion. Diabetes Care. 2009;32:1591–1593. DOI: 10.2337/ DOI: 10.1016/S0140‐6736(15)60098‐8.
dc09‐0177. 134. Esfahani K, Roudaia L, Buhlaiga N, Del Rincon SV, Papneja
123. Kleinberger JW, Pollin TI. Undiagnosed MODY. Time for N, Miller WH, Jr. A review of cancer immunotherapy: from
Action. Curr Diab Rep. 2015;15:110. DOI: 10.1007/s11892‐ the past, to the present, to the future. Curr Oncol.
015‐0681‐7. 2020;27:S87–S97. DOI: 10.3747/co.27.5223.
124. O’Rahilly S. Human genetics illuminates the paths to meta 135. Higham CE, Olsson‐Brown A, Carroll P, Cooksley T,
bolic disease. Nature. 2009;462:307–314. DOI: nature08532 Larkin J, Lorigan P et al. Society for Endocrinology
[pii] 10.1038/nature08532. Endocrine Emergency Guidance. Acute management of
125. Philipson LH, Murphy R, Ellard S, Hattersley AT, Stoy J, the endocrine complications of checkpoint inhibitor ther
Greeley SAW et al. Chapter 2 – genetic testing in diabetes apy. Endocr Connect. 2018;7:G1–G7. DOI: 10.1530/EC‐
mellitus: a clinical guide to monogenic diabetes. In: Weiss 18‐0068.
RE, Refetoff S, eds. Genetic Diagnosis of Endocrine 136. de Filette JMK, Pen JJ, Decoster L, Vissers T, Bravenboer B,
Disorders. (First Edition). San Diego: Academic Press; Van der Auwera BJ et al. Immune checkpoint inhibitors and
2010:17–26. type 1 diabetes mellitus: a case report and systematic review.
126. Barrett TG, Bundey SE, Macleod AF. Neurodegeneration and Eur J Endocrinol. 2019;181:363–374. DOI: 10.1530/EJE‐
diabetes: UK nationwide study of Wolfram (DIDMOAD) 19‐0291.
syndrome. Lancet. 1995;346:1458–1463. DOI: 10.1016/ 137. Stamatouli AM, Quandt Z, Perdigoto AL, Clark PL, Kluger
s0140‐6736(95)92473‐6. H, Weiss SA et al. Collateral damage: insulin‐dependent
127. Naylor R, Knight Johnson A, del Gaudio D. Maturity‐Onset diabetes induced with checkpoint inhibitors. Diabetes. 2018;
Diabetes of the Young Overview. 2018 May 24. In: Adam 67:1471–1480. DOI: 10.2337/dbi18‐0002.
MP, Ardinger HH, Pagon RA et al. editors. GeneReviews® 138. Wright JJ, Salem JE, Johnson DB, Lebrun‐Vignes B,
[Internet]. Seattle (WA): University of Washington, Seattle; Stamatouli A, Thomas JW et al. Increased reporting of
1993–2020. Available from: https://www.ncbi.nlm.nih.gov/ immune checkpoint inhibitor‐associated diabetes. Diabetes
books/NBK500456/. Care. 2018;41:e150–e151. DOI: 10.2337/dc18‐1465.
128. McDonald TJ, Ellard S. Maturity onset diabetes of the 139. Castinetti F, Borson‐Chazot F. [Immunotherapy‐induced
young: identification and diagnosis. Ann Clin Biochem. endocrinopathies: insights from the 2018 French Endocrine
2013;50:403–415. DOI: 10.1177/0004563213483458. Society Guidelines]. Bull Cancer. 2019;106:492–496. DOI:
129. Dickens LT, Naylor RN. Clinical management of women 10.1016/j.bulcan.2019.02.003.
with monogenic diabetes during pregnancy. Curr Diab Rep. 140. Barroso‐Sousa R, Barry WT, Garrido‐Castro AC, Hodi FS,
2018;18:12. DOI: 10.1007/s11892‐018‐0982‐8. Min L, Krop IE et al. Incidence of endocrine dysfunction fol
130. Kapoor RR, Locke J, Colclough K, Wales J, Conn JJ, lowing the use of different immune checkpoint inhibitor regi
Hattersley AT et al. Persistent hyperinsulinemic hypoglyce mens: a systematic review and meta‐analysis. JAMA Oncol.
mia and maturity‐onset diabetes of the young due to 2018;4:173–182. DOI: 10.1001/jamaoncol.2017.3064.
heterozygous HNF4A mutations. Diabetes. 2008;57:1659– 141. Puzanov I, Diab A, Abdallah K, Bingham CO, 3rd, Brogdon
1663. DOI: 10.2337/db07‐1657. C, Dadu R et al. Managing toxicities associated with
131. Malecki MT, Skupien J, Gorczynska‐Kosiorz S, Klupa T, immune checkpoint inhibitors: consensus recommenda
Nazim J, Moczulski DK et al. Renal malformations may be tions from the Society for Immunotherapy of Cancer
(SITC) Toxicity Management Working Group. J Immunother oma: impact of using supine reference intervals for plasma
Cancer. 2017;5:95. DOI: 10.1186/s40425‐017‐0300‐z. metanephrines with samples collected from fasted/seated
142. Griffin TP, Bogdanet D, Navin P, Callagy G, O’Shea PM, patients. Ann Clin Biochem. 2017;54:170–173. DOI:
Bell M. The importance of standardisation of measurement 10.1177/0004563216646395.
and reference intervals for detection of phaeochromocy 144. Griffin TP, Casey R, Wall D, Bell M, O’Shea PM.
toma and paraganglioma (PPGL). Ir J Med Sci. 2018;187: Evaluating the optimum rest period prior to blood col
993–998.DOI: 10.1007/s11845‐018‐1756‐7. lection for fractionated plasma free metanephrines
143. Casey R, Griffin TP, Wall D, Dennedy MC, Bell M, O’Shea analysis. Pract Lab Med. 2016;5:39–46. DOI: 10.1016/j.
PM. Screening for phaeochromocytoma and paragangli plabm.2016.05.001.
5 How to Screen Appropriately
for Monogenic Diabetes
Adrian Vella
LE A R N I N G P OI NT S diagnosis may also facilitate genetic counselling to affected

individuals and their families.
●● The term “Maturity Onset Diabetes of the Young” has Affected individuals were initially characterized on the
been superseded by the term “Monogenic Diabetes” basis of a family history suggesting an autosomal mode of
for reasons related to the underlying pathophysiology inheritance and the presence of residual insulin secretion –
and heterogeneity of the condition.
●● The commonest forms of monogenic diabetes are those
hence the term “Maturity‐Onset” implying that the disease
caused by mutations in GCK, HNF1A and HNF4A. behaves like type 2 diabetes albeit presenting at a younger
●● Extra‐pancreatic manifestations may help raise the age (at least one member of an affected kindred diagnosed
clinical suspicion of monogenic diabetes and suggest a before age 25). Mutations in at least seven different genes
genetic diagnosis.
have been identified thereby accounting for the significant
●● A logistic regression model using demographic data
and disease history can be used to predict whether
heterogeneity in clinical presentation with different age at
genetic testing is indicated. presentation, differences in the pattern of hyperglycemia
(due in part to the severity of insulin deficiency) and the
presence or absence of extra‐pancreatic manifestations [4].
Accordingly, monogenic diabetes has become the preferred
term for describing disease(s) previously characterized as
Maturity onset diabetes of the young
MODY [5].
(MODY)
Diabetes is a clinical syndrome characterized by hypergly-
Differentiating monogenic diabetes
cemia [1]. The causation of diabetes in a given individual
from type 1 diabetes
may be heterogeneous but is broadly categorized into insu-
lin‐mediated diabetes (type 1 diabetes) and non‐insulin The diagnosis of neonatal diabetes (< 6 months of age) is
mediated diabetes (type 2 diabetes) [2]. However, it is read- beyond the scope of this chapter. However, such forms of
ily apparent that a small minority of patients have a mono- diabetes often have a monogenic cause and have may have
genic disorder underlying their diabetes [3]. These associated (“syndromic”) clinical features. Diabetes may be
disorders have been collectively termed “Maturity‐Onset transient in some monogenic forms of neonatal diabetes
Diabetes of the Young (MODY)”. The challenge to recog- (e.g. mutations in KCNJ11) [6]. Type 1 diabetes rarely
nizing such patients is balancing expensive genetic screen- occurs before 6 months of age and, in fact, patients with per-
ing against the potential harm of misdiagnosis and manent diabetes presenting before 6 months of age do not
consequent failure to optimize therapy. Making a correct have increased genetic susceptibility to type 1 diabetes [7].

68
How to screen appropriately for Monogenic Diabetes 69
In contrast, in diabetes presenting after 6 months of age, However, applying such testing to a group of patients that
type 1 diabetes is the commonest underlying cause. are less rigorously selected decreases the prior probability
As discussed elsewhere, the presence of detectable C‐pep- of monogenic diabetes being present and increases the
tide – a marker of endogenous insulin secretion – is unhelp- likelihood of false positive findings. Also, benign polymor-
ful in making a diagnosis of type 1 diabetes at the time of phisms may be reported as disease‐causing variants in the
presentation with hyperglycemia and/or ketosis. The pres- absence of an understanding of the structural conse-
ence of autoantibodies associated with an autoimmune pro- quences of a coding change [11].
cess in the pancreatic islets is certainly helpful in making a Traditional clinical expertise has focused on features of
diagnosis. However, in their absence, a genetic cause should the history (and family history), a physical examination
be considered especially if there is a two‐ or three‐genera- and specialized testing to identify features associated with
tion family history of diabetes. Persistence of endogenous the syndromes of monogenic diabetes. C‐peptide and islet
insulin secretion for more than 5 years after initial presenta- antibody testing has significant limitations and family his-
tion should also question a diagnosis of type 1 diabetes [8]. tory may not be well characterized [12, 13].
The decreasing costs of sequencing large swathes of the Using research databases where type 1 diabetes was
genome has resulted in the development of genetic panels defined by the use of insulin within 6 months of diagnosis
that can simultaneously test for multiple sequence changes and where a genetic diagnosis was present (mutation test-
associated with different forms of monogenic diabetes. This ing in GCK, HNF1A and HNF4A) Shields et al. developed a
has decreased the need for a priori selection of genes for clinical prediction model to identify patients with a high
sequencing most likely to explain such presentations. likelihood of having monogenic diabetes. This model
Nevertheless, mutations in HNF1A are most likely to explain improves the sensitivity and specificity of the diagnosis
a presentation similar to that of type 1 diabetes [5]. compared to using the prior criteria of having an affected
parent and an age < 25 at the time of diagnosis [14].
Differentiating monogenic diabetes The factors that discriminated monogenic diabetes from
from type 2 diabetes type 1 diabetes were an older age at diagnosis, better glycemic
control (lower HbA1c), female sex and a parent with diabetes.
Type 2 diabetes is an imprecisely defined disease where age at
To discriminate from type 2 diabetes, the model identified
presentation has decreased because of the increased preva-
body mass index (lower), female sex, better glycemic control
lence of obesity in the population [1]. There are numerous
(lower HbA1c), the absence of treatment with insulin or oral
reports of presentation in adolescents and young adults with a
hypoglycemic agents, and a parent with diabetes. A version of
subsequent relatively rapid progression to severe β‐cell dys-
the logistic regression models that identified these associa-
function [9]. The absence of obesity, the presence of a multi-
tions is now available online (www.diabetesgenes.org) and as
generational family history of diabetes and no evidence of
Diabetes Diagnostic application for mobile platforms.
insulin resistance such as acanthosis nigricans may suggest a
This probalistic assessment should be combined with
diagnosis other than that of type 2 diabetes. Disorders of tri-
islet autoantibody testing and C‐peptide analysis in
glyceride or lipid metabolism do not exclude monogenic dia-
patients treated with insulin. The presence of autoantibod-
betes since these dyslipidemias can be observed in diabetes
ies or a C‐peptide < 200pmol/l makes a diagnosis of mono-
due to GCK or HNF4A mutations respectively [5, 10].
genic diabetes unlikely [15].
When to screen for monogenic diabetes? Common forms of monogenic diabetes

In the absence of blanket testing, the challenge is to identify GCK mutations
the 1–2% of the population with diabetes who may have an Glucokinase catalyzes the phosphorylation of glucose
underlying monogenic form of diabetes. The falling cost of which in practice is the rate‐limiting step of glycolysis and
genetic characterization using a gene panel approach decreases of glucose uptake in the liver, skeletal muscle and in the
the need to define the genetic etiology prior to testing. β‐cell. It acts as a glucose sensor in the islet. Heterozygous
inactivating mutations of this enzyme produce an altered and is characterized by diabetes, diabetes insipidus and
set point, increasing the threshold for insulin secretion. optic atrophy. Common variants in the same gene are asso-
Consequently, affected individuals typically have mild ciated with type 2 diabetes [19]. Wolfram syndrome is an
hyperglycemia and are asymptomatic. Since this is a set‐ autosomal recessive disorder.
point abnormality, the β‐cell response to an oral challenge HNF1B mutations are associated with the presence of renal
is intact so that HbA1c rarely exceed 7.5% and microvascu- disease that may range from cysts to renal dysplasia and other
lar complications are rare. forms of developmental renal abnormalities. Genitourinary
Unfortunately, a subset of patients with prediabetes have tract abnormalities may also be present. Some common vari-
impaired fasting glucose and normal glucose tolerance – in ants in HNF1B alter predisposition to type 2 diabetes and to
effect behaving like people with GCK mutations [16]. prostate cancer [19]. The degree of pancreatic endocrine dys-
Recent evidence suggests that these individuals progress to function that is usually present means that sulfonylureas are
type 2 diabetes at the same rate as do those with impaired often ineffective in these patients.
glucose tolerance so that a genetic test may be necessary to
counsel patients appropriately [17]. Typically, patients with Conclusions
GCK mutations do not require glucose‐lowering therapy.
However, pregnant women with this mutation run the risk Monogenic diabetes is often overlooked as a diagnostic
of fetal macrosomia if their child does not inherit this possibility but should be considered when patients present
mutation and secretes excess insulin (stimulated by mater- with diabetes at an “atypical” age for either type 1 or type 2
nal hyperglycemia) which promotes fetal growth [18]. diabetes, have a multigenerational family history, no evi-
dence of autoantibodies and evidence of residual insulin
HNF1A mutations secretion. The commonest causes of monogenic diabetes
Affected patients typically develop diabetes in their teens encountered in clinical practice are those caused by muta-
and experience progressive β‐cell failure. HNF1A is a tran- tions in GCK, HNF1A, and HNF4A. In such circumstances
scription factor necessary for pancreatic islet development use of a probalistic model [14] may help determine whether
and functioning. The risk of microvascular complications it is reasonable to screen for these disorders. In other situa-
in affected individuals is related to poor glycemic control as tions, appropriate clinical suspicion coupled with targeted
observed in other forms of diabetes. The insulin secretory detection of syndromic extra‐pancreatic abnormalities
apparatus is intact in affected β‐cells and therefore sulfony- may help guide the decision to perform genetic testing and
lureas are able to stimulate insulin secretion even when identify other less common forms of monogenic diabetes.
patients are treated with insulin.
HNF4A mutations
References
The clinical features of affected patients overlap signifi- 1. Smushkin G, Vella A. What is type 2 diabetes? Medicine
cantly with those affected by HNF1A mutations and are (Baltimore). 2010;38:597–601.
characterized by progressive β‐cell dysfunction which is 2. Gale EA. Declassifying diabetes. Diabetologia.
responsive to sulfonylurea treatment. Dyslipidemia with 2006;49:1989–1995.
elevated LDL‐cholesterol and decreased HDL‐cholesterol 3. Fajans SS, Bell GI, Polonsky KS. Molecular mechanisms and
is a common accompanying feature. clinical pathophysiology of maturity‐onset diabetes of the
young. N Engl J Med. 2001;345:971–980.
Around 10% of families with autosomal dominant β‐cell
4. Hattersley AT, Pearson ER. Minireview: pharmacogenetics
dysfunction do not have an identified genetic diagnosis
and beyond: the interaction of therapeutic response, beta‐
implying the presence of undiscovered genetic mutations.
cell physiology, and genetics in diabetes. Endocrinology.
Rare forms of monogenic diabetes include those caused by 2006;147:2657–2663.
mutations in NEUROD1 and PAX1 – both genes encoding 5. Murphy R, Ellard S, Hattersley AT. Clinical implications of a
transcription factors important in early β‐cell develop- molecular genetic classification of monogenic beta‐cell
ment. Wolfram syndrome is caused by mutations in WFS1 diabetes. Nat Clin Pract Endocrinol Metab. 2008;4:200–213.
How to screen appropriately for Monogenic Diabetes 71
6. Stanley CA. Perspective on the Genetics and Diagnosis of Voight BF, Hansen T, Tuomi T, Boehm BO, Groop L, Leslie
Congenital Hyperinsulinism Disorders. J Clin Endocrinol RD, Grant SFA. First genome‐wide association study of
Metab. 2016;101:815–826. latent autoimmune diabetes in adults reveals novel insights
7. Flanagan SE, Patch AM, Mackay DJ, Edghill EL, Gloyn AL, linking immune and metabolic diabetes. Diabetes Care.
Robinson D, Shield JP, Temple K, Ellard S, Hattersley AT. 2018;41:2396–2403.
Mutations in ATP‐sensitive K+ channel genes cause 13. Mishra R, Akerlund M, Cousminer DL, Ahlqvist E, Bradfield
transient neonatal diabetes and permanent diabetes in JP, Chesi A, Hodge KM, Guy VC, Brillon DJ, Pratley RE,
childhood or adulthood. Diabetes. 2007;56:1930–1937. Rickels MR, Vella A, Ovalle F, Harris RI, Melander O, Varvel
8. Effect of intensive therapy on residual beta‐cell function in S, Hakonarson H, Froguel P, Lonsdale JT, Mauricio D,
patients with type 1 diabetes in the diabetes control and Schloot NC, Khunti K, Greenbaum CJ, Yderstraede KB,
complications trial. A randomized, controlled trial. The Tuomi T, Voight BF, Schwartz S, Boehm BO, Groop L, Leslie
Diabetes Control and Complications Trial Research Group. RD, Grant SFA. Genetic discrimination between LADA and
Ann Intern Med. 1998;128:517–523. childhood‐onset type 1 diabetes within the MHC. Diabetes
9. Mechanick JI, Adams S, Davidson JA, Fergus IV, Galindo RJ, Care. 2020;43:418–425.
McKinney KH, Petak SM, Sadhu AR, Samson SL, Vedanthan 14. Shields BM, McDonald TJ, Ellard S, Campbell MJ, Hyde C,
R, Umpierrez GE. Transcultural Diabetes Care. in the United Hattersley AT. The development and validation of a clinical
States – a Position Statement by the American Association of prediction model to determine the probability of MODY in
Clinical Endocrinologists. Endocr Pract. 2019;25:729–765. patients with young‐onset diabetes. Diabetologia.
10. Agius L. Targeting hepatic glucokinase in type 2 diabetes: 2012;55:1265–1272.
weighing the benefits and risks. Diabetes. 2009;58:18–20. 15. Hattersley AT, Patel KA. Precision diabetes: learning from
11. Awa WL, Thon A, Raile K, Grulich‐Henn J, Meissner T, monogenic diabetes. Diabetologia. 2017;60:769–777.
Schober E, Holl RW, DPV‐Wiss. Study Group. Genetic and 16. Bock G, Chittilapilly E, Basu R, Toffolo G, Cobelli C,
clinical characteristics of patients with HNF1A gene varia- Chandramouli V, Landau BR, Rizza RA. Contribution of
tions from the German‐Austrian DPV database. Eur J hepatic and extrahepatic insulin resistance to the pathogen-
Endocrinol. 2011;164:513–520. esis of impaired fasting glucose: role of increased rates of
12. Cousminer DL, Ahlqvist E, Mishra R, Andersen MK, Chesi gluconeogenesis. Diabetes. 2007;56:1703–1711.
A, Hawa MI, Davis A, Hodge KM, Bradfield JP, Zhou K, Guy 17. Han SJ, Kim HJ, Kim DJ, Lee KW, Cho NH. Incidence and
VC, Akerlund M, Wod M, Fritsche LG, Vestergaard H, predictors of type 2 diabetes among Koreans: a 12‐year fol-
Snyder J, Hojlund K, Linneberg A, Karajamaki A, Brandslund low up of the Korean Genome and Epidemiology Study.
I, Kim CE, Witte D, Sorgjerd EP, Brillon DJ, Pedersen O, Diabetes Res Clin Pract. 2017;123:173–180.
Beck‐Nielsen H, Grarup N, Pratley RE, Rickels MR, Vella A, 18. Spyer G, Hattersley AT, Sykes JE, Sturley RH, MacLeod KM.
Ovalle F, Melander O, Harris RI, Varvel S, Grill VER, Bone Influence of maternal and fetal glucokinase mutations in ges-
Mineral Density in Childhood S, Hakonarson H, Froguel P, tational diabetes. Am J Obstet Gynecol. 2001;185:240–241.
Lonsdale JT, Mauricio D, Schloot NC, Khunti K, Greenbaum 19. Smushkin G, Vella A. Genetics of type 2 diabetes. Curr Opin
CJ, Asvold BO, Yderstraede KB, Pearson ER, Schwartz S, Clin Nutr Metab Care. 2010;13:471–477.
PART II
Initial Evaluation and Management
of Diabetes
6 Managing Gestational Diabetes During
and After Pregnancy
Aoife M. Egan1 and Fidelma P. Dunne2
1
Division of Endocrinology, Diabetes, Metabolism and nutrition, Mayo Clinic, Rochester, MN, USA
2
Galway Diabetes Research Centre, National University of Ireland, Galway, Ireland
LE A R NI NG P OI NT S In order to preserve carbohydrate for the growing fetus,

insulin resistance increases in the second and early third
●● Gestational Diabetes Mellitus (GDM) is a common trimesters of pregnancy [4]. GDM will develop if the com
disorder associated with adverse pregnancy outcomes. pensatory pancreatic response with increased insulin
●● Although the ideal diagnostic strategy remains unclear, secretion is inadequate. Due to the transfer of glucose
it is typically identified using an oral glucose tolerance across the placenta down a concentration gradient, mater
test (OGTT) at 24–28 weeks’ gestation.
nal hyperglycemia will lead to fetal hyperglycemia with
●● The majority of women can successfully treat GDM
with lifestyle changes. subsequent fetal hyperinsulinemia, which drives the devel
●● Women with GDM have a high future risk of type 2 opment of pregnancy complications [5]. The largest impact
diabetes and require lifelong surveillance. of GDM involves increasing birth weight and the propor
tion of infants that are classified as large for gestational age
(LGA) [6]. Beyond this, a diagnosis of GDM also confers
an increased risk of multiple additional complications
Introduction including pre‐eclampsia [7], perineal trauma [8], and
cesarean delivery [9] for the mother; and neonatal birth
Gestational diabetes mellitus (GDM) is defined as glucose injury, neonatal hypoglycemia [10], jaundice, and even
intolerance with onset or first recognition during preg stillbirth [11] for the infant. GDM is a strong risk factor for
nancy, excluding those with diabetes likely to represent future type 2 diabetes in the mother (relative risk (RR)
overt diabetes mellitus [1]. It affects approximately 13% compared to women without GDM 7.43, 95% confidence
pregnancies and is associated with adverse pregnancy out interval (CI) 4.79–11.51) [12], and offspring of women
comes [2]. Risk factors for GDM reflect those for type 2 with GDM are also at high risk of metabolic complications
diabetes and include: overweight or obese body mass including obesity and diabetes [13, 14].
index, GDM in a prior pregnancy (or previous delivery of a
macrosomic baby), increasing age, non‐European ethnic
ity, family history of Type 2 diabetes and conditions associ
GDM diagnosis
ated with insulin resistance such as polycystic ovarian One might argue that an accurate diagnosis is critical for
syndrome [3]. optimal management of any medical condition. However,

75
76 Initial Evaluation and Management of Diabetes
in the case of GDM this is fraught with difficulty. Central to women with mean glucose levels at 24–28 weeks in the
this is the strong, continuous association between increas Hyperglycemia and Neonatal Outcomes (HAPO)
ing maternal glucose levels and key adverse outcomes such study [10, 17, 18]. These criteria based on pregnancy out
as increased birth weight, neonatal hypoglycemia and pri comes are also endorsed by a number of additional organi
mary cesarean delivery [10]. The lack of an obvious thresh zations including the European Board and College of
old at which risk increases makes it challenging to choose Obstetrics and Gynaecology (EBCOG) [19], the
an appropriate diagnostic cut‐off, especially once addi International Federation of Gynecology and Obstetrics
tional factors such as clinical benefit and cost‐effectiveness (FIGO) [20] and the European Association of Perinatal
of treatments are considered. This has resulted in the use of Medicine (EAPM) [21]. The American Diabetes
countless strategies for screening and diagnosing GDM Association (ADA) also endorses this approach but state
that vary from country to country and even center to that alternative, two‐step approaches are also accepta
center [15, 16]. Table 6.1 outlines the more commonly used ble [22]. The diagnostic cut‐offs used in the two‐step
approaches internationally. One should note that the one‐ approaches are largely based on thresholds that optimize
step, International Association of Diabetes and Pregnancy predictive ability for subsequent development of maternal
Study Groups (IADPSG)/World Health Organization diabetes, rather than pregnancy outcomes [23].
(WHO) criteria were chosen primarily based on an odds The majority of clinical centers conduct GDM screening
ratio for adverse outcomes of at least 1.75 compared to at 24–28 weeks gestation, but it is known that accelerated
TABLE 6.1 Options for GDM screening and diagnosis.
Number of abnormal Glucose cut‐offs

values required Oral glucose load mg/dL (mmol/L)
Two‐step strategies:
Non‐fasting glucose challenge test 1 50 g ≥ 130, 135 or 140
(7.2, 7.5 or 7.8)
Followed by either of the options below:
1. Carpenter and Coustan ≥2a 100 g Fasting ≥ 95 (5.3)
1 hour ≥ 180 (10.0)
2 hour ≥ 155 (8.6)
3 hour ≥ 140 (7.8)
2. NDDG ≥2a 100 g Fasting ≥ 105 (5.8)
1 hour ≥ 190 (10.6)
2 hour ≥ 165 (9.2)
3 hour ≥ 145 (8.0)
One‐step strategies:
2010 IADPSG/2013 WHO ≥1 75 g Fasting ≥ 92 (5.1)
1 hour ≥ 180 (10.0)
2 hour ≥ 153 (8.5)
2015 NICE ≥1 75 g Fasting ≥ 5.6 (100)
2 hour ≥ 7.8 (140)
EASD ≥1 75 g Fasting ≥ 6.0 (108)
2 hour ≥ 9.0 (162)
GDM: Gestational diabetes mellitus; IADPSG: International Association of Diabetes in Pregnancy Study Groups; NICE: National
Institute for Health and Care Excellence; OGTT: oral glucose tolerance test; WHO: World Health Organization; EASD: European
Association for the Study of Diabetes.
a
ACOG notes that one elevated value may be used for diagnosis.
Managing Gestational Diabetes During and After Pregnancy 77
fetal growth may occur earlier in pregnancy in women with with weight management. For most women, an exercise
GDM [24]. A significant amount of work including a large program with a goal of moderate‐intensity exercise of
randomized controlled trial is underway to examine 20–30 minutes per day is reasonable [32]. Observational
whether diagnosing and treating GDM in early pregnancy studies suggest that in routine clinical practice, the intro
improves pregnancy outcomes [25]. In the meantime, the duction of lifestyle modifications and remaining within
American College of Obstetricians and Gynecologists recommended guidelines for gestational weight gain are
(ACOG) accept the 24‐week criteria applied to a 100 g oral associated with improved clinical outcomes including a
glucose tolerance test (OGTT) in early pregnancy, but reduction in large‐for‐gestational age infants [33, 34]. A
there is minimal evidence to support this practice and the meta‐analysis of trials involving lifestyle interventions also
IADPSG have indicated that their criteria are not suitable demonstrated a reduction in large‐for‐gestational age neo
for use in early pregnancy [26, 27]. What is clear, is that nates and neonatal fat mass in those randomized to the
women with risk factors for type 2 diabetes should be tested intervention [35].
for undiagnosed pre‐gestational diabetes using standard Fasting and postprandial self‐monitoring of blood glu
diagnostic criteria at their initial prenatal visit [22, 26]. cose is recommended in women diagnosed with GDM [22].
Intensification of therapy should occur within 1–2 weeks if
glycemic targets are not met following lifestyle interven
GDM management
tion – this is generally defined as > 15–20% glucose read
Glycemic goals ings above goal. In addition, it is common practice for
The following glycemic goals are recommended for women pharmacological therapy to be introduced at diagnosis if
with GDM: fasting glucose ≤ 95 mg/dL (5.3 mmol/L), 1 the fasting glucose level is ≥ 126 mg/dL (7.0 mmol/L) as
hour after eating ≤ 140 mg/dL (7.8 mmol/L) and 2 hours this makes the presence of type 2 diabetes highly likely [36].
after eating ≤ 120 mg/dL (6.7 mmol/L) [22, 26, 28].
Insulin therapy
Lifestyle modification Insulin is the most commonly used agent for treatment of
Over 70% women with GDM diagnosed using a two‐step GDM and reassuringly, it has not been demonstrated to
approach can be treated with lifestyle modifications cross the placenta [37–39]. Traditionally isophane insulin
alone [22], and use of the more stringent IADPSG criteria (neutral protamine Hagedorn, NPH) was used as interme
increases this proportion [9]. Women should meet with a diate‐acting insulin (two‐three times per day) but the long‐
registered dietician or registered dietician nutritionist who acting insulin analogues detemir and glargine are
is familiar with the management of GDM [22]. An indi increasingly used and appear safe [40, 41]. Although more
vidualized plan should be developed, taking into account extensively studied in the setting of pregestational diabetes,
baseline BMI, cultural, and other dietary preferences. A aspart and lispro are also safe during pregnancy and are
minimum of 175 g of carbohydrates, 71 g of protein and 28 frequently used to treat GDM [42, 43]. A randomized con
g fiber is recommended based on the Dietary Reference trolled, open‐label trial is currently underway comparing
Intakes for pregnant women [22]. Although there is no aspart with faster‐acting aspart insulin (Fiasp) in women
clear guidance on the ideal total calorie intake, limiting with pre‐existing diabetes with completion estimated in
carbohydrates to 34–45% of total calories per day is recom 2022 [44]. In addition, a randomized trial is underway to
mended by the Endocrine Society [29]. Added sugars and evaluate insulin degludec versus insulin detemir in the
refined carbohydrates should be excluded from the diet treatment of pregnant women with type 1 diabetes and is
with a focus on carbohydrate sources such as vegetables, expected to be complete in 2021 [45]. Results from these
lower glycemic index fruits and whole grains [30]. trials will be useful when considering these newer insulin
Even light postprandial exercise decreases postprandial options for women with GDM.
glucose excursions in women with GDM [31] and physical Insulin regimens are typically tailored according to the
activity will maintain or improve physical fitness and assist pattern of maternal hyperglycemia – for example, if the
primary concern is fasting hyperglycemia, intermediate or based case‐control study (using controls with congenital
long‐acting insulin is initiated initially, with rapid‐acting malformations) did not find any evidence for an increased
insulin introduced to manage postprandial hyperglycemia. risk of all non‐genetic congenital anomalies combined fol
If a combination of long‐acting/intermediate‐acting and lowing first trimester metformin exposure, however, the
short‐acting insulin is required at the outset, ACOG sug risk of pulmonary valve atresia was increased among met
gest a starting total daily dosage of 0.7–1.0units/kg which formin exposed babies (aOR 3.54, CI 1.05 – 12.00 com
may be divided with a regimen of multiple injections [26]. pared with controls with non‐genetic anomalies) [54]. The
Women requiring basal‐bolus insulin therapy should moni Metformin in Gestational Diabetes (MiG) trial randomized
tor fasting, pre meal and postprandial glucose levels to allow 751 women to metformin or insulin and there was no sig
for better adjustment of insulin [22]. As pregnancy pro nificant difference in the composite fetal outcome (neona
gresses, postprandial glucose control can be impaired by tal hypoglycemia, respiratory distress, need for
slower glucose disposal [46], and the time‐to‐peak prandial phototherapy, birth trauma, 5‐minute Apgar < 7 or prema
insulin concentration is increasingly delayed [47]. Bolusing turity) between the two groups. However, there was a
meal insulin at least 15–30 minutes before eating may there higher rate of preterm birth among the metformin‐exposed
fore ameliorate postprandial glycemic excursions. infants (RR 1.6, CI 1.02–2.52, p=0.04) [55]. Another point
Insulin therapy is very effective for treating hyperglyce to note is that within the metformin group, 46.3% required
mia and in “real‐world” clinical settings, it appears to supplemental insulin, but the women preferred metformin
reduce rates of large‐for‐gestational age towards that of to insulin treatment.
women without GDM [48]. The main clinical concern with A follow‐up study (MiG trial participants) observed
insulin use is risk of maternal hypoglycemia and additional that children exposed to metformin (or a combination of
instruction is necessary to educate women on self‐injection insulin and metformin) in utero had similar total and
and recognition and treatment of hypoglycemia. abdominal body fat percent and metabolic markers at 7–9
years compared to those exposed to insulin only. However,
Oral glucose‐lowering agents at 9 years, metformin exposed offspring were larger by
Although not approved for use in pregnancy, metformin measures of weight, arm and waist circumferences [56]. A
and glyburide (glibenclamide) are commonly used for meta‐analysis examining outcomes of GDM‐affected preg
treatment of GDM [49]. Metformin is typically introduced nancies randomized to treatment with metformin versus
at a dose of 500 mg with the evening meal and titrated up insulin (28 studies, 3976 participants) found that met
to 1000 mg twice daily. It is a biguanide and exhibits its glu formin exposed children had a lower average birth weight
cose‐lowering effect by suppressing gluconeogenesis and (mean difference −107.7 g, 95% confidence interval −182.3
enhancing insulin suppression of endogenous glucose pro to −32.7, p = 0.005) but appeared to experience accelerated
duction in the liver. In addition, it possibly improves glu postnatal growth resulting in higher BMI by mid‐child
cose uptake and utilization by peripheral tissues and hood (mean difference 0.78 kg/m2, 95% confidence
reduces intestinal glucose absorption [50]. Slow upward interval 0.23 to 1.33, p=0.005) [57]. Whether these dif
titrations in the dose of metformin can reduce the risk of ferences will result in long‐term cardio‐metabolic out
gastrointestinal disturbance. It has been used since the late comes remains unknown. We await the results of the
1970s for women with type 2 diabetes and GDM, particu first placebo‐controlled metformin trial in GDM, which
larly in South Africa; and is commonly used in women is currently recruiting with estimated completion in
with polycystic ovarian syndrome pre‐conceptually and 2022 [58].
during pregnancy [51, 52]. Glyburide is a sulfonylurea and exerts its glucose‐lower
Metformin does cross the placenta in significant quanti ing effect by stimulating insulin secretion from the pancre
ties, and this has raised concerns in relation to both long‐ and atic beta cell. The primary mechanism of action is through
short‐term offspring effects [53]. A large population‐ closure of the APT‐sensitive potassium channels in the
plasma membrane of the beta cell. This initiates a chain of glyburide [63]. There are no data with respect to long‐
events culminating in insulin release [59]. Glyburide is term outcomes in infants exposed to glyburide in utero.
generally used at total daily doses of 2.5–20.0mg, although Guidelines and providers vary in terms of their approach to
a dose‐response relationship in pregnancy is not clear [60]. pharmacological intervention for GDM treatment. The ADA
In 2000, the first randomized controlled trial compared recommends insulin as first‐line therapy but do not express a
glyburide to insulin in 404 women with GDM [61]. Four preference of metformin over glyburide in situations where
mothers in the glyburide group required supplemental insulin may not be feasible. Although metformin is likely
insulin and there were no significant difference in multiple associated with a reduction in hypertensive disorders of preg
adverse outcomes including large‐for‐gestational age, nancy [64], it is not recommended for women with estab
hypoglycemia and admission to a neonatal intensive care lished preeclampsia, hypertension or those at risk of
unit. The study also reported that glyburide was not intrauterine growth restriction [22]. ACOG also recommend
detected in the cord serum of any infant in the glyburide insulin as first‐line therapy but suggest metformin (or rarely
group, however, subsequent work has demonstrated that glyburide) as a reasonable alternative [26]. On the other hand,
glyburide does cross the placenta [62]. Likely based on the the National Institute for Health and Care Excellence (NICE)
reassuring results of the aforementioned randomized con in the United Kingdom recommend metformin as first‐line
trolled trial, glyburide use for treatment of GDM increased therapy with insulin as the next step [65].
from 7.4% to 64.5% between 2000 and 2011 in the United
States, replacing insulin as the most common pharmaco Gestational weight gain
therapy [49]. More recently, concern has been raised about Despite what appears to be excellent glycemic control,
safety of glyburide in pregnancy [60]. A meta‐analysis women with diabetes can still have a fetus that is mac
including 2509 randomized‐controlled trial participants rosomic [66]. This may be due to the presence of hypergly
reported that glyburide resulted in higher birth weight cemia overlooked on standard glucose testing. However, in
and more macrosomia and neonatal hypoglycemia com addition to hyperglycemia, prepregnancy BMI and exces
pared with insulin therapy [63]. Interestingly, the same sive gestational weight gain play key roles in fetal growth
study reported that metformin was associated with less (Figure 6.1). This is demonstrated in study findings where
maternal weight gain, macrosomia and lower birth‐ overweight women with well controlled GDM have a 50%
weight but increased risk of preterm birth compared to greater risk of having a macrosomic baby than women of
Excessive
Maternal
gestational weight
hyperinsulinemia
gain
Maternal
Maternal obesity
hyperglycemia
Excessive
fetal
growth
FIGURE 6.1 Key factors contributing to

excessive fetal growth.
TABLE 6.2 Guidelines for gestational weight gain based be to manage expectantly until term with induction taking
on prepregnancy BMI. placing between 39 + 0 and 40 + 6 weeks’ gestation. For
women requiring pharmacological intervention, delivery
Recommended total
BMI (kg/m2) weight gain (kg) between 39 + 0 and 39 + 6 weeks is generally accepted.
Earlier delivery should be considered on a case‐by‐case basis
< 18.5 12.5 – 18.0 for women with additional complications. For those with an
18.5–24.9 11.5 – 16.0
estimated fetal weight of ≥ 4500 g on ultrasound, the option
25.0–29.9 7.0 – 11.5
≥ 30.0 5.0 – 9.0 of elective cesarean delivery should be discussed [26]. Third
trimester anesthesiology consultation is helpful for many
BMI: body mass index women to discuss the birth plan and discuss options for pain
management during labor and delivery. It should be particu
normal weight with well‐controlled GDM [67]. Increasing larly considered in women with additional complications
evidence points towards factors other than glucose as play including morbid obesity. Women with diabetes should give
ing a key role in modulating fetal overgrowth – including birth in a location where advanced neonatal resuscitation
maternal lipids [68]. While it is not considered appropriate skills are available continuously [65].
to advise weight loss once pregnancy is established, a
restriction on gestational weight gain may help mitigate
Glucocorticoid therapy
risk of fetal overgrowth, particularly in women who are
Glucocorticoids are commonly used to accelerate fetal lung
overweight or obese at baseline. In women with GDM, ges
maturation when early delivery is anticipated, but they
tational weight gain that remains within Institute of
have the propensity to significantly exacerbate hyperglyce
Medicine Guidelines [69] (Table 6.2) is associated with
mia in women with GDM. The hyperglycemia usually
lower risk of having a large‐for‐gestational age infant and
starts about 6 hours after initial glucocorticoid dosing but
gestational hypertension [33].
remains for at least 24 hours after the last dose. Additional
glucose monitoring is advised following steroid adminis
Fetal monitoring
tration and supplemental insulin is typically administered
Antenatal fetal testing in women with poorly controlled or
intravenously with infusion rates adjusted hourly, targeting
medication‐requiring GDM without other complications is
a glucose of 70–126 mg/dL (3.9–7.0 mmol/L) [72]. It is
typically initiated at 28–32 weeks’ gestation [26, 65].
important to monitor infants carefully for neonatal hypo
Guidance does not specify the approach to monitoring
glycemia, particularly if the glucocorticoids were adminis
women with GDM who are well controlled with lifestyle
tered close to delivery.
interventions. However, antenatal fetal testing is generally
started at some point from 28 weeks’ gestation [26, 65]. The
optimal delivery time for women with GDM is not clear. A Intrapartum management
recent trial randomized 425 women with GDM to induc Intrapartum hyperglycemia has been linked with neonatal
tion at 38 weeks or expectant management up to 41 weeks’ hypoglycemia and most local policies specify a glycemic
gestation. Although the study was underpowered, no dif goal range that falls between 70–126 mg/dL (3.9–
ference was detected in birth outcomes between 7.0 mmol/L) [73]. IV insulin is frequently used in women
groups [70]. An older systematic review (including one who required insulin during their pregnancy and insulin
randomized trial and four observational studies), found should also be initiated in insulin‐naïve women who are
that active rather than expectant management of labor at outside of this glycemic range. A randomized controlled
term for women with GDM may reduce rates of large‐for‐ trial published in 2019 randomized 76 women with GDM
gestational age infants. However, heterogeneity of included to tight and liberalized glucose goal ranges during labor.
studies was a significant issue in the review [71]. A reason “Tight control” constituted glucose measurements hourly
able approach in the setting of uncomplicated GDM would and treatment for maternal glucose levels lower than
60 mg/dL (3.3 mmol/L) or > 100 mg/dL (5.6 mmol/L). 12 weeks postpartum. This should include a 75 g OGTT
“Liberalized control” involved glucose measurements with standard criteria used to diagnose diabetes or pre‐dia
every 4 hours and treatment for maternal glucose levels < betes [26]. Women with persistent diabetes should receive
60 mg/dL (3.3 mmol/L) or > 120 mg/dL (6.7 mmol/L). The appropriate treatment and metformin should be consid
primary outcome of neonatal blood glucose levels was sim ered in women with prediabetes as it may delay progres
ilar between both groups [53 mg/dL (2.9 mmol/L) versus sion to type 2 diabetes in this situation [78]. Attendance at
58 mg/dL (3.2 mmol/L), mean difference −4.18, 95% CI postpartum glucose testing is often poor. However, pro
−12.66 to 4.29]. However, mean neonatal glucose level grams with a dedicated coordinator who makes verbal and
within the first 24 hours of life was lower in the tight con written contact can achieve up to a 75% recall rate [79].
trol group [54 mg/dL (3.0 mmol/L) vs 58 mg/dL At a median follow‐up of 11.4 years, 52.2% women with
(3.2 mmol/L), mean difference −3.39, 95% CI −7.07 to GDM developed a disorder of glucose metabolism versus
0.29] [74]. 20.1% without GDM (OR 3.44, 95% CI 2.85 – 4.14).
Women with GDM should therefore have lifelong, regular
Postpartum neonatal care assessment of glucose status at 1‐ to 3‐year intervals [22].
Ideally, infants will remain with the mother on delivery and Inter‐pregnancy weight gain is a risk for adverse obstetric
feeding should be encouraged as soon as possible after outcomes including GDM, pre‐eclampsia and still
birth. Capillary blood glucose testing should take place 2–4 birth [80]. The postpartum visit is an ideal opportunity to
hours after birth and if capillary plasma glucose values are discuss future pregnancy planning, review contraceptive
< 36 mg/dL (2 mmol/L) on two consecutive occasions options, and provide assistance with weight loss if
despite feeding, or if there are abnormal clinical signs, indicated.
additional supports such as intravenous glucose or tube
feeding will be required [65]. Future directions
A Canadian‐based retrospective cohort study identified As the prevalence of GDM continues to rise in line with the
that GDM was associated with lower rates of breastfeeding worldwide epidemic of type 2 diabetes, a large number of
in hospital and on discharge [75]. With this in mind, high‐quality trials have attempted to develop effective
women with GDM should receive extra lactation support interventions for GDM prevention [6]. A 2017 Cochrane
postpartum, as breastfeeding is associated with nutritional meta‐analysis of trials evaluating the combination of diet
and immunological benefits for the baby and a reduced and exercise interventions for GDM prevention concluded
risk of postpartum dysglycemia (8.2% vs 18.4%, p< 0.001 in that there was a possible reduced risk of GDM in the inter
a prospective cohort study) [76]. Benefits of breastfeeding vention groups (RR 0.85, 95% CI 0.71–1.01; 6633 women;
were also demonstrated in the 30‐year CARDIA study 19 RCTs) but study heterogeneity limited ability to inform
where lactation duration showed an inverse association practice [81]. One observation is that women at risk of
with diabetes incidence – adjusted HR for > 0 to 6 months: GDM are often reluctant to engage in lifestyle changes and
0.75 (95% CI, 0.51–1.09); > 6 months to < 12 months: 0.52 find it difficult to adhere to recommended interventions
(95% CI, 0.31–0.87); and ≥ 12 months 0.53 (0.29–0.98) vs even within a trial setting [6]. Metformin has been exam
no breastfeeding [77]. ined for GDM prevention in two randomized controlled
trials and it was ineffective in achieving this outcome [82,
Postpartum maternal care 83]. Indeed, GDM is a heterogenous disorder, in terms of
Postpartum, insulin resistance decreases dramatically and its time of onset and the underlying pathophysiology that
any glucose lowering therapies should be discontinued. It is may involve predominantly insulin resistance, impaired
recommended that women with GDM should have blood insulin secretion or both. This makes it difficult to envisage
glucose testing prior to discharge from hospital to exclude a “one size fits all approach”, and successful prevention and
persistent hyperglycemia [65]. Formal screening for a per treatment strategies will likely require personalized pro
sistent glucose disorder typically takes place between 4 and grams depending on individual phenotype.
The optimal strategy for diagnosis of GDM continues to normal fetal growth and reduce maternal complications.
be debated. Each approach will identify different degrees of Postpartum efforts should focus on screening for persistent
hyperglycemia and risk for adverse outcomes [60]. diabetes and reducing risk of future type 2 diabetes. It is
Identifying GDM in early pregnancy is particularly chal hoped that ongoing work will help identify the best diag
lenging as early pregnancy hyperglycemia may have nostic and treatment strategies for GDM.
resolved by late second trimester [84], and as already men
tioned, the clinical benefit of treating women from early
pregnancy is not clear. Numerous studies have evaluated References
alternative biomarkers for detection of GDM in maternal 1. Egan AM, Vellinga A, Harreiter J et al. Epidemiology of
plasma and urine [85]. One example is plasma glycated gestational diabetes mellitus according to IADPSG/WHO
CD59, which has high sensitivity and specificity for identi 2013 criteria among obese pregnant women in Europe.
fying the risk for LGA when measured at 24–28 weeks’ ges Diabetologia. 2017;60(10):1913–1921.
tation [86]. Another diagnostic option under investigation 2. International Diabetes Federation. IDF Atlas. Brussels,
involves the use of machine‐learning to predict GDM. This Belgium 2013.
has the potential to provide a cost‐effective screening 3. Avalos GE, Owens LA, Dunne F, Collaborators AD.
Applying current screening tools for gestational diabetes
approach that avoids the need for a cumbersome
mellitus to a European population: is it time for change?
OGTT [87].
Diabetes Care. 2013;36(10):3040–3044.
While it is clear that much work is needed to develop 4. Catalano P, Huston L, Amini S, Kalhan S. Longitudinal
novel approaches for GDM treatment and evaluate long‐ changes in glucose metabolism during pregnancy in obese
term effects of current treatment strategies, future efforts women with normal glucose tolerance and gestational
should also address the risk of type 2 diabetes. Our under diabetes. Am J Obstst Gyneol. 1999;180(4):903–916.
standing of the factors driving β‐cell failure or recovery 5. Pedersen J. Diabetes and pregnancy: blood sugar of newborn
after GDM is limited [88], and further mechanistic studies infants (PhD Thesis). Copenhagen, Danish Science Press; 1952.
are needed to evaluate islet cell function postpartum. For 6. Egan AM, Simmons D. Lessons learned from lifestyle
example, prior work indicates that compared to women prevention trials in gestational diabetes mellitus. Diabet
without GDM, women with GDM have elevated fasting Med. 2019;36(2):142–150.
7. Yogev, Chen, Hod et al. Hyperglycemia and Adverse
proinsulin levels during pregnancy and 8–14 weeks post
Pregnancy Outcome (HAPO) study: preeclampsia. Am J
partum [89, 90]. An increase in the fasting proinsulin to
Obstet Gynecol. 2010;202(3):255.e251–257.
insulin ratio has been used as a marker of impaired β‐cell 8. Jastrow N, Roberge S, Gauthier RJ et al. Effect of birth weight
function and is associated with rapid conversion to type 2 on adverse obstetric outcomes in vaginal birth after cesarean
diabetes in people with prediabetes [91]. With this in mind, delivery. Obstet Gynecol. 2010;115(2 Pt 1):338–343.
measurement of how prandial proinsulin and insulin 9. O’Sullivan EP, Avalos G, O’Reilly M et al. Atlantic Diabetes
secretion change relative to one another, and changes in the in Pregnancy (DIP): the prevalence and outcomes of
proinsulin to insulin ratio over time may provide insights gestational diabetes mellitus using new diagnostic criteria.
into β‐cell function postpartum. Ultimately, identifying Diabetologia. 2011;54(7):1670–1675.
accurate markers of postpartum β‐cell health will generate 10. Metzger BE, Lowe LP, Dyer AR et al. Hyperglycemia and
data to assist in developing effective type 2 diabetes preven adverse pregnancy outcomes. N Engl J Med. 2008;358(19):
1991–2002.
tion programs for women with GDM.
11. Sweeting AN, Ross GP, Hyett J et al. Gestational diabetes
mellitus in early pregnancy: evidence for poor pregnancy
Conclusion outcomes despite treatment. Diabetes Care. 2016;39(1):
75–81.
The prevalence of GDM is high and represents a significant 12. Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 dia
public health concern. Women with GDM benefit from a betes mellitus after gestational diabetes: a systematic review
multidisciplinary approach to care in order to support and meta‐analysis. Lancet. 2009;373(9677):1773–1779.
13. Guerrero‐Romero F, Aradillas‐García C, Simental‐Mendia 25. Simmons D, Hague WM, Teede HJ et al. Hyperglycaemia in
LE, Monreal‐Escalante E, de la Cruz Mendoza E, Rodríguez‐ early pregnancy: the Treatment of Booking Gestational dia
Moran M. Birth weight, family history of diabetes, and met betes Mellitus (TOBOGM) study. A randomised controlled
abolic syndrome in children and adolescents. J Pediatr. trial. Med J Aust. 2018;209(9):405–406.
2010;156(5):719–723, 723.e711. 26. ACOG Practice Bulletin No. 190 Summary: gestational dia
14. Whincup PH, Kaye SJ, Owen CG et al. Birth weight and risk betes mellitus. Obstet Gynecol. 2018;131(2):406–408.
of type 2 diabetes: a systematic review. JAMA. 27. McIntyre HD, Sacks DA, Barbour LA et al. Issues with the
2008;300(24):2886–2897. diagnosis and classification of hyperglycemia in early preg
15. Benhalima K, Mathieu C, Van Assche A et al. Survey by the nancy. Diabetes Care. 2016;39(1):53–54.
European Board and College of Obstetrics and Gynaecology 28. Metzger BE, Buchanan TA, Coustan DR et al. Summary and
on screening for gestational diabetes in Europe. Eur J Obstet recommendations of the Fifth International Workshop‐
Gynecol Reprod Biol. 2016;201:197–202. Conference on Gestational Diabetes Mellitus. Diabetes Care.
16. Bilous R. Diagnosis of gestational diabetes, defining the net, 2007;30 Suppl 2:S251–260.
refining the catch. Diabetologia. 2015;58(9):1965–1968. 29. Blumer I, Hadar E, Hadden DR et al. Diabetes and preg
17. Metzger BE, Gabbe SG, Persson B et al. International asso nancy: an endocrine society clinical practice guideline.
ciation of diabetes and pregnancy study groups recommen J Clin Endocrinol Metab. 2013;98(11):4227–4249.
dations on the diagnosis and classification of hyperglycemia 30. Castorino K, Jovanovič L. Pregnancy and diabetes manage
in pregnancy. Diabetes Care. 2010;33(3):676–682. ment: advances and controversies. Clin Chem. 2011;57(2):
18. World Health Organization. Diagnostic criteria and classifica 221–230.
tion of hyperglycaemia first detected in pregnancy: a World 31. García‐Patterson A, Martín E, Ubeda J, María MA, de
Health Organization guideline. 2013. Available at: https:// Leiva A, Corcoy R. Evaluation of light exercise in the treat
apps.who.int/iris/bitstream/handle/10665/85975/WHO_ ment of gestational diabetes. Diabetes Care. 2001;24(11):
NMH_MND_13.2_eng.pdf. Accessed 27 March, 2020. 2006–2007.
19. Benhalima K, Mathieu C, Damm P et al. A proposal for the 32. Gynecologists TACoOa. Committee Opinion Number 650:
use of uniform diagnostic criteria for gestational diabetes in Physical Activity and Exercise During Pregnancy and the
Europe: an opinion paper by the European Board & College Postpartum Period. 2015.
of Obstetrics and Gynaecology (EBCOG). Diabetologia. 33. Egan AM, Dennedy MC, Al‐Ramli W, Heerey A, Avalos G,
2015;58(7):1422–1429. Dunne F. ATLANTIC‐DIP: excessive gestational weight gain
20. Hod M, Kapur A, Sacks DA et al. The International and pregnancy outcomes in women with gestational or
Federation of Gynecology and Obstetrics (FIGO) Initiative pregestational diabetes mellitus. J Clin Endocrinol Metab.
on gestational diabetes mellitus: a pragmatic guide for diag 2014;99(1):212–219.
nosis, management, and care. Int J Gynaecol Obstet. 2015;131 34. Kgosidialwa O, Egan AM, Carmody L, Kirwan B, Gunning P,
Suppl 3:S173–211. Dunne FP. Treatment With Diet and Exercise for Women
21. Hod M, Pretty M, Mahmood T, FIGO, EAPMand EBCOG. With Gestational Diabetes Mellitus Diagnosed Using
Joint position statement on universal screening for GDM in IADPSG Criteria. J Clin Endocrinol Metab. 2015;100(12):
Europe by FIGO, EBCOG and EAPM. Eur J Obstet Gynecol 4629–4636.
Reprod Biol. 2018;228:329–330. 35. Brown J, Alwan NA, West J et al. Lifestyle interventions for
22. American Diabetes Association. 14. Management of diabe the treatment of women with gestational diabetes. Cochrane
tes in pregnancy: standards of medical care in diabetes – Database Syst Rev. 2017;5:CD011970.
2020. Diabetes Care. 2020;43(Suppl 1):S183–S192. 36. Egan AM, Dunne FP. Optimal management of gestational
23. Ferrara A, Hedderson M, Quesenberry C, Selby J. Prevalence diabetes. Br Med Bull. 2019;131(1):97–108.
of gestational diabetes mellitus detected by the National 37. Suffecool K, Rosenn B, Niederkofler EE et al. Insulin detemir
Diabetes Data Group or the Carpenter and Coustan plasma does not cross the human placenta. Diabetes Care. 2015;38(2):
glucose thresholds. Diabetes Care. 2020;25(9):1625–1630. e20–21.
24. Sovio U, Murphy HR, Smith GC. Accelerated fetal growth 38. McCance DR, Damm P, Mathiesen ER et al. Evaluation of
prior to diagnosis of gestational diabetes mellitus: a prospec insulin antibodies and placental transfer of insulin aspart in
tive cohort study of nulliparous women. Diabetes Care. pregnant women with type 1 diabetes mellitus. Diabetologia.
2016;39(6):982–987. 2008;51(11):2141–2143.
39. Boskovic R, Feig DS, Derewlany L, Knie B, Portnoi G, Koren G. 50. Foretz M, Guigas B, Bertrand L, Pollak M, Viollet B.
Transfer of insulin lispro across the human placenta: in vitro; Metformin: from mechanisms of action to therapies. Cell
perfusion studies. Diabetes Care. 2003;26(5):1390–1394. Metab. 2014;20(6):953–966.
40. Callesen NF, Damm J, Mathiesen JM, Ringholm L, Damm P, 51. Coetzee EJ, Jackson WP. Metformin in management of preg
Mathiesen ER. Treatment with the long‐acting insulin ana nant insulin‐independent diabetics. Diabetologia. 1979;
logues detemir or glargine during pregnancy in women with 16(4):241–245.
type 1 diabetes: comparison of glycaemic control and preg 52. Feig DS, Moses RG. Metformin therapy during pregnancy:
nancy outcome. J Matern Fetal Neonatal Med. 2013;26(6): good for the goose and good for the gosling too? Diabetes
588–592. Care. 2011;34(10):2329–2330.
41. Mathiesen ER, Hod M, Ivanisevic M et al. Maternal efficacy 53. Vanky E, Zahlsen K, Spigset O, Carlsen SM. Placental pas
and safety outcomes in a randomized, controlled trial com sage of metformin in women with polycystic ovary syn
paring insulin detemir with NPH insulin in 310 pregnant drome. Fertil Steril. 2005;83(5):1575–1578.
women with type 1 diabetes. Diabetes Care. 2012;35(10): 54. Given JE, Loane M, Garne E et al. Metformin exposure in
2012–2017. first trimester of pregnancy and risk of all or specific con
42. Mathiesen ER, Kinsley B, Amiel SA et al. Maternal glycemic genital anomalies: exploratory case‐control study. BMJ.
control and hypoglycemia in type 1 diabetic pregnancy: a 2018;361:k2477.
randomized trial of insulin aspart versus human insulin in 55. Rowan JA, Hague WM, Gao W, Battin MR, Moore MP,
322 pregnant women. Diabetes Care. 2007;30(4):771–776. Investigators MT. Metformin versus insulin for the treat
43. Garg SK, Frias JP, Anil S, Gottlieb PA, MacKenzie T, Jackson ment of gestational diabetes. N Engl J Med. 2008;358(19):
WE. Insulin lispro therapy in pregnancies complicated by 2003–2015.
type 1 diabetes: glycemic control and maternal and fetal out 56. Rowan JA, Rush EC, Plank LD et al. Metformin in gesta
comes. Endocr Pract. 2003;9(3):187–193. tional diabetes: the offspring follow‐up (MiG TOFU): body
44. Insulin Fiasp vs. Insulin Novorapid during pregnancy and lac composition and metabolic outcomes at 7–9 years of age.
tation in women with pre‐existing diabetes. ClinicalTrials.gov. BMJ Open Diabetes Res Care. 2018;6(1):e000456.
Available at: https://clinicaltrials.gov/ct2/show/NCT03770767. 57. Tarry‐Adkins JL, Aiken CE, Ozanne SE. Neonatal, infant,
Accessed 27 March, 2020. and childhood growth following metformin versus insulin
45. Research Study Comparing Insulin Degludec to Insulin treatment for gestational diabetes: a systematic review and
Detemir, Together With Insulin Aspart, in Pregnant Women meta‐analysis. PLoS Med. 2019;16(8):e1002848.
With Type 1 Diabetes (EXPECT). clinicaltrials.gov. Available 58. Dunne F. A Clinical Trial of the Effectiveness of Metformin
at: https://clinicaltrials.gov/ct2/show/NCT03377699. Accessed in Addition to Usual Care in the Reduction of Gestational
27 March, 2020. Diabetes Mellitus (EMERGE). clinicaltrials.gov Available at:
46. Murphy HR, Elleri D, Allen JM et al. Pathophysiology of https://clinicaltrials.gov/ct2/show/NCT02980276. Accessed
postprandial hyperglycaemia in women with type 1 diabetes 27 March, 2020.
during pregnancy. Diabetologia. 2012;55(2):282–293. 59. Ashcroft FM. Mechanisms of the glycaemic effects of sulfo
47. Goudie RJ, Lunn D, Hovorka R, Murphy HR. Pharma nylureas. Horm Metab Res. 1996;28(9):456–463.
cokinetics of insulin aspart in pregnant women with type 1 60. Wexler DJ, Powe CE, Barbour LA et al. Research gaps in ges
diabetes: every day is different. Diabetes Care. 2014;37(6): tational diabetes mellitus: executive summary of a national
e121–122. institute of diabetes and digestive and kidney diseases work
48. Bogdanet D, Egan AM, Reddin C et al. ATLANTIC DIP. shop. Obstet Gynecol. 2018;132(2):496–505.
Insulin Therapy for Women With IADPSG‐Diagnosed 61. Langer O, Conway DL, Berkus MD, Xenakis EM, Gonzales
Gestational Diabetes Mellitus. Does It Work? J Clin Endocrinol O. A comparison of glyburide and insulin in women with
Metab. 2017;102(3):849–857. gestational diabetes mellitus. N Engl J Med. 2000;343(16):
49. Camelo Castillo W, Boggess K, Stürmer T, Brookhart MA, 1134–1138.
Benjamin DK, Jonsson Funk M. Trends in glyburide com 62. Hebert MF, Ma X, Naraharisetti SB et al. Are we optimizing
pared with insulin use for gestational diabetes treatment in gestational diabetes treatment with glyburide? The pharma
the United States, 2000–2011. Obstet Gynecol. 2014;123(6): cologic basis for better clinical practice. Clin Pharmacol
1177–1184. Ther. 2009;85(6):607–614.
63. Balsells M, García‐Patterson A, Solà I, Roqué M, Gich I, 76. O’Reilly MW, Avalos G, Dennedy MC, O’Sullivan EP, Dunne
Corcoy R. Glibenclamide, metformin, and insulin for the F. Atlantic DIP. high prevalence of abnormal glucose toler
treatment of gestational diabetes: a systematic review and ance post partum is reduced by breast‐feeding in women
meta‐analysis. BMJ. 2015;350:h102. with prior gestational diabetes mellitus. Eur J Endocrinol.
64. Kalafat E, Sukur YE, Abdi A, Thilaganathan B, Khalil A. 2011;165(6):953–959.
Metformin for prevention of hypertensive disorders of preg 77. Gunderson EP, Lewis CE, Lin Y et al. Lactation duration and
nancy in women with gestational diabetes or obesity: sys progression to diabetes in women across the childbearing
tematic review and meta‐analysis of randomized trials. years: the 30‐year CARDIA study. JAMA Intern Med.
Ultrasound Obstet Gynecol. 2018;52(6):706–714. 2018;178(3):328–337.
65. Excellence NIfHaC. Diabetes in pregnancy: management 78. Ratner RE, Christophi CA, Metzger BE et al. Prevention of
from preconception to the postnatal period. In:2015. diabetes in women with a history of gestational diabetes:
66. Evers IM, de Valk HW, Mol BW, ter Braak EW, Visser GH. effects of metformin and lifestyle interventions. J Clin
Macrosomia despite good glycaemic control in Type I dia Endocrinol Metab. 2008;93(12):4774–4779.
betic pregnancy; results of a nationwide study in The Nether 79. Carmody L, Egan AM, Dunne FP. Postpartum glucose test
lands. Diabetologia. 2002;45(11):1484–1489. ing for women with gestational diabetes mellitus: improving
67. Langer O, Yogev Y, Xenakis EM, Brustman L. Overweight regional recall rates. Diabetes Res Clin Pract. 2015.
and obese in gestational diabetes: the impact on pregnancy 80. Villamor E, Cnattingius S. Interpregnancy weight change
outcome. Am J Obstet Gynecol. 2005;192(6):1768–1776. and risk of adverse pregnancy outcomes: a population‐based
68. Catalano PM, Hauguel‐De Mouzon S. Is it time to revisit the study. Lancet. 2006;368(9542):1164–1170.
Pedersen hypothesis in the face of the obesity epidemic? Am 81. Shepherd E, Gomersall JC, Tieu J, Han S, Crowther CA,
J Obstet Gynecol. 2011;204(6):479–487. Middleton P. Combined diet and exercise interventions for
69. Institute of Medicine (US) and National Research Council preventing gestational diabetes mellitus. Cochrane Database
(US) Committee to Reexamine IOM Pregnancy Weight Syst Rev. 2017;11:CD010443.
Guidelines. In: Rasmussen KM, Yaktine AL, ed. Weight Gain 82. Chiswick C, Reynolds RM, Denison F et al. Effect of met
During Pregnancy: Reexamining the Guidelines. Washington, formin on maternal and fetal outcomes in obese pregnant
DC. National Acadamies Press (US); 2009. women (EMPOWaR): a randomised, double‐blind, placebo‐
70. Alberico S, Erenbourg A, Hod M et al. Immediate delivery controlled trial. Lancet Diabetes Endocrinol. 2015;3(10):
or expectant management in gestational diabetes at term: the 778–786.
GINEXMAL randomised controlled trial. BJOG. 2017;124(4): 83. Syngelaki A, Nicolaides KH, Balani J et al. Metformin versus
669–677. placebo in obese pregnant women without diabetes mellitus.
71. Witkop CT, Neale D, Wilson LM, Bass EB, Nicholson WK. N Engl J Med. 2016;374(5):434–443.
Active compared with expectant delivery management in 84. Zhu WW, Yang HX, Wei YM et al. Evaluation of the value of
women with gestational diabetes: a systematic review. Obstet fasting plasma glucose in the first prenatal visit to diagnose
Gynecol. 2009;113(1):206–217. gestational diabetes mellitus in china. Diabetes Care.
72. Egan AM, Murphy HR, Dunne FP. The management of 2013;36(3):586–590.
type 1 and type 2 diabetes in pregnancy. QJM. 2015;108(12): 85. Lorenzo‐Almorós A, Hang T, Peiró C et al. Predictive and
923–927. diagnostic biomarkers for gestational diabetes and its associ
73. Ryan EA, Al‐Agha R. Glucose control during labor and ated metabolic and cardiovascular diseases. Cardiovasc
delivery. Curr Diab Rep. 2014;14(1):450. Diabetol. 2019;18(1):140.
74. Hamel MS, Kanno LM, Has P, Beninati MJ, Rouse DJ, 86. Ghosh P, Luque‐Fernandez MA, Vaidya A et al. Plasma gly
Werner EF. Intrapartum glucose management in women cated CD59, a novel biomarker for detection of pregnancy‐
with gestational diabetes mellitus: a randomized controlled induced glucose intolerance. Diabetes Care. 2017;40(7):
trial. Obstet Gynecol. 2019. 981–984.
75. Finkelstein SA, Keely E, Feig DS, Tu X, Yasseen AS, Walker 87. Artzi NS, Shilo S, Hadar E et al. Prediction of gestational
M. Breastfeeding in women with diabetes: lower rates diabetes based on nationwide electronic health records. Nat
despite greater rewards. A population‐based study. Diabet Med. 2020;26(1):71–76.
Med. 2013;30(9):1094–1101.
88. Xiang AH, Peters RK, Kjos SL et al. Effect of pioglitazone on 90. Kautzky‐Willer A, Thomaseth K, Ludvik B et al. Elevated
pancreatic beta‐cell function and diabetes risk in Hispanic islet amyloid pancreatic polypeptide and proinsulin in lean
women with prior gestational diabetes. Diabetes. 2006;55(2): gestational diabetes. Diabetes. 1997;46(4):607–614.
517–522. 91. Haffner SM, Gonzalez C, Mykkänen L, Stern M. Total
89. Genova M, Todorova‐Ananieva K, Atanasova Bea. immunoreactive proinsulin, immunoreactive insulin and
Proinsulin in healthy pregnancy, pregnancy with gestational specific insulin in relation to conversion to NIDDM. the
diabetes and after delivery. Acta Medica Bulgarica. 2014(41): Mexico City Diabetes Study. Diabetologia. 1997;40(7):
13–21. 830–837.
7 What Is the Role of Self‐Monitoring
in Diabetes? Is There a Role
for Postprandial Glucose Monitoring?
How Does Continuous Glucose
Monitoring Integrate into Clinical
Practice?
Rami Almokayyad1 and Robert Cuddihy2
1
ndocrine Fellow Division of Endocrinology, Department of Medicine, University of Minnesota Medical School,
E
University of Minnesota, Minneapolis, MN, USA
2
edical Director, International Diabetes Center, World Health Organization Collaborating Center for Diabetes
M
Education, Translation and Computer Technology, Minneapolis, MN, USA
LE A R NI NG P OI NT S Self‐monitoring of blood glucose

Self‐monitoring of blood glucose (SMBG) can be consid-
●● There is general consensus that the use of SMBG in ered one of the bigger advances in diabetes management
most individuals with T1DM or those with T2DM
after the discovery of insulin and oral agents. Bedside
treated with insulin is of benefit.
●● The benefit of SMBG in individuals with T2DM, not
measurement of plasma glucose was introduced in the
on insulin therapy, remains inconclusive and mid‐1920s as a necessity, after the discovery and use of the
controversial. first insulin. At that time, the only modality to monitor dia-
●● The role of postprandial glucose monitoring has betes control was qualitative urine glucose measurements,
previously been firmly established in gestational
which only gave patients an estimate of their levels over the
diabetes but remains controversial in T1DM or
T2DM.
proceeding 1–2 days, revealing hyperglycemia only when
●● Continuous Glucose Monitoring (CGM) devices have serum glucose had exceeded the renal threshold (∼200 mg/
documented benefit in adults with T1DM on an insulin dl) resulting in glycosuria. The procedure of measuring
pump therapy who use the data it provides to modify glucose would typically take 20 minutes, as described by
their therapy.
Maclean in his book Modern Methods in the Diagnosis and
●● There is a clear need for well‐controlled clinical trials to
assess the value of SMBG in T2DM for those on other
Treatment of Glycosuria and Diabetes, published in
than insulin therapy. 1924 [1]. The use of this method was limited to use by the
●● The lack of a simple universal output of SMBG and clinician.
CGM data are an impediment to their appropriate By the early 1940s, semiquantitative urine glucose
uptake and use in primary care.
measurement kits became commercially available for home

87
use, and they became the major modality for glucose self‐ tality risk with assignment to a strategy aiming to achieve
monitoring at home. near normoglycemia. However, somewhat surprisingly,
In the mid‐1960s, reagent strips were introduced as a this risk was inversely proportional to the rate of fall in
semiquantitative method of measuring capillary whole HbA1c, being lowest in those individuals who rapidly cor-
blood glucose; a drop of blood was added to these reagents rected their HbA1c and highest in those who lowered their
that utilized a glucose oxidase reaction to effect a colori- HbA1c slowly, or not at all, and the highest mortality in
metric change. The resultant color from the reaction could both the standard and intensive glycemic control arms of
then be compared with a standard chart and this would ACCORD occurred in those patients with higher HbA1c
give a rough estimate of the blood glucose level. These (publicly presented ACCORD data in [9]). The available
strips were initially limited to office and hospital use, with evidence would suggest that in most patients with diabetes
limited patient access for home use. one should maintain “tight” glycemic control as early in the
By the late 1960s reflectance meters were introduced course of the disease as possible and to continue to main-
that utilized the same type of reagent strips where a light tain this level of control as tightly as possible [10], generally
source was reflected on the strip (instead of reading them targeting an A1c level < 7.0% by ADA guidelines (or < 6.5%
manually) enabling the reflected light to be read by a pho- by IDF or AACE guidelines), if this can be accomplished
toelectric cell, which in turn gave a readout of the estimated safely without exposing the patient to undue risk of severe
glucose (using a swinging needle at that time). In the late hypoglycemia.
1970s meters were developed for patient use. SMBG therefore can play a pivotal role, when used
With advances in technology, electrochemical meters properly, as an adjunctive tool to enhance patient self‐man-
were introduced, and as the technology evolved, improve- agement of their diabetes. SMBG provides day‐to‐day data
ments were made to reduce the size of these meters, reduce on glycemic control; it provides an immediate feedback
their reaction and reading times, simplify blood sampling about the effect of nutrition, physical activity, and medica-
techniques, and to reduce the discomfort associated with tions on blood glucose. It allows prompt determination
blood sampling [2]. Today, glucose meters have a central of hypoglycemia or hyperglycemia that not only can
role in diabetes management, with nearly 70% of all improve patient safety, but also can motivate individuals
patients with diabetes performing some SMBG [3]. to make appropriate changes in diet, exercise, and
Long‐term follow‐up of participants in the Diabetes medications.
Complication and Control Trial (DCCT) as part of the By shaping meal and activity patterns and providing
Epidemiology of Diabetes Interventions and Complications feedback on medication dosing or titration while simulta-
(EDIC) in type 1 diabetes mellitus (T1DM), and the 10‐ neously providing useful information on the potential
year follow‐up of the United Kingdom Prospective occurrence or risks of hypoglycemia, the information
Diabetes Study (UKPDS) in type 2 diabetes mellitus gleaned from SMBG monitoring can help optimize the
(T2DM) provide clear evidence that the early and intensive delicate balance between the benefits and risks of tight gly-
control of hyperglycemia reduces the long‐term risks of cemic control.
microvascular and macrovascular complications in diabe- In this section, we will discuss the utility of SMBG in
tes. The effect of good glycemic control seems to persist for diabetes management, and we will discuss the evidence
years, to some extent independently of subsequent glyce- behind its use in different patient groups.
mic control. This phenomenon is referred to as “metabolic
memory” or “the legacy effect” [4–6]. However, tight glyce-
SMGB in type 1 DM and insulin‐requiring
mic control does come at a cost, which is a threefold
T2DM patients
increased risk of severe hypoglycemia seen in the DCCT
trial (T1DM), as well as the Action to Control The use of SMBG in patients with type 1 diabetes and in
Cardiovascular Risks in Diabetes (ACCORD) trial in those with type 2 diabetes treated with intensive insulin
T2DM [7, 8]. ACCORD also suggested an increased mor- therapies or multiple daily insulin injections (MDI) appears
What Is the Role of Self-Monitoring in Diabetes? 89
logical for several reasons, is supported by clinical trial nity and the strategies for use of the results from SMBG
data, and thus is not controversial. were not clearly defined. The relative benefit in terms of
First, SMGB will provide day‐to‐day information that HbA1c reduction typically was modest, even in those ran-
can help adjust insulin dose to optimize the overall glyce- domized clinical trials which did show benefit, in the range
mic control. Patients on bolus (“prandial or meal‐related”) from one quarter to two thirds of 1%, and most meta‐anal-
insulin use the premeal SMBG data to adjust the dose of ysis of various combinations of these RCTs favor SMBG.
their prandial insulin, correcting it for the amount of car- Epidemiologic studies utilizing large patient databases
bohydrate consumed, as well for any variance of the pre- have generally noted improved glycemic control in indi-
meal glucose above or below the desired target range viduals with diabetes who monitor more often, but such
(so‐called corrective dose insulin). While various insulin studies can only implicate an association between more
adjustment methods may differ in their specific titration frequent SMBG and improved glycemic control, but not a
schemes, they all utilize premeal SMBG values as an causal relationship. Selection biases and other confounding
actionable item aiding in appropriately adjusting the insu- variables may also affect these results [16]. Other cross‐
lin dosing before a meal so as to achieve tighter glycemic sectional studies, such as the Fremantle Diabetes Study, do
control. not show benefit in terms of glycemic control with
One of the earliest studies that demonstrated the effec- SMBG [17].
tiveness of SMBG in improving glycemic control was con- The Cochrane Collaborative best‐evidence‐based
ducted in 1982 [11]. This concept was emphasized in the review of SMBG use in patients with T2DM on non‐insulin
Diabetes Complication and Control Trial, in which inten- therapies also appears to favor SMBG but noted that more
sive insulin therapy has shown to improve glycemic con- evidence is needed [18].
trol and reduce microvascular complications, and this was Farmer and colleagues conducted a randomized con-
achieved by intensifying insulin regimens utilizing trolled trial (Diabetes Glycaemic Education and
SMBG [7]. Monitoring [DiGEM] study) that aimed to test whether
Also, SMBG is an important safety tool in insulin treated SMBG, used with or without instruction in incorporating
patients to detect hypoglycemia, especially in patients with findings into self‐care, can improve glycemic control in
hypoglycemia unawareness. non‐insulin‐treated diabetes patients compared with
The recommendation from the DCCT was to perform standardized usual care [19]. A total of 453 patients were
SMBG at least four times per day; however, observational individually randomized to one of three groups: (1)
studies have shown that most patients with type 1 diabetes standardized usual care with 3‐monthly HbA1c (control);
on intensive (physiological) insulin regimen fall short of (2) blood glucose self‐testing with patient training
these recommendations [12]. focused on clinician interpretation of results in addition
to usual care (less intensive self‐monitoring); or (3)
SMBG with additional training of patients in interpreta-
SMBG in non‐insulin‐requiring T2DM
tion and application of the results to enhance motivation
Although SMBG is now widely accepted as a part of the and maintain adherence to a healthy lifestyle (more inten-
management of patients with non‐insulin‐treated type 2 sive self‐monitoring).
diabetes, its efficacy and rationale are more controversial. There was no evidence of glycemic benefit between the
Most of the existing data for the efficacy of SMBG in these three groups at the end of 12 months (no difference in the
subjects come from cross‐sectional and retrospective primary outcome; Hemoglobin A1C). In addition, there
studies, which slightly favored more frequent SMBG was no evidence of a significantly different impact of self‐
[13–15]. monitoring on glycemic control when comparing sub-
Many randomized clinical trials (RCTs) have been car- groups of patients defined by duration of diabetes,
ried out in small groups of patients. Participants were not therapy, and diabetes‐related complications. Patients who
recruited from representative populations in the commu- were in the more intensive SMBG arm detected more
hypoglycemia. The economic analysis suggested that in glycemic control in such a population of individuals and
SMBG resulted in extra health care costs and was unlikely at least raises the stakes for proponents of SMBG to prove
to be cost‐effective if used routinely. There was an initial its worth in non‐insulin‐treated populations. Several
negative impact associated with more frequent use of groups have formed to attempt to outline the necessary
SMBG on the quality of life [20]. components of a large scale RCT to better evaluate the role
The potential clinical ramifications from this study have and utility of SMBG in T2DM [16, 21].
been huge and called into question the utility, cost‐effec- From a philosophical standpoint it is of vital importance
tiveness, and effect on quality of life of SMBG individuals to understand the use of SMBG as a useful diagnostic tool
with T2DM who are not on insulin therapy. In an era of to enhance patient self‐management of diabetes rather
tightening financial resources, an epidemic of T2DM, and than as a direct therapeutic intervention targeting glucose
attempts to curb health care expenditures in general, this levels. As such, there are multiple aspects of the use of
and subsequent reports from the DiGEM study had lead to SMBG that must be in place for the data it generates to be
a wide reappraisal of benefits or SMBG in individuals who accurate, beneficial to glycemic control and, most impor-
are not yet treated with insulin. Reappraisal of the need to tantly, used by the patient. Important potential barriers to
cover SMBG testing supplies for individuals with T2DM appropriate SMBG include proper technique, correct cod-
not treated with insulin by payers and national groups such ing of glucose meters to match the testing strips, correct
as the National Health Service in the U.K. and the Centers setting of the time and date of the meter to aid in SMBG
for Medicare and Medicaid Services (CMS) in the United review of downloaded meter data and, most importantly,
States have reportedly taken place. Thus, a careful look at appropriate patient education and understanding of the
some of the potential concerns or criticism of the DiGEM timing of SMBG and the use of the data derived from it to
study is warranted. modify the patient’s self‐management. The data can then
This study enrolled individuals with relatively recent become an actionable item leading to modification in ther-
onset of diabetes (median duration of 3 years) treated with apy (whether leading to changes in diet or activity through
diet or oral agents with reasonably good glycemic control behavioral change, or adjustment in medication). Ideally,
(mean HbA1c of 7.5%), and specifically selected individu- to maximize the impact of SMBG, the patients themselves
als who were either not monitoring SMBG at all or moni- should be educated on how to appropriately use SMBG.
toring no more than a single one‐time SMBG per week. Such a tool could be used at various time points to optimize
Thus, the study may have inadvertently selected a biased management throughout the day, uncovering needed
population less geared toward, or less compliant with, behavioral modifications in diet and activity, providing
SMBG monitoring and with potentially less to gain from feedback for potential problematic periods of marked
improvements in glucose control. Of concern is the case that hyperglycemia or hypoglycemia, and providing the patient
it was in the intensive SMBG cohort that more individuals an early detection of worsening overall glycemic control
quit SMBG monitoring than the less intensive cohort. Also, due either to situational factors such as intercurrent illness
for those who were to utilize the SMBG data to modify their or steroid usage, or the progressive nature of T2DM itself.
lifestyle or medication, a delineation of a specified action Such information would indicate the need to titrate ther-
plan in response to the SMBG data is lacking and not deline- apy to reestablish target levels of glycemic control in a
ated. While there was a minimal decline in HbA1c by 0.17 % timely manner.
in this group, it was not statistically significant. Of potential There is a great need for specific algorithms instructing
concern was the issue that the reduction in HbA1c in this patients what do in terms of altering their management or
study was far less than those reported from the majority of therapy in response to their SMBG data. Some early pre-
other RCT evaluating the effect of SMBG. liminary data suggest that patient‐driven algorithms based
Criticisms aside, this study was a careful attempt to get on SMBG can be more effective in helping them reach tar-
at the issue of the value of SMBG in terms of improvements get than management that relies on the patient’s health care
provider or clinician to review the data and recommend hydrate consumed during the day or to add more physical
changes in therapy. For example, in the commonly seen activity prior to this meal. If unable to correct this issue
patient with good control of AM fasting blood glucose with dietary or activity changes, such findings could lead
(BG), but HbA1c’s above target, it is likely that BG values the patient to engage their physician to consider the addi-
are higher at other time points throughout the day. tion of another pharmacologic agent targeting postpran-
Bergenstal et al. showed that utilizing an algorithm for dial BGs. The individual could “learn” from the response of
titrating premeal insulin solely based on preprandial SMBG just how various meals and meal composition
SMBG values, rather than utilizing strict carbohydrate affect their BGs and what the affect of various activities
counting with matching insulin to carbohydrate dosing, are on their BGs, acting as a useful tool and reinforce-
was equally effective in controlling glycemic levels in insu- ment for beneficial behavioral modification. If the post-
lin‐treated individuals with T2DM [22]. prandial BG rise was more pronounced following one
The technical aspects of glucose meters that serve as specific meal, then the pattern of SMBG monitoring
barriers are the easiest to correct and, in fact, meters might temporarily change while the individual “works on
that do not require coding of the meter to match the that particular problem area,” perhaps using many of
testing strips are beginning to enter the market. Time their weekly SMBG determinations before and after that
and date stamping of SMBG values will become auto- particular meal to assess the result of various attempted
mated. Meal markers are available on some units to aid interventions, returning to a widespread surveillance pat-
in interpreting fasting or premeal glucose patterns from tern of SMBG once the “problem is solved” or corrected.
postprandial patterns. Many meters can inform the This surveillance SMBG would then inform the individ-
patient if the blood sample is inadequate for an accurate ual if and where the next issue in fine‐tuning glycemic
test result. control should occur.
The lack of patient education and training in diabetes If individuals are on agents that can cause hypoglyce-
self‐management, including proper use of SMBG, remains mia, such as a sulfonylurea (SU), then surveillance
a significant barrier that requires more effort to correct. In SMBG can indicate if there is a problematic period of
clinical practice, one frequently encounters patients treated increased risk of hypoglycemia during the day, such as
with lifestyle modification and Metformin monotherapy the late afternoon or predinner period or overnight,
who have been instructed to perform SMBG and dutifully which should ideally lead to corrective intervention to
obtain one fasting blood glucose value each morning. No reduce this risk.
alternative testing schema is offered if the AM fasting glu- Utilizing SMBG for continuous surveillance and quality
coses are within goal but the HbA1c remains above goal. improvement of their diabetes self‐management can help
Clearly, rather than obtain seven “normal fasting readings” counteract the clinical inertia currently seen in our health
per week the utility of SMBG would be markedly enhanced systems, which are not properly designed to manage non‐
if such patients were educated to use these same seven acute chronic diseases such as diabetes. Brown et al. [23]
weekly readings in a more dispersed fashion, sampling have demonstrated how this inertia can result in patients’
once daily, but at alternating times each day. Perhaps encountering 8–10 years of exposure to significant chronic
obtaining AM fasting blood glucose on Monday, 2‐hour hyperglycemia, with the increased risk for complications
postprandial BG after breakfast on Tuesday, prelunch BG that this entails while their medical regimen is very slowly
on Wednesday, 2‐hour post‐lunch BG on Thursday, progressed through the different available therapies.
predinner BG on Friday, post‐dinner BG on Saturday, and Patients should be empowered to contact their health care
bedtime BG on Sunday would provide a wealth of action- provider as soon as they encounter problematic hypergly-
able information? For instance, a significant postprandial cemia that has not responded to their attempts at correc-
rise in BG following a specific meal could signal the patient tion, as medication may need to be advanced. Rather than
to make modifications in the timing or quantity of carbo- await their next regularly scheduled 3‐ to 6‐month appoint-
ment before advancement in therapy is undertaken, the statistical summaries, is certainly an improvement.
therapies could be quickly optimized until the glycemic However, the multitude of differing proprietary software
goals or treatment targets are achieved. programs needed to download the data from each compa-
Another barrier to the optimal use of patient derived ny’s meter and the slightly different presentation format of
SMBG data in diabetes management occurs in physician each of these software programs severely inhibit the broad
offices. In today’s environment where health care providers generalizability of this important tool and thus limits its
have less and less time in which to see their patients, they uptake in offices, especially in primary care settings. As
are often forced to complete the entire visit in 15 to 20 min- opposed to another diagnostic tool, the electrocardiogram
utes. In such a scenario, the use of the glucose logbook to (ECG), which has the same standardized universal output,
look for patterns from which to make therapeutic recom- no such common format and universal output exists for
mendations is problematic. It is difficult to expect busy SMBG data. Thus, while the standard 12‐lead ECG tracing
providers to visually scan often messy, hand‐scribbled col- enjoys widespread use in clinical practice, SMBG use
umns of individual glucose values and try and make some remains most widely relegated to the use of a handwritten
sense of any emerging pattern after flipping through sev- glucose logbook. This is a major gap in our ability to teach
eral pages in a standard glucose logbook (Figure 7.1). The glycemic pattern recognition to our patients and fellow cli-
ability to download glucose meter data and present verified nicians, severely impeding the great potential of SMBG to
aggregate glucose data in an organized fashion, with basic help shape therapeutic interventions and improve overall
BREAKFAST LUNCH DINNER BEDTIME

SMBG values
Date Night BG Med BG BG Med BG BG Med BG BG Med Notes/Ketones tranlated in
BG
mg/dL mg/dL mg/dL mmol/L
2/24 179 199 227 9.9-11.0-15.1
2/25 198 177 189 Shopping 11.0-9.8-10.5
2/26 165 188 182 9.2-10.4-10.1
2/27 195 209 248 10.8-11.6-13.8
2/28 187 225 10.4-12.5
3/1 153 167 288 20 min walk 8.5-9.3-16.0
3/2 182 159 219 10.1-8.9-12.1
3/3 197 155 203 10.9-8.6-11.3
3/4 189 182 199 20 min walk 10.5-10.1-11.1
3/5 210 153 262 20 min walk 11.7-8.5-14.6
3/6 173 9.6
3/7
FIG 7.1 Typical patient glucose logbook with self‐reported data.

220 Overall target

200
180
160
140
Glucose
120
100
80
60
40
20
0
0 AM 0 AM 0 AM 0 PM 0 PM 0 PM 0 PM 0 PM 0 PM 0 AM 0 AM 0 AM 0 AM
6:0 8:0 10:0 12:0 2:0 4:0 6:0 8:0 10:0 12:0 2:0 4:0 6:0
Time of day
Statistics
Glucose average: 127 Target type: Personal

% Within target: 50 Before meal target: 70–110
# of glucose readings: 16 After meal target: 90–140
# of hypo. readings: 1 Hypoglycemic: 67
Standard deviation: 48
FIG 7.2 Modal Day presentation of downloaded meter with SMBG data.
levels of control. The representation of all SMBG data in increased risks for hypoglycemia, could many patients
expressed as a single day over 24 hours was an attempt to use these therapies singly or in combination without the
have some more common output from the multitude of need for any SMBG?
meters, many of which do allow data to also be expressed in These patients could be followed by period HbA1c
this form (Figure 7.2). Unfortunately, the widespread use of measurement to be sure that they are achieving their over-
such a modal day output never became commonplace, all glycemic targets. Might the cost savings based on forgo-
especially in primary care practice where they may be most ing the need for expensive glucose testing strips one or
useful. many times daily, as well as the ease of daily self‐manage-
The many barriers to the proper use of SMBG and the ment of their diabetes and potential perceived improve-
controversy still remaining around the utility of SMBG in ment in quality of life, justify such an approach?
patients with T2DM who are not on insulin have given rise Thus, there remains an urgent and ongoing need for
to an interesting question. Given that many available properly designed, well‐controlled RCTs to evaluate just
therapies including dietary and activity modification, what is the benefit of SMBG in those individuals with
or pharmacologic agents such as metformin, dipeptidyl T2DM treated with non‐insulin therapies. This popula-
peptidase‐IV (DDP‐4) inhibitors, glucagon‐like peptide‐1 tion, given their greater numbers, likely makes up the
(GLP‐1) receptor agonists, and thiazolidinediones (TZDs) lion’s share of the SMBG market and thus this question
all can improve overall glycemic control and do not result has enormously important ramifications for general pub-
lic health care policy in dealing with the growing diabetes with either T1DM or T2DM to use their next or following
epidemic. premeal SMBG, rather than a 2‐hour PPG, to gauge the
efficacy of their previous premeal insulin dose or other
therapy directed toward glycemic control. This assumes
Postprandial SMBG
that if the next premeal SMBG is within target, then the
The role of postprandial SMBG is firmly established in postprandial glucose values were likely “acceptable”. While
women with gestational diabetes mellitus (GDM). this method does provide a crude, indirect estimate of gly-
Adjusting insulin therapy in mothers with GDM, based cemic control over this period, it cannot guarantee glucose
on postprandial BG, resulted in improved HbA1c, lower excursions, or overall glycemic exposure, are well con-
birth weights (i.e., less macrosomia), less neonatal hypo- trolled. Thus, while this method is less labor‐intensive and
glycemia and less need for a cesarean section at may be an appropriate compromise in terms of “quality of
delivery [24]. life” for the individual, more intensive SMBG monitoring,
Meal‐based SMBG is a valuable tool for improving out- including PPG, should be considered if HbA1c is not
comes in pregnancy complicated by diabetes and has been within the desired target range.
shown to improve fetal perinatal outcomes [25]. This is In T2DM the role of postprandial SMBG is even less
more evident in insulin‐treated patients, and less clear with clear. T2DM usually has a more indolent onset, slowly pro-
diet‐controlled patients [26], although such monitoring gressing through phases of glucose intolerance to frank
may be useful in providing feedback for behavioral modifi- diabetes, and is closely associated with the metabolic syn-
cation and surveillance as to the adequacy of glucose con- drome and the increased cardiovascular risk that this
trol and the need to intensify therapies if not within target. entails. There have been several studies, such as the
Also, it encourages patients to actively participate in their DECODE, and others, that have found correlations with
own care. elevated postprandial glucoses and cardiovascular disease
The role of postprandial SMBG in T1DM and especially risk extending from impaired glucose tolerance right down
in T2DM is controversial. The landmark trials, such as the into the normal range for glucose, and in fact indicate
DCCT‐EDIC and UKPDS and its 10‐year follow‐up trials, greater association with risks of cardiovascular disease for
clearly showed that improving overall glycemic control, elevations in postprandial glucose (PPG) than fasting
thus reducing chronic exposure to hyperglycemia as meas- plasma glucose (FPG) [28, 29].
ured by the HbA1c, reduced the risks of microvascular and Again, most of the trials correlated improvement in
macrovascular complications [4–7]. Thus, the current rec- chronic glucose exposure as measured by HbA1c with
ommendations by the ADA are to target an HbA1c of < 7% decreased risk of long‐term complications. Some studies
in most individuals with diabetes [27]. These trials did not suggest a closer correlation of postprandial or post‐chal-
typically require postprandial SMBG, nor did they com- lenge glucose with HbA1c and mortality than fasting glu-
pare the effects of targeting fasting and premeal SMBG ver- cose [30], but this remains controversial and is not seen in
sus postprandial SMBG in reducing complications. yet other studies. Therapies that specifically target PPG
In T1DM it is not uncommon for individuals to monitor such as Acarbose have been shown to reduce CV disease
postprandial SMBG regularly to assess the adequacy of and all‐cause mortality in the Stop‐NIDDM trial [31, 32].
control exerted by their premeal short‐ or rapid‐acting Studies by Ceriello, and others, have provided indirect evi-
insulin on an MDI program or their bolus, if they are on dence that glycemic variability or excursions (typically
insulin‐pump continuous subcutaneous insulin infusion most marked in the immediate postprandial period)
(CSII) therapy. Monitoring postprandial SMBG provides directly contribute to diabetic complications through oxi-
these individuals with feedback both on insulin dosage and dative stress, and the generation of free radical formation,
self‐management behavior such as carbohydrate counting activation of the polyol pathways, and generation of PKC β
and their insulin to carbohydrate ratio. Nonetheless, it and advanced glycosylation end products (AGES) [33, 34].
remains a common clinical practice for many individuals The role PPG plays in the development of long‐term dia-
betic complications remains very controversial. One often‐ as an actionable item to potentially modify therapy or in
referenced paper, used to support this evidence that PPG surveillance to ensure adequate glycemic control is being
plays a role in complications in T1DM, noted that in the maintained.
DCCT trial those individuals in the intensive glycemic
treatment arm suffered from less diabetes‐related compli-
Continuous glucose monitoring systems
cations than did their counterparts in the standard arm,
even when matched for HbA1c, which is actually incor- The relatively recent advent of commercially available sub-
rect [35]. This was actually a hypothesized extrapolation or cutaneous continuous glucose monitoring (CGM) systems
modeling of the data and not directly representative of the that continuously measure interstitial fluid glucose has
clinical data itself [36]. While the controversy continues to added to the armamentarium of tools potentially useful for
rage, some groups like the International Diabetes the self‐management of diabetes. Devices from three man-
Federation (IDF) have published guidelines for the control ufacturers are currently available and licensed as an adjunc-
of PPG [37]. tive tool for diabetes management, most commonly in
Given the relative lack of strong RCT data comparing T1DM for persons on MDI or CSII insulin regimens. The
the benefits of postprandial SMBG versus fasting and FDA has mandated that these devices not be used for
immediate premeal SMBG measurement, why might post- directly calculating an insulin dose based on the most cur-
prandial SMBG measurement be of importance in indi- rent CGM reading, as their accuracy in comparison to
viduals with T2DM who are not treated with insulin? It those from approved glucose meters measuring capillary
may be necessary to target PPG to get more of these indi- venous whole blood has not been firmly established.
viduals to their glycemic goal. Monnier et al. have shown in Rather, the data generated from CGM can be useful in
a population of individuals with T2DM not treated with guiding necessary SMBG testing by delineating the real‐
insulin that the contribution of PPG to overall glycemic time trending in sequential glucose values, either indicat-
exposure rises progressively the closer one approaches a ing worsening hyperglycemia that may result in the need to
target HbA1c < 7% [38]. At HbA1c < 7.6% the contribution confirm with SMBG and potentially take an added supple-
of PPG to overall glycemic exposure totals approximately mental dose of insulin; or in indicating rapid glucose low-
70%. This, in turn, may help to explain why in several RCTs ering that may result in eventual hypoglycemia, warranting
that targeted interventions aimed at achieving an AM fast- treatment. These systems also make available alarms that
ing blood sugar within a specific goal range (typically less can call one’s attention to developing hyperglycemia or
than 100 mg/dL), many individuals who have achieved this hypoglycemia prompting corrective action.
target still have HbA1c that remains above 7%. With nearly Most studies of CGM to date have utilized this tool in
40% of the U.S. population not in target in terms of HbA1c, patients with T1DM on MDI or CSII insulin therapies, tak-
more focus on PPG through proper postprandial SMBG ing advantage of the real‐time glucose trending data to
monitoring and patient education on what action to take if fine‐tune insulin therapy. Just as with SMBG, CGM is but a
they are not in target may be necessary! As noted, the IDF tool to help guide one’s diabetes self‐management, and not
has published a guideline of the targeting of postprandial an antihyperglycemic intervention in and of itself. The
SMBG [37], and the ADA and AACE guidelines give rec- effects of CGMonHbA1c have been relatively mild but real.
ommended postprandial target glucose levels. The recent JDRF CGM study [39] showed that the use of
A reasonable approach may be to recommend targeted CGM in patients with T1DM on CSII did improve HbA1c
PPG or intensification of SMBG in individuals who are in adult patients (over 25 years of age), but not in children
either not within their goal range for HbA1c on their stabi- and teenagers who had much higher rates of non‐adher-
lized maintenance diabetes regimen, or in newly treated ence with proper CGM use (i.e., wearing a sensor) or using
patients with initially high HbA1c’s, as their HbA1c’s are the provided data for decision making.
lowered by initial therapies and are approaching 7%. The The Star 1 trial [40] did not demonstrate dramatic
key is to use SMBG when it serves a discreet purpose either HbA1c lowering in individuals with T1DM using CSII who
used CGM. It did suggest that in those who used the CGM, The potential future growth of CGM may well depend
the increased usage was associated with an increased prob- largely on its intermittent use for collection and summary
ability of HbA1c lowering. Many other moderate‐sized tri- of large quantities of glucose data for “glycemic pattern rec-
als (n of 100–200) assessing CGM use have come to similar ognition” in individuals with T2DM (or T1DM). The use of
conclusions, i.e., increased use of the CGM devices and the periodic CGM in this way would function as a “diagnostic
data they provide are associated with higher probability of biopsy” providing much more detailed and potentially
HbA1c lowering. The Star 3 trial compared the use of sen- more useful information than an HbA1c. This glycemic
sor‐augmented pump (SAP) therapy with a conventional pattern data would aid in targeting appropriate therapeutic
MDI program [41]. The trial demonstrated a ~0.8% decrease changes as well as delineating the optimal and appropriate
in HbA1c compared to a 0.2% decrease in the control group. timing of SMBG once the short period of CGM monitoring
Of course, by design, the study could not distinguish between is concluded.
the effect of CGM and CSII on the outcome. However, just as with SMBG, similar barriers are arising
Like almost any intervention, there has been a “down- with CGM, as each company promotes its own proprietary
side” occasionally noted in some individuals with T1DM software for data downloading. With no standardized uni-
who use a CGM. In some individuals there is a tendency to versal output and at least three competing and differing
“overbolus” in response to continuing high readings viewed software outputs, the widespread adoption of CGM tools in
on the device. This occurs when individuals feel compelled T2DM, especially in primary care, will face similar barriers
to frequently re‐bolus with short‐ or rapid‐acting insulin as the proper use of SMBG. Its use will inevitably be
when high glucose readings are viewed on the CGM device, retarded by the increased effort, logistical issues, and more
most of which provide a new reading every 1–5 minutes. complicated training and education the widespread use of
This impatience, upon frequently viewing high glucose fol- such varying outputs in the CGM will require of busy phy-
lowing an intervention that has not yet had adequate time sician practices. These issues are made even more complex
to work, can lead to “insulin stacking”. This phenomenon by the sheer amount of data generated by CGM devices (up
occurs where the effect of the most recent insulin dose is in to 15K glucose values over a 2‐week period) making even
addition to (“on top of ”) the ongoing insulin effect from organizing the data as modal day challenging to make clin-
the residual insulin on board from the previous injection, ically relevant sense of (Figures 7.3–7.5).
resulting in hypoglycemia. Proper education and training
of individuals as well as careful patient selection is required
Flash glucose monitoring systems
in choosing individuals who will benefit from CGM.
Another effect that has been seen is “sensor burnout”, A flash glucose monitoring system is akin to an intermit-
where it becomes difficult to maintain the ongoing effort tently scanned CGM and has been available for clinical use
required to respond to the wealth of data provided by a since 2017. Large, randomized controlled trials suggest
device in “real time”, which may or may not require an superiority over conventional SMBG in avoiding hypogly-
action on the individual’s part. In some studies, utilizing cemia and decreasing glycemic variability [42, 43]. The fre-
CGM the HbA1c can be seen to improve over the first quency of scanning and therefore the amount of
3–9 months only to regress toward the mean after more information available to guide management is patient
prolonged follow‐up because the information is no longer dependent and optimal benefit requires careful patient
being used with the same intensity. selection as well as education in how to use the data avail-
The past decade has seen significant advances towards able [44]. In such circumstances, this system provides
the adoption of pump therapy controlled by the continuous many of the benefits of CGM with less complexity.
data derived from a CGM device. The devices currently However, it cannot fulfil the sentinel functions of most
available and the ongoing work in this field are reviewed in conventional CGM – alerting patients when glucose con-
a separate chapter. centrations fall.
400
300
Guardian RT
Glucose-mg/dL
200
100
3:00 AM 6:00 AM 9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM
Modal Day
DexCom Seven 420
390
360
330
300
270
240
mg/dl
210
180
150
120
90
60
30
2.00 AM 4.00 AM 6.00 AM 8.00 AM 10.00 AM 12.00 PM 2.00 PM 4.00 PM 6.00 PM 8.00 PM 10.00 PM
Tue 2/13 Wed 2/14 Thu 2/15 Sat 2/17 Sun 2/18 Mon 2/19
FreeStyle Navigator 550

500
450
Glucose (mg/dL)
400
350
300
250
200
150
100
50
6:00 AM 12:00 PM 6:00 PM 12:00 AM

Time
FIG 7.3 Varying outputs from commercially available CGM devices.
350
Modal Day
300
250
200
150
100
50
0
12:00 AM
2:00 AM
4:00 AM
6:00 AM
8:00 AM
10:00 AM
12:00 PM
2:00 PM
4:00 PM
6:00 PM
8:00 PM
100:00 PM
120:00 AM
FIG 7.4 CGM device output as a Modal Day.

350
AGP (without diabetes)
300
AGP (patient with normal
glucose tolerance)
250
Glucose (mg/dL)
200
150
100
50
0
12:00 AM
2:00 AM
4:00 AM
6:00 AM
8:00 AM
10:00 AM
12:00 PM
2:00 PM
4:00 PM
6:00 PM
8:00 PM
100:00 PM
120:00 AM
350
AGP (with T2DM)
300
250
90th
Glucose (mg/dL)
200
75th
150 median
25th
100 10th
50
0
12:00 AM
2:00 AM
4:00 AM
6:00 AM
8:00 AM
10:00 AM
12:00 PM
2:00 PM
4:00 PM
6:00 PM
8:00 PM
100:00 PM
120:00 AM
N TARGETS AB O V E W IT H IN BELO W ME N SD M AX M IN AUC DAY

3632 140 70 43.0 53.1 4.0 137.3 44.2 279.0 39.0 3213.5 2372
Days Basal Bolus HbA1c 10th 25th 50th 75th 90th IQ Range NRM NORM
30 40 LA 45 RA 7.20% 87.0 107.9 134.0 164.0 198.1 56.0 133.89583 139.52941
FIG 7.5 Universal CGM output as ambulatory glucose profile (AGP). A. Subject without diabetes. B. Subject with T2DM.
It is important to keep in mind that both SMBG and the benefits of SMBG or CGM, it must be used with appro-
CGM are but tools to aid in diabetes self‐management and to priate understanding of how the information gleaned from
provide actionable data for patients and their caregivers. these tools is to be used to monitor, and if need be alter, one’s
They are not a therapeutic intervention expected to improve current therapies. These tools need to be viewed in the larger
overall glycemic control in and of themselves. To optimize context of the patient’s knowledge, abilities, financial
resources, and desire to utilize these tools appropriately. Control Cardiovascular Risk in Diabetes (ACCORD)
Current data, especially on the use of SMBG in patients with Investigators. Diabetes Care. 2010;33(5):983–990.
T2DM not utilizing insulin therapy remains fraught with 10. Del Prato S, LaSalle J, Matthaei S, Bailey CJ, on behalf of the
controversy. Opinions are driven by contradictory conclu- Global Partnership for Effective Diabetes Management.
Tailoring treatment to the individual in type 2 diabetes prac-
sions from multiple small, sometimes poorly designed
tical guidance from the Global Partnership for Effective
clinical trials, as well as various meta‐analysis. There
Diabetes Management. Int J Clin Pract. 2010;64:295–304.
remains a real need for well‐controlled, thoughtfully
11. Schiffrin A, Belmonte M. Multiple daily self‐glucose moni-
designed randomized clinically controlled trials to help toring: its essential role in long‐term glucose control in insu-
answer these questions. Groups are now attempting to lin‐dependent diabetic patients treated with pump and
define those important characteristics to be included in multiple subcutaneous injections. Diabetes Care. 1982;5:
such trials [16, 21]. 479–484.
12. Hansen MV, Pedersen‐Bjergaard U, Heller SR et al.
Frequency and motives of blood glucose self‐monitoring in
type 1 diabetes. Diabetes Res Clin Pract. 2009;85(2):183–
References
188. Epub 2009 Jun 3.
1. MacLean H. Modern Methods in the Diagnosis and Treatment 13. Evans JM, Newton RW, Ruta DA et al. Frequency of blood
of Glycosuria and Diabetes. London: Constable, 1924. glucose monitoring in relation to glycaemic control: obser-
2. Dufaitre‐Patouraux L, Vague P, Lassmann‐Vague V. History, vational study with diabetes database. BMJ. 1999;319(7202):
accuracy and precision of SMBG devices. Diabetes Metab. 83–86.
2003;29(2 Pt 2):S7–14. 14. Davidson PC, Bode BW, Steed RD, Hebblewhite HR. A
3. Centers for Disease Control and Prevention. MMWR self‐ cause‐and‐effect‐based mathematical curvilinear model that
monitoring of blood glucose among adults with diabetes – predicts the effects of self‐monitoring of blood glucose fre-
United States, 1997–2006. Morb Mortal Wkly Rep. 2007; quency on hemoglobin A1c and is suitable for statistical cor-
56(43):1133–1137. relations. J Diabetes Sci Technol. 2007;1(6):850–856.
4. The Diabetes Control and Complications Trial/ 15. Murata GH, Shah JH, Hoffman RM et al. Intensified blood
Epidemiology of Diabetes Interventions and Complications glucose monitoring improves glycemic control in stable,
(DCCT/EDIC) Study Research Group. Intensive diabetes insulin‐treated veterans with type 2 diabetes: the Diabetes
treatment and cardiovascular disease in patients with type 1 Outcomes in Veterans Study (DOVES). Diabetes Care.
diabetes. N Engl J Med. 2005;353:2643–2653. 2003;26(6):1759–1763.
5. The Diabetes Control and Complications Trial/Epidemiology 16. Klonoff DC, Bergenstal R, Blonde R et al. Consensus report
of Diabetes Interventions and Complications Research of the coalition for clinical research – self‐monitoring of
Group. Sustained effect of intensive treatment of type 1 blood glucose. J Diabetes Sci Technol. 2008;2(6):
diabetes mellitus on development and progression of 1030–1053.
diabetic nephropathy. JAMA. 2003;290: 2159–2167. 17. Davis WA, Bruce DG, Davis TME. Does self‐monitor of
6. Holman RR, Paul SK, Bethel MA et al. 10‐year follow‐up of blood glucose improve outcome in type 2 diabetes? The
intensive glucose control in type 2 diabetes. N Engl J Med. Fremantle Study. Diabetologia. 2007;50:510–515.
2008;359:1577–1589. 18. Welschen LMC, Bloemendal E, Nijpels G et al. Self‐monitor-
7. The Diabetes Control and Complications Trial Research ing of blood glucose in patients with type 2 diabetes mellitus
Group. The effect of intensive treatment of diabetes on the who are not using insulin. Cochrane Database Syst Rev.
development and progression of long‐term complications in 2005;2 (Art. No.: CD005060). DOI: 10.1002/14651858.
insulin‐dependent diabetes mellitus. N Engl J Med. CD005060.pub2.
1993;329:977–986. 19. Farmer A, Wade A, Goyder E et al. Impact of self monitoring
8. The Action to Control Cardiovascular Risk in Diabetes Study of blood glucose in the management of patients with non‐
Group. Effects of intensive glucose lowering in type 2 insulin treated diabetes: open parallel group randomized
diabetes. N Engl J Med. 2008;358:2545–2559. trial. BMJ. 2007;335(7611):132.
9. Riddle MC, Ambrosius WT, Brillon DJ, Buse JB, Byington 20. Farmer A, Wade A, French DP et al. Blood glucose self‐mon-
RP, Cohen RM, Goff DC Jr, Malozowski S, Margolis KL, itoring in type 2 diabetes: a randomised controlled trial
Probstfield JL, Schnall A, Seaquist ER, and for the Action to Health Technol Assess. 2009;13(15).
21. Hirsch IB, Bode BW, Childs BP et al. Self‐Monitoring of 33. Ceriello A, Ihnat M, Thorpe J. Clinical review 2: the “meta-
Blood Glucose (SMBG) in insulin‐ and non‐insulinusing bolic memory”: is more than just tight glucose control nec-
adults with diabetes: consensus recommendations for essary to prevent diabetic complications? Journal Of Clinical
improving SMBG accuracy, utilization, and research. Endocrinology And Metabolism. 94(2):410–415, 2009.
Diabetes Technol Ther. 2008;10(6):419–439. 34. Brownlee M, Hirsch I. Glycemic variability: a hemoglobin
22. Bergenstal R, Johnson M, Powers M et al. Adjust to target in A1c‐independent risk factor for diabetic complications.
type 2 diabetes. Diabetes Care. 2008;31(7):1305–1310. JAMA. 295(14):1707–1708, 2006.
23. Brown LB, Nichols GA, Perry A. The burden of treatment 35. DCCT Study Group. The relationship of glycemic exposure
failure in type 2 diabetes. Diabetes Care. 2004;27: (HbA1c) to the risk of development and progression of
1535–1540. retinopathy in the diabetes control and complications trial.
24. DeVeciana M, Major C, Morgan M et al. Postprandial versus Diabetes. 1995;44:968–983.
preprandial blood glucose monitoring in women with gesta- 36. Lachin JM, Genuth S, Nathan DM, Zinman B, Rutledge BN
tional diabetes mellitus requiring insulin therapy. N Engl J and for the DCCT/EDIC Research Group. Effect of glycemic
Med. 1995;333(19):1237–1241. exposure on the risk of microvascular complications in the
25. Wecher DJ, Kaufmann RC, Amankwah KS et al. Prevention diabetes control and complications trial – Revisited.
of neonatal macrosomia in gestational diabetes by the use of Diabetes. 2008;57:995–1001.
intensive dietary therapy and home glucose monitoring. Am 37. Guideline for Management of Postmeal Glucose.
J Perinatol. 1991;8:131–134. International Diabetes Federation, 2007; www.idf.org.
26. Homko CJ, Sivan E, Reece EA. The impact of self‐monitor- 38. Monnier L, Lapinski H, Colette C. Contributions of fasting
ing of blood glucose on self‐efficacy and pregnancy out- and postprandial plasma glucose increments to the overall
comes in women with diet‐controlled gestational diabetes. diurnal hyperglycemia of type 2 diabetic patients: variations
Diabetes Educ. 2002;28(3):435–443. with increasing levels of HbA1c. Diabetes Care. 2003;26:
27. 2010 ADA Clinical Practice Recommendations. 881–885.
28. The DECODE Study Group. Glucose tolerance and mortal- 39. Juvenile Diabetes Research Foundation Continuous Glucose
ity: comparison of WHO and American Diabetes Monitoring Study Group. Continuous glucose monitoring
Association diagnostic criteria. The DECODE study group. and intensive treatment of type 1 diabetes. N Engl J Med.
European Diabetes Epidemiology Group. Diabetes epidemi- 2008;359:1464–1476. DOI: 10.1056/NEJMoa0805017.
ology: collaborative analysis of diagnostic criteria in Europe. 40. Hirsch IB, Abelseth J, Bode BW et al. Sensor‐augmented
Lancet. 1999;354:617–621. insulin pump therapy: results of the first randomized treat‐
29. Ohkubo Y, Kishikawa H, Araki E et al. Intensive insulin to‐target study. Diabetes Technol Ther. 2008;10:377–383.
therapy prevents the progression of diabetic microvascular 41. Bergenstal et al. Effectiveness of sensor‐augmented insulin‐
complications in Japanese patients with non‐insulin‐ pump therapy in type 1 diabetes. N Engl J Med. 2010;363:
dependent diabetes mellitus: a randomized prospective 6‐ 311–320.
year study. Diabetes Res Clin Pract. 1995;28:103–117. 42. Bolinder et al. Novel glucose‐sensing technology and hypo-
30. Avignon A, Radauceanu A, Monnier L. Nonfasting plasma glu- glycaemia in type 1 diabetes: a multicentre, non‐masked,
cose is a better marker of diabetic control than fasting plasma randomised controlled trial. Lancet. 2016;388(10057):
glucose in type 2 diabetes. Diabetes Care. 1997;20: 1822–1826. 2254–2263.
31. Chiasson JL, Josse RG, Gomis R et al. Acarbose treatment 43. Haak et al. Flash glucose‐sensing technology as a replace-
and the risk of cardiovascular disease and hypertension in ment for blood glucose monitoring for the management of
patients with impaired glucose tolerance: the STOP‐NIDDM insulin‐treated type 2 diabetes: a multicenter, open‐label
trial. JAMA. 2003;290:486–494. randomized controlled trial. Diabetes Ther. 2017;8(1):
32. Hanefeld M, Cagatay M, Petrowitsch T et al. Acarbose 55–73.
reduces the risk for myocardial infarction in type 2 diabetic 44. Kudva et al. Approach to using trend arrows in the FreeStyle
patients: meta‐analysis of seven long‐term studies. Eur Libre flash glucose monitoring systems in adults. J Endocr
Heart J. 2004;25(1):10–16. Soc. 2018;2(12):1320–1337.
8 Does HbA1c Remain the Most Important
Therapeutic Target in Outpatient
Management of Diabetes?
Kristen Gonzales1 and Steve A. Smith2
1
Department of Internal Medicine, Division of Endocrinology, Diabetes, and Metabolism, University of New Mexico
Health Sciences Center, Albuquerque, NM, USA
2
intermittent and/or sustained hyperglycemia [9]. These

have included impaired endothelial cell dysfunction [10],
endothelium‐dependent relaxation, endothelial cell apop-
●● HbA1c is a measure for not only characterizing
diabetes, but also in relaying predictive information
tosis [11], altered nitric oxide and aldose reductase, oxygen
regarding the effects of chronic hyperglycemia derived free radicals [12, 13], protein kinase C, Na+K+‐
●● Various clinical states can interfere with accurate ATPase, and advanced glycosylation end products [14].
interpretation of HbA1c. Some of the most highly‐researched mechanisms for
●● Going beyond HbA1c as a measure implies the microvascular disruption related to hyperglycemia have
utilization of novel glycemic measures to guide more
comprehensive assessments that place the patient at
been the aldose reductase theory, advanced glycation end‐
the front‐lines of management. product theory, reactive oxygen intermediate pathway, and
the protein kinase C pathway. However, clinical evidence is
lacking for these proposed mechanisms, making develop-
ment of targeted therapies challenging [9]. Thus, minimi-
The glucose hypothesis and vascular risk
zation and prevention of vascular complications through
Hyperglycemia is strongly associated with both micro‐ and glycemic control is critical, and accurate measures of
macrovascular events in people with diabetes. The Diabetes hyperglycemia are imperative to this goal.
Control and Complication (DCCT) and the Kumamoto Ideal clinical measures of hyperglycemia should inform
trials were the first to demonstrate the direct effects of tar- the patient and clinician in their management and be pre-
geted improvements in glycemic control in both Type 1 dictive of both short‐ and long‐term vascular complica-
(DM1) and Type 2 (DM2) diabetes, respectively [1, 2]. tions. In this chapter, we will review the historical and
Since those early studies, many randomized control trials future assessments of glycemic control in the care of people
have demonstrated favorable impacts on cardiovascular with diabetes. This requires precise methods for measuring
and renal outcomes with improved glycemic control, fur- (1) chronic and intermittent hyperglycemia, (2) hypoglyce-
ther confirming the notion that glycemic control is strongly mia, and (3) glycemic variability. We will address measures
linked to vascular risk [3–8]. for each of these and further evaluate the evidence sup-
A number of hypotheses have been proposed for the porting standard or more novel measures of glycemic
development of vascular complications as a result of control.

101
Measuring glycemic control in the clinical tion were laborious and inconsistent, thus prolonging its
management of diabetes acceptance into routine clinical practice [15].
The impetus for standardization of the HbA1c assay was
Hemoglobin A1c: Discovery and integration
borne out of the landmark findings of the DCCT. In this
into clinical practice
study, HbA1c was used as a primary measurement of glyce-
Discussion of the hemoglobin A1c (HbA1c) and its role in
mic control, and measurements were obtained through a
diabetes management is key to the examination of optimal
centralized laboratory using high‐performance liquid
modalities for measuring glycemic control. HbA1c was
chromatography. With the demonstration of marked ben-
first discovered in 1968 and introduced into clinical prac-
efit in microvascular outcomes through improved glyce-
tice in 1977 [15, 16]. It has continued to be used as the
mic control, the HbA1c quickly became recognized as a
main outcome for diabetes management in both clinical
key target for diabetes care. As a result, national programs
and research settings given its value in predicting risk asso-
in the United States, Japan, and Sweden were established to
ciated with long‐term glycemic exposure [17]. Historically,
standardize the assay such that techniques and results
the American Diabetes Association (ADA) has maintained
would be directly comparable to those of the DCCT.
that the primary measures of glycemic control should include
Ultimately, the International Federation of Clinical
both self‐monitored blood glucose values (SMBG) and
Chemistry and Laboratory Medicine (IFCC) Working
HbA1c. However, with the advent of continuous glucose
Group on Standardization of HbA1c established an inter-
monitoring (CGM) and the recognition that interstitial glu-
national HbA1c assay employing either HPLC mass spec-
cose as measured through CGM is well‐correlated with blood
trometry or capillary electrophoresis. The IFCC also
glucose, the ADA officially incorporated CGM as a standard established a reporting system using units of either mmol/
measure of glycemic control in 2008 [18]. CGM, as will be dis- dL or percent glycosylation [15]. In total, these efforts
cussed, has paved the way for enhanced precision in glycemic allowed for great strides to be made in the minimization of
measures which have been associated with improved out- inter‐laboratory variability and propelled the HbA1c for-
comes. We will discuss each of these measures in the subse- ward as a meaningful population measure of glycemic
quent sections, beginning with an overview of the role for control.
HbA1c in both past and present‐day diabetes care.
HbA1c is a hemoglobin A moiety that has undergone HbA1c: Role in predicting micro‐ and
valine glycosylation at the N‐terminus of its beta‐chain macrovascular complications
through a non‐enzymatic ketoamine linkage. This ketoam- Multiple prospective trials have employed HbA1c as a pri-
ine linkage creates a stable HbA1c complex which forms mary measure for glucose control when assessing out-
slowly and continuously throughout the life span of the red comes associated with long‐standing hyperglycemia. The
blood cell, thus reflecting hemoglobin glycosylation over a DCCT (1982–1993), was a prospective clinical trial which
2‐ to 3‐month span. It is therefore an indirect measure of randomized patients with DM1 to receive either intensive
average glycemic exposure. Rhabar and Trivelli first insulin therapy, where the average HbA1c achieved was
described its importance in diabetes by noting a two‐fold 7.1%, or conventional insulin therapy, where the average
increase in this moiety among individuals with diabetes HbA1c achieved was 9.1%. Primary outcomes were the
compared to those without [16, 19]. Trivelli and colleagues development and progression of retinopathy, nephropathy,
also described a relationship between HbA1c, mean blood and neuropathy. The study demonstrated a strong correla-
glucose, and long‐term complications of diabetes, thus tion between both prevention and progression of retinopa-
establishing HbA1c as a measure not only for characteriz- thy with HbA1c reduction in the intensive treatment
ing diabetes, but also in relaying predictive information group, and progression was likewise linked exponentially
regarding the effects of chronic hyperglycemia [19]. with rising HbA1c levels [1]. In a long‐term observational
However, validation and standardization of the HbA1c follow‐up of the DCCT cohort, the Epidemiology of
assay took several decades, as the methods used for isola- Diabetes Interventions and Complications (EDIC) trial,
Is HbA1c the Most Important Therapeutic Target in the Outpatient Management of Diabetes? 103
HbA1c values converged between both groups. Yet, there these groups [25–27]. Thus, the benefits of HbA1c lower-
remained a marked risk reduction in microvascular disease ing based on established cutoffs for development of micro-
in the intensive group compared to the conventional group vascular complications must be counterbalanced by the
over 2 decades of follow‐up. There were also significant risks of hypoglycemia, particularly in select populations,
reductions in fatal and non‐fatal myocardial infarctions and thus broadening the scope of glycemic control.
strokes. These data suggested that early, intensive lowering
of HbA1c portended long‐term vascular benefits in people Clinical utility of HbA1c
with DM1, a phenomenon known as “metabolic mem- There are several reasons for why the HbA1c has so deeply
ory.” [20] Similar findings regarding micro‐ and macrovas- permeated general understanding of chronic glycemic con-
cular risk reduction among people with DM2 who achieved trol and the complications associated with diabetes.
early HbA1c lowering were demonstrated in the United HbA1c, being a measure of hemoglobin glycosylation, pro-
Kingdom Prospective Diabetes Study (UKPDS) [21, 22]. vides insight into the average degree of glycemic exposure
Notably, neither the DCCT nor the UKPDS were over a 2‐ to 3‐month span [28]. It is widely accepted that
designed to establish HbA1c thresholds beyond which reducing blood glucose levels corresponds to a reduction
microvascular disease progression was likely to occur. in microvascular and macrovascular complications, at least
Rather, insight into these HbA1c cutoffs was gleaned from when using HbA1c and SMBG values as surrogate meas-
the 6‐year Kumamoto Study, a randomized control trial ures, and these generally serve as good predictors for devel-
involving individuals with DM2 randomized to receive opment of diabetes complications [1, 2, 20, 21, 25].
either intensive or conventional insulin therapy. As In 2009, the International Expert Committee on diabe-
expected, rates of microvascular complications were tes officially incorporated HbA1c into the diagnostic crite-
reduced in those randomized to the intensive insulin ther- ria for diabetes, establishing a cutoff of ≥ 6.5% as being
apy group, and the threshold for microvascular progression consistent with the diagnosis and correlating with a point
was correlated with an average HbA1c > 6.5%, a FPG < in the disease beyond which microvascular risk markedly
110 mg/dL, and a 2‐hour postprandial glucose > 180 mg/ increased [24]. Thus, based on the available evidence to
dL [2]. Similarly, the Stockholm Diabetes Intervention that point, HbA1c was and continues to be a valuable
Study, a smaller 7.5‐year prospective trial, demonstrated measure of glycemic exposure. From a clinical perspective,
that individuals with mild retinopathy at study entry did performance of the HbA1c assay is convenient, timely,
not develop severe retinopathy, nephropathy, or progres- does not require fasting, portends a high degree of pre‐ana-
sive eye disease if mean HbA1c was maintained < 7% for lytical stability, and is less impacted by daily perturbations
the duration of the study [23]. Finally, a compilation of epi- in other metabolic factors. From a population health stand-
demiologic data has shown that HbA1c is useful for pre- point, the HbA1c is perhaps the best available tool for
dicting development of retinopathy, and an HbA1c of ≥ establishing glycemic goals which can then be further indi-
6.5% represents an inflection point beyond which retinop- vidualized according to patient‐specific factors. The use of
athy risk substantially increases, thus justifying this as a HbA1c to set individualized therapeutic targets, however,
threshold for the diagnosis of diabetes [24]. is limited. As we will explore in the subsequent sections, a
With identification of HbA1c thresholds below which focus on HbA1c targets alone may be insufficient for opti-
vascular risks are significantly reduced, a new challenge in mizing diabetes management and preventing related vas-
glycemic control and diabetes management emerged: that cular outcomes.
of hypoglycemia. Perhaps the most robust trials to high-
light this risk with intensive HbA1c lowering were the Limitations of HbA1c measurement
VADT, ADVANCE and ACCORD trials, each involving Due to the fact that HbA1c measures glycosylation of
individuals with DM2. These trials targeted HbA1c goals at hemoglobin, there are various clinical states that can inter-
least less than 7% in the intensive treatment arms and dem- fere with accurate interpretation of HbA1c. Conditions
onstrated significantly higher rates of hypoglycemia in such as sickle cell anemia, pregnancy and post‐partum
states, glucose‐6‐phosphate dehydrogenase deficiency, 14 days, fructosamine provides a semi‐chronic assessment

HIV, hemodialysis, recent blood loss or transfusion, eryth- of glucose exposure [31]. Finally, 1,5‐anhydroglucitol is
ropoietin therapy, and B12, iron, or folate deficiencies may another measure that provides some insight into hypergly-
significantly impact the reported values [29]. Thus, alter- cemia, traditionally declining in the serum with progres-
nate measures should be employed for management deci- sive hyperglycemia and rising with improved glycemic
sions in such cases. In clinical practice, fasting plasma control [32]. While these latter measures are reasonable to
glucose concentrations, and/or other measures of glycemia consider in the presence of medical conditions which may
which will be discussed should be utilized. interfere with accurate HbA1c measurements, their prog-
Additionally, HbA1c does not reflect rapid excursions in nostic value for diabetes complications is not well‐estab-
glucose levels and, and it provides limited value in guiding lished, thus limiting their utility as ideal measures of
daily self‐management. As daily self‐management entails glycemic control.
use of data derived from either SMBG measurements or
CGM, correlation of these values with HbA1c is critical for Self‐monitored blood glucose (SMBG)
disease management. SMBG, as measured through hand‐held glucometers, has
largely replaced traditional measures of glycemia in the
Traditional measures of blood glucose day‐to‐day care of individuals with diabetes [33, 34]. Use of
and glucose exposure SMBG, particularly when used with increased frequency,
Hyperglycemia can be detected through both clinical and bio- provides details regarding daily glycemic patterns, thus
chemical assessments. Polyuria, polydipsia, and nocturia are enabling patients and clinicians to tailor therapies toward
perhaps the most common symptoms. Depending upon the specific glucose goals. These glycemic “snapshots” have
duration and severity of hyperglycemia, other clinical features been deeply embedded into clinical management of diabe-
may include fatigue, lethargy, blurred vision, weight loss, keto- tes, and their use is strongly correlated with reductions in
sis, and infections. However, reliance on clinical symptoms is HbA1c.
insufficient for the assessment of glycemic control. [30] Currently, the ADA recommends that patients on an
Prior to the mid‐1970s, glucose control was measured pri- intensive insulin regimen perform SMBG measurements
marily through urine testing. While random or 24‐hour urine before meals and snacks, fasting, before bedtime, if needed
glucose collections provide some insight into glycemic bur- postprandial, before exercise, when experiencing a low,
den, they are relatively inaccurate and fail to adequately guide until the low is resolved, and before driving. It is recog-
targeted treatment goals. Renal glucose thresholds are highly nized that frequent SMBG measures can guide medical
variable between individuals, and lack of glucosuria provides nutrition therapy, physical activity, insulin dose adjust-
little information regarding the presence of hypoglycemia, ments, and prevention/treatment of hypoglycemia [35].
euglycemia, or even mild hyperglycemia [31]. Thus, there is In the Diabetes Control and Complications Trial
poor correlation between urinary and blood glucose values. (DCCT), participants randomized to the intensive treat-
The development of capillary glucose testing largely ment group performed SMBG measurements at least 4
replaced home urinary measurements by the 1980s and times daily, in addition to a 3 a.m. check once weekly, and
revolutionized management of diabetes by allowing for tar- 7 times daily once every 3 months [1, 36]. In contrast the
geted goals for blood glucose [31]. SMBG with capillary conventional treatment group performed SMBG measure-
glucose measurements remains a mainstay of measuring ments twice daily. While SMBG frequency was not corre-
glycemic control and provides day‐to‐day guidance for lated directly to the primary outcomes of the trial, the
management decisions. The evidence for SMBG will be intensive treatment group overall demonstrated delayed
discussed in the subsequent section. progression and development of microvascular complica-
In the 1990s, discovery of glycated proteins, namely fruc- tions [1]. This data would suggest that SMBG frequency
tosamine, provided an alternative measurement of glycemic may in fact be a significant component of improved glyce-
control. Given the protein half life in circulation is roughly mic control.
Several large registry studies have since demonstrated metrics such as fasting, pre‐ and postprandial levels. As
that SMBG frequency is strongly correlated with a decrease these measures are further examined in relation to diabetes
in HbA1c for all age groups among both insulin pump and outcomes, it becomes apparent that a focus on HbA1c
multiple daily injection (MDI) users. In one study using alone may obscure other meaningful therapeutic targets.
the Type 1 Diabetes Exchange Registry, a repository of 20
555 children and adults with type 1 diabetes, the individu- Post prandial glucose (PPG)
als who tested 3–4 times daily had a mean HbA1c of 8.6%, Population studies have proposed a possible relationship
whereas those who tested ≥ 10 times daily had a mean between risk of cardiovascular events and postprandial
HbA1c of 7.6%. The HbA1c lowering correlated with hyperglycemia (PPG) [39, 40]. Initial studies designed to
increasing frequency of SMBG measurements and reached investigate this possible relationship have been limited to
a plateau at 10–12 checks daily [37]. In a similar study cross sectional observations and associations of elevated
using an Austrian database of both patients with type 1 and values at 2 hours on oral glucose tolerance testing (OGTT)
2 diabetes, HbA1c reduction was again correlated with as opposed to mixed‐meal or real‐life glucose expo-
SMBG frequency. The HbA1c reduction in this study was sure [41–43]. Studies that have compared outcomes for
shown to be 0.26% for every one additional SMBG check/ individuals that emphasize PPG; meglitinides
day, p < 0.0001. Notably, this effect was only seen in patients (NAVIGATOR) [44], alpha glucosidase inhibitors (STOP‐
using insulin therapy. Indeed, there was a negative correla- NIDDM) [45], and basal insulin only (HEART2D) have
tion of SMBG monitoring frequency and glycemic control been limited based on technical challenges in assessing
noted among patients with type 2 diabetes treated with oral duration of glycemic exposure [42]. In the San Luigi
agents [38]. Gonzaga Diabetes Study with 14 years of follow‐up, HbA1c
While SMBG frequency has been correlated with and glucose levels 2 hours after lunch were predictive of
improvement in glycemic control, this alone is not the sole first cardiovascular events and all‐cause mortality, [46]
factor in HbA1c improvement. Rather, accurate interpreta- while Hopper and colleagues have completed a meta analy-
tion of the readings and integration of the data obtained sis that included 23 152 patients with impaired glucose
from SMBG into daily management is an integral compo- intolerance from 10 controlled trials randomized to diet,
nent of diabetes management. Additional factors to con- exercise, or hypoglycemic agent. After an average of 3.75 of
sider regarding SMBG as a measure of glycemic control are years of follow‐up, there was no reduction in all‐cause or
patient technique when testing, device accuracy, and the cardiovascular mortality [47].
patient burden of frequent monitoring. With this back- Just as the Kumamoto trial demonstrated a reduced pro-
ground, how is the information obtained through gression of microvascular complications in patients who
SMBG measurement linked to overall outcomes in diabe- maintained an average HbA1c of < 6.5% during the 6‐year
tes, and is this information valuable in influencing vascular follow‐up period, it also demonstrated that this reduction
outcomes? was correlated with 2‐hour postprandial levels < 180 mg/
dL. Rising PPG levels were associated with worsening
retinopathy and nephropathy [2].
Additional therapeutic targets to consider
PPG, like HbA1c, offers little insight into the assessment
for hyperglycemia in the outpatient
of glycemic variability and, therefore, provides little addi-
management of diabetes
tional information beyond HbA1c. Because of the observa-
The term “beyond A1c” has recently been introduced into tion that the best predictor of postprandial glucose control
the diabetes lexicon as a concept describing the advance- is preprandial levels, it has been argued that the best
ment of glycemic measures toward more comprehensive treatment approach for hyperglycemia would be to limit
assessments that place the patient at the front‐lines of man- the incidence of hypoglycemia by lowering preprandial
agement. SMBG has paved the way for insights into these fasting hyperglycemia prior to emphasis of postprandial
more nuanced measures, providing insight into valuable hyperglycemia [48].
Continuous glucose monitoring (CGM) and range (70–180 mg/dL), time spent per day in hypoglyce-
glycemic control mia, and time spent in hyperglycemia [50]. The magnitude
While SMBG has been crucial in diabetes care and fre- of HbA1c benefit over 6 months in the CGM vs. control
quency of use linked to enhanced glycemic control, many group was similar to that seen in the SWITCH study. In the
individuals on insulin are unable to achieve glycemic tar- latter, there was a 0.43% mean difference in HbA1c between
gets even with standard 4‐times‐daily SMBG monitoring. patients with insulin pumps using CGM vs. those not using
One reason for this is the rarity with which postprandial CGM. Secondary endpoints in SWITCH included changes
and overnight glucose profiles are able to be examined with in glycemic patterns, and time spent hypoglycemia, hyper-
standard SMBG measures and home glucometers. Over the glycemia, and normoglycemia [51]. As with any intensive
last few decades, CGM has been developed in an effort to method of achieving improved glycemic control and
overcome the monitoring limitations of conventional HbA1c reduction, the counterbalance of hypoglycemia is
SMBG. of paramount importance.
CGM first came into use in 1999 and, since that time,
technological advances have led to marked improvements CGM and hypoglycemia
in both accuracy and ease of use [49]. There has subse- Several major trials examining the effectiveness and safety
quently been an increasing real‐time use of CGM in the of CGM have either reported no increase or significant
management of glucose control. More recently, intermit- reduction in hypoglycemia events compared to non‐CGM
tently‐scanned CGM devices, which minimize the burden users despite universal HbA1c‐lowering in the studies
of frequent fingerstick checks, have been introduced into among CGM users. The Juvenile Diabetes Research
clinical practice. These devices have all allowed for Foundation (JDRF) study was perhaps one of the first to
enhanced utilization of glycemic measurements while design a randomized trial in which individuals with DM1
minimizing the daily burden of disease on individuals with and HbA1c values < 7.0% at study entry were recruited to
diabetes. either CGM‐use or standard monitoring. The study dem-
CGM mathematically derives estimates of plasma glu- onstrated a decrease in time spent with a blood glucose ≤
cose from the measurement of interstitial glucose (which 70 mg/dL among CGM‐users (54 vs. 91 mins/day),
correlates well with plasma glucose), providing up to 288 although the difference did not reach statistical signifi-
glucose measurements daily. Sensors are calibrated with cance [52]. The SWITCH trial documented an overall
SMBG measurements and, as of 2013, the latter is recom- improvement in HbA1c and no difference in hypoglycemia
mended for making acute adjustments in insulin. Several events among CGM‐users vs. non‐users [51]. Similarly, the
studies of CGM use have been conducted in patients with (STAR) 3 trial examined sensor‐augmented pump (SAP)
DM1, although more robust data is needed in those with therapy vs. standard MDI therapy in participants with
DM2. Using insulin pumps with CGM has demonstrated DM1 and found no significant differences in hypoglyce-
benefit in achieving glycemic control, as well. Still, fewer mia, but overall greater reduction in HbA1c among the
studies have looked at the use of CGM in patients with SAP group [53]. However, the ASPIRE trial included DM1
DM1 using MDI. This is an important issue given the patients randomized to SAP vs. standard pump therapy
majority of patients with DM1 use MDI rather than pump without CGM. In this study the interventional group had
therapy. 31.8% reduced frequency of nocturnal hypoglycemia over
The DIAMOND trial sought to specifically determine 3‐month use with no difference in HbA1c compared to
the role of CGM in patients with DM1 using MDI. It found standard group over that time [54].
that patients using CGM had a 1% reduction in in HbA1c One crucial element of modern CGMs that has been
after 6 months compared to 0.4% reduction in the control shown to reduce hypoglycemia is that of automated low
group (MDI without CGM) or SMBG checks 4 times daily. glucose suspend (LGS). LGS is an automated feature of
Notably, this trial also measured many secondary out- SAP therapy in which insulin delivery is automatically sus-
comes including glucose variability (CV), time per day in pended when interstitial glucose values fall below a certain
threshold. Indeed, the ASPIRE study utilized this LGS involving CGM has been the hybrid closed‐loop pump, a
feature [53]. Studies involving pediatric populations system in which basal rates are determined and adjusted
with DM1 have also demonstrated promising results automatically by an individualized algorithm derived from
regarding reduced time spent in hypoglycemia (< CGM. Use of this pump has not only shown improve-
70 mg/dl, 101 ± 68 versus 58 ± 33 min/day; p < .002), as ment in HbA1c, but also in many of the newly‐standard-
well as reduced number of hypoglycemic excursions ized CGM metrics: Time in target range increased from
during the day (< 70 mg/dl, 1.27 ± 0.75 vs. 0.95 ± 0.49, p 64% to > 70%; time below target range (≤ 70 mg/dL)
< .01; < 40 mg/dl, 0.28 ± 0.18 versus 0.13 ± 0.14, p < decreased by 30%; and coefficient of variation decreased
.005) [55]. from 33% to 30% [57]. Thus, familiarity with how these
metrics relate to glycemic control is critical as diabetes
Utility of CGM metrics in the management of care transitions beyond HbA1c. We will discuss a few of
diabetes the major metrics and the evidence for their use in dia-
CGM use has been strongly linked to improved glycemic betes care below.
control as measured using the primary outcome of HbA1c.
However, these benefits stem from the detailed informa- Metric 1: Duration of CGM use
tion obtained from these devices, as they allow for immedi- One caveat to the benefits seen in studies involving
ate detection of glycemic excursions and daily patterns not CGM was that sufficient use of the CGM device was
otherwise attainable through standard SMBG using finger- necessary. Among CGM‐users in the SWITCH trial,
stick checks. Using CGM data, a series of metrics have been mean sensor use was 80% of the required time, and
developed which have proven correlation with glycemic those with < 70% use demonstrated smaller decreases in
control. Some of the more common ones include time HbA1c [51]. In the JDRF study, the group that achieved
within/above/below range, glycemic variability as meas- the greatest HbA1c benefit with CGM were those ≥ 25
ured through a coefficient of variation or standard devia- years of age, and roughly 83% of the participants in this
tion (CV or SD), and duration of CGM use. Table 8.1 group utilized the CGM at least 6 days per week. Use
provides a comprehensive list of standardized metrics for was significantly less in the younger age groups (8–24
CGM reporting [56]. Notably, the most recent technology years old), and there were no differences in glycemic
control noted between CGM‐users and non‐users (con-
trol) in these groups [58].
Duration of CGM use, in addition to having proven
TABLE 8.1 International consensus on use of CGM.
benefit on glycemic control during the study time‐frames,
Mean glucose eA1c has also been correlated with long‐term glycemic control.
Xing et al. showed that CGM use between 12–15 days was
% time level 2 TBR Glucose metrics
well‐correlated (R2 range 0.66–0.75) with 3‐month glyce-
< 54 mgm% (3 mmol/L) Sleep, Wake, 24 hours
% time level 1 TBR Data sufficiency (2 weeks) mic control as measured by mean glucose, percentage of
< 70‐54 mgm% (3.9–3.0 mmol/L) values between 71–180 mg/dL and < 70 mg/dL, and glyce-
% time TIR Data sufficiency mic variability [59]. More recently, data using newer CGM
70‐180mgm% (3.9–10 mmol/L) 70–80% CGM data over technology has demonstrated an even greater correlation
2 weeks
(R2 range 0.84–0.86) of 14‐day CGM use with 3‐month
% time level 1 TAR Hypoglycemia standard
> 180mgm% (10 mmol//L)1 definition measures of glycemic control [60]. Beyond 14 days of use,
Hyperglycemia standard there was little incremental change noted in correlation
definition with long‐term outcome measures. Thus, 10‐ to 14‐day
% time level 2 TAR Hypoglycemia index sensor use is a reasonable target from which the available
> 250mgm% (13.9 mmol/L)
data can be used to derive estimations regarding average
Glycemic variability Hyperglycemia index
glucose and HbA1c.
Metric 2: Time in range (TIR) the ADA and European Association for the Study of
TIR is a key metric that has been identified as being per- Diabetes (EASD).
haps the most feasible therapeutic target in day‐to‐day
management of diabetes with use of CGM. It typically Metric 3: Glycemic variability
refers to glucose values between 70 – 180 mg/dL. Although Glycemic goals in diabetes care are to maintain normal glu-
the metrics listed in Table 8.1 address not only TIR, but cose patterns while avoiding hypoglycemia [64]. Table 8.2
also time above range (TAR) and time below range (TBR), lists 14 different measures describing glycemic variability.
it is the TIR metric that has been most heavily analyzed in While variability as a concept is intuitive, the calculation
recent studies. However, in clinical practice, achievement and interpretation of these measures have not been trans-
of TIR targets practically involves first a foremost an assess- parent or intuitive to clinicians and patients unfamiliar
ment of TBR in order to prevent serious hypoglycemia, fol- with the metrics. To date, they have not been universally
lowed then with a focus on reducing TAR. These used to provide guidance to patients or clinicians in their
approaches, in turn, lead to increased TIR. glycemic management [65].
Perhaps the most robust data for TIR in patients with Mathematical models attempting to quantify glycemic
DM1 comes from a post‐hoc analysis of the DCCT trial data variability have been focused on the statistical middle, and
in which 7‐point blood glucose profiles collected quarterly range of glycemic measures have used non‐parametric
were analyzed to develop Cox proportional hazard models analyses due to lack of normal distribution of the data [66].
for assessing the relationship between TIR and microvascu- The graphic presentation of trends over time have helped
lar outcomes. Indeed, the hazard ratios for retinopathy and
microalbuminuria progression increased by 64% and 30%, TABLE 8.2 Measures used to quantify glycemic variation
respectively, for each 10% reduction in TIR [36]. Similar, (modified from [64]).
although less robust, data has been collected for patients
Lability index mean +3 SD of the urinary glucose in gm/hr
with DM2 in an observational trial which demonstrated a
Mean glucose Arithmetic mean within a time frame
correlation between diabetic retinopathy severity, reduced Glucose index Low/High glucose index
TIR, and increased variability [61]. Risk index Blood glucose risk index
Observational studies using CGM metrics in ambula- Rate of change Blood glucose rate of change
tory patients have found that pre‐ and postprandial hyper- M value Stability of glucose compared to ideal
MAGE Each blood glucose decreases from peak to
glycemia and glycemic variation contribute equally to
nadir
HbA1c levels and presumed vascular risk [62]. Despite MODD or Across The absolute value of the difference
these observational data, however, secondary analysis of Day Variation between glucose values at the same time on
glycemic measures from DCCT/EDIC trials have not (ADV) consecutive days
shown an independent effect for cardiovascular risk J index J = 0.001 (MBG + SD)2 for BG
measurements in mg/dl, J = 0.324(MBG +
beyond mean glucose and HbA1c [63]. With the develop-
SD)2 for BG measurements in mmol/l.
ment of CGM and the extent of more complete capture of VGA Variable Grid Analysis (VGA) – minimum/
clinical glycemia, it is anticipated that “big data” analysis maximum plot over a certain period of
methods will provide quantification and familiarity with observation
predictive methods and tools that go beyond HbA1c. CONGA n or Continuous Overall Net Glycemic Action
Within Day (CONGA and CONGA n) or Within Day
In clinical practice, the ability to provide targeted
Variation (WDV) Variation (WDV) of the standard deviation of
goals for TAR, TIR, and TBR for patients is crucial in the difference between two time points
optimizing the data collected with CGM. Targets should (e.g. work, school, meals, other)
be safe, achievable, and individualized according to TIR Time In Range 4 to 12 mmol/l 7‐180 mgm%
patient‐specific factors. Age, comorbidities, and preg- TBR Time Below Range < 4 mmol/l < 70 mgm%
(TAR) Time Above Range > 12 mmol/l >
nancy will all affect target glycemic goals for TIR. Using
180 mgm%
CGM data, TIR recommendations are now endorsed by
with interpretation and identification of corrective action extensively in the clinical management of diabetes [65]. Mean
in decision making. For example: Case Study 1a is an of Daily Difference (MODD) or Across Day Variation
example of variability due to frequent reactive across day (MODD or ADV) and Continuous Overlapping Net Glycemic
insulin dosing (ADV) while Case Study 1b demonstrates Action (CONGAn) or Within Day Variation (CONGn or
the within day variability (WDV) associated with elevated WDV) can be used not only as a continued effort in the quan-
cortisol levels from Cushing’s Syndrome. Variability Grid titative assessment of within day and across day variation, but
Analysis (VGA) or Poincare Plot recognizes the impor- they graphically offer the opportunity for qualitative assess-
tance of assessing quantitative and qualitative patterns to ment. Most importantly, they can serve as clinical tools for
assist the clinician and patient in management decisions. review and discussion with individual patients and their man-
While glucose variation has been stated to have two com- agement behavior. For example, Figure 8.1a is a patient that
ponents, amplitude and timing, that are predictive of hyper dominantly uses sliding scale versus preventive insulin dose
and hypoglycemia, these parameters have been mostly adjustment. Figure 8.1b demonstrates a pattern of significant
directed towards hyperglycemia and have not been used postprandial hyperglycemia (WDV) related to steroids.
FIG 8.1A Glycemic variation with frequent insulin supplements. 40‐year‐old woman with Type 1 diabetes on MDI. She is measuring
pre and postprandial glucose levels and notes significant variation at the same time each day despite similar diet and activity. Her
reactive supplemental insulin should be considered as a cause of her Across Day Variation (ADV).
400
350
300
250
mg/dL
200
150
100
50
12:00 3:00 6:00 9:00 12:00 3:00 6:00 9:00 12:00
FIG 8.1B Glycemic variation due to within day variation (WDV) from steroids.
A 30‐year‐old male with type 2 diabetes on a single Summary

long‐acting daily insulin. He is undergoing an evaluation
The evidence for HbA1c as a measure of long‐term glyce-
for the question of Cushing’s Syndrome from an adrenal
mic control and development of vascular outcomes is
adenoma. He is observed to have within day variation
robust, and it remains a valuable metric in the manage-
(WDV) that includes a progressive rise in postprandial
ment of both DM1 and DM2. However, as diabetes tech-
reflectance meter glucoses with highest values at the end of
the day and progressive fall nocturnally that places him at nology has advanced with development of CGM and
risk for early morning hypoglycemia. closed‐loop systems of insulin delivery, the limitations of
Recognizing that there are many causes of glycemic HbA1c have become even more pronounced. CGM has
variation to include people and medication factors, the allowed for a real‐time view into actionable metrics that
Beyond A1c Working Group has offered an attempt to permit clinicians, patients, and researchers to tailor man-
quantitate a more predictive measure of glycemic control agement toward specific goals based on patient‐specific
they call the Comprehensive Glucose Pentagon and the factors. This focused approach using metrics such as TIR,
glycemic risk parameter (GRP). The GRP reflects the five glucose variability, and magnitude of excursions has led
main axes of glucose control: time out of range, coefficient to reductions in hyperglycemia while minimizing risks of
of variation, severity of hyperglycemia, severity of hypogly- hypoglycemia, both cornerstones of glycemic control.
cemia, and mean glucose [67]. HbA1c is not considered as Preliminary studies have demonstrated the role for met-
part of the main axes of their model, and validation will rics such as variability and TIR in reducing risk of at least
be necessary to determine whether it has a prognostic microvascular complications, and future studies are
value greater than HbA1c. Currently the authors provide needed to further identify their role in other diabetes‐
three case examples to describe its use. Sufficiently pow- related outcomes such as macrovascular disease. Thus,
ered long‐term study will be necessary in which assess- standardized CGM metrics are appropriate complements
ment of these measures maybe be possible due to the to the traditional measure of HbA1c and provide valuable
wide range of outcomes of information collected from tools for both patients and clinicians in the day‐to‐day
CGMs and pattern analysis. management of diabetes.
References 8. Zinman B, Marso SP, Christiansen E, Calanna S, Rasmussen

S, Buse JB, Investigators LEADER Publication Committee on
1. Diabetes C, Complications Trial Research G, Nathan DM, behalf of the LEADER Trial Investigators. Hypoglycemia,
Genuth S, Lachin J, Cleary P, Crofford O, Davis M, Rand L, cardiovascular outcomes, and death: the LEADER
Siebert C. The effect of intensive treatment of diabetes on the experience. Diabetes Care. 41(8):1783–1791. PubMed PMID:
development and progression of long‐term complications in 29903847.
insulin‐dependent diabetes mellitus. N Engl J Med. 9. Sheetz MJ, King GL. Molecular understanding of hyperglyce-
1993;329(14):977–986. Epub 1993/09/30. doi: 10.1056/ mia’s adverse effects for diabetic complications. JAMA.
NEJM199309303291401. PubMed PMID: 8366922. 2002;288(20):2579–2588. Epub 2002/11/28. doi: 10.1001/
2. Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, jama.288.20.2579. PubMed PMID: 12444865.
Motoyoshi S, Kojima Y, Furuyoshi N, Shichiri M. Intensive 10. Cohen R. Dysfunction of vascular endothelium in diabetes
insulin therapy prevents the progression of diabetic mellitus. Circulation. 1993;87(Suppl V):V67–V76.
microvascular complications in Japanese patients with non‐ 11. Risso A, Mercuri F, Quagliaro L, Damante G, Ceriello A.
insulin‐dependent diabetes‐mellitus – a randomized Intermittent high glucose enhances apoptosis in human
prospective 6‐year study. Diabetes Res Clin Pr. umbilical vein endothelial cells in culture. American Journal
1995;28(2):103–117. doi: 10.1016/0168‐8227(95)01064‐K. of Physiology – Endocrinology & Metabolism. 281(5):E924–
PubMed PMID: WOS:A1995RR63400004. 930. PubMed PMID: 11595647.
3. Mann JFE, Orsted DD, Brown‐Frandsen K, Marso SP, 12. Nishikawa T, Edelstein D, Du XL, Yamagishi S, Matsumura
Poulter NR, Rasmussen S, Tornoe K, Zinman B, Buse JB, T, Kaneda Y, Yorek MA, Beebe D, Oates PJ, Hammes HP,
LEADER Steering Committee and Investigators. Liraglutide Giardino I, Brownlee M. Normalizing mitochondrial super-
and renal outcomes in type 2 diabetes. N Engl J Med. oxide production blocks three pathways of hyperglycaemic
377(9):839–848. PubMed PMID: 28854085. damage. Nature. 404(6779):787–790. PubMed PMID:
4. Marso SP, Daniels GH, Brown‐Frandsen K, Kristensen P, Mann 10783895.
JF, Nauck MA, Nissen SE, Pocock S, Poulter NR, Ravn LS, 13. Piconi L, Quagliaro L, Assaloni R, Da Ros R, Maier A,
Steinberg WM, Stockner M, Zinman B, Bergenstal RM, Buse JB, Zuodar G, Ceriello A. Constant and intermittent high glu-
LEADER Steering Committee, LEADER Trial Investigators. cose enhances endothelial cell apoptosis through mitochon-
Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes. drial superoxide overproduction. Diabetes/Metabolism
N Engl J Med. 375(4):311–322. PubMed PMID: 27295427. Research Reviews. 22(3):198–203. PubMed PMID: 16453381.
5. Tamborlane WV, Barrientos‐Perez M, Fainberg U, Frimer‐ 14. Quagliaro L, Piconi L, Assaloni R, Martinelli L, Motz E,
Larsen H, Hafez M, Hale PM, Jalaludin MY, Kovarenko M, Ceriello A. Intermittent high glucose enhances apoptosis
Libman I, Lynch JL, Rao P, Shehadeh N, Turan S, Weghuber related to oxidative stress in human umbilical vein endothe-
D, Barrett T, Ellipse Trial Investigators. Liraglutide in lial cells: the role of protein kinase C and NAD(P)H‐oxidase
children and adolescents with type 2 diabetes. N Engl J Med. activation. Diabetes. 52(11):2795–2804. PubMed PMID:
381(7):637–646. PubMed PMID: 31034184. 14578299.
6. Verma S, Bhatt DL, Bain SC, Buse JB, Mann JFE, Marso SP, 15. John WG, Mosca A, Weykamp C, Goodall I. HbA1c stand-
Nauck MA, Poulter NR, Pratley RE, Zinman B, Michelsen ardisation: history, science and politics. Clin Biochem Rev.
MM, Monk Fries T, Rasmussen S, Leiter LA, LEADER 2007;28(4):163–168. Epub 2008/04/09. PubMed PMID:
Publication Committee on behalf of the LEADER Trial 18392123; PMCID. PMC2282401.
Investigators. Effect of liraglutide on cardiovascular events in 16. Rahbar S. An abnormal hemoglobin in red cells of diabetics.
patients with type 2 diabetes mellitus and polyvascular Clin Chim Acta. 1968;22(2):296–298. Epub 1968/10/01.
disease: results of the LEADER trial. Circulation. PubMed PMID: 5687098.
137(20):2179–2183. PubMed PMID: 29760228. 17. Smith SA. HbA1c: Is it the most important therapeutic tar-
7. Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, get in outpatient management of diabetes? In: Vella A and
Hantel S, Mattheus M, Devins T, Johansen OE, Woerle HJ, Rizza RA (eds) Clinical Dilemmas in Diabetes. Blackwell
Broedl UC, Inzucchi SE for the EPMPA‐REG OUTCOME Publishing; 2011.
Investigators. Empagliflozin, cardiovascular outcomes, and 18. American Diabetes A. Standards of medical care in diabetes –
mortality in type 2 diabetes. N Engl J Med. 2015;Nov 2008. Diabetes Care. 2008;31(Suppl 1):S12–54. Epub 2008/
26;373(22):2117–2128 01/10. doi: 10.2337/dc08‐S012. PubMed PMID: 18165335.
19. Trivelli LA, Ranney HM, Lai HT. Hemoglobin components 430. doi: 10.1016/S0140‐6736(10)60576–60574. PubMed
in patients with diabetes mellitus. N Engl J Med. PMID: WOS:000281012900027.
1971;284(7):353–357. Epub 1971/02/18. doi: 10.1056/ 27. Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N,
NEJM197102182840703. PubMed PMID: 5539916. Reaven PD, Zieve FJ, Marks J, Davis SN, Hayward R, Warren
20. Diabetes Control and Complications Trial/Epidemiology of SR, Goldman S, McCarren M, Vitek ME, Henderson WG,
Diabetes Interventions and Complications Research Group, Huang GD, VADT Investigators. Glucose control and vascu-
Lachin JM, Genuth S, Cleary P, Davis MD, Nathan DM. lar complications in veterans with type 2 diabetes. New Eng J
Retinopathy and nephropathy in patients with type 1 diabe- Med. 2009;360(2):129–39. doi: 10.1056/NEJMoa0808431.
tes four years after a trial of intensive therapy. N Engl J Med. PubMed PMID: WOS:000262258300006.
2000;342(6):381–389. Epub 2000/02/10. doi: 10.1056/ 28. Sacks DB, Bruns DE, Goldstein DE, Maclaren NK,
NEJM200002103420603. PubMed PMID: 10666428; McDonald JM, Parrott M. Guidelines and recommendations
PMCID. PMC2630213. for laboratory analysis in the diagnosis and management of
21. Effect of intensive blood‐glucose control with metformin on diabetes mellitus. Clin Chem. 2002;48(3):436–472. Epub
complications in overweight patients with type 2 diabetes 2002/02/28. PubMed PMID: 11861436.
(UKPDS 34). UK Prospective Diabetes Study (UKPDS) 29. Beck RW, Connor CG, Mullen DM, Wesley DM, Bergenstal
Group. Lancet. 1998;352(9131):854–865. Epub 1998/09/22. RM. The fallacy of average: how using HbA1c alone to assess
PubMed PMID: 9742977. glycemic control can be misleading. Diabetes Care.
22. Intensive blood‐glucose control with sulphonylureas or 2017;40(8):994–999. Epub 2017/07/25. doi: 10.2337/
insulin compared with conventional treatment and risk of dc17‐0636. PubMed PMID: 28733374; PMCID.
complications in patients with type 2 diabetes (UKPDS 33). PMC5521971.
UK Prospective Diabetes Study (UKPDS) Group. Lancet. 30. Nathan DM, Singer DE, Hurxthal K, Goodson JD. The
1998;352(9131):837–853. Epub 1998/09/22. PubMed PMID: clinical information value of the glycosylated hemoglobin
9742976. assay. N Engl J Med. 1984;310(6):341–346. Epub 1984/
23. Reichard P. Are there any glycemic thresholds for the serious 02/09. doi: 10.1056/NEJM198402093100602. PubMed PMID:
microvascular diabetic complications. J Diabetes Complicat. 6690962.
1995;9(1):25–30. doi: 10.1016/1056‐8727(94)00008‐C. 31. Goldstein DE, Little RR, Lorenz RA, Malone JI, Nathan D,
PubMed PMID: WOS:A1995RK84900004. Peterson CM. Tests of glycemia in diabetes. Diabetes Care.
24. International Expert Committee. International Expert 1995;18(6):896–909. Epub 1995/06/01. doi: 10.2337/diac-
Committee report on the role of the A1C assay in the diag- are.18.6.896. PubMed PMID: 7555528.
nosis of diabetes. Diabetes Care. 2009;32(7):1327–1334. 32. Kishimoto M, Yamasaki Y, Kubota M, Arai K, Morishima T,
Epub 2009/06/09. doi: 10.2337/dc09–9033. PubMed PMID: Kawamori R, Kamada T. 1,5‐Anhydro‐D‐glucitol evaluates
19502545; PMCID. PMC2699715. daily glycemic excursions in well‐controlled NIDDM.
25. Patel A, MacMahon S, Chalmers J, Neal B, Billot L, Diabetes Care. 1995;18(8):1156–1159. Epub 1995/08/01. doi:
Woodward M, Marre M, Cooper M, Glasziou P, Grobbee D, 10.2337/diacare.18.8.1156. PubMed PMID: 7587851.
Hamet P, Harrap S, Heller S, Liu LS, Mancia G, Mogensen 33. Judd SL, Sonksen PH. Teaching diabetics about self‐man-
CE, Pan CY, Poulter N, Rodgers A, Williams B, Bompoint S, agement. Diabetes Care. 1980;3(1):134–139.
de Galan BE, Joshi R, Travert F, Grp AC. Intensive blood glu- 34. Sonksen PH, Judd SL, Lowy C. Home monitoring of blood‐
cose control and vascular outcomes in patients with type 2 glucose. Method for improving diabetic control. Lancet.
diabetes. New Eng J Med. 2008;358(24):2560–2572. PubMed 1978;1(8067):729–732. PubMed PMID: 76745.
PMID: WOS:000256593600004. 35. American Diabetes A. 6. Glycemic targets: standards of medi-
26. Ismail‐Beigi F, Craven T, Banerji MA, Basile J, Calles J, cal care in diabetes – 2019. Diabetes Care. 2019;42(Suppl 1):
Cohen RM, Cuddihy R, Cushman WC, Genuth S, Grimm S61–S70. Epub 2018/12/19. doi: 10.2337/dc19‐S006. PubMed
RH, Hamilton BP, Hoogwerf B, Karl D, Katz L, Krikorian A, PMID: 30559232.
O’Connor P, Pop‐Busui R, Schubart U, Simmons D, Taylor 36. Beck RW, Bergenstal RM, Riddlesworth TD, Kollman C, Li Z,
H, Thomas A, Weiss D, Hramiak I, for the ACCORD trial Brown AS, Close KL. Validation of time in range as an out-
group. Effect of intensive treatment of hyperglycaemia on come measure for diabetes clinical trials. Diabetes Care.
microvascular outcomes in type 2 diabetes: an analysis of the 2019;42(3):400–405. Epub 2018/10/26. doi: 10.2337/dc18‐
ACCORD randomised trial. Lancet. 2010;376(9739):419– 1444. PubMed PMID: 30352896.
37. Miller KM, Beck RW, Bergenstal RM, Goland RS, Haller MJ, 45. Chiasson JL, Gomis R, Hanefeld M, Josse RG, Karasik A,
McGill JB, Rodriguez H, Simmons JH, Hirsch IB, T1D Laakso M. The STOP‐NIDDM Trial: an international study
Exchange Clinic Network. Evidence of a strong association on the efficacy of an alpha‐glucosidase inhibitor to prevent
between frequency of self‐monitoring of blood glucose and type 2 diabetes in a population with impaired glucose toler-
hemoglobin A(1c) levels in T1D Exchange Clinic Registry ance: rationale, design, and preliminary screening data.
participants. Diabetes Care. 2013;36(7):2009–2014. doi: Study to prevent non‐insulin‐dependent diabetes mellitus.
10.2337/dc12‐1770. PubMed PMID: WOS:000321472700034. Diabetes Care. 21(10):1720–1725. PubMed PMID: 9773737.
38. Schutt M, Kern W, Krause U, Busch P, Dapp A, Grziwotz R, 46. Cavalot F, Pagliarino A, Valle M, Di Martino L, Bonomo K,
Mayer I, Rosenbauer J, Wagner C, Zimmermann A, Kerner Massucco P, Anfossi G, Trovati M. Postprandial blood glu-
W, Holl RW, DPV Initiative. Is the frequency of self‐moni- cose predicts cardiovascular events and all‐cause mortality
toring of blood glucose related to long‐term metabolic con- in type 2 diabetes in a 14‐year follow‐up: lessons from the
trol? Multicenter analysis including 24,500 patients from San Luigi Gonzaga Diabetes Study. Diabetes Care.
191 centers in Germany and Austria. Exp Clin Endocrinol 34(10):2237–2243. PubMed PMID: 21949221.
Diabetes. 2006;114(7):384–388. Epub 2006/08/18. doi: 10. 47. Hopper I, Billah B, Skiba M, Krum H. Prevention of diabetes
1055/s‐2006‐924152. PubMed PMID: 16915542. and reduction in major cardiovascular events in studies of
39. Brownlee M. The pathobiology of diabetic complications: a subjects with prediabetes: meta‐analysis of randomised con-
unifying mechanism. Diabetes. 54(6):1615–1625. PubMed trolled clinical trials. European Journal of Cardiovascular
PMID: 15919781. Prevention & Rehabilitation. 18(6):813–823. PubMed PMID:
40. Hirsch IB, Brownlee M. Should minimal blood glucose vari- 21878448.
ability become the gold standard of glycemic control? 48. Davidson MB. Counterpoint: postprandial glucose levels are
Journal of Diabetes & its Complications. 19(3):178–181. Pub not a clinically important treatment target. Diabetes Care.
Med PMID: 15866065. 2010;33:1908–1910.
41. Ceriello A. Postprandial hyperglycemia and diabetes com- 49. Liebl A, Henrichs HR, Heinemann L, Freckmann G,
plications: is it time to treat? Diabetes. 54(1):1–7. PubMed Biermann E, Thomas A, Continuous Glucose Monitoring
PMID: 15616004. Working Group of the Working Group Diabetes Technology
42. Ceriello A. Point: postprandial glucose levels are a clinically of the German Diabetes A. Continuous glucose monitoring:
important treatment target. Diabetes Care. 2010;33(8): evidence and consensus statement for clinical use. J Diabetes
1905–1907. Sci Technol. 2013;7(2):500–519. Epub 2013/04/10. doi:
43. Meier JJ, Baller B, Menge BA, Gallwitz B, Schmidt WE, 10.1177/193229681300700227. PubMed PMID: 23567009;
Nauck MA. Excess glycaemic excursions after an oral glu- PMCID. PMC3737652.
cose tolerance test compared with a mixed meal challenge 50. Beck RW, Riddlesworth T, Ruedy K, Ahmann A, Bergenstal
and self‐measured home glucose profiles: is the OGTT a R, Haller S, Kollman C, Kruger D, McGill JB, Polonsky W,
valid predictor of postprandial hyperglycaemia and vice Toschi E, Wolpert H, Price D, DIAMOND Study Group.
versa? Diabetes, Obesity & Metabolism. 11(3):213–222. Pub Effect of continuous glucose monitoring on glycemic control
Med PMID: 18564177. in adults with type 1 diabetes using insulin injections. The
44. NAVIGATOR Study Group, Holman RR, Haffner SM, DIAMOND Randomized Clinical Trial. JAMA.
McMurray JJ, Bethel MA, Holzhauer B, Hua TA, Belenkov Y, 2017;317(4):371–378. doi: 10.1001/jama.2016.19975.
Boolell M, Buse JB, Buckley BM, Chacra AR, Chiang FT, PubMed PMID: WOS:000392509600019.
Charbonnel B, Chow CC, Davies MJ, Deedwania P, Diem P, 51. Battelino T, Conget I, Olsen B, Schutz‐Fuhrmann I, Hommel
Einhorn D, Fonseca V, Fulcher GR, Gaciong Z, Gaztambide E, Hoogma R, Schierloh U, Sulli N, Bolinder J, SWITCH
S, Giles T, Horton E, Ilkova H, Jenssen T, Kahn SE, Krum H, Study Group. The use and efficacy of continuous glucose
Laakso M, Leiter LA, Levitt NS, Mareev V, Martinez F, monitoring in type 1 diabetes treated with insulin pump
Masson C, Mazzone T, Meaney E, Nesto R, Pan C, Prager R, therapy: a randomised controlled trial. Diabetologia.
Raptis SA, Rutten GE, Sandstroem H, Schaper F, Scheen A, 2012;55(12):3155–3162. doi: 10.1007/s00125‐012‐2708‐9.
Schmitz O, Sinay I, Soska V, Stender S, Tamas G, Tognoni G, PubMed PMID: WOS:000310381800003.
Tuomilehto J, Villamil AS, Vozar J, Califf RM. Effect of nat- 52. Juvenile Diabetes Research Foundation Continuous Glucose
eglinide on the incidence of diabetes and cardiovascular Monitoring Study G, Beck RW, Hirsch IB, Laffel L,
events. New Eng J Med. 362(16):1463–1476 Tamborlane WV, Bode BW, Buckingham B, Chase P,
Clemons R, Fiallo‐Scharer R, Fox LA, Gilliam LK, Huang New Eng J Med. 2008;359(14):1464–U65. PubMed PMID:
ES, Kollman C, Kowalski AJ, Lawrence JM, Lee J, Mauras N, WOS:000259631700007.
O’Grady M, Ruedy KJ, Tansey M, Tsalikian E, Weinzimer 59. Xing D, Kollman C, Beck RW, Tamborlane WV, Laffel L,
SA, Wilson DM, Wolpert H, Wysocki T, Xing D. The effect Buckingham BA, Wilson DM, Weinzimer S, Fiallo‐Scharer
of continuous glucose monitoring in well‐controlled type 1 R, Ruedy KJ, Juvenile Diabetes Research Foundation
diabetes. Diabetes Care. 2009;32(8):1378–1383. Epub 2009/ Continuous Glucose Monitoring Study Group. Optimal
05/12. doi: 10.2337/dc09‐0108. PubMed PMID: 19429875; sampling intervals to assess long‐term glycemic control
PMCID. PMC2713649. using continuous glucose monitoring. Diabetes Technol
53. Bergenstal RM, Tamborlane WV, Ahmann A, Buse JB, Ther. 2011;13(3):351–358. Epub 2011/02/09. doi: 10.1089/
Dailey G, Davis SN, Joyce C, Peoples T, Perkins BA, Welsh dia.2010.0156. PubMed PMID: 21299401; PMCID.
JB, Willi SM, Wood MA, STAR3 Study Group. Effectiveness PMC6468940.
of sensor‐augmented insulin‐pump therapy in type 1 diabe- 60. Riddlesworth TD, Beck RW, Gal RL, Connor CG, Bergenstal
tes. N Engl J Med. 2010;363(4):311–320. Epub 2010/07/01. RM, Lee S, Willi SM. Optimal Sampling duration for continu-
doi: 10.1056/NEJMoa1002853. PubMed PMID: 20587585. ous glucose monitoring to determine long‐term glycemic con-
54. Bergenstal RM, Klonoff DC, Garg SK, Bode BW, Meredith trol. Diabetes Technol Ther. 2018;20(4):314–316. Epub 2018/
M, Slover RH, Ahmann AJ, Welsh JB, Lee SW, Kaufman FR, 03/23. doi: 10.1089/dia.2017.0455. PubMed PMID: 29565197.
ASPIRE In‐Home Study Group. Threshold‐based insulin‐ 61. Lu J, Ma X, Zhou J, Zhang L, Mo Y, Ying L, Lu W, Zhu W, Bao
pump interruption for reduction of hypoglycemia. New Eng Y, Vigersky RA, Jia W. Association of time in range, as
J Med. 2013;369(3):224–232. doi: 10.1056/NEJMoa1303576. assessed by continuous glucose monitoring, with diabetic
PubMed PMID: WOS:000321945800008. retinopathy in type 2 diabetes. Diabetes Care. 2018;41(11):
55. Danne T, Kordonouri O, Holder M, Haberland H, 2370–2376. Epub 2018/09/12. doi: 10.2337/dc18‐1131. Pub
Golembowski S, Remus K, Blasig S, Wadien T, Zierow S, Med PMID: 30201847.
Hartmann R, Thomas A. Prevention of hypoglycemia by 62. Faerch K, Alssema M, Mela DJ, Borg R, Vistisen D. Relative
using low glucose suspend function in sensor‐augmented contributions of preprandial and postprandial glucose expo-
pump therapy. Diabetes Technol Ther. 2011;13(11):1129–1134. sures, glycemic variability, and non‐glycemic factors to
Epub 2011/08/11. doi: 10.1089/dia.2011.0084. PubMed PMID: HbA1c in individuals with and without diabetes. Nutrition &
21827318. Diabetes. 2018;8(1):38. PubMed PMID: 29855488.
56. Danne T, Nimri R, Battelino T, Bergenstal RM, Close KL, 63. Siegelaar SE, Kerr L, Jacober SJ, Devries JH. A decrease in
DeVries JH, Garg S, Heinemann L, Hirsch I, Amiel SA, Beck glucose variability does not reduce cardiovascular event
R, Bosi E, Buckingham B, Cobelli C, Dassau E, Doyle FJ, 3rd, rates in type 2 diabetic patients after acute myocardial
Heller S, Hovorka R, Jia W, Jones T, Kordonouri O, Kovatchev infarction: a reanalysis of the HEART2D study. Diabetes
B, Kowalski A, Laffel L, Maahs D, Murphy HR, Norgaard K, Care. 2011;34(4):855–857. PubMed PMID: 21447661.
Parkin CG, Renard E, Saboo B, Scharf M, Tamborlane WV, 64. McDonnell CM, Donath SM, Vidmar SI, Werther GA,
Weinzimer SA, Phillip M. International consensus on use of Cameron FJ. A novel approach to continuous glucose analy-
continuous glucose monitoring. Diabetes Care. 2017;40(12): sis utilizing glycemic variation. Diabetes Technology &
1631–1640. Epub 2017/11/23. doi: 10.2337/dc17‐1600. Pub Therapeutics. 2005;7(2):253–263. PubMed PMID: 15857227.
Med PMID: 29162583; PMCID. PMC6467165. 65. Cobelli C, Facchinetti A. Yet another glucose variability
57. Vigersky RA. Going beyond HbA1c to understand the ben- index: time for a paradigm change? Diabetes Technology &
efits of advanced diabetes therapies. Journal of Diabetes. 2019; Therapeutics. 2018;20(1):1–3. PubMed PMID: 29320256.
11(1):23–31. doi: 10.1111/1753‐0407.12846. PubMed PMID: 66. Zhou T, Knopp JL, Chase G. The state of variability: a vision
WOS:000451868200005. for descriptors of glycaemia. Annual Reviews in Control.
58. Tamborlane WV, Beck RW, Bode BW, Buckingham B, Chase 2019;48:472–484. doi.org/10.1016/j.arcontrol.2019.06.004.
HP, Clemons R, Fiallo‐Scharer R, Fox LA, Gilliam LK, 67. Thomas A, Heinemann L. Prediction of the risk to develop
Hirsch IB, Huang ES, Kollman C, Kowalski AJ, Laffel L, diabetes‐related late complications by means of the glucose
Lawrence JM, Lee J, Mauras N, O’Grady M, Ruedy KJ, pentagon model: analysis of data from the Juvenile Diabetes
Tansey M, Tsalikian E, Weinzimer S, Wilson DM, Wolpert Research Foundation continuous glucose monitoring study.
H, Wysocki T, Xing DY, Continuous JDRF. continuous glu- Journal of Diabetes Science & Technology. 2012;6(3):572–580.
cose monitoring and intensive treatment of type 1 diabetes. PubMed PMID: 22768888.
9 Technology Issues: Continuous Glucose
Monitoring, Insulin Pumps, and Closed
Loop Control for Patients with Diabetes
Ravinder Jeet Kaur, Shafaq R. Rizvi and Yogish C. Kudva
Division of Endocrinology, Diabetes, Metabolism and Nutrition, Mayo Clinic, Rochester, MN
delivery. Based on deficiency in insulin secretion, three

stages occur in the natural history of T1D, normal glucose
●● There is enough evidence to indicate that optimal
in stage 1, dysglycemia (hyperglycemia) diagnosed with
management of type 1 diabetes requires a personal sophisticated laboratory testing in stage 2 and symptomatic
continuous glucose monitor. hyperglycemia in stage 3 [3, 4]. For stage 2 disease onwards,
●● The best current treatment for patients with type 1 insulin therapy can be delivered by multiple daily injec-
diabetes and severe insulin deficiency is a closed loop tions (MDI), continuous subcutaneous insulin infusion
glucose insulin control (CLC) system.
●● Currently, the closed loop system has only been tested
(CSII) or insulin pump with insulin dosing adjusted fre-
and approved in patients with severe or stage III type 1 quently based on glucose patterns. T1DM patients have
diabetes. used MDI (basal-bolus) for more than 40 years to ade-
●● So far CLC has not been tested in stage I and II type quately maintain their blood glucose (BG) guided by self-
1diabetes. monitoring of blood glucose (SMBG) four or more times
●● CLC is being tested in specific subgroups and contexts
such as pregnancy with type 1 diabetes at present.
daily (before meals and at bedtime) using capillary blood
●● CLC is expensive and may only be available in certain sampling. The association of such management with fre-
countries at this time. quent hypoglycemia, hypoglycemia unawareness and
●● Technology use for type 2 diabetes mellitus has to be severe hypoglycemia accelerated the development of con-
studied further. tinuous glucose monitors (CGM). Insulin pumps to deliver
insulin continuously were developed in the 1970s. The
Introduction
maturation of these device technologies and the develop-
Type 1 diabetes mellitus (T1DM) is a disease characterized ment of control algorithms that could automatically adjust
by immune- mediated destruction of insulin producing insulin delivery have led to the testing and approval of
pancreatic β-cells leading to insulin deficiency and hyper- Closed loop control (CLC) in the last decade [5]. Recently,
glycemia [1, 2], thus requiring lifelong insulin therapy to international committees have developed consensus about
achieve glucose control. Despite complex insulin therapy, managing diabetes to achieve maximal time in target (TIR)
T1DM is characterized by high glucose variability (GV). on CGM. The consensus opinion is that CGM is used to
GV occurs due to severe deficiency in insulin secretion and guide optimization of therapy to increase TIR (70–180mg/
the difficulty of replacing insulin without frequent glucose dL) and minimize hypoglycemia (< 70mg/dL) and hyper-
measurements that enable prompt change in insulin glycemia (> 250mg/dL) in patients with T1DM.

115
Type 2 diabetes mellitus (T2DM) is a disease with differ- At 26-weeks of study, reduction in HbA1c with HbA1c <
ent pathophysiology and therefore the role of technology in 7.0%, occurred in a significantly larger proportion in the
the management of the disease is different. Also, technol- younger group (8–14 yr, 27 vs 12%) and adults (34 vs 9%).
ogy testing for its role in the treatment of the disease is at Another RCT conducted in the USA with 158 T1D sub-
an early stage and therefore the amount of information jects on MDI and HbA1c 7.5–9.9% randomized subjects
provided in this chapter regarding the role of technology in 2:1 to CGM (Dexcom G4, n=105) and usual care (controls,
the treatment of T2DM is limited. 53) [7]. During 24 weeks of the study, a greater decrease in
We restricted our content to ambulatory care and have HbA1c level was seen with increased TIR, decreased time
not discussed inpatient use of technologies for diabetes below 70mg/dL decreased glucose variability and high
management. degree of participant satisfaction for CGM use (Table 9.1).
A similar study performed in 161 adults with T1D using
Continuous Glucose Monitoring (CGM) MDI with HbA1c of at least 7.5% in Sweden also showed
for Diabetes Care reduction in HbA1c (7.92% vs 8.35%) with CGM use
(Dexcom G4 PLATINUM) compared to usual care [8].
Currently, CGM provides real-time data every 5–15 min-
utes depending on the device. CGM allows for real-time Impact in a Higher Risk Population: Patients
assessment of glucose concentrations and trends by meas- with T1D and Impaired Hypoglycemia
uring interstitial glucose concentrations. Interstitial glu- Awareness
cose concentrations are usually close to blood glucose In 2018, another RCT (n=149), was conducted using CGM
concentrations when glucose concentrations are stable and (Libre, n=75) in T1D adults with impaired hypoglycemia
lag by just a few minutes, but this difference is greater dur- awareness or severe hypoglycemia treated with MDI
ing rapid changes in blood glucose such as after food intake (HypoDE) to observe the effect on clinical hypoglycemic
or during exercise. Current CGM systems may provide events (glucose ≤3.0 mmol/L for ≥20 min) compared to the
data on demand (Libre) or continuously (all other sys- control group (n=74) [9]. Significant reductions in mean
tems). These systems are placed subcutaneously and hypoglycemic events (glucose values of 3.0 mmol/l
replaced every 7–14 days or may be implanted for (≤54 mg/dL) or lower for at least 20 min, preceded by a
3–6 months (Eversense). The subcutaneous systems may minimum of 30 min with glucose concentration greater
be factory-calibrated or calibrated daily. than 3.0 mmol/l (> 54 mg/dL)) were observed in rtCGM
CGM data can be used to prevent and protect against group (10.8 vs 3.5 events from baseline to follow-up phase)
extremes of glucose control and improve glucose control compared with SMBG group (14.4 vs 13.7 events).
either with MDI or CSII or with a CLC system. The alerts Additionally, rtCGM use decreased glycemic variability
associated with hypo-and hyperglycemia prevent and with slight improvement in HbA1c (Table 9.1).
decrease severe hypo-and hyperglycemia. Efficacy of CGM in the Elderly: The WISDM
Study
Impact of CGM in Patients on MDI or CSII Most recently, a RCT was performed on 203 older adults at
The impact of CGM on diabetes management has been least 60 years of age with T1DM [10]. Participants were
observed in both T1D and T2D and in select populations randomly assigned in a 1:1 ratio to use CGM (103) and
with diabetes. Real-time CGM has been available since standard BGM (100). Results from this study showed a
2005 and the major CGM systems were used in a multiple- small but significant improvement in hypoglycemia (<
center randomized control trail (RCT) sponsored by the 70mg/dL) over 6 months in the CGM group (Table 9.1)
JDRF, called the JDRF CGM trial and published in 2008 [6]. along with decrease in mean HbA1c (7.6% at baseline to
This trial showed reduction in HbA1c and hypoglycemia 7.2% at 6 months). There was also significant improvement
in adults (≥25 years of age) from baseline to 26 weeks of the in TIR (56% to 63%) along with reduction in time in hyper-
study with the benefit strongly related to age (Table 9.1) [6]. glycemia with CGM use.
Closed Loop Systems in Type 1 Diabetes 117
TABLE 9.1 Continuous glucose monitoring (CGM) impact in diabetes (randomized control trials)
Hypoglycemia
Number (≤70mg/dL)
of Insulin (control vs
Studies subjects delivery CGM Groups Control CGM group) HbA1C
JDRF [6] 322 Pump ad MDI Dexcom 8–14 yr 8–14 yr Decreased (9% Decreased
users Seven, 15–24 yr 15–24 yr vs 34%) (age (0.02±0.45% vs
(Predominately Medtronic ≥25 yr ≥25 yr group ≥25 yr) −0.50±0.56%)
pump) Paradigm, (CGM (SMBG users (age group ≥25
Abbott users=165) =157) yr)
FreeStyle
Navigator
WISDM [10] 203 Pump and MDI Dexcom CGM BGM Decreased Decreased
G5 (N=103) (N=100) (BGM: 4.9% vs (BGM: 7.4% vs
CGM –2.7%) CGM – 7.2%)
Baseline: Baseline:
5.1% – CGM, 7.65 – CGM,
4.7% – BGM 7.5% – BGM
HypoDE [9] 149 100%MDI Dexcom rtCGM SMBG Decreased Improved
G5 mobile (N=75) (N=74) (6.4% vs 1.6%) Slightly but not
system (N=66 in significant
follow up (from 7.4 to
phase) 7.3% (rtCGM)
vs from 7.6 to
7.4%(SMBG))
DIAMOND [7] 158 100% MDI Dexcom CGM users Usual care with Decreased Decreased
G4 with with MDI MDI (home (minutes per (8.2 vs 7.7%)
software blood glucose day < 70mg/
505 monitoring at dL – median of
least 4 times 80 vs 43)
daily – Bayer
Contour Next
USB meter)
BGM – Blood glucose monitoring, MDI – Multiple daily injections, CGM – Continuous glucose monitoring.
CGM in Pregnancy of 11.1% for reference samples > 4.2mmol/L (75mg/dL). As

CGM has been tested in special contexts such as pregnancy expected, performance in the hypoglycemic range (≤
with preexisting T1D (Medtronic iPRO2) and showed 75mg/dL) was less accurate, with a MARD of 21.7 % com-
reduction in HbA1c, improved TIR, less time in hypergly- pared to 11.6 % when glucose was 40–400mg/dL, and
cemia, and improved neonatal outcomes [11]. HbA1c significantly improved from 7.5% to 7.2%. CGM
accuracy significantly decreased in the last month of use,
Implantable CGM but overall showed safety and accurate performance over
A multinational, multicenter pivotal trial was performed the full sensor life. The implantable Eversense XL CGM
on 71 participants with T1DM and T2DM was conducted was approved in Europe (180-day version) in 2017 and in
with the Eversense CGM (implanted under the skin for the USA (90-day version) in 2018.
180 days) [12]. Results showed a median sensor lifetime of CGM has improved greatly over the past decade with
149 days, overall mean absolute relative difference (MARD) longer sensor life (7 days – 6 months), factory calibration
(Dexcom G6 and Abbott FreeStyle Libre) or twice a day in patients with stage III T1DM and are in commercial use.
calibration (Medtronic Guardian Sensor and Senseonics These systems have predictive low-glucose insulin-suspend
Eversense), improved alarms and alerts, easy insertion and functionality and increase basal insulin delivery rates using
ability to share glucose values in real time. This has an algorithm in response to predicted hyperglycemia or
improved quality of life [13] immensely in T1D patients hyperglycemia signals received from real-time CGM. They
with a decrease in hypoglycemia (< 70mg/dL) [6, 7, 10] and are called hybrid because of patient engagement at meal-
improvement in A1c [13]. Over the last decade, CGM use time that is similar to SAP therapy or CSII.
increased from 7% (2010–2012) to 30% (2016–2018) as per
the T1D Exchange registry [14], especially in children < 12 MiniMed 670G
years old (> 10 fold) [12]. MiniMed 670G (Medtronic, Northridge, CA, USA) was the
The evidence base for CGM use in T2DM is significantly first hybrid CLC system available in the USA after 3 months
less. Studies in T2DM using older CGM devices have been of a single-arm safety trial completed on 124 T1DM par-
published [15]. Personal or intermittent clinic-requested ticipants [17]. Patients had to periodically calibrate CGM
RT CGM has been shown to improve A1c by 0.3 % (meta- in this study and dose mealtime insulin using SMBG. The
analysis of 4 RCT) with benefits to diet, lifestyle, and ther- Control Algorithm utilized is the Proportional Integrated
apy adjustment [16]. Although the use of CGM in T2DM Derivative (PID). Participants spent around 87.2% time in
would be intuitively beneficial, more investigation is closed loop mode. Patients improved A1c and every CGM
needed. metric (Table 9.2). There were no episodes of severe hypo-
glycemia and DKA during this study. The lack of a control
group is a major limitation of the study. The approval of the
Currently Approved CLC for T1DM
device was contingent on the conduct of an RCT compar-
The optimal current treatment for patients with T1DM and ing hybrid-closed loop to three different treatments: MDI,
severe insulin deficiency is a CLC system. So far CLC has CSII, and SAP. This trial is scheduled to be completed by
not been tested in clinical stages 1 and 2 of T1DM. December 2021 (NCT02748018).
Currently, two hybrid CLC systems are approved for use in Most recently, a 6-month International Diabetes Closed
the USA by the FDA. Both CLC were tested and approved loop (iDCL) RCT enrolled 168 T1DM patients at seven US
TABLE 9.2 Pivotal studies of closed loop system.
Baseline versus hybrid closed loop
Type of > 180 > 250 > 300

Studies study Geography T1R < 70mg/dL mg/dL mg/dL mg/dL HbA1c (%)
Medtronic Non RCT USA, Israel 66.7% vs 5.9% vs 3.3% 27.4% vs N/A 2.3% vs 7.4% vs
670G [17] (n=124) 72.2% 24.5% 1.7% 6.9%
Tandem RCT USA 61% vs 71% 3.6% vs 1.6% 36% vs N/A N/A 7.4% vs
Control (n=168) (HCL) 2.8% vs 2.2% 27% 7.1%
IQ [18] 59% vs 59% 38% vs 7.4% vs
(control) 38% 7.4%
FLAIR RCT USA, 57% vs 67% 2.3% vs 2.1% 41% vs 13% vs N/A 7.9% vs
Trail [19] (n=113) Germany, (AHCL) 2.3% vs 2.1% 31% 9% 7.4%
Israel, 575 vs 63% 41% vs 13% vs 7.9% vs
Slovenia (670G) 34% 10% 7.6%
Medtronic RCT USA 68.8% vs 3.3% vs 2.3% 27.9% vs 6.2% vs N/A 7.5% vs
780G [20] (n=157) 74.5% 23.1% 4.6% 7.0%
TIR – Time in Range (70–180mg/dL), glycated hemoglobin (HbA1c) %.

centers to compare safety and efficacy of an advanced 100mg/dL target and 2–3 hours of AIT, TIR increased to
hybrid closed-loop system (Control IQ, Tandem Diabetes 76% (74% daytime and 84% overnight) and even more
Care) SAP with both groups using Dexcom G6 CGM [18]. (79%) with 2hours of AIT (76% daytime and 87% over-
The algorithm used in this system was a Model Predictive night). Data from more challenging patients including
Control (MPC) with hypoglycemia safety module, auto- those with less controlled diabetes and a younger popula-
mated correction boluses, and overnight intensification of tion from New Zealand also showed improved outcomes
basal insulin delivery to keep glucose within normal lim- compared to predictive low- glucose management algo-
its – the algorithm was constrained by a fixed active insulin rithm [21]. There was overall improvement in TIR and
time. All CGM metrics improved in the CLC cohort decrease in time below range with participants staying in
(Table 9.2). No serious hypoglycemia events occurred in closed loop 95% of the time.
either group. One episode of diabetic ketoacidosis occurred T:slim X2 pump users in Canada will have the option to
in the CLC group because of a pump infusion set failure. add Control-IQ technology to their existing pump begin-
ning in March 2021 via a remote software update and the
FLAIR trial 780G has received approval in select European countries
Recently, the Medtronic MiniMedTM 780G Advanced
Hybrid Closed Loop (AHCL) automated insulin delivery
Post-marketing Studies
system was compared to 670G in adolescents and young
adults [19]. This RCT was associated with better glucose
control and patient satisfaction. One hundred and thirteen SmartGuardTM Auto-mode enabled MiniMedTM
patients aged 14 to 29 years were randomized 1:1 to receive 670G System
either a commercially approved 670G system (n=57) or an After it became available in the USA in 2017, CareLinkTM
AHCL system (n=56) at seven sites in four countries with system data that were uploaded by real-world patients were
12 weeks in each arm. CGM metrics improved in the analyzed [22] and published retrospectively using data for
AHCL (Table 9.2). Patients without CGM or insulin pump the first 3 months from 3141 670G users. These subjects
prior to study experienced the greatest improvement in showed adherent satisfied patients with results sufficiently
TIR. A single patient experienced severe hypoglycemia close to the pivotal 670G study. Percentage time in manual
(SH) and two patients experienced hyperglycemia or keto- mode and subsequent auto mode in real world cohort were
sis in the AHCL arm related to insulin pump use compared TIR: 66 to 73%, < 70mg/dL=2.7 to 2.1%, > 180mg/dL =31.4
to zero patients with SH, three had hyperglycemia or keto- to 24.6% and > 250mg/dL =8.1 to 5.4%. These data confirm
sis related to pump and one with hyperglycemia or ketosis the robustness of the system in subjects that uploaded to
unrelated to insulin pump use in the 670G arm. Fewer sys- the manufacturer’s website.
tem alerts, more time spent in Auto Mode and reduced
Auto Mode exits occurred in the AHCL cohort. The t:slim X2TM Insulin Pump with Control IQ
Technology
780G pivotal trial After becoming available in the US in early 2020, real-world
Medtronic Diabetes conducted a 90-day at-home trial with outcomes were reported in 1435 study participants at time
the MiniMedTM 780G Advanced Hybrid Closed Loop point 1 (at least 3 weeks after starting control-IQ technol-
(AHCL) and showed exciting results with no severe hypo- ogy) versus time point 2 (24 weeks from T1) [23]. The results
glycemia and DKA [20]. Participants stayed in closed loop showed improved TIR over from T1 to T2 (76.7±11.6% vs
(SmartGuardTM) 95% of the time with improved CGM 77.2±12.3%), reduction in time below 70mg/dL (1.3(T1) vs
metrics. This study showed average A1C of 7% with high 1.2(T2) – not significant), reduction in time above 180mg/
user satisfaction. This system features a default target of dL (19.8 (T1) vs 19.0 (T2)) and time above 250mg/dL (2.9
100mg/dL with the option of 120mg/dL and two to eight (T1) vs 2.5 (T2)). Participants expressed high device-related
hours of programmable insulin action time (AIT). With satisfaction and reduced d iabetes impact. This system
showed improvement in psychosocial outcomes and preservation of islet function at 1 year [25]. CLC was used
improvement in TIR glycemic outcomes in people with for 3 days in patients, followed by randomization to SAP vs
T1DM. Currently this system is approved in 18 countries MDI. A1c achieved was 7.3 % and Area Under the Curve
worldwide (United States, Australia, Canada, UK, Denmark, (AUC) of Mixed Meal (MM) stimulated C-peptide was did
Nederland, Czech Republic, France, Italy, New Zealand, not differ between groups. A larger, more definitive study is
South Africa, Finland, Sweden, Norway, Spain, Belgium, currently being performed (see below).
Germany, and Luxembourg). Currently, a randomized parallel assignment RCT
(NCT02871089) is being conducted to assess the effect of CLC
from onset in T1D (CLOuD). The hypothesis is to improve
Ongoing Pivotal Studies/Anticipated
glucose control with the anticipated improvements in AUC
Closed Loop System
MM stimulated C-peptide secretion. Participants will be ran-
domized to CLC (Florence M or CamAPS FX) or MDI, with
HorizonTM Pivotal study (Omnipod HorizonTM) each treatment lasting 24 months. All participants that com-
Omnipod recently showed data on a hybrid, Omnipod® plete the 24-month study period will be invited to continue in
5automated insulin delivery system. Two hundred and thirty- an optional extension phase with the treatment allocated at
five participants have been enrolled in a single arm, multi- randomization for a further 24 months. Estimated 24-month
center, prospective clinical study. Study design includes a study completion date is June 2021.
14-day outpatient, standard therapy phase, followed by 13-
week (94-day) HCL phase in an outpatient setting. After com- CLC is Being Tested in Specific Subgroups
pleting HCL phase, subjects have the option to continue for an with Type 1 Diabetes at Present
additional 6 months. During the HCL phase, a subset of sub- T1D Pregnancy
jects will participate in 5-days of supervised meal and exercise Two CLC studies have been conducted in pregnancy
challenges. Primary outcome measures are Incidence rate of occurring in pre-existing T1D in the UK. The first study
SH, DKA, A1c, and TIR (70–180mg/dl). The estimated study utilized CLC overnight improving with TIR 15 % higher
completion date is December 2021. (74.7 vs 59.5%). The second study used CLC for 24 hours
and showed a non-significant change in TIR (62.3 vs 60.1
Patch Pump Series: Diabeloop Generation 1 %) [26, 27]. Currently, three CLC clinical trials are being
(DBLG1) Hybrid Closed Loop
conducted in pregnancy with preexisting T1D:
A recent RCT done on 71 patients (65 completed) using
DBLG1 (Cellnovo Generation 1; Cellnovo, Paris, France) OO NCT04492566: In this trial, CLC is being evaluated in 21
hybrid CLC system (tubeless patch pump) or SAP over pregnant adult subjects at three US centers for a 48–60
12 weeks of free living [24], followed by 8-week washout hour closed loop session in a supervised outpatient envi-
period and then cross-over to the other intervention for ronment using a CLC with glucose targets similar to
12 weeks showed higher TIR (68.5%) in comparison with pregnancy specific targets.
SAP (59.4%). 2% time below 70mg/dL in DBLG1 is com- OO NCT03774186: In these two US centers trial, up to
pared to 4.3% in SAP group. There were no DKA noted in 47 women would be enrolled at ≤ 9 weeks’ gestation to
each group but nine severe hyperglycemic events (capillary compare 670G with SAP therapy throughout most of
blood glucose > 20mmol/L) in CLC and five SH events in gestation and the first 6 weeks of the postpartum
closed loop, compared to three SH events in SAP. period.
OO NCT04520971: In this trial starting in Europe (2021), up
to 92 women would be recruited at up to 12 weeks gesta-
Closed Loop Control Initiation at Onset
tion and randomized to the 780G pump or standard of
of T1D
care and followed up till delivery. The same CGM sys-
Hybrid CLC was attempted for 72 hours in one study tem (Guardian™ Sensor 3) will be used to account for
within 7 days of onset of T1D to test its effectiveness in differences in type of CGM in both groups for control
group and data are masked with collection at at least to pay threshold of $50,000 per QALY gained [28]. The
four different time points in pregnancy: at 14–17 weeks, technologies significantly improve daily quality of life, with
20–23 weeks, 26–29 weeks, and 33–36 weeks. burden of self-care decreasing iteratively.
Closed Loop in Adults Prone to Hypoglycemia

Insulin Pump Use for People With T2DM
(DCLP2)
NCT04266379: A clinical trial is currently being conducted Though studies of insulin pump use in T2DM have been
in France testing Tandem Control-IQ driven by CGM published, they pre-date pivotal studies of newer medica-
(Dexcom G6) through an algorithm, in free-living conditions such as GLP1 Receptor Agonists and SGLT2 inhibitors
tions vs SAP for 3 months, with a 3-month extension that have improved therapy for T2DM. Therapeutic technol-
phase. Estimated enrollment is of 72 participants and the ogy use for T2DM needs to be addressed in future studies.
estimated study completion date is January 2021.
Conclusion
Cost of Diabetes Technologies
In conclusion, technologies to treat diabetes have signifi-
These technologies are expensive and have been used more cantly advanced in the last decade. The next decade is
in Western Europe, USA, and Australia/New Zealand. expected to provide further advances. Please refer to
There is a lag between approval of technologies and availa- Table 9.3, which describes the technologies currently avail-
bility in each country. A detailed cost benefit analysis done able and being developed further. These include advances
by Pease et al. reported that HCL were associated with a in glucose sensing, utilization of additional sensors, deliv-
mean of $37,767 per QALY gained and 86% of stimulations ery of additional hormones in addition to insulin, espe-
predicted the HCLC to be cost-effective at the willingness cially glucagon and more advanced control algorithms.
TABLE 9.3 Brief synopsis of current and future technologies in closed loop.
Sensors Glucose CGM Implantable Eversense

Non-implantable Dexcom G6, Medtronic Guardian
3, Abbott FreeStyle Libre
Non-glucose Ketone Detects: β-hydroxybutyrate (HB)
single
Non-glucose CKM Detects: HB + glucose or HB/lactate
dual
Hormonal Single hormone Insulin
delivery Dual hormone Insulin + glucagon
Insulin + GLP-1/SGLT2 inhibitors
Triple hormone Insulin + glucagon + pramlintide
Control Proportional integral derivative Medtronic 670G and Medtronic 780G
algorithm (PID)
Model predictive control (MPC) Tandem Control-IQ
Fuzzy logic (FL) Omnipod
Target Single set point
glucose Single zone set point
Varying set points Overnight and daytime targets differ
Varying Day time set points Day time set points may differ between post-
prandial and post-absorptive
CGM – Continuous Glucose Monitoring, CKM – Continuous Ketone bodies monitoring

References with insulin-treated diabetes in the landmark study. Diabetes

Technol Ther. 2021.
1. Atkinson MA and Maclaren NK. The pathogenesis of 14. Foster NC, Beck RW, Miller KM et al. State of type 1 diabetes
insulin-dependent diabetes mellitus. N Engl J Med. management and outcomes from the T1D exchange in
1994;331(21):1428–36. 2016–2018. Diabetes Technol Ther. 2019;21(2):66–72.
2. Atkinson MA, Eisenbarth GS, and Michels AW. Type 1 dia- 15. Carlson AL, Mullen DM, and Bergenstal RM. Clinical use of
betes. Lancet. 2014;383(9911):69–82. continuous glucose monitoring in adults with type 2 diabe-
3. Eisenbarth GS. Type I diabetes mellitus. A chronic autoim- tes. Diabetes Technol Ther. 2017;19(S2):S4-s11.
mune disease. N Engl J Med. 1986;314(21):1360–8. 16. Poolsup N, Suksomboon N, and Kyaw AM. Systematic
4. Insel RA, Dunne JL, Atkinson MA et al. Staging presympto- review and meta-analysis of the effectiveness of continuous
matic type 1 diabetes: a scientific statement of JDRF, the glucose monitoring (CGM) on glucose control in diabetes.
Endocrine Society, and the American Diabetes Association. Diabetol Metab Syndr. 2013;5:39.
Diabetes Care. 2015;38(10):1964–74. 17. Bergenstal RM, Garg S, Weinzimer SA et al. Safety of a
5. Kovatchev B. The artificial pancreas in 2017: The year of hybrid closed-loop insulin delivery system in patients with
transition from research to clinical practice. Nat Rev type 1 diabetes. JAMA. 2016;316(13):1407–8.
Endocrinol. 2018;14(2):74–6. 18. Brown SA, Kovatchev BP, Raghinaru D et al. Six-month ran-
6. Tamborlane WV, Beck RW, Bode BW et al. Continuous glu- domized, multicenter trial of closed-loop control in type 1
cose monitoring and intensive treatment of type 1 diabetes. diabetes. N Engl J Med. 2019;381(18):1707–17.
N Engl J Med. 2008;359(14):1464–76. 19. Bergenstal RM, Nimri R, Beck RW et al. A comparison of
7. Beck RW, Riddlesworth T, Ruedy K et al. Effect of continu- two hybrid closed-loop systems in adolescents and young
ous glucose monitoring on glycemic control in adults with adults with type 1 diabetes (FLAIR): a multicentre, ran-
type 1 diabetes using insulin injections: The DIAMOND domised, crossover trial. Lancet.
randomized clinical trial. JAMA. 2017;317(4):371–8. 2021;397(10270):208–19.
8. Lind M, Polonsky W, Hirsch IB et al. Continuous glucose 20. Carlson AL, Bode BW, Brazg RL et al. 97-LB: Safety and gly-
monitoring vs conventional therapy for glycemic control in cemic outcomes of the Minimed advanced hybrid closed-
adults with type 1 diabetes treated with multiple daily insu- loop (AHCL) system in subjects with T1D. Diabetes.
lin injections: The GOLD randomized clinical trial. JAMA. 2020;June;69(Supplement 1).
2017;317(4):379–87. 21. Collyns O, Meier R, Betts Z et al. 199-OR: Improved glyce-
9. Heinemann L, Freckmann G, Ehrmann D et al. Real-time mic outcomes with Medtronic Minimed advanced hybrid
continuous glucose monitoring in adults with type 1 diabe- closed-loop delivery: Results from a randomized crossover
tes and impaired hypoglycaemia awareness or severe hypo- trial comparing automated insulin delivery with predictive
glycaemia treated with multiple daily insulin injections low glucose suspend in people with type 1 diabetes. Diabetes.
(HypoDE): a multicentre, randomised controlled trial. 2020;June;69 (Supplement 1).
Lancet. 2018;391(10128):1367–77. 22. Stone MP, Agrawal P, Chen X et al. Retrospective analysis of
10. Pratley RE, Kanapka LG, Rickels MR et al. Effect of continu- 3-month real-world glucose data after the MiniMed 670G
ous glucose monitoring on hypoglycemia in older adults system commercial launch. Diabetes Technol Ther.
with type 1 diabetes: A randomized clinical trial. JAMA. 2018;20(10):689–92.
2020;323(23):2397–406. 23. Pinsker JE, Müller L, Constantin A et al. Real-world patient
11. Feig DS, Donovan LE, Corcoy R et al. Continuous glucose reported outcomes and glycemic results with initiation of
monitoring in pregnant women with type 1 diabetes Control-IQ technology. Diabetes Technol Ther.
(CONCEPTT): A multicentre international randomised 2020;23(2):120–7.
controlled trial. Lancet. 2017;390(10110):2347–59. 24. Benhamou PY, Franc S, Reznik Y et al. Closed-loop insulin
12. Kropff J, Choudhary P, Neupane S et al. Accuracy and lon- delivery in adults with type 1 diabetes in real-life conditions:
gevity of an implantable continuous glucose sensor in the a 12-week multicentre, open-label randomised controlled
PRECISE study: A 180-day, prospective, multicenter, pivotal crossover trial. Lancet Digit Health. 2019;1(1):e17-e25.
trial. Diabetes Care. 2017;40(1):63–8. 25. Buckingham B, Beck RW, Ruedy KJ et al. Effectiveness of
13. Gilbert T, Noar A, Blalock O, and Polonsky W. Change in early intensive therapy on β-cell preservation in type 1 dia-
HbA1c and quality of life with real-time CGM use by people betes. Diabetes Care. 2013;36(12):4030–5.
26. Stewart ZA, Wilinska ME, Hartnell S et al. Closed-loop insu- 28. Pease A, Zomer E, Liew D, Earnest A et al. Cost-effectiveness
lin delivery during pregnancy in women with type 1 diabe- analysis of a hybrid closed-loop system versus multiple daily
tes. N Engl J Med. 2016;375(7):644–54. injections and capillary glucose testing for adults with type 1
27. Stewart ZA, Wilinska ME, Hartnell S et al. Day-and-night diabetes. Diabetes Technol Ther. 2020;22(11):812–21.
closed-loop insulin delivery in a broad population of preg-
nant women with type 1 diabetes: A randomized controlled
crossover trial. Diabetes Care. 2018;41(7):1391–9.
10 Optimizing Diet in Patients with Diabetes
Meera Shah
Division of Endocrinology, Diabetes, Metabolism, and Nutrition, Mayo Clinic, Rochester, MN, USA
particularly to food intake and activity, resulting in more

LE A R N I N G P OI NT S
sedentary lifestyles and a rise in the prevalence of over-
●● A plant‐based diet can lower the incidence of type 2
weight and obesity.
diabetes in an average risk population. Medical nutrition therapy is a cornerstone in the man-
●● A program of modest calorie restriction, resulting in agement of type 2 diabetes, and has been shown to be at
5–10% weight loss, should be the initial goal for least as efficacious as anti‐hyperglycemic drugs in lowering
patients with glucose intolerance and an elevated BMI. hemoglobin A1c [2]. Patients obtain dietary advice for the
Time‐restricted feeding, which naturally aligns with the
biological circadian rhythm, may also be beneficial, at
management of type 2 diabetes from various sources; how-
least in the short term. ever, not all are reliable or individualized for that particular
●● Aggressive calorie restriction in early type 2 diabetes patient. A common question from a patient with diabetes
may be beneficial in the remission of type 2 diabetes; for the practicing clinician is “What should I eat?”, and the
however, this approach may only be suitable for a answer can vary based on where the individual is in terms
select few. Overall, emphasis on healthy eating
patterns, and avoidance of ultra‐processed foods,
of glycemic disturbance, what their body weight is, and
added sugars and refined grains should be the focus of what their own personal and cultural preferences are.
a medical nutrition therapy plan for type 2 diabetes.
●● Patients with type 2 diabetes have similar macronutri- A 40‐year‐old man with no family history of type 2 diabetes
ent needs to the general population. Emphasis should asks you about how he can prevent the onset of diabetes
be on quality of the individual macronutrient, rather
than quantity. Sugar substitutes may have a role in
through a dietary plan. He is of normal body weight and has
reducing overall carbohydrate and calorie intake in normal fasting glucose.
patients with type 2 diabetes.
●● Patients with type 1 diabetes should be counseled on a The Nurses Health Study [3] was an early reporter on
personalized dietary plan that accounts for other the impact of lifestyle on the risk of developing type 2 dia-
comorbidities that may impact food intake.
betes. Being overweight or having obesity was the single
best predictor of developing diabetes in this study.
Type 2 diabetes mellitus was once called a “disease of However, when controlled for other variables, diet quality
affluence” because of its initial rise in societies that acquired had a measurable impact on the risk of developing the dis-
material wealth. Now, type 2 diabetes affects populations at ease. Specifically, the group with the highest intake of
all levels of the socio‐economic strata worldwide and in cereal fiber reduced the risk of the disease by approxi-
fact disproportionately affects populations in lower‐ mately 40%; while the group that had the highest consump-
income countries [1]. Across all cultures and countries, the tion of trans fats increased their risk of developing the
common theme has been changes in traditional lifestyles, disease by approximately 40%.

124
Optimizing Diet in Patients with Diabetes 125
Plant‐based diets have been extensively studied. This million people were followed over approximately two dec-
type of diet is variably described across studies, but com- ades and the incidence of diabetes was recorded. Compared
mon elements are: to low adherence to the Mediterranean diet plan, high
adherence to the Mediterranean diet plan reduced the risk
●● Foods derived from plant sources with lower consump-
of developing type 2 diabetes by 12%. The association was
tion of or exclusion of animal products
less strong in subjects with obesity, or those younger than
●● May exclude potentially beneficial yogurt and fish
50 years of age, however.
●● May include consumption of refined grains, starches,
In another European population with traditional cardi-
and sugars
ovascular risk factors, adhering to a Mediterranean style
The Adventist Health Study enrolled approximately 40 diet reduced the risk of developing type 2 diabetes by 50%
000 Adventist Church members in the United States and when compared to a low‐fat diet, an effect that was inde-
Canada in the early 2000s [4]. Over a two‐year period, the pendent of changes in weight or physical activity [8].
incidence of diabetes was higher in members not following Cultural norms, food availability, native produce and
a vegetarian diet compared to members following a semi‐ food preferences are some of the factors that determine
vegetarian or vegetarian diet. Similar findings were seen in how different populations eat. In a large meta‐ analysis
a combined analysis of participants from the Nurses’ combining studies from across the globe (Asia, Europe,
Health Study, Nurses’ Health Study II and Health North American) a 20% reduced risk of developing diabe-
Professionals Follow‐up Study totaling approximately 200 tes was observed in populations with greatest adherence to
000 participants. Subjects who adhered to a plant‐based a plant‐based diet, independent of weight changes [9].
diet or healthy plant‐based diet reduced their risk of devel- Hence, any form of a healthy plant‐based food plan,
oping type 2 diabetes by 50%. Interestingly, participants adjusted to an individual’s personal and cultural prefer-
scoring high on the unhealthy plant‐based index (examples ences, can be beneficial in reducing the overall risk of
include consuming fruit juices, refined grains, potato prod- developing type 2 diabetes.
ucts, and sweets and desserts) had a higher risk of develop- Several mechanisms have been proposed to explain the
ing type 2 diabetes even after adjustment for BMI [5]. A benefit of a plant‐based diet in the risk reduction of type 2
recent meta‐analysis and systematic review combining diabetes. Components commonly present in fruits and
nine large prospective studies showed similar results, con- vegetables such as fiber, vitamins and minerals, antioxi-
firming both the benefit a healthy plant‐based diet which dants, phenolic compounds, and unsaturated fatty acids
minimizes refined carbohydrates, and the detrimental improve insulin sensitivity and blood pressure, reduce
effects of consuming an unhealthy plant‐based diet [6]. long‐term weight gain, and ameliorate systemic inflamma-
A popular version of the plant‐based diet is the tion. There is also the de‐emphasis or avoidance of red and
Mediterranean diet. This food plan is also variably processed meats, which can adversely affect risk of type 2
described across studies, but common elements include: diabetes, possibly owing to high heme iron or dietary cho-
lesterol content [10]. The interaction of animal‐derived
●● primarily plant‐based
compounds (e.g. choline and l‐carnitine) with the gut
●● high consumption of foods of vegetable origin, such as
microbiome and the production of trimethylamine‐N‐
cereals and whole grains, fruits, vegetables, legumes, nuts
oxide have been implicated in the risk of developing gesta-
●● olive oil as the principal source of fat
tional diabetes and type 2 diabetes, although the current
●● fish and poultry consumed in low to moderate amounts
evidence remains inconclusive [11, 12].
●● relatively low consumption of red meat
●● moderate consumption of wine, normally with meals.
The impact of the Mediterranean diet on the develop-

C lin ical pear l
A plant‐based diet can lower the incidence of type 2
ment type 2 diabetes in the general European population diabetes in an average risk population.
was studied by the InterAct Consortium [7]. Almost half
A 45‐year‐old woman with impaired fasting glucose asks you diabetes. An important consideration before recommend-
if she can reduce her risk of developing diabetes with a diet ing such an eating pattern will be to ensure that the patient
plan alone. She has a BMI of 31 kg/m2. is proficient in adjusting anti‐diabetic medications, espe-
cially insulin, because of the risk of precipitating
Being overweight or having obesity is the most impor- hypoglycemia [22].
tant risk factor for the development of type 2 diabetes.
Patients with type 2 diabetes who have an elevated BMI (>
C lin ical pear l
than 25 kg/m2 or > 23 kg/m2 in the Asian population) A program of modest calorie restriction, resulting in
should initially be counseled on weight loss through life- 5–10% weight loss, should be the initial goal for patients
style change. A modest reduction in weight of approxi- with glucose intolerance and an elevated BMI. Time‐
mately 5–10% has been shown to provide many health restricted feeding, which naturally aligns with the
biological circadian rhythm, may also be beneficial, at
benefits with improvement in metabolic parameters [13,
least in the short term.
14]. Calorie restriction is a core component of any weight
loss program. However, no one food plan has been shown
to be superior to another when it comes to efficacy or sus- A 55‐year‐old man with recently diagnosed type 2 diabetes
tainability of weight loss. Hence, a patient’s personal prefer- asks if he can “reverse” diabetes with dietary change alone.
ence should be the most important determinant of what He has an A1c of 7.5% and is on a maximally tolerated dose
dietary plan is advised. of metformin.
Intermittent fasting is one popular way of managing calo-
rie consumption. The health benefits of intermittent fasting The Look AHEAD study enrolled over 5000 partici-
result from its impact on cellular adaptations that promote pants with type 2 diabetes and randomized them to inten-
metabolic flexibility, i.e. the ability of the body to switch sive lifestyle modification or standard medical care [23].
from using glucose as a primary fuel source to fatty acids and Participants in the intervention group were provided a
ketone bodies; increase stress reduction and reduce inflam- calorie‐restricted meal plan with the use of meal replace-
mation [15]. The concept relies on the total absence of calo- ment products in the first year of the intervention. At the
rie consumption over a consecutive period of time, typically end of an almost 10‐year follow‐up period, participants in
at least 16 hours, during which these advantageous cellular the intensive lifestyle intervention group lost more weight
adaptations occur [16, 17]. An observational study in mem- and had better improvements in hemoglobin A1c [23].
bers of the Church of Latter Day Saints, who routinely However, the primary outcome of composite of death from
observe one 24‐hour fast a month as part of their religious cardiovascular causes, non‐fatal MI, non‐fatal stroke, or
obligations, reported the risk of developing type 2 diabetes hospitalization for angina was not different between the
was half that of non‐fasters [18]. two groups. Nevertheless, other clinically important out-
A corollary to intermittent fasting is time‐restricted eat- comes as a result of the intensive lifestyle intervention pro-
ing (TRE), which is based on the concept that endogenous gram, such as an improvement in glycemic control with the
metabolic rhythms (glucose, lipids and energy metabo- subsequent reduction in the need for insulin, improve-
lism) have circadian patterns [19, 20]. A small but rigor- ments in physical fitness, and overall improvement in
ously conducted study on eight men with pre‐diabetes mobility and quality of life should be acknowledged.
showed an improvement in insulin sensitivity and ß cell Significant calorie restriction in the early phases of type
function after TRE, when compared to an isocaloric, tradi- 2 diabetes may be able to achieve disease remission, as
tional “three meals a day” eating pattern suggesting that the demonstrated by the DiRECT trial [24]. This study enrolled
beneficial impact of TRE is independent of weight loss [21]. approximately 300 participants with type 2 diabetes diag-
Data is not as strong in patients with established type 2 dia- nosed in the last six years, with variable glycemic control
betes, thus no firm conclusions can be made about the use (hemoglobin A1c 6% to 12%) and not on insulin. All
of intermittent fasting or TRE in the management of type 2 patients were at least in the overweight category of body
mass index and randomized to an intervention arm or con- benefit of inducing ketonemia was first described by Drs
trol arm. Patients enrolled in the intervention arm were Wilder and Winter at the Mayo Clinic in the 1920s, in the
taken off all diabetes medications and placed on a very low context of treating epilepsy. Numerous studies since then
calorie diet (approximately 850 kilocalories) in the form of have focused on its role in the management of type 2 diabe-
meal replacement products over a period of 12 to 20 weeks. tes. Ketogenic diets are described differently in the litera-
Following the initial weight loss phase, patients were tran- ture; however, common components include a reduction in
sitioned to a stepped food re‐introduction program over carbohydrates (usually to less than 50 g/day) and a relative
the next 2–8 weeks with a goal of weight maintenance. increase in the proportions of protein and fat. Others have
Patients who gained weight during the weight maintenance also variably described an intake between 50–150 g/day, or
phase were offered products to attenuate this. less than 30% of calories from carbohydrates as “low carb
Approximately 25% of the intervention group achieved at intake”. There are many commercial diet programs that uti-
least 15 kg weight loss. Diabetes remission, defined as lize the concept of a low carbohydrate/ketogenic diet for
hemoglobin A1c of less than 6.5% after at least two months weight loss, each with its own macronutrient composition
off all antidiabetic medications, was achieved in 46% of the and other proprietary products. This does make the
intervention group, and sustained for one year. At the two‐ research space heterogeneous and complex and hence firm
year follow‐up mark, this proportion was approximately conclusions cannot be drawn about ketogenic diets in the
36%. In the study population as a whole, the likelihood of management of type 2 diabetes or impaired glucose
diabetes remission was related to the degree of weight loss tolerance.
achieved. In the same vein, sustained remission of diabetes A common confounder in these studies is what role
was more likely in participants with sustained weight loss. weight loss alone plays in explaining the beneficial effects
Of note, there was a 22% drop‐out rate after 24 months in seen by following the diet plan. Additionally, these diets
the intervention arm (versus 6% in the control group), can be difficult to adhere to in the long term. A recent
which speaks to the difficulty in adhering to such an inten- meta‐analysis showed that the metabolic advantages seen
sive program. with the low‐carbohydrate diet (versus a high carbohydrate
The Mediterranean diet plan has been studied in the diet) at 6 months were lost at 12 months [27]. Because of
management of early type 2 diabetes. Patients with early this well recognized phenomenon of difficulty with long‐
type 2 diabetes and on no medications were randomized to term adherence, individuals interested in this diet plan
either an iso‐caloric low‐carbohydrate Mediterranean diet should be regularly assessed and their diet plans individu-
or a low‐fat diet. Participants in the low‐carbohydrate alized as needed. Currently, the American Diabetes
Mediterranean diet group were able to postpone the initia- Association dietary guidelines recommend avoiding the
tion of an anti‐diabetes medication by about two years, ketogenic diet plan in patients with type 2 diabetes who are
when compared to the low‐fat group, largely independent pregnant or lactating; have disordered eating; have renal
of weight loss [25]. disease; or are taking an SGLT‐2 inhibitor because of the
In a one year, prospective randomized trial consisting of potential risk of ketoacidosis [28].
approximately 250 participants with class I obesity, the On the opposite side of the coin, ultra‐processed foods
low‐carbohydrate Mediterranean diet (less than 35% low may have a detrimental effect on glycemic control. Ultra‐
glycemic index carbohydrates) and standard Mediterranean processed foods have been defined as “formulations mostly
diet were superior to the 2003 American Diabetes of cheap industrial sources of dietary energy and nutrients
Association (50–55% carbohydrate, 30% fat and 20% pro- plus additives, using a series of processes” and containing
tein) diet in terms of improvements in A1c. This result minimal whole foods [29]. Ultra‐processed foods are typi-
occurred despite no significant between‐group differences cally ready‐to‐eat formulations with five or more ingredi-
in calorie consumption or achieved weight loss [26]. ents, often including flavor‐enhancing additives, dyes, or
Interest in the ketogenic or very low carbohydrate diet stabilizers (NOVA classification [30]) and can include
for the management of type 2 diabetes remains strong. The foods undergoing physical and chemical processes such as
extruding, molding, pre‐frying, and hydrogenation. Ultra‐

processed foods currently constitute the majority of calo- C lin ical pear l
Aggressive calorie restriction in early type 2 diabetes may
ries consumed in America [31]. In a large population‐based
be beneficial in inducing remission of type 2 diabetes;
prospective study, comprising over 100 000 participants, however, this approach may only be suitable for a select
consuming a higher proportion of ultra‐processed foods in few. Overall, emphasis on healthy eating patterns, and
the diet was associated with a higher risk of developing avoidance of ultra‐processed foods, added sugars and
type 2 diabetes. This was true even when adjusting for the refined grains should be the focus of a medical nutrition
therapy plan for type 2 diabetes.
consumption of unprocessed or minimally processed foods
concurrently [32]. Several theories have been put forward
to explain this observation, including the propensity to A 60‐year‐old man with type 2 diabetes wonders if he should
consume more calories when eating a diet high in ultra‐ limit carbohydrates or include a certain amount of protein or
processed foods [33]. However, no definitive mechanism fiber in his diet to optimize glycemic control. He also wonders
has been identified. about the role of artificial sweeteners, as he has read a lot
The American Diabetes Association in its most recent about their detrimental health effects.
statement on medical nutrition therapy in the management
of type 2 diabetes, emphasizes that there are no comparison Overall, consensus guidelines for macronutrient intake
studies between eating patterns [28]. However, certain in people with type 2 diabetes are similar to the general
commonalities between seemingly beneficial eating pat- population [28]. Carbohydrate quality, rather than quan-
terns such as an emphasis on fruits, non‐starchy vegetables, tity, should be emphasized when providing dietary advice,
whole grains and minimizing refined grains and added as quality of carbohydrate has the greatest impact on glyce-
sugar should form the basis of any dietary intervention in mic response [34] (Figure 10.1). Optimal health benefits
the management of type 2 diabetes. are seen when carbohydrate intake is from whole grains,
Glycemic Index
Calorie density
FIG 10.1 Schematic representation of the relationship between caloric density and glycemic index of typical foods. It is important to
remember that degree of ripeness, texture and cooking method can affect the glycemic index of some foods. Source: Adapted
from The University of Sydney, from: http://www.glycemicindex.com/index.php
fruits, and vegetables. Some patients with diabetes count supplements, to reap other benefits such as micronutrients
carbohydrates to manage post‐meal glycemic excursions and phytochemicals naturally found in foods.
and for others, prandial insulin dosing. There are two Sugar substitutes, which include non‐nutritive sweeten-
assumptions when carbohydrate counting is utilized: one, ers, high‐intensity sweeteners, low‐calorie sweeteners, and
that carbohydrate is the main macronutrient affecting gly- artificial sweeteners, have been deemed safe for consump-
cemic response and two, that carbohydrates are quickly tion by humans. When sugar substitutes are used to replace
converted to glucose after a meal. Carbohydrate counting sugar or other caloric sweeteners, they may have a role in
involves detailed education, consistent practice and regular reducing overall carbohydrate and caloric intake [40]. To
continued visits with a registered dietitian to ensure ongo- date, there is no conclusive evidence that sugar substitutes
ing benefit with this practice. are associated with any health risk to humans [40]. Sugar
Based on current evidence, protein needs in patients alcohols such as sorbitol or xylitol are associated with lower
with type 2 diabetes are no different to that in the general postprandial glucose rise and lower calorie content com-
population, assuming normal renal function. Protein pared to sucrose. Because sugar alcohol is less sweet, a
intake should represent approximately 20% of calories con- greater quantity is often required to achieve an equivalent
sumed, and adjusted to renal function. level of sweetness, thus negating any calorie benefit when
Similar to the general population, guidelines for dietary compared to sugar [41, 42]. Additionally, large amounts of
fat intake in people with diabetes is approximately 20–35% sugar alcohols (30–49 g) can be associated with osmotic
of total calorie intake. As is the case with carbohydrates, diarrhea; hence use of sugar alcohols should be balanced
quality of fat matters more than quantity, and certain with gastrointestinal side effects.
higher‐fat dietary patterns have been shown to improve
cardiovascular risk factors and glycemia in people with
C lin ical pear l
type 2 diabetes [35]. There is clear evidence that consump- Patients with type 2 diabetes have similar macronutrient
tion of trans‐fat is detrimental to health and should be needs to the general population. Emphasis should be on
avoided [36, 37]. An important consideration in people quality of the individual macronutrient, rather than
with diabetes who are also trying to manage weight is the quantity. Sugar substitutes may have a role in reducing
overall carbohydrate and calorie intake in patients with
energy density associated with fat. These patients should
type 2 diabetes.
be mindful of total caloric consumption, even if from
“healthier fats” so as not to promote weight gain.
Regular dietary fiber intake is associated with lower all‐ A 20‐year‐old lady with Type 1 diabetes asks if there is a
cause mortality in people with type 2 diabetes [34, 38]. specific dietary plan she should follow to better manage dia-
Current dietary guidelines recommend 14 g of fiber for betes. She has no complications from diabetes. Her BMI is
every 1000 kilocalories consumed, with no distinction 30 kg/m2.
made between soluble and non‐soluble fiber [37]. Fiber
supplementation may improve glycemic control. When Current dietary guidelines published by the American
overweight and obese volunteers were placed on an iso‐ Diabetes Association are applicable to all people living with
caloric, weight‐maintaining diet that differed only by the diabetes. Some groups have suggested that patients with
amount of dietary fiber from whole grains consumed for type 1 diabetes, who tend to have greater glycemic variabil-
6 weeks (28 g versus 17.8 g), the group on the whole grain ity, could benefit from a low‐carbohydrate diet [43, 44].
diet had a significant improvement in HOMA‐IR and Unfortunately, studies of dietary patterns in patients with
other insulin sensitivity indices as measured by the eugly- type 1 diabetes are often of short duration and in small
cemic‐hyperinsulinemic clamp, compared to individuals numbers of participants, and hence firm conclusions can-
on the refined grain diet [39]. The American Diabetes not be drawn. Patients with type 1 diabetes may also be
Association recommends that fiber be incorporated living with other conditions that alter dietary intake, for
through foods that are naturally high in fiber, as opposed to example celiac disease or gastroparesis. In recent times, the
proportion of patients with type 1 diabetes and concurrent 7. InterAct C, Romaguera D, Guevara M, Norat T, Langenberg C,
obesity has increased, and it is currently estimated at Forouhi NG et al. Mediterranean diet and type 2 diabetes
50% [45]. These patients would benefit from a weight loss risk in the European Prospective Investigation into Cancer
program to both improve glycemic control as well as other and Nutrition (EPIC) study: the InterAct project. Diabetes
Care. 2011;34(9):1913–1918.
aspects of health [46–48].
8. Salas‐Salvado J, Bullo M, Babio N, Martinez‐Gonzalez MA,
Ibarrola‐Jurado N, Basora J et al. Reduction in the incidence
Clinical pearl of type 2 diabetes with the Mediterranean diet: results of the
Patients with type 1 diabetes should be counseled on a PREDIMED‐Reus nutrition intervention randomized trial.
personalized dietary plan that accounts for other Diabetes Care. 2011;34(1):14–19.
comorbidities that may impact food intake. 9. Esposito K, Chiodini P, Maiorino MI, Bellastella G,
Panagiotakos D, Giugliano D. Which diet for prevention of
type 2 diabetes? A meta‐analysis of prospective studies.
Summary
Endocrine. 2014;47(1):107–116.
Medical nutrition therapy remains the foundation in the 10. Micha R, Michas G, Mozaffarian D. Unprocessed red and
management of patients with diabetes mellitus. Dietary processed meats and risk of coronary artery disease and type
guidelines favor healthy eating plans over specific macro- 2 diabetes – an updated review of the evidence. Curr
Atheroscler Rep. 2012;14(6):515–524.
nutrients. Ultimately, the dietary plan that may best benefit
11. Li P, Zhong C, Li S, Sun T, Huang H, Chen X et al. Plasma
the individual patient should take into account available
concentration of trimethylamine‐N‐oxide and risk of gesta-
evidence, and the patient’s own personal and cultural
tional diabetes mellitus. Am J Clin Nutr. 2018;108(3):
preferences. 603–610.
12. Svingen GF, Schartum‐Hansen H, Pedersen ER, Ueland PM,
Tell GS, Mellgren G et al. Prospective associations of sys-
References
temic and urinary choline metabolites with incident type 2
1. Roglic G, World Health Organization. Global Report on diabetes. Clin Chem. 2016;62(5):755–765.
Diabetes. Geneva, Switzerland: World Health Organization; 13. Pan XR, Li GW, Hu YH, Wang JX, Yang WY, An ZX et al.
2016. Effects of diet and exercise in preventing NIDDM in people
2. Evert AB, Franz MJ, American Diabetes Association. with impaired glucose tolerance. The Da Qing IGT and
American Diabetes Association Guide to Nutrition Therapy Diabetes Study. Diabetes Care. 1997;20(4):537–544.
for Diabetes. Third edition. Arlington: American Diabetes 14. Diabetes Prevention Program Research G. The Diabetes
Association; 2017. xvii. Prevention Program (DPP): description of lifestyle interven-
3. Hu FB, Manson JE, Stampfer MJ, Colditz G, Liu S, Solomon tion. Diabetes Care. 2002;25(12):2165–2171.
CG et al. Diet, lifestyle, and the risk of type 2 diabetes 15. Mattson MP, Moehl K, Ghena N, Schmaedick M, Cheng A.
mellitus in women. N Engl J Med. 2001;345(11):790–797. Intermittent metabolic switching, neuroplasticity and brain
4. Tonstad S, Stewart K, Oda K, Batech M, Herring RP, health. Nat Rev Neurosci. 2018;19(2):63–80.
Fraser GE. Vegetarian diets and incidence of diabetes in 16. Panda S. Circadian physiology of metabolism. Science.
the Adventist Health Study‐2. Nutr Metab Cardiovasc Dis. 2016;354(6315):1008–1015.
2013;23(4):292–299. 17. Di Francesco A, Di Germanio C, Bernier M, de Cabo R. A
5. Satija A, Bhupathiraju SN, Rimm EB, Spiegelman D, Chiuve time to fast. Science. 2018;362(6416):770–775.
SE, Borgi L et al. Plant‐based dietary patterns and incidence 18. Horne BD, Muhlestein JB, May HT, Carlquist JF, Lappe DL,
of type 2 diabetes in us men and women: results from three Bair TL et al. Relation of routine, periodic fasting to risk of
prospective cohort studies. PLoS Med. 2016;13(6): diabetes mellitus, and coronary artery disease in patients
e1002039. undergoing coronary angiography. Am J Cardiol. 2012;109(11):
6. Qian F, Liu G, Hu FB, Bhupathiraju SN, Sun Q. Association 1558–1562.
between plant‐based dietary patterns and risk of type 2 19. Poggiogalle E, Jamshed H, Peterson CM. Circadian regula-
diabetes: a systematic review and meta‐analysis. JAMA tion of glucose, lipid, and energy metabolism in humans.
Intern Med. 2019. Metabolism. 2018;84:11–27.
20. Scheer FA, Hilton MF, Mantzoros CS, Shea SA. Adverse meta- energy intake in the USA. Public Health Nutr.
bolic and cardiovascular consequences of circadian misalign- 2018;21(1):114–124.
ment. Proc Natl Acad Sci USA. 2009;106(11):4453–4458. 32. Srour B, Fezeu LK, Kesse‐Guyot E, Alles B, Debras C,
21. Sutton EF, Beyl R, Early KS, Cefalu WT, Ravussin E, Peterson Druesne‐Pecollo N et al. Ultraprocessed food consumption
CM. Early time‐restricted feeding improves insulin sensitiv- and risk of type 2 diabetes among participants of the
ity, blood pressure, and oxidative stress even without weight NutriNet‐Santé prospective cohort. JAMA Intern Med.
loss in men with prediabetes. Cell Metab. 2018;27(6): 2020;180(2):283–291.
1212–1221 e3. 33. Hall KD, Ayuketah A, Brychta R, Cai H, Cassimatis T, Chen
22. Carter S, Clifton PM, Keogh JB. The effect of intermittent KY et al. Ultra‐processed diets cause excess calorie intake
compared with continuous energy restriction on glycaemic and weight gain: an inpatient randomized controlled trial of
control in patients with type 2 diabetes: 24‐month follow‐up ad libitum food intake. Cell Metab. 2019;30(1):226.
of a randomised noninferiority trial. Diabetes Res Clin Pract. 34. He M, van Dam RM, Rimm E, Hu FB, Qi L. Whole‐grain,
2019;151:11–19. cereal fiber, bran, and germ intake and the risks of all‐cause
23. Cardiovascular Effects of Intensive Lifestyle Intervention in and cardiovascular disease‐specific mortality among women
Type 2 Diabetes. New Eng J Med. 2013;369(2):145–154. with type 2 diabetes mellitus. Circulation. 2010;121(20):
24. Lean ME, Leslie WS, Barnes AC, Brosnahan N, Thom G, 2162–2168.
McCombie L et al. Primary care‐led weight management for 35. Qian F, Korat AA, Malik V, Hu FB. Metabolic effects of mon-
remission of type 2 diabetes (DiRECT): an open‐label, clus- ounsaturated fatty acid‐enriched diets compared with car-
ter‐randomised trial. Lancet. 2018;391(10120):541–551. bohydrate or polyunsaturated fatty acid‐enriched diets in
25. Esposito K, Maiorino MI, Petrizzo M, Bellastella G, patients with type 2 diabetes: a systematic review and meta‐
Giugliano D. The effects of a Mediterranean diet on the need analysis of randomized controlled trials. Diabetes Care.
for diabetes drugs and remission of newly diagnosed type 2 2016;39(8):1448–1457.
diabetes: follow‐up of a randomized trial. Diabetes Care. 36. Aronis KN, Khan SM, Mantzoros CS. Effects of trans fatty
2014;37(7):1824–1830. acids on glucose homeostasis: a meta‐analysis of rand-
26. Elhayany A, Lustman A, Abel R, Attal‐Singer J, Vinker S. A omized, placebo‐controlled clinical trials. Am J Clin Nutr.
low carbohydrate Mediterranean diet improves cardiovascu- 2012;96(5):1093–1099.
lar risk factors and diabetes control among overweight 37. U.S. Department of Health and Human Services and U.S.
patients with type 2 diabetes mellitus: a 1‐year prospective Department of Agriculture. 2015–2020 Dietary Guidelines
randomized intervention study. Diabetes Obes Metab. for Americans. December 2015.
2010;12(3):204–209. 38. Burger KN, Beulens JW, van der Schouw YT, Sluijs I,
27. Korsmo‐Haugen HK, Brurberg KG, Mann J, Aas AM. Spijkerman AM, Sluik D et al. Dietary fiber, carbohydrate
Carbohydrate quantity in the dietary management of type 2 quality and quantity, and mortality risk of individuals with
diabetes: a systematic review and meta‐analysis. Diabetes diabetes mellitus. PLoS One. 2012;7(8):e43127.
Obes Metab. 2019;21(1):15–27. 39. Pereira MA, Jacobs DR, Jr., Pins JJ, Raatz SK, Gross MD,
28. Evert AB, Dennison M, Gardner CD, Garvey WT, Lau KHK, Slavin JL et al. Effect of whole grains on insulin sensitivity in
MacLeod J et al. Nutrition therapy for adults with diabetes or overweight hyperinsulinemic adults. Am J Clin Nutr.
prediabetes: a consensus report. Diabetes Care. 2019;42(5): 2002;75(5):848–855.
731–754. 40. Gardner C, Wylie‐Rosett J, Gidding SS, Steffen LM, Johnson
29. Monteiro CA, Cannon G, Moubarac JC, Levy RB, Louzada RK, Reader D et al. Nonnutritive sweeteners: current use
MLC, Jaime PC. The UN Decade of Nutrition, the NOVA and health perspectives: a scientific statement from the
food classification and the trouble with ultra‐processing. American Heart Association and the American Diabetes
Public Health Nutr. 2018;21(1):5–17. Association. Diabetes Care. 2012;35(8):1798–1808.
30. Monteiro CA CG, Levy RB, Moubarac J‐C, Jaime P, Martins AP, 41. Wiebe N, Padwal R, Field C, Marks S, Jacobs R, Tonelli M. A
Canella D, Louzada ML, Parra D; with Ricardo C, Calixto G, systematic review on the effect of sweeteners on glycemic
Machado P, Martins C, Martinez E, Baraldi L, response and clinically relevant outcomes. BMC Med.
Garzillo J, Sattamini I. NOVA. The star shines bright [Food 2011;9:123.
classification. Public health]. World Nutrition. 2016:1–3, 28–38. 42. Fitch C, Keim KS, Academy of N, Dietetics. Position of the
31. Martinez Steele E, Raubenheimer D, Simpson SJ, Baraldi LG, Academy of Nutrition and Dietetics: use of nutritive and non-
Monteiro CA. Ultra‐processed foods, protein leverage and nutritive sweeteners. J Acad Nutr Diet. 2012;112(5):739–758.
43. Nielsen JV, Gando C, Joensson E, Paulsson C. Low carbohy- 46. Chillaron JJ, Benaiges D, Mane L, Pedro‐Botet J, Flores Le‐
drate diet in type 1 diabetes, long‐term improvement and Roux JA. Obesity and type 1 diabetes mellitus management.
adherence: a clinical audit. Diabetol Metab Syndr. 2012; Minerva Endocrinol. 2015;40(1):53–60.
4(1):23. 47. de Ferranti SD, de Boer IH, Fonseca V, Fox CS, Golden SH,
44. Ranjan A, Schmidt S, Damm‐Frydenberg C, Holst JJ, Madsbad Lavie CJ et al. Type 1 diabetes mellitus and cardiovascular
S, Norgaard K. Short‐term effects of a low carbohydrate diet on disease: a scientific statement from the American Heart
glycaemic variables and cardiovascular risk markers in patients Association and American Diabetes Association. Circulation.
with type 1 diabetes: a randomized open‐label crossover trial. 2014;130(13):1110–1130.
Diabetes Obes Metab. 2017;19(10):1479–1484. 48. Corbin KD, Driscoll KA, Pratley RE, Smith SR, Maahs DM,
45. Conway B, Miller RG, Costacou T, Fried L, Kelsey S, Evans Mayer‐Davis EJ et al. Obesity in type 1 diabetes: pathophysiol-
RW et al. Temporal patterns in overweight and obesity in ogy, clinical impact, and mechanisms. Endocr Rev. 2018;39(5):
Type 1 diabetes. Diabet Med. 2010;27(4):398–404. 629–663.
11 Are Insulin Sensitizers Useful Additions
to Insulin Therapy?
John W. Richard, III1 and Philip Raskin2
1
Endocrinology Fellow, Division of Endocrinology, Diabetes, Nutrition and Metabolism, University of Texas
Southwestern Medical Center at Dallas, Dallas, TX, USA
2
Professor of Medicine, Clifton and Betsy Robinson Chair in Biomedical Research, University of Texas Southwestern
Medical Center at Dallas, Dallas, TX, USA
tion characterized by decreased insulin sensitivity, high

insulin levels, high triglycerides, and low HDL cholesterol
leading to type 2 diabetes and cardiovascular disease risk.
●● Patients with type 2 diabetes often need higher doses
of insulin because of obesity and decreased insulin Considering the role of insulin resistance in the develop-
sensitivity. ment of these disease processes, insulin action at the tar-
●● Insulin sensitizing agents act as useful additions to get tissue level is an attractive therapeutic target in type 2
insulin therapy to combat worsening insulin sensitivity. diabetes [1]. The biguanides (i.e., metformin) and the
Metformin suppresses hepatic glucose production
thiazolidinediones (TZDs) act directly to improve insulin
●●
(gluconeogenesis) and increases peripheral tissue

insulin sensitivity. sensitivity, and so are regarded as insulin sensitizing
●● Thiazolidinediones increase insulin‐stimulated glucose drugs.
uptake in peripheral tissues, hepatic insulin sensitivity,
and insulin sensitivity in adipose tissue through
suppression of fatty acid production. Biguanides
●● Insulin sensitizing agents have proven effectiveness at
improving glucose control and lipid profiles, and In Europe, Galega officinalis (goat’s rue or French lilac) was
reducing insulin dose requirements; however, these used for centuries to treat diabetes. This guanidine‐rich
agents are not without side effects. substance later led to the development of several glucose‐
lowering guanidine derivatives in the 1920s. With the dis-
covery of insulin these antidiabetic agents were forgotten
until the 1950s when the biguanides: metformin, phen-
formin and buformin were reintroduced into diabetes
Insulin sensitizers
treatment [1]. Phenformin was withdrawn in many coun-
Insulin resistance is not a new phenomenon. In the 1930s tries and from the U.S. market in 1975 because of a high
British physician Harold Percival Himsworth coined the incidence of lactic acidosis. Buformin received only limited
term “insulin insensitivity” to describe patients he observed use in a few countries, making metformin the principal
with resistance to injectable insulin. Fifty years later, Dr. biguanide drug used in pharmacotherapy worldwide, espe-
Gerald Reaven described metabolic syndrome X, a condi- cially with its improved safety profile and lower cost [1].

133
Mechanism of action in the splanchnic bed also increases through insulin‐

The mechanism of action of biguanides (i.e., metformin) is dependent mechanisms. Additional metabolic effects
still not fully understood, but this agent is distinctly differ- include the suppression of fatty acid oxidation (Figure 11.2)
ent from oral sulfonylureas. Biguanides do not stimulate as well as a reduction in triglyceride levels in patients with
pancreatic insulin secretion and only partial suppression of hypertriglyceridemia [1, 2]. Combined, the cellular effects
gluconeogenesis occurs in the liver, therefore, hypoglyce- of metformin counter insulin resistance and reduce the
mia with monotherapy is rare [2, 3]. Metformin has a vari- toxic metabolic effects of glucose toxicity and lipotoxicity
ety of metabolic effects, some of which extend beyond in type 2 diabetes.
glucose lowering (Table 11.1). At a cellular level, metformin
improves insulin sensitivity through a mediated modifica-
TABLE 11.1 Direct and indirect effects of metformin therapy.
tion of post‐receptor signaling in the insulin pathway.
Recent data have suggested that a protein, adenosine 5′‐ Decreases hyperglycemia
monophosphate protein kinase (AMPK), has been identi- Improves diastolic function
fied as a possible target of metformin [1]. The mainstay of Decreases total cholesterol levels
action of metformin can be attributed to its hepatic effects. Decreases very low‐density lipoprotein cholesterol levels
Decreases low‐density lipoprotein cholesterol levels
Hepatic sensitivity to insulin is increased by metformin,
Increases high‐density lipoprotein cholesterol levels
thereby reducing gluconeogenesis (Figure 11.1) as well as Decreases oxidative stress
glycogenolysis, which contribute to the postprandial Improves vascular relaxation
plasma glucose lowering effects. Skeletal muscle and adi- Decreases plasminogen activator inhibitor‐1 levels
pocytes undergo upregulation of the insulin‐sensitive Increases tissue plasminogen activator activity
Decreases von Willebrand factor levels
GLUT‐4 and GLUT‐1 transporters to the cell membranes,
Decreases platelet aggregation and adhesion
thereby increasing glucose uptake [2]. Glucose metabolism
Metformin Metformin
Lactate
(–)
Glycerol (–) Pyruvate
Ca++
Amino (+) (–)
(–)
acids CO2 ATP
Ca++ ATP+Pi
(–)
OAA
Glucose PEP
Metformin
(+) TK
Increased glycogen Alanine
GLUT 4 storage
FIG 11.1 Hypothesized mechanism of
Glucose Lactate
action of metformin on hepatic
glucose production.
Are Insulin Sensitizers Useful Additions to Insulin Therapy? 135
Liver (–)
PDH
Acyl CoA Krebs Citrate Pyruvate

β- cycle
oxidation
PFK
FA-CoA (–)
G6P
OAA
CPT Glycolysis
(–) PEP Glucose
Insulin/IGF-1
Glucose receptor
Metformin PO4
n
Glucose
tio
ca
transport GLUT 4
slo
Glucose
n
Tra
Adipose
tissues (–) (+) P85 IRS-1
(–) Pl-3
Kinase P110
(–)
FFA
Muscle Cell
FIG 11.2 Effects of metformin on fatty acid metabolism.
Intriguingly, a relatively recent study intended to explore etformin is not metabolized, there is no interference
m
the observation that intravenous administration of with the metabolism of coadministered drugs. Renal
metformin seemed less efficacious, utilized a formulation clearance is approximately 3.5 times greater than creati-
of metformin designed to target the ileum where met- nine clearance, which suggests that tubular secretion is
formin absorption is low. Despite much lower bioavailabil- the main route of metformin elimination. With an oral
ity, therapeutic efficacy was similar to the equivalent oral administration of metformin, patients with normal
dose of conventional metformin formulations. This would renal function have a plasma half‐life of 2–5 hours, and
suggest that actions within the gut e.g. altering entero‐ almost 90% of an absorbed dose is eliminated within 12
endocrine secretion are key to metformin’s mechanism(s) hours [1].
of action and deserve further study [4].
Efficacy of combination therapy
Pharmacokinetics Many people with type 2 diabetes are overweight and insu-
In normal subjects, studies demonstrate that metformin is lin‐resistant, making high doses of insulin often necessary
excreted unchanged in the urine and does not undergo to achieve adequate glucose control. However, insulin ther-
hepatic metabolism or biliary excretion. Because apy is associated with weight gain, which could impede any
Case
A 52‐year‐old woman with obesity and a 9‐year history of type 2 diabetes presents with complaints of fatigue and difficulty
losing weight. She denies polyuria, polydipsia, polyphagia, blurred vision, or vaginal infections.
She states that she has gained an enormous amount of weight since being placed on insulin 6 years ago. Her weight has
continued to increase over the past 5 years, and she is presently at the highest weight she has ever been. She states that
every time she tries to cut down on her eating, she has symptoms of shakiness, diaphoresis, and increased hunger. She
does not follow any specific diet and has been so fearful of hypoglycemia that she often eats extra snacks.
Her health care practitioners have repeatedly advised weight loss and exercise to improve her health status. She
complains that the pain in her knees and ankles makes it difficult to do any exercise.
Her blood glucose values on capillary blood glucose testing have been 170–200 mg/d1 before breakfast. Before supper
and bedtime values range from 150 mg/dl to > 300 mg/dl. Her current insulin regimen is 45 U of NPH plus 10 U of regular
insulin before breakfast and 35 U of NPH plus 20 U of regular insulin before supper. This dose was recently increased after
her HbA1c, was found to be 8.9% (goal < 7.0 %).
Past medical history is remarkable for hypertension, hypertriglyceridemia, and arthritis. Current medications include only
insulin, lisinopril (Prinivil), and hydrochlorthiazide (Dyazide) with triamterene.
On physical exam, her height is 5′ 1½” and her weight is 265 lb. Her blood pressure is 160/88 mmHg. The remainder of
the physical exam is unremarkable.
On laboratory testing, chemistries, BUN, creatinine, and liver function tests are normal. Thyroid function tests and urine
microalburnin are also normal.
After an explanation that the increasing insulin doses were contributing to her weight gain and that she would need to
decrease her insulin dose along with her food intake to prevent hypoglycemia, the patient agreed to follow a restricted‐
calorie diet and to decrease her insulin to 30 U of NPH and 10 U of regular insulin twice daily. As she had no contraindica-
tions to metformin (Glucophage), she was also started on 500 mg orally and it was increased to twice daily after one week.
She returned to the clinic 3 months later, still on the same dose of insulin. She continued to complain of fear of
hypoglycemia in the middle of the night and was overeating at night. Despite this she had lost 7 lb. Her blood glucose
values were still elevated in a range of 120–275 mg/dl before meals.
She was reassured that further insulin reduction would prevent hypoglycemia. Her insulin dosage was decreased to 25 U
of NPH and 5 U of regular insulin twice daily and metformin was increased to 500 mg three times daily. Two months later,
she returned to the clinic with an average blood glucose level of 160 mg/dl. Her weight was now 246 lb, and her HbA1c
was 7.5%. She was feeling much more energetic and was able to start a walking program.
Questions
1. Are insulin sensitizers useful additions to insulin therapy?
progress in achieving glucose control that would normally Further studies have verified these findings, like this
be expected. In a double‐blind, placebo‐controlled study randomized trial comparing insulin monotherapy to com-
where patients were randomly assigned to receive placebo bined therapy with insulin and metformin, which showed
or metformin in combination with insulin for 24 weeks, clear benefit to using metformin with insulin compared to
the data showed that adding metformin to an intensified insulin alone. In this study, insulin as a monotherapy
insulin regimen resulted in an 11% reduction in hemo- resulted in a reduction in HbA1c 8.7 ± 1.6 to 7.0 ± 1.0%
globin A1c compared to insulin therapy alone. The study with approximately 69% more insulin needed from base-
also reported improved glucose control using 29% less line to achieve this effect. Patients in this arm of the study
insulin, and a less complicated insulin regimen with no also required a more complicated insulin regimen in
increase in the occurrence of hypoglycemia or weight approximately 25% of the cases that became time consum-
gain [5]. ing for both the patient and providers. These patients also
gained 4.4 kg of weight. In the second arm of this study many of the symptoms. In clinical studies, women with
using insulin in combination with metformin resulted in a PCOS being treated with insulin sensitizers like metformin
reduction in HbA1c of 8.8 ± 1.2 to 7.1± 1.0%, which was saw a reduction in circulating insulin levels, a reduction in
comparable to the monotherapy arm, however, these androgen levels, improvements in ovarian function, and
results were achieved without an increase in the total daily improved lipid profiles and severity of hirsutism [8].
dose of insulin. In this arm, complexity of the insulin regi-
men was not affected, and there was virtually no weight Contraindications
gain or hypoglycemia reported. The only drawbacks were Metformin therapy is contraindicated in patients with liver
that two‐thirds of patients experienced gastrointestinal failure, alcoholism, and active moderate to severe infec-
side effects, but they were mild and transient [6]. tion [1]; these conditions predispose to development of
Because studies have shown that decreasing the required lactic acidosis, either by increased production or decreased
daily dose of exogenous insulin is associated with a metabolism of lactic acid [1]. Metformin is also contraindi-
decreased risk of cardiovascular disease, another placebo‐ cated in people with kidney disorders (creatinine levels
controlled, randomized, double‐blind trial was designed to above 1.4 mg/dl in women and 1.5 mg/dl in men, depend-
look at effects on glucose control and insulin requirements ing on lean body mass [1], according to the package insert),
when metformin was added in patients with type 2 diabe- and lung disease. Heart failure has long been considered a
tes intensively treated with insulin. After 16 weeks of com- contraindication for metformin use, although a 2007 sys-
bination therapy, the data showed that when metformin tematic review showed metformin to be the only antidia-
was compared with placebo use, there was a statistically betic drug not associated with harm in people with heart
significant improvement in glucose control shown by a failure [9].
hemoglobin A1c of 6.9 versus 7.6%. The data also showed Current recommendations suggest that metformin should
that the use of metformin was associated with less weight be temporarily discontinued before any radiographic study
gain (−1.6 kg compared with placebo) and with a small involving iodinated contrast, as contrast dye may tempo-
decrease in LDL cholesterol (– 0.19 mmol/l) [7]. rarily impair kidney function, indirectly leading to lactic
Taken together, these data suggest that metformin can acidosis by causing retention of metformin in the body [10].
be an effective adjunct to insulin therapy. Studies have Once metformin is withheld, hydration should be main-
shown that adding metformin to insulin therapy reduces tained until preserved kidney function is documented at 24
hemoglobin A1c, total daily dose of insulin, and reduces or and 48 hours after the intervention [1]. General anesthesia
prevents the weight gain associated with intensive insulin should be used cautiously to prevent hypotension, which
treatment. However, metformin has gastrointestinal side leads to renal hypoperfusion and peripheral tissue hypoxia
effects that can hinder titration to its maximum effective with subsequent lactate accumulation [1]. Metformin
dose, but these tend to be mild and transient. should also be used cautiously in the elderly, whose
decreased lean body mass leads to reduced serum creati-
Additional indications nine concentrations that often mask impaired glomerular
Polycystic ovary syndrome filtration rates [1].
We now understand that the main feature of polycystic
ovary syndrome (PCOS), as it relates to the use of insulin Adverse effects
sensitizers, is the decreased insulin sensitivity and com- Lactic acidosis
pensatory hyperinsulinemia that is associated with the Lactic acidosis is a life‐threatening complication of
condition. Women suffering from PCOS have an increased biguanide therapy that carries a mortality rate of
risk of diabetes later in life, and an increased incidence of 30–50% [1]. The estimated incidence of metformin‐
gestational diabetes if they conceive. The pathophysiology associated lactic acidosis is 0.03 cases per 1000 patient‐
of the disorder is not fully understood, but the use of an years [1], which is 10 to 20 times lower than that seen with
insulin sensitizer appears to have a profound effect on phenformin therapy [1], which was withdrawn because of
an increased risk of lactic acidosis (up to 60 cases per mil- est meal to prevent gastrointestinal symptoms. These
lion patient‐years). Development of lactic acidosis is almost symptoms generally disappear within 2 weeks of treat-
always related to coexistent hypoxic conditions that are ment [12]. Medication doses should be increased in 500‐
probably responsible for the associated high mortality rate. mg increments given after meals every 1 to 2 weeks until a
In one report, 91% of patients who developed lactic acido- desirable blood glucose level or the maximal effective daily
sis while being treated with metformin had a predisposing metformin dose of 2000 mg is reached [13].
condition, such as congestive heart failure, renal insuffi-
ciency, chronic lung disease with hypoxia, or age older than
Thiazolidinediones
80 years [1]. Therefore, patients with impaired renal func-
tion or coexistent hypoxic conditions should not be given The peroxisome‐proliferator–activated receptors (PPARs)
metformin. Excessive alcohol consumption may also are a subfamily of super‐receptors that regulate gene
potentiate the effect of metformin on lactate metabolism. A expression in response to ligand binding [14]. Three
careful history of alcohol use is therefore important before PPARs, identified as PPARα, PPARδ (also known as
starting metformin therapy [1]. PPARβ), and PPARγ, have been discovered to date. PPARs
regulate gene transcription by two mechanisms: transacti-
Gastrointestinal vation, which is DNA‐dependent and transrepression,
Side effects of metformin are mostly limited to digestive which is DNA‐independent and may explain the anti‐
tract symptoms, such as diarrhea, flatulence, and abdomi- inflammatory actions of PPARs [14].
nal discomfort [1, 11]. Approximately 5% of patients can- PPARα is expressed mainly in the liver, heart, and mus-
not tolerate treatment because of gastrointestinal side cle, as well as in the vascular wall [2]. Fibrates (i.e., fenofi-
effects [1]. The mechanisms of these side effects remain brate) act as full or partial PPARα agonists. In general, the
unclear but probably are related to accumulation of high activation of PPARα enhances free fatty acid oxidation,
amounts of metformin in the intestinal tissue [2], with controls how multiple genes regulate lipoprotein concen-
resultant elevation of local lactate production. These symp- trations, and has anti‐inflammatory effects (Table 11.2).
toms are dose dependent and can usually be avoided by Studies have shown that PPARα agonists can retard or even
slow titration and, in some cases, reduction of the dose [1]. prevent atherosclerosis in both mice and humans [14].
Metformin therapy should be initiated with a single dose of PPARδ is expressed in several tissue types, with the most
medication (usually 500 mg) taken with the patient’s larg- expression in the skin, brain, and adipose tissue. PPARγ is
TABLE 11.2 Molecular targets of PPARγ and PPARα action.
Liver Skeletal muscle Adipose tissue Vascular wall
PPARγ • Decreased • Increased • Increased fatty acid transport protein‐1 • Decreased intercellular
C‐reactive GLUT4 • Increased acyl–coenzyme A synthetase adhesion molecule‐1
protein • Increased • Increased adiponectin • Decreased vascular‐cell
phosphatidyl • Increased LPL adhesion molecule‐1
3‐kinase • Increased phosphatidyl 3‐kinase • Decreased iNOS
• Decreased • Increased GLUT4 • Decreased interleukin‐6
PDK‐4
PPARα • Decreased • Decreased vascular‐cell
C‐reactive adhesion molecule‐1
protein • Decreased cyclooxygenase‐2
• Decreased • Decreased TNFα
fibrinogen B • Decreased interleukin‐6
• Decreased tissue factor
mainly found in adipose tissue but can also be found in Thiazolidinediones

pancreatic beta cells, vascular endothelium, and mac-
rophages [14]. Its expression is low in tissues that express
predominantly PPARα, such as the liver, the heart, and
skeletal muscle. PPARγ was discovered as a target for thia- Skeletal muscle Adipose tissue Liver
zolidinediones only after testing many agents during mul-
tiple large clinical trials.
↑ Glucose uptake Adipogenesis ↓ Gluconeogenesis
The first thiazolidinedione, troglitazone, was approved in
1997 for use in treating patients with type 2 diabetes in the ↑ Fatty acid uptake
United States. This glucose‐lowering agent was subsequently ↑ Lipogenesis
withdrawn from the market, in March 2000, because of cases ↑ Glucose uptake
of hepatotoxicity. In 1999, two γ agonists, rosiglitazone and
pioglitazone, were approved in the United States for use in
treating hyperglycemia in type 2 diabetes.
↓ Plasma free fatty acids
Mechanism of action ↓ Hyperglycemia

The simulation of PPARγ is considered a major mechanism
through which thiazolidinediones enhance insulin sensi- FIG 11.3 Effects of thiazolidinediones.
tivity. As mentioned previously, PPARγ is mainly found in

adipose tissue and less concentrated in skeletal muscle and found in human adipose tissue are listed in Table 11.2.
liver. PPARγ is essential for normal adipocyte differentia- Several adipokines are regulated by PPARγ but we will
tion and proliferation as well as fatty acid uptake and stor- focus mainly on adiponectin, an adiopocytokine produced
age [14]. Because PPARγ has such a profound effect on exclusively by adipose tissue, because it has been shown to
adipose tissue it has been postulated that thiazolidinedi- increase insulin sensitivity in rodents [14]. Whether
ones achieve their insulin‐sensitizing actions through increases in adiponectin increase hepatic insulin sensitivity
direct means or indirectly, through altered adipokine in humans is still not clear, only rodents have shown this
release affecting insulin sensitivity beyond the adipose tis- effect of PPARγ on adiponectin levels.
sue itself. Basically, thiazolidinediones act directly by pro-
moting fatty acid uptake and storage in adipose tissue Pharmacokinetics
(Figure 11.3). Through this process, there is an increase in The thiazolidinediones are absorbed completely and rap-
adipose‐tissue mass and other insulin‐sensitive tissues, like idly, with peak concentrations achieved within 1–2 hours,
skeletal muscle, the liver, and possibly pancreatic beta cells, but this may be slightly delayed when taken with food.
are spared from the harmful effects of high concentrations Both rosiglitazone and pioglitazone are extensively metab-
of free fatty acids [14]. Therefore, thiazolidinediones keep olized by the liver. The metabolites of rosiglitazone are
fat in its proper place. Data from studies of PPARγ knock- weakly active and are excreted mainly in the urine.
out mouse models supports the notion that thiazolidinedi- Pioglitazone produces metabolites that are more active and
ones act mainly on adipose tissue as long as normal are excreted mostly in the bile. Both thiazolidinediones are
amounts of adipose tissue exist [14]. metabolized by the cytochrome P450 system and no clini-
Although thiazolidinediones keep fat in its proper place cally significant reduction in plasma concentrations of
(in adipose tissue), there are still indirect processes that other drugs has been reported [16].
may be responsible for enhancing insulin sensitivity.
Studies have shown that select thiazolidinediones (i.e., Efficacy of combination therapy
pioglitazone) regulate the expression of more than 100 Studies have shown that thiazolidinediones used in combi-
genes [9] and some of the established PPARγ target genes nation with insulin therapy result in improved blood glu-
cose control, reduced total daily insulin dose, and reduced lin therapy, when patients assigned to pioglitazone 15 or
triglyceride levels. In a randomized, double‐blind, placebo‐ 30 mg daily in combination with their baseline doses of insu-
controlled study of more than 300 insulin‐treated type 2 lin for 16 weeks, HbAlc levels were reduced by 1.0 and 1.3%
diabetic patients, the addition of troglitazone at doses of respectively, compared to a reduction of 0.3% in patients on
200 mg and 600 mg daily resulted in a reduction in HbA1c insulin and placebo [14]. Those patients on insulin and
level of 0.8 and 1.4%, while insulin dose was reduced by 11 pioglitazone gained 2.3 and 3.7 kg, compared with virtually
and 29% respectively. On the other hand, subjects taking no change in weight in those using insulin and placebo.
insulin and placebo experienced a decrease of 0.1% in This data suggests that thiazolidinediones can be effec-
HbAlc level and a 1% increase in insulin dose [17]. Subjects tive adjuncts to insulin therapy. A prospective, randomized
randomized to placebo therapy gained 1.5 kg, while those trial of cardiovascular outcomes, called Prospective
taking the study drug in 200 mg and 600 mg doses in com- Pioglitazone Clinical Trial in Macrovascular Events
bination with insulin gained 1.9 and 3.6 kg. Total choles- (PROACTIVE) has looked at coronary and peripheral vas-
terol, LDL, and HDL cholesterol levels were found to be cular events as a primary outcome and myocardial infarc-
increased with troglitazone use at 600 mg/day. An open‐ tion, stroke, and death as secondary outcomes and the data
label study showed that troglitazone had insulin‐sparing support pioglitazone as having a favorable effect, especially
effects that persisted up to 24 months [18]. In a rand- on lipids, particularly triglycerides, more so than rosiglita-
omized, double‐blind, placebo‐controlled study of more zone [21]. In contrast to the previous study, the
than 200 type 2 diabetic patients who were given troglita- Rosiglitazone Evaluated for Cardiac Outcomes and
zone 400 mg daily in combination with their baseline insu- Regulation of Glycemia in Diabetes (RECORD) trial, a
lin dose, the results showed a significantly greater reduction long‐term, multicenter, randomized, open‐label study,
in total daily dose of insulin and HbAlc level compared found conflicting results when looking at cardiovascular
with those patients using insulin and placebo [19]. In this outcomes in patients with type 2 diabetes treated with
study, end points were defined as a 50% reduction in rosiglitazone plus metformin or sulfonylurea, as compared
injected insulin or either a reduction in blood glucose by > with the combination of metformin and sulfonylurea. The
15%, and only 7% of patients on insulin and placebo data showed that the rate of primary end points (hospitali-
achieved these goals, while 22% of patients taking insulin zation or death from cardiovascular causes) was low at
and troglitazone 400 mg daily achieved these goals. 3.1% per year, while secondary end points like acute myo-
Studies using rosiglitazone or pioglitazone in combina- cardial infarction, death from cardiovascular causes or any
tion with insulin therapy have yielded similar results. In a cause, or the composite of cardiovascular death, myocar-
randomized control study of inadequately controlled insu- dial infarction, and stroke showed no statistically signifi-
lin‐treated type 2 diabetes patients, those receiving cant difference between the rosiglitazone group and the
26 weeks of treatment with rosiglitazone 4 and 8 mg expe- control group [22]. However, the fact remains that TZDs
rienced a reduction in HbAlc level of 0.6 and 1.2% respec- are still responsible for significant weight gain and
tively, as compared with virtually no change in the group increased peripheral edema.
taking insulin and placebo [20]. In this same study, insulin
doses also decreased by 4.8 and 9.4% in the groups taking Additional indications
insulin plus study drug compared to a decrease of 0.6% in Nonalcoholic fatty liver disease
patients taking insulin and placebo. Total cholesterol, Type 2 diabetes has a strong association with nonalcoholic
HDL, and LDL cholesterol levels were found to increase fatty liver disease (NAFLD), which has a spectrum of liver
significantly on treatment with rosiglitazone. Unfortunately, damage that ranges from simple fatty liver (steatosis) to
significant weight gain occurred in the placebo, rosiglita- irreversible, advanced scarring of the liver (cirrhosis) [14].
zone 4mg, and rosiglitazone 8mg groups with correspond- There are an estimated 6.4 million adults in the United
ing values of 0.9 kg, 4.0 kg, and 5.3 kg. In a randomized, States diagnosed with NAFLD, which is the most common
placebo‐controlled study in patients receiving stable insu- cause of elevated levels of liver enzymes [14]. Elevated lev-
els of alanine amino‐transferase (ALT) have been shown to therapy with pioglitazone or combination therapy with
predict type 2 diabetes independently of obesity [14]. Fatty pioglitazone and sulfonylurea, metformin, or insulin has
liver disease is associated with decreased hepatic insulin been used. However, some studies have shown increases in
sensitivity and correlates with insulin requirements during LDL cholesterol levels, between 8 to 16% higher with
insulin therapy in patients with type 2 diabetes [14]. rosiglitazone use. Studies have also shown an approximate
Several recent studies have shown that thiazolidinedi- 10% increase in high‐density lipoprotein (HDL) choles-
ones actually reduce fat accumulation in the liver in terol levels with the use of both drugs. The effects of TZDs
patients with type 2 diabetes as well as in patients with lipo- on triglycerides have been more variable, with decreases in
dystrophy associated with the use of highly active antiret- triglyceride levels having been observed more often with
roviral therapy (HAART). Studies have also shown that pioglitazone than with rosiglitazone [14]. In a direct com-
liver enzymes actually decrease rather than increase during parison of rosiglitazone and pioglitazone, one study of 127
treatment with pioglitazone and rosiglitazone [14]. patients previously treated with troglitazone, supports the
notion that both drugs have similar effects on glucose lev-
Polycystic ovary syndrome els and body weight [14]. This same study supported the
The polycystic ovary syndrome is a disorder that affects notion that pioglitazone is more effective than rosiglita-
approximately 4% of women of reproductive age [14]. zone at lowering LDL cholesterol and serum triglyceride
Women with PCOS frequently develop insulin resistance levels. The difference in efficacy of these two drugs on
and hence have an increased risk for type 2 diabetes [14]. lipids cannot be attributed to the effect they have on serum
Hyperinsulinemia, which accompanies insulin resistance, free fatty acid concentrations, which decreases by approxi-
is thought to contribute to the hyperandrogenism that is mately 20 to 30% in both [14]. Pioglitazone appears to be a
seen in patients with PCOS [14]. Interventions with the partial agonist of PPARα in vitro;, and rosiglitazone
role of reducing insulin levels, like weight loss and medica- appears to be a pure PPARγ agonist [14]. So far, however,
tions (i.e., metformin), can decrease hyperandrogenism the data on mechanisms underlying the effects of the TZDs
and reduce insulin resistance [14]. A large‐scale placebo‐ on lipids in humans is quite limited.
controlled trial of 410 women showed that the use of trogl-
itazone showed significant improvements in ovulatory Lipodystrophies
function, hirsutism, hyperandrogenism, and insulin resist- The most common form of lipodystrophy is that associated
ance [14]. A more recent, small placebo‐controlled study with highly active antiretroviral therapy use in patients
that randomized women to either rosiglitazone and pla- with human immunodeficiency virus (HIV) disease. After
cebo or to rosiglitazone and clomiphene has shown similar only 12 to 18 months of therapy with HAART, approxi-
results. This study demonstrated that 56% of women previ- mately half of patients develop a lipodystrophy‐related
ously resistant to clomiphene were able to ovulate [14]. symptom like facial lipoatrophy [14]. Facial lipoatrophy
Although metformin is considered safe for women who can be disfiguring and stigmatizing, especially since there
become pregnant, rosiglitazone and pioglitazone are classi- is no pharmacologic therapy for this condition, which is
fied as pregnancy category C, which indicates toxic effects usually accompanied by marked insulin resistance.
in studies in animal models, but the results in human stud- Thiazolidinediones would seem to be a wonderful solution
ies are inadequate. If these agents are used during preg- to insulin resistance and lipoatrophy caused by HAART
nancy, then the potential benefit must justify the potential because these drugs increase both insulin sensitivity and
risk to the fetus. Polycystic ovary syndrome is currently not subcutaneous fat mass. Unfortunately, there has been only
an approved indication for the use of TZDs. one placebo‐controlled trial in which patients with
HAART‐associated lipodystrophy were treated with rosigl-
Lipid lowering itazone 8 mg per day for six months, and there was no
Studies have shown that low‐density lipoprotein (LDL) increase in adipose tissue or body weight, in contrast to
cholesterol levels have remained unchanged when mono- studies in patients with type 2 diabetes [14].
Carotid intima‐media thickness there are no guidelines on the use of TZDs in patients with
Carotid IMT is a well‐established surrogate marker for car- diabetes who have any degree of heart disease or for those
diovascular risk. A thickened carotid intima‐media layer already on a TZD who develop CHF. What makes this clin-
not only correlates with increased cardiovascular risk but ical dilemma more perplexing is the more common side
also with the risk of future macrovascular events like myo- effect of peripheral edema associated with TZDs, which
cardial infarction and stroke [23]. Several studies using makes the origin of the development of edema or weight
multiple agents (i.e., ACE inhibitors, calcium channel gain more difficult to decipher.
blockers, β‐blockers, and statins) have shown a reduction Physicians should be cautious before prescribing TZDs
or even regression of carotid IMT in patients without dia- to patients with diabetes who have been previously diag-
betes [23]. However, carotid IMT appears to be more sig- nosed with a condition that increases risk of bone fractures
nificant in patients with type 2 diabetes reflecting a more (i.e., osteoporosis, hyperparathyroidism, Paget’s disease,
dramatic cardiovascular risk in this patient popula- etc.). Recent studies have shown a correlation between
tion [23]. Limited data currently exists about the effect of TZD use and an increased risk of bone fractures. In fact,
intervention in type 2 diabetes on carotid IMT. However, a according to data taken from the ADOPT study group and
recent randomized control study of 192 patients showed others, the relative risk of fractures with thiazolidinediones
that treatment with pioglitazone for 24 weeks led to a sig- remained consistently elevated irrespective of age or meno-
nificant decrease in carotid IMT in patients with type 2 pausal status of women [25]. In addition, a recent meta‐
diabetes, and this was found to be independent of glucose analysis showed that the long‐term use of thiazolidinediones
control [23]. (rosiglitazone, pioglitazone or troglitazone) doubles the
risk of fractures among women with type 2 diabetes, and
Contraindications the overall use of thiazolidinediones significantly increased
In patients with diabetes, hypertension and coronary artery the risk of fractures among patients with type 2 diabetes.
disease occur frequently. These conditions are risk factors Also identified in this meta‐analysis was that thiazolidine-
for the development of congestive heart failure (CHF) [24]. dione use was also associated with significant changes in
Diabetes can affect cardiac structure and systolic or dias- bone mineral density at the lumbar spine and the hip [25].
tolic function, independent of other established risk factors
for CHF because of diabetic cardiomyopathy [24]. This Adverse effects
phenomenon makes diabetes an independent risk factor Multiple studies have correlated weight gain and glyco-
for CHF, which was supported by an analysis of 9591 peo- sylated hemoglobin with thiazolidinedione use – it appears
ple with type 2 diabetes in the Kaiser Permanente that TZDs lead to an increase in body weight of 2 to 3 kg for
Northwest Division that demonstrated that 11.8% of dia- every 1% decrease in glycosylated hemoglobin values. This
betic subjects had CHF at baseline, and an additional 7.7% increase in body weight is the same irrespective of TZD use
developed CHF during a 30‐month follow‐up period [24]. as a monotherapy or in combination with insulin or met-
This suggests that CHF may be present prior to physicians formin in patients with type 2 diabetes. One of the mecha-
beginning therapy with TZDs or that this condition may nisms proposed to explain this phenomenon is that the
develop during the course of treatment. increase in body weight is attributed to expansion of the
TZDs can still be used in patients with underlying subcutaneous fat depot, and in some patients to edema,
asymptomatic heart disease, although its safety has not whereas the mass of visceral fat remains unchanged or even
been fully established. The package inserts for both rosigli- decreases [14].
tazone and pioglitazone indicate that patients with more The use of TZDs is not only associated with weight gain,
advanced heart disease (NYHA class III or IV) were but some patients experience fluid retention and plasma
excluded in pre‐marketing clinical trials, and hence, these volume expansion, which lead to the development of
drugs are not recommended in such patients. Currently, peripheral edema. The development of peripheral edema
has been reported in 4 to 6% of patients undergoing treat- be taken when considering treatment of type 2 diabetic
ment with TZDs as compared with 1 to 2% of those receiv- patients with TZDs who have an increased risk for
ing placebo or other hypoglycemic therapies. Edema fractures.
development appears to occur most when either of the
TZDs is used in combination with insulin. Studies have
Discussion
shown that the use of rosiglitazone 4 or 8 mg per day in
combination with insulin was associated with a 13.1% and So the question remains, “Are insulin sensitizers useful
16.2% incidence of edema, respectively, compared with additions to insulin therapy?” Currently, the data suggest
4.7% in those taking insulin alone [20], and with pioglita- that metformin is a feasible option in all patients with type
zone used at 15 mg or 30 mg daily in combination with 2 diabetes and should be initiated at the onset of insulin
insulin resulted in a combined 15.3% incidence of edema, therapy unless a specific contraindication to metformin
compared with 7.0% for insulin alone [24]. Edema appears use exists. The benefits of metformin therapy in these
to occur at a higher incidence when either of the TZDs is patients is an improvement in hemoglobin A1c, a reduc-
combined with insulin, as opposed to when TZDs are used tion in total daily dose of insulin, and most importantly, a
in combination therapy with other oral hypoglycemic reduction in weight gain despite intensive insulin regi-
agents. This increase in body weight and edema has been mens [5–7]. The only potential disadvantage to metformin
associated with an increase in the incidence of congestive therapy is its side effect profile. This medication is con-
heart failure in patients treated with TZDs and insulin. traindicated in patients with impaired renal function due
As a result, the Food and Drug Administration added a to risk of lactic acidosis, but renal impairment can be a
warning in the drug packet information for those common condition in those with a prolonged diabetes
patients taking rosiglitazone and pioglitazone. The course [1, 10]. The other disadvantage to metformin ther-
European Agency for the Evaluation of Medicinal apy relates to its gastrointestinal side effects which may
Products actually considers insulin therapy a contraindi- impede titration to its maximal effective dose. However,
cation to the use of TZDs. According to the data this these side effects are often mild and transient, and usually
agency presents, the frequency of CHF was 2.5 times can be avoided by slow titration, and in some cases, reduc-
greater in combination therapy of insulin and thiazoli- tion of the dose [1]. Therefore, metformin appears to be an
dinediones than with insulin alone. However, the cause appropriate choice when considering adding an inexpen-
for this result remains unclear. sive insulin sensitizer with few side effects to all patients
In a meta‐analysis of 42 trials containing more than 27 with type 2 diabetes currently on insulin therapy or plan-
000 patients, over 15 000 of them on treatment with rosigli- ning its initiation.
tazone, the data showed an overall odds ratio of 1.43 for Similarly, studies have shown that thiazolidinediones
myocardial infarction (MI) and 1.64 for death from cardio- used in combination with insulin therapy result in
vascular cause. From this meta‐analysis, compared with improved blood glucose control, reduced total daily insulin
placebo or with other hypoglycemic agents, treatment dose, and reduced triglyceride levels [17–20]. However, the
with rosiglitazone was associated with a significant increase side effects of these agents can prove to significantly reduce
in the risk of MI and with an increase in the risk of death any benefit that may have been gained from their use. In
from cardiovascular causes that was of borderline some studies, rosiglitazone has been associated with a sig-
significance [15]. nificant increase in the risk of MI and with an increase in
A recent study suggests an elevated risk of fracture asso- the risk of death from cardiovascular causes [15].
ciated with TZDs compared with other oral hypoglycemic Rosiglitazone has also been found to significantly increase
agents. The risk of fracture with TZD use was present at total cholesterol, HDL, and LDL cholesterol levels [14].
multiple anatomic sites and these results were potentially This data suggests that extreme caution should be used
clinically significant [26]. This suggests that caution should when treating patients with this medication, and likely
even avoided. In addition, recent studies have suggested References

that TZDs have an elevated risk of bone fracture at multiple
1. Krentz AJ, Bailey CJ. Oral antidiabetic agents current role in
anatomic sites compared with other oral hypoglycemic
type 2 diabetes mellitus. Drugs. 2005;65.
agents [26]. Studies of pioglitazone have shown more favora- 2. Klip A, Leiter LA. Cellular mechanism of action of met-
ble outcomes than rosiglitazone with less risk of myocardial formin. Diabetes Care. 1990;13(6):696–704.
infarction and other cardiovascular‐related deaths, as well as 3. Kirpichnikov D, McFarlane SI, Sowers JR. Metformin: an
a better lipid profile response, especially with triglycer- update. Ann Intern Med. 2002;137:25–33.
ides [15, 21]. However, both medications cause significant 4. Buse J B. et al. The primary glucose‐lowering effect of met-
weight gain and peripheral edema [14, 20]. formin resides in the gut, not the circulation: results from
With dual therapy including insulin plus metformin or short‐term pharmacokinetic and 12‐week dose‐ranging
insulin plus thiazolidinediones, significant reductions have studies. Diabetes Care. 2016;39(2):198–205.
5. Aviles‐Santa ML, Sinding J, Raskin P. Effects of metformin
been shown in hemoglobin A1c, total daily dose of insulin,
in patients with poorly controlled, insulin‐treated type 2 dia-
and weight gain. These results would suggest that triple
betes mellitus. Ann Intern Med. 1999;131:182–188.
therapy including insulin plus metformin and TZDs would
6. Strowig SM, Aviles‐Santa ML, Raskin P. Comparison of
produce an even more profound effect, and select studies insulin monotherapy and combination therapy with insulin
have considered these combinations. For example, in this and metformin or insulin and troglitazone in type 2 diabe-
randomized study of 28 type 2 patients with diabetes using tes. Diabetes Care. 2002;25:1691–1698.
insulin monotherapy, 4 months of triple therapy was initi- 7. Strowig SM, Raskin P. Combination therapy using met-
ated by adding metformin to insulin, then troglitazone formin or thiazolidinediones and insulin in the treatment of
(TGZ) versus adding TGZ to insulin, then metformin. diabetes mellitus. Diabetes Obes Metabol. 2005;7:633–641.
Researchers found that hemoglobin A1c decreased with 8. Muth S, Norman J, Sattar N, Fleming R. Women with poly-
dual therapy but improved more during triple therapy cystic ovary syndrome (PCOS) often undergo protracted
treatment with metformin and are disinclined to stop: indi-
(insulin + metformin 7.0%, insulin + TGZ 6.2%; insulin +
cations for a change in licensing arrangements? Hum Reprod.
metformin, adding TGZ 6.1%; insulin + TGZ, adding met-
2004;19:2718–2720.
formin 5.8%). Total daily dose of insulin was significantly
9. Eurich DT, McAlister FA, Blackburn DF et al. Benefits and
reduced in the insulin + TGZ group (−14.1 units), insulin harms of antidiabetic agents in patients with diabetes and
+ TGZ, adding metformin (−13.7 units), and the insulin + heart failure: systematic review. BMJ. 2007;335:497.
metformin, adding TGZ (−17.3 units), but not in the insulin 10. Weir J. Guidelines with regard to metformin‐induced lactic
+ metformin group (−3.2 units). However, patients in the acidosis and X‐ray contrast medium agents. R Coll Radiol.
insulin + TGZ group experienced significant weight gain 1999;54:29–33.
(4.4 kg), and those in the insulin + metformin, insulin + 11. Garber AJ, Duncan TG, Goodman AM, Mills DJ, Rohlf JL.
metformin, adding TGZ, and insulin + TGZ; adding met- Efficacy of metformin in type II diabetes: results of a double‐
formin saw no weight gain. Thus, the order in which met- blind, placebo‐controlled, dose response trial. Am J Med.
1997;103:491–497.
formin and TZDs are added to the treatment regimen may
12. Haupt E, Knick B, Koschinsky T, Liebermeister H, Schneider J,
be important, especially as it relates to weight gain. Adding a
Hirche H. Oral antidiabetic combination therapy with sul-
TZD to a patient who is already on metformin seems to pro-
phonylureas and metformin. Diabetes Metab. 1991;17:
tect the patient from TZD‐induced weight gain [27]. 224–231.
Ultimately, metformin is an excellent choice for all indi- 13. DeFronzo RA, Goodman AM. Efficacy of metformin in
viduals with type 2 diabetes seeking to reach glucose and patients with noninsulin‐dependent diabetes mellitus. The
weight targets, and perhaps should be used in combination Multicenter Metformin Study Group. NEJM. 1995;333:
with insulin unless there is some specific contraindication 541–549.
to its use. Thiazolidinediones, however, appear to be more 14. Yki‐Jarvinen H. Thiazolidinediones. NEJM. 2004;351:
problematic. 1106–1118.
15. Nissen SE, Nissen SE, Wolski K. Effect of rosiglitazone on Clinical Trial In macroVascular Events): a randomized con-
the risk of myocardial infarction and death from cardiovas- trolled trial. Lancet. 2005;366:1279–1289.
cular causes. NEJM. 2007;356:2457–2471. 22. Home PD, Pocock SJ, Beck‐Nielson H et al. Rosiglitazone
16. Baldwin SJ, Clarke SE, Chenery RJ. Characterisation of the evaluated for cardiovascular outcomes – an interim analysis.
cytochrome P450 enzymes involved in the in vitro; metabo- NEJM. 2007;357:28–38.
lism of rosiglitazone. Br J Clin Pharmacol. 1999;48: 424–432. 23. Langenfeld MR, Forst T, Hohberg C et al. Pioglitazone
17. Schwartz S, Raskin P, Fonseca V, Graveline JF. Effect of tro- decreases carotid intima‐media thickness independently of
glitazone in insulin‐treated patients with type II diabetes glycemic control in patients with type 2 diabetes mellitus:
mellitus. NEJM. 1998;338:861–866. results from a controlled randomized study. Circulation.
18. Fonseca V, Foyt HL, Shen K, Whitcomb R. Long‐term effects 2005;111:2525–2531.
of troglitazone: open‐label extension studies in type 2 dia- 24. Nesto RW, Bell D, Bonow RO et al. Thiazolidinedione use,
betic patients. Diabetes Care. 2000;23:354–359. fluid retention, and congestive heart failure. Diabetes Care.
19. Buse JB, Gumbiner B, Mathias NP, Nelson DM, Faja BW, 2004;27(1):256–263.
Whitcomb RW. Troglitazone use in insulin‐treated type 2 25. Loke YK, Singh S, Furberg CD. Long‐term use of thiazoli-
diabetic patients. The Troglitazone Insulin Study Group. dinediones and fractures in type 2 diabetes: a meta‐analysis.
Diabetes Care. 1998;21:1455–1461. CMAJ. 2009;180:32–39.
20. Raskin P, Rendell M, Riddle MC, Dole JF, Freed MI, 26. Solomon DH, Cadarette SM, Choudhry NK, Canning C,
Rosenstock J. A randomized trial of rosiglitazone therapy in Levin R, Sturmer T. A cohort study of thiazolidinediones
patients with inadequately controlled insulin‐treated type 2 and fractures in older adults with diabetes. JCEM.
diabetes. Diabetes Care. 2001;24:1226–1232. 2009;94:2792–2798.
21. Dormandy JA, Charbonnel B, Eckland DJ et al. Secondary 27. Strowig SM, Aviles‐Santa ML, Raskin P. Improved glycemic
prevention of macrovascular events in patients with type 2 control without weight gain using triple therapy in type 2
diabetes in the PROactive Study (PROspective pioglitAzone diabetes. Diabetes Care. 2004;27:1577–1583.
12 Incretin‐Based Therapy for the
Management of Type 2 Diabetes
Kristin Gonzales1 and Adrian Vella2
1
Department of Internal Medicine, Division of Endocrinology, Diabetes, and Metabolism, University of New Mexico
Health Sciences Center, Albuquerque, NM, USA
2
LE A R N I N G P OI NT S production (EGP). In the fasting state, reduced plasma glu

cose concentrations signal the release of glucagon, inhibit
●● Incretin‐based therapy comprises Dipeptidyl
insulin secretion, and promote hepatic EGP. In order for
Peptidase‐4 inhibitors (DPP‐4i) which raise concentra- glucose concentrations to be maintained within a strict
tions of endogenous incretin hormones or Glucagon‐ physiologic range during each of these states, the rate of
Like Peptide‐1 (GLP‐1) Receptor Agonists which EGP must match that of glucose utilization by peripheral
activate the endogenous GLP‐1 receptor.
tissues. Thus, normal glucose tolerance is dependent upon
●● While DPP‐4i are weight‐neutral, GLP‐1 receptor
agonists induce (some) weight loss.
five key factors: (1) insulin secretion, (2) glucagon suppres
●● GLP‐1 receptor agonists have also been associated sion, (3) insulin action i.e. the ability of insulin to stimulate
with cardiovascular benefit. glucose uptake and suppress EGP, (4) glucose effectiveness
i.e. the ability of glucose to stimulate its own uptake, and
suppress EGP, and (5) gastric emptying.
In diabetes or states of impaired glucose tolerance,
defects in some or all of these parameters can result in dys
Introduction
regulation of glucose homeostasis. This leads to chronic
Diabetes is characterized by an absolute or relative insulin elevations in plasma glucose concentrations and predis
deficiency. Type 2 diabetes is a heterogeneous disorder poses to systemic vascular complications [1].
resulting from dysregulation of multiple processes involved
in glucose homeostasis. Under normal conditions, glucose
The incretin effect
homeostasis is tightly maintained through a combination
of nutrient sensing and hormonal signaling to balance Under states of normal glucose tolerance, the insulin secre
peripheral glucose utilization and production. In the fed tory response to an oral glucose load is robust, and the pat
state, elevated plasma glucose concentrations, which are tern of insulin secretion differs substantially from that
factors of gastric emptying and systemic glucose appear observed following intravenous glucose infusions. During
ance, suppress glucagon secretion, stimulate insulin release, intravenous glucose challenges, insulin release is biphasic,
and directly stimulate hepatic glucose uptake for storage as with an initial rapid release followed by a more steady
glycogen. Insulin serves as the main regulatory hormone in release over several hours [2]. This biphasic response has
this state to inhibit hepatic gluconeogenesis and glycogen been observed to be blunted in patients with diabetes dur
olysis, the main processes involved in endogenous glucose ing intravenous glucose challenges [3].

146
Incretin-Based Therapy for the Management of Type 2 Diabetes 147
In contrast, during an oral glucose challenge, patients minal dipeptides His7‐Ala8 of GLP‐1, thus rendering it
with normal glucose tolerance demonstrate a substantially inactive. The result of this rapid degradation is a biological
greater insulin response compared to those seen during half‐life of 2 minutes, a limiting factor when considering
intravenous glucose challenges, despite similar plasma glu GLP‐1 for use in clinical pharmacology [11].
cose concentrations [4]. This differential in insulin
responses is due to a phenomenon known as the incretin
Early development of GLP‐1 RA therapy
effect, a process mediated by both hormonal and neural gut
stimuli. The hormonal stimulus for this phenomenon orig The limitations regarding use of early GLP‐1 analogs were
inates from endogenous incretins – enteroendocrine hor circumvented following the discovery of Exendin‐4, or
mones secreted by K and L cells of the intestine in response exenatide. Exenatide is a 39‐amino acid polypeptide dis
to meal ingestion. They exert direct regulatory actions on covered in nature as having 53% sequence homology with
insulin and glucagon secretion, as well as on gastric empty GLP‐1 and demonstrating similar binding capability to the
ing and satiety. Glucagon‐like peptide‐1 (GLP‐1) and glu GLP‐1 receptor. However, due to alterations in its amino
cose‐dependent insulinotropic polypeptide (GIP) are the acid chain with substitution of an alanine residue for gly
main incretin hormones implicated in these processes and cine at position 8, exenatide is not susceptible to DPP‐4
exert their insulinotropic and glucagon‐suppressive actions degradation. This results in a relatively prolonged half‐life
in a glucose‐dependent fashion [5, 6]. Among patients with of 3.3 to 4 hours and enables intermittent rather than con
impaired glucose tolerance and diabetes, the insulin‐secre tinuous administration [12].
tory and glucagon‐suppressive effects of incretins are nota Early studies in diabetic mice demonstrated that once
bly impaired [7]. daily administration of intravenous exenatide resulted in a
Several hypotheses have been proposed regarding the significant insulinotropic and glucagon‐lowering effect, as
mechanisms for these diminished incretin effects, as well did twice daily subcutaneous injections of exenatide.
as the role of incretins in development of diabetes. Similar findings were observed in humans with type 2 dia
Although some studies suggest alterations in secretion of betes who received intravenous exenatide during hyper
GIP and GLP‐1 in response to glucose loads, there remains glycemic insulin clamps [13–16]. Between 2003 and 2005,
to be conclusive evidence implicating changes to endoge randomized controlled trials (RCTs) found that exenatide
nous incretin hormone secretion in the pathogenesis of could be feasibly and effectively administered as a twice‐
type 2 diabetes [8, 9]. daily subcutaneous injection to patients with type 2, and it
Due to the insulinotropic and glucagon‐lowering effects could be safely added to metformin for glucose‐lowering
of endogenous GIP and GLP‐1, research efforts have and weight reduction [17, 18]. Thus, immediate‐release
sought to determine their relevance in the clinical manage exenatide was the first GLP‐1 receptor agonist (GLP‐1
ment of diabetes. Pharmacologic doses of GIP have dem RA) to be FDA‐approved for use in type 2 diabetes in
onstrated little effect on diabetic islets with regard to 2005. In 2015, an extended‐release formulation requiring
insulin secretion and glucagon suppression, although only once weekly subcutaneous administration was
pharmacologic GLP‐1 has shown substantial promise in approved [19].
patients with type 2 diabetes [10]. In 1993, Nauck and col To date, multiple other GLP‐1 RAs have been approved
leagues demonstrated that pharmacologic doses of syn for use in type 2 diabetes. These include liraglutide, lixi
thetic GLP‐1 (7‐36 amide), when administered via senatide, dulaglutide, and semaglutide. Albiglutide has
continuous infusion, lowered plasma glucose and glucagon received FDA approval but is not currently marketed for
concentrations, and raised insulin concentrations in a glu use. These agents share sequence homology ranging from
cose‐mediated fashion, among patients with type 2 diabe 53–97% with native GLP‐1. Each is administered as subcu
tes. Similar to native GLP‐1, however, these early synthetic taneous injections and available in pre‐filled pens [20].
forms of GLP‐1 underwent rapid degradation by dipepti Oral semaglutide received FDA‐approval in 2019, making
dyl peptidase‐4 (DPP‐4). This enzyme cleaves the N‐ter it the first oral GLP‐1 RA available in its class [21].
Clinical benefits and side effects of secretagogues, GLP‐1 RAs rarely cause hypoglycemia, as
GLP‐1 RAs their influence on beta cell insulin release is largely glu
cose‐dependent. However, increased incidence of hypogly
GLP‐1 RAs share common features unique to this class
cemia has been noted in trials employing combined use of
including HbA1c‐lowering and weight loss. GLP‐1 RAs are
these agents with sulfonylureas. Thus, caution is recom
known to suppress appetite and reduce gastric emptying,
mended when initiating GLP‐1 RAs in the setting of con
although there is tachyphylaxis to the gastric‐emptying effects,
current insulin or sulfonylurea use.
suggesting that weight loss is independent of this process.
In addition to the common GI side effects of GLP‐1
Table 12..1 delineates the ranges of weight loss and HbA1c
RAs, preclinical safety studies in rats have implicated these
reduction observed in head‐to‐head trials of GLP‐1 RAs. agents in the development of c‐cell hyperplasia, c‐cell ade
In addition to the insulinotropic and glucagon‐lowering nomas, and medullary thyroid carcinoma. Although the
effects of GLP‐1 RAs, cell and animal models have impli relevance of these findings to humans remains unclear, the
cated these agents in the trophic regulation of beta cells. FDA has issued a black‐box warning on all GLP‐1 RAs cau
Exenatide has been shown to augment beta cell mass, pro tioning against use in patients who have a personal or fam
mote beta cell differentiation, and to inhibit beta cell apop ily history of medullary thyroid cancer or multiple
tosis. These data would collectively imply a potential role endocrine neoplasia 2 (MEN 2) [23]. Notably, in large
for incretin mimetics in attenuation of disease progression. RCTs involving GLP‐1 RAs in humans, there have been no
However, human studies have yet to demonstrate any sig documented cases of medullary thyroid cancer related to
nificant effect in this regard [12, 22, 23]. pharmacological use of these agents.
Although data is limited for use of GLP‐1 RAs in the Post‐marketing reports have also implicated incretin
setting of mild to moderate chronic kidney disease (CKD), mimetics in the development of acute pancreatitis, prompt
several RCTs have shown them to be safe and effective in ing the FDA to issue warning labels regarding this potential
these settings [24]. The majority of data pertaining to use risk [23]. Significant controversy has also emerged regard
of these agents in CKD is derived from large cardiovascular ing the risk for dysplastic pancreatic lesions in patients
outcome trials (CVOTs) wherein renal outcomes were treated with GLP‐1 RAs. However, diabetes, itself, is asso
examined as secondary endpoints. In these trials, signifi ciated with a 2‐fold increased risk of acute pancreatitis and
cant improvements in composite renal endpoints were a higher baseline risk of pancreatic cancer compared to the
noted with liraglutide, semaglutide, dulaglutide, extended‐ general population. These factors preclude causal associa
release exenatide, and lixisenatide; however, these improve tions from being made between clinically‐significant pan
ments were largely driven by reductions in new‐onset creatic pathology and GLP‐1 RAs. Furthermore, adverse
macroalbuminuria [25–29]. The extent of impact on other event reporting in large‐scale RCTs have not reported sig
renal outcomes such as progression to end‐stage renal dis nificant increases in these outcomes among patients treated
ease (ESRD), progressive microalbuminuria, or change in with GLP‐1 RAs or DPP‐4 inhibitors. Yet, these associa
estimated glomerular filtration rate (eGFR) remains tions should also be factored into clinical decision‐making
unclear, as many of these trial did not demonstrate sub when considering use of GLP‐1 RAs [23, 25–29, 31–34].
stantial improvement in these endpoints. There have been
some reports of acute renal impairment in association with
GLP‐1 RAs, although these findings are likely confounded The evidence for GLP‐1 ras in
by underlying conditions such as dehydration due to the management of type 2 diabetes
gastrointestinal (GI) side effects of GLP‐1 RAs [30]. Exenatide
The most common adverse events associated with Exenatide is available in both immediate and extended‐
GLP‐1 RAs are nausea and vomiting, both effects of their release formulations, allowing for either twice daily or once
actions on gastric emptying and the paraventricular weekly administration. Dosing options for immediate‐
nucleus of the hypothalamus. Unlike insulin and insulin release exenatide are 5μg and 10 μg, and extended‐release
exenatide is available as a 2mg dose. In RCTs with sulfony 1.7%, respectively). Major hypoglycemic events did not
lureas, metformin, thiazolinediones (TZDs), and insulin as differ between groups, although more minor episodes of
comparators, immediate‐release exenatide was shown to hypoglycemia occurred in the placebo [41]. These findings
lower HbA1c by approximately 0.4–1.0% and to reduce indicated a pertinent role for GLP‐1 RAs as adjunct ther
weight by 1.6–2.8 kg among patients with type 2 diabe apy to basal insulin in patients with suboptimal glycemic
tes [17, 35–38]. Extended‐release exenatide, however, dem control.
onstrated a greater impact on HbA1c‐lowering of 1.4% vs.
1.0% when compared in a head‐to‐head fashion with Liraglutide
immediate‐release exenatide. Weight loss was similar Liraglutide was the second FDA‐approved GLP‐1 RA and
between groups receiving the two drugs, although GI side is available in three dosing regimens: 0.6 mg daily, 1.2 mg
effects were less common with extended‐release exenatide. daily, and 1.8 mg daily. Standard practice is to initiate ther
The only reported events of hypoglycemia occurred in apy at the lowest dose and gradually escalate to the maxi
patients receiving GLP‐1 RA therapy in combination with mally‐tolerated dose. In a series of clinical trials conducted
sulfonylureas [39]. to examine the efficacy of liraglutide either as mono‐ or
Exenatide has also been studied as add‐on therapy to combination therapy with sulfonylureas, TZDs, or met
basal insulin. One RCT of immediate‐release exenatide formin, liraglutide reduced HbA1c on average by 0.8–1.5%.
added to basal insulin, with or without metformin and/or When compared directly to immediate‐release exenatide,
pioglitazone, demonstrated an average HbA1c reduction of liraglutide 1.8mg daily lowered HbA1c by a mean of 1.12%
1.74% in the exenatide‐treated group compared to a 1.04% vs. 0.79%. Weight loss was similar between the two
reduction in the group managed without exenatide. groups [42]. Similarly, when compared head‐to‐head with
Average weight loss was 1.8 kg in the exenatide group, extended‐release exenatide, liraglutide demonstrated sig
whereas a weight gain of 1.0 kg occurred in the comparator nificantly greater HbA1c and weight reduction (1.48% vs.
group. There were no significant differences in hypoglyce 1.28% and 3.57 kg vs. 2.68 kg, respectively) [43].
mia [40]. Similarly, an RCT examining the efficacy of Several studies have also examined the utility of liraglu
extended‐release exenatide added to insulin glargine with tide as add‐on therapy to basal insulin. One such RCT ana
or without metformin and/or a sulfonylurea demonstrated lyzed the effect of liraglutide added to insulin degludec.
an average decline in HbA1c of 0.96% vs. 0.23%. Weight The results noted an HbA1c reduction of 1.9% in the group
also declined significantly in the exenatide‐treated group treated with liraglutide vs. 0.9% in the comparator group,
by 1.04 kg vs. 0.48 kg. However, more individuals in the along with significant reductions in overall 9‐point self‐
weekly exenatide group experienced GI‐related adverse monitored blood glucose values. Average weight loss of
events prompting withdrawal from the study (3.9% vs. 2.7 kg occurred in the liraglutide‐treated group vs. no
TABLE 12.1 GLP‐1 RAs: effects on glucose and weight in head‐to‐head trials.
Immediate‐ Extended‐
release release
exenatide [42, exenatide [43, Liraglutide Dulaglutide Semaglutide
Drug 45, 48, 82, 83] 52, 82, 83] [42, 43, 49, 84] Lixisenatide [45] [48, 49, 53] [52, 53]
Average A1c 0.79–1.5 0.9–1.9 0.99–1.48 0.79 1.42–1.51 1.5–1.8

reduction (%)*
Average weight 1.24–3.8 1.9–3.7 2.91–3.6 2.8 1.3–2.9 5.6–6.5
loss (kg)*
Average HbA1c and weight reductions reported with maximum doses of respective GLP‐1 RAs
*
weight loss in the comparator group [44]. These findings Semaglutide

paved the way for use of long‐acting GLP‐1 RAs such as Semaglutide is the newest GLP‐1 RA to be FDA‐approved
liraglutide in combination with basal insulin. for use in type 2 diabetes. It is currently available as a once
weekly injection at doses of 0.5 mg and 1.0 mg. Structurally,
Lixisenatide it is similar to liraglutide, although has a longer half‐life of
Lixisenatide is available as a once‐daily injection in doses approximately 7 days.
of 10 μg and 20 μg. When compared head‐to‐head with In the first RCT to examine the effects of this GLP‐1 RA,
immediate‐release exenatide, lixisenatide proved to be semaglutide was compared to diet and exercise in patients
non‐inferior with regard to reduction in HbA1c [45]. with type 2 diabetes who were treatment naïve and who
When compared to liraglutide as add‐on therapy to met had failed to achieve glycemic control. Average HbA1c
formin, liraglutide lowered HbA1c significantly by 1.8% reductions in the semaglutide groups were 1.45% and
compared to 1.2% with lixisenatide. Weight loss effects of 1.55% with 0.5mg and 1.0mg, respectively vs. 0.02% in the
each were not statistically different [46]. placebo group. Weight decreased by 3.73 kg and 4.54 kg
One RCT compared lixisenatide to rapid‐acting insulin with 0.5mg and 1.0mg of semaglutide, respectively vs.
glusiline, in combination with basal insulin glargine. This 0.98 kg with placebo [51]. Semaglutide has also demon
study demonstrated a significant reduction in 2‐hour post strated greater weight loss when compared head‐to‐head
prandial glucose concentrations in the lixisenatide group with extended‐release exenatide and dulaglutide [52, 53].
compared to that receiving insulin glusiline three times Thus, to date, semaglutide portends the greatest weight loss
daily. These findings further supported earlier data dem effect of all available GLP‐1RAs.
onstrating that GLP‐1 RAs could be effectively added to Like other GLP‐1 RAs when examined as add‐on ther
basal insulin regimens as a safe and viable replacement for apy to basal insulin, semaglutide plus basal insulin showed
bolus meal‐time insulin in select patient populations [47]. substantial HbA1c‐lowering of 1.4% and 1.8% with 0.5mg
and 1.0mg semaglutide weekly. This was in contrast to a
Dulaglutide 0.1% reduction in the group managed without semaglu
Dulaglutide is available as a weekly GLP‐1 RA in doses of tide. Reductions were greater with higher baseline HbA1c
0.75 mg or 1.5mg. When compared to immediate‐release values. Significantly more patients in the semaglutide
exenatide with metformin and pioglitazone, HbA1c groups achieved HbA1c levels < 7% without an increase in
decreased by 1.51% and 1.3% with 1.5 mg and 0.75 mg, hypoglycemia. Overall reductions in 7‐point self‐moni
respectively vs. 0.99% with exenatide. Weight reductions tored blood glucose values were greater in the semaglutide
were similar between groups in this study [48]. In a non‐ groups, as well. Weight reductions were also greater in the
inferiority RCT comparing dulaglutide with liraglutide, semaglutide group (3.7 kg and 6.4 kg with semaglutide
HbA1c reductions were comparable (1.42% vs. 1.36 %), but 0.5 mg and 1.0 mg, respectively vs. 1.4 kg in the comparator
significantly greater weight reduction was noted with lira group). However, there were a higher number of hypogly
glutide vs. dulaglutide (3.6 kg vs. 2.9 kg) [49]. cemic events noted in the semaglutide groups among
Dulaglutide 1.5mg weekly was studied as add‐on ther patients whose baseline HbA1c levels were < 8.0%. No dif
apy to basal insulin glargine alone in a prospective RCT to ferences in rates of hypoglycemia were noted between
determine its effect on HbA1c longitudinally over groups among individuals whose baseline HbA1c values
28 weeks. During this time, HbA1c declined by 1.44% in were > 8.0% [54].
the dulaglutide group vs. 0.67% in the comparator group. A notable finding that has emerged from studies of
Weight declined by 1.91 kg with dulaglutide vs. a gain of semaglutide is its association with progressive diabetic
0.5 kg in the comparator group. Additionally, hypoglyce retinopathy. In a major CVOT of semaglutide compared
mia rates were similar between the two, although more with placebo, there were significantly higher rates of dia
patients treated with dulaglutide discontinued the study for betic retinopathy complications (HR 1.76) among patients
GI‐related adverse events [50]. treated with semaglutide. Post‐hoc analysis demonstrated
that rates were highest among those with pre‐existing dia pared to wild type mice, exhibited higher plasma GLP‐1
betic retinopathy at study entry, and the trial was not and insulin levels, as well as lower plasma glucose levels, in
designed to adequately assess progression of retinal response to a glucose challenge. Administration of a
changes over time. While signals of early retinopathy pro DPP‐4 inhibitor to wild type mice resulted in similar out
gression were noted in trials involving both liraglutide and comes to those of DPP‐4 knockout mice. This suggested
exenatide, these findings did not reach statistical signifi that the absence of DPP‐4 augments both the plasma con
cance, and the trials were similarly not designed to ade centrations and insulinotropic effects of endogenous
quately examine this outcome. Despite these limitations, GLP‐1 [56].
there is substantial evidence to suggest that rapid glucose The GLP‐1‐augmenting effect of DPP‐4 inhibitors has
lowering with agents such as insulin and GLP‐1 RAs does been well‐established, with several studies demonstrating a
result in an “early worsening” phenomenon of diabetic doubling of active GLP‐1 in the postprandial state [57–59].
retinopathy. This is related to both the magnitude and In association with this elevation is a lowering of postpran
rapidity of glucose lowering. Thus, pre‐existing diabetic dial glucose and glucagon concentrations, although the
retinopathy should be carefully evaluated and managed direct effect on insulin secretion is controversial. Some
prior to initiation of GLP‐1 RAs, and risks of progression studies report an increase in insulin levels, while others
should be thoroughly discussed with patients prior to com show no change with DPP‐4 inhibition [57, 59, 60].
mencement of therapy [26, 55]. Additionally, gastric emptying does not seem to be signifi
With glycemic control and weight reduction being the cantly impaired by DPP‐4 inhibition, suggesting that
cornerstones of type 2 diabetes care, GLP‐1 RAs demon delayed meal appearance is not a suitable explanation for
strate efficacy in each of these realms. Although direct the glucose‐lowering effect of these agents [57].
comparisons cannot be made between many of the trials All FDA‐approved DPP‐4 inhibitors interact with vari
examining utility of these agents individually, general con ous sites of the DPP‐4 enzyme to exert > 80% inhibi
clusions about this class can be made based on the available tion [61]. As demonstrated in early human studies
evidence: (1) reductions in HbA1c and weight occur along regarding mechanisms of action, this degree of inhibition
a spectrum within the class of GLP‐1 RAs, with longer‐act is sufficient to raise active GLP‐1 levels to produce both
ing agents demonstrating greater overall HbA1c reduction insulinotropic and glucagon‐lowering effects, yet insuffi
compared to shorter‐acting agents, and semaglutide exhib cient to alter gastric emptying [57]. Collectively, this data
iting the most potent effect on weight loss; (2) GLP‐1RAs suggests that the glucose lowering effects of DPP‐4 inhibi
are useful as both monotherapy and in combination with tors are likely due to actions on postprandial beta cell func
basal insulin or other oral antidiabetic agents, and (3) their tion rather than on mediation of gastric emptying. This is
relative safety, efficacy, and ease of use allow for glycemic unique from GLP‐1 RAs and may explain the glucose‐low
optimization while minimizing risks of hypoglycemia and ering differential observed between these two drug classes.
overall burden of disease. A multitude of DPP‐4 inhibitors have been FDA‐
approved for use in type 2 diabetes. Currently available
DPP‐4 inhibitors include sitagliptin, saxagliptin, linaglipi
The role of DPP‐4 inhibitors in type 2
tin, vildagliptin, and alogliptin, all of which can be admin
diabetes management
istered orally. When compared to long‐acting GLP‐1 RAs,
Endogenous DPP‐4 functions to degrade GLP‐1, thus they exert a lesser effect on glycemic control and are con
reducing its biological half‐life. The class of incretin sidered to be weight‐neutral. Yet, despite these differences
mimetics can, therefore, be expanded to include drugs spe in efficacy, both GLP‐1 RAs and DPP‐4 inhibitors share a
cifically designed for inhibiting this enzymatic degradation low risk for hypoglycemia, with the highest propensity for
and augmenting the endogenous effects of native GLP‐1: hypoglycemia being in the setting of combined use with
DPP‐4 inhibitors. Rationale for the use of DPP‐4 inhibitors insulin or sulfonylureas. In accordance with their lack of
stems from studies of DPP‐4 knockout mice which, com effect on gastric emptying, DPP‐4 inhibitors are also not
known to cause significant GI side effects and are generally enhanced beta cell responsiveness, a function that was not
well‐tolerated [61]. preserved with use of sulfonylureas [67]. Several RCTs have
DPP‐4 inhibitors have been shown to be safe for use in also compared the glycemic effectiveness of DPP‐4 inhibi
the setting of renal impairment or ESRD. As sitagliptin, tors to TZDs, namely pioglitazone, and demonstrated simi
saxagliptin, vildagliptin, and alogliptin are renally‐ lar if not enhanced efficacy in glucose‐lowering with
excreted, doses should be adjusted when used in these set DPP‐4 inhibitors. However, TZDs continue to have greater
tings. Linaglitptin does not undergo renal excretion but is impact on insulin resistance through their actions on
rather hepatically‐metabolized. Thus, dose adjustments are PPARℽ in adipocytes and hepatocytes, thus distinguishing
not necessary when this agent is initiated in the setting of them as a unique therapeutic class to be considered in
CKD. Both RCTs and meta‐analyses have implicated select cases [70–72].
DPP‐4 inhibitors and GLP‐1 RAs as having potential Although decisions regarding second and third‐line
nephroprotective effects in patients with type 2 diabetes, therapies for type 2 diabetes are dependent upon a balance
although the extent of this remains unclear. While several of both risks and benefits of various drug classes,
RCTs have demonstrated improvements in development of DPP‐4 inhibitors have demonstrated substantial efficacy
macroalbuminuria with use of these agents, there remains for their intended use. They are associated with minimal
to be consistent data demonstrating overall improvement hypoglycemia and can be administered in such a way that
in progression to ESRD or microalbuminuria [62]. Thus, limits the daily burden of disease for patients. Given the
further research is necessary to better define the role of robust data available, it is reasonable to consider
incretin‐based therapies on clinically‐important renal out DPP‐4 inhibitors in advance of sulfonylureas as add‐on
comes in diabetes. therapy, although cost of these medications continues to
Similar to GLP‐1 RAs, considerable debate has ensued limit more widespread use.
regarding the potential risk for acute pancreatitis in asso
ciation with DPP‐4 inhibitor use. While robust data has not
Cardiovascular outcomes with incretin‐
substantiated these concerns, the FDA has issued a warn
based therapies for type 2 diabetes
ing label on all incretin‐based therapies regarding risk of
acute pancreatitis [23, 63]. Unlike GLP‐1 RAs, however, Diabetes confers significant morbidity and mortality, with
there have been no signals indicating a heightened risk for atherosclerotic cardiovascular disease (ASCVD) being the
c‐cell hyperplasia or medullary thyroid cancers with use of leading cause of death among those affected. Patients with
DPP‐4 inhibitors. diabetes possess approximately a 2‐fold increased risk of
Traditionally, oral diabetic agents such as sulfonylureas coronary artery disease, stroke, and death from other vas
and TZDs have been advocated for as adjuncts to met cular causes [73]. In 2008, amid concerns for increased car
formin therapy if glycemic targets were not met. However, diovascular risk with rosiglitazone, the FDA mandated that
these agents are associated with inherent weight gain and, cardiovascular safety trials be conducted for all agents
in the case of sulfonylureas, significant risk of hypoglyce approved and marketed for type 2 diabetes [74]. The major
mia. Thus, given their beneficial effect on hyperglycemia, criterion for safety was achievement of a minimum hazard
low risk of hypoglycemia, weight neutrality, and ease of ratio (HR) of 1.30 with composite endpoints of major
administration, DPP‐4 inhibitors have become viable con adverse cardiovascular events (MACE) including non‐fatal
tenders as second or third‐line agents in the management myocardial infarction, non‐fatal stroke, and death from
of type 2 diabetes. In a series of RCTs, DPP‐4 inhibitors cardiovascular disease [74]. Subsequently, this led to a
were shown to be non‐inferior to sulfonylureas with regard robust influx of CVOTs involving incretin‐based therapies,
to HbA1c‐lowering and demonstrated statistically signifi as well a novel class of drugs targeting renal sodium‐glu
cant benefits over sulfonylureas with regard to hypoglyce cose transporters, or SGLT‐2 inhibitors. These CVOTs
mia and weight loss [63–69]. One of these trials also were designed to demonstrate non‐inferiority to standard
suggested that DPP‐4 inhibitors were associated with of care. The emergence of these trials dramatically trans
formed the landscape of diabetes management and opened signals were not observed in CVOTs involving sitagliptin
the door to novel therapies portending both glycemic and (TECOS), alogliptin (EXAMINE), or linagliptin
cardiovascular benefit. (CARMELINA) [33, 34, 77]. Similarly, there have been no
To date, CVOTs have been published for all FDA‐ demonstrable associations with heart failure among the
approved GLP‐1 RAs. These CVOTs are prospective, GLP‐1 RAs, suggesting that this is unlikely to be a class
multi‐center RCTs designed to demonstrate non‐inferior effect among incretin‐based therapies.
ity to standard of care in patients with type 2 diabetes. All In the last decade, there has been increased recognition
have achieved the FDA‐mandated criteria for safety in the of the effects of incretin‐based therapy beyond glycemic
primary specified cardiovascular outcomes. Standard dia control. Based on their proven impact on weight and car
betes care in these trials has generally consisted of a combi diovascular disease, incretin‐based therapies have been
nation of oral hypoglycemic agents, excluding incretin‐based firmly integrated into the mainstream of diabetes care.
therapies, and/or basal insulin. Inclusion criteria have
encompassed patients with cardiovascular risk factors and/
Incorporating incretin‐based therapy into
or established ASCVD.
clinical practice
Of the GLP‐1 RA CVOTs published to date, LEADER,
SUSTAIN‐6, HARMONY, and REWIND have reported The approach to management of type 2 diabetes has tradi
liraglutide, semaglutide, albiglutide, and dulaglutide, tionally followed a hierarchical algorithm in accordance
respectively, to have statistically significant reductions in with disease severity and rate of progression. As insulin
cardiovascular events compared to placebo [25–27, 31]. resistance is strongly correlated with obesity, weight reduc
Trials involving sustained‐release exenatide (EXSCEL) and tion through lifestyle modifications is an early intervention
lixisenatide (ELIXA) have demonstrated non‐inferiority which should be maintained throughout the disease
for these cardiovascular endpoints [28, 29]. Clinical sub course [79]. However, given the inevitability of disease pro
group analyses have also provided insights regarding use of gression despite patients’ best efforts, pharmacologic ther
these agents in clinical practice. In a meta‐analysis of the apy is ultimately crucial in optimizing diabetes care.
major GLP‐1 RA CVOTs, it was suggested that patients Unless contraindicated or poorly‐tolerated, metformin
with baseline characteristics of BMI > 30 kg/m2, male gen remains the first‐line pharmacologic agent of choice for
der, and pre‐existing ASCVD were most likely to benefit type 2 diabetes. In addition to modulating insulin resist
from GLP‐1 RAs, and specifically those that have been ance and improving glycemic control, metformin has also
shown to significantly improve MACE outcomes [75]. been shown to have modest cardiovascular benefit, an
Thus, based on these trials, the ADA recommends that advantage which has not been reported with sulfonylu
GLP‐1 RAs demonstrating cardiovascular benefit be initi reas [80, 81].
ated in patients with type 2 diabetes who have established If glycemic goals remain unmet with metformin mono
ASCVD or multiple risk factors for such [76]. therapy, initiation of additional agents is then based on fac
The CVOTs for DPP‐4 inhibitors have established non‐ tors pertaining to patient preference, comorbidities, disease
inferiority to standard care with regard to MACE out severity, cost‐effectiveness, and risk of associated adverse
comes, although superiority has not been demonstrated [33, events. Insulin should be initiated early if hyperglycemia is
34, 77, 78]. As a result, DPP‐4 inhibitors are considered to severe (plasma glucose > 300 mg/dL or HbA1c > 10%), or
have neutral effects on cardiovascular disease in patients if signs of hyperglycemia or increased catabolism are pre
with type 2 diabetes. One signal for increased heart failure sent. In the absence of these findings, oral agents can be
hospitalizations was noted in the SAVOR‐TIMI 53 trial, a considered as add‐on therapy to metformin. Sulfonylureas
CVOT that compared saxagliptin to standard care with and TZDs have traditionally been the recommended sec
placebo. In this study, saxagliptin was associated with a ond and third‐line therapies in these settings. How
heart failure hospitalization incidence of 3.5% compared to ever, while generally well‐tolerated and effective for
2.4% in the placebo group (HR 1.27).(78) However, similar glucose‐lowering, these agents portend undesirable risks
including weight gain and, in the case of sulfonylureas, rable efficacy for major cardiovascular outcomes and being
hypoglycemia [81]. associated with weight gain and hypoglycemia, sulfonylu
In light of the recent data that has emerged for incretin‐ reas remain potent glucose‐lowering medications and can
based therapies, institutional guidelines regarding stand be carefully titrated to maximize both safety and efficacy.
ard pharmacotherapy for type 2 diabetes have shifted TZDs are also associated with a variety of adverse events
substantially. Current ADA guidelines recommend consid including heart failure, weight gain, fractures, and bladder
eration of GLP‐1 RAs for patients not adequately con cancer. Yet, pioglitazone, the only available TZD, remains
trolled with metformin who (1) are at increased risk for effective for hyperglycemia and can be considered as a
ASCVD, (2) have established ASCVD, (3) are in need of means of cost‐effective treatment in select clinical scenar
therapies that minimize hypoglycemia, or (4) would bene ios.(81)
fit from weight reduction or minimization of weight
gain [81]. For patients with heightened risk of ASCVD, or
Conclusions
for those with established ASCVD, GLP‐1 RAs with proven
cardiovascular benefit such as liraglutide, dulaglutide, and Type 2 diabetes is characterized by chronic hyperglycemia
semaglutide are recommended. Although albiglutide has resulting from impaired beta cell function, insulin resist
proven cardiovascular benefit in CVOTs, it is not currently ance, and diminished endogenous incretin effects. Incretin
marketed for use in the United States. Worth mentioning, hormones are integral to glycemic control and thus have
although outside the scope of this chapter, are been the targets of considerable research for their role in
SGLT‐2 inhibitors. These agents have gained significant both treatment and attenuation of disease progression.
traction in diabetes care, having proven cardiovascular and High‐quality evidence has shown GLP‐1 RAs to be effec
renal efficacy in patients with type 2 diabetes. Therefore, tive in multiple aspects of type 2 diabetes, from glucose‐
they have also been incorporated into current guidelines lowering and weight reduction to cardiovascular outcome
for consideration in patients with or at risk for ASCVD, improvement. DPP‐4 inhibitors have demonstrated mod
heart failure, and chronic kidney disease [81]. est efficacy in glycemic control while minimizing weight
In the absence of ASCVD or significant risk thereof, gain, and both carry a low risk for hypoglycemia. Indeed,
DPP‐4 inhibitors can be considered as second‐line therapy. many questions remain unanswered regarding the dura
These agents may also be added when GLP‐1 RAs are not tion of glycemic control that can be expected with incretin‐
concurrently used [81]. Due to its association with based therapies, as well as their role in mitigating
increased heart failure hospitalizations, saxagliptin should progression of disease. However, available evidence sug
be avoided in the setting of heart failure [78]. However, all gests that these agents exhibit substantial promise in reduc
other DPP‐4 inhibitors, given their neutral status in the ing diabetes‐related morbidity, and they have dramatically
role of cardiovascular disease, are considered safe for modified the approach to care for patients with diabetes.
patients at higher risk for or with established ASCVD. Of
note, combined use of GLP‐1 RAs and DPP‐4 inhibitors is
Key points
not recommended due to lack of expected benefit with
additive therapy and paucity of data regarding their com ●● The incretin effect mediates insulin release and gluca
bined use. gon suppression in response to meal ingestion. This
Although incretin‐based therapies and SGLT‐2 inhibi effect is disrupted in diabetes.
tors now occupy a dominant role in type 2 diabetes care, it ●● Incretin mimetics enhance islet insulinotropic and
is important for clinicians to remain familiar with the role glucagon suppressive responses in a glucose‐mediated
of sulfonylureas and TZDs in select cases. According to fashion.
current ADA guidelines, these medications should be con ●● Hypoglycemia is rare with GLP‐1 RAs and DPP‐4 inhib
sidered as second‐line therapy to metformin when cost is a itors, although may occur when used in combination
major consideration for patients. While lacking in compa with insulin or sulfonylureas.
●● Weight loss is expected with GLP‐1 RAs, with semaglu 2016;4(6):525–536. doi: 10.1016/S2213‐8587(15)00482‐9. Pub
tide being most effective for weight reduction. Med PMID: WOS:000376462800018.
●● DPP‐4 inhibitors are weight‐neutral and less effective 8. Calanna S, Christensen M, Holst JJ, Laferrere B, Gluud LL,
Vilsboll T, Knop FK. Secretion of glucagon‐like peptide‐1 in
for glycemic control compared to GLP‐1 RAs.
patients with type 2 diabetes mellitus: systematic review and
●● Liraglutide, semaglutide, albiglutide, and dulaglutide
meta‐analyses of clinical studies. Diabetologia.
have been shown to significantly reduce cardiovascular
2013;56(5):965–972. Epub 2013/02/05. doi: 10.1007/s00125‐
events in major cardiovascular outcome trials. 013‐2841‐0. PubMed PMID: 23377698; PMCID. PMC3
●● DPP‐4 inhibitors have neutral effects on cardiovascular 687347.
outcomes. Saxagliptin should be avoided in the setting 9. Calanna S, Christensen M, Holst JJ, Laferrere B, Gluud LL,
of heart failure. Vilsboll T, Knop FK. Secretion of glucose‐dependent insuli
notropic polypeptide in patients with type 2 diabetes: sys
tematic review and meta‐analysis of clinical studies. Diabetes
References Care. 2013;36(10):3346–3352. Epub 2013/09/26. doi:
10.2337/dc13‐0465. PubMed PMID: 24065842; PMCID.
1. Sharabi K, Tavares CD, Rines AK, Puigserver P. Molecular
PMC3781498.
pathophysiology of hepatic glucose production. Mol Aspects
10. Nauck MA, Heimesaat MM, Orskov C, Holst JJ, Ebert R,
Med. 2015;46:21–33. Epub 2015/11/10. doi: 10.1016/j.mam
Creutzfeldt W. Preserved incretin activity of glucagon‐like
.2015.09.003. PubMed PMID: 26549348; PMCID. PMC46
peptide 1 [7‐36 amide] but not of synthetic human gastric
74831.
inhibitory polypeptide in patients with type‐2 diabetes mel
2. Brunzell JD, Robertson RP, Lerner RL, Hazzard WR, Ensinck
litus. J Clin Invest. 1993;91(1):301–307. Epub 1993/01/01.
JW, Bierman EL, Porte D. Relationships between fasting plasma doi: 10.1172/JCI116186. PubMed PMID: 8423228; PMCID.
glucose levels and insulin‐secretion during intravenous glu PMC330027.
cose‐tolerance tests. J Clin Endocr Metab. 1976;42(2): 11. Nauck MA, Kleine N, Orskov C, Holst JJ, Willms B,
222–229. doi:. DOI 10.1210/jcem‐42‐2‐222. PubMed PMID: Creutzfeldt W. Normalization of fasting hyperglycaemia by
WOS:A1976BG96800003. exogenous glucagon‐like peptide 1 (7‐36 amide) in type 2
3. Del Prato S, Marchetti P, Bonadonna RC. Phasic insulin (non‐insulin‐dependent) diabetic patients. Diabetologia.
release and metabolic regulation in type 2 diabetes. Diabetes. 1993;36(8):741–744. Epub 1993/08/01. doi: 10.1007/
2002;51 Suppl 1:S109–116. Epub 2002/01/30. doi: 10.2337/ bf00401145. PubMed PMID: 8405741.
diabetes.51.2007.s109. PubMed PMID: 11815468. 12. Chia CW, Egan JM. Incretin‐based therapies in type 2 diabe
4. Elrick H, Stimmler L, Hlad CJ, Jr., Arai Y. Plasma insulin tes mellitus. J Clin Endocrinol Metab. 2008;93(10):3703–
response to oral and intravenous glucose administration. 3716. Epub 2008/07/17. doi: 10.1210/jc.2007–2109. PubMed
J Clin Endocrinol Metab. 1964;24:1076–1082. Epub 1964/ PMID: 18628530; PMCID. PMC2579648.
10/01. doi: 10.1210/jcem‐24‐10‐1076. PubMed PMID: 13. Greig NH, Holloway HW, De Ore KA, Jani D, Wang Y, Zhou
14228531. J, Garant MJ, Egan JM. Once daily injection of exendin‐4 to
5. Toft‐Nielsen MB, Damholt MB, Madsbad S, Hilsted LM, diabetic mice achieves long‐term beneficial effects on blood
Hughes TE, Michelsen BK, Holst JJ. Determinants of the glucose concentrations. Diabetologia. 1999;42(1):45–50.
impaired secretion of glucagon‐like peptide‐1 in type 2 dia Epub 1999/02/23. doi: 10.1007/s001250051111. PubMed
betic patients. J Clin Endocrinol Metab. 2001;86(8): PMID: 10027577.
3717–3723. Epub 2001/08/15. doi: 10.1210/jcem.86.8.7750. 14. Young AA, Gedulin BR, Bhavsar S, Bodkin N, Jodka C,
PubMed PMID: 11502801. Hansen B, Denaro M. Glucose‐lowering and insulin‐sensi
6. Nauck M, Stockmann F, Ebert R, Creutzfeldt W. Reduced tizing actions of exendin‐4: studies in obese diabetic (ob/ob,
incretin effect in type 2 (non‐insulin‐dependent) diabetes. db/db) mice, diabetic fatty Zucker rats, and diabetic rhesus
Diabetologia. 1986;29(1):46–52. Epub 1986/01/01. doi: monkeys (Macaca mulatta). Diabetes. 1999;48(5):1026–
10.1007/bf02427280. PubMed PMID: 3514343. 1034. Epub 1999/05/20. doi: 10.2337/diabetes.48.5.1026.
7. Nauck MA, Meier JJ. The incretin effect in healthy individuals PubMed PMID: 10331407.
and those with type 2 diabetes: physiology, pathophysiology, and 15. Egan JM, Clocquet AR, Elahi D. The insulinotropic effect of
response to therapeutic interventions. Lancet Diabetes Endo. acute exendin‐4 administered to humans: comparison of
nondiabetic state to type 2 diabetes. J Clin Endocrinol Metab. secondary prevention of kidney disease. Diabetes Res Clin
2002;87(3):1282–1290. Epub 2002/03/13. doi: 10.1210/jcem.87. Pract. 2019;157:107907. Epub 2019/11/05. doi: 10.1016/j.dia
3.8337. PubMed PMID: 11889200. bres.2019.107907. PubMed PMID: 31676332.
16. Fehse F, Trautmann M, Holst JJ, Halseth AE, Nanayakkara 25. Marso SP, Daniels GH, Brown‐Frandsen K, Kristensen P,
N, Nielsen LL, Fineman MS, Kim DD, Nauck MA. Exenatide Mann JF, Nauck MA, Nissen SE, Pocock S, Poulter NR, Ravn
augments first‐ and second‐phase insulin secretion in LS, Steinberg WM, Stockner M, Zinman B, Bergenstal RM,
response to intravenous glucose in subjects with type 2 diabe Buse JB, for the LEADER Steering Committee on behalf of
tes. J Clin Endocrinol Metab. 2005;90(11):5991–5997. Epub the LEADER Trial Investigators. Liraglutide and cardiovas
2005/09/08. doi: 10.1210/jc.2005‐1093. PubMed PMID: cular outcomes in type 2 diabetes. N Engl J Med. 2016;375(4):
16144950. 311–322. Epub 2016/06/14. doi: 10.1056/NEJMoa1603827.
17. DeFronzo RA, Ratner RE, Han J, Kim DD, Fineman MS, PubMed PMID: 27295427; PMCID. PMC4985288.
Baron AD. Effects of exenatide (exendin‐4) on glycemic 26. Marso SP, Bain SC, Consoli A, Eliaschewitz FG, Jodar E,
control and weight over 30 weeks in metformin‐treated Leiter LA, Lingvay I, Rosenstock J, Seufert J, Warren ML,
patients with type 2 diabetes. Diabetes Care. 2005;28(5):1092– Woo V, Hansen O, Holst AG, Pettersson J, Vilsboll T, for the
1100. Epub 2005/04/28. doi: 10.2337/diacare.28.5.1092. Pub SUSTAIN‐6 Investigators. Semaglutide and cardiovascular
Med PMID: 15855572. outcomes in patients with type 2 diabetes. N Engl J Med.
18. Kolterman OG, Buse JB, Fineman MS, Gaines E, Heintz S, 2016;375(19):1834–1844. Epub 2016/09/17. doi: 10.1056/NEJ
Bicsak TA, Taylor K, Kim D, Aisporna M, Wang Y, Baron Moa1607141. PubMed PMID: 27633186.
AD. Synthetic exendin‐4 (Exenatide) significantly reduces 27. Gerstein HC, Colhoun HM, Dagenais GR, Diaz R,
postprandial and fasting plasma glucose in subjects with Lakshmanan M, Pais P, Probstfield J, Riesmeyer JS, Riddle
type 2 diabetes. J Clin Endocr Metab. 2003;88(7):3082–3089. MC, Ryden L, Xavier D, Atisso CM, Dyal L, Hall S, Rao‐
doi: 10.1210/jc.2002‐021545. PubMed PMID: WOS:0001 Melacini P, Wong G, Avezum A, Basile J, Chung N, Conget I,
83926300020. Cushman WC, Franek E, Hancu N, Hanefeld M, Holt S,
19. Knop FK, Bronden A, Vilsboll T. Exenatide: pharmacokinet Jansky P, Keltai M, Lanas F, Leiter LA, Lopez‐Jaramillo P,
ics, clinical use, and future directions. Expert Opin Cardona Munoz EG, Pirags V, Pogosova N, Raubenheimer
Pharmaco. 2017;18(6):555–571. doi: 10.1080/14656566. PJ, Shaw JE, Sheu WH, Temelkova‐Kurktschiev T, for the
2017.1282463. PubMed PMID: WOS:000399620000003. REWIND Investigators. Dulaglutide and cardiovascular out
20. Aroda VR. A review of GLP‐1 receptor agonists: evolution comes in type 2 diabetes (REWIND): a double‐blind, ran
and advancement, through the lens of randomised con domised placebo‐controlled trial. Lancet. 2019;394(10193):
trolled trials. Diabetes Obes Metab. 2018;20 Suppl 1:22–33. 121–130. Epub 2019/06/14. doi: 10.1016/S0140‐6736(19)
Epub 2018/01/25. doi: 10.1111/dom.13162. PubMed PMID: 31149‐3. PubMed PMID: 31189511.
29364586. 28. Holman RR, Bethel MA, Mentz RJ, Thompson VP,
21. Anderson SL, Beutel TR, Trujillo JM. Oral semaglutide in Lokhnygina Y, Buse JB, Chan JC, Choi J, Gustavson SM,
type 2 diabetes. J Diabetes Complicat. 2020:107520. Epub Iqbal N, Maggioni AP, Marso SP, Ohman P, Pagidipati NJ,
2020/01/19. doi: 10.1016/j.jdiacomp.2019.107520. PubMed Poulter N, Ramachandran A, Zinman B, Hernandez AF, for
PMID: 31952996. the EXSCEL Study Group. Effects of once‐weekly exenatide
22. Lee YS, Lee C, Choung JS, Jung HS, Jun HS. Glucagon‐like on cardiovascular outcomes in type 2 diabetes. N Engl J Med.
peptide 1 increases beta‐cell regeneration by promoting 2017;377(13):1228–1239. Epub 2017/09/15. doi: 10.1056/
alpha‐ to beta‐cell transdifferentiation. Diabetes. 2018;67(12): NEJMoa1612917. PubMed PMID: 28910237.
2601–2614. Epub 2018/09/28. doi: 10.2337/db18‐0155. Pub 30. Scheen AJ. Pharmacokinetics and clinical use of incretin‐
Med PMID: 30257975. based therapies in patients with chronic kidney disease and
23. Tasyurek HM, Altunbas HA, Balci MK, Sanlioglu S. type 2 diabetes. Clin Pharmacokinet. 2015;54(1):1–21. doi:
Incretins: their physiology and application in the treatment 10.1007/s40262‐014‐0198‐2. PubMed PMID: WOS:000347147
of diabetes mellitus. Diabetes Metab Res Rev. 2014;30(5):354– 300001.
371. Epub 2014/07/06. doi: 10.1002/dmrr.2501. PubMed 31. Hernandez AF, Green JB, Janmohamed S, D’Agostino RB,
PMID: 24989141. Sr., Granger CB, Jones NP, Leiter LA, Rosenberg AE, Sigmon
24. De Cosmo S, Viazzi F, Piscitelli P, Leoncini G, Mirijello A, KN, Somerville MC, Thorpe KM, McMurray JJV, Del Prato
Bonino B, Pontremoli R. Impact of CVOTs in primary and S, for the Harmony Outcomes committees and investigators.
Albiglutide and cardiovascular outcomes in patients with 38. Nauck MA, Duran S, Kim D, Johns D, Northrup J, Festa A,
type 2 diabetes and cardiovascular disease (Harmony Brodows R, Trautmann M. A comparison of twice‐daily
Outcomes): a double‐blind, randomised placebo‐controlled exenatide and biphasic insulin aspart in patients with type 2
trial. Lancet. 2018;392(10157):1519–1529. Epub 2018/10/07. diabetes who were suboptimally controlled with sulfonylu
doi: 10.1016/S0140‐6736(18)32261‐X. PubMed PMID: 302 rea and metformin: a non‐inferiority study. Diabetologia.
91013. 2007;50(2):259–267. Epub 2006/12/13. doi: 10.1007/s00125
32. Husain M, Birkenfeld AL, Donsmark M, Dungan K, ‐006‐0510‐2. PubMed PMID: 17160407.
Eliaschewitz FG, Franco DR, Jeppesen OK, Lingvay I, 39. Wysham CH, Rosenstock J, Vetter ML, Dong F, Ohman P,
Mosenzon O, Pedersen SD, Tack CJ, Thomsen M, Vilsboll Iqbal N. Efficacy and tolerability of the new autoinjected
T, Warren ML, Bain SC, for the PIONEER 6 Investigators. suspension of exenatide once weekly versus exenatide
Oral semaglutide and cardiovascular outcomes in patients twice daily in patients with type 2 diabetes. Diabetes Obes
with type 2 diabetes. N Engl J Med. 2019. Epub 2019/06/12. Metab. 2018;20(1):165–172. Epub 2017/07/08. doi:
doi: 10.1056/NEJMoa1901118. PubMed PMID: 10.1111/dom.13056. PubMed PMID: 28685973; PMCID.
31185157. PMC5724491.
33. White WB, Cannon CP, Heller SR, Nissen SE, Bergenstal 40. Buse JB, Bergenstal RM, Glass LC, Heilmann CR, Lewis
RM, Bakris GL, Perez AT, Fleck PR, Mehta CR, Kupfer S, MS, Kwan AYM, Hoogwerf BJ, Rosenstock J. Use of twice‐
Wilson C, Cushman WC, Zannad F, for the EXAMINE daily exenatide in basal insulin‐treated patients with type 2
Investigators. Alogliptin after acute coronary syndrome in diabetes a randomized, controlled trial. Annals of Internal
patients with type 2 diabetes. N Engl J Med. 2013;369(14): Medicine. 2011;154(2):103‐+. doi: 10.7326/0003‐4819‐
1327–1335. Epub 2013/09/03. doi: 10.1056/NEJMoa1305889. 154‐2‐201101180‐00300. PubMed PMID: WOS:000286
PubMed PMID: 23992602. 223500004.
34. Green JB, Bethel MA, Armstrong PW, Buse JB, Engel SS, 41. Guja C, Frias JP, Somogyi A, Jabbour S, Wang H, Hardy E,
Garg J, Josse R, Kaufman KD, Koglin J, Korn S, Lachin JM, Rosenstock J. Effect of exenatide QW or placebo, both added
McGuire DK, Pencina MJ, Standl E, Stein PP, Suryawanshi S, to titrated insulin glargine, in uncontrolled type 2 diabetes:
Van de Werf F, Peterson ED, Holman RR, for the TECOS the DURATION‐7 randomized study. Diabetes Obes Metab.
Study Group. Effect of sitagliptin on cardiovascular outcomes 2018;20(7):1602–1614. Epub 2018/02/24. doi: 10.1111/
in type 2 diabetes. N Engl J Med. 2015;373(3):232–242. Epub dom.13266. PubMed PMID: 29473704; PMCID. PMC6
2015/06/09. doi: 10.1056/NEJMoa1501352. PubMed PMID: 032936.
26052984. 42. Buse JB, Rosenstock J, Sesti G, Schmidt WE, Montanya E,
35. Buse JB, Henry RR, Han J, Kim DD, Fineman MS, Baron Brett JH, Zychma M, Blonde L, Group L‐S. Liraglutide once
AD, for the Exenatide‐113 Clinical Study Group. Effects of a day versus exenatide twice a day for type 2 diabetes: a 26‐
exenatide (exendin‐4) on glycemic control over 30 weeks in week randomised, parallel‐group, multinational, open‐label
sulfonylurea‐treated patients with type 2 diabetes. Diabetes trial (LEAD‐6). Lancet. 2009;374(9683):39–47. Epub 2009/
Care. 2004;27(11):2628–2635. Epub 2004/10/27. doi: 10.2337/ 06/12. doi: 10.1016/S0140‐6736(09)60659‐0. PubMed PMID:
diacare.27.11.2628. PubMed PMID: 15504997. 19515413.
36. Kendall DM, Riddle MC, Rosenstock J, Zhuang D, Kim DD, 43. Buse JB, Nauck M, Forst T, Sheu WH, Shenouda SK,
Fineman MS, Baron AD. Effects of exenatide (exendin‐4) on Heilmann CR, Hoogwerf BJ, Gao A, Boardman MK,
glycemic control over 30 weeks in patients with type 2 diabe Fineman M, Porter L, Schernthaner G. Exenatide once weekly
tes treated with metformin and a sulfonylurea. Diabetes versus liraglutide once daily in patients with type 2 diabetes
Care. 2005;28(5):1083–1091. Epub 2005/04/28. doi: 10.2337/ (DURATION‐6): a randomised, open‐label study. Lancet.
diacare.28.5.1083. PubMed PMID: 15855571. 2013;381(9861):117–124. Epub 2012/11/13. doi: 10.1016/
37. Zinman B, Hoogwerf BJ, Duran Garcia S, Milton DR, S0140‐6736(12)61267‐7. PubMed PMID: 23141817.
Giaconia JM, Kim DD, Trautmann ME, Brodows RG. The 44. Buse JB, Vilsboll T, Thurman J, Blevins TC, Langbakke IH,
effect of adding exenatide to a thiazolidinedione in subopti Bottcher SG, Rodbard HW, Investigators NNT. Contribution
mally controlled type 2 diabetes: a randomized trial. Ann of liraglutide in the fixed‐ratio combination of insulin deglu
Intern Med. 2007;146(7):477–485. Epub 2007/04/04. doi: 10. dec and liraglutide (IDegLira). Diabetes Care. 2014;37(11):
7326/0003‐4819‐146‐7‐200704030‐00003. PubMed PMID: 2926–2933. Epub 2014/08/13. doi: 10.2337/dc14‐0785.
17404349. PubMed PMID: 25114296.
45. Rosenstock J, Raccah D, Koranyi L, Maffei L, Boka G, diabetes (SUSTAIN 1): a double‐blind, randomised, placebo‐
Miossec P, Gerich JE. Efficacy and safety of lixisenatide once controlled, parallel‐group, multinational, multicentre phase
daily versus exenatide twice daily in type 2 diabetes inade 3a trial. Lancet Diabetes Endocrinol. 2017;5(4):251–260. Epub
quately controlled on metformin: a 24‐week, randomized, 2017/01/24. doi: 10.1016/S2213‐8587(17)30013‐X. PubMed
open‐label, active‐controlled study (GetGoal‐X). Diabetes PMID: 28110911.
Care. 2013;36(10):2945–2951. Epub 2013/05/24. doi: 10.2337/ 52. Ahmann AJ, Capehorn M, Charpentier G, Dotta F, Henkel
dc12‐2709. PubMed PMID: 23698396; PMCID. PMC E, Lingvay I, Holst AG, Annett MP, Aroda VR. Efficacy and
3781502. safety of once‐weekly semaglutide versus exenatide ER in
46. Nauck M, Rizzo M, Johnson A, Bosch‐Traberg H, Madsen J, subjects with type 2 diabetes (SUSTAIN 3): a 56‐week, open‐
Cariou B. Once‐daily liraglutide versus lixisenatide as add‐ label, randomized clinical trial. Diabetes Care. 2018;41(2):258–
on to metformin in type 2 diabetes: a 26‐week randomized 266. Epub 2017/12/17. doi: 10.2337/dc17‐0417. PubMed
controlled clinical trial. Diabetes Care. 2016;39(9):1501–1509. PMID: 29246950.
doi: 10.2337/dc15‐2479. PubMed PMID: WOS:000383663 53. Pratley RE, Aroda VR, Lingvay I, Ludemann J, Andreassen
000017. C, Navarria A, Viljoen A, investigators S. Semaglutide versus
47. Rosenstock J, Guerci B, Hanefeld M, Gentile S, Aronson R, dulaglutide once weekly in patients with type 2 diabetes
Tinahones FJ, Roy‐Duval C, Souhami E, Wardecki M, Ye J, (SUSTAIN 7): a randomised, open‐label, phase 3b trial.
Perfetti R, Heller S, on behalf of the GetGoal Duo‐2 Trial Lancet Diabetes Endocrinol. 2018;6(4):275–286. Epub 2018/
Investigators. Prandial options to advance basal insulin glar 02/06. doi: 10.1016/S2213‐8587(18)30024‐X. PubMed PMID:
gine therapy: testing lixisenatide plus basal insulin versus 29397376.
insulin glulisine either as basal‐plus or basal‐bolus in type 2 54. Rodbard HW, Lingvay I, Reed J, de la Rosa R, Rose L,
diabetes: the GetGoal Duo‐2 Trial. Diabetes Care. 2016;39(8): Sugimoto D, Araki E, Chu PL, Wijayasinghe N, Norwood P.
1318–1328. Epub 2016/05/26. doi: 10.2337/dc16‐0014. PubMed Semaglutide added to basal insulin in type 2 diabetes
PMID: 27222510. (SUSTAIN 5): a randomized, controlled trial. J Clin
48. Wysham C, Blevins T, Arakaki R, Colon G, Garcia P, Atisso C, Endocrinol Metab. 2018;103(6):2291–2301. Epub 2018/04/25.
Kuhstoss D, Lakshmanan M. Efficacy and safety of dulaglutide doi: 10.1210/jc.2018‐00070. PubMed PMID: 29688502; PMCID.
added onto pioglitazone and metformin versus exenatide in PMC5991220.
type 2 diabetes in a randomized controlled trial (AWARD‐1). 55. Vilsboll T, Bain SC, Leiter LA, Lingvay I, Matthews D, Simo
Diabetes Care. 2014;37(8):2159–2167. Epub 2014/06/01. doi: R, Helmark IC, Wijayasinghe N, Larsen M. Semaglutide,
10.2337/dc13‐2760. PubMed PMID: 24879836. reduction in glycated haemoglobin and the risk of diabetic
49. Dungan KM, Povedano ST, Forst T, Gonzalez JG, Atisso C, retinopathy. Diabetes Obes Metab. 2018;20(4):889–897. Epub
Sealls W, Fahrbach JL. Once‐weekly dulaglutide versus once‐ 2017/11/28. doi: 10.1111/dom.13172. PubMed PMID: 2917
daily liraglutide in metformin‐treated patients with type 2 dia 8519; PMCID. PMC5888154.
betes (AWARD‐6): a randomised, open‐label, phase 3, non‐ 56. Marguet D, Baggio L, Kobayashi T, Bernard AM, Pierres M,
inferiority trial. Lancet. 2014;384(9951):1349–1357. Epub 2014/ Nielsen PF, Ribel U, Watanabe T, Drucker DJ, Wagtmann N.
07/16. doi: 10.1016/S0140‐6736(14)60976‐4. PubMed PMID: Enhanced insulin secretion and improved glucose tolerance
25018121. in mice lacking CD26. Proc Natl Acad Sci USA. 2000;97(12):
50. Pozzilli P, Norwood P, Jodar E, Davies MJ, Ivanyi T, Jiang 6874–6879. Epub 2000/05/24. doi: 10.1073/pnas.120069197.
HH, Woodward DB, Milicevic Z. Placebo‐controlled, rand PubMed PMID: 10823914; PMCID. PMC18768.
omized trial of the addition of once‐weekly glucagon‐like 57. Vella A, Bock G, Giesler PD, Burton DB, Serra DB, Saylan
peptide‐1 receptor agonist dulaglutide to titrated daily insu ML, Dunning BE, Foley JE, Rizza RA, Camilleri M. Effects
lin glargine in patients with type 2 diabetes (AWARD‐9). of dipeptidyl peptidase‐4 inhibition on gastrointestinal
Diabetes Obesity & Metabolism. 2017;19(7):1024–1031. doi: function, meal appearance, and glucose metabolism in
10.1111/dom.12937. PubMed PMID: WOS:0004032366 type 2 diabetes. Diabetes. 2007;56(5):1475–1480. Epub
00013. 2007/02/17. doi: 10.2337/db07‐0136. PubMed PMID:
51. Sorli C, Harashima SI, Tsoukas GM, Unger J, Karsbol JD, 17303799.
Hansen T, Bain SC. Efficacy and safety of once‐weekly sema 58. Bergman AJ, Stevens C, Zhou Y, Yi B, Laethem M, De Smet
glutide monotherapy versus placebo in patients with type 2 M, Snyder K, Hilliard D, Tanaka W, Zeng W, Tanen M, Wang
AQ, Chen L, Winchell G, Davies MJ, Ramael S, Wagner JA, 65. Del Prato S, Camisasca R, Wilson C, Fleck P. Durability of
Herman GA. Pharmacokinetic and pharmacodynamic the efficacy and safety of alogliptin compared with glipizide
properties of multiple oral doses of sitagliptin, a dipeptidyl in type 2 diabetes mellitus: a 2‐year study. Diabetes Obes
peptidase‐IV inhibitor: a double‐blind, randomized, pla Metab. 2014;16(12):1239–1246. Epub 2014/08/19. doi:
cebo‐controlled study in healthy male volunteers. Clin Ther. 10.1111/dom.12377. PubMed PMID: 25132212.
2006;28(1):55–72. Epub 2006/02/24. doi: 10.1016/j. 66. Goke B, Gallwitz B, Eriksson JG, Hellqvist A, Gause‐Nilsson
clinthera.2006.01.015. PubMed PMID: 16490580. I. Saxagliptin vs. glipizide as add‐on therapy in patients with
59. Ahren B, Landin‐Olsson M, Jansson PA, Svensson M, type 2 diabetes mellitus inadequately controlled on met
Holmes D, Schweizer A. Inhibition of dipeptidyl peptidase‐4 formin alone: long‐term (52‐week) extension of a 52‐week
reduces glycemia, sustains insulin levels, and reduces gluca randomised controlled trial. Int J Clin Pract. 2013;67(4):307–
gon levels in type 2 diabetes. J Clin Endocrinol Metab. 316. Epub 2013/05/03. doi: 10.1111/ijcp.12119. PubMed
2004;89(5):2078–2084. Epub 2004/05/06. doi: 10.1210/ PMID: 23638466.
jc.2003‐031907. PubMed PMID: 15126524. 67. Seck T, Nauck M, Sheng D, Sunga S, Davies MJ, Stein PP,
60. Herman GA, Bergman A, Stevens C, Kotey P, Yi B, Zhao P, Kaufman KD, Amatruda JM, Sitagliptin Study G. Safety and
Dietrich B, Golor G, Schrodter A, Keymeulen B, Lasseter efficacy of treatment with sitagliptin or glipizide in patients
KC, Kipnes MS, Snyder K, Hilliard D, Tanen M, Cilissen C, with type 2 diabetes inadequately controlled on metformin:
De Smet M, de Lepeleire I, Van Dyck K, Wang AQ, Zeng W, a 2‐year study. Int J Clin Pract. 2010;64(5):562–576. Epub
Davies MJ, Tanaka W, Holst JJ, Deacon CF, Gottesdiener 2010/05/12. doi: 10.1111/j.1742‐1241.2010.02353.x.
KM, Wagner JA. Effect of single oral doses of sitagliptin, a PubMed PMID: 20456211.
dipeptidyl peptidase‐4 inhibitor, on incretin and plasma glu 68. Filozof C, Gautier JF. A comparison of efficacy and safety of
cose levels after an oral glucose tolerance test in patients vildagliptin and gliclazide in combination with metformin
with type 2 diabetes. J Clin Endocrinol Metab. in patients with Type 2 diabetes inadequately controlled
2006;91(11):4612–4619. Epub 2006/08/17. doi: 10.1210/ with metformin alone: a 52‐week, randomized study. Diabet
jc.2006‐1009. PubMed PMID: 16912128. Med. 2010;27(3):318–326. Epub 2010/06/12. doi:
61. Nauck M. Incretin therapies: highlighting common features 10.1111/j.1464‐5491.2010.02938.x. PubMed PMID:
and differences in the modes of action of glucagon‐like pep 20536495.
tide‐1 receptor agonists and dipeptidyl peptidase‐4 inhibi 69. Matthews DR, Dejager S, Ahren B, Fonseca V, Ferrannini E,
tors. Diabetes Obes Metab. 2016;18(3):203–216. Epub Couturier A, Foley JE, Zinman B. Vildagliptin add‐on to
2015/10/23. doi: 10.1111/dom.12591. PubMed PMID: metformin produces similar efficacy and reduced hypogly
26489970; PMCID. PMC4785614. caemic risk compared with glimepiride, with no weight gain:
62. Scheen AJ. Effects of glucose‐lowering agents on surrogate results from a 2‐year study. Diabetes Obes Metab.
endpoints and hard clinical renal outcomes in patients with 2010;12(9):780–789. Epub 2010/07/24. doi:
type 2 diabetes. Diabetes Metab. 2019;45(2):110–121. Epub 10.1111/j.1463‐1326.2010.01233.x. PubMed PMID:
2018/11/28. doi: 10.1016/j.diabet.2018.10.003. PubMed 20649630.
PMID: 30477733. 70. Liu SC, Chien KL, Wang CH, Chen WC, Leung CH. Efficacy
63. Deacon CF, Holst JJ. Dipeptidyl peptidase‐4 inhibitors for and safety of adding pioglitazone or sitagliptin to patients
the treatment of type 2 diabetes: comparison, efficacy and with type 2 diabetes insufficiently controlled with met
safety. Expert Opin Pharmacother. 2013;14(15):2047–2058. formin and a sulfonylurea. Endocr Pract. 2013;19(6):980–
Epub 2013/08/08. doi: 10.1517/14656566.2013.824966. 988. Epub 2013/06/29. doi: 10.4158/EP13148.OR. PubMed
PubMed PMID: 23919507. PMID: 23807528.
64. Gallwitz B, Rosenstock J, Rauch T, Bhattacharya S, Patel S, 71. Takihata M, Nakamura A, Tajima K, Inazumi T, Komatsu Y,
von Eynatten M, Dugi KA, Woerle HJ. 2‐year efficacy and Tamura H, Yamazaki S, Kondo Y, Yamada M, Kimura M,
safety of linagliptin compared with glimepiride in patients Terauchi Y. Comparative study of sitagliptin with pioglita
with type 2 diabetes inadequately controlled on metformin: zone in Japanese type 2 diabetic patients: the COMPASS
a randomised, double‐blind, non‐inferiority trial. Lancet. randomized controlled trial. Diabetes Obes Metab.
2012;380(9840):475–483. Epub 2012/07/04. doi: 10.1016/ 2013;15(5):455–462. Epub 2013/01/03. doi: 10.1111/
S0140‐6736(12)60691‐6. PubMed PMID: 22748821. dom.12055. PubMed PMID: 23279373.
72. Bae J, Kim G, Lee YH, Lee BW, Kang ES, Cha BS. Differential TIMI 53 Steering Committee and Investigators. Saxagliptin
effects of thiazolidinediones and dipeptidyl pepti and cardiovascular outcomes in patients with type 2 diabetes
dase‐4 inhibitors on insulin resistance and beta‐cell function mellitus. N Engl J Med. 2013;369(14):1317–1326. Epub 2013/
in type 2 diabetes mellitus: a propensity score‐matched anal 09/03. doi: 10.1056/NEJMoa1307684. PubMed PMID: 239
ysis. Diabetes Ther. 2019;10(1):149–158. Epub 2018/12/07. 92601.
doi: 10.1007/s13300‐018‐0541‐y. PubMed PMID: 30506494; 79. Schnabel CA, Wintle M, Kolterman O. Metabolic effects of
PMCID. PMC6349276. the incretin mimetic exenatide in the treatment of type 2
73. Emerging Risk Factors C, Sarwar N, Gao P, Seshasai SR, diabetes. Vasc Health Risk Manag. 2006;2(1):69–77. Epub
Gobin R, Kaptoge S, Di Angelantonio E, Ingelsson E, Lawlor 2007/02/27. doi: 10.2147/vhrm.2006.2.1.69. PubMed PMID:
DA, Selvin E, Stampfer M, Stehouwer CD, Lewington S, 17319471; PMCID. PMC1993968.
Pennells L, Thompson A, Sattar N, White IR, Ray KK, 80. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA.
Danesh J. Diabetes mellitus, fasting blood glucose concen 10‐year follow‐up of intensive glucose control in type 2 dia
tration, and risk of vascular disease: a collaborative meta‐ betes. N Engl J Med. 2008;359(15):1577–1589. Epub 2008/
analysis of 102 prospective studies. Lancet. 2010;375(9733): 09/12. doi: 10.1056/NEJMoa0806470. PubMed PMID: 1878
2215–2222. Epub 2010/07/09. doi: 10.1016/S0140‐6736(10) 4090.
60484‐9. PubMed PMID: 20609967; PMCID. PMC2904878. 81. American Diabetes Association. 9. Pharmacologic
74. Home P. Cardiovascular outcome trials of glucose‐lowering approaches to glycemic treatment: standards of medical care
medications: an update. Diabetologia. 2019;62(3):357–369. in diabetes – 2020. Diabetes Care. 2020;43(Suppl 1):S98–S110.
Epub 2019/01/05. doi: 10.1007/s00125‐018‐4801‐1. PubMed Epub 2019/12/22. doi: 10.2337/dc20‐S009. PubMed PMID:
PMID: 30607467. 31862752.
75. Wang Q, Liu L, Gao L, Li Q. Cardiovascular safety of GLP‐1 82. Drucker DJ, Buse JB, Taylor K, Kendall DM, Trautmann M,
receptor agonists for diabetes patients with high cardiovas Zhuang D, Porter L, for the DURATION‐1 Study Group.
cular risk: a meta‐analysis of cardiovascular outcomes trials. Exenatide once weekly versus twice daily for the treatment
Diabetes Res Clin Pract. 2018;143:34–42. Epub 2018/06/24. of type 2 diabetes: a randomised, open‐label, non‐inferiority
doi: 10.1016/j.diabres.2018.06.009. PubMed PMID: 29935211. study. Lancet. 2008;372(9645):1240–1250. Epub 2008/09/11.
76. American Diabetes A. 10. Cardiovascular disease and risk doi: 10.1016/S0140‐6736(08)61206‐4. PubMed PMID: 18
management: standards of medical care in diabetes – 2020. 782641.
Diabetes Care. 2020;43(Suppl 1):S111–S34. Epub 2019/12/22. 83. Blevins T, Pullman J, Malloy J, Yan P, Taylor K, Schulteis C,
doi: 10.2337/dc20‐S010. PubMed PMID: 31862753. Trautmann M, Porter L. DURATION‐5: exenatide once
77. Rosenstock J, Perkovic V, Johansen OE, Cooper ME, Kahn weekly resulted in greater improvements in glycemic control
SE, Marx N, Alexander JH, Pencina M, Toto RD, Wanner C, compared with exenatide twice daily in patients with type 2
Zinman B, Woerle HJ, Baanstra D, Pfarr E, Schnaidt S, diabetes. J Clin Endocrinol Metab. 2011;96(5):1301–1310.
Meinicke T, George JT, von Eynatten M, McGuire DK, Epub 2011/02/11. doi: 10.1210/jc.2010‐2081. PubMed PMID:
Investigators C. Effect of linagliptin vs placebo on major car 21307137.
diovascular events in adults with type 2 diabetes and high 84. Pratley RE, Nauck MA, Barnett AH, Feinglos MN, Ovalle F,
cardiovascular and renal risk: the CARMELINA rand Harman‐Boehm I, Ye J, Scott R, Johnson S, Stewart M,
omized clinical trial. JAMA. 2019;321(1):69–79. Epub 2018/ Rosenstock J, group Hs. Once‐weekly albiglutide versus
11/13. doi: 10.1001/jama.2018.18269. PubMed PMID: 3041 once‐daily liraglutide in patients with type 2 diabetes inad
8475; PMCID. PMC6583576. equately controlled on oral drugs (HARMONY 7): a ran
78. Scirica BM, Bhatt DL, Braunwald E, Steg PG, Davidson J, domised, open‐label, multicentre, non‐inferiority phase 3
Hirshberg B, Ohman P, Frederich R, Wiviott SD, Hoffman study. Lancet Diabetes Endocrinol. 2014;2(4):289–297. Epub
EB, Cavender MA, Udell JA, Desai NR, Mosenzon O, 2014/04/08. doi: 10.1016/S2213‐8587(13)70214‐6. PubMed
McGuire DK, Ray KK, Leiter LA, Raz I, for the SAVOR‐ PMID: 24703047.
PART III
Diagnosis and Management
of Cardiovascular Risk Factors
and Cardiovascular Disease
13 Screening Patients with Prediabetes
and Diabetes for Cardiovascular Disease
Sabreen Ahmed, Saritha Tirumalasetty and Vivian Fonseca
Section of Endocrinology, Tulane University School of Medicine, New Orleans, LA, USA
threshold of fasting glucose to ≥ 126 mg/dL in 1997 and the

addition of hemoglobin A1c (HbA1c) ≥ 6.5% in 2010 [2, 4, 5].
●● Patients with dysglycemia are at risk of developing
CVD continues to be a significant cause of mortality
cardiovascular disease before and after the develop- worldwide and in those with DM [6].
ment of overt diabetes. Patients with prediabetes in the San Antonio Heart
●● Risk factor assessment will help stratify risk and guide Study had atherogenic risk factors suggesting that risk for
treatment. CVD may start before the diagnosis of diabetes [7]. In a
●● Appropriate screening likely requires a combination of
risk factor assessment in combination with appropriate
study of Finnish patients with and without diabetes fol-
functional or imaging studies intended to assess lowed for seven years, the risk of a recurrent event was
cardiovascular risk. greatest in diabetic patients with a known history of
●● Screening tests are not recommended for people with CVD [8]. The risk was similar in patients with diabetes and
very high risk or very low risk. However, they may be no history of CVD and patients without diabetes who had
useful for people at intermediate risk allowing
re-stratification and appropriate therapeutic approaches.
a history of known CVD. This gave rise to the notion of
“cardiac risk equivalent” and the need to evaluate and treat
patients with diabetes in the same way as those with pre‐
Introduction existing heart disease. However, more recent data suggest
Cardiovascular disease (CVD) continues to be the leading that risk varies depending on several factors including age
cause of death accounting for up to 31% (17.9 million) of and duration and control of diabetes. Therefore, predicting
global deaths [1]. The majority of this mortality is due to risk of events or screening for CVD in asymptomatic
myocardial infarction and strokes. The increasing preva- patients may be an important strategy in appropriate pre-
lence correlates to a higher amount of procedural care. In vention of CV events.
the United States alone, the economic burden has doubled The aim of this review is to discuss optimal identifica-
from 1996–1997 to 2014–2015 at which time direct costs tion of patients at risk, current recommendations for such
were estimated to be up to $213.8 billion [2]. People with identification, recent developments, and possible future
diabetes mellitus (DM) have been shown to have twice the directions in such efforts.
risk of CVD as the general population [3]. The CDC’s
National Health and Nutrition Examination Survey
Guidelines
(NHANES) has shown an increase in the prevalence of
diagnosed and undiagnosed DM in both men and women Within guidelines published for the United States there is
from 1988–1994 to 2013 to 2016 which corresponds to the variability in screening recommendations with the
changes in diagnostic testing for DM namely the lowered American Diabetes Association (ADA) at this time not rec-

163
164 Diagnosis and Management of Cardiovascular Risk Factors and Cardiovascular Disease
ommending CVD screening in diabetics without symp- In contrast, possible utility in screening for cardiovascu-
toms and the American College of Cardiology/American lar disease may influence prevention treatments and deter-
Heart Association (ACC/AHA) as well as ACC‐Imaging mining risk for CVD events. The PREDICT and MESA
Council proclaiming possible usefulness in screening [9, studies in particular showed a direct correlation between
10, 11]. The ADA guidance is based on findings of the high coronary artery calcium (CAC) scores and CVD risk
Detection of Ischemia in Asymptomatic Diabetics (DIAD) for both non‐diabetic and diabetic patients. Both studies
study, which describes resolution to ischemia on follow‐up also may help with revealing low risk patients with very low
perfusion imaging. DIAD randomized asymptomatic CAC scores requiring avoidance of invasive therapies.
patients into “no imaging” or “cardiac perfusion imaging” These trials will be discussed in further depth in the CAC
groups. Of all the patients who underwent perfusion imag- section.
ing, 22% had abnormal testing. Of these, 40% had less than
two risk factors for CVD. Three years later when the testing
Hyperglycemia and risks of CVD
was repeated, of those that had a negative test 10% now had
a positive test [12]. A more striking finding was the resolu- Glucose is a continuous variable in the population with an
tion of inducible ischemia in 79% of those that had almost linear association with CVD. This is seen even in
ischemia initially and no intervening revascularization glucose concentrations below that used to diagnose diabe-
procedure. On follow‐up, 4.8 years after the original study, tes. DECODE, an epidemiologic study in Europe, looked at
the incidence of CV events was low. There was a significant over 25 000 subjects and demonstrated that mortality was
increase in primary prevention efforts in both groups and increased not just in patients with diabetes but also in those
there was no difference in mortality between those that had with impaired glucose tolerance (IGT). The study also
myocardial perfusion testing and those that did not [13]. It demonstrated that IGT carries a greater risk than impaired
is important to recognize that the DIAD cohort had very fasting glucose ((IFG), a finding of practical importance as
good control of CV risk factors, perhaps better than the glucose tolerance testing is often not carried out in patients
general population with diabetes. at possible risk [16]. Norhammer et al. demonstrated that
Aggressive risk factor control is supported by the in patients with established coronary artery disease (CAD),
Courage trial, which randomized patients into percutane- IGT is common and may be missed [17]. This has led to
ous coronary intervention (PCI) with medical therapy vs recommendations for more widespread use of GTTs.
medical therapy alone in order to gauge changes in future Therefore, patients at risk for dysglycemia should be
cardiovascular event rates. At 5‐year follow‐up both study screened early and have risk factor management early and
groups met similar targets in pressure control < 130/85, those with CVD should be screened for dysglycemia.
low‐density lipoprotein (LDL) < 85, and HbA1c goal of 7%. In patients with known diabetes, treatment of hypergly-
Rates of primary outcomes including death from any cause cemia reduces the risk of CVD Early prospective studies
and non‐fatal myocardial infarctions (MI) were about such as DCCT and UKPDS [18, 19] showed improvement
equal [14]. It is arguable that screening may lead to invasive in microvascular complications in patients with type 1 and
intervention which may be unnecessary in those with good type 2 diabetes, respectively. When these patients were fol-
medical treatment. Even further secondary outcomes of lowed, a decrease in macrovascular complications was also
event rates were also similar despite the treatment group. demonstrated [20, 21]. The EDIC trial [20] describes long‐
The BARI 2D study group produced similar findings with term follow‐up of the patients in DCCT trial. In EDIC,
no significant differences in all‐cause mortality or major lower HbA1c while the patients were actively enrolled in
cardiac events even within different DM management the DCCT trial appeared to correlate with improved car-
styles except for patients within the CABG arm [15]. Those diovascular outcomes. It is important to recognize that the
who underwent coronary artery bypass surgery (CABG) DCCT and UKPDS trials enrolled relatively young patients
had fewer major cardiovascular events when DM manage- with short duration of DM. In addition, after the original
ment included insulin sensitization. DCCT and UKPDS trials were terminated, the patients no
Screening Asymptomatic Patients with Prediabetes and Diabetes for Cardiovascular Disease 165
longer maintained their randomization. Despite this, there Intensive treatment patients had a 20% absolute reduction
appeared to be an improvement in CVD outcomes in of CV death, nonfatal MI, CABG, PCI, peripheral vascular
intensively treated patients during long‐term follow‐up. In disease, cerebrovascular accident (CVA) and amputations
three large studies (ACCORD, ADVANCE and VADT) after a mean of 7.8 years as well as a significant reduction in
where the patients studied were older than those in UKPDS microvascular complications. In a follow‐up study after 5.5
and had DM for a longer duration tight blood sugar control years where all patients were told to follow intensive treat-
did not improve CV outcomes in the short term. Those ment regimen, the initial intensive therapy group had
with long‐standing DM may have already developed irre- fewer CV deaths and showed the importance of early risk
versible and progressive atherosclerosis and management factor reduction.
of non‐glycemic factors takes on more importance.
The evidence supports the notion that elevated blood
Diabetes‐specific clinical risk predictors of
sugars do increase risk for CVD and control of hyperglyce-
CVD
mia does improve that risk. The control of other risk fac-
tors also appears to be important, especially in those with Individuals with diabetes have increased risk of cardiovas-
long‐standing type 2 diabetes. cular disease even when hyperglycemia is controlled [24].
A recent large observational cohort study evaluated 32611
patients with diabetes type 1 over 10.4 years and 17 risk
Traditional risk factors
factors for MI, CVA, heart failure (HF) and death [25]. At
Several risk factors have been associated with CVD. The the end of the study, the strongest predictors of these end
INTERHEART study evaluated a number of these risks points were albuminuria, glycohemoglobin, duration of
and tried to assess the contribution of these risk factors to diabetes, systolic blood pressure (SBP), and LDL choles-
the presence of CVD [22]. Nine commonly recognized risk terol. Estimated glomerular filtration rate was also one of
factors (smoking, abnormal lipids (apolipoprotein B1/ the most important risk factors of HF hospitalization.
apolipoprotein A1 ratio), hypertension (self‐reported), Lower HbA1c, SBP, and LDL levels had a linear correlation
diabetes, abdominal obesity, psychosocial factors) were with the outcomes other than HbA1c and all‐cause mortal-
evaluated. The percent attributable risk (PAR) for each risk ity, which raises the question of whether with tighter con-
factor was calculated. In other words, these risks accounted trol the current guidelines might be beneficial.
for 90% of the risk for CVD. The vast majority of risk is Patients with diabetes are screened for nephropathy
accounted for by what have come to be termed “tradi- with albumin‐to‐creatinine ratio. There are many trials
tional” risk factors. that demonstrate moderate albuminuria ≥ 30 mg/g of cre-
The COURAGE and BARI 2D trials as described previ- atinine is a risk factor for CVD, although the mechanism is
ously showed that there was no significant difference in not well understood. As seen in the Heart Outcomes
death, non‐fatal MI and other cardiovascular event rates Prevention Evaluation (HOPE) trial, moderate albuminu-
between intensive medical therapy alone and medical ther- ria was associated with MI, CV death, CVA and HF hospi-
apy plus PCI, although the patient populations had estab- talizations even in patients without diabetes [26]. Of the
lished CAD [14, 15]. When it comes to intensive or 1140 patients with diabetes and microalbuminuria, the use
conventional medical therapy, the Steno‐2 trial suggests of an ACE‐I reduced albuminuria and a 21% decrease in
that intensive treatment of multiple risk factors reduced the CV outcomes according to the Microalbuminuria,
risk of vascular events [23]. 160 patients with diabetes and Cardiovascular, and Renal Outcomes (MICRO)‐HOPE
microalbuminuria received conventional or intensive ther- substudy [27]. This held true after adjusting for changes in
apy, which consisted of targets of HbA1c < 6.5%, blood blood pressure.
pressure < 140/85 mmHg and later < 130/80 mmHg, tri- Peripheral artery disease (PAD) is an indication that
glycerides < 150 mg/dl, total cholesterol < 190 mg/dl, angi- there might be atherosclerosis in other areas of the body
otensin‐converting enzyme inhibitor (ACE‐I) and aspirin. and diabetes mellitus is a known risk factor. The UKPDS
showed the prevalence correlated with longer diabetes event risk and developing treatment plans for said patients.
duration, increased diabetes severity and female, African In the 2004 Raggi et al. study referred > 10 000 patients
American, and Hispanic patients with diabetes [19]. (903 of them with DM) for EBT to determine CAC scores.
Theoretically, identifying PAD in asymptomatic patients In a 5‐year follow‐up all‐cause mortality was higher in the
with diabetes would benefit from aggressive risk factor DM population, which also had higher CAC scores [29].
modification. The ADA currently recommends that a Of significance, the DM population in Raggi’s study also
screening ankle brachial index be done if patients with dia- had older patients with more hypertension and tobacco use
betes over the age of 50 every 5 years. However, a rand- than those without DM.
omized clinical trial is needed to determine which patients In the 2008 PREDICT study 589 patients with DM type
with diabetes to screen and if screening in these asympto- 2 and no history of CVD from Central and West London
matic patients would be clinically beneficial. had CAC scores measured and were followed up annually
for 4 years. Black African patients were excluded in this
study due to low coronary heart disease (CHD) rates at the
Nontraditional risk factors
time of this study. In analysis of CAC scores and primary
Although the majority of CVD appears to be explained by endpoints, there is a logarithmic association between
traditional risk factors, there is residual risk that may be increasing CAC scores and first CHD or stroke event in
due to the presence of other factors. These have been these patients [30]. Furthermore when taking into account
termed “non‐traditional” risk factors. These may be useful risk calculator models, CAC scores seem to further enhance
in determining risk in patients who by history would be at the prediction models of Framingham and UKPDS.
high risk for CVD but have normal values for the tradi- More recently the Multi‐Ethnic Study of Atherosclerosis
tional risk factors. As our understanding of the mecha- (MESA) study reviewed a large population of varying eth-
nisms of CVD in diabetes has improved, the role of nicities including Caucasian, African American, Hispanic,
inflammation in development of both DM and CVD has and Chinese from 6 communities across the United States.
been recognized. In addition, abnormal fibrinolysis, Over 6600 people with metabolic syndrome (MetS) with-
endothelial function, alteration in adipokine release, and out DM, DM, and individuals without MetS or DM were
vascular wall abnormalities may all play a role in vascular assessed for CAC (using EBT vs CT) and Carotid Intimal
dysfunction in patients with diabetes [28]. As these mecha- Medial Thickness (CIMT) using ultrasound. After about a
nisms are investigated, novel risk factors have been identi- 6.4‐year follow‐up CHD and CVD endpoints were
fied. A number of interventions we institute for traditional assessed. They had a median of 6.4‐year follow‐up. CAC
risk factors such as statins and ACE‐I appear to improve with traditional risk factors were found to significantly be
these nontraditional risk factors. predictive of CHD and CVD events at all CAC levels with
Although promising, the nontraditional factors are not increased events at higher CAC scores [31]. Although
ready for routine use in all patients. There are multiple reasons CIMT did show an increase in events over time this
for this, such as lack of standardized commercially available increase was not statistically significant except for in CHD
tests, lack of data on CVD event reduction following risk events in the 4th quartile for patients without MetS or DM.
improvement etc. Of all the novel risks, high‐sensitivity C‐ Of note in diabetic patients a doubling of events occurred
reactive protein (hsCRP) and coronary calcium scores may be when CAC scores reached above 400. Further follow‐up
the only ones that are currently clinically applicable. after 11–12 years was done within the MESA study. Patients
with DM and MetS had significantly higher rates of CHD
and ASCVD events compared to DM alone, MetS alone,
Screening tools
and those with neither MetS nor DM [32]. It is notable that
Coronary artery calcium score those with DM alone had more of these events than MetS
As described in the ACC/AHA and ACC‐Cardiac Imaging alone; those with neither condition had the lowest event
guidelines, CAC may be useful in ascribing cardiovascular rate. While specifically looking at the range of CAC scores,
those with higher scores specifically CAC > 400 showed found to have severe stenosis. All patients with ≥ 10% ste-
the same trend in event rates. Upon extrapolation of the nosis received aggressive therapy (SBP < 120 mmHg, LDL
Diabetic population data with CAC scores > 400, those < 70 mg/dL, HDL > 50 mg/dL, triglyceride < 150 mg/dL,
with higher CHD/ASCVD event rates were found in and HbA1c < 6.0%). Patients with severe stenosis ≥ 70%
patients with long‐term insulin use and DM duration of had diagnostic coronary angiography and patients with
over 10 years. Furthermore DM patients with a CAC score moderate stenosis ≥ 50% received stress cardiac imaging
of 0 had less CHD and ASCVD events compared to CAC of and then angiography if needed. This resulted in 8% of
> 0 with Hazard Ratios of 0.35 and 0.43 respectively. patients who had CCTA undergoing angiography and 5.8%
Interestingly DM patients with CAC scores of > 400 and needing revascularization. After a mean follow‐up of 4 ±
tight control of HbA1c to < 7 were found to have higher 1.7 years, there was no statistically significant difference in
CHD/ASCVD rates. primary event outcomes (all‐cause mortality, nonfatal
Thus, it appears that CAC is a highly predictive for coro- myocardial infarction and hospitalization for unstable
nary atherosclerosis, and has emerged as the primary rec- angina) in the CCTA arm vs. standard care, 6.2% and 7.6%
ommended test to evaluate risk for future atherosclerotic respectively.
cardiovascular disease, as it is an easy to perform, cost CCTA may be valuable in providing long‐term prognos-
effective and reproducible means of quantifying plaque in tic information like CAC for asymptomatic patients with
the coronary arteries. As discussed above, the ACC and diabetes as demonstrated in several studies. An observa-
AHA cholesterol and prevention guidelines now include tional study of 591 patients with diabetes and without signs
CAC as part of the clinical algorithm in asymptomatic peo- or symptoms of CVD studied the event‐free survival at the
ple for planning primary prevention interventions such as end of a median of 5.3 years, defined as no cardiac death,
statins and aspirin. On the other hand for people with dia- nonfatal MI, unstable angina, or late revascularization [35].
betes such therapy is appropriate anyway and the test may After a CCTA, patients were categorized into no luminal
be less imperative. Nevertheless, when done it is important stenosis (28%), nonobstructive CAD (40%) and obstruc-
for clinicians and patients alike to be aware that a calcium tive CAD ≥ 50% (32%). The event‐free survivals were 99.3
score of zero indicates virtually no risk, and the recom- ± 0.7%, 96.7 ± 1.2% and 86.2 ± 3.0% (p < 0.001), respec-
mendation is to consider no statin, with re‐assessment in tively. Similarly, in a prospective study, 390 patients with
5–10 years. A CAC score of 1–99 favors statin (especially type 1 and 2 diabetes with suspected CAD (new chest pain,
after age 55) and a CAC of 100+ or > 75th percentile is an abnormal stress test, multiple cardiovascular risk factors)
indication to initiate statin therapy. had CCTA and were followed up after 62 ± 9 months for
end points of cardiac death, nonfatal MI and unstable
Coronary computed tomography angiography angina [36]. It also showed patients without any CAD had
The development of high‐resolution multidetector coro- an excellent prognosis and can help predict cardiac events.
nary computed tomography angiography (CCTA) is able to The event‐free survival was 100% for normal coronary
provide coronary anatomy noninvasively and determine arteries, 78% for nonobstructive CAD, and 60% for
the severity of atherosclerosis [33]. The CORE 64 study obstructive CAD. These conclusions are supported by
demonstrated that CCTA and invasive coronary angiogra- other studies [37, 38].
phy have a high correlation of determining CAD severity In a prospective multicenter observational cohort study
and extent [34]. In the FACTOR‐64 study, a randomized (400 of the patients had diabetes and were asymptomatic
clinical trial, 900 patients with DM type 1 or 2 of at least for CVD), CCTA findings were classified into per‐patient
3–5 years’ duration and without CAD were randomized to maximal stenosis, number of vessels with ≥ 50% stenosis
either CCTA screening or standard of care (SBP < and segment stenosis score [39]. It was determined that
130 mmHg, LDL < 100 mg/dL and HbA1c < 7.0%) [33]. Of increases in these were associated with incremental risk
the patients who received CCTA, 31% had no luminal dis- estimates of MACE after adjusting for CAD risk factors
ease, 58% had mild, 12% had moderate, and 11% were and CACS, improving risk classification. Although these
studies suggest that CCTA is predictive of adverse events, it while in the DE group, it was 11% and 5%, respectively.
does not address whether there is a benefit of invasive These findings were not significant. Furthermore, the posi-
revascularization in the obstructive group. tive predictive value of detecting significant CAD between
the two also was not significant. After 3 years, cardiovascu-
Stress testing – Myocardial perfusion imaging lar death and MI rates were comparable. Of note, 4% of all
The DIAD study noted no significant differences in CV the participants had significant CAD revealed on
events and mortality rates in patients receiving myocardial angiography.
perfusion screening versus no imaging [12, 13]. Similarly, Though promising, currently it is not recommended to
in the Do You Need to Assess Myocardial Ischemia in screen with rMPI or stress echo. Many studies have shown
Type‐2 diabetes (DYNAMIT) trial of 631 asymptomatic prevalence and diagnosis of CHD in diabetes patients.
patients with diabetes with at least two additional CV risk However, most prognostic studies were in patients who
factors, 21.5% of the patients screened with symptom‐lim- were symptomatic and more studies are needed in asymp-
ited bicycle exercise and radionuclide myocardial perfu- tomatic patients and of the prognosis after invention.
sion imaging (rMPI) had silent ischemia and after mean
follow‐up of 3.5 years, there was no significant difference Carotid ultrasound and endothelial function
in death from all‐cause, nonfatal MI, nonfatal stroke, or HF studies
requiring intervention [40]. The BARDOT trial took 400 CIMT was originally proposed as an alternative method to
asymptomatic type 2 diabetic patients and performed myo- determine cardiovascular risk without ionizing radiation.
cardial perfusion single‐photon emission computed It also had the advantage of being closely related to insulin
tomography (MPS) at the beginning of the study and after resistance [44]. However, concerns over reproducibility,
two years [41]. A substudy of patients with abnormal MPS and difficulty in obtaining high quality imaging in practice
were randomly assigned to medical versus invasive and have resulted in this test to a Class III (no benefit) recom-
medical treatment. As expected, intervention on a stenosed mendation from the ACC and AHA for general screening
lesion resulted in reduction of subclinical progression to of CVD risk. CIMT can therefore be used to assess cardio-
silent CAD and scintigraphic ischemia, though not rate or vascular risk in populations where risk estimation with tra-
MACE, cardiac death, and MI. ditional risk factors does not apply, such as populations
with specific genetic mutations in lipid metabolism.
Stress testing – Echocardiogram for CAD Younger patients (where CAC testing is not useful, due to
An open label randomized pilot study of 141 people with lack of time to develop calcification) with familial hyper-
type 2 diabetes without CVD either were in the control arm cholesterolemia (FH) represent a population in which
or had screening with an exercise electrocardiogram and CIMT has proven to be a valuable tool.
dipyridamole stress echocardiography (DSE) with subse- Similarly, methods for evaluating endothelial function
quent coronary angiography if either testing was abnor- (brachial artery reactivity) and tonometry etc. are useful in
mal [42]. 21.1% of patients had a positive DSE. After a research studies but less useful in clinical practice [45].
mean of 53.5 month, the absolute risk reduction in the
screened group was 15.8% and the study concluded that a Screening for HF
proportion of all cardiac events was significantly lower in Patients with diabetes have over two times increased risk
this group (P = 0.018). of HF compared to patients without DM [46]. In a cross‐
One randomized study compared the positive predictive sectional study in the Netherlands of 581 patients with
value of dobutamine stress echocardiography (DE) to type 2 diabetes and over 60 years old, 27.7% were found to
stress myocardial scintigraphy (SPECT) [43]. It consisted have HF that was not previously diagnosed [47]. 22.9%
of 204 individuals with diabetes type 2 without symptoms had preserved ejection fraction and 4.8% had reduced
but at high risk of CVD. Of the patients that underwent ejection fraction. However, the patients were not
SPECT, 13% had silent MI and 4% had significant CAD asymptomatic.
Currently, there are not any recommendations to screen Further studies are needed to recommend if NT‐
for HF in patients with diabetes. In a recent cluster analysis proBNP and microalbuminuria can identity patients with
based on baseline echocardiography variables, data from diabetes who may benefit for HF screening
two large prospective cohorts were used of 745 subjects
with type 2 diabetes without heart disease [48]. Cluster 1
Risk calculators
patients had the lowest hypertrophy, high LVEF and myo-
cardial strain. They were predominantly male with the low- Risk calculators have been used to stratify patients with
est comorbidity rates including obesity and HTN. Cluster CVD risk factors. Previous prediction tools include those
2 had the worst diastolic dysfunction, the patient popula- based on the Framingham Heart Study and UKPDS.
tion were mostly elderly and female with the highest BP, Limitations to these calculators included small diabetic
BMI, and heart rate. Finally, cluster 3 had the lowest hyper- population size and short length of disease diagnosis for
trophic systolic dysfunction and strain where patients were each study, respectively leading to difficulty in assessing
largely male and similar to cluster 1 in terms of age, HTN risk. With more recent studies, newer calculators have been
and obesity. After a median follow‐up of 67 months, CV developed.
mortality and hospitalization (arrhythmia, HF, CAD) were In 2013 the ACC/AHA guidelines developed a new
increased in clusters 2 and 3 with similar hazard ratios. model for risk assessment based on pooled data from the
Based on these results, a patient’s clinical features are not following studies: ARIC (Atherosclerosis Risk in
helpful in risk stratifying patients. Communities) study, the CHS (Cardiovascular Health
Studies have suggested that microalbuminuria may pos- Study), and CARDIA (Coronary Artery Risk Development
sibly be used as a prescreening tool for echocardiography in Young Adults) study, combined with applicable data
(ECHO) [49]. The Strong Heart Study divided 1748 from the Framingham Original and Offspring Study.
American Indian patients with type 2 diabetes into no Together these studies consisted of 9098 non‐Hispanic
albuminuria, microalbuminuria, and macroalbuminuria. white men, 1647 African‐American men, 11 240 non‐
After adjusting for age, sex, SBP, BMI, diabetes duration, Hispanic white women, and 2641 African‐American
CAD, and LV mass, results showed that the degree of dias- women between 40 and 79 years of age who had no previ-
tolic dysfunction was proportional to the level of microal- ous history of an MI, stroke, CHF, PCI, CABG, or afib [51].
buminuria. The HOPE study mentioned earlier Key components to the ACC/AHA calculator include the
demonstrated a correlation between albuminuria and addition of race and diabetes which were not included in
CVD, including HF [27]. the Framingham risk score [52]. Furthermore this 2013
Another proposed tool is N‐terminal pro‐B‐type natriu- ASCVD evaluation tool includes stroke outcome in risk
retic peptide (NT‐proBNP). 300 patients with diabetes prediction has led to improved identification of at‐risk
type 2 and without heart disease but NT‐proBNP > 125 pg/ female patients.
ml were randomized into a control group or intervention The large population of the MESA trial, patients of vary-
group in the NT‐proBNP selected prevention of cardiac ing ethnicities from across the US were followed for an
events in a population of diabetic patients without a history average of 10 years to cardiovascular mortality endpoints.
of cardiac disease (PONTIAC) trial [50]. The intervened From this follow‐up and integration of CAC scores algo-
group had beta blockers and renin‐angiotensin system rithm, The Mesa Risk score, was developed. The algorithm
(RAS) antagonists titrated up until the maximum recom- tool takes into account traditional Framingham risk factors
mended or tolerated dose was reached or NT‐proBNP con- (age, sex, smoking, total cholesterol, HDL cholesterol, sys-
centrations were below normal or decreased by 50%. At the tolic blood pressure, and antihypertensive medication use)
end of one year, there was no significant decrease in NT‐ and includes diabetes, family history of MI, race, and CAC
proBNP concentrations between the two but a significant score. Notably CAC score integration with traditional risk
decrease in HF hospitalization, CV hospitalization and factors led to improvement in risk prediction (Harrell’s C‐
CV death. statistic 0.8 vs 0.75; p‐value < 0.0001) [53]. This improve-
ment to prediction was evidenced in two external against the ASPEN, ADVANCE, and CARDS trials. This
validations studies Heinz Nixdorf Recall Study (HNR) and process when plotted had an estimated R2 of 0.86. With F‐
Dallas Heart Study (DHS). For the validation study > 3500 test application of the fitted line to the line of 100% accu-
Caucasian patients from Germany and > 1000 patients of racy, there was no statistically significant difference (p =
varying ethnicities from Dallas County in Texas, US 0.82).
respectively were included [54–56]. The DHS cohort repre- Another interesting risk prediction tool new in 2018 is
sents the diversity of the MESA population more than the the Astro‐CHARM calculator which was developed to esti-
HNR cohort. However, the prediction results still improved mate CV risk for astronauts. When building this estimator,
in both patient populations. Use of The MESA Risk score patient populations were pooled from the MESA trial,
on HNR and DHS studies resulted in good discrimination DHS, Prospective Army Coronary Calcium project
and calibration with Harrell’s C‐statistic scores of 0.779 and (PACC) and aimed to calculate risk in a middle‐aged popu-
0.816 respectively. lation (ages 40–65 years). All pooled cohorts measured
Based on another large population derivation and vali- CAC; some participants also had high sensitivity c‐reactive
dation study, the PREDICT risk estimator enrolls a large protein (hs‐CRP) measured. The prediction tool for the
New Zealand’s primary care population (over 400 000 par- first time incorporates CAC scores and hs‐CRP measure-
ticipants) in order to estimate cardiovascular risk [57]. This ments, along with various risk factors similar to the ones
population included the Maori population as well as Asian, mentioned for other calculators, to develop a novel estima-
Pacific, and Caucasian New Zealanders. Patients with tor [59]. In internal validation against the risk factor model
known cardiovascular disease, renal disease, and conges- only, c‐statistic scores were 0.784, 0.720, 0.813 (P=0.0002),
tive HF were excluded. Apart from other risk calculators, and 0.817 (P ≤ 0.0001) for the risk factor model, CAC
this estimator includes the traditional CHD, MI, and stroke model, risk factor model+CAC, and full Astro‐CHARM
outcomes as well as congestive HF. model respectively. A modified model without hs‐CRP was
Specifically in patients with diabetes, risk models devel- found to have a c‐statistic of 0.826. In external validation
oped in the past are derived from data from the United against the FHS cohorts demonstrated c‐statistics of 0.78
Kingdom Prospective Diabetes Study (UKPDS) which and 0.79 for models with and without hs‐CRP
dates back to populations of the UK in the 1970s. At that respectively.
time around 80% of the UK’s population was Caucasian [19].
The 2018 BRAVO study developed an updated risk estima-
Summary
tion tool based on data from the ACCORD trial which
derived data from the US consisting of over 10 000 partici- Recommendations for screening for CVD have undergone
pants with diabetes who were followed for about 4 many changes in the last decade, and guidelines vary
years [58]. The BRAVO calculator includes the importance among medical organizations. Better quality evidence to
of racial/ethnic background and medical history such as support some recommendations have emerged that help us
history of MI, stroke, angina, CHF, and previous revascu- individualize risk prediction to guide therapy. For most
larization for estimations in a variety of outcomes. people with diabetes, screening is still of limited value, as
Endpoint measurements of this study included stroke, screening imaging modalities may lead to both false posi-
CHF, non‐fatal MI, angina end‐stage renal disease (ESRD), tives and negatives and high‐risk patients may be better
blindness, revascularization surgery, severe pressure sensa- served by aggressive risk factor modification as suggested
tion loss, all‐cause mortality, and CVD death. Notably a by a number of guidelines.
history of severe hypoglycemia was included for predictors Studies have shown that the sodium‐glucose co‐trans-
of certain outcomes. Internal validation for this study porter 2 (SGLT‐2) inhibitors and some glucagon‐like‐pep-
showed close fit Kaplan Meier curves within 95% confi- tide‐1 (GLP‐1) receptor agonists, specifically liraglutide,
dence intervals for microvascular and macrovascular com- dulaglutide and semaglutide, decreased the risk of cardiovas-
plications. The external validation process was conducted cular death, nonfatal MI and nonfatal stroke. SGLT‐2 inhib-
itor dapagliflozin reduced HF hospitalizations. However, 9. American Diabetes Association. 10. Cardiovascular disease
most of the patients in these studies already had established and risk management: standards of medical care in diabetes
cardiovascular disease or multiple CV risk factors. There is – 2020. Diabetes Care. 2020;43(Suppl 1):
not enough data to suggest that these medications are also S111–S134.
10. Arnett DK, Blumenthal RS, Albert MA et al. 2019 ACC/
beneficial for primary prevention in patients with asymp-
AHA Guideline on the primary prevention of cardiovascular
tomatic type 2 diabetes, with fewer risk factors. In such
disease: Executive Summary: a report of the American
patients, screening may determine the appropriateness of
College of Cardiology/American Heart Association Task
deferring such therapy or other risk lowering therapy if Force on clinical practice guidelines. J Am Coll Cardiol.
patients so prefer. For example, a negative CAC may allow 2019;74(10):1376–1414.
a clinician to defer aggressive therapy for a few years. For 11. Budoff MJ, Raggi P, Beller GA et al. Noninvasive cardiovas-
patients with atypical symptoms screening for CVD may cular risk assessment of the asymptomatic diabetic patient:
help determine the need for more invasive testing and risk the Imaging Council of the American College of Cardiology.
factor reduction. JACC Cardiovasc Imaging. 2016;9(2):176–192.
12. Wackers FJ, Chyun DA, Young LH et al. Resolution of
asymptomatic myocardial ischemia in patients with type 2
diabetes in the Detection of Ischemia in Asymptomatic
References
Diabetics (DIAD) study. Diabetes Care. 2007;30(11):
1. World Health Organization. Cardiovascular Disease (CVDs). 2892–2898.
https://www.who.int/news‐room/fact‐sheets/detail/ 13. Young LH, Wackers FJ, Chyun DA et al. Cardiac outcomes
cardiovascular‐diseases‐(cvds) (accessed 29 September 2020). after screening for asymptomatic coronary artery disease in
2. Heart disease and stroke statistics‐2019 update: a report patients with type 2 diabetes: the DIAD study: a randomized
from the American Heart Association. Circulation. controlled trial. JAMA. 2009;301(15),1547–1555. 2009.
2019;139(10):e56–e528. 14. Boden WE, O’Rourke RA, Teo KK et al. COURAGE Trial
3. Buse JB, Ginsberg HN, Bakris GL et al. Primary prevention Research Group. Optimal medical therapy with or without
of cardiovascular diseases in people with diabetes mellitus: a PCI for stable coronary disease. N Engl J Med.
scientific statement from the American Heart Association 2007;356:1503–1516.
and the American Diabetes Association. Diabetes Care. 15. BARI 2D Study Group, Frye RL, August P et al. A rand-
2007;30(1):162–172. omized trial of therapies for type 2 diabetes and coronary
4. Expert Committee on the Diagnosis and Classification of artery disease. N Engl J Med. 2009;360:2503–2515.
Diabetes Mellitus. Report of the Expert Committee on the 16. The Decode Study Group. Is the current definition for dia-
Diagnosis and Classification of Diabetes Mellitus. Diabetes betes relevant to mortality risk from all causes and cardio-
Care. 1997;20(7):1183–1197. vascular and noncardiovascular disease. Diabetes Care.
5. American Diabetes Association. Standards of medical care in 2003;26(3):688–696.
diabetes – 2010 [published correction appears in Diabetes 17. Norhammer A, Tenerz A, Nilsson G et al. Glucose metabo-
Care. 2010;33(3):692]. Diabetes Care. 2010;33(Suppl lism in patients with acute myocardial infarction and no pre-
1):S11–S61. vious diagnosis of diabetes mellitus: a prospective study.
6. Pasquel FJ, Gregg EW, Ali MK. The evolving epidemiology Lancet. 2002;359(9324):2140–2144.
of atherosclerotic cardiovascular disease in people with 18. The Diabetes Control and Complications Trial Research Group.
diabetes. Endocinol Metab Clin N Am. 2018;47(1):1–32. The effect of intensive treatment of diabetes on the develop-
7. Haffner SM, Stern MP, Hazuda HP et al. Cardiovascular risk ment and progression of long‐term complications in insulin‐
factors in confirmed prediabetic individuals: does the clock dependent diabetes mellitus. N Engl J Med. 1993;329(14):
for coronary heart disease start ticking before onset of 977–986.
clinical diabetes? JAMA. 1990;263(21):2893–2898. 19. UK Prospective Diabetes Study (UKPDS) Group. Intensive
8. Haffner SM, Lehto S, Ronnemaa T et al. Mortality from blood sugar control with sulphonylureas or insulin com-
coronary heart disease in subjects 0with type 2 diabetes pared with conventional treatment and risk of complications
mellitus and in nondiabetic subjects with and without prior in patients with type 2 diabetes (UKPDS 33). Lancet.
myocardial infarction. N Engl J Med. 1998;339(4):229–234. 1998;352:837–853.
20. The Diabetes Control and Complications Trial/Epidemiology 2 diabetes and metabolic syndrome from the multi‐ethnic
of Diabetes Interventions and Complications (DCCT/EDIC) study of atherosclerosis. JAMA Cardiol. 2017;2(12):
Study Research Group. Intensive diabetes treatment and car- 1332–1340.
diovascular disease in patients with type 1 diabetes. N Engl J 33. Muhlestein JB, Lappé DL, Lima JAC et al. Effect of screening
Med. 2005;353(25):2643–2653. for coronary artery disease using ct angiography on mortal-
21. Holman RR, Paul SK, Bethel AM et al. 10‐year follow‐up of ity and cardiac events in high‐risk patients with diabetes: the
intensive glucose control in type 2 diabetes. N Engl J Med. FACTOR‐64 randomized clinical trial. JAMA. 2014;312(21):
2008;359(15):1577–1589. 2234–2243.
22. Yusuf S, Hawken S, Ounpuu S et al. Effect of potentially 34. Miller JM, Rochitte CE, Dewey M et al. Diagnostic perfor-
modifiable risk factors associated with myocardial infarction mance of coronary angiography by 64‐row CT. N Engl J Med.
in 52 countries (the interheart study):case‐control study. 2008;359:2324–2336.
Lancet. 2004;364(9438):937–952. 35. Kang SH, Park G‐M, Lee S‐W et al. Long‐term prognostic
23. Gaede P, Vedel P, Larsen N et al. Multifactorial intervention value of coronary ct angiography in asymptomatic type 2
and cardiovascular disease in patients with type 2 diabetes. diabetes mellitus. JACC. Cardiovascular Imaging. 2016;9(11):
N Engl J Med. 2003;348(5):383–393. 1292–1300.
24. Lind M, Svensson AM, Kosiborod M et al. Glycemic control 36. Andreini D, Pontone G, Mushtaq S et al. Prognostic value of
and excess mortality in type 1 diabetes. N Engl J Med. 2014; multidetector computed tomography coronary angiography
371:1972–1982. in diabetes. Diabetes Care. 2013;36(7):1834–1841.
25. Rawshani A, Rawshani A, Sattar N et al. Relative prognostic 37. Blanke P, Naoum C, Ahmadi A et al. Long‐term prognostic
importance and optimal levels of risk factors for mortality utility of coronary ct angiography in stable patients with dia-
and cardiovascular outcomes in type 1 diabetes mellitus. betes mellitus. JACC. Cardiovascular Imaging. 2016;9(11):
Circulation. 2019;139(16):1900–1912. 1289–1291.
26. Gerstein HC, Mann JF, Yi Q et al. Albuminuria and risk of 38. Werkhoven JMV, Cademartiri F, Seitun S et al. Diabetes:
cardiovascular events, death, and heart failure in diabetic prognostic value of CT coronary angiography – comparison
and nondiabetic individuals. JAMA. 2001;286(4):421–426. with a nondiabetic population. Radiology. 2010;256(1):
27. Heart Outcomes Prevention Evaluation Study Investigators. 83–92.
Effects of ramipril on cardiovascular and microvascular out- 39. Min, JK, Labounty TM, Gomez MJ et al. Incremental prog-
comes in people with diabetes mellitus: results of the HOPE nostic value of coronary computed tomographic angiogra-
study and MICRO‐HOPE substudy. Lancet. 2000;355: phy over coronary artery calcium score for risk prediction of
253–259. major adverse cardiac events in asymptomatic diabetic indi-
28. Fonseca V, Desouza C, Asnani S et al. Nontraditional risk viduals. Atherosclerosis. 2014;232(2):298–304.
factors for cardiovascular disease in diabetes. Endocrine 40. Lièvre, MM, Moulin P, Thivolet C et al. Detection of silent
Reviews. 2004;25(1):153–175. myocardial ischemia in asymptomatic patients with diabe-
29. Raggi, P, Shaw LJ, Berman DS et al. Prognostic value of coro- tes: results of a randomized trial and meta‐analysis assessing
nary artery calcium screening in subjects with and without the effectiveness of systematic screening. Trials. 2011;12:23.
diabetes. J Am Coll Cardiol. 2004;43(9):1663–1669. 41. Zellweger MJ, Maraun M et al. Progression to overt or silent
30. Elkeles RS, Godsland IF, Feher MD et al. Coronary calcium CAD in asymptomatic patients with diabetes mellitus at high
measurement improves prediction of cardiovascular events coronary risk: main findings of the prospective multicenter
in asymptomatic patients with type 2 diabetes: the PREDICT BARDOT trial with a pilot randomized treatment substudy.
study. Eur Heart J. 2008;29(18):2244–2251. JACC Cardiovasc Imaging. 2014;7(10):1001–1010.
31. Malik, S, Budoff MJ, Katz R et al. Impact of subclinical ath- 42. Jacqueminet S, Barthelemy O, Rouzet F et al. A randomized
erosclerosis on cardiovascular disease events in individuals study comparing isotope and echocardiography stress test-
with metabolic syndrome and diabetes: the multi‐ethnic ing in the screening of silent myocardial ischaemia in type 2
study of atherosclerosis. Diabetes Care. 2011;34(10): diabetic patients. Diabetes Metab. 2010;36(6):463–469.
2285–2290. 43. Faglia E, Manuela M, Antonella Q et al. Risk reduction of
32. Malik, S, Zhao Y, Budoff M et al. Coronary artery calcium cardiac events by screening of unknown asymptomatic coro-
score for long‐term risk classification in individuals with type nary artery disease in subjects with type 2 diabetes mellitus
at high cardiovascular risk: an open‐label randomized pilot [published correction appears in J Am Coll Cardiol. 2014;Jul 1;
study. Am Heart J. 2005;149(2):1–6. 63(25 Pt B):3026]. J Am Coll Cardiol. 2014;63(25 Pt
44. Fonseca V, Desouza C, Asnani S et al. Nontraditional risk B):2935–2959.
factors for cardiovascular disease in diabetes. Endocrine 53. McClelland RL, Jorgensen NW, Budoff M et al. 10‐year coro-
Reviews. 2004;25(1):153–175. nary heart disease risk prediction using coronary artery cal-
45. DeSouza CA, Shapiro LF, Clevenger CM et al. Regular aero- cium and traditional risk factors: derivation in the MESA
bic exercise prevents and restores age‐related declines in (Multi‐Ethnic Study of Atherosclerosis) with validation in
endothelium‐dependent vasodilation in healthy men. the HNR (Heinz Nixdorf Recall) Study and the DHS (Dallas
Circulation. 2000;102(12):1351–1357. Heart Study). J Am Coll Cardiol. 2015;66(15):1643–1653.
46. Dei Cas A, Khan SS, Butler J et al. Impact of diabetes on 54. Schmermund A, Möhlenkamp S, Stang A et al. Assessment
epidemiology, treatment, and outcomes of patients with of clinically silent atherosclerotic disease and established
heart failure. JACC Heart Fail. 2014;3(2):126–145. and novel risk factors for predicting myocardial infarction
47. Boonman‐de Winter LJ, Rutten FH, Cramer MJ et al. High and cardiac death in healthy middle‐aged subjects: rationale
prevalence of previously unknown heart failure and left ven- and design of the Heinz Nixdorf RECALL Study. Am Heart
tricular dysfunction in patients with type 2 diabetes. J. 2002;144 (2):212–218.
Diabetologia. 2012;55(8):2154–2162. 55. Stang A, Moebus S, Dragano N et al. Baseline recruitment
48. Ernande L, Audureau E, Jellis CL et al. Clinical implications and analyses of nonresponse of the Heinz Nixdorf Recall
of echocardiographic phenotypes of patients with diabetes Study: identifiability of phone numbers as the major deter-
mellitus. J Am Coll Cardiol. 2017;70(14):1704–1716. minant of response. Eur J Epidemiol. 2005;20(6):489–496.
49. Liu JE, Robbins DC, Palmieri V et al. Association of albumi- 56. Victor RG, Haley RW, Willett DL et al. The Dallas Heart
nuria with systolic and diastolic left ventricular dysfunction Study: a population‐based probability sample for the multi-
in type 2 diabetes: the Strong Heart Study. J Am Coll Cardiol. disciplinary study of ethnic differences in cardiovascular
2003;41(11):2022–2028. health. Am J Cardiol. 2004;93(12):1473–1480.
50. Huelsmann M, Neuhold S, Resl M et al. PONTIAC (NT‐ 57. Pylypchuk R, Wells S, Kerr A et al. Cardiovascular disease
proBNP selected prevention of cardiac events in a popula- risk prediction equations in 400 000 primary care patients in
tion of diabetic patients without a history of cardiac disease): New Zealand: a derivation and validation study. The Lancet.
a prospective randomized controlled trial. J Am Coll Cardiol. 2018;391(10133):1897–1907.
2013;62(15):1365–1372. 58. Shao H, Fonseca V, Stoecker C et al. Novel risk engine for
51. Karmali KN, Goff DC, Ning H et al. A systematic examina- diabetes progression and mortality in USA. Building,
tion of the 2013 ACC/AHA pooled cohort risk assessment Relating, Assessing, and Validating Outcomes (BRAVO).
tool for atherosclerotic cardiovascular disease. J Am Coll PharmacoEconomics. 2018;36:1125–1134.
Cardiol. 2014;64(10):959–968. 59. Khera A, Budoff MJ, O’Donnell CJ et al. Astronaut
52. Goff DC Jr, Lloyd‐Jones DM, Bennett G et al. 2013 ACC/ Cardiovascular Health and Risk Modification (Astro‐
AHA guideline on the assessment of cardiovascular risk: a CHARM) coronary calcium atherosclerotic cardiovascular
report of the American College of Cardiology/American disease risk calculator. Circulation. 2018;138(17):1819–1827.
Heart Association Task Force on Practice Guidelines doi:10.1161/CIRCULATIONAHA.118.033505.
14 Choosing Medications for Type 2 Diabetes –
What Weighting Should Be Given
to Cardiovascular Risk Reduction?
Adrian Vella
the effects of glycemic control in type 2 diabetes on cardio-

vascular risk are less certain. In fairness, similar findings
●● Cardiovascular risk reduction is an integral part of
have been reported by the United Kingdom Prospective
diabetes management. Diabetes Study group (UKPDS) for type 2 diabetes [3, 4].
●● This should be achieved with lipid‐lowering and Since hypoglycemia can also produce acute and chronic,
antihypertensive agents in addition to pharmacother- potentially catastrophic, complications, the goal of diabetes
apy intended to achieve glucose lowering. therapy has been to safely achieve glucose concentrations
●● Some agents may have direct effects on cardiovascular
outcomes over and above their effects on glycemic
as close to the normal physiologic range as possible while
control. avoiding hypoglycemia.
Cardiovascular disease (CVD) is the leading cause of
morbidity and mortality in the world regardless of race,
ethnicity, or gender [5], and costs related to CVD in the
United States surpassed $300 billion in 2010 [6]. In spite of
Diabetes and cardiovascular risk:
success in reducing the mortality from acute complications
how does it affect therapeutic goals?
of heart disease, the incidence of atherosclerosis and its
Diabetes mellitus causes morbidity, mortality and health- burden has continued to grow. The estimated annual inci-
care costs in excess of ~$170 billion/year [1]. Chronic dence of a myocardial infarction is 785 000 while 470
hyperglycemia leads to the micro‐ and macrovascular 000 will suffer a recurrent acute coronary syndrome.
complications that frequently accompany diabetes. The Depending on their gender and clinical outcome, individu-
Diabetes Control and Complications Trial (DCCT) was als who survive the acute stage of a myocardial infarction
able to show that “tight” glucose control i.e. glucose con- have 1.5 to 15 times higher probability of illness and death
centrations closer to the normal physiologic range (as compared to the general population.
measured by Hemoglobin A1c) compared to “conventional” Patients with T2DM are at an increased risk of adverse
control prevented or delayed retinopathy and nephropathy cardiovascular events compared to those without diabe-
in people with type 1 diabetes [2]. Since that landmark tes [7], and despite efforts at risk reduction including
study the goal of diabetes management has been to safely smoking cessation and blood pressure and cholesterol opti-
achieve good glycemic control while avoiding hypoglyce- mization, the majority of people with T2DM continue to
mia. Although it is reasonable to extrapolate the conclu- die from cardiovascular disease [8]. Observational studies
sions of the DCCT to the management of type 2 diabetes, have shown a significant association between glycemic

174
Choosing Medications for Type 2 Diabetes 175
control and CVD [9–12]. However, results from rand- ostprandial insulin secretion decrease with metformin
p
omized clinical trials have been more controversial. The therapy. Hypoglycemia in patients treated with metformin
UK Prospective Diabetes Study (UKPDS) showed that monotherapy is unlikely and should prompt a search for
HbA1c, was an independent predictor of cardiovascular other causes of hypoglycemia. Because of its ease of use and
outcomes [13]; the ADVANCE study showed that intensive general safety, metformin is usually considered to be first‐
glucose control (HbA1c < 6.5 %) was associated with a line pharmacotherapy in the treatment of type 2 diabetes.
reduction in major microvascular events, but not in major Its use has also been advocated in people with
macrovascular events [14]; the VADT study showed that prediabetes [18].
intensive glucose control had no significant effect on rates The primacy of this compound in the treatment of type
of major cardiovascular events, death, or microvascular 2 diabetes arises out of the UK Prospective Diabetes Study
complications [15] while the ACCORD trial showed that (UKPDS) which demonstrated that it decreased the risk of
aggressive glycemic control targeting normal HbA1c was microvascular complications with less associated weight
associated with an increased mortality but did not reduce gain and fewer episodes of hypoglycemia compared to
major cardiovascular events [16]. insulin or sulfonylureas [4]. It may be the first‐line phar-
There are multiple caveats to these observations. Over macological therapy of choice in these patients. In this
the past two decades non‐diabetes therapies that success- study, subjects in the intensive treatment arm had less fatal
fully mitigate cardiovascular risk by targeting cardiovascu- and non‐fatal myocardial infarction after 10 years of fol-
lar risk factors such as hypertension and dyslipidemia as low‐up although no significant differences attributable to
well as the proliferation of diabetes therapies have improved treatment modality could be ascertained [13]. Other direct
glycemic control and cardiovascular risk to the point that comparisons with sulfonylureas or other compounds
demonstrating benefit of other interventions has paradoxi- have shown no clear evidence of superiority for
cally made it more difficult to demonstrate benefit. It is metformin [19].
also important to appreciate that the prediabetic state is
associated with high cardiovascular risk conferred by dys-
Sulfonylureas
lipidemia, obesity, and hypertension despite comparatively
minimal hyperglycemia [17]. Sulfonylureas bind to the potassium channel complex
The logical conclusions one can make from these obser- (encoded by KCJN11/ABCC8) causing decreased potas-
vations is that diabetes is associated with adverse cardio- sium influx, depolarization of the β‐cell membrane and a
vascular risk and ameliorating cardiovascular risk is a subsequent increase of calcium flux into the cell. This
central goal of therapy. Glycemic control is also important results in exocytosis of insulin from secretory gran-
and should not be neglected. Some agents used to treat dia- ules [20]. Unlike incretin‐based therapy, discussed later in
betes may affect cardiovascular risk beyond their ability to this chapter, this effect is not glucose‐dependent and all
lower glucose concentrations (see Figure 14.1). The evi- patients treated with sulfonylureas are theoretically at risk
dence for each therapeutic class of “non‐insulin” therapy of hypoglycemia. Good initial response to therapy followed
will be discussed in the relevant sections of this chapter. by subsequent failure is often referred to as secondary fail-
ure. This may be due to patient‐related factors (poor com-
pliance with therapy/lifestyle, co‐morbidities that preclude
Metformin
sulfonylurea use) as well disease‐progression. Although a
Metformin is the only biguanide approved for use in the similar rate of secondary failure with metformin has been
United States and seems to improve insulin action in observed in the UKPDS [21], durability of effect was infe-
hepatic and peripheral tissues. The mechanism of action of rior to metformin or thiazolidinedione therapy in the
this compound remains unknown. However, as would be ADOPT study [22].
anticipated from its known efficacy of improving insulin The safety of this class of compounds was first ques-
action in patients with type 2 diabetes, fasting and tioned by the University Group Diabetes Program trial
SU Metformin GLP-1RA
Magnitude of Glycemic lowering
TZDs
SGLT-2i
DPP-4i
Bromocriptine
Acarbose
Pramlintide Colesevelam
Evidence of direct Cardiovascular benefit*
FIG 14.1 Therapeutic classes clustered by therapeutic effect versus the level of evidence supporting a direct cardiovascular benefit.
SU = sulfonylurea; TZDs = Thiazolidinediones; GLP‐1RA = Glucagon-Like Peptide‐1 Receptor Agonists; SGLT‐2i = Sodium‐
Glucose cotransporter‐2 inhibitors; DPP‐4i = DiPeptidyl Peptidase‐4 inhibitors. *Position on the axis represents the strength of the
current evidence not the magnitude of cardiovascularbenefit.
where tolbutamide was associated with increased mortal- It exists in three isoforms with different tissue specificity.
ity [23]. Tolbutamide is no longer used in clinical practice. When activated, PPAR acts as a transcriptional enhancer of
However, in the UKPDS outcomes were similar to those of genes with PPAR‐response elements [25]. The PPAR‐α
metformin [3]. In a meta‐analysis of 116 randomized clini- isoform is expressed in the liver, heart, and brown fat where
cal trials, Monami et al. reported no effect of major cardio- active fatty acid catabolism occurs. PPAR‐γ is expressed in
vascular events and a trend towards benefit in myocardial adipose tissue although it is also expressed in the vascula-
infarction [24]. On the other hand, glimepiride increased ture. PPAR‐δ is more generally expressed. PPAR‐α regu-
the risk of stroke. More notable are that these observations lates the transcription of genes important in fatty acid
depend on multiple relatively underpowered studies. uptake and oxidation. Fibrates used to treat dyslipidemia,
While comparisons versus placebo or no therapy are act as PPAR‐α agonists [26].
favorable, this was not the case against other comparators Thiazolidinediones are PPAR‐γ agonists and stimulate
such as dipeptidyl peptidase‐4 inhibitors. Nevertheless, adipocyte differentiation as well as the expression of glu-
these compounds are associated with a risk of hypoglyce- cose transporters (GLUT‐1 and GLUT‐4), increasing
mia and they should be used cautiously in patients with hepatic and peripheral glucose uptake [27]. The first agent
established, active heart disease. Well‐powered cardiovas- in this class, troglitazone, was withdrawn after reports of
cular outcome trials are still needed to establish the safety hepatotoxicity [28]. This seems to be unique to troglita-
of this class of compounds. zone. Clinical use of these agents has been associated with
weight gain. This is in part due to stimulation of adipocyte
differentiation as well as fluid retention which can be sig-
Thiazolidinediones
nificant and may be to a degree that results in hemodilu-
The peroxisome proliferator activated receptor (PPAR) is tion [25]. For this reason, thiazolidinediones are
activated by fatty acids and fatty acid‐derived eicosanoids. contraindicated in congestive cardiac failure [29].
The discontinuation of troglitazone allowed randomiza- The other potential adverse event associated with this
tion of previously treated patients to either rosiglitazone or therapeutic class is that of fracture risk. The ADOPT (A
pioglitazone [30]. This study reported that in the absence Diabetes Outcome Progression Trial) study demonstrated
of differences in glycemic control pioglitazone lowered an increased risk of limb fractures in pre‐ and postmeno-
total and LDL‐cholesterol whereas rosiglitazone produced pausal women [22]. At present it appears that the benefits
a slight increase in these parameters. Nissen et al. subse- of decreased cardiovascular mortality outweigh the mor-
quently reported that a meta‐analysis of published studies tality attributed to osteoporotic fracture [43].
suggested that rosiglitazone increased risk of myocardial
infarction and cardiovascular mortality [31]. Because of
Acarbose
this and other data the Food and Drug Administration
(FDA) placed significant restrictions on rosiglitazone use. Acarbose is an α‐glycosidase inhibitor which prevents the
These restrictions were only rescinded after and open‐label luminal digestion of polysaccharides and, therefore, delays
non‐inferiority study specifically designed to assess cardio- the absorption of monosaccharides such as glucose from
vascular outcomes [32] showed no increased risk of mor- the gut. It has also been argued that part of its effect arises
tality or heart attack. The BARI 2D trial also reported no from the increased delivery of undigested carbohydrate to
increase in risk of myocardial infarction associated with the terminal ileum thereby stimulating Glucagon‐like pep-
rosiglitazone [33]. tide‐1 (GLP‐1) secretion by the intestinal L‐cells [44].
There is significant in vitro; and rodent data to support Since acarbose does not directly interfere with insulin
a favorable effect of thiazolidinediones on vascular biol- secretion, hypoglycemia does not occur in diabetic patients
ogy [25]. Indeed, in humans, pioglitazone slows progres- using this medication as monotherapy. However, if hypo-
sion of carotid intima‐media thickness [34] and decreased glycemia (due to another medication e.g. sulfonylurea)
coronary atheroma [35]. A secondary cardiovascular pre- does occur, pure glucose should be used to treat this, as
vention study utilizing pioglitazone failed to show benefit absorption of disaccharides or more complex carbohy-
for a primary composite cardiovascular endpoint. However, drates will be delayed [45].
the study was able to fulfil its secondary endpoint which The Study to Prevent Non‐Insulin Dependent Diabetes
was a composite of death, heart attack, and cerebrovascular Mellitus (STOP‐NIDDM Trial) randomized ~1400 sub-
events compared to placebo [36]. Post hoc analyses in jects with impaired fasting glucose or diabetes (≥ 100 mg/
patients with a previous heart attack and stroke show a dL (5.6 mmol/L) but ≤ 140 mgd/L (7.8 mmol/L) and with
decrease in recurrent events with pioglitazone [37, 38]. impaired glucose tolerance to placebo or acarbose taken
This was borne out by the Insulin Resistance Intervention with each meal (three times a day). The investigators sub-
after Stroke (IRIS) trial [39]. sequently reported that acarbose decreased the incidence
Against these benefits one must balance two potential of diabetes, cardiovascular disease and hypertension [46].
adverse events associated with thiazolidinedione use. The The incidence of pre‐specified) 2° endpoints of myocardial
first is that of bladder cancer which seems to be increased infarction (1 vs 12) and other cardiovascular events (15 vs.
after more than 2 years of drug exposure [40]. Bladder can- 32) was decreased. These results have been the subject of
cer is so uncommon, however, that randomized controlled some controversy [47].
trials are insufficiently powered to detect these effects. For Subsequently, the Acarbose Cardiovascular Evaluation
example, the PROactive study cited above suggested an (ACE) study in ~6500 patients with established coronary
association that did not persist after 6 years of follow‐ heart disease and impaired glucose tolerance demonstrated
up [41]. In the interests of framing the prospective risk no effect of acarbose on the prevention of cardiovascular
appropriately, an individual’s annual risk of bladder cancer events [48]. In this cohort of people with impaired glucose
goes from 0.043% to 0.049% when on pioglitazone or more tolerance, acarbose decreased the incidence of diabetes and
than 20 000 would need to be treated to cause 1 new case of improved glucose tolerance. There is no evidence to sug-
bladder cancer [25, 42]. gest that this benefit persists after cessation More recently,
acarbose was compared to sulfonylureas as adjunctive ther- of pramlintide differ between type 1 and type 2 diabetes.
apy for people with type 2 diabetes [49]. The investigators Currently it is approved as adjunctive therapy, adminis-
demonstrated a decrease in major atherosclerotic events tered pre‐prandially with prandial insulin. Its use is associ-
and decreased hospitalization for hypoglycemia but no ated with satiety and a mild decrease in weight [57, 58].
effect on all‐cause mortality. There are no published cardiovascular outcomes data for
this compound. However, some relatively small studies
suggest that there are no direct effects on cardiovascular
Colesevelam
risk factors such as blood pressure and lipoprotein
Bile‐acid sequestrants such as colesevelam and cholesty- concentrations [59].
ramine increase fecal excretion of bile acids by interrupting
the enterohepatic circulation. This leads to a decrease in Bromocriptine
the bile acid pool resulting in diversion of hepatic choles-
A quick‐release formulation of bromocriptine (a central
terol synthesis to the production of bile acids with concom-
Dopamine D2 receptor agonist used to treat prolactinoma)
itant upregulation of hepatic (low‐density lipoprotein)
is approved for the treatment of type 2 diabetes. The mecha-
LDL‐cholesterol receptors [50]. These actions result in a
nism by which bromocriptine produces glucose lowering is
small, but significant, lowering of cholesterol concentra-
uncertain. Gaziano et al. randomized 3095 patients with
tions. Intriguingly, their use in people with type 2 diabetes
type 2 diabetes (BMI of < 43 kg/m2, and an A1C concentra-
has resulted in ~0.3% decrease in HbA1c [51–53].
tion ≤ 10.0%) to bromocriptine or placebo as an adjunct to
Consequently, colesevelam has been approved as adjunctive
the patient’s usual diabetes therapy [60]. The drug caused
therapy for type 2 diabetes. In humans it lowers fasting
nausea but did not alter weight. However, a composite end-
glucose concentrations likely through an insulin‐sensitizing
point consisting of stroke, coronary revascularization, myo-
effect on the liver. Postprandial hepatic glucose uptake is
cardial infarction, and hospitalization for angina or heart
also enhanced [54].
failure was significantly decreased by bromocriptine therapy.
Although there is general consensus as regards the cho-
Intriguingly, in this study, bromocriptine resulted in a 0.1%
lesterol‐lowering properties of these compounds, evidence
increase in HbA1c (versus a 0.2% increase in the placebo
that this results in improved outcomes is relatively sparse.
arm) and a significant decrease (5%) in triglycerides. Other
For example, analysis of the healthcare insurance claims
studies have reported a ~0.6% decrease in HbA1c [61].
data of a third‐party payer compared a composite cardio-
The mechanism of action whereby bromocriptine use
vascular event score which included myocardial infarction,
results in glucose‐lowering is somewhat speculative. It is
stroke, angina, or revascularization in patients taking
suggested by the observation in hibernating mammals that
ezetimibe or colesevelam. After 12 months of treatment,
the transition to an insulin‐resistant state during hiberna-
subjects taking colesevelam had a lower event score (odds
tion is governed by dopamine concentrations in hypotha-
ratio (OR) 0.54, 95 % confidence interval (CI) 0.30–0.97)
lamic nuclei. In humans who do not experience such
compared to those taking ezetimibe [55].
seasonal changes in their metabolic state it is speculated
that stimulation of central dopamine signaling may restore
Pramlintide circadian misalignment [62]. Small, short‐term studies in
humans have suggested an effect on insulin sensitivity but
Pramlintide is a synthetic derivative of amylin – a product this has not been consistently observed [63].
of the β‐cell that is co‐secreted with insulin [56].
Concentrations are elevated in people with type 2 diabetes
Glucagon‐like peptide‐1‐based therapy
and low or undetectable in people with type 1 diabetes [57].
Pharmacologic concentrations of amylin and its analogue The physiological basis for Glucagon‐like peptide‐1
inhibit gastric emptying through a central vagally‐mediated (GLP‐1)‐based therapy is outside the scope of this chapter
mechanism. Intriguingly, the dose‐response characteristics [64]. Nevertheless, it is important to recognize the basis of
the two therapeutic classes in this category [65]. GLP‐1 has talization for unstable angina, coronary revascularization,
a short half‐life in the circulation (secondary to its rapid or heart failure was similarly unaffected [73]. However,
inactivation by dipeptidyl peptidase 4 – DPP‐4 – an more patients in the saxagliptin group than in the placebo
enzyme which is widely distributed and cleaves all peptides group were hospitalized for heart failure (3.5% vs. 2.8%,
whose second N‐terminal amino acid is a proline or ala- according to 2‐year Kaplan–Meier estimates; hazard ratio,
nine). Accordingly, inhibition of DPP‐4 impairs clearance 1.27; 95% CI, 1.07 to 1.51; p = 0.007). Non‐inferiority for
of all DPP‐4 substrates. From a glycemic point of view the safety compared to placebo was also reported for aloglip-
most important is GLP‐1. The rise in active GLP‐1 seen tin [74] and sitagliptin [75] with no significant effects on
with DPP‐4 inhibition is transient and there are no associ- cardiovascular endpoints. No signal in terms of heart fail-
ated gastrointestinal side‐effects or weight loss [66, 67]. ure was observed in these trials.
In contrast, GLP‐1 receptor agonists comprise modified The mechanism underlying the observation of heart
forms of GLP‐1 that render them resistant to the actions of failure associated with saxagliptin remains unexplained. Of
DPP‐4 or analogues of GLP‐1 capable of binding and stimu- note, a study of 254 patients with a left ventricular ejection
lating the GLP‐1 receptor. For example, exenatide is a natu- fraction of ~30% were randomized to placebo or vildaglip-
rally occurring analogue of GLP‐1 that is found in the saliva of tin. After a year of therapy vildagliptin slightly, but signifi-
large reptiles such as Heloderma horridum. Its second amino cantly increased left ventricular end‐diastolic and
acid is a glycine, making it resistant to degradation by DPP‐4, end‐systolic volume. Ejection fraction was unchanged [76].
conferring a half‐life of ~2 hours as opposed to the 2–5 min- Overall, to date, the data would support cardiovascular
utes of native GLP‐1. Detectable concentrations are present safety but little direct benefit of DPP‐4 inhibitors on car-
for up to 10 hours after administration. GLP‐1 receptor ago- diovascular disease. Recent data has suggested that the
nists have more powerful glucose‐lowering effects than do increased risk of hospitalization for heart failure observed
DPP‐4 inhibitors (a decrease of HbA1c of ~1% vs. ~0.5% asso- in SAVOR‐TIMI could be explained by deteriorating renal
ciated with DPP‐4i monotherapy). They produce weight loss function [77].
although there is a significant incidence of gastrointestinal Marso et al. reported that liraglutide decreased the inci-
side‐effects such as nausea and vomiting [68, 69]. dence of a composite endpoint (cardiovascular death, non-
DPP‐4 inhibition can affect circulating concentrations fatal myocardial infarction or nonfatal stroke). There was
of other peptide mediators that are potentially substrates also a lower risk of all‐cause mortality and microvascular
for DPP‐4 [70]. Indeed, this might explain the (uncom- complications than were observed in the placebo group.
mon) interaction between angiotensin‐converting enzyme The authors estimated that 66 people would need to be
inhibitors and DPP‐4 inhibitors to produce angi- treated for 3 years to prevent a cardiovascular death [78].
oedema [71]. On the other hand, differences in selectivity This has not been observed with lixisenatide [79] or exena-
for other members of the dipeptidyl peptidase family were tide [80], which, while non‐inferior to placebo, had no
originally thought to explain potential differences in side‐ effects on cardiovascular endpoints.
effect profiles of these therapeutic agents [72]. In practice, In addition to liraglutide, semaglutide studied in
significant side‐effects attributable to this class of agents patients with cardiovascular disease (~80% of the study
has been hard to document. However, it is possible that (a population) or with multiple risk factors, successfully
lack of) dipeptidyl peptidase might explain between‐com- decreased the primary composite endpoint (cardiovascular
pound differences in cardiovascular outcomes trials. death, nonfatal myocardial infarction or nonfatal stroke).
For example, in the SAVOR‐TIMI study the primary This was driven by a decrease in cerebrovascular events
end point was a composite of death from cardiovascular and a trend towards decreased myocardial infarction [81].
causes, myocardial infarction, or ischemic stroke. Albiglutide also decreased the identical composite primary
Saxagliptin did not have any effect on this endpoint. The endpoint – this time driven by a decrease in myocardial
secondary endpoint, a composite of death from cardiovas- infarction [82]. Dulaglutide also produced similar
cular causes, myocardial infarction, ischemic stroke, hospi- effects [83].
Although the tolerability of GLP‐1 receptor agonists has in diabetic patients with heart failure, 6 months after initia-
improved over time, gastrointestinal distress remains the tion of treatment [93].
most prominent side‐effect that leads to discontinuation. The CREDENCE (Canagliflozin and Renal Events in
Cost may also be a barrier to their use. The other potential Diabetes With Established Nephropathy Clinical
consideration is patient resistance to injectable therapy, Evaluation) study randomized 4401 participants with type
although the advent of once‐weekly injectable therapy may 2 diabetes mellitus and chronic kidney disease to canagli-
make this less of an obstacle. flozin or placebo and demonstrated improved cardiovas-
cular endpoints regardless of whether it was used in
patients without or with established cardiovascular dis-
SGLT‐2i
ease [94]. The DECLARE‐TIMI (Dapagliflozin and
Sodium glucose cotransporter 2 inhibitors (SGLT‐2i) are a Cardiovascular outcomes in Type 2 Diabetes) studied
new class of anti‐hyperglycemic agents that function to patients with cardiovascular risk factors or established car-
concomitantly inhibit the reabsorption of glucose and diovascular disease (~40%) of the study population over a
sodium in the renal proximal convoluting tubule [84]. 4‐year period. It failed to reduce the composite of stroke,
These compounds produce glycosuria and natriuresis, myocardial infarction, or cardiovascular death. However, a
which translates into approximately a 0.7–1.0 % reduction composite of cardiovascular death and hospitalization for
in circulating Hemoglobin A1c, approximately a 5/2 mm Hg heart failure was significantly reduced. This was a pre‐
blood pressure reduction, and a 2–3 kg loss in body weight, specified endpoint where the decrease was driven by
a 30–40% reduction in albuminuria via a reduction in decreased heart failure [95].
intra‐glomerular pressure, and other favorable metabolic The findings from these studies need to be kept in per-
effects [85]. By promoting glycosuria, these compounds spective. The cardiovascular benefit of these compounds,
also induce calorie wastage and induce a mild degree of while impressive, may have been driven by the relatively
weight loss. Nevertheless, the use of an SGLT‐2i may result high‐risk populations (established CVD at baseline) stud-
in potential risk, with increased risk for genitourinary tract ied. In the above referenced studies, the difference in glyce-
infections, diabetic ketoacidosis, acute kidney injury, frac- mia between the treatment groups was minimal, suggesting
tures, and atraumatic below‐knee lower extremity amputa- that extra‐glycemic effects of the drugs are responsible for
tion [86, 87]. the reduction in CVD events [96]. Indeed, the Dapa‐HF
It is notable that the actions of SGLT‐2i are independent trial, designed to study the effect of dapagliflozin on car-
of the integrity of the insulin secretory apparatus, and can diovascular mortality and worsening heart failure, demon-
theoretically lower glucose in people with type 1 diabetes. strated benefit in patients without (and with) type 2
Unexpectedly, in addition to an improvement in HbA1c, diabetes [97]. The other notable observation for this class
treatment with SGLT‐2i decreases cardiovascular out- of compounds was that the side‐effects observed were rela-
comes. Trials in patients with type 2 diabetes mellitus and tively minor and did not lead to discontinuation of
high CVD risk have recently reported reductions in major therapy.
adverse cardiovascular events, specifically the composite of
CV mortality, nonfatal myocardial infarction, and nonfatal
Conclusions
stroke, and particular benefit in reducing hospitalization
for heart failure, namely the EMPA‐REG OUTCOME trial, It is important to appreciate that modern diabetes therapy
and the CANVAS Program (CANVAS and is multi‐faceted and the clinician has multiple options that
CANVASS‐R) [86, 88, 89]. The mechanisms of these may allow better individualization of pharmacotherapy for
actions are uncertain at the present time but are the subject a given individual. Recently, the emergence of data to sug-
of active investigation [90–92]. Intriguingly, the EMPA‐ gest that specific compounds may have direct cardiovascu-
HEART CardioLink‐6 Randomized Clinical Trial demon- lar benefit has re‐emphasized the importance of
strated that empagliflozin decreased left ventricular mass cardiovascular risk reduction in the management of type 2
diabetes (Figure 14.1). Accordingly, it is also important to References

remember the importance of lifestyle modification and
1. Albright AL, Gregg EW. Preventing type 2 diabetes in
appropriate management of dyslipidemia and
communities across the U.S.: the National Diabetes
hypertension. Prevention Program. American Journal of Preventive
Glycemic control per se has established benefits on Medicine. 2013;44:S346–351.
micro‐ and macrovascular disease although the benefits on 2. Diabetes C, Complications Trial Research G, Nathan DM,
the latter are less clear in type 2 diabetes than they are in Genuth S, Lachin J, Cleary P, Crofford O, Davis M, Rand L,
type 1 diabetes. While recent data suggests direct benefits Siebert C. The effect of intensive treatment of diabetes on the
of SGLT‐2i and GLP‐1 receptor agonists it is important to development and progression of long‐term complications in
remember that there is evidence of direct benefit for other insulin‐dependent diabetes mellitus. N Engl J Med.
compounds and also that there are some limitations to the 1993;329:977–986.
3. UK Prospective Diabetes Study (UKPDS) Group. Intensive
generalizability of these findings – especially so, given the
blood‐glucose control with sulphonylureas or insulin
reliance on composite endpoints.
compared with conventional treatment and risk of
With these caveats, however, it may be reasonable to
complications in patients with type 2 diabetes (UKPDS 33).
give preferential consideration to therapeutic classes Lancet. 1998;352:837–853.
depending on the clinical context (see Table 14.1). For 4. UK Prospective Diabetes Study (UKPDS) Group. Effect of
example, the presence of weight gain or abnormal eating intensive blood‐glucose control with metformin on
patterns, the presence of heart failure or renal disease or complications in overweight patients with type 2 diabetes
lipodystrophy [98, 99] may all suggest a therapeutic choice (UKPDS 34). Lancet. 1998;352:854–865.
in addition to standard first‐line therapy. 5. Roger VL, Go AS, Lloyd‐Jones DM, Benjamin EJ, Berry JD,
Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, Fullerton
TABLE 14.1 Patient characteristics that might affect prescribing. HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, Kissela
BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD,
Patient factors that Therapeutic class Makuc DM, Marcus GM, Marelli A, Matchar DB, Moy CS,
might affect
To avoid To use Mozaffarian D, Mussolino ME, Nichol G, Paynter NP,
prescribing
Soliman EZ, Sorlie PD, Sotoodehnia N, Turan TN, Virani SS,
Hypoglycemia risk Sulfonylureas DPP‐4i Wong ND, Woo D, Turner MB, and on behalf of the
Acarbose* GLP‐1RA American Heart Association Statistics Committee and Stroke
SGLT‐2i Statistics Subcommittee. Executive summary: heart disease
Lipodystrophy TZDs and stroke statistics – 2012 update: a report from the
Heart failure TZDs SGLT‐2i American Heart Association. Circulation. 2012;125:188–197.
Obesity TZDs† GLP‐1RA
6. Lloyd‐Jones D, Adams RJ, Brown TM, Carnethon M, Dai S,
Nephropathy Sulfonylureas SGLT‐2i
De Simone G, Ferguson TB, Ford E, Furie K, Gillespie C, Go
Active cardiac disease Sulfonylureas TZDs
GLP‐1RA A, Greenlund K, Haase N, Hailpern S, Ho PM, Howard V,
SGLT‐2i Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A,
Dyslipidemia despite Colesevelam McDermott MM, Meigs J, Mozaffarian D, Mussolino M,
statin use Nichol G, Roger VL, Rosamond W, Sacco R, Sorlie P, Stafford
Multiple daily Pramlintide R, Thom T, Wasserthiel‐Smoller S, Wong ND, Wylie‐Rosett
injections of insulin J, and on behalf of the American Heart Association Statistics
Committee and Stroke Statistics Subcommittee. Executive
TZDs = Thiazolidinediones; GLP‐1RA = Glucagon‐Like Peptide‐1
summary: heart disease and stroke statistics – 2010 update: a
Receptor Agonists; SGLT‐2i = Sodium‐Glucose cotrans-
report from the American Heart Association. Circulation.
porter‐2 inhibitors; DPP‐4i = DiPeptidyl Peptidase‐4 inhibitors.
*
Acarbose might alter the ability to treat hypoglycemia not 2010;121:948–954.
cause hypoglycemia. 7. Sarwar N, Gao P, Seshasai SR, Gobin R, Kaptoge S, Di
†
TZDs may exacerbate weight gain but still improve insulin Angelantonio E, Ingelsson E, Lawlor DA, Selvin E, Stampfer
action and glycemic control overall. M, Stehouwer CD, Lewington S, Pennells L, Thompson A,
Sattar N, White IR, Ray KK, Danesh J. Diabetes mellitus, 18. Knowler WC, Hamman RF, Edelstein SL, Barrett‐Connor
fasting blood glucose concentration, and risk of vascular E, Ehrmann DA, Walker EA, Fowler SE, Nathan DM, Kahn
disease: a collaborative meta‐analysis of 102 prospective SE. Prevention of type 2 diabetes with troglitazone in
studies. Lancet. 2010;375:2215–2222. the Diabetes Prevention Program. Diabetes 2005;54:
8. Gu K, Cowie CC, Harris MI. Mortality in adults with and 1150–1156.
without diabetes in a national cohort of the U.S. population, 19. Gnesin F, Thuesen ACB, Kahler LKA, Madsbad S,
1971–1993. Diabetes Care. 1998;21:1138–1145. Hemmingsen B. Metformin monotherapy for adults with
9. Khaw KT, Wareham N, Bingham S, Luben R, Welch A, Day type 2 diabetes mellitus. Cochrane Database Syst Rev 2020;6:
N. Association of hemoglobin A1c with cardiovascular CD012906.
disease and mortality in adults: the European prospective 20. Siconolfi‐Baez L, Banerji MA, Lebovitz HE. Characterization
investigation into cancer in Norfolk. Ann Intern Med. and significance of sulfonylurea receptors. Diabetes Care.
2004;141:413–420. 1990;13 Suppl 3:2–8.
10. Selvin E, Marinopoulos S, Berkenblit G, Rami T, Brancati 21. United Kingdom Prospective Diabetes Study Group. United
FL, Powe NR, Golden SH. Meta‐analysis: glycosylated Kingdom Prospective Diabetes Study 24: a 6‐year, rand-
hemoglobin and cardiovascular disease in diabetes mellitus. omized, controlled trial comparing sulfonylurea, insulin,
Ann Intern Med. 2004;141:421–431. and metformin therapy in patients with newly diagnosed
11. Lawes CM, Parag V, Bennett DA, Suh I, Lam TH, Whitlock G, type 2 diabetes that could not be controlled with diet ther-
Barzi F, Woodward M, Asia Pacific Cohort Studies apy. Ann Intern Med. 1998;128:165–175.
Collaboration. Blood glucose and risk of cardiovascular disease 22. Kahn SE, Haffner SM, Heise MA, Herman WH, Holman
in the Asia Pacific region. Diabetes Care. 2004;27:2836–2842. RR, Jones NP, Kravitz BG, Lachin JM, O’Neill MC, Zinman
12. Tancredi M, Rosengren A, Svensson AM, Kosiborod M, B, Viberti G. Glycemic durability of rosiglitazone, met-
Pivodic A, Gudbjornsdottir S, Wedel H, Clements M, formin, or glyburide monotherapy. N Engl J Med. 2006;
Dahlqvist S, Lind M. Excess mortality among persons with 355:2427–2443.
type 2 diabetes. N Engl J Med. 2015;373:1720–1732. 23. The University Group Diabetes Program. A study of the
13. Chalmers J, Cooper ME. UKPDS and the legacy effect. N effects of hypoglycemic agents on vascular complications in
Engl J Med. 2008;359:1618–1620. patients with adult‐onset diabetes. V. Evaluation of pheni-
14. Group AC, Patel A, MacMahon S, Chalmers J, Neal B, Billot L, formin therapy. Diabetes. 1975;24(Suppl 1):65–184.
Woodward M, Marre M, Cooper M, Glasziou P, Grobbee D, 24. Monami M, Genovese S, Mannucci E. Cardiovascular safety
Hamet P, Harrap S, Heller S, Liu L, Mancia G, Mogensen CE, of sulfonylureas: a meta‐analysis of randomized clinical tri-
Pan C, Poulter N, Rodgers A, Williams B, Bompoint S, de als. Diabetes Obes Metab. 2013;15:938–953.
Galan BE, Joshi R, Travert F. Intensive blood glucose control 25. Soccio RE, Chen ER, Lazar MA. Thiazolidinediones and the
and vascular outcomes in patients with type 2 diabetes. N promise of insulin sensitization in type 2 diabetes. Cell
Engl J Med. 2008;358:2560–2572. Metab. 2014;20:573–591.
15. Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N, 26. Staels B, Fruchart JC. Therapeutic roles of peroxisome pro-
Reaven PD, Zieve FJ, Marks J, Davis SN, Hayward R, Warren liferator‐activated receptor agonists. Diabetes. 2005;54:
SR, Goldman S, McCarren M, Vitek ME, Henderson WG, 2460–2470.
Huang GD, VADT Investigators. Glucose control and vascu- 27. Inzucchi SE, Maggs DG, Spollett GR, Page SL, Rife FS,
lar complications in veterans with type 2 diabetes. N Engl Walton V, Shulman GI. Efficacy and metabolic effects of
J Med. 2009;360:129–139. metformin and troglitazone in type II diabetes mellitus. N
16. Gerstein HC, Miller ME, Byington RP, Goff DC, Jr., Bigger Engl J Med. 1998;338:867–872.
JT, Buse JB, Cushman WC, Genuth S, Ismail‐Beigi F, Grimm 28. Vella A, de Groen PC, Dinneen SF. Fatal hepatotoxicity asso-
RH, Jr., Probstfield JL, Simons‐Morton DG, Friedewald WT. ciated with troglitazone. Ann Intern Med. 1998;129:1080.
Effects of intensive glucose lowering in type 2 diabetes. 29. Nesto RW, Bell D, Bonow RO, Fonseca V, Grundy SM,
N Engl J Med. 2008;358:2545–2559. Horton ES, Le Winter M, Porte D, Semenkovich CF, Smith S,
17. Meigs JB, Nathan DM, D’Agostino RB, Sr., Wilson PW. Young LH, Kahn R. Thiazolidinedione use, fluid retention,
Fasting and postchallenge glycemia and cardiovascular dis- and congestive heart failure: a consensus statement from the
ease risk: the Framingham Offspring Study. Diabetes Care. American Heart Association and American Diabetes
2002;25:1845–1850. Association. Diabetes Care. 2004;27:256–263.
30. Khan MA, St Peter JV, Xue JL. A prospective, randomized patients with type 2 diabetes and previous myocardial
comparison of the metabolic effects of pioglitazone or infarction: results from the PROactive (PROactive 05) Study.
rosiglitazone in patients with type 2 diabetes who were pre- J Am Coll Cardiol. 2007;49:1772–1780.
viously treated with troglitazone. Diabetes Care. 2002;25: 38. Wilcox R, Bousser MG, Betteridge DJ, Schernthaner G,
708–711. Pirags V, Kupfer S, Dormandy J, Investigators PR. Effects of
31. Nissen SE, Wolski K. Effect of rosiglitazone on the risk of pioglitazone in patients with type 2 diabetes with or without
myocardial infarction and death from cardiovascular causes. previous stroke: results from PROactive (PROspective piogl-
N Engl J Med. 2007;356:2457–2471. itAzone Clinical Trial In macroVascular Events 04). Stroke.
32. Home PD, Pocock SJ, Beck‐Nielsen H, Curtis PS, Gomis R, 2007;38:865–873.
Hanefeld M, Jones NP, Komajda M, McMurray JJ, Team RS. 39. Kernan WN, Viscoli CM, Furie KL, Young LH, Inzucchi SE,
Rosiglitazone evaluated for cardiovascular outcomes in oral Gorman M, Guarino PD, Lovejoy AM, Peduzzi PN, Conwit
agent combination therapy for type 2 diabetes (RECORD): a R, Brass LM, Schwartz GG, Adams HP, Jr., Berger L, Carolei
multicentre, randomised, open‐label trial. Lancet. 2009;373: A, Clark W, Coull B, Ford GA, Kleindorfer D, O’Leary JR,
2125–2135. Parsons MW, Ringleb P, Sen S, Spence JD, Tanne D, Wang D,
33. Bach RG, Brooks MM, Lombardero M, Genuth S, Donner Winder TR, for the IRIS Trial Investigators. Pioglitazone
TW, Garber A, Kennedy L, Monrad ES, Pop‐Busui R, Kelsey after ischemic stroke or transient ischemic attack. N Engl
SF, Frye RL, Investigators BD. Rosiglitazone and outcomes for J Med. 2016;374:1321–1331.
patients with diabetes mellitus and coronary artery disease in 40. Lewis JD, Ferrara A, Peng T, Hedderson M, Bilker WB,
the Bypass Angioplasty Revascularization Investigation Quesenberry CP, Jr., Vaughn DJ, Nessel L, Selby J, Strom BL.
2 Diabetes (BARI 2D) trial. Circulation. 2013;128: Risk of bladder cancer among diabetic patients treated with
785–794. pioglitazone: interim report of a longitudinal cohort study.
34. Mazzone T, Meyer PM, Feinstein SB, Davidson MH, Kondos Diabetes Care. 2011;34:916–922.
GT, D’Agostino RB, Sr., Perez A, Provost JC, Haffner SM. 41. Erdmann E, Song E, Spanheimer R, van Troostenburg de
Effect of pioglitazone compared with glimepiride on carotid Bruyn AR, Perez A. Observational follow‐up of the PROactive
intima‐media thickness in type 2 diabetes: a randomized study: a 6‐year update. Diabetes Obes Metab. 2014;16:
trial. JAMA. 2006;296:2572–2581. 63–74.
35. Nissen SE, Nicholls SJ, Wolski K, Nesto R, Kupfer S, Perez A, 42. Ferwana M, Firwana B, Hasan R, Al‐Mallah MH, Kim S,
Jure H, De Larochelliere R, Staniloae CS, Mavromatis K, Saw Montori VM, Murad MH. Pioglitazone and risk of bladder
J, Hu B, Lincoff AM, Tuzcu EM, Investigators P. Comparison cancer: a meta‐analysis of controlled studies. Diabet Med.
of pioglitazone vs glimepiride on progression of coronary 2013;30:1026–1032.
atherosclerosis in patients with type 2 diabetes: the 43. Brauer CA, Coca‐Perraillon M, Cutler DM, Rosen AB.
PERISCOPE randomized controlled trial. JAMA. 2008;299: Incidence and mortality of hip fractures in the United States.
1561–1573. JAMA. 2009;302:1573–1579.
36. Dormandy JA, Charbonnel B, Eckland DJ, Erdmann E, 44. Seifarth C, Bergmann J, Holst JJ, Ritzel R, Schmiegel W,
Massi‐Benedetti M, Moules IK, Skene AM, Tan MH, Nauck MA. Prolonged and enhanced secretion of glucagon‐
Lefebvre PJ, Murray GD, Standl E, Wilcox RG, Wilhelmsen like peptide 1 (7‐36 amide) after oral sucrose due to alpha‐
L, Betteridge J, Birkeland K, Golay A, Heine RJ, Koranyi L, glucosidase inhibition (acarbose) in Type 2 diabetic patients.
Laakso M, Mokan M, Norkus A, Pirags V, Podar T, Scheen Diabet Med. 1998;15:485–491.
A, Scherbaum W, Schernthaner G, Schmitz O, Skrha J, Smith 45. DeFronzo RA. Pharmacologic therapy for type 2 diabetes
U, Taton J. Secondary prevention of macrovascular events in mellitus. Ann Intern Med. 1999;131:281–303.
patients with type 2 diabetes in the PROactive Study 46. Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A,
(PROspective pioglitAzone Clinical Trial In macroVascular Laakso M, STOP‐NIDDM Trial Research Group. Acarbose
Events): a randomised controlled trial. Lancet. 2005;366: treatment and the risk of cardiovascular disease and hyper-
1279–1289. tension in patients with impaired glucose tolerance: the
37. Erdmann E, Dormandy JA, Charbonnel B, Massi‐Benedetti STOP‐NIDDM trial. JAMA. 2003;290:486–494.
M, Moules IK, Skene AM, Investigators PR. The effect of 47. Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A,
pioglitazone on recurrent myocardial infarction in 2,445 Laakso M. Acarbose for the prevention of Type 2 diabetes,
hypertension and cardiovascular disease in subjects with 57. Vella A, Lee JS, Camilleri M, Szarka LA, Burton DD,
impaired glucose tolerance: facts and interpretations con- Zinsmeister AR, Rizza RA, Klein PD. Effects of pramlintide,
cerning the critical analysis of the STOP‐NIDDM Trial data. an amylin analogue, on gastric emptying in type 1 and 2 dia-
Diabetologia. 2004;47:969–975; discussion 976–967. betes mellitus. Neurogastroenterol Motil. 2002;14:123–131.
48. Holman RR, Coleman RL, Chan JCN, Chiasson JL, Feng H, 58. Samsom M, Szarka LA, Camilleri M, Vella A, Zinsmeister
Ge J, Gerstein HC, Gray R, Huo Y, Lang Z, McMurray JJ, AR, Rizza RA. Pramlintide, an amylin analog, selectively
Ryden L, Schroder S, Sun Y, Theodorakis MJ, Tendera M, delays gastric emptying: potential role of vagal inhibition.
Tucker L, Tuomilehto J, Wei Y, Yang W, Wang D, Hu D, Pan Am J Physiol Gastrointest Liver Physiol. 2000;278:G946–951.
C, ACE Study Group. Effects of acarbose on cardiovascular 59. Hoogwerf BJ, Doshi KB, Diab D. Pramlintide, the synthetic
and diabetes outcomes in patients with coronary heart dis- analogue of amylin: physiology, pathophysiology, and effects on
ease and impaired glucose tolerance (ACE): a randomised, glycemic control, body weight, and selected biomarkers of
double‐blind, placebo‐controlled trial. Lancet Diabetes vascular risk. Vascular Health & Risk Management. 4:355–362.
Endocrinol. 2017;5:877–886. 60. Gaziano JM, Cincotta AH, O’Connor CM, Ezrokhi M, Rutty
49. Hsu PF, Sung SH, Cheng HM, Shin SJ, Lin KD, Chong K, Yen D, Ma ZJ, Scranton RE. Randomized clinical trial of quick‐
FS, Yu BH, Huang CT, Hsu CC. Cardiovascular benefits of release bromocriptine among patients with type 2 diabetes
acarbose vs sulfonylureas in patients with type 2 diabetes on overall safety and cardiovascular outcomes. Diabetes
treated with metformin. Journal of Clinical Endocrinology & Care. 2010;33:1503–1508.
Metabolism. 103:3611–3619. 61. Cincotta AH, Meier AH, Cincotta M, Jr. Bromocriptine
50. Staels B, Fonseca VA. Bile acids and metabolic regulation: improves glycaemic control and serum lipid profile in obese
mechanisms and clinical responses to bile acid sequestra- Type 2 diabetic subjects: a new approach in the treatment of
tion. Diabetes Care. 2009;32 Suppl 2:S237–245. diabetes. Expert Opin Investig Drugs. 1999;8:1683–1707.
51. Zieve FJ, Kalin MF, Schwartz SL, Jones MR, Bailey WL. 62. DeFronzo RA. Bromocriptine: a sympatholytic, d2‐dopa-
Results of the glucose‐lowering effect of WelChol study mine agonist for the treatment of type 2 diabetes. Diabetes
(GLOWS): a randomized, double‐blind, placebo‐controlled Care. 2011;34:789–794.
pilot study evaluating the effect of colesevelam hydrochlo- 63. Pijl H, Ohashi S, Matsuda M, Miyazaki Y, Mahankali A,
ride on glycemic control in subjects with type 2 diabetes. Kumar V, Pipek R, Iozzo P, Lancaster JL, Cincotta AH,
Clin Ther. 2007;29:74–83. DeFronzo RA. Bromocriptine: a novel approach to the treat-
52. Fonseca VA, Rosenstock J, Wang AC, Truitt KE, Jones MR. ment of type 2 diabetes. Diabetes Care. 2000;23:1154–1161.
Colesevelam HCl improves glycemic control and reduces 64. Kieffer TJ. Gastro‐intestinal hormones GIP and GLP‐1. Ann
LDL cholesterol in patients with inadequately controlled Endocrinol (Paris). 2004;65:13–21.
type 2 diabetes on sulfonylurea‐based therapy. Diabetes 65. Drucker DJ. Evolving concepts and translational relevance
Care. 2008;31:1479–1484. of enteroendocrine cell biology. Journal of Clinical
53. Beysen C, Murphy EJ, Deines K, Chan M, Tsang E, Glass A, Endocrinology and Metabolism. 2016;101(3):778–786.
Turner SM, Protasio J, Riiff T, Hellerstein MK. Effect of bile acid 66. Vella A, Bock G, Giesler PD, Burton DB, Serra DB, Saylan
sequestrants on glucose metabolism, hepatic de novo lipogen- ML, Deacon CF, Foley JE, Rizza RA, Camilleri M. The effect
esis, and cholesterol and bile acid kinetics in type 2 diabetes: a of dipeptidyl peptidase‐4 inhibition on gastric volume, satia-
randomised controlled study. Diabetologia. 2012;55:432–442. tion and enteroendocrine secretion in type 2 diabetes: a double‐
54. Smushkin G, Sathananthan M, Piccinini F, Dalla Man C, blind, placebo‐controlled crossover study. Clin Endocrinol
Law JH, Cobelli C, Zinsmeister AR, Rizza RA, Vella A. The (Oxf). 2008;69:737–744.
effect of a bile acid sequestrant on glucose metabolism in 67. Vella A, Bock G, Giesler PD, Burton DB, Serra DB, Saylan
subjects with type 2 diabetes. Diabetes. 2013;62:1094–1101. ML, Dunning BE, Foley JE, Rizza RA, Camilleri M. Effects of
55. Schwab P, Louder A, Li Y, Mallick R, Bays H. Cholesterol dipeptidyl peptidase‐4 inhibition on gastrointestinal func-
treatment patterns and cardiovascular clinical outcomes tion, meal appearance, and glucose metabolism in type 2
associated with colesevelam HCl and ezetimibe. Drugs Aging diabetes. Diabetes. 2007;56:1475–1480.
2014;31:683–694. 68. Holst JJ. Implementation of GLP‐1 based therapy of type 2
56. Ryan GJ, Jobe LJ, Martin R. Pramlintide in the treatment of type diabetes mellitus using DPP‐IV inhibitors. Adv Exp Med
1 and type 2 diabetes mellitus. Clin Ther. 2005;27:1500–1512. Biol. 2003;524:263–279.
69. Kieffer TJ, Habener JF. The glucagon‐like peptides. Endocr LEADER Trial Investigators. Liraglutide and cardiovascular
Rev. 1999;20:876–913. outcomes in type 2 diabetes. N Engl J Med. 2016;375:311–322.
70. Fadini GP, Avogaro A. Cardiovascular effects of DPP‐4 79. Pfeffer MA, Claggett B, Diaz R, Dickstein K, Gerstein HC,
inhibition: beyond GLP‐1. Vascul Pharmacol. 2011;55: Kober LV, Lawson FC, Ping L, Wei X, Lewis EF, Maggioni
10–16. AP, McMurray JJ, Probstfield JL, Riddle MC, Solomon SD,
71. Brown NJ, Byiers S, Carr D, Maldonado M, Warner BA. Tardif JC, for the ELIXA Investigators. Lixisenatide in patients
Dipeptidyl peptidase‐IV inhibitor use associated with with type 2 diabetes and acute coronary syndrome. N Engl J
increased risk of ACE inhibitor‐associated angioedema. Med. 2015;373:2247–2257.
Hypertension. 2009;54:516–523. 80. Holman RR, Bethel MA, Mentz RJ, Thompson VP,
72. Baetta R, Corsini A. Pharmacology of dipeptidyl pepti- Lokhnygina Y, Buse JB, Chan JC, Choi J, Gustavson SM,
dase‐4 inhibitors: similarities and differences. Drugs. Iqbal N, Maggioni AP, Marso SP, Ohman P, Pagidipati NJ,
2011;71:1441–1467. Poulter N, Ramachandran A, Zinman B, Hernandez AF, for
73. Scirica BM, Bhatt DL, Braunwald E, Steg PG, Davidson J, the EXSCEL Study Group. Effects of once‐weekly exenatide
Hirshberg B, Ohman P, Frederich R, Wiviott SD, Hoffman on cardiovascular outcomes in type 2 diabetes. N Engl J Med.
EB, Cavender MA, Udell JA, Desai NR, Mosenzon O, 2017;377:1228–1239.
McGuire DK, Ray KK, Leiter LA, Raz I, for the SAVOR‐ 81. Marso SP, Bain SC, Consoli A, Eliaschewitz FG, Jodar E,
TIMI 53 Steering Committee and Investigators. Saxagliptin Leiter LA, Lingvay I, Rosenstock J, Seufert J, Warren ML,
and cardiovascular outcomes in patients with type 2 diabetes Woo V, Hansen O, Holst AG, Pettersson J, Vilsboll T, for the
mellitus. N Engl J Med. 2013;369:1317–1326. SUSTAIN‐6 Investigators. Semaglutide and cardiovascular
74. White WB, Cannon CP, Heller SR, Nissen SE, Bergenstal outcomes in patients with type 2 diabetes. N Engl J Med.
RM, Bakris GL, Perez AT, Fleck PR, Mehta CR, Kupfer S, 2016;375:1834–1844.
Wilson C, Cushman WC, Zannad F, for the EXAMINE 82. Hernandez AF, Green JB, Janmohamed S, D’Agostino RB,
Investigators. Alogliptin after acute coronary syndrome in Sr., Granger CB, Jones NP, Leiter LA, Rosenberg AE, Sigmon
patients with type 2 diabetes. N Engl J Med. 2013;369: KN, Somerville MC, Thorpe KM, McMurray JJV, Del Prato
1327–1335. S, Harmony Outcomes committees and investigators.
75. Green JB, Bethel MA, Armstrong PW, Buse JB, Engel SS, Albiglutide and cardiovascular outcomes in patients with
Garg J, Josse R, Kaufman KD, Koglin J, Korn S, Lachin JM, type 2 diabetes and cardiovascular disease (Harmony
McGuire DK, Pencina MJ, Standl E, Stein PP, Suryawanshi S, Outcomes): a double‐blind, randomised placebo‐controlled
Van de Werf F, Peterson ED, Holman RR, for the TECOS trial. Lancet. 2018;392:1519–1529.
Study Group. Effect of sitagliptin on cardiovascular out- 83. Gerstein HC, Colhoun HM, Dagenais GR, Diaz R,
comes in type 2 diabetes. N Engl J Med. 2015;373:232–242. Lakshmanan M, Pais P, Probstfield J, Riesmeyer JS, Riddle
76. McMurray JJV, Ponikowski P, Bolli GB, Lukashevich V, MC, Ryden L, Xavier D, Atisso CM, Dyal L, Hall S, Rao‐
Kozlovski P, Kothny W, Lewsey JD, Krum H, VIVIDD Trial Melacini P, Wong G, Avezum A, Basile J, Chung N, Conget I,
Committees and Investigators. Effects of vildagliptin on ven- Cushman WC, Franek E, Hancu N, Hanefeld M, Holt S,
tricular function in patients with type 2 diabetes mellitus Jansky P, Keltai M, Lanas F, Leiter LA, Lopez‐Jaramillo P,
and heart failure: a randomized placebo‐controlled trial. Cardona Munoz EG, Pirags V, Pogosova N, Raubenheimer
JACC Heart Fail. 2018;6:8–17. PJ, Shaw JE, Sheu WH, Temelkova‐Kurktschiev T, for the
77. Scirica BM, Mosenzon O, Bhatt DL, Udell JA, Steg PG, REWIND Investigators. Dulaglutide and cardiovascular
McGuire DK, Im K, Kanevsky E, Stahre C, Sjostrand M, Raz outcomes in type 2 diabetes (REWIND): a double‐blind,
I, Braunwald E. Cardiovascular outcomes according to uri- randomised placebo‐controlled trial. Lancet.
nary albumin and kidney disease in patients with type 2 dia- 2019;394:121–130.
betes at high cardiovascular risk: observations from the 84. Perkins BA, Udell JA, Cherney DZ. No need to sugarcoat the
SAVOR‐TIMI 53 trial. JAMA Cardiol. 2018;3:155–163. message: is cardiovascular risk reduction from SGLT2 inhi-
78. Marso SP, Daniels GH, Brown‐Frandsen K, Kristensen P, Mann bition related to natriuresis? Am J Kidney Dis. 2016;68:
JF, Nauck MA, Nissen SE, Pocock S, Poulter NR, Ravn LS, 349–352.
Steinberg WM, Stockner M, Zinman B, Bergenstal RM, Buse 85. Heerspink HJ, Perkins BA, Fitchett DH, Husain M, Cherney
JB, for the LEADER Steering Committee on behalf of the DZ. Sodium glucose cotransporter 2 inhibitors in the treat-
ment of diabetes mellitus: cardiovascular and kidney effects, 94. Mahaffey KW, Jardine MJ, Bompoint S, Cannon CP, Neal B,
potential mechanisms, and clinical applications. Circulation. Heerspink HJL, Charytan DM, Edwards R, Agarwal R,
2016;134:752–772. Bakris G, Bull S, Capuano G, de Zeeuw D, Greene T, Levin
86. Neal B, Perkovic V, Matthews DR. Canagliflozin and cardio- A, Pollock C, Sun T, Wheeler DC, Yavin Y, Zhang H, Zinman
vascular and renal events in type 2 diabetes. N Engl J Med. B, Rosenthal N, Brenner BM, Perkovic V. Canagliflozin and
2017;377:2099. cardiovascular and renal outcomes in type 2 diabetes melli-
87. Fralick M, Schneeweiss S, Patorno E. Risk of diabetic ketoac- tus and chronic kidney disease in primary and secondary
idosis after initiation of an SGLT2 inhibitor. N Engl J Med. cardiovascular prevention groups. Circulation. 2019;140:
2017;376:2300–2302. 739–750.
88. Zinman B, Lachin JM, Inzucchi SE. Empagliflozin, cardio- 95. Wiviott SD, Raz I, Bonaca MP, Mosenzon O, Kato ET, Cahn
vascular outcomes, and mortality in type 2 diabetes. N Engl J A, Silverman MG, Zelniker TA, Kuder JF, Murphy SA, Bhatt
Med. 2016;374:1094. DL, Leiter LA, McGuire DK, Wilding JPH, Ruff CT, Gause‐
89. Mahaffey KW, Neal B, Perkovic V, de Zeeuw D, Fulcher G, Nilsson IAM, Fredriksson M, Johansson PA, Langkilde AM,
Erondu N, Shaw W, Fabbrini E, Sun T, Li Q, Desai M, Sabatine MS, for the DECLARE–TIMI 58 Investigators.
Matthews DR, CANVAS Program Collaborative Group. Dapagliflozin and cardiovascular outcomes in type 2 diabe-
Canagliflozin for primary and secondary prevention of car- tes. N Engl J Med. 2019;380:347–357.
diovascular events: results from the CANVAS Program 96. Scheen AJ. Cardiovascular effects of new oral glucose‐lower-
(Canagliflozin Cardiovascular Assessment Study). Circulation. ing agents: DPP‐4 and SGLT‐2 inhibitors. Circ Res.
2018;137:323–334. 2018;122:1439–1459.
90. Mudaliar S, Alloju S, Henry RR. Can a shift in fuel energetics 97. McMurray JJV, Solomon SD, Inzucchi SE, Kober L,
explain the beneficial cardiorenal outcomes in the EMPA‐ Kosiborod MN, Martinez FA, Ponikowski P, Sabatine MS,
REG OUTCOME Study? A unifying hypothesis. Diabetes Anand IS, Belohlavek J, Bohm M, Chiang CE, Chopra VK,
Care. 2016;39:1115–1122. de Boer RA, Desai AS, Diez M, Drozdz J, Dukat A, Ge J,
91. Ferrannini E, Baldi S, Frascerra S, Astiarraga B, Heise T, Howlett JG, Katova T, Kitakaze M, Ljungman CEA, Merkely
Bizzotto R, Mari A, Pieber TR, Muscelli E. Shift to fatty sub- B, Nicolau JC, O’Meara E, Petrie MC, Vinh PN, Schou M,
strate utilization in response to sodium‐glucose cotrans- Tereshchenko S, Verma S, Held C, DeMets DL, Docherty KF,
porter 2 inhibition in subjects without diabetes and patients Jhund PS, Bengtsson O, Sjostrand M, Langkilde AM, for the
with type 2 diabetes. Diabetes. 2016;65:1190–1195. DAPA‐HF Trial Committees and Investigators. Dapagliflozin
92. Ferrannini E, Muscelli E, Frascerra S, Baldi S, Mari A, Heise in patients with heart failure and reduced ejection fraction.
T, Broedl UC, Woerle H‐J. Metabolic response to sodium‐ N Engl J Med. 2019;381:1995–2008.
glucose cotransporter 2 inhibition in type 2 diabetic patients. 98. Arioglu E, Duncan‐Morin J, Sebring N, Rother KI, Gottlieb
Journal of Clinical Investigation. 2014;124:499–508. N, Lieberman J, Herion D, Kleiner DE, Reynolds J,
93. Verma S, Mazer CD, Yan AT, Mason T, Garg V, Teoh H, Zuo Premkumar A, Sumner AE, Hoofnagle J, Reitman ML,
F, Quan A, Farkouh ME, Fitchett DH, Goodman SG, Taylor SI. Efficacy and safety of troglitazone in the treatment
Goldenberg RM, Al‐Omran M, Gilbert RE, Bhatt DL, Leiter of lipodystrophy syndromes. Ann Intern Med. 2000;133:
LA, Juni P, Zinman B, Connelly KA. Effect of empagliflozin 263–274.
on left ventricular mass in patients with type 2 diabetes mel- 99. Hegele RA, Joy TR, Al‐Attar SA, Rutt BK. Thematic review
litus and coronary artery disease: the EMPA‐HEART series: adipocyte biology. Lipodystrophies: windows on adi-
CardioLink‐6 randomized clinical trial. Circulation. pose biology and metabolism. J Lipid Res. 2007;48:
2019;140:1693–1702. 1433–1444.
15 Choosing Medications for Weight Loss
in Type 2 Diabetes Mellitus
Shubhada Jagasia1, Chase Dean Hendrickson2 and Alexander J. Williams3
1
rofessor of Medicine, Division of Diabetes, Endocrinology and Metabolism, Vanderbilt University Medical Center,
P
Nashville, TN, USA
2
ssistant Professor of Medicine, Division of Diabetes, Endocrinology, and Metabolism, Vanderbilt University Medical
A
Center, Nashville, TN, USA
3
ellow, Division of Diabetes, Endocrinology, and Metabolism, Vanderbilt University Medical Center, Nashville,
F
TN, USA
classified as obesity class 1, BMI 35.0–39.9 is classified as

obesity class 2, and BMI of 40.0 or more is classified as obe
●● Intensive lifestyle interventions in people with type 2
sity class 3 [2]. In the National Health and Nutrition
diabetes result in weight loss and are associated with Examination Survey (NHANES), obesity is defined as a
significant improvements in glycemic control and BMI of 30 or more, and severe obesity is defined as a BMI
cardiovascular disease risk factors. of 40 or more. In adults 20 years old or greater, the preva
●● Therapy for type 2 diabetes and improved glycemic lence of obesity increased from 33.7% in 2007–2008 to
control can result in weight gain.
●● Newer therapies for type 2 diabetes can decrease
39.6% in 2015–2016, while the prevalence of severe obesity
caloric intake or increase caloric wastage causing also increased, from 5.7% in 2007–2008 to 7.7% in
weight loss or at minimum prevent weight gain while 2015–2016 [3]. Abdominal obesity has been described as a
improving glycemic control stronger independent predictor for type 2 diabetes [4]. The
prevalence of type 2 diabetes in people with abdominal
obesity increased from 16.0% in 1999–2000 to 20.1% in
2013–2014, compared to 6.2% to 6.3% in non‐obese adults
45–64 years old [5].
Introduction: Prevalence and relationship
The 20‐year Nurses’ Health Study followed 84 941
of obesity and type 2 diabetes
female nurses from 1980–1996 and found that the risk of
Type 2 diabetes develops with increasing frequency at developing diabetes was 38.8 times greater if BMI was 35.0
higher weight and body mass index (BMI). Even when or more and 20.1 times greater if BMI was 30.0–34.9, com
compared to other known risk factors such as lack of exer pared to a BMI less than 23.0 [1]. The data from the
cise, smoking and poor diet, elevated BMI as a marker for NHANES study between 1999–2006 also showed the prev
excess body fat remains the most important risk factor for alence of diabetes increased with BMI: 8% with normal
developing type 2 diabetes [1]. BMI, 15% in overweight, 23% in obesity class 1, 33% in
The National Heart, Lung, and Blood Institute bases obesity class 2, and 43% in obesity class 3 [6]. These studies
obesity definitions on BMI: BMI < 18.5 is classified as show that the prevalence of both obesity and type 2 diabe
underweight, BMI 18.5–24.9 is classified as normal weight, tes are continuing to increase and that the risk of deve
BMI 25.0–29.9 is classified as overweight, BMI 30.0–34.9 is loping diabetes is strongly tied to excess body weight.

187
The Look AHEAD (Action for Health in Diabetes) trial thiazolidinediones (TZDs), and insulin were all shown to
demonstrated that after one year, intensive lifestyle inter reduce the risk of microvascular and macrovascular com
ventions in people with type 2 diabetes resulted in weight plications but also gave insight into the effects of different
loss (8.6% of their initial weight) and was associated with medications classes on body weight in people with diabe
significant improvements in glycemic control and risk fac tes [10]. In the ACCORD trial, the intensive arm gained an
tors for cardiovascular disease (CVD). However, the differ average of 3.5 kg compared to 0.4 kg in the conventional
ences in cardiovascular morbidity and mortality between group, thought to be mostly due to insulin and TZDs [11].
the intervention and control group were not significant at Table 15.1 summarizes some of the pertinent literature on
10 years. The patients in the intervention group that lost the effects of pharmacotherapy for type 2 diabetes on
weight were prescribed fewer glucose‐lowering medica weight and can be used as a reference to accompany the
tions and had a lower glycated hemoglobin (A1c). In the discussion below.
intervention group about one‐third of patients were unable
to attain the goal of 5% weight loss after 1 year, and of all Insulin
the patients that lost about 5% of their body weight about The weight gain associated with insulin use has been
half regained some or all of their weight by year 8 [7]. explained by various mechanisms, primarily by the increase
Given this relationship between diabetes and obesity, it in glucose utilization in peripheral tissues and glycemic
is important to consider the effects of glucose‐lowering levels that fall below the renal threshold for excretion, lead
therapy on weight when choosing medications to treat ing to increased conservation of ingested calories [12].
diabetes. Insulin has also been shown to inhibit protein catabolism
and promote lipogenesis [13]. Another suggested explana
tion is that insulin resistance in the central nervous system
Therapeutic classes for the treatment
causes dysregulation of catabolic and anorexic neuronal
of type 2 Diabetes and their effects
responses [14]. Patients’ fears of hypoglycemia may also
on weight
contribute to weight gain via increased intake of
Understanding the intimate relationship between weight calories [15].
gain and type 2 diabetes led to an appropriate focus on In the UKPDS, patients treated with insulin monother
weight loss to help prevent and treat type 2 diabetes. The apy gained an average of 4.0 kg during the 10‐year study
American Diabetes Association (ADA) recommends that period, with most of the weight gain occurring in the first
patients with prediabetes engage in lifestyle changes to 12 months [16]. Henry et al. found that patients with type
achieve and maintain 7–10% loss of initial body weight to 2 diabetes treated with an intensive insulin regimen to
reduce the risk of conversion to type 2 diabetes [8]. The obtain tight glycemic control over a 6‐month study period,
first intervention for most overweight/obese patients with the average weight gained was approximately 9 kg [17].
diabetes is to improve their diet, increase physical activity, Yki‐Jarvinen showed that weight gain was more substantial
and ultimately reduce caloric intake. However, diet‐ in obese people with type 2 diabetes treated with insulin,
induced weight loss has been shown to increase orexigenic with an average 5.1 kg increase in the 12‐month study
hormones such as ghrelin and gastric inhibitory polypep period [18].
tide (GIP) and decrease anorexigenic hormones such as
leptin, amylin, cholecystokinin (CCK), and glucagon‐like Sulfonylureas
peptide 1 (GLP‐1) [9]. Elucidating this hormonal dysregu Sulfonylureas work by stimulating pancreatic beta‐cells to
lation, coupled with the fact that some of the earliest glu secrete insulin, and, therefore, the mechanisms by which
cose‐lowering medications, including insulin, cause weight they cause weight gain are thought to be similar to exoge
gain, led to changes in the suggested pharmaco‐therapeutic nous insulin administration [19]. In the UKPDS, patients
algorithms for type 2 diabetes. In the UK Prospective given chlorpropamide or glibenclamide had an average of
Diabetes Study (UKPDS), sulfonylureas, biguanides, 2.15 kg (1.7 for glibenclamide and 2.6 for chlorpropamide)
TABLE 15.1
Study (reference Average net Duration of Average A1C Primary side

No.) weight change (kg) treatment change (%) effects Comparator
Insulin
UKPDS (16) +4.0 10 years ‐0.9 Hypoglycemia, compared to placebo/ lifestyle
Henry RR (17) +9.0 6 months NR weight gain compared to placebo/ lifestyle
Yki (18) +5.1 1 year ‐2.1 compared to placebo/ lifestyle
Sulfonylureas
UKPDS (16) +2.2 10 years ‐0.9 Hypoglycemia, compared to placebo
Advance (20) +0.7 5 years NR weight gain compared to placebo
Bautista (21) +4.4 3 months ‐1.1 compared to placebo
Simonson (22) ~0 16 weeks ‐1.5 compared to placebo
Metformin
DPPOS (26) ‐1.5 10 years ‐0.2 Diarrhea, compared to placebo
DeFronzo (25) ‐3.8 29 weeks ‐1.5 nausea compared to glyburide and
placebo
Goldstein (27) ‐2.1 18 weeks ‐0.1 compared to glipizide
Thiazolidinediones
Miyazaki (29) +3 16 weeks ‐1.4 Edema, compared to placebo
Basu (30) +2.6 12 weeks +0.4 weight gain compared to glipizide
Alpha‐glucosidase
inhibitors
Feinbock (34) ‐1.5 20 weeks ‐1.8 Flatulence compared to glimepiride
Lindstrom (35) ~0 24 weeks ‐0.9 compared to placebo
Wolever (36) ‐1.0 1 year ‐1.4 compared to placebo
Phillips (33) ‐0.9 24 weeks ‐0.2 compared to placebo
Amylin
Riddle (39) ‐2.3 16 weeks ‐0.7 Nausea compared to placebo
Thompson (38) ‐0.9 4 weeks ‐0.6 compared to placebo
Hollander (40) ‐2.1 52 weeks ‐0.7 compared to placebo
Hollander (41) ‐1.8 26 weeks ‐0.6 compared to placebo
GLP‐1 receptor agonists
Buse (43) ‐1.6 30 weeks ‐0.8 Nausea, compared to placebo
DeFronzo (44) ‐2.8 30 weeks ‐1.0 vomiting, compared to placebo
Kendall (45) ‐0.5 30 weeks ‐0.8 diarrhea compared to placebo
Heine (46) ‐4.1 26 weeks ‐1.1 compared to insulin glargine
Zinman (47) ‐2.0 26 weeks ‐1.6 compared to placebo
Garber (48) ‐3.3 52 weeks ‐0.8 compared to glimepiride
Kim (49) ‐3.8 15 weeks ‐1.4 compared to placebo
Davies (50) ‐6.4 26 weeks ‐1.9 compared to placebo
Henry (51) ‐1.2 26 weeks ‐2.8 compared to placebo
DPP‐4 inhibitors
Ristic (53) ‐0.1 4 weeks ‐0.5 Nausea compared to placebo
Goldstein (55) +1.0 24 weeks ‐2.9 compared to placebo
Bolli (54) ~0 24 weeks ‐0.9 compared to placebo
Ahren (56) ‐2.0 104 weeks ‐0.4 compared to glimepiride
SGLT‐2 inhibitors
Bolinder (58) ‐4.5 102 weeks ‐0.3 Urinary tract compared to placebo
Sakai (61) ‐2.6 52 weeks ‐0.7 infections, compared to placebo
Leiter (60) ‐4.4 104 weeks ‐0.6 genital compared to glimepiride
Lavelle‐Gonzalez (59) ‐2.5 52 weeks ‐0.8 infections compared to sitagliptin
Kawata (62) ‐2.6 24 weeks ‐0.8 compared to placebo
Zinman (63) ‐3.0 3.1 years ‐0.5 compared to placebo
Cherney (64) ‐2.1 24 weeks ‐0.9 compared to placebo
weight gain over the 10‐year study period [16]. The to be at least partially due to plasma volume expansion and
ADVANCE trial saw a smaller weight gain of 0.71 kg (gli peripheral edema [27]. Miyazaki et al. reported that after
clazide) over a 5‐year period [20]. Bautista et al. demon 16 weeks of pioglitazone patients gained 3.0 kg due to an
strated that patients with type 2 diabetes taking glimepiride increase in fat mass compared to placebo in patients
gained 2.3 kg compared to a 2.1 kg decrease in the placebo already taking a sulfonylurea [28]. These findings were
group after 3 months [21]. Conversely, Simonsen et al. similar to those reported by Basu et al., where pioglitazone
reported that extended‐release glipizide was not found to increased body weight by 3.1 kg, 2.4 kg of which was total
have any significant effect on weight over a 16‐week period, body water compared to 0.5 kg in the glipizide group over
potentially related to altered pharmacokinetics that pro 12 weeks [29]. Despite the effects of weight gain, TZDs
duce smaller fluctuations in insulin levels [22]. generally have neutral effects on or even decrease the
amount of visceral fat, which may have a positive effect on
Biguanides/Metformin cardiovascular disease [30].
Metformin has been in use since the 1950s although its
mechanism of action is uncertain. It has been shown to Alpha‐glucosidase Inhibitors
decrease hepatic glucose production and increase insulin Alpha‐glucosidase inhibitors slow the rise in postprandial
sensitivity. In the UKPDS and in subsequent studies, met glucose and insulin levels by delaying glucose absorp
formin has been shown to have weight loss effects in people tion [31]. Phillips et al. demonstrated that patients taking
with type 2 diabetes [23]. DeFronzo et al. compared met acarbose lost 1.32 kg compared to 0.43 kg in the placebo
formin to glyburide and placebo and found that patients group in patients already taking metformin over a 24‐week
with type 2 diabetes taking metformin lost 3.8 kg com study period [32]. Feinbock et al. showed that patients with
pared to those taking glyburide, who experienced no sig type 2 diabetes taking acarbose lost an average of 1.9 kg
nificant change in weight. There was a 0.4 kg weight gain compared to 0.4 kg in the glimepiride group over
observed in the metformin‐glyburide group over 20 weeks [33]. Conversely, Lindstrom et al. showed that
29 weeks [24]. In the Diabetes Prevention Program patients taking acarbose had no significant weight change
Outcomes Study (DPPOS), patients taking metformin lost compared to placebo after 24 weeks [34]. Finally, Wolever
an average of 2.5 kg compared to a loss of 1 kg in the pla et al. demonstrated that after 1 year of treatment patients
cebo group over 10 years, with the placebo group subse with diabetes (not on insulin) lost 0.46 kg compared to a
quently regaining the weight while the metformin group 0.33 kg weight gain in the placebo group [35].
maintained their weight loss [25]. Goldstein et al. demon
strated in a multicenter, double‐masked, parallel‐group, Amylin Mimetics
active controlled study in which patients with type 2 diabe Amylin is a neuroendocrine hormone secreted by pancre
tes were randomized to receive glipizide, metformin, or atic beta cells along with insulin and suppresses postpran
glipizide‐metformin that patients in the metformin group dial glucagon secretion and slows gastric emptying [36].
lost an average of 2.7 kg compared to a loss of 0.4 kg in the Pramlintide is an amylin analog and reduces postprandial
glipizide and 0.3 kg in the metformin‐glipizide groups [26]. glucose levels. Thompson et al. showed a modest reduction
in weight of pramlintide (0.89 kg decrease) after 4 weeks in
Thiazolidinediones people with type 2 diabetes [37]. While Riddle et al. showed
Thiazolidinediones (TZDs) act as insulin sensitizers by that people with type 2 diabetes taking pramlintide lost an
increasing the expression of perioxisome‐proliferator‐acti average of 1.6 kg after 16 weeks compared to placebo,
vated receptor gamma (PPARγ). This is a nuclear receptor which had a 0.7 kg weight increase [38]. Hollander et al.
which binds, and is activated by, numerous naturally found a 1.4 kg weight loss in people with type 2 diabetes
occurring substances such as arachidonic acid metabolites started on pramlintide compared to a 0.7 kg weight increase
and polyunsaturated fatty acids. The improvement in insu after 52 weeks [39]. In a post hoc analysis of overweight/
lin sensitivity is also accompanied by weight gain, thought obese patients with type 2 diabetes, Hollander et al.
Choosing Medications for Weight Loss in Type 2 Diabetes Mellitus 191
reported that pramlintide given for 26‐weeks caused 9% of been administered subcutaneously, Davies et al. investigated
patients to lose more than 5% of their body weight com oral semaglutide compared to subcutaneous delivery and to
pared to 3% of patients in the placebo group. This trans placebo. Oral semaglutide had up to a 6.9 kg weight reduction
lated to an average of 1.8 kg weight loss and was greatest in at 40‐mg a day compared to a 6.4 kg weight reduction in the
patients with a BMI over 40 (3.2 kg) and in patients also subcutaneous form at 1.0 mg once a week [49]. Henry et al.
taking metformin (2.5 kg) [40]. showed that an osmotic mini‐pump that continuously deliv
ers exenatide subcutaneously reduced weight in people with
GLP‐1 Receptor agonists uncontrolled type 2 diabetes by 1.2 kg after 26 weeks [50].
Glucagon‐like peptide (GLP)‐1 is an incretin hormone
released from the intestinal L‐cell in response to nutrients DPP‐4 inhibitors
and inhibits glucagon secretion and stimulates insulin exo GLP‐1 and glucose‐dependent insulinotropic peptide
cytosis in a glucose‐dependent manner. It also reduces the (GIP) are incretins that are inactivated by dipeptidyl pepti
gastric‐emptying rate and stimulates satiety via actions at dase IV (DPP‐4). In 2003 DPP‐4 inhibitors were formu
the hypothalamus [41]. Buse et al. reported that exenatide, lated as a therapy to increase endogenous incretin
the first studied GLP‐1 receptor agonist, decreased weight activity [51]. Ristic et al. found that vildagliptin had mini
by 1.6 kg after 30 weeks compared to placebo in people mal effects on body weight compared to placebo in people
with type 2 diabetes already on a sulfonylurea [42]. with type 2 diabetes, with a loss of 0.07 kg after 4 weeks [52].
DeFronzo et al. found that the addition of exenatide to Bolli et al. also showed that vildagliptin had no effects on
metformin in people with type 2 diabetes caused a dose‐ body weight when added to metformin [53]. Similarly,
dependent decrease in weight: 2.8 kg in the group receiving Goldstein et al. did not show any meaningful effects on
10 mcg twice a day, compared to 1.6 kg in the group receiv body weight in people with type 2 diabetes after 24 weeks
ing 5 mcg twice a day over a 30 week period [43]. Kendall with sitagliptin; a loss of 0.1 kg compared to a 1.1 kg
et al. showed that exenatide use in people with type 2 dia decrease in the placebo group. In a separate trial, sitagliptin
betes on both metformin and sulfonylureas resulted in a again showed no changes in weight [54]. Ahren et al. com
1.6 kg weight loss in both dosage strengths compared to a pared the addition of albiglutide, sitagliptin, glimepiride
0.9 kg weight loss in the placebo group after 30 weeks [44]. and metformin in people with type 2 diabetes already on
Heine et al. compared exenatide to insulin glargine in peo metformin for 104‐weeks in the HARMONY‐3 Trial. This
ple with type 2 diabetes, and found that in the exenatide showed that sitagliptin reduced weight by 0.86 kg, com
arm there was a 2.3 kg weight decrease compared to a pared to a loss of 1.21 kg in the albiglutide group and a gain
1.8 kg weight increase in the arm treated with insulin over of 1.17 kg in the glimepiride group [55].
26 weeks of treatment [45]. Zinman et al. studied liraglu
tide in a 26‐week double‐blind, placebo‐controlled trial in SGLT‐2 inhibitors
people with type 2 diabetes already on metformin and Sodium‐glucose cotransporters (SGLT) are a family of pro
rosiglitazone, and reported that the patients receiving lira teins that actively transport glucose across cell membranes.
glutide lost 1.0 and 2.0 kg in the 1.2 mg and 1.8 mg daily SGLT‐2 is primarily expressed in the proximal tubule of
regimens, respectively [46]. When Garber et al. compared the kidney, but is also present at low levels in the liver and
liraglutide to glimepiride in people with type 2 diabetes not brain. The role of this protein is to reabsorb glucose from
on other medications, the liraglutide group lost 1.9 kg the urine, thus inhibition lowers the renal threshold for
(1.2 mg daily) and 2.1 kg (1.8 mg daily), whereas the glime glucose reabsorption from the collecting system resulting
piride group gained 1.2 kg [47]. Kim et al. showed that the in glycosuria [56]. Bolinder et al. reported that dapagliflo
long‐acting formulation of exenatide (2 mg once a week) zin reduced body weight by 4.54 kg over 102 weeks in peo
caused a 3.8 kg weight loss in people with type 2 diabetes ple with type 2 diabetes uncontrolled on metformin [57].
uncontrolled on metformin compared to placebo at Lavelle‐Gonzalez compared canagliflozin to sitagliptin in
15 weeks [48]. Since all prior GLP‐1 receptor agonists had people with type 2 diabetes and found that canagliflozin
resulted in a 3.3 (100 mg) and 3.7 (300 mg) weight reduc obese patients but also reduced weight by 2.1 kg, though
tion, whereas sitagliptin caused only a 1.2 kg weight reduc the lifestyle‐intervention group lost more weight (5.6 kg
tion after 52 weeks [58]. Leiter et al. found that canagliflozin after 4 years) [66]. Seifarth et al. found that metformin
in addition to metformin in people with type 2 diabetes therapy in obese patients caused an average of 5.8 kg
resulted in a 3.6 kg weight reduction in both the 100 mg weight loss compared to a 0.8 kg weight gain in the pla
and 300 mg groups compared to a 0.8 kg weight increase in cebo group, which resulted in 47.4% of the 154 met
the glimepiride group at 104 weeks [59]. Sakai et al. formin‐treated patients achieving a loss of 5% of body
reported that people with type 2 diabetes started on luse weight at the end of 6 months [67]. Wu et al. used met
ogliflozin lost an average of 2.56 kg after 52 weeks, with the formin to prevent weight gain associated with antipsy
largest weight reductions in the group with a BMI of 30 or chotic medications and reported a 4.7 kg weight reduction
above [60]. Kawata et al. demonstrated that ipragliflozin in the metformin group compared to a 3.1 weight gain in
caused a 2.6 kg weight reduction in patients with type 2 the placebo group [68].
diabetes after 24 weeks [61]. Zinman et al. showed that
people with type 2 diabetes started on empagliflozin lost GLP‐1 Receptor agonists
2.0 kg (10 mg) and 3.0 kg (25 mg) compared to placebo Astrup et al. compared orlistat to liraglutide in obese patients
after 3.1 years in the EMPA‐REG outcomes study [62]. without diabetes and found that liraglutide caused up to a
Cherney et al. stated that in patients with reduced glomer 7.2 kg weight reduction compared to 4.1 kg with orlistat
ular filtration rates (< 30 ml/min), the degree of weight loss (lipase inhibitor) and 2.8 kg with placebo after 20 weeks. A
is attenuated (1.4 kg) compared with normal GFR (> 90 ml/ subsequent analysis compared the number of subjects who
min) (2.1 kg) but was still greater than placebo [63]. The lost more than 5% of their body weight; 76% of people on
proposed mechanism by which weight is lost is through liraglutide lost more than 5% body weight, compared to 44%
direct osmotic water loss but more importantly loss of of those treated with orlistat and 30% of those receiving pla
body fat. This is thought to be accomplished by stimulation cebo [69]. During a 2‐year extension the investigators
of lipolysis and lipid oxidation due to the decreased levels reported that liraglutide‐treated patients sustained a 7.8 kg
of plasma glucose and insulin [64]. weight loss and the incidence of metabolic syndrome and
prediabetes decreased by 59% and 52%, respectively [70].
Use of diabetic medications for weight Wadden et al. reported similar results: in obese patients
loss in patients without diabetes without diabetes treated with liraglutide a 6.0 kg weight
loss was achieved. This compared favorably to a 0.1 kg
Obesity is not only associated with a risk for the develop weight loss in the placebo group after 56 weeks of treat
ment of type 2 diabetes but is also a risk factor for stroke, ment [71]. O’Neil et al. compared semaglutide to liraglu
heart disease, sleep apnea, osteoarthritis, and hyperlipi tide and found semaglutide caused a 17.4 kg weight
demia [65]. With the prevalence of obesity continuing to reduction, whereas liraglutide caused an 8.5 kg weight
increase [3], there has been an increased emphasis placed reduction [72].
on medical therapies to help induce weight loss. This has
led to the use of medications for type 2 diabetes for indi SGLT‐2 inhibitors
viduals without diabetes to help induce weight loss and Bays et al. reported that canagliflozin reduced body weight
decrease the incidence of comorbid conditions. For exam by 2.4 kg in obese patients without diabetes after
ple, a 5% reduction in body weight in people with a BMI of 12 weeks [73]. Ramirez‐Rodriguez et al. reported a 3.0 kg
34 decreased the incidence of type 2 diabetes by 58% [66]. weight reduction in obese patients with prediabetes after
12 weeks compared to placebo [74]. Hollander et al. inves
Metformin tigated the combination of canagliflozin with phentermine
In the DPP study, Knowler et al. initially showed that met in obese patients without diabetes and found up to a 7.3 kg
formin not only reduced the incidence of type 2 diabetes in reduction in body weight [75].
A Practical Guide to Choosing medication reas $578.54. However, the authors also concluded that
for type 2 diabetes with weight loss the overall cost‐effectiveness (cost per major adverse
in mind cardiovascular events avoided) was higher in oral sema
glutide compared to sitagliptin and liraglutide but not
It has been shown that a 5% weight reduction by diet and
empagliflozin [80].
exercise can decrease the risk of transitioning from pre‐
While medication costs will be variable over time, the
diabetes to diabetes by ~50% [25]. While the same degree
evidence to date suggests that GLP‐1 receptor agonists and
of weight reduction in people with type 2 diabetes results in
SGLT‐2 inhibitors provide clinically relevant weight loss
small changes in A1c [76], this degree of weight reduction
and provide positive overall health benefits compared to
reduces mortality by 25% [77]. Thus, lifestyle changes
sulfonylureas. DPP‐4 inhibitors have not shown the same
remain an important component in the treatment of type 2
beneficial health effects, have lower weight‐loss potential,
diabetes. The development of multiple, novel therapeutic
and are expensive compared to sulfonylureas. Alpha‐glu
classes for the treatment of type 2 diabetes over the past two
coside inhibitors and amylin mimetics provide weight loss
decades has given clinicians more options to choose from
but are not considered second‐line therapy due to clear
when deciding which pharmacological approach is optimal
benefits in decreasing cardiovascular endpoints. TZDs are
for a given patient. Metformin continues to be considered
associated with weight gain [78] therefore limiting their
the universal first‐line medication for type 2 diabetes due to
use as first‐line agents.
its low cost, effectiveness, and potential for weight loss [78].
When considering second‐line therapy or add‐on therapy,
the ADA recommends a patient‐centric approach, taking References
into account cardiovascular risk, avoidance of side effects, 1. Hu FB, Manson JE, Stampfer MJ, Colditz G, Liu S, Solomon
and cost [78]. The cost of each medication can vary based on CG et al. Diet, lifestyle, and the risk of type 2 diabetes mellitus
insurance coverage and payer source. in women. N Engl J Med. 2001;345(11):790–797.
In 2014 The Institute for Clinical and Economic 2. NHLBI Obesity Education Initiative Expert Panel on the
Review (ICER) reported that the combination of met Identification, Evaluation, and Treatment of Obesity in Adults.
formin and a GLP‐1 receptor agonist is superior to met Clinical Guidelines on the Identification, Evaluation, and
formin and a sulfonylurea, and that the combination of Treatment of Overweight and Obesity in Adults. Bethesda,
metformin, a GLP‐1 receptor agonist, and a sulfonylurea MI: NHLBI, 1998.
3. Hales CM, Fryar CD, Carroll MD, Freedman DS, Ogden CL.
is superior to the combination of metformin, a GLP‐1
Trends in obesity and severe obesity prevalence in us youth
receptor agonist, and NPH insulin. These recommenda
and adults by sex and age, 2007–2008 to 2015–2016. JAMA.
tions were based on the clinical benefits of weight reduc
2018;319(16):1723–1725.
tion, risk of hypoglycemia, and cost to patients. They 4. Mamtani M, Kulkarni H, Dyer TD, Almasy L, Mahaney MC,
did not find adequate evidence to recommend Duggirala R et al. Waist circumference independently associates
DPP‐4 inhibitors over sulfonylureas due to lower clini with the risk of insulin resistance and type 2 diabetes in Mexican
cal benefits with higher cost [79]. In 2019 the ICER American families. PLoS One. 2013;8(3):e59153.
updated their recommendations comparing oral sema 5. Caspard H, Jabbour S, Hammar N, Fenici P, Sheehan JJ,
glutide to other second line therapies for type 2 diabetes Kosiborod M. Recent trends in the prevalence of type 2 diabe
and reported that there was inconclusive evidence to tes and the association with abdominal obesity lead to grow
recommend oral semaglutide or empagliflozin over each ing health disparities in the USA. An analysis of the NHANES
surveys from 1999 to 2014. Diabetes Obes Metab. 2018;20(3):
other but there was high certainty that semaglutide has
667–671.
at least a small health benefit over sitagliptin. The net
6. Nguyen NT, Nguyen XM, Lane J, Wang P. Relationship
price per year was also compared: oral semaglutide
between obesity and diabetes in a US adult population: find
$6520.02, sitagliptin $1505.07, empagliflozin $2088.13, ings from the National Health and Nutrition Examination
liraglutide $5341.90, metformin $917.19, and sulfonylu Survey, 1999–2006. Obes Surg. 2011;21(3):351–355.
7. Wing RR, Bolin P, Brancati FL, Bray GA, Clark JM, Coday M predictors amongst 11 140 patients with type 2 diabetes in
et al. Cardiovascular effects of intensive lifestyle interven the ADVANCE trial. Diabetes Obes Metab. 2012;14(5):
tion in type 2 diabetes. N Engl J Med. 2013;369(2):145–154. 464–469.
8. American Diabetes Association. Standards of medical care 21. Luis Bautista J, Bugos C, Dirnberger G, Atherton T. Efficacy
in diabetes – 2020. Diabetes Care. 2020;43(Suppl 1). and safety profile of glimepiride in Mexican American
9. Sumithran P, Prendergast LA, Delbridge E, Purcell K, Patients with type 2 diabetes mellitus: a randomized, placebo‐
Shulkes A, Kriketos A et al. Long‐term persistence of hor controlled study. Clin Ther. 2003;25(1):194–209.
monal adaptations to weight loss. N Engl J Med. 22. Simonson DC, Kourides IA, Feinglos M, Shamoon H,
2011;365(17):1597–1604. Fischette CT. Efficacy, safety, and dose‐response characteris
10. United Kingdom Prospective Diabetes Study (UKPDS). 13: tics of glipizide gastrointestinal therapeutic system on glyce
Relative efficacy of randomly allocated diet, sulphonylurea, mic control and insulin secretion in NIDDM. Results of two
insulin, or metformin in patients with newly diagnosed non‐ multicenter, randomized, placebo‐controlled clinical trials.
insulin dependent diabetes followed for three years. BMJ. The Glipizide Gastrointestinal Therapeutic System Study
1995;310(6972):83–88. Group. Diabetes Care. 1997;20(4):597–606.
11. Fonseca V, McDuffie R, Calles J, Cohen RM, Feeney P, 23. Effect of intensive blood‐glucose control with metformin on
Feinglos M et al. Determinants of weight gain in the action complications in overweight patients with type 2 diabetes
to control cardiovascular risk in diabetes trial. Diabetes Care. (UKPDS 34). UK Prospective Diabetes Study (UKPDS)
2013;36(8):2162–2168. Group. Lancet. 1998;352(9131):854–865.
12. Carlson MG, Campbell PJ. Intensive insulin therapy and 24. DeFronzo RA, Goodman AM. Efficacy of metformin in
weight gain in IDDM. Diabetes. 1993;42(12):1700–1707. patients with non‐insulin‐dependent diabetes mellitus. The
13. Umpleby AM, Boroujerdi MA, Brown PM, Carson ER, Multicenter Metformin Study Group. N Engl J Med. 1995;
Sonksen PH. The effect of metabolic control on leucine 333(9):541–549.
metabolism in type 1 (insulin‐dependent) diabetic patients. 25. Knowler WC, Fowler SE, Hamman RF, Christophi CA,
Diabetologia. 1986;29(3):131–141. Hoffman HJ, Brenneman AT et al. 10‐year follow‐up of dia
14. Schwartz MW, Niswender KD. Adiposity signaling and bio betes incidence and weight loss in the Diabetes Preven
logical defense against weight gain: absence of protection or tion Program Outcomes Study. Lancet. 2009;374(9702):
central hormone resistance? J Clin Endocrinol Metab. 2004; 1677–1686.
89(12):5889–5897. 26. Goldstein BJ, Pans M, Rubin CJ. Multicenter, randomized,
15. Weight gain associated with intensive therapy in the diabetes double‐masked, parallel‐group assessment of simultaneous
control and complications trial. The DCCT Research Group. glipizide/metformin as second‐line pharmacologic treat
Diabetes Care. 1988;11(7):567–573. ment for patients with type 2 diabetes mellitus that is inade
16. Intensive blood‐glucose control with sulphonylureas or quately controlled by a sulfonylurea. Clin Ther. 2003;25(3):
insulin compared with conventional treatment and risk of 890–903.
complications in patients with type 2 diabetes (UKPDS 33). 27. Yki‐Jarvinen H. Thiazolidinediones. N Engl J Med. 2004;
UK Prospective Diabetes Study (UKPDS) Group. Lancet. 351(11):1106–1118.
1998;352(9131):837–853. 28. Miyazaki Y, Mahankali A, Matsuda M, Mahankali S, Hardies
17. Henry RR, Gumbiner B, Ditzler T, Wallace P, Lyon R, J, Cusi K et al. Effect of pioglitazone on abdominal fat distri
Glauber HS. Intensive conventional insulin therapy for type bution and insulin sensitivity in type 2 diabetic patients.
II diabetes. Metabolic effects during a 6‐mo outpatient trial. J Clin Endocrinol Metab. 2002;87(6):2784–2791.
Diabetes Care. 1993;16(1):21–31. 29. Basu A, Jensen MD, McCann F, Mukhopadhyay D, Joyner
18. Yki‐Jarvinen H, Ryysy L, Kauppila M, Kujansuu E, Lahti J, MJ, Rizza RA. Effects of pioglitazone versus glipizide on
Marjanen T et al. Effect of obesity on the response to insulin body fat distribution, body water content, and hemodynam
therapy in noninsulin‐dependent diabetes mellitus. J Clin ics in type 2 diabetes. Diabetes Care. 2006;29(3):510–514.
Endocrinol Metab. 1997;82(12):4037–4043. 30. Wilding J. Thiazolidinediones, insulin resistance and obe
19. Ashcroft FM. Mechanisms of the glycaemic effects of sulfo sity: finding a balance. Int J Clin Pract. 2006;60(10):
nylureas. Horm Metab Res. 1996;28(9):456–463. 1272–1280.
20. van Dieren S, Czernichow S, Chalmers J, Kengne AP, 31. Bischoff H. Pharmacology of alpha‐glucosidase inhibition.
de Galan BE, Poulter N et al. Weight changes and their Eur J Clin Invest. 1994;24 Suppl 3:3–10.
32. Phillips P, Karrasch J, Scott R, Wilson D, Moses R. Acarbose 43. DeFronzo RA, Ratner RE, Han J, Kim DD, Fineman MS,
improves glycemic control in overweight type 2 diabetic Baron AD. Effects of exenatide (exendin‐4) on glycemic
patients insufficiently treated with metformin. Diabetes control and weight over 30 weeks in metformin‐treated
Care. 2003;26(2):269–273. patients with type 2 diabetes. Diabetes Care.
33. Feinbock C, Luger A, Klingler A, Egger T, Bielesz GK, 2005;28(5):1092–1100.
Winkler F et al. Prospective multicentre trial comparing the 44. Kendall DM, Riddle MC, Rosenstock J, Zhuang D, Kim DD,
efficacy of, and compliance with, glimepiride or acarbose Fineman MS et al. Effects of exenatide (exendin‐4) on glyce
treatment in patients with type 2 diabetes not controlled mic control over 30 weeks in patients with type 2 diabetes
with diet alone. Diabetes Nutr Metab. 2003;16(4):214–221. treated with metformin and a sulfonylurea. Diabetes Care.
34. Lindstrom J, Tuomilehto J, Spengler M. Acarbose treatment 2005;28(5):1083–1091.
does not change the habitual diet of patients with type 2 dia 45. Heine RJ, Van Gaal LF, Johns D, Mihm MJ, Widel MH,
betes mellitus. The Finnish Acargbos Study Group. Diabet Brodows RG. Exenatide versus insulin glargine in patients
Med. 2000;17(1):20–25. with suboptimally controlled type 2 diabetes: a randomized
35. Wolever TM, Chiasson JL, Josse RG, Hunt JA, Palmason trial. Ann Intern Med. 2005;143(8):559–569.
C, Rodger NW et al. Small weight loss on long‐term acar 46. Zinman B, Gerich J, Buse JB, Lewin A, Schwartz S, Raskin P
bose therapy with no change in dietary pattern or nutri et al. Efficacy and safety of the human glucagon‐like pep
ent intake of individuals with non‐insulin‐dependent tide‐1 analog liraglutide in combination with metformin
diabetes. Int J Obes Relat Metab Disord. 1997;21(9): and thiazolidinedione in patients with type 2 diabetes
756–763. (LEAD‐4 Met+TZD). Diabetes Care.
36. Weyer C, Maggs DG, Young AA, Kolterman OG. Amylin 2009;32(7):1224–1230.
replacement with pramlintide as an adjunct to insulin ther 47. Garber A, Henry R, Ratner R, Garcia‐Hernandez PA,
apy in type 1 and type 2 diabetes mellitus: a physiological Rodriguez‐Pattzi H, Olvera‐Alvarez I et al. Liraglutide ver
approach toward improved metabolic control. Curr Pharm sus glimepiride monotherapy for type 2 diabetes
Des. 2001;7(14):1353–1373. (LEAD‐3 Mono): a randomised, 52‐week, phase III, double‐
37. Thompson RG, Pearson L, Schoenfeld SL, Kolterman OG. blind, parallel‐treatment trial. Lancet.
Pramlintide, a synthetic analog of human amylin, improves 2009;373(9662):473–481.
the metabolic profile of patients with type 2 diabetes using 48. Kim D, MacConell L, Zhuang D, Kothare PA, Trautmann M,
insulin. The Pramlintide in Type 2 Diabetes Group. Diabetes Fineman M et al. Effects of once‐weekly dosing of a long‐
Care. 1998;21(6):987–993. acting release formulation of exenatide on glucose control
38. Riddle M, Frias J, Zhang B, Maier H, Brown C, Lutz K et al. and body weight in subjects with type 2 diabetes. Diabetes
Pramlintide improved glycemic control and reduced weight Care. 2007;30(6):1487–1493.
in patients with type 2 diabetes using basal insulin. Diabetes 49. Davies M, Pieber TR, Hartoft‐Nielsen ML, Hansen OKH,
Care. 2007;30(11):2794–2799. Jabbour S, Rosenstock J. Effect of oral semaglutide com
39. Hollander PA, Levy P, Fineman MS, Maggs DG, Shen LZ, pared with placebo and subcutaneous semaglutide on glyce
Strobel SA et al. Pramlintide as an adjunct to insulin therapy mic control in patients with type 2 diabetes: a randomized
improves long‐term glycemic and weight control in patients clinical trial. JAMA. 2017;318(15):1460–1470.
with type 2 diabetes: a 1‐year randomized controlled trial. 50. Henry RR, Rosenstock J, Denham DS, Prabhakar P, Kjems L,
Diabetes Care. 2003;26(3):784–790. Baron MA. Clinical impact of ITCA 650, a novel drug‐
40. Hollander P, Maggs DG, Ruggles JA, Fineman M, Shen L, device GLP‐1 receptor agonist, in uncontrolled type 2 diabe
Kolterman OG et al. Effect of pramlintide on weight in over tes and very high baseline HbA1c: the FREEDOM‐1 HBL
weight and obese insulin‐treated type 2 diabetes patients. (High Baseline) Study. Diabetes Care. 2018;41(3):613–619.
Obes Res. 2004;12(4):661–668. 51. Drucker DJ. Therapeutic potential of dipeptidyl peptidase
41. Deacon CF. Therapeutic strategies based on glucagon‐like IV inhibitors for the treatment of type 2 diabetes. Expert
peptide 1. Diabetes. 2004;53(9):2181–2189. Opin Investig Drugs. 2003;12(1):87–100.
42. Buse JB, Henry RR, Han J, Kim DD, Fineman MS, Baron 52. Ristic S, Byiers S, Foley J, Holmes D. Improved glycaemic
AD. Effects of exenatide (exendin‐4) on glycemic control control with dipeptidyl peptidase‐4 inhibition in patients
over 30 weeks in sulfonylurea‐treated patients with type 2 with type 2 diabetes: vildagliptin (LAF237) dose response.
diabetes. Diabetes Care. 2004;27(11):2628–2635. Diabetes Obes Metab. 2005;7(6):692–698.
53. Bolli G, Dotta F, Rochotte E, Cohen SE. Efficacy and tolera mortality in type 2 diabetes. N Engl J Med. 2015;373(22):
bility of vildagliptin vs. pioglitazone when added to metformin: 2117–2128.
a 24‐week, randomized, double‐blind study. Diabetes Obes 63. Cherney DZI, Cooper ME, Tikkanen I, Pfarr E, Johansen
Metab. 2008;10(1):82–90. OE, Woerle HJ et al. Pooled analysis of Phase III trials indi
54. Goldstein BJ, Feinglos MN, Lunceford JK, Johnson J, cate contrasting influences of renal function on blood pres
Williams‐Herman DE. Effect of initial combination therapy sure, body weight, and HbA1c reductions with empagliflozin.
with sitagliptin, a dipeptidyl peptidase‐4 inhibitor, and met Kidney Int. 2018;93(1):231–244.
formin on glycemic control in patients with type 2 diabetes. 64. Ferrannini E, Baldi S, Frascerra S, Astiarraga B, Heise T,
Diabetes Care. 2007;30(8):1979–1987. Bizzotto R et al. Shift to fatty substrate utilization in response
55. Ahren B, Johnson SL, Stewart M, Cirkel DT, Yang F, Perry C to sodium‐glucose cotransporter 2 inhibition in subjects
et al. HARMONY 3: 104‐week randomized, double‐blind, without diabetes and patients with type 2 diabetes. Diabetes.
placebo‐ and active‐controlled trial assessing the efficacy 2016;65(5):1190–1195.
and safety of albiglutide compared with placebo, sitagliptin, 65. Li Z, Maglione M, Tu W, Mojica W, Arterburn D, Shugarman
and glimepiride in patients with type 2 diabetes taking met LR et al. Meta‐analysis: pharmacologic treatment of obesity.
formin. Diabetes Care. 2014;37(8):2141–2148. Ann Intern Med. 2005;142(7):532–546.
56. Abdul‐Ghani MA, Norton L, DeFronzo RA. Role of sodium‐ 66. Knowler WC, Barrett‐Connor E, Fowler SE, Hamman RF,
glucose cotransporter 2 (SGLT 2) inhibitors in the treatment Lachin JM, Walker EA et al. Reduction in the incidence of
of type 2 diabetes. Endocr Rev. 2011;32(4):515–531. type 2 diabetes with lifestyle intervention or metformin. N
57. Bolinder J, Ljunggren O, Johansson L, Wilding J, Langkilde Engl J Med. 2002;346(6):393–403.
AM, Sjostrom CD et al. Dapagliflozin maintains glycaemic 67. Seifarth C, Schehler B, Schneider HJ. Effectiveness of
control while reducing weight and body fat mass over 2 years metformin on weight loss in non‐diabetic individuals
in patients with type 2 diabetes mellitus inadequately con with obesity. Exp Clin Endocrinol Diabetes. 2013;121(1):
trolled on metformin. Diabetes Obes Metab. 2014;16(2): 27–31.
159–169. 68. Wu RR, Zhao JP, Jin H, Shao P, Fang MS, Guo XF et al.
58. Lavalle‐Gonzalez FJ, Januszewicz A, Davidson J, Tong C, Lifestyle intervention and metformin for treatment of antip
Qiu R, Canovatchel W et al. Efficacy and safety of canagliflo sychotic‐induced weight gain: a randomized controlled trial.
zin compared with placebo and sitagliptin in patients with JAMA. 2008;299(2):185–193.
type 2 diabetes on background metformin monotherapy: a 69. Astrup A, Rossner S, Van Gaal L, Rissanen A, Niskanen L,
randomised trial. Diabetologia. 2013;56(12):2582–2592. Al Hakim M et al. Effects of liraglutide in the treatment of
59. Leiter LA, Yoon KH, Arias P, Langslet G, Xie J, Balis DA et al. obesity: a randomised, double‐blind, placebo‐controlled study.
Canagliflozin provides durable glycemic improvements and Lancet. 2009;374(9701):1606–1616.
body weight reduction over 104 weeks versus glimepiride in 70. Astrup A, Carraro R, Finer N, Harper A, Kunesova M, Lean
patients with type 2 diabetes on metformin: a randomized, dou ME et al. Safety, tolerability and sustained weight loss over 2
ble‐blind, phase 3 study. Diabetes Care. 2015;38(3):355–364. years with the once‐daily human GLP‐1 analog, liraglutide.
60. Sakai S, Kaku K, Seino Y, Inagaki N, Haneda M, Sasaki T Int J Obes (Lond). 2012;36(6):843–854.
et al. Efficacy and safety of the SGLT2 inhibitor luseogliflo 71. Wadden TA, Hollander P, Klein S, Niswender K, Woo V,
zin in Japanese patients with type 2 diabetes mellitus strati Hale PM et al. Weight maintenance and additional weight
fied according to baseline body mass index: pooled analysis loss with liraglutide after low‐calorie‐diet‐induced weight
of data from 52‐week phase iii trials. Clin Ther. 2016;38(4): loss: the SCALE Maintenance randomized study. Int J Obes
843–862.e9. (Lond). 2013;37(11):1443–1451.
61. Kawata T, Iizuka T, Iemitsu K, Takihata M, Takai M, 72. O’Neil PM, Birkenfeld AL, McGowan B, Mosenzon O,
Nakajima S et al. Ipragliflozin improves glycemic control Pedersen SD, Wharton S et al. Efficacy and safety of sema
and decreases body fat in patients with type 2 diabetes mel glutide compared with liraglutide and placebo for weight
litus. J Clin Med Res. 2017;9(7):586–595. loss in patients with obesity: a randomised, double‐blind,
62. Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, placebo and active controlled, dose‐ranging, phase 2 trial.
Hantel S et al. Empagliflozin, cardiovascular outcomes, and Lancet. 2018;392(10148):637–649.
73. Bays HE, Weinstein R, Law G, Canovatchel W. Canagliflozin: 77. Gregg EW, Gerzoff RB, Thompson TJ, Williamson DF.
effects in overweight and obese subjects without diabe Trying to lose weight, losing weight, and 9‐year mortality in
tes mellitus. Obesity (Silver Spring). 2014;22(4):1042–1049. overweight U.S. adults with diabetes. Diabetes Care. 2004;
74. Ramirez‐Rodriguez AM, Gonzalez‐Ortiz M, Martinez‐ 27(3):657–662.
Abundis E. Effect of dapagliflozin on insulin secretion and 78. American Diabetes Association. 9. Pharmacologic
insulin sensitivity in patients with prediabetes. Exp Clin Approaches to Glycemic Treatment: Standards of Medical
Endocrinol Diabetes. 2018. Care in Diabetes – 2019. 2019.
75. Hollander P, Bays HE, Rosenstock J, Frustaci ME, Fung A, 79. Alzaid AA, Dinneen SF, Turk DJ, Caumo A, Cobelli C, Rizza
Vercruysse F et al. Coadministration of canagliflozin and RA. Assessment of insulin action and glucose effectiveness
phentermine for weight management in overweight and in diabetic and nondiabetic humans. J Clin Invest.
obese individuals without diabetes: a randomized clinical 1994;94(6):2341–2348.
trial. Diabetes Care. 2017;40(5):632–639. 80. O’Brien TD, Rizza RA, Carney JA, Butler PC. Islet amyloido
76. Pastors JG, Franz MJ, Warshaw H, Daly A, Arnold MS. How sis in a patient with chronic massive insulin resistance due to
effective is medical nutrition therapy in diabetes care? J Am antiinsulin receptor antibodies. J Clin Endocrinol Metab.
Diet Assoc. 2003;103(7):827–831. 1994;79(1):290–292.
16 Are Statins the Optimal Therapy
for Cardiovascular Risk in Patients
with Diabetes? What Newer Agents Are
There for the Treatment for Dyslipidemia
in Diabetes? Are Triglycerides
an Important Risk Factor for Diabetes?
Recie Davern and Timothy O’Brien
Department of Medicine and Endocrinology/Diabetes Mellitus, University College Hospital Galway and National
University of Ireland, Galway, Ireland
33 kg/m2 [20–25]. Lipid profile shows a total cholesterol of

4.4 mmol/l (< 4.6 mmol/l), LDL 1.4 mmol/l (< 2.6 mmol/l),
●● Treating dyslipidemia is central to the prevention of
HDL of 0.78 mmol/l (> 1.0 mmol/l) and fasting triglycer
cardiovascular disease in diabetes. ides of 4.4 mmol/l (< 1.7 mmol/l). His weekly alcohol con
●● There is strong epidemiological evidence to date that sumption is 16–20 units per week.
hypertriglyceridemia is an independent risk factor for The above clinical scenario is very commonly encoun
ischemic heart disease. tered in outpatient diabetes care. Although evidence‐based
●● Combination therapy may be necessary to adequately
address dyslipidemia in diabetes.
targets such as HbA1C, blood pressure and LDL choles
terol are all within desired limits, uncertainty arises over
whether treatment should be tailored to account for ele
vated fasting plasma triglycerides and low HDL choles
Introduction
terol, which are the two most common lipid derangements
A 54‐year‐old man with a four‐year history of Type 2 in Type 2 diabetes mellitus. In this chapter we will discuss
Diabetes Mellitus is reviewed in the outpatient department. the pathophysiology of dyslipidemia in Type 2 diabetes
Past medical history is otherwise significant for hyperten mellitus, review the arguments for use of statins in both
sion and dyslipidemia. His medications are metformin primary and secondary prevention of cardiovascular dis
500 mg tds, ezetimibe/simvastatin 10/20 mg od, and tel ease (CVD) in diabetes mellitus, discuss the evidence base
misartan 80 mg od. Glycosylated hemoglobin (HbA1C) is for elevated triglycerides as an independent risk factor for
5.4% (4.0–6.0). Blood pressure is well‐controlled at CVD in diabetes mellitus and discuss new emerging thera
110/78 mmHg. Body mass index (BMI) is elevated at pies for the treatment of dyslipidemia.

198
Are Statins the Optimal Therapy for Cardiovascular Risk in Patients with Diabetes? 199
Dyslipidemia in type 2 diabetes mellitus Insulin resistance also promotes the conversion LDL to
smaller lipoproteins termed small dense LDL [8]. These
Dyslipidemia is one of the major risk factors for cardiovas
small dense particles have been shown to have atherogenic
cular disease (CVD) in type 2 diabetes mellitus. Despite
properties [9]. Small dense LDL particles have reduced
mounting data from clinical trials, the management of dys
affinity for the LDL B/E receptor and are preferentially
lipidemia other than lowering low‐density lipoprotein (LDL)
taken up by macrophages, through the scavenger receptor,
cholesterol continues to be debated. While the term “dyslipi
leading to the formation of foam cells. Small dense LDL
demia” describes a vast array of lipoprotein abnormalities,
particles have higher affinity for intimal proteoglycans
diabetic dyslipidemia is a type of secondary dyslipidemia.
than large LDL particles which may favor the penetration
Both quantitative and qualitative lipoprotein abnormali
of LDL particles into the arterial wall [10]. This promotes
ties are observed in type 2 diabetic patients [1]. The quanti
atherosclerosis and contributes to their increased cardio
tative abnormalities are hypertriglyceridemia and low levels
vascular risk.
of cardioprotective high‐density lipoprotein (HDL) choles
To fully understand the role of the lipid abnormalities
terol levels. The qualitative abnormalities include the pres
seen in type 2 diabetes mellitus we must review the meta
ence of large very low‐density lipoproteins (VLDL) with
bolic syndrome. The metabolic syndrome refers to a group
relatively high levels of triglycerides and small dense LDL.
of metabolic abnormalities that are frequently seen in type
Insulin plays a central role in the regulation of lipid
2 diabetes mellitus that increase a person’s risk of CVD,
metabolism. In adipose tissue, insulin inhibits hormone
both individually and collectively [11]. There are different
sensitive lipase, therefore, it has anti‐lipolytic action. It pro
sets of diagnostic criteria for metabolic syndrome that have
motes storage of triglyceride (TG) in the adipocytes and
been published by the WHO, the European Group for the
reduces secretion in the circulation of free fatty acids (FFA)
Study of Insulin Resistance (EGIR), the National
from adipose tissue.
Cholesterol Education Program (NCEP) Adult Treatment
The hypertriglyceridemia seen in type 2 diabetic
Panel III (ATP III), and the International Diabetes
patients is a product of insulin resistance. Insulin resistance
Foundation (IDF). The NCEP ATP III is the most com
leads to excessive amounts of FFA, many of which build up
monly used and we have summarized those criteria in
in the liver as TG [2]. The liver has three different handling
Box 16.1.
mechanisms for the excess, it either: stores it, burns it
through β‐oxidation in the mitochondria or exports it by
Treatment of dyslipidemia in type 2
synthesizing VLDL particles, which are TG rich.
diabetes mellitus
The decrease in HDL cholesterol is caused by increased
catabolism due to insulin resistance. This is due to an increased There is a large body of evidence supporting treatment of
pool of TG rich VLDL. This increase leads to the transfer of dyslipidemia in type 2 diabetes mellitus for primary and
TG from TG rich lipoproteins to HDL leading to the forma secondary prevention of CVD [12, 13].
tion of TG‐rich HDL particles [3]. Cholesteryl ester transfer
protein (CETP) is responsible for these transfers to HDL and Box 16.1 NCEP guidelines for diagnosis
of the metabolic syndrome
CETP inhibition has been studied as a mechanism to increase
HDL and reduce CVD risk. Unfortunately, many clinical tri Presence of three of the following five risk factors:
als [4, 5, 6] have not shown that CETP inhibition reduces OO Abdominal obesity (waist circumference > 40 inches in
men, > 35 inches in women)
CVD risk. Though a recent trial [7] of anacetrapib showed a
OO Plasma triglycerides ≥ 150 mg/dl (1.7 mmol/l)
modest reduction in CVD risk after a long follow‐up period. OO Plasma HDL cholesterol ≤ 40 mg/dl in men
Hepatic lipase whose activity is increased in insulin resistance (1.02 mmol/l) and ≤ 50 mg/dl in women (1.27 mmol/l)
and type 2 diabetes leads to increased catabolism of these TG OO Blood pressure > 130/85 mmHg
rich HDL particles. OO Fasting plasma glucose > 110 mg/dl (6.1 mmol/l)
Statins are the mostly extensively studied treatment for 20 536 patients with CHD, other occlusive vascular disease
dyslipidemia. Two studies have examined statin use for or diabetes mellitus, and included 5963 subjects with diabe
the primary prevention of CVD exclusively in those with tes mellitus. Simvastatin 40 mg or placebo was given over the
diabetes mellitus. The Collaborative Atorvastatin course of 5 years. 2912 of the participants with diabetes mel
Diabetes Study (CARDS) involved 2838 subjects aged litus had no prior vascular disease (49%) (primary preven
40–75 years from centers around the UK and Ireland who tion) and in this subgroup the rate of the defined endpoint
were randomized to either 10 mg atorvastatin or placebo. for subcategories which was major vascular events (cardio
Participants had type 2 diabetes mellitus with no history vascular death, non‐fatal myocardial infarction, stroke,
of CVD, but had at least one other cardiovascular risk revascularization) was reduced significantly from 13% to 9%
factor. The study was terminated 2 years early because the in the group receiving simvastatin [16].
pre‐specified stopping rule for efficacy in reducing car The Anglo‐Scandinavian Cardiac Outcomes Trial
diovascular events had been achieved. Subjects allocated (ASCOT) was primarily aimed at investigating treatment
to atorvastatin had a 37% reduction in the primary end of hypertension in 19 342 hypertensive patients, but half of
point of acute coronary heart disease (CHD) events the study group (n = 10 305), including 2532 with type two
(myocardial infarction, unstable angina, acute CHD diabetes mellitus, were allocated to a lipid sub‐study with
death, resuscitated cardiac arrest), coronary revasculari atorvastatin 10 mg or placebo. Overall results for the lipid‐
zation or stroke. All‐cause mortality was numerically lowering trial showed a significant reduction in the pri
lower in the atorvastatin group but was not statistically mary endpoint of non‐fatal myocardial infarction and fatal
significantly reduced [14]. CHD with atorvastatin. No statistically significant reduc
The Atorvastatin Study for Prevention of CHD tion in this endpoint was found when looking at the sub
Endpoints in Non‐insulin‐dependent diabetes mellitus group with diabetes mellitus. However, a subsequent
(ASPEN) study included 2410 subjects with type two dia pre‐specified analysis of an expanded composite of total
betes mellitus who were also assigned to receive 10 mg cardiovascular outcomes (major cardiovascular events plus
atorvastatin or placebo. This international study was origi procedures) demonstrated a significant reduction for the
nally intended to look at secondary prevention with a fol subgroup with diabetes mellitus [17].
low‐up of 4 years; however, due to changing guidelines The Management of Elevated Cholesterol in the Primary
impairing recruitment, the protocol was changed and a Prevention Group of Adult Japanese study (MEGA)
primary prevention cohort were also recruited. For the pri recruited 7832 subjects with dyslipidemia and assigned
mary prevention cohort a significant 30% reduction in them to diet modification only or diet plus open‐label low‐
LDL levels was observed, but no beneficial effect was found dose pravastatin (10–20 mg). 1632 subjects (21%) were
on the primary outcome which was a composite of cardio reported by their physicians to have diabetes mellitus at
vascular death, non‐fatal myocardial infarction, non‐fatal baseline. Overall, the rate of CHD events (fatal or non‐fatal
stroke, recanalization, coronary artery bypass surgery, myocardial infarction, angina, revascularization, sudden
resuscitated cardiac arrest, and hospitalization for unstable cardiac death) was significantly reduced in the pravastatin‐
angina. Similar results were found in the small secondary treated group compared with diet only. Subgroup analysis
prevention cohort. There are multiple limitations of this did not show significant interaction in any subgroup,
study including the change in the study design, the multi including the diagnosis of diabetes mellitus or not. A post‐
ple endpoints, the low risk of events in the primary preven hoc analysis looked at the subgroup with diabetes mellitus
tion cohort and the extensive use of non‐study lipid‐lowering and categorized them according to reductions in LDL cho
therapy in both groups [15]. lesterol and increases in HDL cholesterol. The greatest
Several studies have investigated statin use for primary reduction in risk of CVD was found where LDL cholesterol
prevention with substantial subgroups of subjects with dia decreased more than 15% and HDL cholesterol increased
betes mellitus. The Heart Protection Study (HPS) recruited by more than 5% [18].
The Antihypertensive and Lipid‐Lowering treatment to and average cholesterol levels. The incidence of the pri
prevent Heart Attack Trial (ALLHAT) [19] looked at treat mary endpoint of a fatal coronary event or non‐fatal myo
ment of hypertension in subjects aged 55 years or older. cardial infarction was significantly reduced by 24% in the
Within this study there was a Lipid‐Lowering Trial compo treatment group overall, demonstrating benefit of statin
nent (ALLHAT‐LLT) which randomized 10 355 subjects, use in secondary prevention even in the absence of raised
(of which 3638 (35%) had type two diabetes mellitus), to LDL levels. In the 586 subjects with diabetes mellitus
open label pravastatin 40 mg treatment or usual care. No (14%), the reduction in the primary endpoint was not sig
significant difference was found in all‐cause mortality (pri nificant but there was a 25% significant reduction in an
mary outcome) or CHD outcomes (secondary outcome of expanded coronary endpoint.
fatal CHD, non‐fatal myocardial infarction) in the whole There is also evidence comparing high‐dose versus low‐
study group or in the subgroup with diabetes mellitus. It is dose statin in diabetes mellitus. The Treating to New
felt that one of the reasons for the negative results seen in Targets (TNT) study [24] compared atorvastatin 80 mg
this trial may be the relatively high rates of statin use in the versus atorvastatin 10 mg in 10 001 people with stable
usual care group, reaching nearly 25%, as intensive lipid CHD. 1501 (15%) of the subjects had diabetes mellitus and
lowering became a routine part of clinical care following subgroup analysis showed high‐dose atorvastatin was ben
the publication of other statin studies. eficial in significantly reducing the risk of major cardiovas
When it comes to secondary prevention there are also cular events (CHD death, myocardial infarction, stroke,
many landmark studies. The Scandinavian Simvastatin resuscitation after cardiac arrest) by 25%. Safety analysis
Survival Study (4S) recruited 4444 patients with a prior his demonstrated a significantly greater number of adverse
tory of CHD and hypercholesterolemia who were rand events in the intensively treated group, particularly an
omized to treatment with 20–40 mg simvastatin or placebo. increase in abnormal liver function tests, with no signifi
Over a median duration of 5.4 years of follow‐up, those cant differences between the rates of reported myalgia and
randomized to simvastatin had a significant reduction in rhabdomyolysis in both groups.
all‐cause mortality and major CHD events, and this The Pravastatin or Atorvastatin Evaluation and
included a significant reduction in CHD events in the Infection Therapy (PROVE IT) Thrombolysis in
small subgroup of 202 with diabetes mellitus [20]. A later Myocardial Infarction (TIMI) 22 trial [25] compared
post‐hoc analysis was performed using newer diagnostic 40 mg pravastatin versus 80 mg atorvastatin in people with
criteria for type two diabetes mellitus and confirmed a sig recent acute coronary syndromes. High intensity statin
nificant reduction in CHD events in 483 subjects with treatment significantly reduced the primary extended car
baseline diabetes mellitus [21]. diovascular composite end point by 16% in the general
The Long‐term Intervention with Pravastatin in study population. Nearly one quarter of the subjects in
Ischemic Disease (LIPID) trial involved a larger cohort of PROVE IT had diabetes mellitus and, although subgroup
9014 patients with CHD, with 1077 (12%) diagnosed as analysis showed a reduction in primary end points in the
having diabetes mellitus at baseline. Overall, pravastatin high intensity group, this did not reach statistical signifi
40 mg significantly reduced the primary outcome of mor cance. The authors attributed this to the sub study being
tality from CHD, and in a pre‐specified subgroup analysis, underpowered. A higher number of people in the intensive
pravastatin reduced the occurrence of major CHD events arm had abnormal liver function tests, but the rates of dis
and major coronary events (CHD death or non‐fatal myo continuation of statin therapy due to side effects was not
cardial infarction) in the subgroup with diabetes melli significantly different between groups.
tus [22]. The effects shown on triglyceride levels in this There is also evidence of benefit in non‐statin therapies
study will be discussed later in the chapter. for dyslipidemia. Ezetimibe reduces cholesterol absorption
The Cholesterol and Recurrent Events (CARE) [23] trial by inhibiting Niemann‐Pick C1‐Like 1 (NPC1L1) protein
was a secondary prevention study comparing pravastatin in the small intestine and hepatocytes. In the IMPROVE‐
40 mg with placebo in 4159 patients with a history of CHD IT (Improved Reduction of outcomes: Vytorin Efficacy
International Trial) [26], the first trial to show an improve binds to LDL receptors expressed on the surface of hepato
ment in cardiovascular (CV) outcomes with the addition of cytes and is internalized by endocytosis. LDL receptors are
a non‐statin drug to a statin, 18 144 patients who had been usually recycled to the cell surface, but when PCSK9 binds
hospitalized for an acute coronary syndrome (ACS) within with LDL receptors, LDL receptors are delivered to lys
the preceding 10 days were randomized to simvastatin– osomes for degradation, resulting in lower expression of
ezetimibe combination therapy or simvastatin monother LDL receptors and an increase in LDL cholesterol. The
apy. Participants with diabetes mellitus was a prespecified relationship between LDL cholesterol levels and athero
subgroup in this trial. It had a median follow‐up of 6 years, sclerosis has been extensively studied. Higher LDL choles
the addition of ezetimibe to simvastatin reduced LDL‐C by terol levels are strongly correlated with atherosclerosis and
16 mg/dL and resulted in a 6.4% reduction (32.7% vs. higher cardiovascular comorbidities and mortality [32].
34.7%; hazard ratio (HR) 0.936; 95% confidence interval Whereas gain‐of‐function mutations were associated with
(CI) 0.89–0.99; p = 0.016) in the primary endpoint, which increased levels of LDL cholesterol and early onset of ath
was a composite of CV death, MI, unstable angina requir erosclerosis, loss‐of‐function mutations on the other hand
ing hospitalization, coronary revascularization, or stroke, was linked to a lower LDL cholesterol and a subsequent
compared to simvastatin monotherapy. decrease in cardiovascular risk [33]
Fibrates are agonists of peroxisome proliferator‐acti The FOURIER [34] trial included 27 564 patients with
vated receptor alpha (PPARα), mediating transcription ASCVD and LDL ≥ 70 mg/dL who were receiving statin
factors that control lipoprotein metabolism. They improve therapy. Patients were randomly assigned to evolocumab
the lipid profile of diabetic dyslipidemia by decreasing TG (140 mg every 2 weeks or 420 mg monthly) or placebo, and
level and increasing HDL‐C level, but recent trials have at 48 weeks, evolocumab reduced LDL cholesterol by 59%
failed to show a benefit in outcomes, both with monother compared to placebo, from a median baseline value of
apy [27] and in addition to a statin [28]. In the ACCORD 92 mg/dL to 30 mg/dL. With a median follow‐up of 2.2
(Action to Control Cardiovascular Risk in Diabetes) lipid years, evolocumab significantly reduced the primary end
trial, the addition of fenofibrate to simvastatin in high‐risk point, which was a composite of CV death, MI, stroke, hos
patients with type 2 diabetes mellitus did not reduce the pitalization for unstable angina, or coronary
primary endpoint, which was a composite of MI, stroke, or revascularization, by 15%, and the key secondary endpoint,
CV death, with a mean follow‐up of 4.7 years. However, in which was a composite of CV death, MI, or stroke, by 20%.
a prespecified subgroup analysis, a possible benefit for Diabetes mellitus was present at baseline in 11 031 (40%)
patients with both a high baseline TG level ≥ 204 mg/dL patients, and in a prespecified secondary analysis, similar
and a low baseline level of HDL‐C ≤ 34 mg/dL was sug efficacy in the primary and key secondary endpoints were
gested, with similar post hoc subgroup analysis in the observed in patients with and without diabetes mellitus.
FIELD (Fenofibrate Intervention and Event Lowering in However, since patients with diabetes mellitus had a higher
Diabetes) study [29]. baseline risk, they seemed to have a greater absolute risk
reduction in the primary endpoint over 3 years. This ben
efit was driven largely by a greater absolute risk reduction
Evidence that PSCK 9 is associated with
in coronary revascularization, and there was no difference
increased CVD risk
in the absolute risk reduction for the key secondary
The newest agents in the treatment of dyslipidemia are the endpoint.
Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9) ODYSSEY Outcomes [35] included 18 924 patients who
Inhibitors. In 2003, Abifadel et al. were the first to identify had been hospitalized for an ACS 1 to 12 months prior to
mutations in the genes encoding for PCSK9 as a cause of randomization. After a run‐in period of 2 to 16 weeks on
autosomal familial hypercholesterolemia [30] Investigating high‐intensity statins, patients with LDL cholesterol ≥
the role of PCSK9 showed its association with LDL recep 70 mg/dL, non‐HDL cholesterol ≥ 100 mg/dL, or apolipo
tor intracellular degradation [31]. Plasma LDL cholesterol protein B ≥ 80 mg/dL were randomized to alirocumab
(75 mg every 2 weeks) or placebo. A target LDL cholesterol ing LDL cholesterol levels. Also, as it is only a twice‐yearly
level of 25 to 50 mg/dL was specified, with up‐titration of injection it offers a more practical alternative to the mono
alirocumab to 150 mg every 2 weeks in patients with LDL clonal antibody therapies targeting PCSK9.
≥ 50 mg/dL and a blinded switch to placebo in patients
who consistently had LDL < 15 mg/dL. In the on‐treatment
Evidence linking hypertriglyceridemia to
analysis, which excluded LDL cholesterol values after pre
CVD risk in diabetes mellitus
mature treatment discontinuation or blinded switch to pla
cebo, alirocumab reduced LDL by 61% from a mean LDL While there is substantial evidence that raised LDL choles
of 96.4 mg/dL to 42.3 mg/dL at 1 year, and by 54.7% from a terol levels are a risk factor for CVD in patients with diabe
mean LDL of 101.4 mg/dL to 53.3 mg/dL at 4 years. With a tes mellitus, whether hypertriglyceridemia independently
median follow‐up of 2.8 years, alirocumab significantly is a risk factor for CVD is a matter for debate. This is prob
reduced the primary endpoint, which was a composite of ably due to conflicting results in the studies performed.
CHD death, MI, ischemic stroke, or unstable angina In most studies assessing the role of hypertriglyceri
requiring hospitalization, by 15%, and the secondary com demia in CVD development in diabetes mellitus patients
posite endpoint of all‐cause death, MI, or ischemic stroke the triglyceride levels are obtained after an 8–12 hour fast.
by 14%. Although all‐cause death was significantly lower It is important to note this when assessing triglyceride lev
with alirocumab, there were no significant differences in els in our patients as many present with triglyceride results
CHD death and CV death. Roughly 30% of patients had from non‐fasting samples. However, there is some evi
diabetes mellitus. This led to the ODYSSEY DM – dence that non‐fasting triglyceride levels may have stronger
DYSLIPIDEMIA and ODYSSEY DM – INSULIN [36] tri associations with CVD risk than fasting levels [38]. In the
als being undertaken. In ODYSSEY DM – DYSLIPIDAEMIA Women’s Health Study, non‐fasting triglyceride levels were
participants with type 2 diabetes mellitus and mixed dys found to be independently associated with CVD events
lipidemia (defined as non‐ HDL‐C > 100 mg/dl) were ran while fasting triglycerides were not [39].
domized to open label alirocumab 75 mg every two weeks The combination of raised triglyceride levels, low HDL
or usual care for 24 weeks. 142 participants were included levels and the presence of VLDL particles has been shown
for analysis and 95.1% of these participants had evidence of to be associated with increased CVD risk [40] but the inde
coronary heart disease. In the treatment arm, 64.6% pendent role of triglyceride levels has proved
achieved non‐HDL‐C of < 100 mg/dl compared with 23.8% controversial.
in the usual treatment arm. On review of the available evidence, when differences in
In ODYSSEY DM – INSULIN, participants had type 2 HDL and LDL cholesterol levels are taken into account the
diabetes mellitus treated with insulin and had an LDL> association between triglycerides and CVD risk is less
70 mg/dl. They were randomized in a double‐blind fashion impressive. For example, in patients being treated with
to alirocumab 75 mg every two weeks or placebo for statins, triglyceride level was not associated with CVD risk
24 weeks. The analysis included 177 participants and 86.4% in the Air Force/Texas Coronary Atherosclerosis
of them had evidence of coronary heart disease. The results Prevention Study (AFCAPS/TexCAPS) [41]. Triglyceride
showed 65.4% of participants achieved non‐HDL‐C of < levels were also not shown to have an association with
100 mg/dl compared with 14.9% in the placebo group. CVD risk in the Department of Veterans Affairs High‐
Recently, a trial looking at the use of inclisiran, an inter Density Lipoprotein Intervention Trial (VA‐HIT) [42].
fering RNA shown to inhibit the hepatic synthesis of It has also been shown that the risk of CVD does not
PCSK9, in patients with heterozygous familial hypercho appear to be elevated in patients with inherited forms of
lesterolemia has been published [37]. This study contained severe hypertriglyceridemia, unless the inherited form is
482 patients, of which only 48 had diabetes mellitus. The associated with increased apolipoprotein production or
treatment group showed a 38.1% reduction in LDL choles associated with an increase in triglyceride‐rich remnant
terol levels. This is another interesting pathway for lower particles [43].
However, more recent data increasingly shows an inde Bezafibrate Infarction Prevention Study and included both
pendent role of triglycerides in the development of CVD. diabetic and non‐diabetic patients. The increased risk for
For example, a 2016 study that controlled for body mass CHD related to elevated triglycerides was seen in both dia
index showed that an elevated triglyceride level was associ betic and non‐diabetic subgroups.
ated with increased coronary plaque development in Nordestgaard et al. examined the hypothesis that very
patients whose LDL‐ Cholesterol was well controlled with high levels of non‐fasting triglycerides were predictors of
lipid‐lowering therapy [44]. myocardial infarction, ischemic heart disease and
There are also studies showing that triglyceride levels death [48]. This was a prospective cohort study of
are an independent marker for recurrent disease after MI 7587 women and 6394 men from the general Danish popu
in those with controlled LDL [45]. It has also been shown lation followed from 1976–2004. Baseline non‐fasting tri
that non fasting triglycerides were associated with an glycerides were stratified into categories of increasing
increased risk of CVD of 2.8 times for each 1 mmol/L severity of hypertriglyceridemia, and compared to triglyc
(89 mg/dL) increase in triglyceride levels [46]. eride levels of less than 88.5mg/dl (1 mmol/l). Non‐fasting
The Hoorne study [47] was a population‐based cohort triglyceride levels of 442.5mg/dl (≥ 5 mmol/l) or more were
study with 869 men and 948 women participants which found to be predictive of CHD. This study also supported
showed hypertriglyceridemia to be an independent risk the concept that non‐fasting triglyceride levels more
factor for CVD in those with abnormal blood glucose. strongly predict CHD risk than levels measured after a
After 10 years of follow‐up, they found the age‐ and sex‐ 12–14 hour fast. Postprandial lipoproteins are generally tri
adjusted hazard ratios for cardiovascular disease were 1.35 glyceride‐rich, and if an individual has a predisposition to
(1.11–1.64) and 1.71 (1.40–2.08) for high triglycerides and small, dense LDL or has insulin‐resistance, then clearance
high non‐HDL‐cholesterol respectively after mutual of these lipoprotein particles can be delayed for up to 12
adjustment. After stratification for glucose metabolism sta hours. The authors thus postulate that prolonged exposure
tus, the hazard ratios for cardiovascular disease for non‐ of the endothelium to triglyceride‐rich, atherogenic rem
HDL‐cholesterol were 1.70 (1.31–2.21) in normal glucose nant particles, or the associated states in which atherogenic
metabolism and 1.56 (1.12–2.18) in abnormal glucose lipoprotein particles occur, such as obesity or the metabolic
metabolism. Triglycerides were not a risk factor in subjects syndrome, may explain why non‐fasting triglycerides are
with normal glucose metabolism, with a hazard ratio of more predictive of CHD risk.
0.94 (0.73–1.22), but in subjects with abnormal glucose In summary, there is strong epidemiological evidence to
metabolism, the hazard ratio for cardiovascular disease date that hypertriglyceridemia is an independent risk fac
was 1.54 (1.07–2.22). In subjects with abnormal glucose tor for CHD. Prospective randomized controlled trials
metabolism, the hazard ratio for the combined presence of examining this link are difficult given the large number of
high triglycerides and non‐HDL‐cholesterol was 2.12 patients with diabetes mellitus who are already receiving
(1.35–3.34). statin treatment (which modestly lowers triglycerides) and
In 2001 the influence of low high‐density lipoprotein due to the presence of other lipoprotein and metabolic
cholesterol and elevated triglyceride on coronary heart dis derangements which may confound results.
ease events and response to simvastatin therapy in 4S (the
Scandinavian Simvastatin Survival Study) reported that
Evidence that treating
patients with elevated LDL cholesterol, low HDL choles
hypertriglyceridemia will decrease CVD
terol, and elevated triglycerides were more likely than
risk in DM
patients with isolated LDL cholesterol elevation to have
other characteristics of the metabolic syndrome, had Now that we have reviewed the evidence linking elevated
increased risk for CHD events on placebo, and received triglyceride levels to increased cardiovascular risk in diabe
greater benefit with simvastatin therapy. This was a sub tes mellitus patients, we next need to establish if there is
study analyses of the Helsinki Heart Study and the evidence that treating this lipid abnormality decreases this
risk. Fenofibrates, niacin and omega 3 fish oils are some diabetes mellitus patients, of which 418 were randomized
of the pharmacological interventions for treating to either receive fenofibrate 200 mg od or placebo for three
hypertriglyceridemia. years. This was both a primary and secondary prevention
In the Bezafibrate Infarction Prevention (BIP), 3090 trial as one‐half of the participants had a history of CVD.
patients with a previous myocardial infarction or stable Total plasma cholesterol, HDL‐cholesterol, LDL‐choles
angina, total cholesterol of 180 to 250 mg/dl, HDL‐choles terol, and triglyceride concentrations all changed signifi
terol < 45 mg/dl, triglycerides < 300 mg/dl, and LDL‐cho cantly more than baseline in the fenofibrate group (n=207)
lesterol < 180 mg/dl were randomized to receive either than in the placebo group (n=211). The fenofibrate group
400 mg of bezafibrate daily or placebo, and followed for a showed a significantly smaller increase in percentage
mean of 6.2 years [49]. Bezafibrate reduced triglycerides by diameter stenosis on coronary angiogram than the placebo
21% and raised HDL cholesterol by 18%. No difference was group (mean 2.11 vs 3.65, P=0.02). Although the trial was
apparent in the all‐cause and cardiac mortality between the not powered to examine clinical endpoints, there were
bezafibrate and placebo groups. However, on post‐hoc fewer in the fenofibrate than placebo group (38 vs 50).
analysis, there was a significant reduction in the primary Although there was angiographic benefit in the setting of
end point in 459 patients with high baseline triglycerides reduction of hypertriglyceridemia, as in most trials on tri
(≥ 200 mg/dl or 2.26 mmol/l). The reduction in the cumu glycerides it is difficult to deduce if this is a direct triglycer
lative probability of the primary endpoint by bezafibrate ide‐lowering effect or the result of correction of other
was 39.5% (P=0.02). Bezafibrate may thus have a promi lipoprotein abnormalities.
nent role in the management of dyslipidemia and CHD Niacin or nicotinic acid has been shown to reduce tri
when targeted to a subgroup of patients with CHD. This glyceride levels and increase HDL by up to 30% when
supports the evidence from meta‐analyses and epidemio doses of 1.5–3 g daily are used [52]. In older niacin trials
logical studies that triglycerides are indeed an independent for secondary cardiovascular prevention before the advent
risk factor for CHD in patients with both normal and of statin therapy, absolute mortality reductions of 6.2% and
abnormal glucose metabolism. 7.8% were demonstrated, compared to the best absolute
The authors subsequently evaluated the effect of bezafi mortality reduction of 3.5% with a statin in the
brate on the incidence of MI in patients enrolled in the BIP Scandinavian Simvastatin Survival Study [53].
study who met the criteria for the metabolic syndrome [50]. However, there is no evidence to suggest outcomes are
The study sample for this post hoc subgroup analysis com improved with niacin for those already established on statin
prised 1470 patients aged 42 to 74 years who were randomly therapy. The HPS2‐THRIVE (Heart Protection Study
assigned to receive bezafibrate 400 mg daily (740 patients) or 2‐Treatment of HDL to Reduce the Incidence of Vascular
placebo (730 patients). The follow‐up period was 6.2 years Events) study of 25 673 patients, which examined the impact
for events and 8.1 years for mortality data. New myocardial of niacin 2 g daily plus laropiprant (to reduce flushing) ver
infarction was recorded in 82 patients from the bezafibrate sus double placebo [54], was conducted on background sim
group (11.1%) and 111 patients (15.2%) from the placebo vastatin ± ezetimibe therapy. After a median follow‐up of 3.9
group (P=0.02). Bezafibrate was associated with a reduced years, the primary endpoint of first major vascular event was
risk of any MI and non‐fatal MI with hazard ratios of 0.71 not significantly lower in niacin‐treated patients. Patients on
(95% CI, 0.54–0.95) and 0.67 (95% CI, 0.49–0.91) respec niacin experienced more serious adverse events including
tively. The cardiac mortality risk tended to be lower in myopathy, gastrointestinal effect, rash, and an increase in
patients taking bezafibrate (HR, 0.74; 95% CI, 0.54–1.03). infections and bleeding. This study also demonstrated unfa
However, posthoc analysis on the effect on MI risk of specifi vorable glycemic effects, with an increase in both new cases
cally lowering triglycerides within this group with metabolic of diabetes mellitus and significant worsening of glycemic
syndrome was not performed. control in patients with established disease.
Regarding fenofibrate, HDL, and cardiovascular disease Omega‐3 fatty acid preparations containing eicosapen
in Type‐2 diabetes, the DAIS trial [51] screened 731 type 2 taenoic acid (EPA) and/or docosahexaenoic acid (DHA)
reduce fasting and postprandial triglycerides through sup 1.28 mmol/L) based on physiologically higher HDL levels
pression of hepatic VLDL production [55]. They are fre in the female population. Increasing HDL cholesterol levels
quently used as an adjutant to other therapies. Doses of 3–4 pharmacologically in patients with diabetes mellitus is dif
g per day of EPA ± DHA are required to reduce triglycerides ficult since the most effective agent is nicotinic acid, which
by up to 45% [56]. Outcomes trials yielded mixed results for is relatively contraindicated in diabetes mellitus. We agree
cardiovascular benefit with the use of omega‐3 fatty with current ADA guidelines that triglycerides are a recog
acids [57, 58]. However, doses used in initial studies were nized target for intervention in diabetes and that optimal
much lower than currently recommended for hypertriglyc triglyceride levels are < 150 mg/dL (1.7 mmol/L).
eridemia (1 g instead of 3–4 g), and these low doses were Now to relate these guidelines to common clinical sce
examined on a background statin therapy. A subsequent narios encountered in our practice. Firstly, we will look at
landmark study in Japan utilized a moderate dose of EPA the scenario described at the beginning of the chapter
ethyl esters (1.8 g) in conjunction with low‐dose statin ther where our patient’s LDL was at target but they had an iso
apy versus statin therapy alone, and observed a 19% reduc lated hypertriglyceridemia. The initial therapeutic inter
tion in major coronary events in those receiving EPA [59]. vention for hypertriglyceridemia is behavioral modification
with weight loss, increased physical activity and modera
tion of alcohol consumption. Though not required in our
Discussion
scenario, should the patient have poor glycemic control we
There are exciting new developments in the treatment of would advise intensification of their diabetic treatment
dyslipidemia in diabetes mellitus patients with the arrival regimen.
of the PCSK 9 inhibitors. These agents are changing the Another clinical scenario to consider is the patient with
landscape of lipid management but cost remains a barrier mixed dyslipidemia as is commonly seen in the type 2 dia
to widespread adoption. As we have shown above, hyper betes mellitus population. If the patient is not receiving
triglyceridemia is associated with the metabolic syndrome. treatment, we would advise the behavioral modifications
While LDL is an established independent risk factor for outlined above and improvements in glycemic control. If
CVD in type 2 diabetes mellitus, there are still controver this fails to achieve the required targets, we would advise
sies surrounding the evidence that suggests aggressively statin as first‐line therapy. The decision as to whether to
lowering triglycerides with pharmacotherapy is of benefit start high‐dose or low‐dose statin therapy, we would argue,
in the diabetic population. needs to be individualized to each patient after reviewing a
We would recommend statin treatment for all patients number of factors. Initially, we need to look for the pres
with type 2 diabetes mellitus to lower LDL, thus reducing ence of other CVD risk factors and, if present, assess how
cardiovascular risk. The ADA recommends initiating a sta well are these risk factors are controlled. The greater the
tin for primary prevention when LDL is > 130 mg/dL CVD risk the more likely we would be to initiate high‐dose
(3.36 mmol/L) with the goal of reducing LDL to 100 mg/dl statin therapy, especially if the modifiable risk factors are
(< 2.57 mmol/l). Target LDL in current guidelines is < poorly controlled. The trials looking at high‐dose statins
2.6 mmol/l and < 1.8 mmol/L for secondary prevention. versus low‐dose statins were undertaken in participants
We recommend the lowest possible LDL for secondary pre with established CVD and so these are the individuals that
vention based on available trial data. We would suggest that should be selected for high‐dose therapy in our patient
a target of LDL cholesterol < 70 mg/dl (1.8 mmol/l) may be cohort. Next, we would consider the degree of the derange
reasonable in patients with type two diabetes mellitus and ment in LDL and TG. Those with only modest increase
CVD. This, however, must be taken in the context of with a low CVD risk and no evidence of CVD should be
increased risk of adverse events related to statin‐lowering started on low‐dose statin therapy. Lastly, we need to con
therapy on higher doses of these agents. Optimal HDL lev sider if this patient is at risk from any of the described sta
els are > 40 mg/dL (1.02 mmol/L) in men; a higher target tin side effects and consider whether gradual dose titration
may be more desirable in women (> 50 mg/dL or would be more successful in this population.
The third scenario is the patient on statin therapy who for cholesteryl ester transfer protein. Am J Physiol. 1998;274:
has an LDL cholesterol at or below target, but triglycerides E1091–1098.
are elevated and HDL suboptimal. The physician may con 4. Barter PJ, Caulfield M, Eriksson M et al.; ILLUMINATE
sider adding a fibrate or fish oil therapy. We would not rec Investigators. Effects of torcetrapib in patients at high risk
for coronary events. N Engl J Med. 2007;357:2109–2122.
ommend niacin be used routinely in the diabetic mellitus
5. Kastelein JJ, van Leuven SI, Burgess L et al.; RADIANCE
population. Unfortunately, the combination of a statin and
1 Investigators. Effect of torcetrapib on carotid atherosclero
fibrate increases the risk of side effects such as myopathy.
sis in familial hypercholesterolemia. N Engl J Med.
Therefore, for more modest hypertriglyceridemia (1.7– 2007;356:1620–1630.
3.0 mmol/l), we would recommend the addition of fish oil 6. Lincoff AM, Nicholls SJ, Riesmeyer JS et al.; ACCELERATE
therapy to help achieve targets. A Cochrane review of 23 Investigators. Evacetrapib and cardiovascular outcomes
randomized control trials involving 1075 participants with in high‐risk vascular disease. N Engl J Med. 2017;376:
type 2 diabetes mellitus showed that supplementation with 1933–1942.
fish oil (average dose of 3.5 g/day for an average of 7. Bowman L, Hopewell JC, Chen F et al.; HPS3/TIMI55–
8.9 weeks) lowered TG levels by 8.1 mg/dl and VLDL by REVEAL Collaborative Group. Effects of anacetrapib in
1.26 mg/dl [60]. LDL levels increased by 1.98 mg/dl and patients with atherosclerotic vascular disease. N Engl J Med.
there was no change in HDL levels. There is a lack of strong 2017;377:1217–1227pmid:28847206.
8. Krauss RM. Heterogeneity of plasma low‐density lipopro
evidence to guide management in this scenario. It would be
teins and atherosclerosis risk. Curr Opin Lipidol. 1994;5:
reasonable to monitor the TG levels and only intervene if
339–349.
approaching levels where pancreatitis could develop (TG >
9. Lamarche B, Chernof A, Moorjani S et al. Small, dense low‐
5 mmol/L) or in case of existing CAD (although no evi density lipoprotein particles as a predictor of the risk of
dence to support this). Features consistent with familial ischemic heart disease in men. Prospective results from the
combined hyperlipidemia (varying phenotype over time or Quebec Cardiovascular Study. Circulation. 1997;95:69–75.
family history) might also prompt intervention. However, 10. Chapman MJ, Guerin M, Bruckert E. Atherogenic, dense
should the above patient have an LDL and TG level above low‐density lipoproteins. Pathophysiology and new thera
target then the risk/benefit balance could be in favor of peutic approaches. Eur Heart J. 1998;19(Suppl A): A24–30.
combination statin and fibrate therapy. Fibrates such as 11. Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH,
fenofibrate, gemfibrozil, or bezafibrate can be considered Franklin BA, Gordon DJ, Krauss RM, Savage PJ, Smith SC Jr,
Spertus JA, Costa F, American Heart Association, National
as an adjunct to statin therapy.
Heart, Lung, and Blood Institute. Diagnosis and manage
These are our current recommendations for managing
ment of the metabolic syndrome: an American Heart
dyslipidemia in type 2 diabetes mellitus. In the future,
Association/National Heart, Lung, and Blood Institute
when further long‐term studies emerge relating the PCSK9 Scientific Statement. Circulation. 2005;Oct 25;112(17):
agents, we may find that our LDL targets have been too 2735–2752.
conservative and more aggressive targets may become 12. Giugliano, RP, Cannon CP, Blazing MA, Nicolau JC,
standard practice. Corbalan R, Spinar J, Park JG, White JA, Bohula E,
Braunwald E. Benefit of adding ezetimibe to statin therapy
on cardiovascular outcomes and safety in patients with vs.
References
without diabetes: results from IMPROVE‐IT. Circulation.
1. Taskinen MR. Diabetic dyslipidaemia: from basic research to 2018;Apr 10;137(15):1571–1582.
clinical practice. Diabetologia. 2003;46:733–749. 13. Goff DC, Gerstein HC, Ginsberg HN, Cushman WC,
2. Savage DB, Petersen KF, Shulman GI. Disordered lipid Margolis KL, Byington RP, Buse JB, Genuth S, Probstfield JL,
metabolism and the pathogenesis of insulin resistance. Simons‐Morton DG. Prevention of cardiovascular disease in
Physiol Rev. 2007;87:507–520. persons with type 2 diabetes mellitus: current knowledge
3. C Caste, Kuiper S, Blake W, Paigen B, Marotti K, Melchior G. and rationale for the Action to Control Cardiovascular Risk
Remodeling of the HDL in NIDDM has a fundamental role in Diabetes (ACCORD) trial. Am J Cardiol. 2007;99: 4i–20i.
14. Colhoun HM, Betteridge DJ, Durrington PN et al. on behalf 23. Sacks FM, Pfeffer MA, Moye LA et al. for the Cholesterol
of the CARDS Investigators. Primary prevention of cardio And Recurrent Events Trial Investigators. The effect of
vascular disease with atorvastatin in type 2 diabetes in the pravastatin on coronary events after myocardial infarction
Collaborative Atorvastatin Diabetes Study (CARDS): multi in patients with average cholesterol levels. N Engl J Med.
centre randomised placebo‐controlled trial. Lancet. 2004; 1996;335:1001–1009.
364:685–696. 24. LaRosa JC, Grundy SM, Waters DD et al. for the Treating to
15. Knopp RH, d’Emden M, Smilde JG, Pocock SJ, on behalf of New Targets (TNT) Investigators. Intensive lipid lowering
the ASPEN Study Group. Efficacy and safety of atorvastatin with atorvastatin in patients with stable coronary disease. N
in the prevention of cardiovascular end points in subjects Engl J Med. 2005;352:1425–1435.
with type 2 diabetes: the Atorvastatin Study for Prevention 25. Cannon CP, Braunwald E, McCabe CH et al. for the
of Coronary Heart Disease Endpoints in Non‐insulin‐ Pravastatin or Atorvastatin Evaluation and Infection
dependent diabetes mellitus (ASPEN). Diabetes Care. 2006; Therapy‐Thrombolysis in Myocardial Infarction 22 Investi
29:1478–1485. gators. Intensive versus conventional lipid lowering with
16. Heart Protection Study Collaborative Group. MRC/BHF statins after acute coronary syndromes. N Engl J Med. 2004;
Heart Protection Study of cholesterol lowering with simvas 350:1495–1504.
tatin in 5963 people with diabetes: a randomised placebo‐ 26. Cannon CP, Blazing MA, Giugliano RP, McCagg A, White
controlled trial. Lancet. 2003;361:2005–2016. JA, Theroux P, Darius H, Lewis BS, Ophuis TO, Jukema JW
17. Sever PS, Poulter NR, Dahlöf B et al. for the ASCOT et al. Ezetimibe added to statin therapy after acute coronary
Investigators. Reduction in cardiovascular events with ator syndromes. N Engl J Med. 2015;372:2387–2397.
vastatin in 2,532 patients with type 2 diabetes. Anglo‐ 27. Keech A, Simes RJ, Barter P, Best J, Scott R, Taskinen MR,
Scandinavian Cardiac Outcomes Trial‐Lipid Lowering Arm Forder P, Pillai A, Davis T, Glasziou P et al. Effects of long‐
(ASCOT‐LLA). Diabetes Care. 2005;28:1151–1157. term fenofibrate therapy on cardiovascular events in 9795
18. Nakamura H, Arakawa K, Itakura H et al. for the MEGA people with type 2 diabetes mellitus (the FIELD study):
Study Group. Primary prevention of cardiovascular disease Randomised controlled trial. Lancet. 2005;366:1849–1861.
with pravastatin in Japan (MEGA Study): a prospective ran 28. Ginsberg HN, Elam MB, Lovato LC, Crouse JR, 3rd, Leiter
domised controlled trial. Lancet. 2006;368:1155–1163. LA, Linz P, Friedewald WT, Buse JB, Gerstein HC,
19. ALLHAT Collaborative Research Group. Major outcomes in Probstfield J et al. Effects of combination lipid therapy in
moderately hypercholesterolemic, hypertensive patients type 2 diabetes mellitus. N Engl J Med. 2010;362:
randomized to pravastatin vs usual care. The 1563–1574.
Antihypertensive and Lipid‐Lowering Treatment to Prevent 29. Ginsberg HN, Elam MB, Lovato LC, Crouse JR, 3rd, Leiter
Heart Attack Trial (ALLHAT‐LLA). JAMA. 2002;288: LA, Linz P, Friedewald WT, Buse JB, Gerstein HC,
2998–3007. Probstfield J et al. Effects of combination lipid therapy in
20. Pyörälä K, Pedersen TR, Kjekshus J, Faergeman O, Olsson type 2 diabetes mellitus. N Engl J Med. 2010;362:
AG, Thorgeirsson G. Cholesterol lowering with simvastatin 1563–1574.
improves prognosis of diabetic patients with coronary heart 30. Abifadel M, Varret M, Rabès JP, Allard D, Ouguerram K,
disease: a subgroup analysis of the Scandinavian Simvastatin Devillers M, Cruaud C, Benjannet S, Wickham L, Erlich D,
Survival Study (4S). Diabetes Care. 1997;20:614–620. Derré A, Villéger L, Farnier M, Beucler I, Bruckert E,
21. Haffner SM, Alexander CM, Cook TJ et al. for the Chambaz J, Chanu B, Lecerf JM, Luc G, Moulin P,
Scandinavian Simvastatin Survival Study Group. Reduced Weissenbach J, Prat A, Krempf M, Junien C, Seidah NG,
coronary events in simvastatin‐treated patients with coro Boileau C. Mutations in PCSK9 cause autosomal dominant
nary heart disease and diabetes or impaired fasting glucose hypercholesterolemia. Nat Genet. 2003;Jun;34(2):154–156.
levels. Arch Intern Med. 1999;159:2661–2667. 31. Abifadel M, Varret M, Rabes JP, Allard D, Ouguerram K,
22. The Long‐term Intervention with Pravastatin in Ischaemic Devillers M, Cruaud C, Benjannet S, Wickham L, Erlich D
Disease (LIPID) Study Group. Prevention of cardiovascular et al. Mutations in PCSK9 cause autosomal dominant hyper
events and death with pravastatin in patients with coronary cholesterolemia. Nat Genet. 2003;34:154–156.
heart disease and a broad range of initial cholesterol levels. N 32. Maiolino G, Rossitto G, Caielli P, Bisogni V, Rossi GP, Calò
Engl J Med. 1998;339:1349–1357. LA. The role of oxidized low‐density lipoproteins in athero
sclerosis: the myths and the facts. Mediators Inflamm. Affairs High‐Density Lipoprotein Intervention Trial.
2013;2013:714653. Relation of gemfibrozil treatment and lipid levels with major
33. Cohen JC, Boerwinkle E, Mosley TH, Jr, Hobbs HH. coronary events: VA‐HIT: a randomized controlled trial.
Sequence variations in PCSK9, low LDL, and protection JAMA. 2001;Mar 28;285(12):1585–1591.
against coronary heart disease. N Engl J Med. 2006;354(12): 43. Miller M, Stone NJ, Ballantyne C, Bittner V, Criqui MH,
1264–1272. Ginsberg HN, Goldberg AC, Howard WJ, Jacobson MS, Kris‐
34. Sabatine MS, Giugliano RP, Keech AC, Honarpour N, Etherton PM, Lennie TA, Levi M, Mazzone T, Pennathur S,
Wiviott SD, Murphy SA, Kuder JF, Wang H, Liu T, American Heart Association Clinical Lipidology, Thrombosis,
Wasserman SM et al. Evolocumab and clinical outcomes in and Prevention Committee of the Council on Nutrition,
patients with cardiovascular disease. N Engl J Med. 2017;376: Physical Activity, and Metabolism, Council on Arteriosclerosis,
1713–1722. Thrombosis and Vascular Biology, Council on Cardiovascular
35. Steg PG. Evaluation of cardiovascular outcomes after an Nursing, Council on the Kidney in Cardiovascular Disease.
acute coronary syndrome during treatment with alirocumab Triglycerides and cardiovascular disease: a scientific statement
– ODYSSEY OUTCOMES. In Proceedings of the American from the American Heart Association. Circulation. 2011;May
College of Cardiology Annual Scientific Session (ACC 24;123(20):2292–2333.
2018), Orlando, FL, USA, 10–12 March 2018. 44. Puri R, Nissen SE, Shao M, Elshazly MB, Kataoka Y, Kapadia
36. Ray KK et al. Alirocumab therapy in individuals with type 2 SR, Tuzcu EM, Nicholls SJ. Non‐HDL Cholesterol and tri
diabetes mellitus and atherosclerotic cardiovascular disease: glycerides: implications for coronary atheroma progression
analysis of the ODYSSEY DM‐DYSLIPIDEMIA and DM‐ and clinical events. Arterioscler Thromb Vasc Biol.
INSULIN studies. Cardiovasc Diabetol. 2019;18:149 https:// 2016;Nov;36(11):2220–2228.
doi.org/10.1186/s12933‐019‐0951‐9. 45. Miller M, Cannon CP, Murphy SA, Qin J, Ray KK, Braunwald
37. Raal, FJ, Kallend DR, Kausik K, Turner T, Koenig W, Wright E, PROVE IT‐TIMI 22 Investigators. Impact of triglyceride
R Scott, Wijngaard, PLJ, Curcio D, Jaros MJ, Leiter LA, levels beyond low‐density lipoprotein cholesterol after acute
Kastelein JJP. Inclisiran for the treatment of heterozygous coronary syndrome in the PROVE IT‐TIMI 22 trial. J Am
familial hypercholesterolemia, 2020/03/18, N Engl J Med. Coll Cardiol. 2008;Feb 19;51(7):724–730.
1520–1530, 382, 16, 10.1056/NEJMoa1913805. 46. Watts GF, Karpe F. Triglycerides and atherogenic dyslipidae
38. Di Angelantonio E, Sarwar N, Perry P, Kaptoge S, Ray KK, mia: extending treatment beyond statins in the high‐risk
Thompson A, Wood AM, Lewington S, Sattar N, Packard CJ, cardiovascular patient. Heart. 2011;97(5):350–356.
Collins R, Thompson SG, Danesh J. Emerging risk factors 47. Bos G, Dekker JM, Nijpels G, de Vegt FA, Diamant M,
collaboration – major lipids, apolipoproteins, and risk of Stehouwer CDA, Bouter LM, Heine RJ. A combination of
vascular disease. JAMA. 2009;302(18):1993–2000. high concentrations of serum triglyceride and non‐HDL
39. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg‐Hansen A. cholesterol is a risk factor for cardiovascular disease in sub
Nonfasting triglycerides and risk of myocardial infarction, jects with abnormal glucose metabolism: The Hoorn Study.
ischemic heart disease, and death in men and women. Diabetologia. 2003;46:910–916.
JAMA. 2007;298(3):299–308. 48. Nordestgaard BG, Benn M, Schnohr P et al. Nonfasting triglyc
40. Musunuru K. Atherogenic dyslipidemia: cardiovascular risk erides and risk of myocardial infarction, ischemic heart disease
and dietary intervention. Lipids. 2010;Oct;45(10):907–914. and death in men and women. JAMA. 2007;298(3):299–308.
41. Gotto AMJ Jr, Whitney E, Stein EA, Shapiro DR, Clearfield 49. The Bezafibrate Infarction Prevention (BIP) study.
M, Weis S, Jou JY, Langendörfer A, Beere PA, Watson DJ, Secondary prevention by raising HDL cholesterol and
Downs JR, de Cani JS. Relation between baseline and on‐ reducing triglycerides in patients with coronary artery dis
treatment lipid parameters and first acute major coronary ease. Circulation. 2000;102:21–27.
events in the Air Force/Texas Coronary Atherosclerosis 50. Tenenbaum A, Motro M, Enrique Z et al. Bezafibrate for the
Prevention Study (AFCAPS/TexCAPS). Circulation. 2000; secondary prevention of myocardial infarction in patients
101(5):477–484. with metabolic syndrome. Arch Intern Med. 2005;165:
42. Robins SJ, Collins D, Wittes JT, Papademetriou V, Deedwania 1154–1160.
PC, Schaefer EJ, McNamara JR, Kashyap ML, Hershman JM, 51. Steiner G et al. Effect of fenofibrate on progression of
Wexler LF, Rubins HB, VA‐HIT Study Group. Veterans coronary‐artery disease in type 2 diabetes: the Diabetes
Atherosclerosis Intervention Study (DAIS), a randomised in severe hypertriglyceridemia. J Cardiovasc Risk.

study. Lancet. 2001;357:905–910. 1997;4(5–6):385–391.
52. Chapman MJ, Redfern JS, McGovern ME, Giral P. Niacin 57. The ORIGIN Trial Investigators, Bosch J, Gerstein HC,
and fibrates in atherogenic dyslipidemia: pharmacotherapy Dagenais GR, Diaz R, Dyal L et al. n‐3 fatty acids and cardio
to reduce cardiovascular risk. Pharmacol Ther. vascular outcomes in patients with dysglycemia. N Engl J
2010;126(3):314–345. Med. 2012;367(4):309–318.
53. Superko HR, Zhao XQ, Hodis HN, Guyton JR. Niacin and 58. Risk and Prevention Study Collaborative Group, Roncaglioni
heart disease prevention: engraving its tombstone is a mis MC, Tombesi M, Avanzini F, Barlera S et al. n‐3 fatty acids in
take. J Clin Lipidol. 2017;11(6):1309–1317. patients with multiple cardiovascular risk factors. N Engl J
54. Group HTC, Landray MJ, Haynes R, Hopewell JC, Parish S, Med. 2013;368(19):1800–1808.
Aung T et al. Effects of extended‐release niacin with laro 59. Yokoyama M, Origasa H, Matsuzaki M, Matsuzawa Y, Saito
piprant in high‐risk patients. N Engl J Med. 2014;371(3): Y, Ishikawa Y et al. Effects of eicosapentaenoic acid on major
203–212. coronary events in hypercholesterolaemic patients (JELIS): a
55. Nestel PJ, Connor WE, Reardon MF, Connor S, Wong S, randomised open‐label, blinded endpoint analysis. Lancet.
Boston R. Suppression by diets rich in fish oil of very low 2007;369(9567):1090–1098.
density lipoprotein production in man. J Clin Invest. 60. Hartweg J, Perera R, Montori V, Dinneen S, Neil HA, Farmer
1984;74(1):82–89. A. Omega‐3 polyunsaturated fatty acids (PUFA) for type 2
56. Harris WS, Ginsberg HN, Arunakul N, Shachter NS, diabetes mellitus. Cochrane Database Syst Rev. 2008;(1):
Windsor SL, Adams M et al. Safety and efficacy of Omacor CD003205.
17 New Agents for Treatment
of Dyslipidemia
Vinaya Simha
Associate Professor, Division of Endocrinology, Diabetes, Metabolism and nutrition, Mayo Clinic, Rochester, MN, USA
●● Diabetes mellitus is associated with elevated risk for atherosclerotic cardiovascular disease (ASCVD), and statin therapy has
been shown to reduce this risk.
●● Despite optimal statin therapy, there is considerable residual risk, and additional lipid‐lowering therapies may help
ameliorate it.
●● Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a secreted protein which modulates the concentration of low‐den-
sity lipoprotein (LDL)‐cholesterol by promoting lysosomal degradation of LDL receptors. Gain‐of‐function mutations in
PCSK9 is a cause of familial hypercholesterolemia, while loss‐of‐function mutations are associated with low LDL‐cholesterol
and reduced ASCVD risk.
●● Monoclonal antibodies to PCSK9 have been developed (evolocumab and alirocumab), and shown to reduce LDL‐choles-
terol by 50–60% when administered as monotherapy or added to statin therapy.
●● Cardiovascular outcome trails have shown a significant risk reduction (HR = 0.85, ARR = 1.6–2) with both evolocumab and
alirocumab when added on to statin therapy. These benefits were similar in patients with and without diabetes, and were
not associated with significant adverse effects.
●● Current guidelines recommend the addition of PCSK9 inhibitors to maximally tolerated statin therapy in: 1. patients with
familial hypercholesterolemia, 2. patients with established ASCVD whose LDL‐cholesterol ≥ 70 mg/dL on statin +
ezetimibe.
●● PCSK9 inhibitor therapy should be considered in patients with diabetes and ASCVD whose non‐HDL cholesterol ≥ 100 mg/
dL on maximally tolerated statin therapy.
●● Novel triglyceride‐lowering therapies which are being developed also have the potential for lowering ASCVD risk when
added to statin therapy, but need further investigation.
Introduction
aggressively address cardiovascular risk in diabetes. Stain
Diabetes mellitus increases the risk for atherosclerotic car therapy has been consistently shown to reduce ASCVD
diovascular disease (ASCVD) at least two‐fold, with risk in patients with diabetes [5], and has been strongly rec
increasing risk as the duration of diabetes increases [1]. ommended in most patients with diabetes [6]. Together
Indeed ASCVD is the leading cause for morbidity and with appropriate lifestyle changes and glucose‐lowering
mortality in patients with type 2 diabetes [2], and the therapies, this will significantly reduce the burden of
recent heartening trend of decreasing ASCVD mortality in ASCVD. However, considerable residual risk persists even
the overall population [3] is unfortunately not seen in after adequate statin therapy, with an estimated 1 in 7
patients with diabetes [4]. It is therefore imperative to patients experiencing an ASCVD event within 5 years

211
despite ongoing statin therapy [7]. This has prompted an LDL receptor participates in receptor‐mediated endocyto
intense search for additional and novel therapies that sis of circulating LDL particles, thus helping to restore cel
reduce the levels of atherogenic lipoproteins and overall lular cholesterol levels, while simultaneously decreasing
ASCVD risk. LDL particles in the circulation. After internalization
Patients with type 2 diabetes have characteristic lipid through clathrin‐coated pits on the hepatocyte membrane,
abnormalities marked by hypertriglyceridemia, low HDL‐ the receptor dissociates from the LDL particle and recycles
cholesterol and elevated levels of small, dense LDL‐choles back to the surface to continue this process nearly 100–150
terol [8, 9]. Since hypertriglyceridemia is the most common times during its lifecycle. However, when the LDL receptor
lipid abnormality in type 2 diabetes, considerable attention is bound to PCSK9, it is unable to dissociate from the LDL
has focused on primary triglyceride lowering therapies particle, and undergoes degradation along with the LDL
such as fibrates, omega‐3 fatty acid and niacin. Except for particle and PCSK9 within the lysosome (Figure 17.1). The
the recent REDUCE‐IT trial showing the benefit of 4 g net effect of PCSK9 is reduced recycling of LDL receptor to
eicosapent‐ethyl [10], other studies have not shown a sig the hepatocyte membrane and resultant decrease in clear
nificant benefit in statin treated patients, though sub‐group ance of circulating LDL particles. Indeed, overexpression
analysis has suggested some risk reduction with fibrate of PCSK9 in mice is associated with high circulating cho
therapy [11]. Additional LDL‐cholesterol lowering using lesterol concentrations [17], while PCSK9 deficiency leads
ezetimibe has been shown to reduce risk in post‐Acute to low cholesterol levels [18]. These experiments were fol
Coronary Syndrome (ACS) patients [12]. However, the lowed by epidemiological studies which clearly showed
most promising development in recent years has been the decreased LDL‐cholesterol and ASCVD risk in individuals
introduction of a novel class of lipid‐lowering medications, harboring deleterious sequence variations in PCSK9 [19–
Proprotein convertase subtilisin/kexin type 9 (PCSK9) 21]. This has been the basis for development of pharmaco
inhibitors. This review briefly summarizes the role of logical approaches to decrease PCSK9 activity and thus
PCSK9 in regulation of circulating cholesterol levels, clini circulating LDL‐cholesterol levels.
cal trials of PCSK9 inhibitors, and their optimal use in
clinical practice, followed by a brief discussion on emerg
PCSK9 inhibitors: Efficacy and safety
ing triglyceride‐lowering therapies.
Several monoclonal antibodies and a small inhibiting RNA
(siRNA) have been developed for inhibition of PCSK9
PCSK9 and regulation of LDL‐Cholesterol
activity. Evolocumab and alirocumab are two fully human
levels
monoclonal antibodies currently approved for clinical use.
The role of PCSK9 in regulation of circulating cholesterol They are administered by subcutaneous injection, usually
levels was first appreciated around the turn of the century, every two weeks (evolocumab 140 mg and alirocumab 75
when Abifadel and colleagues identified gain‐of‐function or 150 mg). They can also be administered every 4 weeks,
mutations in PCSK9 as a cause for autosomal dominant evolocumab at a dose of 420 mg, and alirocumab at a dose
familial hypercholesterolemia [13]. Since then, there has of 300 mg, respectively.
been considerable progress in understanding the physio Several phase III trials have been conducted in a variety
logical role of this protein [14, 15] which has also led to of patient populations including those with familial hyper
rapid therapeutic advances [16]. cholesterolemia, statin intolerance and moderate hyper
PCSK9 is a circulating serine protease which binds to cholesterolemia, both as monotherapy and in combination
the LDL receptor, leading to their lysosomal degradation. with statin therapy. In general, the reduction in LDL‐cho
Interestingly, the expression of both LDL receptors and lesterol was about 60%, and was accompanied by a roughly
PCSK9 are regulated by the same transcription factor, 50% decrease in Apo B levels and 25% decrease in Lp (a)
sterol regulatory element binding protein (SREBP), and is levels. There was a modest 15% reduction in serum triglyc
increased in states of cellular cholesterol deficiency. The erides with negligible effect on HDL‐cholesterol [22, 23].
New Agents for Treatment of Dyslipidemia 213
L
LDL L P PCSK9 Ab
L
L
P
PCSK9
L P L P
LDL-R
HM
L Endosome L P
Lysosome
L L P
FIG 17.1 Role of PCSK9 in hepatic uptake of LDL particles. Low density lipoprotein (LDL) particles bind to LDL receptors (LDL‐R) on
the hepatocyte membrane (HM), and the receptor‐ligand complex is internalized to form an endosome. LDL‐R dissociates from the
complex and recycles back to the surface while the LDL particle undergoes lysosomal degradation. PCSK9 is a secreted protein
which also binds to the LDL‐R, and prevents dissociation of the LDL/LDL‐R complex leading to lysosomal degradation of both the
receptor and LDL particle. Also shown are PCSK9 antibodies (Ab) which can bind to circulating PCSK9 and thereby decrease LDL‐R
degradation.
No effect on highly sensitive C‐reactive protein was noted PCSK9 inhibitors: Cardiovascular outcome
when added on to statin therapy [24]. As expected, the trails
lipid‐lowering effects are much less robust in patients with
Both evolocumab and alirocumab have been evaluated for
homozygous familial hypercholesterolemia, and depend
on the genotype and number of LDL receptor‐null cardiovascular outcomes in large, randomized, double‐
alleles [25]. blind, placebo‐controlled trials, namely FOURIER [28]
Very few consistent adverse effects beyond injection site and ODYSSEY Outcomes Trial [29], respectively.
reactions have been observed in these trials. Specifically, The FOURIER trial was a secondary prevention trial
there has been no evidence of increased incidence of liver which enrolled 27 564 patients with established ASCVD
function abnormalities, myopathy or new onset diabetes. A (myocardial infarction, non‐hemorrhagic stroke or periph
substantial number of patients achieved very low LDL‐ eral vascular disease) who had LDL‐cholesterol ≥ 70 mg/
cholesterol levels (< 25 mg/dL), but did not appear to expe dL or non HDL‐cholesterol ≥ 100 mg/dL on stable, opti
rience higher rates of adverse effects, except for a slight mized lipid‐lowering therapy (moderate – high intensity
increase in the incidence of cataracts [26]. The initial con statin ± ezetimibe). The baseline plasma LDL‐cholesterol
cerns about increased risk for neurocognitive adverse was 91 mg/dL and nearly 70% of the patients were on high
events have been largely allayed by a dedicated neurocog intensity statin therapy. The study was initially planned for
nitive sub‐study involving nearly 2000 patients which did 4 years, but ended after a mean follow‐up of 2.2 years as
not show any effect on cognitive functions [27]. there was a faster than expected accrual of events.
Evolocumab therapy decreased LDL‐cholesterol by about patients with diabetes compared to those without diabe
59%, and the mean LDL‐cholesterol in the treatment group tes [33, 34].
at the end of the study was about 30 mg/dL. There was a In addition, there have been a few phase 3 clinical stud
15% relative risk reduction of the primary composite car ies of PCSK9 inhibitors exclusively in patients with diabe
diovascular end point, with an absolute risk reduction of tes which have also demonstrated safety and
2%, which was projected to be 3.3% if the trial were to be efficacy [35–37]. The ODYSSEY DM‐INSULIN rand
extended to 5 years. There was a significant reduction in omized trial studied the effect of alirocumab in insulin‐
risk for myocardial infarction and stroke, but no difference treated patients with both type 1 and type 2 diabetes, and
in overall mortality or cardiovascular deaths. Except for showed about 50% reduction in LDL‐cholesterol without
injection‐site reactions, there was no difference in the inci raising any concerns about co‐administration of insulin
dence of any adverse events. and alirocumab [35]. The ODYSSEY DM‐DYSLIPIDEMIA
The ODYSSEY Outcomes Trial was also a similar sec trial demonstrated significant reduction in non HDL‐cho
ondary prevention trial in 18 924 patients who had a recent lesterol in patients with type 2 diabetes and mixed hyper
(1–12 months) Acute Coronary Syndrome, and had LDL‐ lipidemia who were on maximally tolerated statin therapy.
cholesterol ≥ 70 mg/dL, non HDL‐cholesterol ≥ 100 mg/dL Compared to “usual care” which included the option to add
or Apo B ≥ 80 mg/dL on high intensity or maximally toler either fenofibrate, omega‐3 fatty acids or niacin, treatment
ated statin therapy. The baseline plasma LDL‐cholesterol with alirocumab led to a mean 38% reduction in non‐HDL
was 87 mg/dL and nearly 90% of the patients were on high cholesterol in this 24‐week study [36]. Similarly evel
intensity statin therapy. The mean follow‐up duration was ocumab has been shown to reduce LDL‐cholesterol by
2.8 years. Alirocumab dose was adjusted during the study about 65% and non HDL‐cholesterol by about 56%, besides
to target an LDL‐cholesterol concentration of 25–50 mg/ other atherogenic lipid markers in patients with type 2 dia
dL. There was a significant 15% relative risk reduction of betes and dyslipidemia on background statin therapy [37].
the primary composite cardiovascular end point, with an In all these studies, there was no change in measures of gly
absolute risk reduction of 1.6%. Reduction in all‐cause cemic control.
mortality was also noted (HR = 0.85, 95% CI = 0.73–0.98), While there are no outcome trials exclusively in patients
but was not considered statistically significant as it was with diabetes, a pre‐specified sub‐group analysis of the
tested after two non‐significant outcomes (CHD death and FOURIER trial showed that evolocumab reduced cardio
CVD death) in the testing hierarchy. Similar to the evo vascular risk equally in patients with and without diabetes,
locumab trial, there was no difference in the incidence of without increasing the risk for new onset diabetes or dete
adverse events except for local injection‐site reactions. rioration in glucose control [38]. Of the 27 564 study
patients, 40% had diabetes at baseline, and the hazard
ratios for the primary end point were 0.83 (95% CI, 0.75–
PCSK9 inhibitor use in patients with
0.93) for patients with diabetes and 0.87 (95% CI, 0.79–
diabetes
0.96) for patients without diabetes. There was no change in
The use of PCSK9 inhibitors has been found to be safe and hemoglobin A1c or fasting plasma glucose levels between
efficacious in patients with diabetes. Sub‐group analysis of the evolocumab and placebo groups over time in patients
multiple phase 3 trials of both evolocumab and alirocumab with diabetes, prediabetes, or normoglycemia. These data
have shown that lipid lowering in patients with diabetes is suggest that evelocumab use is safe and efficacious in
similar to those without diabetes, and was not associated patients with ASCVD and diabetes.
with worsening of diabetes control or other adverse
effects [30]. Similarly, the reduction in LDL‐cholesterol did
PCSK9 inhibitors: Clinical use and future
not differ in patients with pre‐diabetes, impaired fasting
perspectives
glucose or metabolic syndrome compared to normoglyce
mic individuals [31, 32]. Interestingly, the incidence of The available data clearly show the ability of this class of
local injection‐site adverse effects was noted to be less in lipid‐lowering drugs to safely and effectively reduce the
level of LDL‐cholesterol and other atherogenic lipoproteins of such aggressive measures. Similarly the effect of
leading to reduced incidence of major vascular events. PCSK9 inhibitors on new onset diabetes also needs to be
However, this has to be tempered with other factors includ examined in longer term studies, especially since
ing the current high cost of these drugs, and the lack of Mendelian randomization studies do show an association
more long‐term safety studies, before recommending a between PCSK9 variants which cause low cholesterol and
broader use. The current guidelines recommend using increased risk for diabetes [41]. If long‐term safety is estab
them in patients who are likely to see the greatest absolute lished, and cost reduction is achieved, PCSK9 inhibitors do
risk reduction. These would be patients with extremely have the potential to significantly reduce the burden of
high cholesterol, such as those with familial hypercholes ASCVD in both patients with and without diabetes. Future
terolemia, or patients with established ASCVD and sub‐ advances including the use of siRNA to decrease PCSK9
optimally treated dyslipidemia who remain at very high activity, immunization to develop antibodies against native
risk for future ASCVD events. The ODYSSEY ESCAPE PCSK9, and CRISPR‐Cas9 gene editing of PCSK9 may fur
study in patients with heterozygous familial hypercholes ther facilitate this goal. While the siRNA, Inclisiran, has
terolemia showed that alirocumab therapy was effective in already shown promising results in phase III trials with a
substantially decreasing or even eliminating the need for 50% reduction in LDL‐cholesterol when administered
LDL apheresis [39]. Indeed the number of patients requir every 6 months [42], the other options are yet to be tested
ing LDL apheresis for severe hypercholesterolemia has in humans [43, 44].
been steadily declining since the advent of PCSK9 inhibitor
therapy.
Other novel lipid‐lowering therapies
In patients with diabetes, those with concomitant
ASCVD who have non HDL‐cholesterol ≥ 100 mg/dL on Besides ezetimibe and PCSK9 inhibitors, bempedoic acid
maximally tolerated statin therapy should be considered can be used to decrease LDL‐cholesterol further in statin
for PCSK9 inhibitor therapy. It is quite likely that other treated (or intolerant) individuals. It inhibits the enzyme
patients with long duration of diabetes, and other risk fac ATP‐citrate lyase in the cholesterol biosynthetic pathway,
tors, will also benefit from earlier initiation of such therapy, and has been shown to reduce LDL‐cholesterol by about
even for primary prevention. It is well recognized that both 21% [45]. It was recently approved for clinical use by the
the magnitude of hyperlipidemia, and its duration are the US Food and Drug Administration (FDA) for patients with
key determinants of atherogenesis. Initiation of effective, heterozygous familial hypercholesterolemia or those with
safe lipid‐lowering therapy at a younger age has the poten established ASCVD who need additional LDL‐cholesterol
tial to greatly decrease the incidence of ASCVD in the reduction. It is administered orally at a dose of 180 mg
population, but its cost effectiveness (under current pric daily.
ing), feasibility and long‐term safety needs to be deter Recently, there has been growing interest in novel tri
mined. Many patients on PCSK9 inhibitors achieve glyceride lowering therapies which has focused on increas
extremely low LDL‐cholesterol levels. In the FOURIER ing lipoprotein lipase (LPL) mediated clearance of
trial, patients with baseline LDL‐cholesterol < 70 mg/dL triglyceride‐rich‐lipoproteins (TGRL). Figure 17.2 depicts
achieved a mean LDL‐cholesterol of 21 mg/dL which was the metabolism of chylomicrons and very low‐density lipo
associated with a 30% relative risk reduction. Secondary proteins (VLDL), the two principal TGRLs. After initial
analysis of data from this trial also show a monotonic rela hydrolysis by LPL, the remnants of TGRL are either taken
tionship between major ASCVD events and achieved LDL‐ up by the liver or further hydrolyzed by the hepatic lipase
cholesterol concentrations down to < 10 mg/dL without to generate LDL particles. However, the remnant particles
any concomitant increase in adverse effects [40]. This has can also permeate the vascular endothelium, akin to LDL
prompted suggestions for more aggressive treatment particles, and promote inflammation and atherogenesis by
thresholds, even below 20 mg/dL [16]. However, the mean a variety of mechanisms including abnormal endothelial
follow‐up in the FOURIER trial was 2.2 years, and we secretion of inflammatory cytokines and impaired flow‐
clearly need more long‐term studies to establish the safety mediated dilatation [46–48]. This assumes great signifi
B-48
B-100
E TG A-V
E TG A-V
C
C
CE
DL
CE CH
VL YL
O
Apo C-II
Insulin
FFA HSL LPL

FFA
P
LR FFA
HL Apo C-III
ANGPTL3
ANGPTL4
LDL
TG
CE CE
TG
CE
Vas TG
c TG
infla ular CE
mm CE
atio
n
RLP: CMR. VLDLr, lDL
FIG 17.2 Metabolism of triglyceride‐rich lipoproteins (TGRL). The two principal triglyceride‐rich lipoproteins, Chylomicrons (Chylo)
and Very Low Density Lipoproteins (VLDL) are secreted by the intestine and liver respectively. They undergo hydrolysis by the
Lipoprotein lipase (LPL) predominantly expressed in adipose tissue and skeletal muscle releasing Free Fatty Acids (FFA) for these
tissues and multiple remnant lipoproteins (RLP) including Chylomicron Remnants (CR), VLDL remnants (VLDLr), and Intermediate
Density Lipoproteins (IDL). These are either cleared by the liver through LDL receptor related proteins (LRP), or undergo further
hydrolysis by the Hepatic Lipase (HL) leading to generation of Low Density Lipoprotein (LDL) particles. Similar to LDL, RLP can also
be taken up into the vessel wall and promote vascular inflammation and atherogenesis. Also shown in the Figure are positive and
negative influencers of LPL activity, and the role of insulin in suppressing adipose tissue Hormone Sensitive Lipase (HSL). Other
abbreviations: TG, triglyceride; CE, Cholesteryl ester; B‐100, Apo B‐100; B‐48, Apo B‐48; E, Apo E; A‐V, Apo A‐V; C, Apo C.
cance in patients with type 2 diabetes who often have remnant lipoprotein concentrations which are
hypertriglyceridemia due to excess hepatic VLDL synthesis primarily responsible for the increased risk for ASCVD
and secretion, driven by enhanced delivery of free fatty despite normal LDL cholesterol concentrations in patients
acids released by adipose tissue hormone sensitive lipase. with dia betes and hypertriglyceridemia. The novel
Adiposity, insulin resistance and insulin deficiency, which triglyceride‐lowering medications work by decreasing the
are the hallmarks of type 2 diabetes, serve to amplify this activity of proteins that inhibit LPL such as Apo C3 and
process. Excess TGRL generation also results in excess ANGPTL 3/4. Volanesorsen is an antisense oligonucleotide
which inhibits Apo C3 and has been shown in Phase II tri References
als to reduce Apo C3 levels by 40–80% and serum triglyc
1. Hu FB, Stampfer MJ, Solomon CG, Liu S, Willett WC,
erides by 31–71% in a dose dependent effect [49]. It was
Speizer FE et al. The impact of diabetes mellitus on mortality
recently approved for clinical use in Europe for patients
from all causes and coronary heart disease in women: 20
with severe hypertriglyceridemia due to Familial
years of follow‐up. Arch Intern Med. 2001;161(14):
Chylomicronemia Syndrome, but was not approved for 1717–1723.
clinical use by the US Food and Drug Administration due 2. Fox CS, Golden SH, Anderson C, Bray GA, Burke LE, de
to risk of thrombocytopenia. An n‐acetyl galactosamine Boer IH et al. Update on prevention of cardiovascular dis
conjugated version of this drug has been developed and ease in adults with type 2 diabetes mellitus in light of recent
shown to significantly reduce triglyceride levels in healthy evidence: a scientific statement from the American Heart
volunteers [50]. Both a monoclonal antibody (Evinacumab) Association and the American Diabetes Association.
and an anti‐sense oligonucleotide to ANGPTL3 have also Circulation. 2015;132(8):691–718.
been developed, and are awaiting clinical trials. While 3. Yeh RW, Sidney S, Chandra M, Sorel M, Selby JV, Go AS.
these drugs are primarily being studied for their triglycer Population trends in the incidence and outcomes of acute
ide‐lowering effect, they do have the potential to reduce myocardial infarction. N Engl J Med. 2010;362(23):
ASCVD in patients with hypertriglyceridemia and elevated 2155–2165.
postprandial remnant lipoprotein cholesterol levels, as sug 4. Rawshani A, Rawshani A, Franzen S, Eliasson B, Svensson
gested by genetic studies linking Apo C3 and ANGPTL4 AM, Miftaraj M et al. Mortality and cardiovascular disease
variations with serum triglycerides and ASCVD [51, 52]. in type 1 and type 2 diabetes. N Engl J Med. 2017;376(15):
Patients with diabetes and hypertriglyceridemia do have 1407–1418.
elevated remnant lipoproteins, and future studies will help 5. Cholesterol Treatment Trialists C, Kearney PM, Blackwell L,
understand the role of these therapies as an adjunct to LDL Collins R, Keech A, Simes J et al. Efficacy of cholesterol‐low
cholesterol‐lowering therapies to further reduce cardiovas ering therapy in 18,686 people with diabetes in 14 ran
cular morbidity and mortality. domised trials of statins: a meta‐analysis. Lancet.
2008;371(9607):117–125.
6. Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK,
Conclusions Blumenthal RS et al. 2018 AHA/ACC/AACVPR/AAPA/
A B C / AC P M / A DA / AG S / A Ph A / ASP C / N L A / P C NA
Patients with diabetes, both type 1 and type 2, are at high
Guideline on the management of blood cholesterol.
risk for ASCVD. Besides healthy lifestyle changes and good
Circulation. 2018:CIR0000000000000625.
glucose control, addition of statin therapy has been shown
7. Cholesterol Treatment Trialists C, Fulcher J, O’Connell R,
to reduce the risk, and should be initiated in most patients
Voysey M, Emberson J, Blackwell L et al. Efficacy and safety
with diabetes. To further reduce the residual risk after sta
of LDL‐lowering therapy among men and women: meta‐
tin therapy, additional LDL‐cholesterol lowering therapies
analysis of individual data from 174,000 participants in 27
such as ezetimibe, bempedoic acid, and PCSK9 inhibitors
randomised trials. Lancet. 2015;385(9976):1397–1405.
can be considered, especially for patients with established 8. Jacobs MJ, Kleisli T, Pio JR, Malik S, L’Italien GJ, Chen RS
ASCVD. PCSK9 inhibitors have been shown to robustly et al. Prevalence and control of dyslipidemia among persons
reduce LDL‐cholesterol by about 60% with minimal with diabetes in the United States. Diabetes Res Clin Pract.
adverse events. They have also been shown to reduce car 2005;70(3):263–269.
diovascular outcomes in patients with established ASCVD, 9. Hirano T. Pathophysiology of diabetic dyslipidemia. J
both with and without diabetes. Novel triglyceride‐lower Atheroscler Thromb. 2018;25(9):771–782.
ing therapies targeting LPL‐mediated clearance of triglyc 10. Bhatt DL, Steg PG, Miller M, Brinton EA, Jacobson TA,
eride‐rich‐lipoproteins are being developed which have the Ketchum SB et al. Cardiovascular risk reduction with icosa
potential for additional ASCVD risk reduction and are also pent ethyl for hypertriglyceridemia. N Engl J Med.
deserving of further study. 2019;380(1):11–22.
11. Sacks FM, Carey VJ, Fruchart JC. Combination lipid therapy 24. Bohula EA, Giugliano RP, Leiter LA, Verma S, Park JG, Sever
in type 2 diabetes. N Engl J Med. 2010;363(7):692–694; PS et al. Inflammatory and Cholesterol Risk in the FOURIER
author reply 4–5. Trial. Circulation. 2018;138(2):131–140.
12. Cannon CP, Blazing MA, Giugliano RP, McCagg A, White 25. Raal FJ, Hovingh GK, Blom D, Santos RD, Harada‐Shiba M,
JA, Theroux P et al. Ezetimibe added to statin therapy after Bruckert E et al. Long‐term treatment with evolocumab
acute coronary syndromes. N Engl J Med. 2015;372(25): added to conventional drug therapy, with or without apher
2387–2397. esis, in patients with homozygous familial hypercholesterol
13. Abifadel M, Varret M, Rabes JP, Allard D, Ouguerram K, aemia: an interim subset analysis of the open‐label TAUSSIG
Devillers M et al. Mutations in PCSK9 cause autosomal study. Lancet Diabetes Endocrinol. 2017;5(4):280–290.
dominant hypercholesterolemia. Nat Genet. 2003;34(2): 26. Robinson JG, Rosenson RS, Farnier M, Chaudhari U, Sasiela
154–156. WJ, Merlet L et al. Safety of very low low‐density lipoprotein
14. Horton JD, Cohen JC, Hobbs HH. PCSK9: a convertase that cholesterol levels with alirocumab: pooled data from rand
coordinates LDL catabolism. J Lipid Res. 2009;50(Suppl): omized trials. J Am Coll Cardiol. 2017;69(5):471–482.
S172–177. 27. Giugliano RP, Mach F, Zavitz K, Kurtz C, Im K, Kanevsky E
15. Tavori H, Rashid S, Fazio S. On the function and homeosta et al. Cognitive function in a randomized trial of evo
sis of PCSK9: reciprocal interaction with LDLR and addi locumab. N Engl J Med. 2017;377(7):633–643.
tional lipid effects. Atherosclerosis. 2015;238(2):264–270. 28. Sabatine MS, Giugliano RP, Keech AC, Honarpour N,
16. Sabatine MS. PCSK9 inhibitors: clinical evidence and imple Wiviott SD, Murphy SA et al. Evolocumab and clinical out
mentation. Nat Rev Cardiol. 2019;16(3):155–165. comes in patients with cardiovascular disease. N Engl J Med.
17. Maxwell KN, Breslow JL. Adenoviral‐mediated expression 2017;376(18):1713–1722.
of Pcsk9 in mice results in a low‐density lipoprotein receptor 29. Schwartz GG, Steg PG, Szarek M, Bhatt DL, Bittner VA, Diaz R
knockout phenotype. Proc Natl Acad Sci USA. 2004;101(18): et al. Alirocumab and cardiovascular outcomes after
7100–7105. acute coronary syndrome. N Engl J Med. 2018;379(22):
18. Rashid S, Curtis DE, Garuti R, Anderson NN, Bashmakov Y, 2097–2107.
Ho YK et al. Decreased plasma cholesterol and hypersensi 30. Handelsman Y, Lepor NE. PCSK9 Inhibitors in lipid man
tivity to statins in mice lacking Pcsk9. Proc Natl Acad Sci agement of patients with diabetes mellitus and high cardio
USA. 2005;102(15):5374–5379. vascular risk: a review. J Am Heart Assoc. 2018;7(13):e008953.
19. Cohen J, Pertsemlidis A, Kotowski IK, Graham R, Garcia 31. Leiter LA, Muller‐Wieland D, Baccara‐Dinet MT, Letierce
CK, Hobbs HH. Low LDL cholesterol in individuals of A, Samuel R, Cariou B. Efficacy and safety of alirocumab in
African descent resulting from frequent nonsense mutations people with prediabetes vs those with normoglycaemia at
in PCSK9. Nat Genet. 2005;37(2):161–165. baseline: a pooled analysis of 10 phase III ODYSSEY clinical
20. Cohen JC, Boerwinkle E, Mosley TH, Jr., Hobbs HH. trials. Diabet Med. 2018;35(1):121–130.
Sequence variations in PCSK9, low LDL, and protection 32. Blom DJ, Koren MJ, Roth E, Monsalvo ML, Djedjos CS,
against coronary heart disease. N Engl J Med. 2006;354(12): Nelson P et al. Evaluation of the efficacy, safety and glycae
1264–1272. mic effects of evolocumab (AMG 145) in hypercholesterol
21. Kathiresan S, Myocardial Infarction Genetics C. A PCSK9 aemic patients stratified by glycaemic status and metabolic
missense variant associated with a reduced risk of early‐ syndrome. Diabetes Obes Metab. 2017;19(1):98–107.
onset myocardial infarction. N Engl J Med. 2008;358(21): 33. Sattar N, Preiss D, Robinson JG, Djedjos CS, Elliott M,
2299–2300. Somaratne R et al. Lipid‐lowering efficacy of the PCSK9 inhi
22. Sabatine MS, Giugliano RP, Wiviott SD, Raal FJ, Blom DJ, bitor evolocumab (AMG 145) in patients with type 2 diabetes:
Robinson J et al. Efficacy and safety of evolocumab in reducing a meta‐analysis of individual patient data. Lancet Diabetes
lipids and cardiovascular events. N Engl J Med. 2015;372(16): Endocrinol. 2016;4(5):403–410.
1500–1509. 34. Leiter LA, Zamorano JL, Bujas‐Bobanovic M, Louie MJ,
23. Robinson JG, Farnier M, Krempf M, Bergeron J, Luc G, Lecorps G, Cannon CP et al. Lipid‐lowering efficacy and
Averna M et al. Efficacy and safety of alirocumab in reduc safety of alirocumab in patients with or without diabetes: A
ing lipids and cardiovascular events. N Engl J Med. sub‐analysis of ODYSSEY COMBO II. Diabetes Obes Metab.
2015;372(16):1489–1499. 2017;19(7):989–996.
35. Leiter LA, Cariou B, Muller‐Wieland D, Colhoun HM, Del 43. Landlinger C, Pouwer MG, Juno C, van der Hoorn JWA,
Prato S, Tinahones FJ et al. Efficacy and safety of alirocumab Pieterman EJ, Jukema JW et al. The AT04A vaccine against
in insulin‐treated individuals with type 1 or type 2 diabetes proprotein convertase subtilisin/kexin type 9 reduces total
and high cardiovascular risk: The ODYSSEY DM‐INSULIN cholesterol, vascular inflammation, and atherosclerosis in
randomized trial. Diabetes Obes Metab. 2017;19(12): APOE*3Leiden.CETP mice. Eur Heart J. 2017;38(32):
1781–1792. 2499–2507.
36. Ray KK, Leiter LA, Muller‐Wieland D, Cariou B, Colhoun 44. Ding Q, Strong A, Patel KM, Ng SL, Gosis BS,
HM, Henry RR et al. Alirocumab vs usual lipid‐lowering Regan SN et al. Permanent alteration of PCSK9 with in vivo;
care as add‐on to statin therapy in individuals with type 2 CRISPR‐Cas9 genome editing. Circ Res. 2014;115(5):
diabetes and mixed dyslipidaemia: The ODYSSEY DM‐ 488–492.
DYSLIPIDEMIA randomized trial. Diabetes Obes Metab. 2018; 45. Laufs U, Banach M, Mancini GBJ, Gaudet D, Bloedon LT,
20(6):1479–1489. Sterling LR et al. Efficacy and safety of bempedoic acid in
37. Lorenzatti AJ, Eliaschewitz FG, Chen Y, Lu J, Baass A, patients with hypercholesterolemia and statin intolerance.
Monsalvo ML et al. Randomised study of evolocumab in J Am Heart Assoc. 2019;8(7):e011662.
patients with type 2 diabetes and dyslipidaemia on back 46. Doi H, Kugiyama K, Oka H, Sugiyama S, Ogata N, Koide SI
ground statin: Primary results of the BERSON clinical trial. et al. Remnant lipoproteins induce proatherothrombogenic
Diabetes Obes Metab. 2019;21(6):1455–1463. molecules in endothelial cells through a redox‐sensitive
38. Sabatine MS, Leiter LA, Wiviott SD, Giugliano RP, mechanism. Circulation. 2000;102(6):670–676.
Deedwania P, De Ferrari GM et al. Cardiovascular safety and 47. Chait A, Ginsberg HN, Vaisar T, Heinecke JW, Goldberg IJ,
efficacy of the PCSK9 inhibitor evolocumab in patients with Bornfeldt KE. Remnants of the triglyceride‐rich lipoproteins,
and without diabetes and the effect of evolocumab on gly diabetes, and cardiovascular disease. Diabetes. 2020;69(4):
caemia and risk of new‐onset diabetes: a prespecified analy 508–516.
sis of the FOURIER randomised controlled trial. Lancet 48. Zheng XY, Liu L. Remnant‐like lipoprotein particles impair
Diabetes Endocrinol. 2017;5(12):941–950. endothelial function: direct and indirect effects on nitric
39. Moriarty PM, Parhofer KG, Babirak SP, Cornier MA, Duell oxide synthase. J Lipid Res. 2007;48(8):1673–1680.
PB, Hohenstein B et al. Alirocumab in patients with hete 49. Gaudet D, Alexander VJ, Baker BF, Brisson D, Tremblay K,
rozygous familial hypercholesterolaemia undergoing lipo Singleton W et al. Antisense inhibition of apolipoprotein c‐
protein apheresis: the ODYSSEY ESCAPE trial. Eur Heart J. iii in patients with hypertriglyceridemia. N Engl J Med.
2016;37(48):3588–3595. 2015;373(5):438–447.
40. Giugliano RP, Pedersen TR, Park JG, De Ferrari GM, Gaciong 50. Alexander VJ, Xia S, Hurh E, Hughes SG, O’Dea L, Geary RS
ZA, Ceska R et al. Clinical efficacy and safety of achieving very et al. N‐acetyl galactosamine‐conjugated antisense drug to
low LDL‐cholesterol concentrations with the PCSK9 inhibi APOC3 mRNA, triglycerides and atherogenic lipoprotein
tor evolocumab: a prespecified secondary analysis of the levels. Eur Heart J. 2019;40(33):2785–2796.
FOURIER trial. Lancet. 2017;390(10106):1962–1971. 51. The TG and HDL Working Group of the Exome Sequencing
41. Ference BA, Robinson JG, Brook RD, Catapano AL, Project, National Heart, Lung, and Blood Institute, Crosby J,
Chapman MJ, Neff DR et al. Variation in PCSK9 and Peloso GM, Auer PL et al. Loss‐of‐function mutations in
HMGCR and risk of cardiovascular disease and diabetes. APOC3, triglycerides, and coronary disease. N Engl J Med.
N Engl J Med. 2016;375(22):2144–2153. 2014;371(1):22–31.
42. Ray KK, Wright RS, Kallend D, Koenig W, Leiter LA, Raal FJ 52. Dewey FE, Gromada J, Shuldiner AR. Variants in ANGPTL4
et al. Two phase 3 trials of inclisiran in patients with elevated and the risk of coronary artery disease. N Engl J Med.
LDL cholesterol. N Engl J Med. 2020;382(16):1507–1519. 2016;375(23):2305–2306.
18 The Role of Bariatric Surgery in Obese
Patients with Diabetes: Primary or Rescue
Therapy?
Praveena Gandikota1 and Blandine Laferrère2
1
Endocrine Fellow, Endocrine, Diabetes and Nutrition Division, Department of Medicine, St Luke’s Roosevelt Hospital,
New York, NY, USA
2
Assistant Professor of Medicine, Division of Endocrinology, Diabetes and Nutrition Obesity Research Center
Department of Medicine, St Luke’s Roosevelt Hospital Center, Columbia University College of Physicians and Surgeons,
New York, NY, USA
c onsiderable public health burden. Approximately half of

patients with diabetes in the United States are obese.
●● Obesity and type 2 diabetes mellitus (T2DM) are major
Medical therapy for patients with diabetes combines costly
public health issues that are closely related. Lifestyle and often inefficient lifestyle modifications (LSM), with
modifications (LSM) with diet and exercise leading to various medications, some of them inducing unwanted
weight loss are cornerstones of T2DM management weight gain and/or hypoglycemia. However, in the last few
along with pharmacologic therapy. Limitations are cost, years, management and prognosis of T2DM have changed
effectiveness, and short-term effect for LSM and
compliance, cost, and side effects for medications.
considerably. Novel pharmacotherapies have emerged,
●● Newer medications may help preserve beta-cell both for the management of diabetes as well as for the pre-
function over time. vention and treatment of its cardiometabolic complica-
●● Bariatric surgery is increasingly becoming popular for tions. Glycemic remission of diabetes can be achieved after
management of morbid obesity and its related bariatric surgery, although this is affected by many factors –
comorbidities including T2DM. Diabetes control
improves for most patients after bariatric surgery, with
indeed, the likelihood of remission can be predicted on the
full remission in over 50% of patients. basis of several characteristics such as age, duration of dia-
●● Although bariatric surgery is approved for diabetic betes, medications used to control diabetes etc. [1]. What is
patients with BMI ≥ 35 kg/m2, the use of surgery to increasingly clear is that bariatric surgery does not alter
treat diabetes in less obese individuals is currently beta-cell mass but reverses glucose toxicity (in the short-
proposed. Limitations to this approach are lack of
high-quality data on short and long term complications
term) and improves insulin action – the response to oral
of the surgery; and long-term data on diabetes and intravenous secretagogues does not change [2, 3]. In
remission and cost analysis. view of the spectacular effect of bariatric surgery on T2DM,
it is understandable that bariatric surgery be proposed as
treatment of T2DM, independently of the beneficial weight
loss effect, i.e., in patients with body mass index (BMI)
Introduction
< 35 kg/m2. This could be a particularly attractive alterna-
The prevalence of Type 2 diabetes mellitus (T2DM) and tive in some ethnic groups, like Asians, who have a high
obesity is increasing worldwide and represents a risk of T2DM at lower BMI compared to Caucasians.

220
The Role of Bariatric Surgery in Obese Patients with Diabetes: Primary or Rescue Therapy? 221
The objective of this chapter is to discuss whether bari- oral glucose compared to intravenous glucose. The incre-
atric surgery should be offered as first-line therapy, or as tins play a key role in postprandial glucose-mediated insu-
rescue therapy, when medical options have failed to control lin secretion. In addition, GLP- 1 decreases glucagon
diabetes. In addition, hypotheses on mechanisms of action induces weight loss and, in animal models, decreased
of bariatric surgery on glucose homeostasis, the role of apoptosis and increased beta-cell mass. The incretin effect
bariatric surgery in treatment of patients with T2DM with is blunted in T2DM. Both incretin levels and effect increase
BMI < 35 kg/m2, risks and benefits of different types of after GBP [8, 9], but not after diet [10] or purely restrictive
bariatric surgery, and future techniques will be discussed. procedures [11]. Levels of GLP- 1 increase as early as
2 days [12] and persist up to 20 years [13]. The mechanisms
of increased incretin levels after GBP are not fully under-
Possible Mechanisms by Which Bariatric
stood. The rapid delivery of nutrients to the lower intestine
Surgery Improves T2DM
after bypass surgeries or after vertical sleeve gastrectomy
Effect of Weight Loss (VSG) [14] may increase the production of GLP-1 by direct
T2DM is a very complex disorder with insulin resistance in nutrient exposure of the L cells. This hypothesis has been
the muscle and the liver, impaired beta-and alpha-cell fueled by results of ileal transposition in rodents [15]. The
function, impaired incretin effect, and alteration of glucose foregut exclusion hypothesis emphasizes the role of exclu-
homeostasis in the kidney and the brain. A significant sion of the proximal part of the gut from nutrients [16],
degree of beta-cell failure and vascular complications are suggesting the presence of a would-be anti-incretin factor
often present at the time of diagnosis of T2DM. Reduced secreted by the proximal small intestine. This hypothesis
calorie intake and increased physical activity are essential was the basis for trials with endoluminal sleeves and duo-
for the management of T2DM. denal mucosal resurfacing (see below).
Multiple studies have shown the effectiveness of LSM in Other possible mechanisms of successful weight loss and
preventing T2DM, with a risk reduction up to 58%. metabolic improvement after bariatric surgery include
Similarly, weight loss and LSM in patients with established changes in ghrelin and PYY [11, 17], bile acids [18], changes
T2DM are beneficial [4]. A modest 10% weight loss is asso- in taste patterns, inflammatory markers, gastric emptying
ciated with significant improvement in diabetes and other and intestinal transit time [9], and possibly in gut flora [19].
obesity-related comorbidities such as hypertension and
dyslipidemia. Contrary to LSM that result in weight loss of
Role of Bariatric Surgery in T2DM:
small magnitude and short duration [5] bariatric surgery
First-line Therapy or Rescue Therapy
results in weight loss of greater magnitude, 40–50% excess
When Medical Options Fail?
weight loss (EWL), sustained over time up to 14 years [6,
7]. Weight loss is a key factor in the improvement of T2DM Pharmacological agents used to treat T2DM [20] are not
after bariatric surgery. without risks and side effects. Weight gain, and/or the risk
of hypoglycemia, especially with sulfonylureas and insulin,
Effect of Incretins are a major hindrance in the management of the already
Resolution of T2DM after bypass procedures has been obese patients. Additionally, challenges occur in patients
described within days of the surgery, suggesting that fac- with associated conditions like renal failure or heart failure.
tors other than weight loss could be responsible for the dia- However, in spite of the various excellent pharmacological
betes improvement. The change in incretins may explain agents available in the United States, 43% of patients with
part of diabetes remission after GBP. The two main incre- T2DM fail medical therapy and do not achieve the HbA1C
tins are glucagon- like peptide 1 (GLP1) and glucose- target of < 7% [21]. In a study by Holman et al, only 31.9–
dependent insulinotropic peptide (GIP), produced by the L 44.7% of patients achieved an HbA1c < 6.5% at 3 years with
and K cells, respectively. Together, they are responsible for different insulin regimen [22]. None of the medical clinical
the “incretin effect” or the greater insulin response after trials report diabetes remission.
222 Surgery for Weight Management
Nearly 30% of patients who undergo bariatric surgery better fasting plasma glucose and HbA1c post surgery
have T2DM [6], although about a third of these patients are compared with insulin-requiring patients with more severe
undiagnosed [23]. In the Buchwald meta-analysis [24], cri- disease. Additionally, the longer the duration of diabetes,
teria used to define diabetes resolution was fasting blood the more likely patients were to remain on medications
glucose < 100 mg/dl and/or HbA1c < 6%, off diabetes med- and/or insulin. However, the preoperative duration of
ications. The percentage of diabetes remission varies T2DM did not result in significant differences in postop-
according to surgery: 95% after BPD/duodenal switch, 80% erative glucose control among the groups [29]. Another
after RYGBPP, 80% with gastroplasty and 57–73% after issue in interpreting published data is the definition used
gastric banding (GB). The greater the malabsorption for diabetes remission. As diabetes remission is indeed a
(BPD) and weight loss, the more likely the patient will novel concept [30], no uniform definition was used in pre-
achieve diabetes remission. Significant improvement in vious surgical studies [24, 25]. The recent consensus state-
T2DM has also been shown in the Swedish Obesity Study ment proposed to define complete remission of T2DM as
(SOS), a long-term observational prospective study that “normal glycemic measures (HbA1c in the normal range,
compared surgical versus nonsurgical groups [7]. The sur- fasting glucose < 100 mg/dl) for at least 1 year duration in
gical group fared better at both 2 years’ and 10 years’ fol- the absence of active pharmacologic therapy or ongoing
low-up with respect to decrease in glucose and insulin procedures” [30]. Whether this clinical definition of diabe-
levels, incidence as well as recovery from T2DM. In a ran- tes remission is accompanied by reversal of the pathophysi-
domized controlled trial (RCT), Dixon et al. showed that ological defects seen with diabetes remains to be
GB resulted in a 73% diabetes remission rate at two years demonstrated.
compared to conventional medical interventions According to the current guidelines, bariatric surgery is
(13%) [25]. The benefit of bariatric surgery include not approved for patients with BMI ≥ 40 kg/m2 or BMI
only a ∼40% weight loss, but the resolution and/or signifi- ≥ 35 kg/m2 in the presence of any comorbidity. Less than
cant improvement of diabetes and most obesity-related 3% of the patients who qualify for bariatric surgery
comorbidities like hyperlipidemia (70%), hypertension undergo this treatment of their obesity and over 200,000
(61.7%), and of obstructive sleep apnea (85.7%) [24]. surgeries are performed yearly in the United States. As
Recent studies have shown a decreased mortality [26], par- morbidly obese patients gain significant health and quality
ticularly the mortality related to diabetes, after bariatric of life benefits from bariatric surgery, an effort to increase
surgery. the number of surgeries on patients with higher BMI (>
Would patients failing medical treatment, who tend to 45 kg/m2) is a reasonable approach [31]. According to the
be older, with longer duration of diabetes, with less optimal latest ADA 2010 guidelines [20], bariatric surgery should
access to care, respond to bariatric surgery? Which bariat- be considered for adults with BMI ≥ 35 kg/m2 and T2DM,
ric surgery is more likely to help patients achieve their especially if diabetes or other comorbidities are difficult to
blood glucose target and/or go into remission? Should control with lifestyle and pharmacologic therapy. What
obese patients with BMI ≥ 35 kg/m2 with T2DM be offered defines “difficult to control” and for how long a patient has
bariatric surgery as a first-line therapy? It is difficult to to be failing medical therapy prior to being offered bariat-
apply the current data to individual patients in terms of ric surgery is unclear. Identifying preoperative predictors
balancing risks versus benefits. Available surgical studies of diabetes resolution is critical not only for patient selec-
on diabetes remission do not always report data on diabe- tion but also for determining which type of surgery would
tes control and/or duration [27]. Diabetes duration, when be the best. The amount of weight loss is clearly a major
provided, is short [25, 28] and few patients, 0.5% [25] to determinant. This is particularly true after GB when dia-
39% [29], are insulin- treated. The study by Schauer betes remission is directly proportional to the degree of
et al. [28] is the only one that provides outcome data ana- weight loss [25]. The duration of diabetes and/or preop-
lyzed according to preoperative diabetes status and dura- erative insulin use decreases the chances of diabetes remis-
tion. Patients with less severe disease had significantly sion [28, 29]. This implies that the less the d
eterioration of
beta-cell function at the time of surgery, the higher the nificant, benefits for RYGB at 3 years, at least in terms of
chances of diabetes remission. However, even if patients glycemic control [34].
do not go into remission after bariatric surgery, most of Morbid obesity is associated with an increased mortal-
them will experience significant improvement of their dis- ity. The mortality rate associated with bariatric surgery has
ease and achieve glucose control on less medications [28]. been reported to be 0.1–1.1% [24]. In a recent prospective,
There are no data from RCT comparing aggressive medi- multicenter observational study [31], the 30-day death rate
cal treatment with bariatric surgery in morbidly obese was 0% after laparoscopic GB, 0.2% after laparoscopic
persons. Even with the best medical scenario, however, RYGBP and 2.1% after open RYGBP. These rates are com-
results of bariatric surgery will be likely hard to match, at parable to the 90-day mortality rate of 1% seen after chol-
least in the first 5–10 years after surgery. ecystectomy. Both prospective (e.g., SOS) [26] and
retrospective studies [35] have shown decreased overall
Gastric Banding or Gastric Bypass? mortality after bariatric surgery. Specifically, cause-specific
There are three main categories of bariatric surgery. mortality in the RYGBP surgery group decreased by 92%
Restrictive procedures, with or without gastrectomy, aim for diabetes [35]. Bariatric surgery is associated with early
to reduce gastric volume in order to limit food intake and postoperative complications such as thromboembolism,
induce weight loss. These procedures include laparo- wound complications like infection/dehiscence, pulmo-
scopic adjustable gastric banding (GB), vertical banded nary complications like atelectasis, and late postoperative
gastroplasty (VBG), rarely performed now, and vertical complications like anastomotic stricture, intestinal
sleeve gastrectomy (VSG). The Roux-en-Y gastric bypass obstruction, incisional hernias, and gastrogastric fistula. In
(RYGBP) surgery combines gastric restriction with some addition, dumping syndrome can be disabling. Long-term
malabsorption and is the most commonly performed metabolic and nutritional complications related to altered
procedure in the United States. Finally, biliopancreatic micronutrient and vitamin absorption are frequently
diversion (BPD) is a procedure with significant malab- observed after malabsorptive surgeries [32] and need to be
sorption with (Figure 18.1) or without gastric restriction. aggressively treated, often for a lifetime. Vitamin D defi-
Although surgeons empirically tend to perform malab- ciency may require weekly doses of ergocalciferol 50,000
sorptive procedures in patients with BMI > 50 kg/m2, units weekly or a few times a week and iron and B12 often
rather than GB, there are no data and/or guidelines deter- are needed parenterally. Deficiencies need to be carefully
mining the best surgery for a particular patient. The monitored at least twice yearly, for a lifetime. At present,
choice of the type of surgery should not be based on sur- there are no established guidelines for mineral and vitamin
geons’ preference and/or on early 6–12 months weight supplementation after bariatric surgery.
loss outcome, but rather on short-and long-term data Hyperinsulinemic hypoglycemia associated is a rela-
generated by RCT comparing various surgical proce- tively uncommon complication of RYGB and has variable
dures. Collecting long- term data is essential to learn severity. Dietary management with or without glucosidase
about the effect of bariatric surgery on weight loss, reso- inhibitors, octreotide, and other pharmacological treat-
lution of comorbidities, as well as complications such as ment should always be tried first prior to performing par-
nutritional deficiencies. Data about nutritional deficien- tial pancreatectomy [36].
cies, weight regain, comorbidities resolution, and overall Bariatric surgery is the treatment of choice for morbid
morbidity are essentially poorly known, as loss to follow- obesity, with the additional benefit of diabetes improve-
up is notably high for this patient population. In a recent ment and/or remission in over 80% of cases. As all morbid
review [32] only five studies were identified with data on obese individuals will greatly benefit from bariatric sur-
nutritional deficiencies with at least 12 month follow-up gery, long duration, and/or poor control of diabetes should
and the data were suboptimal. not be contraindications to the surgery. Long-term, high-
Available data suggests reasonably comparable out- quality data (> 10 years) on weight loss, safety, and nutri-
comes of VSG and RYGB at 1 year [33] and slight, but sig- tional complications are essential. These long-term studies
(b)
(a) Bypassed
segment
Adjustable
band
Port
(c) New (d)

stomach
pouch
(Gastric sleeve)
Part of
stomach Stomach
removed removed
Bypassed
segment
FIG 18.1 Bariatric surgeries. (A) Laparoscopic adjustable gastric banding (GB); (B) Roux-en-Y gastric bypass (RYGBP);
(C) biliopancreatic diversion (BPD) with duodenal switch; (D) vertical sleeve gastrectomy (VSG). www.mayoclinic.org/
bariatric-surgery/bariatric-procedures.html. Figure used by permission of Mayo Foundation for Medical Education
and Research. All rights reserved.
will help establish predictors of success, delineate the ric care will grow exponentially. Primary care physicians
choice of surgery, and help implement guidelines for vita- and/or endocrinologists need to be trained appropriately to
min supplementation. As the number of bariatric surgeries optimize patient management both preoperatively and
increases, the number of patients in need of lifetime bariat- postoperatively.
Role of Weight Loss Surgery in the Treatment New Techniques that can be an Option
of Patients with T2DM and BMI < 35 kg/m2 in the Future
There are currently few studies of bariatric surgery on Endoluminal duodenal sleeve either in lean diabetic
patients with T2DM and BMI < 35 kg/m2. In the RCT by rats [17] or in rats with diet-induced obesity [41] improves
Dixon et al. [25], patients with well-controlled T2DM, of glucose metabolism and induces weight loss. The human
< 2 years’ duration, mean BMI of 37.1 kg/m2 went into trials of endoluminal treatments of obesity are limited to
diabetes remission in 73% of cases after GB at 2 years (the restrictive interventions such as intragastric balloons, tran-
dropout rate was low) versus 13% in the diet group [25]. soral gastroplasty, and endoluminal vertical gastroplasty.
The studies from De Paula et al. with complicated surger- Trials with duodenojejunal bypass sleeves have been disap-
ies, including ileal transposition, report high remission pointing both in terms of their efficacy and also because of
rate, but a non-negligible mortality, at least in the initial associated complications such as bolus obstruction and
report [37]. A study reports remission of diabetes and hepatic abscess associated with their use [42].
improvement of insulin sensitivity 18 months after BPD Duodenal mucosal resurfacing employs thermal energy
in five patients with BMI < 35 kg/m2 [38]. A recent non- to ablate the duodenal mucosa – theoretically to alter hor-
randomized study by Shah et al. in Asian Indians reported monal and mucosal signaling. A multicenter study with 46
100% remission rate in 15 patients 3 months after GBP participants reported effects on glycemic control despite
surgery. Preoperatively, 80% of the patients were insulin modest weight changes. These effects were sustained at
treated, diabetes control was poor with HbA1C ∼10%, 12 months. Further mechanistic studies are required to
and diabetes duration was 9 years [39]. The RCT by better understand the mechanism(s) of action and long-
Dixon et al. provides the best evidence-based data to sug- term benefit, if any, on glycemic control [43].
gest the use of GB in patients with diabetes and lower In conclusion, bariatric surgery should be the treatment
BMI [25]. The diabetes remission was directly related to of choice for patients with morbid obesity complicated by
the amount of weight loss and the preoperative HbA1C T2DM and with high metabolic and cardiovascular risks.
levels. Participants had fairly mild diabetes. Whether The surgery has been proven to not only improve quality of
these results can be achieved in a more severe diabetic life but also prolong life in these high-risk patients. There is
population, different ethnic groups, and with a different not enough evidence to perform invasive bariatric surger-
health care system, is unknown. There is currently not ies and/or to experiment with new surgical procedures in
enough evidence to suggest that bariatric surgery, par- patients with lower BMI and/or with milder diabetes, out-
ticularly bypass, should be indicated for patients with side of well-controlled RCT. Optimizing chronic care after
BMI < 35 kg/m2. As for more invasive surgeries (bypass ± GB, a noninvasive procedure, to ensure proper weight loss
ileal transposition), it is unclear whether the risk they and diabetes remission is important. Multidisciplinary
represent is worth their use in clinical practice outside of teams including surgeons, endocrinologists, gastroenter-
very well controlled research RCT. ologists, psychiatrists, nutritionists, and primary care are
Indeed, the diabetes surgery study randomized 120 needed to follow up these patients. Bariatric surgery offers
adult participants, with BMI 30.0-39.9 kg/m2 and HbA1c a unique model to understand the complexity of T2DM
>/=8.0% to either RYGB + intensive medical management and help develop new treatments for this chronic disease.
of diabetes, hypertension, and dyslipidemia or to medical
management alone. None of the subjects managed medi-
cally experienced diabetes remission at 36 months while References
17% of gastric bypass patients had full remission and 19% 1. Still CD, Wood GC, Benotti P et al. Preoperative prediction
had partial remission. However, RYGB was associated with of type 2 diabetes remission after Roux-en-Y gastric bypass
more significant adverse events. In addition, the effect on surgery: a retrospective cohort study. Lancet 2(1): 38–45,
glycemic control decreased with time [40]. 2014.
2. Nguyen KT, Billington CJ, Vella A et al. Preserved insulin 14. Vidal J, Ibarzabal A, Nicolau J et al. Short-term effects of
secretory capacity and weight loss are the predominant pre- sleeve gastrectomy on type 2 diabetes mellitus in severely
dictors of glycemic control in patients with type 2 diabetes obese subjects. Obes Surg. 17(8):1069–1074, 2007.
randomized to Roux-en-Y gastric bypass. Diabetes 64(9): 15. Strader AD, Vahl TP, Jandacek RJ et al. Weight loss through
3104–3110, 2015. ileal transposition is accompanied by increased ileal hor-
3. Dirksen C, Bojsen- Møller KN, Jørgensen NB et al. mone secretion and synthesis in rats. Am J Physiol Endocrinol
Exaggerated release and preserved insulinotropic action of Metab. 288(2):E447–E453, 2005.
glucagon-like peptide-1 underlie insulin hypersecretion in 16. Rubino F, Forgione A, Cummings DE et al. The mechanism
glucose-tolerant individuals after Roux-en-Y gastric bypass. of diabetes control after gastrointestinal bypass surgery
Diabetologia 56(12): 2679–2687, 2013. reveals a role of the proximal small intestine in the patho-
4. Pi-Sunyer X, Blackburn G, Brancati FL et al. Reduction in physiology of type 2 diabetes. Ann Surg. 244(5):741–749,
weight and cardiovascular disease risk factors in individuals 2006.
with type 2 diabetes: one-year results of the look AHEAD 17. Olivan B, Teixeira J, Bose M et al. Effect of weight loss by diet
trial. Diabetes Care. 30(6):1374–1383, 2007. or gastric bypass surgery on peptide YY3-36 levels. Ann
5. Wadden TA, Sternberg JA, Letizia KA et al. Treatment of Surg. 249(6):948–953, 2009.
obesity by very low calorie diet, behavior therapy, and their 18. Patti ME, Houten SM, Bianco AC et al. Serum bile acids are
combination: a five-year perspective. Int J Obes. 13(Suppl higher in humans with prior gastric bypass: potential contri-
2):39–46, 1989. bution to improved glucose and lipid metabolism. Obesity
6. Pories WJ, Swanson MS, MacDonald KG et al. Who would (Silver Spring). 17(9):1671–1677, 2009.
have thought it? An operation proves to be the most effective 19. Zhang H, DiBaise J, Zuccolo A et al. Human gut microbiota
therapy for adult- onset diabetes mellitus. Ann Surg. in obesity and after gastric bypass. PNAS. 106(7):2364–2370,
222(3):339–350, 1995. 2009.
7. Sjostrom L, Lindroos AK, Peltonen M et al. Lifestyle, diabe- 20. American Diabetes Association. Diabetes Care. 33(Suppl 1):
tes, and cardiovascular risk factors 10 years after bariatric S1–S2, 2010.
surgery. N Engl J Med. 351(26):2683–2693, 2004. 21. Cheung BM, Ong KL, Cherny SS et al. Diabetes prevalence
8.. Laferrère B, Heshka S, Wang K et al. Incretin levels and effect and therapeutic target achievement in the United States,
are markedly enhanced 1 month after Roux-en-Y gastric 1999 to 2006. Am J Med. 122(5):443–453, 2009.
bypass surgery in obese patients with type 2 diabetes. 22. Holman RR, Farmer AJ, Davies MJ et al. Three-year efficacy
Diabetes Care. 30(7):1709–1716, 2007. of complex insulin regimens in type 2 diabetes. NEngl J Med.
9. Morinigo R, Moize V, Musri M et al. Glucagon-like peptidel, 361(18):1736–1747, 2009.
peptide YY, hunger, and satiety after gastric bypass surgery 23. Residori L, Garcia-Lorda P, Flancbaum L et al. Prevalence of
in morbidly obese subjects. J Clin Endocrinol Metab. co-morbidities in obese patients before bariatric surgery:
91(5):1735–1740, 2006. effect of race. Obes Surg. 13(3):333–340, 2003.
10. Laferrère B, Teixeira J, McGinty J et al. Effect of weight loss 24. Buchwald H, Avidor Y, Braunwald E et al. Bariatric surgery:
by gastric bypass surgery versus hypocaloric diet on glucose a systematic review and meta-analysis. JAMA. 292(14):1724–
and incretin levels in patients with type 2 diabetes. J Clin 1737, 2004.
Endocrinol Metab. 93(7):2479–2485, 2008. 25. Dixon JB, O’Brien PE, Playfair J et al. Adjustable gastric
11. Bose M, Machineni S, Olivan B et al. Superior appetite hor- banding and conventional therapy for type 2 diabetes: a ran-
mone profile after equivalent weight loss by gastric bypass domized controlled trial. JAMA. 299(3):316–323, 2008.
compared to gastric banding. Obesity (Silver Spring). 26. Sjostrom L, Narbro K, Sjostrom CD et al. Effects of bariatric
18(6):1085–1091, 2010. surgery on mortality in Swedish obese subjects. NEngl J Med.
12. le Roux CW, Welbourn R, Werling M et al. Gut hormones as 357(8):741–752, 2007.
mediators of appetite and weight loss after Roux-en-Y gas- 27. Vetter ML, Cardillo S, Rickels MR, and Iqbal N. Narrative
tric bypass. Ann Surg. 246(5):780–785, 2007. review: effect of bariatric surgery on type 2 diabetes mellitus.
13. Naslund E, Gryback P, Hellstrom PM et al. Gastrointestinal Ann Intern Med. 150(2):94–103, 2009.
hormones and gastric emptying 20 years after jejunoileal 28. Schauer PR, Burguera B, Ikramuddin S et al. Effect of lapa-
bypass for massive obesity. Int J Obes Relat Metab Disord. roscopic Roux-en Y gastric bypass on type 2 diabetes melli-
21(5):387–392, 1997. tus. Ann Surg. 238(4):467–784, 2003.
29. Sugerman HJ, Wolfe LG, Sica DA, and Clore JN. Diabetes sleeve gastrectomy is an effective operation for the treatment
and hypertension in severe obesity and effects of gastric of type 2 diabetes mellitus patients with BMI 21-29. Surg
bypass-induced weight loss. Ann Surg. 237(6):751–756, Endosc. 23(6):1313–1320, 2009.
2003. 38. Chiellini C, Rubino F, Castagneto M et al. The effect of bilio-
30. Buse JB, Caprio S, Cefalu WT et al. How do we define cure pancreatic diversion on type 2 diabetes in patients with BMI
of diabetes? Diabetes Care. 32(11):2133–2135, 2009. <35 kg/m2. Diabetologia. 52(6):1027–1030, 2009.
31. Purnell JQ and Flum DR. Bariatric surgery and diabetes: 39. Shah SS, Todkar JS, Shah PS, and Cummings DE. Diabetes
who should be offered the option of remission? JAMA. remission and reduced cardiovascular risk after gastric
301(15):1593–1595, 2009. bypass in Asian Indians with body mass index <35 kg/m2.
32. Shah M, Simha V, and Garg A. Review: long-term impact of Surg Obes Relat Dis. 6(4):332–338, 2010.
bariatric surgery on body weight, comorbidities, and nutri- 40. Ikramuddin S, Korner J, Lee W-J et al. Durability of addition
tional status. J Clin Endocrinol Metab. 91(11):4223–4231, of Roux-en-Y gastric bypass to lifestyle intervention and
2006. medical management in achieving primary treatment goals
33. Schauer PR, Kashyap SR, Wolski K et al. Bariatric surgery for uncontrolled type 2 diabetes in mild-to-moderate obe-
versus intensive medical therapy in obese patients with dia- sity: a randomized control trial. Diabetes Care 39(9): 1510–
betes. N Engl J Med 366(17): 1567–1576, 2012. 1518, 2016.
34. Schauer PR, Bhatt BL, Kirwan JP et al. Bariatric surgery ver- 41. Aguirre V, Stylopoulos N, Grinbaum R, and Kaplan LM. An
sus intensive medical therapy for diabetes – 3-year out- endoluminal sleeve induces substantial weight loss and nor-
comes. N Engl J Med 370(21): 2002–2013, 2014. malizes glucose homeostasis in rats with diet-induced obe-
35. Adams TD, Gress RE, Smith SC et al. Long-term mortality sity. Obesity (Silver Spring). 16(12):2585–2592, 2008.
after gastric bypass surgery. N Engl J Med. 357(8):753–761, 42. Egan AM and Vella A. Endoscopic treatments for obesity:
2007. the good, the bad, and the ugly. Endocrinol Metab Clin North
36. Salehi, M, Vella A, McLaughlin T et al. Hypoglycemia after Am 49(2): 315–328, 2020.
gastric bypass surgery: current concepts and controversies. J 43. van Baar ACG et al. Endoscopic duodenal mucosal resurfac-
Clin Endocrinol Metab 103(8): 2815–2826, 2018. ing for the treatment of type 2 diabetes mellitus: one year
37. Depaula AL, Macedo AL, Mota BR, and Schraibman V. results from the first international, open-label, prospective,
Laparoscopic ileal interposition associated to a diverted multicentre study. Gut 69(2): 295–303, 2020.
19 Treatment Strategies in Patients
with Diabetes Mellitus and Ischemic Heart
Disease
Madeline K. Mahowald1, Chiam Leker Locker2, Mandeep Singh3 and
Robert L. Frye4
1
Fellow, Cardiovascular Disease, Mayo Clinic
2
Assistant Professor, Cardiovascular Surgery, Mayo Clinic
3
Professor, Cardiovascular Disease, Mayo Clinic
4
Professor, Cardiovascular Disease, Mayo Clinic
Originally considered to be a disease of disordered lipid

metabolism and storage, the role of inflammation in coro
●● Medical therapy is an important part of the revasculari-
nary artery disease (CAD) has become clearer in recent dec
zation strategy in any patient with diabetes and ades [6, 7]. Intracellular hyperglycemia has direct effects on
ischemic heart disease. the vasculature by decreasing endothelial‐dependent vaso
●● Revascularization therapy may be indicated for dilation [8] and also indirect effects via triggering high levels
symptom control as well as improving prognosis and
of reactive oxygen species. Oxidative stress, in turn, stimu
preventing subsequent myocardial infarction.
●● In the absence of contraindications, coronary artery
lates changes in lipid metabolism, altered mitochondrial
bypass grafting seems to be a superior choice to function, and the increased production of inflammatory
percutaneous therapy in people with diabetes. compounds [9]. These reactions affect platelet function,
fibrinolysis, and coagulation cascades. Once present, athero
sclerosis further impairs endothelial function.
As the prevalence of type 2 DM has grown, the contem
porary patient with coronary artery disease (CAD) and
Introduction
DM has other inflammatory conditions, especially obesity,
The synergistic relationship between the altered metabo that contribute to the hostile metabolic milieu. On a cellu
lism and underlying inflammatory state in patients with lar level, adipose cells express inflammatory compounds,
diabetes mellitus (DM) accounts for the high coincidence such as TNF‐alpha, that affect insulin resistance and cellu
of cardiovascular disease in this population [1, 2], which lar uptake of glucose [10]. Specific therapies targeting
remains the leading cause of death among patients with inflammation’s role in atherosclerosis and acute coronary
DM in the United States [3]. Understanding the appropri syndromes (ACS) have shown mixed results and are the
ate management of both diseases is imperative to improv focus of much current research [6, 7, 11, 12].
ing clinical outcomes, especially as the global prevalence of The pattern of atherosclerosis that develops in this met
diabetes and related conditions rises [4, 5]. abolic environment often differs from that in patients with

228
Revascularization Strategies in Patients with Diabetes Mellitus and Ischemic Heart Disease 229
out DM. The volume and extent of atheroma is greater in tion [22]. The precise classes of agents included in OMT
patients with DM, and vascular remodeling is impaired, may change in the near future with the widespread use of
such that diseased vessels are diffusely and extensively PCSK9 inhibitors, SGLT2 inhibitors, and glucagon‐like
involved, revealing narrow lumens on coronary angiogra peptide‐1 agonists, all of which show promise in patients
phy [13]. This pattern of involvement has significant impli with DM and cardiovascular disease [23, 24].
cations for revascularization, as patients with DM may not
have optimal distal vessel targets for bypass graft anasto Antiplatelet therapy
mosis and may not have clear distal “landing zones” for The role, components, and duration of antiplatelet therapy
stents. Additionally, endothelial dysfunction leads to a after revascularization are evolving and situation‐ and
diminished ability to form collateral circulation [14]. The patient‐specific [25, 26] due to research that quantifies the
Bypass Angioplasty Revascularization Investigation benefits of aspirin on preventing cardiovascular death, MI,
(BARI) trial showed that patients with DM had a larger stroke, or TIA at the cost of more major bleeding [27, 28].
amount of jeopardized myocardium during acute MI than At present, the use of daily low‐dose aspirin for primary
patients without DM [15]. Perhaps because of these differ prevention of cardiovascular disease in patients with dia
ent disease patterns, patients with DM have worse out betes is not routinely recommended [29].
comes after myocardial infarction (MI) and in stable CAD For patients with known CAD, indefinite low‐dose daily
with or without revascularization [16–19]. aspirin is recommended regardless of DM status unless con
To guide clinicians on the various treatment options traindicated. The addition of a P2Y12 inhibitor for 12 months
available for diabetic patients with CAD, this chapter will following an ACS is recommended, but the risks of bleeding
focus on: generally outweigh the benefits if dual antiplatelet therapy
(DAPT) is extended beyond that time period.
1 The importance of aggressive medical management
Recommendations after elective revascularization in
regardless of revascularization strategy selected
stable ischemic heart disease are more nuanced and are
2 The indications for coronary revascularization
generally dictated by procedural and patient details,
3 Factors influencing the selection of PCI or CABG if
including susceptibility to bleeding as well as ischemia.
revascularization is pursued.
After PCI, the duration of DAPT can be as brief as 1 month
in patients with high bleeding risk [30]. There is no dedi
Role of medical therapy and glucose
cated randomized trial to guide the duration of DAPT after
control
elective CABG in the absence of recent ACS or PCI, but it
Optimal medical therapy (OMT) is imperative to clinical may be considered for 12 months following surgery to
outcomes for all patients with CAD, perhaps in patients improve graft patency [31]. Procedural techniques are
with DM more than those without. This additional benefit important considerations, and a stronger argument for
may be due to the way in which medications target specific DAPT may be made after off‐pump CABG [32].
pathways affected by metabolic derangements and oxida
tive stress, such as glycoprotein IIb/IIIa inhibitors, which Management of hyperglycemia
reduce the impact of hyperglycemia on promoting platelet After patients with DM are diagnosed with obstructive
aggregation [20]. CAD, abundant evidence points to improved outcomes in
Briefly, medical therapy is generally considered to patients with tight glycemic control. In patients admitted
include antiplatelet agents, statins, beta‐blockers, and angi with ACS or for elective surgical revascularization, the
otensin‐converting enzyme inhibitors or angiotensin degrees of hyperglycemia and glycated hemoglobin eleva
receptor blockers. OMT is under‐prescribed among tion at the time of presentation are predictive of long‐term
patients with DM and CAD [21], despite being indepen mortality [33, 34]. Intensive glucose control after acute
dently associated with improved survival and bestowing a myocardial infarction (MI) can mitigate the poorer out
greater treatment benefit than the method of revasculariza comes for patients with DM.
Several studies have investigated the effects of different TABLE 19.1 Class I indications for myocardial revascularization
strategies for treatment of hyperglycemia in the setting of in stable ischemic heart disease per ACC/AHA and ESC
guidelines.
myocardial revascularization. The BARI 2D trial found
that therapy enhancing insulin sensitization was preferable To improve symptoms
over therapy based on insulin provision for patients with • For patients with ≥ 1 significant coronary artery stenoses
type 2 DM [35]. A subgroup analysis of the FREEDOM amenable to revascularization and unacceptable angina
trial also showed worse outcomes in patients who were despite medical therapy
treated with insulin than those who were treated with other To improve survival
hypoglycemic medications [36]. Even with second‐genera • For survivors of sudden cardiac death with presumed
tion drug‐eluting stents (DES), patients treated with insu ischemia‐mediated ventricular tachycardia caused by
significant stenosis in a major coronary artery (ACC/AHA only)
lin have more target lesion failure [37]. However, in many
• For patients with significant (≥ 50% lumen diameter) left
of these trials (with the notable exception of BARI 2D), main coronary artery stenosis
patients were not randomized to insulin provision versus • For patients with significant stenoses in 3 major coronary
insulin sensitization; patients on insulin therapy may arteries (with or without involvement of the proximal LAD
therefore represent more advanced diabetics with a pro artery) or in the proximal LAD artery plus 1 other major
coronary artery
pensity for worse outcomes. These trials will be discussed
• For patients with multivessel disease and impaired LV
in more detail later in this chapter. function, a large area of ischemia, or a single remaining
patent coronary artery with a significant stenosis (ESC only)
Coronary revascularization Unique recommendations for diabetic patients

• A Heart Team approach to revascularization is
Guidelines from The American College of Cardiology/ recommended in patients with diabetes mellitus and
American Heart Association (ACC/AHA) in 2014 [38] complex multivessel CAD
(and expanded with appropriate use criteria in 2017 [39]) • CABG is generally recommended in preference to PCI to
are essential references. The 2014 document includes the improve survival in patients with diabetes mellitus and
multivessel CAD for which revascularization is likely to
following Class 1 recommendations to improve survival in
improve survival (3‐vessel CAD or complex 2‐vessel CAD
patients with multivessel CAD and DM: (1) a Heart Team involving the proximal LAD), particularly if an ITA graft can be
approach should be utilized in decision‐making regarding anastomosed to the LAD artery and provided the patient is a
revascularization in patients with DM and CAD, and (2) surgical candidate
CABG is generally recommended in preference to PCI in
those for whom revascularization is likely to improve sur
vival (namely, those with the most extensive disease), par ACC/AHA indications for revascularization are summa
ticularly if an internal thoracic artery (ITA) can be rized in Table 19.1 [38, 41].
anastomosed to the left anterior descending (LAD) artery The specific method of revascularization must be tai
and the patient is a good surgical candidate. However, PCI lored to the individual patient and situation. In some
has a role in patients with focal disease who warrant revas patients, the severity, extent, and precise anatomic location
cularization for symptom relief. Additionally, the of obstructive atherosclerotic lesions will dictate the prefer
European Society of Cardiology (ESC) released overlap able technique. In other cases, relative equipoise allows for
ping guidelines in 2014 and, in conjunction with the consideration of additional factors, including the presence
European Association for the Study of Diabetes, recently of symptoms, documented ischemia, myocardial dysfunc
collaborated to release task force guidelines specific to car tion, as well as the patient’s clinical characteristics, comor
diovascular disease in patients with diabetes and pre‐dia bidities, and preferences.
betes. These documents provide guidance on making the Here, we will discuss the three distinct situations in which
diagnosis of DM, risk factor modification, medical man revascularization is indicated and the evidence relevant to
agement, and revascularization [40]. Specific ESC and each one. Regardless of the presence of DM, the two basic
Algorithm to Guide Diagnosis and Management
Clinical evaluation of
patient with DM and
suspected CAD
Acute coronary
syndrome or severe Mild or no ischemic
symptoms of ischemia symptoms
Risk stratification with

Invasive coronary stress test (exercise
angiography preferred), high-sensitivity
Troponin-T, NT-proBNP
Left main disease Large area of ischemia

Focal disease
Severe multivessel disease or significantly Low-risk stress test,
without No obstructive CAD
LV dysfunction elevated biomarkers normal biomarkers
high risk features
SYNTAX score >22
PCI Consider CT coronary

CABG with arterial Aggressive risk Consider CT
or angiogram for young
grafts factor reduction coronary calcium
CABG patients
FIG 19.1 Algorithm to guide diagnosis and management.
reasons to consider elective invasive revascularization are ratio, or intravascular ultrasound, should be routinely
(1) control of symptoms from myocardial ischemia despite employed to assess lesions of borderline significance.
optimal medical therapy, and (2) to reduce mortality and Documentation of ischemia makes symptom improvement
myocardial infarction in high‐risk patients. Many patients more likely after intervention.
may benefit for both reasons. The third situation, ACS, is PCI and CABG both have roles in symptom control.
discussed separately. Figure 19.1 provides an algorithm for PCI is often the first choice for patients with DM and sin
managing patients with DM and CAD. gle‐vessel disease if the lesion is appropriate for stenting.
CABG appears to have a more durable effect, but major
Symptom control for stable patients despite surgery is often delayed until there is multivessel involve
medical therapy ment. For patients with left main or three‐vessel disease,
Randomized trials have demonstrated benefits in symptoms CABG has remained the standard of care [22].
and in quality of life with elective revascularization. The
increasing use of chest computed tomography has increased Prolonging life and preventing subsequent
lead time for the diagnosis of CAD, but determining whether MI in stable patients
nonspecific symptoms are indeed related to anatomic find Since the early days of CABG, trials comparing surgical
ings can be difficult. This challenge is common in patients revascularization with OMT alone demonstrated that the
with DM, who often present with atypical symptoms, such as greatest benefit was derived by patients with the most severe
dyspnea. Functional or physiologic testing, such as determi disease and those with decreased left ventricular systolic
nation of fractional flow reserve, instantaneous wave‐free function. Several landmark trials are reviewed below.
Trials for patients with diabetes and cardiovascular of death from any cause, nonfatal MI, and nonfatal stroke.
disease The CABG arm was superior and significantly reduced
BARI 2D was designed to test treatment strategies in stable, rates of death and MI, although at a higher risk of stroke.
mildly symptomatic or asymptomatic patients with type 2 Of note, the benefits of CABG in this study were driven by
DM and angiographically confirmed CAD to determine if reductions in all‐cause mortality and MI and extended to
prompt, elective revascularization would reduce mortality all patients regardless of categorization by disease com
or subsequent MI compared to OMT. A second layer of plexity [43]. In analyses of patients treated with insulin
randomization compared therapy with insulin‐sensitizing compared with oral agents, CABG was superior to PCI in
agents to an insulin‐provision strategy to achieve target both groups [36]. Subsequent publication of the FREEDOM
glycemic control. Randomization was stratified according Follow‐On study confirmed that the mortality benefits of
to optimal revascularization strategy determined by each CABG over PCI with DES persisted over a median of 7.5
patient’s treating physicians. Neither treatment strategy years [44].
demonstrated differences in mortality, MI, or stroke. CARDia included 510 symptomatic patients with a
However, among the patients with the most extensive ath median follow‐up of 365 days. At the trial’s conclusion,
erosclerosis, there was a significant reduction in the com there was no significant difference in the rate of composite
bined end point of death, MI, and stroke in patients that outcome (death, MI, stroke) but a significant higher rate of
underwent prompt CABG. This trial is notable among repeat revascularization in the PCI arm [45]. The more
studies of revascularization for demonstrating the preven recent BEST trial was stopped early due to slow enrollment,
tion of subsequent nonfatal MI (whereas many trials rely but was on track to support these results [17].
on composite endpoints including repeat revasculariza Common findings among these trials that may help
tion). Updated BARI 2D analysis from 2019 reported that explain the difference in outcomes is the achievement of
the combination of insulin‐sensitization therapy and more complete revascularization in the surgical patients
CABG offered the lowest risk of cardiovascular death, compared with those undergoing PCI; more grafts per
ACS, stroke, transient ischemic attack, repeat coronary patient were utilized than stents per patient. Interestingly,
revascularization, and lower extremity revascularization or the benefit of CABG over PCI has not extended to left main
amputation [42]. In addition to lower levels of hyperglyce disease [18, 46, 47].
mia, the patients in the sensitization group had lower levels
of the inflammatory markers C‐reactive protein and Subgroup analysis of patients with diabetes
fibrinogen. Like the BARI trial, the Synergy between Percutaneous
Since BARI 2D, three additional trials have looked at Coronary Intervention with Taxus and Cardiac Surgery
survival in patients with DM and multivessel disease rand (SYNTAX) trial evaluated intervention for stable ischemic
omized to CABG or PCI that warrant mention. By far, the heart disease. Patients were randomized to PCI or CABG;
largest was the Future Revascularization Evaluation in the primary end point was a composite of death from any
Patients with Diabetes Mellitus: Optimal Management of cause, stroke, MI, or repeat revascularization. Publication
Multivessel Disease trial (FREEDOM); the two smaller of results at 12 months after randomization showed that
studies are the Coronary Artery Revascularization in CABG was superior to PCI; these results were primarily
Diabetics trial (CARDia) and the Randomized Comparison driven by repeat revascularization occurring in patients
of Coronary Artery Bypass Surgery and Everolimus‐ with the highest SYNTAX scores (representing more
Eluting Stent Implantation in the Treatment of Patients extensive and complex disease) who underwent PCI [48].
with Multivessel Coronary Artery Disease (BEST) trial. At study conclusion, there was no significant difference
The FREEDOM trial included 1900 patients with DM between groups except for the subgroup of patients with
and multivessel disease at 140 international centers who treated DM, who benefited from CABG [49].
were randomized to CABG vs PCI and followed for a Most recently, results of the International Study of
median of 3.8 years. The primary outcome was a composite Comparative Health Effectiveness with Medical and
Invasive Approaches (ISCHEMIA) trial showed no benefit STEMI. However, women with DM have not seen the same
of elective revascularization over OMT on all‐cause or car mortality benefit as men and remain undertreated with the
diovascular mortality in a sample of 5179 patients over a cornerstones of medical therapy in ACS [54].
median of 3.2 years [50]. ISCHEMIA trial data for the sub Revascularization within 24 hours is recommended for
group of patients with DM have not been released at the patients with NSTEMI and UA and reduces death and MI
time of publication, but similar earlier trials have not con in all comers. In fact, according to at least one trial, the
sistently shown benefit [19]. benefits of early invasive therapy in patients with UA were
One must proceed cautiously in drawing conclusions more pronounced in patients with DM than in those with
from subgroup analyses. However, one large meta‐analysis out [55]. Emergent PCI is often the preferred strategy for
pooling individual patient‐level data from 11 trials pub patients with NSTEMI complicated by hemodynamic
lished from 2001 to 2016 demonstrated a survival benefit instability, malignant arrhythmia, or refractory ischemic
with CABG in patients with diabetes and multivessel dis chest pain.
ease [46]. PCI has not demonstrated survival benefit over
medical therapy. Coronary artery bypass grafting
For some clinicians, the FREEDOM trial sufficiently con
Biomarkers for risk stratification cluded debate between revascularization strategies among
Identifying patients with low symptom burden but high‐ patients with DM as CABG demonstrated clear benefits in
risk anatomy remains a clinical challenge in daily practice. subsequent cardiac events and survival regardless of dis
Measurement of biomarkers has been proposed as a ease extent and complexity. Additional studies from the
method of risk stratification in patients at high risk for surgical literature suggest that details of the operative strat
CAD but whose coronary anatomy is unknown. For exam egy may have significant consequences. Perhaps most
ple, routine measurement of high‐sensitivity troponin‐T importantly, the use of multiple arterial grafts may confer
levels was recorded among the 2285 patients with DM2 improved long‐term survival, lower risk of adverse cardiac
and stable ischemic heart disease who enrolled in the BARI events, and a higher rate of patency without a difference in
2D trial [51]. Among the 39% of patients who had abnor perioperative mortality compared to the conventional
mal troponin‐T concentrations at baseline, the hazard ratio CABG strategy with the use of the left internal thoracic
for 5‐year event rate of cardiovascular death, MI, or stroke artery (ITA) to the LAD and vein grafts to the non‐LAD
was 1.85 compared to those with normal baseline tro vessels [56–58]. It has been shown that graft patency was
ponin‐T. However, immediate revascularization in this not inferior in diabetics over 20 years follow‐up and the
high risk population showed no benefit compared to OMT. patency rate of BITA remained greater than 90%, outper
Elevated levels of plasma N‐terminal pro‐brain natriuretic forming the SVG patency rate of 40%, and those rates were
peptide (NT‐proBNP) have also been associated with car similar both in diabetics and non‐diabetics [59, 60].
diovascular mortality [52], but how to incorporate this into In a large clinical trial randomizing patients to a surgical
clinical practice is less clear. strategy of single ITA or bilateral internal thoracic artery
(BITA) grafting, there were non‐significant decreases in
Acute coronary syndromes death and in a composite outcome of death, MI or stroke at
ACS include ST‐elevation MI (STEMI), non‐ST‐elevation 10 years favoring the BITA group. However, results of this
MI (NSTEMI), and unstable angina (UA). Emergent PCI is study were confounded by a high crossover rate and a
the recommended revascularization strategy for all patients higher than expected use of the radial artery. In the “as
with STEMI as there is clear benefit in restoring flow as treated cohort,” multiarterial grafting showed significant
soon as possible. Importantly, patients with DM often have advantage compared with single ITA group [61]. Although
delayed presentations for ACS compared with patients the use of BITA has been controversial in patients with DM
without DM [53]. In‐hospital mortality has improved dra due to historically higher rates of deep sternal wound
matically in recent decades for patients with DM and infections, this technique has demonstrated better survival,
less frequent need for revascularization, and longer free 3 A newly perfused distal bed receptive to the vascular
dom from major adverse cardiovascular events in patients growth factors secreted by ITAs and SVGs that promote
with and without diabetes [62, 63]. formation of collateral circulation [76].
The increased incidence of deep sternal wound infec
tion seen in patients with DM (especially women) can be
minimized – if not altogether eliminated – by harvesting Conclusions
ITA in a skeletonized rather than pedicled manner. The Despite the ongoing challenges for patients with DM and
importance of the skeletonized technique over pedicle CAD, evidence suggests that current medical and revascu
grafts is explicitly stated in clinical practice guidelines for larization strategies have significantly improved outcomes
arterial graft harvest [64]. For patients with DM, skeletoni [77]. Patients with non‐severe CAD and mild symptoms
zation of BITA lowers the risk of sternal infection to that of may be safely managed conservatively for many years.
pedicled single graft regardless of treatment with insulin However, regardless of the presence of symptoms, patients
[63, 65–67]. with DM and severe coronary artery disease can derive sig
nificant benefit regarding symptoms, quality of life, mor
Percutaneous coronary intervention
tality, and freedom from revascularization by undergoing
DES are currently the standard of care for both diabetic
early CABG. Patients with less extensive disease but
and non‐diabetic patients; metal struts are typically coated
severe symptoms may benefit from PCI, but there is no
with anti‐proliferative mTOR inhibitors. Although there is
clear evidence that these patients have improved survival
interest in stents comprised of absorbable polymers, out
or freedom from MI or revascularization. Aggressive
comes have not demonstrated significant benefits and their
medical management and risk factor modification is
use is not widespread. Despite advances in stent technol
imperative to improving the outcomes of patients with
ogy and anti‐thrombotic medications, target lesion failure
DM and any degree of CAD, and insulin sensitization
remains more common in patients with DM due to in‐
therapy may confer additional benefits as one compo
stent restenosis and stent thrombosis [68, 69]. Subsequent
nent of OMT.
generations of DES seek to limit the restenosis caused by
intimal hyperplasia, which is particularly exuberant in
patients with diabetes [70], and it remains an area of active References
research.
1. Libby P, Theroux P. Pathophysiology of coronary artery
Why is coronary artery bypass grafting more disease. Circulation. 2005;111(25):3481–3488.
effective? 2. Reardon CA et al. Obesity and insulin resistance promote
atherosclerosis through an IFNγ‐regulated macrophage
The benefits of surgical over percutaneous revasculariza
protein network. Cell Reports. 2018;23(10):3021–3030.
tion are likely multifactorial and secondary to both struc
3. Rodriguez F et al. Diabetes‐attributable mortality in the
tural and biochemical effects on the subtended myocardium
United States from 2003 to 2016 using a multiple‐cause‐of‐
[71]. Proposed mechanisms include the following: death approach. Diabetes Research and Clinical Practice.
1 Reperfusion of the grafted coronary axis with additional 2019;148:169–178.
protection against the progression of native vessel dis 4. Bhupathiraju SN, Hu FB. Epidemiology of obesity and
diabetes and their cardiovascular complications. Circulation
ease [72, 73];
Research. 2016;118(11):1723–1735.
2 Restoration of endothelial function of the grafted and
5. Cho N et al. IDF Diabetes Atlas: Global estimates of diabetes
downstream vessels due to the production of nitric oxide prevalence for 2017 and projections for 2045. Diabetes
of arterial segments [74]. This improvement in regional Research and Clinical Practice. 2018;138:271–281.
endothelial function is in contrast to the changes 6. Tardif J‐C. et al. Efficacy and safety of low‐dose colchicine
induced by stenting, which induced both acute and after myocardial infarction. New Engl J Med.
chronic inflammatory changes [75]; and 2019;381(26):2497–2505.
7. Ridker PM et al. Rosuvastatin to prevent vascular events in 21. Farkouh ME et al. Risk factor control for coronary artery
men and women with elevated C‐reactive protein. New Engl disease secondary prevention in large randomized trials.
J Med. 2008;359(21):2195–2207. Journal of the American College of Cardiology. 2013;61(15):
8. Beckman JA et al. Inhibition of protein kinase Cβ prevents 1607–1615.
impaired endothelium‐dependent vasodilation caused by 22. Iqbal J et al. Optimal medical therapy improves clinical out
hyperglycemia in humans. Circulation Research. comes in patients undergoing revascularization with percu
2002;90(1):107–111. taneous coronary intervention or coronary artery bypass
9. Shah MS, Brownlee M. Molecular and cellular mechanisms grafting: insights from the Synergy Between Percutaneous
of cardiovascular disorders in diabetes. Circulation Research. Coronary Intervention with TAXUS and Cardiac Surgery
2016;118(11):1808–1829. (SYNTAX) trial at the 5‐year follow‐up. Circulation.
10. Hotamisligil GS, Shargill NS, B.M. Spiegelman, Adipose 2015;131(14):1269–1277.
expression of tumor necrosis factor‐alpha: direct role in obe 23. Husain M et al. Oral semaglutide and cardiovascular out
sity‐linked insulin resistance. Science. 1993;259(5091):87–91. comes in patients with type 2 diabetes. N Engl J Med.
11. Ridker PM et al. Antiinflammatory therapy with canaki 2019;381(9):841–851.
numab for atherosclerotic disease. N Engl J Med. 2017;377(12): 24. Perkovic V et al. Canagliflozin and renal outcomes in type 2
1119–1131. diabetes and nephropathy. N Engl J Med. 2019;380(24):
12. Ridker PM et al. Low‐dose methotrexate for the prevention 2295–2306.
of atherosclerotic events. N Engl J Med. 2019;380(8): 25. Valgimigli M et al. 2017 ESC focused update on dual anti
752–762. platelet therapy in coronary artery disease developed in col
13. Nicholls SJ et al. Effect of diabetes on progression of coro laboration with EACTS. European Journal of Cardio‐Thoracic
nary atherosclerosis and arterial remodeling: a pooled anal Surgery. 2018;53(1):34–78.
ysis of 5 intravascular ultrasound trials. Journal of the 26. Levine GN et al. 2016 ACC/AHA guideline focused update
American College of Cardiology. 2008;52(4):255–262. on duration of dual antiplatelet therapy in patients with cor
14. Abaci A et al. Effect of diabetes mellitus on formation of coro onary artery disease: a report of the American College of
nary collateral vessels. Circulation. 1999;99(17):2239–2242. Cardiology/American Heart Association Task Force on
15. Kip KE et al. Differential influence of diabetes mellitus on Clinical Practice Guidelines: an update of the 2011 ACCF/
increased jeopardized myocardium after initial angioplasty AHA/SCAI guideline for percutaneous coronary interven
or bypass surgery: bypass angioplasty revascularization tion, 2011 ACCF/AHA guideline for coronary artery bypass
investigation. Circulation. 2002;105(16):1914–1920. graft surgery, 2012 ACC/AHA/ACP/AATS/PCNA/SCAI/
16. Bangalore S et al. Everolimus‐eluting stents or bypass sur STS guideline for the diagnosis and management of patients
gery for multivessel coronary disease. N Engl J Med. 2015; with stable ischemic heart disease, 2013 ACCF/AHA guide
372(13):1213–1222. line for the management of ST‐elevation myocardial infarc
17. Park S‐J. et al. Trial of everolimus‐eluting stents or bypass tion, 2014 AHA/ACC guideline for the management of
surgery for coronary disease. N Engl J Med. 2015;372(13): patients with non–ST‐elevation acute coronary syndromes,
1204–1212. and 2014 ACC/AHA guideline on perioperative cardiovas
18. Milojevic M et al. Bypass surgery or stenting for left main cular evaluation and management of patients undergoing
coronary artery disease in patients with diabetes. Journal of noncardiac surgery. Circulation. 2016;134(10):e123–e155.
the American College of Cardiology. 2019;73(13):1616–1628. 27. Steg PG et al. Ticagrelor in patients with stable coronary dis
19. Maron DJ et al. Impact of metabolic syndrome and diabetes ease and diabetes. N Engl J Med. 2019;381(14):1309–1320.
on prognosis and outcomes with early percutaneous coro 28. The ASCEND Study Collaborative Group. Effects of aspirin
nary intervention in the COURAGE (Clinical Outcomes for primary prevention in persons with diabetes mellitus. N
Utilizing Revascularization and Aggressive Drug Evaluation) Engl J Med. 2018;379(16):1529–1539.
trial. Journal of the American College of Cardiology. 2011;58(2): 29. Arnett DK et al. 2019 ACC/AHA guideline on the primary
131–137. prevention of cardiovascular disease: a report of the
20. Roffi M et al. Platelet glycoprotein IIb/IIIa inhibitors reduce American College of Cardiology/American Heart Association
mortality in diabetic patients with non‐ST‐segment‐eleva Task Force on Clinical Practice Guidelines. Journal of
tion acute coronary syndromes. Circulation. 2001;104(23): the American College of Cardiology. 2019;74(10):
2767–2771. e177–e232.
30. Windecker S et al. Polymer‐based or polymer‐free stents in Society of Cardiovascular Computed Tomography, and
patients at high bleeding risk. N Engl J Med. 2020;382: Society of Thoracic Surgeons. Journal of the American
1208–1218. College of Cardiology. 2017;69(17):2212–2241.
31. van Diepen S et al. Dual antiplatelet therapy versus aspirin 40. Cosentino F et al. 2019 ESC guidelines on diabetes, pre‐dia
monotherapy in diabetics with multivessel disease undergo betes, and cardiovascular diseases developed in collabora
ing CABG: FREEDOM insights. Journal of the American tion with the EASD: the task force for diabetes, pre‐diabetes,
College of Cardiology. 2017;69(2):119–127. and cardiovascular diseases of the European Society of car
32. Deo SV et al. Dual anti‐platelet therapy after coronary artery diology (ESC) and the European association for the study of
bypass grafting: is there any benefit? A systematic review diabetes (EASD). European Heart Journal. 2020;41(2):
and meta‐analysis. Journal of Cardiac Surgery. 2013;28(2): 255–323.
109–116. 41. Windecker S et al. 2014 ESC/EACTS guidelines on myo
33. Halkos ME et al. Elevated preoperative hemoglobin A1c cardial revascularization. EuroIntervention. 2015;10(9):
level is predictive of adverse events after coronary artery 1024–1094.
bypass surgery. The Journal of Thoracic and Cardiovascular 42. Genuth SM et al. BARI 2D: a reanalysis focusing on cardio
Surgery. 2008;136(3):631–640. vascular events. Mayo Clinic Proceedings. 2019; Elsevier.
34. Robich MP et al. Intensity of glycemic control affects long‐ 43. Farkouh ME et al. Strategies for multivessel revasculariza
term survival after coronary artery bypass graft surgery. The tion in patients with diabetes. N Engl J Med. 2012;367(25):
Annals of Thoracic Surgery. 2019;107(2):477–484. 2375–2384.
35. Group BDS. A randomized trial of therapies for type 2 dia 44. Farkouh ME et al. Long‐term survival following multivessel
betes and coronary artery disease. N Engl J Med. 2009;360(24): revascularization in patients with diabetes: the FREEDOM fol
2503–2515. low‐on study. Journal of the American College of Cardiology.
36. Dangas GD et al. Long‐term outcome of PCI versus CABG 2019;73(6):629–638.
in insulin and non–insulin‐treated diabetic patients: results 45. Kapur A et al. Randomized comparison of percutaneous
from the FREEDOM trial. Journal of the American College of coronary intervention with coronary artery bypass grafting
Cardiology. 2014;64(12):1189–1197. in diabetic patients: 1‐year results of the CARDia (Coronary
37. Pepe M et al. Impact of insulin‐treated and noninsulin‐ Artery Revascularization in Diabetes) trial. Journal of the
treated diabetes mellitus in all‐comer patients undergoing American College of Cardiology. 2010;55(5):432–440.
percutaneous coronary interventions with polymer‐free 46. Head SJ et al. Mortality after coronary artery bypass grafting
biolimus‐eluting stent (from the RUDI‐FREE Registry). The versus percutaneous coronary intervention with stenting for
American Journal of Cardiology. 2019;124(10):1518–1527. coronary artery disease: a pooled analysis of individual
38. Fihn SD et al. 2014 ACC/AHA/AATS/PCNA/SCAI/STS patient data. Lancet. 2018;391(10124):939–948.
focused update of the guideline for the diagnosis and man 47. Giacoppo D et al. Percutaneous coronary intervention vs
agement of patients with stable ischemic heart disease: a coronary artery bypass grafting in patients with left main
report of the American College of Cardiology/American coronary artery stenosis: a systematic review and meta‐anal
Heart Association Task Force on Practice Guidelines, and ysis. JAMA Cardiology. 2017;2(10):1079–1088.
the American Association for Thoracic Surgery, Preventive 48. Serruys PW et al. Percutaneous coronary intervention ver
Cardiovascular Nurses Association, Society for sus coronary‐artery bypass grafting for severe coronary
Cardiovascular Angiography and Interventions, and Society artery disease. N Engl J Med. 2009;360(10):961–972.
of Thoracic Surgeons. Journal of the American College of 49. Kappetein AP et al. Treatment of complex coronary artery
Cardiology. 2014;64(18):1929–1949. disease in patients with diabetes: 5‐year results comparing
39. Patel MR et al. ACC/AATS/AHA/ASE/ASNC/SCAI/SCCT/ outcomes of bypass surgery and percutaneous coronary
STS 2017 appropriate use criteria for coronary revasculari intervention in the SYNTAX trial. European Journal of
zation in patients with stable ischemic heart disease: a report Cardio‐Thoracic Surgery. 2013;43(5):1006–1013.
of the American College of Cardiology appropriate use crite 50. Maron DJ et al. Initial invasive or conservative strategy for
ria task force, American Association for Thoracic Surgery, stable coronary disease. N Engl J Med. 2020;382:1395–1407.
American Heart Association, American Society of 51. Everett BM et al. Troponin and cardiac events in stable
Echocardiography, American Society of Nuclear Cardiology, ischemic heart disease and diabetes. N Engl J Med.
Society for Cardiovascular Angiography and Interventions, 2015;373(7):610–620.
52. Bhalla MA et al. Prognostic role of B‐type natriuretic pep 65. Kieser TM et al. Toward zero: deep sternal wound infection
tide levels in patients with type 2 diabetes mellitus. Journal of after 1001 consecutive coronary artery bypass procedures using
the American College of Cardiology. 2004;44(5):1047–1052. arterial grafts: implications for diabetic patients. The Journal of
53. The BARI Investigators. Influence of diabetes on 5‐year Thoracic and Cardiovascular Surgery. 2014;148(5):1887–1895.
mortality and mobidity in a randomized trial comparing 66. Di Mauro M et al. Bilateral internal mammary artery for
CABG and PTCA in patients with multivessel disease. The multi‐territory myocardial revascularization: long‐term fol
Bypass Angioplasty Revascularization Investigation (BARI). low‐up of pedicled versus skeletonized conduits. European
Circulation. 1997;96:1761–1769. Journal of Cardio‐Thoracic Surgery. 2015;47(4):698–702.
54. Roffi M et al. Gender‐related mortality trends among dia 67. Deo SV et al. Bilateral internal thoracic artery harvest and
betic patients with ST‐segment elevation myocardial infarc deep sternal wound infection in diabetic patients. The
tion: insights from a nationwide registry 1997–2010. Annals of Thoracic Surgery. 2013;95(3):862–869.
European Heart Journal: Acute Cardiovascular Care. 2013;2(4): 68. Kereiakes DJ et al. Outcomes in diabetic and nondiabetic
342–349. patients treated with everolimus‐or paclitaxel‐eluting stents:
55. Invasive compared with non‐invasive treatment in unstable results from the SPIRIT IV clinical trial (Clinical Evaluation
coronary‐artery disease: FRISC II prospective randomised of the XIENCE V Everolimus Eluting Coronary Stent
multicentre study. FRagmin and Fast Revascularisation dur System). Journal of the American College of Cardiology.
ing InStability in Coronary artery disease Investigators. 2010;56(25):2084–2089.
Lancet. 1999;354:708–715. 69. Konigstein M et al. Outcomes among diabetic patients
56. Locker C et al. Multiple arterial grafts improve late survival undergoing percutaneous coronary intervention with con
of patients undergoing coronary artery bypass graft surgery: temporary drug‐eluting stents: analysis from the BIONICS
analysis of 8622 patients with multivessel disease. Circulation. randomized trial. JACC: Cardiovascular Interventions. 2018;
2012;126(9):1023–1030. 11(24):2467–2476.
57. Schwann TA et al. Incremental value of increasing number 70. Kornowski R et al. Increased restenosis in diabetes mellitus
of arterial grafts: the effect of diabetes mellitus. The Annals after coronary interventions is due to exaggerated inti
of Thoracic Surgery. 2018;105(6):1737–1744. mal hyperplasia: a serial intravascular ultrasound study.
58. Gaudino M et al. Radial‐artery or saphenous‐vein grafts in Circulation. 1997;95(6):1366–1369.
coronary‐artery bypass surgery. N Engl J Med. 2018;378(22): 71. Ruel M et al. The SYNTAX score according to diabetic sta
2069–2077. tus: What does it mean for the patient requiring myocardial
59. Raza S et al. Influence of diabetes on long‐term coronary revascularization? The Journal of Thoracic and Cardiovascular
artery bypass graft patency. Journal of the American College Surgery. 2019;159(3):857–860.
of Cardiology. 2017;70(5):515–524. 72. Dimitrova KR et al. Arterial grafts protect the native coro
60. Schwartz L et al. Coronary bypass graft patency in patients nary vessels from atherosclerotic disease progression. The
with diabetes in the Bypass Angioplasty Revascularization Annals of Thoracic Surgery. 2012;94(2):475–481.
Investigation (BARI). Circulation. 2002;106(21):2652–2658. 73. Zhang M et al. Left internal mammary artery versus coro
61. Taggart DP et al. Bilateral versus single internal‐thoracic‐ nary stents: impact on downstream coronary stenoses and
artery grafts at 10 years. N Engl J Med. 2019;380(5): conduit patency. Journal of the American Heart Association.
437–446. 2016;5(9):e003568.
62. Dorman MJ et al. Bilateral internal mammary artery graft 74. Kinoshita T et al. Off‐pump bilateral versus single skele
ing enhances survival in diabetic patients: a 30‐year follow‐ tonized internal thoracic artery grafting in patients with diabe
up of propensity score–matched cohorts. Circulation. tes. The Annals of Thoracic Surgery. 2010;90(4):1173–1179.
2012;126(25):2935–2942. 75. Gomes WJ et al. Coronary artery and myocardial inflamma
63. Zhou P et al. Is the era of bilateral internal thoracic artery tory reaction induced by intracoronary stent. The Annals of
grafting coming for diabetic patients? An updated meta‐ Thoracic Surgery. 2003;76(5):1528–1532.
analysis. 2019. The Journal of Thoracic and Cardiovascular 76. Couffinhal T et al. Vascular endothelial growth factor/vas
Surgery. 158(6):1559–1570. cular permeability factor (VEGF/VPF) in normal and ath
64. Aldea GS et al. The Society of Thoracic Surgeons clinical erosclerotic human arteries. Am J Pathol. 1997;150(5):1673.
practice guidelines on arterial conduits for coronary artery 77. Gregg EW et al. Changes in diabetes‐related complications
bypass grafting. The Annals of Thoracic Surgery. 2016;101(2): in the United States, 1990–2010. N Engl J Med. 2014;370(16):
801–809. 1514–1523.
PART IV
Management of Disease
Complications
20 Diabetes Management in Patients with
Critical and Non-Critical Illness
Kalpana Muthusamy1, Kristin Grdinovac2 and John M. Miles3
1
Division of Endocrinology, Olmsted Medical Center, Rochester, MN, USA
2
Assistant Professor of Medicine, University of Kansas, Kansas City, KS, USA
3
Professor of Medicine, University of Kansas, Kansas City, KS, USA
ill. One of these trials [7] led to a consensus conference

LE A RNI NG P OI NT S
convened in 2004 by the American College of
OO To discuss evidence for and against the use of insulin
Endocrinology and cosponsored by the American
infusion to control hyperglycemia in the critically ill. Diabetes Association, the American Heart Association,
OO To consider the role of hypoglycemia in the disparate the American Society of Anesthesiologists, the Endocrine
results of large clinical trials investigating the use of Society, the Society of Critical Care Medicine, and the
intensive insulin therapy in the critically ill. Society of Thoracic and Cardiovascular Surgeons [8]. The
OO To review the use of nutrition and its potential
influence on insulin requirements and outcomes in the
consensus conference concluded, again on the basis of a
critically ill. single clinical trial [7], that blood glucoses should be
OO To discuss mechanisms by which insulin infusion may maintained < 6.1 mmol/L in critically ill patients, and that
modulate outcomes when used in critically ill patients. insulin infusion was the means to achieve that goal. The
conference also recommended that a preprandial glucose
target < 6.1 mmol/L in patients who are eating on general
medical-surgical wards should be adopted. This latter
Introduction
recommendation, which in our view had and has very lit-
Hyperglycemia is a frequent finding in acute care hospitals, tle evidence to support it, is not the subject of this chapter
with a prevalence rate in excess of one-third of all inpa- and will not be discussed further. Instead, we will address
tients [1, 2]. Among patients with hyperglycemia (admis- mechanisms of hyperglycemia- related morbidity and
sion glucose > 7 mmol/L or random glucose > 11.1 mmol/L), mortality in the critically ill and will undertake a detailed
only two-thirds are known to have diabetes; the remainder review of the evidence supporting the use of insulin infu-
either have newly diagnosed diabetes or temporary sion in the ICU, with emphasis on ten large clinical trials.
(“stress”) hyperglycemia. Hyperglycemia has been identi- We will also discuss the importance of hypoglycemia,
fied as an independent predictor for higher infection nutritional support, and imprecision in blood glucose
rates [3, 4] and increased mortality [5, 6] in critically ill testing in interpreting the results of these trials. We con-
and postoperative patients. A number of large clinical tri- clude that there is strong evidence to support the use of
als have been conducted over the past 15 years to address insulin infusion to control hyperglycemia in the critically
the question of whether intravenous infusion of insulin ill, but very little evidence to indicate what the blood glu-
could improve morbidity and mortality in the critically cose target should be.

241
242 Management of Hyperglycemia in Critical Illness
Mechanisms Mediating Adverse Effects not proving) a role for reduced inflammation in improved
of Hyperglycemia During Critical Illness outcomes [15]. Modulation of nitric oxide synthase expres-
sion, and activity by insulin therapy mitigates the increased
During acute illness, hyperglycemia is exacerbated in peo-
nitric oxide levels seen during stress, thus protecting
ple with diabetes as a consequence of the release of counter-
against endothelial damage [16]. Endothelial adhesion
regulatory hormones and cytokines [9]. The same
molecules such as ICAM-1 and E-selectin, despite being
mechanisms, along with subtle defects in insulin secretion,
comparably elevated at admission are lower during inten-
contribute to stress, or temporary, hyperglycemia in some
sive insulin therapy, indicating reduced endothelial activa-
nondiabetic individuals. The exact mechanisms by which
tion [16]. These factors in combination likely help maintain
hyperglycemia mediates adverse outcomes such as infec-
the integrity of the microvasculature, leading to less organ
tion and increased mortality are not fully understood,
damage and providing survival benefits.
however, nor are the mechanisms responsible for improved
There is also evidence that insulin therapy may favora-
outcomes resulting from correction of hyperglycemia.
bly influence outcomes by improving altered lipid homeo-
Several interesting hypotheses have been proposed involv-
stasis. Plasma-free fatty acids (FFA) are elevated in the
ing both metabolic and non- metabolic pathways [10].
critically ill, in spite of hyperinsulinemia due to resistance
Hyperglycemia can cause direct cellular damage as well as
to insulin’s antilipolytic effects in adipocytes [17]. This is of
indirect effects via increased generation of reactive oxygen
potential importance because elevated FFA contribute to
species. Although the normal protective response of cells
acute endothelial dysfunction [18], have pressor
to hyperglycemia is to down regulate glucose transport,
effects [19], and can contribute to the dyslipidemia of criti-
expression and localization of GLUT-1 and GLUT-3 in sev-
cal illness by driving hepatic VLDL production [20].
eral cell types has been shown to be increased by elevated
Mesotten et al., in a post hoc analysis of a large clinical trial,
levels of cytokines, growth factors, and hypoxia, which
found that intensive insulin therapy in the critically ill pro-
often occur together in the critically ill, potentially leading
duced substantially lower triglyceride levels [21]. There
to cellular glucose overload [11]. Increased substrate avail-
was a four-to fivefold increase in mortality over a range of
ability for glycolysis results in excess superoxide formation
triglyceride levels, although in a multivariate analysis HDL
that can overwhelm the normal detoxification process,
cholesterol had the strongest association with mortality.
leading to mitochondrial dysfunction [11]. Intensive insu-
The mechanism for increased morbidity and mortality
lin therapy has been shown to reverse or prevent hepatic
related to dyslipidemia is not known, but could involve
mitochondrial abnormalities noted with critical ill-
impairment of reticuloendothelial function, as discussed
ness [12]. There is also evidence showing improvement in
elsewhere [22]. Other proposed mechanisms of benefit are
the phagocytic and bactericidal activities of polymorpho-
insulin-induced increase in expression of adiponectin [23]
nuclear leukocytes with improved glycemic control in dia-
and a potential role in enhancing innate immunity [24].
betic subjects, thus enhancing immunity [13]. Glucotoxicity
Collectively, the evidence supports a role for nonglycemic
can also impair β-cell secretory function, potentially initi-
effects of insulin therapy in improving outcomes in critical
ating a vicious cycle of relative insulin insufficiency and
illness, although the relative contribution of these factors is
further increases in glucose concentrations [14].
an important area for future research.
The pleiotrophic effects of insulin, independent of its
glucose-lowering effect, have attracted considerable atten-
Large Intervention Studies
tion as a contributor to clinical benefits. Insulin has been
shown to have anti-inflammatory properties, producing a Background
decrease in concentration of acute phase reactants. As recently as a decade ago, at a time when hyperglycemia
Reduced mortality and organ failure in critically ill patients was recognized as a predictor of adverse outcomes, ele-
treated with insulin infusion is associated with decreased vated blood glucose in the critically ill was often ignored.
high sensitivity C-reactive protein levels, suggesting (but However, since that time the results of large clinical trials
Diabetes Management in Patients with Critical and Non-Critical Illness 243
have forced a reconsideration of this attitude of indiffer- hospitals in the United States, most of which did not have
ence. These results have produced wide swings in the pen- a critical care service, revealed that whereas there was
dulum of expert opinion. Initially, the changes were subtle, a general recognition of the value of insulin infusion
proposed glucose targets were quite liberal and, in fact, in the critically ill, virtually none were strictly adhering
were based more on associations than on harder evidence; to the 6.1 mmol/L benchmark (JMM, unpublished
blood glucose levels of < 12.2 mmol/L were considered by observations).
some to be optimal, as they were associated with improved In March 2009 the results of the NICE-SUGAR trial
postoperative infection rates [25]. were published, demonstrating that excessively tight con-
In 2004, an expert committee declared that a glucose trol (target 6.0 mmol/L, average glucose 6.4 mmol/L) might
target of < 6.1 mmol/L was appropriate for most critically actually be harmful compared to a less stringent target
ill patients [8]. In January, 2009, the ADA stated in a (10 mmol/L, average glucose 8.0 mmol/L) [29]. The odds
position paper that “Critically ill surgical patients’ blood ratio for death in the intensive arm of this study was 1.14 (p
glucose levels should be kept as close to 110 mg/dl = 0.02). In the subset of patients with traumatic brain-
(6.1 mmol/L) as possible. . .” [26]. Four months later, in a injury a later paper concluded that despite a higher inci-
joint statement with the American Association of Clinical dence of hypoglycemia in the intensively treated group,
Endocrinologists, the ADA declared that “a glucose range these patients exhibited no clinically-significant difference
of 140 to 180 mg/dl (7.8 to 10.0 mmol/L) is recommended in outcomes compared to the conventionally treated group.
for the majority of critically ill patients.” [27]. What is the Finfer S et al. Intensive Care Medicine. 41(6):1037-47,
explanation for these wide shifts in prevailing views? 2015 JunThe ADA-AACE consensus statement relaxing the
In 2001, Van den Berghe and colleagues published a guidelines was published less than two months later.
landmark study (hereafter referred to as “Leuven 1”) of Where does this leave us? The results of meta-
more than 1500 patients in a surgical ICU. In this study, analyses [31, 32] conclude that intensive insulin therapy
they aimed for (and achieved) a target blood glucose of does not reduce mortality in critically ill patients and
4.4–6.1 mmol/L in an intensive control group and increases the risk of hypoglycemia, although one
10–11.1 mmol/L in a conventional treatment group [7]. report [32] acknowledges it might benefit surgical patients.
The in-hospital mortality was significantly lower in the Taken at face value, it could be argued that these studies
intensive group at 7.2% compared to 10.9% in the conven- provide a rationale for abandoning intensive treatment
tional arm. They also reported a reduction in mortality with insulin infusion altogether. The fact that the authors
during intensive care from 8% in the conventional group to of the ADA-AACE consensus statement still endorse inten-
4.6% in the intensive group (42% relative reduction). Most sive treatment, albeit in a less aggressive form [26], indi-
of these benefits were observed in the patients who stayed cates that they must see merit in the several studies that
in the intensive care unit (ICU) longer than 5 days. There have reported benefit from this treatment. For this chapter,
was significant improvement in morbidity related to fewer we will review in detail the results of ten large clinical trials
days on mechanical ventilation, shorter ICU stay, reduced that have produced disparate, difficult- to-
reconcile
infection, acute renal failure, critical-illness polyneuropa- results [7, 29, 33–40]. We included all trials of > 500
thy, and fewer blood transfusions. These overwhelmingly patients in which glycemic targets were part of the study
positive results were the sole basis for the target recom- design, but we excluded studies in which glucose infusion
mended in the 2004 consensus statement and obviously was employed in one group but not the other for the pur-
the driving force behind the January, 2009 ADA recom- pose of allowing infusion of insulin at higher rates than
mendation. A survey revealed that the majority of inten- would otherwise be necessary to control hyperglyce-
sivists prefer a target blood glucose of 6.1 mmol/L [28]. mia [41]. Our purpose is to attempt to identify features of
Although there was little published disagreement with the the studies that might explain divergent findings leading to
guidelines, it is not clear that they penetrated to the local conflicting conclusions and recommendations. Specifically,
level; a 2008–2009 informal poll of over 30 medium-sized we hoped to determine whether variables such as acuity of
illness in the study population, hypoglycemia, blood glu- The last five of the studies found no benefit of intensive
cose testing methodology, and nutritional support might insulin therapy, and the NICE- SUGAR study actually
explain differences among studies. reported higher mortality in intensively treated patients [29].
Three of the five studies were smaller and perhaps under-
powered. Furthermore, in two of the five studies the glu-
General Findings (Tables 20.1 and 20.2)
cose target range in the control group was lower than the
Table 20.1 shows the 10 studies in chronological order and Leuven trials. The threshold for initiating insulin infusion
provides details on trial design and study populations, (a higher value than the actual target range in many stud-
together with glucose targets and levels achieved. Eight of ies) in the control group was lower than in the Leuven
the studies were randomized, controlled trials; two used studies in two of the trials [29, 38] and not stated in two
historical controls. Diabetes mellitus was an entry criterion others [39, 40]. It is therefore not surprising that the aver-
for two of the studies; in the other eight, the majority of age glucose in the control group of four of the five negative
patients did not have diabetes. Patients were receiving studies was lower or equal to the average glucose in the
nutritional support in eight of the studies but did not control group of all of the positive studies. Mesotten and
receive artificial nutrition in two [33, 34]. Table 20.2 pro- van den Berghe suggest that this is due to the strong influ-
vides information on severity of illness (APACHE II scores) ence of the positive studies on the design of the subsequent
and mortality. negative studies, and could have minimized differences in
We believe that the first five studies should be consid- mortality between groups [43].
ered to indicate benefit from intensive insulin therapy, In a pediatric ICU setting 1369 patients under the age of
whereas the second five studies should be considered 16 at 13 centers in England were randomized to tight glyce-
negative. Certainly, this interpretation could be ques- mic control (694) and to conventional glycemic control
tioned. The DIGAMI study was likely underpowered, and (675); 60% had undergone cardiac surgery. Participants
showed benefit in the whole cohort only with long-term were expected to require mechanical ventilation and vaso-
follow-up. However, in- hospital mortality was signifi- pressor support for at least 12 hours. As seen in adult stud-
cantly improved with intensive therapy in the diabetic ies, tight glycemic control had no effect on major clinical
patients who were lower risk and not previously treated outcomes, although the incidence of hypoglycemia was
with insulin [33]. The Portland and Stamford studies uti- higher in this group Allen E et al. N Engl J Med. 370(2):107–
lized historical controls, an inherently weaker design, but 118, 2014 Jan 09.
both had the strength of a large sample size. The Leuven 2
study has been interpreted as unequivocally negative by
Role of Hypoglycemia in Interpretation
some authors because there was no difference in the pri-
of Study Results (Table 20.3)
mary end point, i.e., mortality in the aggregate cohort [36].
However, in the subset of individuals who were in the Hypoglycemia (defined as a blood glucose ≤ 2.2 mmol/L, or
ICU for ≥3 days, mortality was significantly reduced with < 40 mg/dl) was reported in the landmark van den Berghe
intensive therapy. This delay in apparent benefit was also study of surgical patients as more common in patients
observed to a lesser extent in the surgical ICU study from receiving intensive treatment than in the control group, but
the same investigators [7]. The benefit in the long-stayers not leading to hemodynamic deterioration or convulsions.
in ICU in Leuven 2 was least apparent in patients in the It was not discussed further. The frequency of hypoglycemia
highest quartile of APACHE II scores [36]. The mortality was remarkably low in the Krinsley study of medical ICU
data in Table 20.2 clearly indicate the Leuven 2 patients to patients, especially considering that a more liberal definition
be the sickest of those in any of the ten studies. It has been of hypoglycemia (< 3.3 mmol/L) was employed. The Leuven
pointed out that a failure of benefit with short-term ther- 2 study was the first of the trials to indicate that hypoglyce-
apy could dilute to insignificance an effect in patients mia might be responsible for mortality in some patients. In
with longer ICU stays [42]. three of the five subsequent negative studies there was an
TABLE 20.1 Trial design, patient characteristics.
Mean blood glucose Mean blood glucose

target (mmoI/L) achieved (mmoI/L)
Type of Patient Number of Diabetes

Study study population patients (%) intensive conventional intensive conventional
DIGAMI, RCT, Coronary care 620 100 7.0–10.9 Usual care 9.6 11.7
1997 [33] multicenter unit
Leuven-1, RCT, Surgical ICU 1548 13 4.4–6.1 10–11.1 5.7 8.5
2001 [7] single-center
Furnary, Historical CABG 3554 100 variable < 11.1 9.8 11.9
2003 [34] controls,
single-center
Krinsley, Historical Medical-surgical 1600 17 < 7.8 Usual care 7.3 8.4
2004 [35] controls, ICU
single-center
Leuven-2, RCT, Medical-surgical 1200 17 4.4–6.1 10–11.1 5.8 8.6
2006 [36] single-center ICU
De La Rosa, RCT, Medical-surgical 504 29–32 4.4–6.1 10–11.1 6.5 8.2
VISEP, RCT, Medical-surgical 537 30 4.4–6.1 10–11.1 6.2 8.4
2008 [38] multicenter ICU
Arabi, RCT, Medical-surgical 523 32–48 4.4–6.1 10-11.1 6.4 9.5
NICE- RCT, Medical-surgical 6104 20 4.5–6.1 < 10 6.4 8.0
SUGAR, multicenter ICU
2009 [29]
Glucontrol, RCT, Medical-surgical 1078 16–21 4.4–6.1 7.8–10 6.5 8.0
2009 [40] multicenter ICU
TABLE 20.2 APACHE scores and mortality data.
Mortality (%)
APACHE Relative reduction in Basis for difference in

Study score Intensive Conventional mortality (%) mortality vs controls
DIGAMI, 1997 [33] NR 19 (at 1 year 26 30 NR

follow-up)
Leuven-1, 2001 [7] 9 (median) 7.2 (in-hospital) 10.9 34 ↓ sepsis, MOF
Furnary, 2003 [34] NR 2.5 (in-hospital) 5.3 47 Cardiac related
Krinsley, 2004 [35] 15 (median) 14.8 20.9 20.9 ↓ in subpopulation with
septic shock, neurological
or surgical diagnosis
Leuven-2, 23 (mean) 37.3 (43 for > 3 40 (52.5 for > 3 ICU NS (18,P< 0.05* ) Multiple
2006 [36] ICU days) days)
(in-hospital)
De La Rosa 16 (mean) 36.6 (in-hospital) 32.4 NS –
2008 [37]
VISEP, 2008 [38] 20 (mean) 39.7 (90 day) 35.4 NS –
Arabi, 2008 [39] 23 (mean) 13.5 17.1 NS –
NICE-SUGAR, 21 (mean) 27.5 24.9 Increased by 14 Increases in cardiovascular
2009 [29] death
Glucontrol, 15 (mean) 15.3 17.2 NS –
2009 [40]
TABLE 20.3 Hypoglycemia data.
Incidence of hypoglycemia (%) Association between

Definition of hypoglycemia and
Study hypoglycemia (mmoI/L) Intensive Conventional mortality?
DIGAMI, 1997 [33] < 3.0 15.0 0.0 NR

Leuven-1, 2001 [7] ≤2.2 5.1 0.8 No
Furnary, 2003 [34] NR NR NR NR
Krinsley, 2004 [35] < 3.3 1.4 0.9 No
Leuven-2, 2006 [36] ≤2.2 18.7 3.1 Yes
De La Rosa, 2008 [37] ≤2.2 8.3 0.8 NR
VISEP, 2008 [38] ≤2.2 12.1 2.1 Yes
Arabi, 2008 [39] ≤2.2 28.6 3.1 Yes
NICE-SUGAR, ≤2.2 6.8 0.5 No
2009 [29]
Glucontrol, 2009 [40] ≤2.2 8.7 2.7 Yes
association between hypoglycemia and mortality [38–40]. the 3.5 mmol/L [46] to 3.9 mmol/L [47] range. It is there-
An association between hypoglycemia and mortality was fore simply not reasonable to dismiss glucose levels of 2.3–
confirmed in the NICE-SUGAR study, and the authors 3.8 mmol/L as clinically unimportant in patients who are
speculated that hypoglycemia could be partly responsible intubated [7], receiving propofol or other sedatives [37,
for adverse effects observed with intensive therapy [29]. 39], or both. Most of the studies reviewed here are encum-
Thus, there is substantial evidence that hypoglycemia could bered by the limitation that clinically important hypoglyce-
have material harmful effects on outcomes in patients mia may have been excluded from consideration by an
receiving intensive treatment with insulin infusion. inappropriately narrow definition of hypoglycemia. A
It is noteworthy that hypoglycemia is conventionally rather high incidence of hypoglycemia (15%) was reported
defined in the critical care literature as a value in the intensive arm of the DIGAMI study [33], but this is
≤2.2 mmol/L [7, 29, 36–40]. It appears that this definition likely because a liberal definition (< 3.0 mmol/L) of hypo-
is used because of a prevailing view that less severe hypoglycemia was used. The incidence of “severe” (<
glycemia is not clinically important. In fact, a recent survey 2.2 mmol/L) hypoglycemia in the DIGAMI study is not
of adult intensivists showed that over 40% thought hyper- known.
glycemia was more dangerous than hypoglycemia, and a In summary, it is clear that hypoglycemia is a potentially
majority preferred a target of 6.1 mmol/L [28]. Ironically, serious limitation to the pursuit of near-euglycemia with
in the same survey the median value for blood glucose that intensive insulin therapy in the critically ill. It is possible,
was considered to be the hypoglycemic threshold was even likely, that hypoglycemia mitigated beneficial effects
3.33 mmol/L. In fact, severe symptoms of hypoglycemia of intensive insulin therapy in negative [29, 36–40] and
can occur at glucose concentrations much higher than equivocal [36] studies. Failure to consider less severe (2.3–
2.2 mmol/L, and the counterregulatory response, including 3.8 mmol/L) hypoglycemia is a serious limitation of these
sympathoadrenal activation, routinely occurs as glucose studies. A recent retrospective analysis of hypoglycemia in
levels descend through the 3.6–3.8 mmol/L range [45]. 1109 patients from two Australian ICUs demonstrated a
Moreover, defective counter-regulation [42] and lack of step-wise and significant increase in mortality associated
clinical cues of hypoglycemia may make hypoglycemia with increasingly severe hypoglycemia, including patients
more difficult to identify, but not necessarily less threaten- with glucose values of 4.0–4.5 mmol/L [48]. This indicates
ing. For this reason, hypoglycemia experts have argued that that even very mild hypoglycemia may have significant
the threshold for clinical important hypoglycemia lies in effects on survival in the critically ill.
Methodology for Glucose Testing The Role of Nutritional Support in Results

of Trials
The observation that glucose values (4.0–4.5 mmol/L) not
considered to be in the hypoglycemic range by even the A physician, so the story goes, once gave 1 mg of thyroxine
most liberal definition of hypoglycemia could be associated daily to an elderly, frail woman who complained of fatigue.
with increased mortality is difficult to understand. In peo- When the patient developed acute atrial fibrillation and
ple with diabetes, there may be a shift upward in the glyce- died, the physician concluded that death was to be expected
mic threshold for the sympathoadrenal response [49]. from thyroxine administration, and that it should never be
Another possible explanation would apply to nondiabetic used again. This kind of reasoning has been attributed to
as well as diabetic patients, and relates to the methods used the “Chagrin Factor,” which derives from aversion to a
to measure glucose in the critically ill. Point-of care (POC) course of action that has produced an unexpected poor
blood glucose testing was used in a number of the studies, outcome [53].
and it has been shown to be very inaccurate and imprecise, Nutrition support is often used in the care of the criti-
especially during hypoglycemia. CLIA defines inaccuracy cally ill, affecting glucose levels and insulin requirements.
with POC glucose devices as values > 20% different from In two of the ten large trials [33, 34], nutrition was given to
the laboratory value except in the case of hypoglycemia, very few patients [22]. Enteral nutrition was used almost
where values > 0.83 mmol/L different are considered inac- exclusively in the study of Krinsley (J Krinsley, personal
curate. Thus, a true glucose level of 4.0 mmol/L could be communication to JMM). Enteral nutrition was used pri-
reported as 3.2 mmol/L, or a true level of 2.8 could be marily in four studies [29, 37, 39, 40], whereas enteral and
reported as 3.6 mmol/L, and both values would be consid- parenteral nutrition were used equally in the VISEP
ered to be accurate. Critchell et al. reported that POC study [38]. In the two Leuven studies, the majority of nutri-
devices, when used by trained laboratory technicians, were tional support was parenteral, at least for the first 7 days in
inaccurate in 19% of samples [50]. Kanji et al. found the the ICU. This has an impact on interpretation of the data
agreement of POC measurements with a central laboratory from the Leuven 1 study, since parenteral nutrition is
to be 56%, and only 26% in hypoglycemia samples [51]. The known to increase infection rates [54], and differences in
potential for error when POC devices are used is thus read- infection, especially sepsis, accounted for most of the dif-
ily apparent. POC devices may also systematically overesti- ferences in mortality between the intensive and control
mate true glucose [50]. Among the large clinical trials groups in Leuven 1 [22, 55]. Enteral feeding does not
reviewed here, a bedside meter was used exclusively in two appear to confer the same risk of infection [55], although
of the negative studies [37, 39]. NICE-SUGAR used both a control data on this issue are not available. In contrast, the
blood gas machine, which has excellent precision and accu- differences in mortality in the NICE-SUGAR study were
racy [52], and a POC device. When a POC device was used based primarily on cardiovascular outcomes, which may be
in NICE-SUGAR, laboratory confirmation was obtained influenced adversely by hypoglycemia [56].
only 60% of the time. Even when hypoglycemic values were In addition to the route of nutrition, the amount of nutri-
confirmed, it should be recognized that the antecedent glu- tion given may influence the results of studies of glycemic
cose level was unconfirmed and thus could have dictated an control in the critically ill. We calculated energy supply in
inappropriate decision about insulin infusion rate that led the Leuven 1 trial in relation to estimated basal energy
to the subsequent hypoglycemic episode. The Leuven 1 expenditure, and thus basal energy requirements, using the
study used a blood gas instrument, and Leuven 2 used the Harris–Benedict equation [22]. This can be done for any
Hemocue, a POC device that is designed for clinical use, study in which the calories given are reported and height
not for self-monitoring. The VISEP study utilized both a and weight are provided. In the absence of height and weight
blood gas device and the Hemocue [38]. Thus, the tech- data, the calculation can be made from body mass index and
niques used for blood glucose measurement in several of estimated height [22]. We therefore calculated energy given
the studies may have contributed to hypoglycemia, failed to in relation to energy requirements in five of the reports
identify hypoglycemia, and even overcalled hypoglycemia. where adequate information was available [7, 29, 36–39],
140
130
Energy supply (% BEE)

120
110
100
90
80
70
Leuven 1 Leuven 2 De La VISEP NICE-
Rosa SUGAR
FIG 20.1 Energy supply expressed as a percentage of estimated basal energy expenditure (BEE, calculated from the Harris–Benedict
equation) in five studies.
150
LEUVEN 1
130
Energy supply (% BEE)
DE LA ROSA LEUVEN 2
110
NICE-SUGAR y = 1.4169x + 34.752

90 R2= 0.923
VISEP
70
30 40 50 60 70 80
Insulin infusion rate (U/day)
FIG 20.2 Energy supply versus insulin infusion rate in five studies.
using data on the intensively treated arm of the studies. values used for daily feeding rate were 27 kcal/kg for Leuven
Height was estimated from online information (e.g., http:// 1 [22], 25 kcal/kg for Leuven 2 [36], 25.5 kcal/kg for De La
forums.interbasket.net/f10/average-m ale-h eight-b y- Rosa [37], 1236 kcal/day for VISEP [38] [from online
country-updated-9287/ and http://en.wikipedia.org/wiki/ appendix], and 1624 kcal for NICE-SUGAR [29] [from
Human height). Including protein in the energy calculation online appendix]. In several of the studies [7, 29, 38] the
(estimated to be 1.0 g/kg per day when not provided), the average energy intake on days 3–5 was used. Energy intake
in the Arabi study was provided only as an average given methodology used for blood glucose testing on frequency
over the first 7 days, and was much lower than expected of hypoglycemia, and did not consider the possibility that
based on the method for feeding described by the authors – (1) hypoglycemia could be avoided, that (2) selection of
using the Harris–Benedict equation and adjusting for stress more reasonable glucose targets and use of better glucose
factors [39]; for this reason, we excluded the Arabi study testing techniques could facilitate avoidance of hypoglyce-
from the analysis. Energy intake expressed as a percentage of mia, and (like the physician who gave up on thyroxine
estimated basal energy requirements is shown in Figure 20.1. rather than consider a different dose) that (3) successful
The feeding rates among the various studies are surprisingly avoidance of hypoglycemia might completely change study
different. The amount of nutrition given in the Leuven stud- outcomes and conclusions.
ies, although within published guidelines, is greater than
probable energy requirements in the critically ill [57] and
Conclusions
may account for the rather high insulin requirements in the
intensively treated groups in those studies, especially note- Available data concerning intensive therapy of hyperglyce-
worthy considering that very few of the participants had mia in the critically ill is inconsistent and confusing.
diabetes. In contrast, studies that relied primarily on enteral However, much of the contradictory data is likely related to
feedings such as the NICE-SUGAR study provided energy differences in experimental design that result in different
more in line with probably energy requirements [57]. This frequency and severity of hypoglycemia. In most of the
may be in part because tube feeding intolerance tends to studies reviewed here, hypoglycemia was poorly defined
limit infusion rate of enteral formulae. The relationship and insufficiently respected. It should be emphasized that
between amount of feeding and insulin infusion rate for the hypoglycemia may exert adverse effects of its own, masking
intensive groups of five studies is shown in Figure 20.2. As beneficial effects of glycemic control. Problems with the
can be seen, there was a strong correlation between feeding accuracy and precision of available blood glucose testing
rate and insulin requirement, particularly impressive since methodology impose considerable limitations on interpre-
the modes of feeding were heterogenous among the studies, tation of data and contribute to safety concerns. Nutritional
and when energy supply is controlled for, insulin require- support may also confound the interpretation of large clini-
ments related to intravenous feeding are greater than cal trials, although avoidance of overfeeding [57] should
requirements during enteral feeding [58]. The data shown in minimize concerns in that regard. Based on available data,
the figures are admittedly estimates, but do serve to illustrate we believe that intensive insulin therapy has a role in the
how variable approaches to nutrition support can impact management of the critically ill, although it is unclear what
insulin requirements in the critically ill. This is an important the optimal target range should be. Pending further
point, as it is common to invoke “stress” as the cause of tem- research and in the interest of minimizing hypoglycemia,
porary hyperglycemia in critically ill nondiabetic patients, we believe that the current ADA/AACE guidelines for
without sufficient attention to the role of overfeeding [57]. treatment of hyperglycemia on the critically ill, which rec-
A recently published meta-analysis concluded that there ommend insulin infusion with a blood glucose target of
is no evidence to support the use of intensive insulin ther- 7.8–10 mmol/L, are appropriate. Robust and carefully
apy in general medical- surgical patients who are fed standardized methods for blood glucose testing should be
according to current guidelines [59]. The study was unu- used, staff should be well trained, and hypoglycemia should
sual in that it attempted to assess the role of nutritional be appropriately defined and assiduously avoided.
support in the outcomes of some of the same clinical trials
reviewed here. It found that there was a high incidence of
Acknowledgments
hypoglycemia and increased risk of death in patients on
intensive insulin therapy not receiving parenteral nutri- This work was supported in part by the USPHS (HL67933)
tion [29, 37, 39, 40]. Unfortunately, the authors did not and the Mayo Foundation. We thank P. E. Cryer for helpful
consider the impact of how hypoglycemia is defined or the comments on the manuscript.
References 16. Langouche L, Vanhorebeek I, Vlasselaers D et al. Intensive

insulin therapy protects the endothelium of critically ill
1. Umpierrez GE, Isaacs SD, Bazargan N et al. Hyperglycemia: patients. J Clin Invest. 115:2277–2286, 2005.
an independent marker of in-hospital mortality in patients 17. Stoner HB, Little RA, Frayn KN et al. The effect of sepsis on
with undiagnosed diabetes. J Clin Endocrinol Metab 87:978– the oxidation of carbohydrate and fat. Br J Surg. 70:32–35,
982, 2002. 1983.
2. Cook CB, Kongable GL, Potter DJ et al. Inpatient glucose 18. Steinberg HO, Tarshoby M, Monestel R et al. Elevated circu-
control: a glycemic survey of 126 U.S. hospitals. J Hosp Med. lating free fatty acid levels impair endothelium-dependent
4:E7–E14, 2009. vasodilation. J Clin Invest. 100:1230–1239, 1997.
3. Golden SH, Peart-Vigilance C et al. Perioperative glycemic 19. Stojiljkovic MP, Zhang D, Lopes HF et al. Hemodynamic
control and the risk of infectious complications in a cohort effects of lipids in humans. Am J Physiol. 280:R1674–R1679,
of adults with diabetes. Diabetes Care 22:1408–1414, 1999. 2001.
4. Zerr KJ, Furnary AP, Grunkemeier GL et al. Glucose control 20. Lewis GF, Uffelman KD, Szeto LW et al. Interaction between
lowers the risk of wound infection in diabetics after open free fatty acids and insulin in the acute control of very low
heart operations. Ann Thorac Surg 63:356–361, 1997. density lipoprotein production in humans. J Clin Invest.
5. Weir CJ, Murray GD, Dyker AG, and Lees KR. Is hypergly- 95:158–166, 1995.
caemia an independent predictor of poor outcome after 21. Mesotten D, Swinnen JV, Vanderhoydonc F et al.
acute stroke? Results of a long-term follow up study. BMJ. Contribution of circulating lipids to the improved outcome
314:1303–1306, 1997. of critical illness by glycemic control with intensive insulin
6. Krinsley JS. Association between hyperglycemia and therapy. J Clin Endocrinol Metab. 89:219–226, 2004.
increased hospital mortality in a heterogeneous population 22. Miles JM, McMagon MM, and Isley WL. For debate: no, the
of critically ill patients. Mayo Clin Proc. 78:1471–1478, 2003. glycemic target in the critically ill should not be <6.1 mmol/L.
7. van den Berghe G, Wouters P, Weekers F et al. Intensive Diabetologia. 51:916–920, 2008.
insulin therapy in critically ill patients. N Engl J Med. 23. Langouche L, Vander Perre S, Wouters PJ et al. Effect of
345:1359–1367, 2001. intensive insulin therapy on insulin sensitivity in the criti-
8. Garber AJ, Moghissi ES, Bransome ED, Jr. et al. ACE posi- cally ill. J Clin Endocrinol Metab. 92:3890–3897, 2007.
tion statement. Endocr Pract 10:5–9, 2004. 24. Weekers F, Van Herck E, Coopmans W et al. A novel in vivo
9. McCowen KC, Malhotra A, and Bistrian BR. Stress-induced rabbit model of hypercatabolic critical illness reveals a
hyperglycemia. Crit Care Clin. 17:107–124, 2001. biphasic neuroendocrine stress response. Endocrinology.
10. Andreelli F, Jacquier D, and Troy S. Molecular aspects of 143:764–774, 2002.
insulin therapy in critically ill patients. Curr Opin Clin Nutr 25. Pomposelli JJ, Baxter JK, 3rd, Babineau TJ et al. Early post-
Metab Care. 9:124–130, 2006. operative glucose control predicts nosocomial infection rate
11. van den Berghe G. How does blood glucose control with in diabetic patients. JPEN: J Parenter Enteral Nutr. 22:77–81,
insulin save lives in intensive care? J Clin Invest. 114:1187– 1998.
1195, 2004. 26. ADA: Standards of Medical Care in Diabetes – 2009.
12. Vanhorebeek I, De Vos R, Mesotten D et al. Protection of Diabetes Care 32:S41, 2009.
hepatocyte mitochondrial ultrastructure and function by 27. Moghissi ES, Korytkowski MT, DiNardo M et al. American
strict blood glucose control with insulin in critically ill Association of Clinical Endocrinologists and American
patients. Lancet. 365:53–59, 2005. Diabetes Association consensus statement on inpatient gly-
13. Bagdade JD, Stewart M, and Walters E. Impaired granulo- cemic control. Diabetes Care. 32:1119–1131, 2009.
cyte adherence. A reversible defect in host defense in 28. Hirshberg E, Lacroix J, Sward K et al. Blood glucose control
patients with poorly controlled diabetes. Diabetes. 27:677– in critically ill adults and children: a survey on stated prac-
681, 1978. tice. Chest 133:1328–1335, 2008.
14. Poitout V and Robertson RP. Glucolipotoxicity: fuel excess 29. Investigators N-SS, Finfer S, Chittock DR et al. Intensive ver-
and beta-cell dysfunction. Endocrine Rev. 29:351–366, 2008. sus conventional glucose control in critically ill patients.
15. Hansen TK, Thiel S, Wouters PJ et al. Intensive insulin ther- NEngl J Med. 360:1283–1297, 2009.
apy exerts antiinflammatory effects in critically ill patients 30. Finfer S, Chittock D, Li Y et al. Intensive versus conventional
and counteracts the adverse effect of low mannose-binding glucose control in critically ill patients with traumatic brain
lectin levels. J Clin Endocrinol Metab. 88:1082–1088, 2003. injury: long-term follow-up of a subgroup of patients from
the NICE- SUGAR study. Intensive Care Medicine. 44. Macrae D, Grieve R, Allen E et al. A randomized trial of
41(6):1037–1047, 2015. hyperglycemic control in pediatric intensive care. N Engl J
31. Wiener RS, Wiener DC, and Larson RJ. Benefits and risks of Med. 370(2):107–118, 2014 Jan 09
tight glucose control in critically ill adults: a meta-analysis. 45. Cryer PE. Hierarchy of physiological responses to hypogly-
[Erratum appeared in JAMA. 2009 Mar4;301(9):936]. cemia: relevance to clinical hypoglycemia in type I (insulin
JAMA. 300:933–944, 2008. dependent) diabetes mellitus. Horm Metab Res. 29:92–96,
32. Griesdale DEG, de Souza RJ, van Dam RM et al. Intensive 1997.
insulin therapy and mortality among critically ill patients: a 46. Frier BM. Defining hypoglycaemia: what level has clinical
meta-analysis including NICE-SUGAR study data. CMAJ. relevance? Diabetologia. 52:31–34, 2009.
180:821–827, 2009. 47. Cryer PE. Preventing hypoglycaemia: what is the appropri-
33. Malmberg K. Prospective randomised study of intensive insulin ate glucose alert value? Diabetologia. 52:35–37, 2009.
treatment on long term survival after acute myocardial infarc- 48. Egi M, Bellomo R, Stachowski E et al. Hypoglycemia and
tion in patients with diabetes mellitus. DIGAMI (Diabetes outcome in critically ill patients. Mayo Clin Proc.
Mellitus, Insulin Glucose Infusion in Acute Myocardial 85:217–224.
Infarction) Study Group. BMJ. 314:1512–1515, 1997. 49. Boyle PJ, Schwartz NS, Shah SD et al. Plasma glucose con-
34. Furnary AP, Gao G, Grunkemeier GL et al. Continuous centrations at the onset of hypoglycemic symptoms in
insulin infusion reduces mortality in patients with diabetes patients with poorly controlled diabetes and in nondiabetics.
undergoing coronary artery bypass grafting. J Thorac N Engl J Med. 318:1487–1492, 1988.
Cardiovasc Surg. 125:1007–1021, 2003. 50. Critchell CD, Savarese V, Callahan A et al. Accuracy of bed-
35. Krinsley JS. Effect of an intensive glucose management pro- side capillary blood glucose measurements in critically ill
tocol on the mortality of critically ill adult patients. Mayo patients. Intens Care Med. 33:2079–2084, 2007.
Clin Proc. 79:992–1000, 2004. 51. Kanji S, Buffie J, Hutton B et al. Reliability of point-of-care
36. Van den Berghe G, Wilmer A, Hermans G et al. Intensive insulin testing for glucose measurement in critically ill adults. Crit
therapy in the medical ICU. N Engl J Med. 354:449–461, 2006. Care Med. 33:2778–2785, 2005.
37. De La Rosa GDC, Donado JH, Restrepo AH et al. Strict gly- 52. Beneteau-Burnat B, Bocque M-C, Lorin A et al. Evaluation
caemic control in patients hospitalised in a mixed medical of the blood gas analyzer Gem PREMIER 3000. Clin Chem
and surgical intensive care unit: a randomised clinical trial. Lab Med. 42:96–101, 2004.
Crit Care (London, England). 12:R120, 2008. 53. Feinstein AR. The ‘chagrin factor’ and quantitative decision
38. Brunkhorst FM, Engel C, Bloos F et al. Intensive insulin analysis. Arch Int Med. 145:1257–1259, 1985.
therapy and pentastarch resuscitation in severe sepsis. N 54. Perioperative total parenteral nutrition in surgical patients.
Engl J Med. 358:125–139, 2008. The Veterans Affairs Total Parenteral Nutrition Cooperative
39. Arabi YM, Dabbagh OC, Tamim HM et al. Intensive versus Study Group. N Engl J Med. 325:525–532, 1991.
conventional insulin therapy: a randomized controlled trial 55. Braunschweig CL, Levy P, Sheean PM, and Wang X. Enteral
in medical and surgical critically ill patients. Crit Care Med. compared with parenteral nutrition: a meta-analysis. Am J
36:3190–3197, 2008. Clin Nutr. 74:534–542, 2001.
40. Preiser J-C, Devos P, Ruiz-Santana S et al. A prospective ran- 56. Adler GK, Bonyhay I, Failing H et al. Antecedent hypoglyce-
domised multi-centre controlled trial on tight glucose control mia impairs autonomic cardiovascular function: implica-
by intensive insulin therapy in adult intensive care units: the tions for rigorous glycemic control. Diabetes. 58:360–366,
Glucontrol study. Intensive Care Med. 35:1738–1748, 2009. 2009.
41. Gray CS, Hildreth AJ, Sandercock PA et al. Glucose- 57. Miles JM. Energy expenditure in hospitalized patients:
potassium-insulin infusions in the management of post- implications for nutritional support. Mayo Clin Proc.
stroke hyperglycaemia: the UK Glucose Insulin in Stroke 81:809–816, 2006.
Trial (GIST-UK). Lancet Neurol. 6:397–406, 2007. 58. van den Berghe G, Wouters PJ, Bouillon R et al. Outcome
42. Cryer PE. Hypoglycaemia: the limiting factor in the glycae- benefit of intensive insulin therapy in the critically ill:
mic management of the critically ill? Diabetologia. 49:1722– Insulin dose versus glycemic control. Crit Care Med. 31:359–
1725, 2006. 366, 2003.
43. Mesotten D, Van den Berghe G. Clinical benefits of tight gly- 59. Marik PE and Preiser J-C. Toward understanding tight gly-
caemic control: focus on the intensive care unit. Best Practice cemic control in the ICU: a systematic review and metaanal-
& Research Clinical Anaesthesiology. 23:421–429, 2009. ysis. Chest. 137:544–551, 2010.
21 Diagnosis and Management
of Ophthalmic Complications of Diabetes
Andrew J. Barkmeier1 and Konstantin Astafurov2
1
Associate Professor of Ophthalmology, Department of Ophthalmology, Mayo Clinic, Rochester, MN, USA
2
Konstantin Astafurov, MD PhD, Fellow in Vitreoretinal Surgery, NJRetina/Rutgers Robert Wood Johnson Medical
School, New Brunswick, NJ, USA
currently available therapies for ocular complications of

diabetes (Table 21.1) and briefly touch upon future poten
●● The most significant threat for vision loss in diabetes is
tial therapies.
due to retinal disease.
Glycemic and metabolic control helps prevent ocular
Risk Factors for Diabetic Retinopathy
●●
microvascular complications of diabetes.

●● In combination with laser photo‐coagulation, The prevalence of diabetic retinopathy increases with the
anti‐VEGF therapy improves outcomes in diabetic
duration of diabetes in patients with both Type 1 and Type
macular edema and proliferative retinopathy.
2 diabetes. Early epidemiologic studies have revealed that
after 20 years of DM, nearly 97% of patients with Type 1
DM and between 50% and 80% of patients with Type 2 DM
exhibit some degree of DR [3]. Two other major factors
Introduction
associated with increased risk of development and progres
Long‐standing DM may affect many parts of the eye sion of DR are poor glycemic control and uncontrolled
including the lens, cornea, and orbit; however, the most hypertension. Additional risk factors include the presence
significant threat for vision loss in DM is due to its effects of other microvascular complications, such as diabetes‐
on the retina. Diabetic retinopathy (DR) is a microvascular related nephropathy and neuropathy, as well as dyslipi
complication of diabetes that develops in substantial pro demia, obstructive sleep apnea, and pregnancy [4–6].
portion of diabetic patients over time. Epidemiologic stud In recent decades, as insulin therapy has become more
ies have estimated that one in three patients with diabetes widespread, the prevalence of DR has decreased [7]. It is
will develop DR [1]. If untreated, advanced stages of DR, noteworthy that intensive insulin therapy may lead to
such as proliferative DR (PDR) and diabetic macular worsening of retinopathy during the first year following
edema (DME) may lead to irreversible blindness. In fact, initiation, yet the long‐term benefits of intensive insulin
DR is the primary cause of visual loss in patients between treatment significantly outweigh the risks [8]. This para
25 and 74 years of age [2]. If detected early and treated in a doxical phenomenon of early diabetic retinopathy acceler
timely manner, however, the most of visual impairment ation following rapid improvement of glycemic control is
from diabetes‐related ocular complications could be mini well‐established, but does not yet have a conclusive patho
mized and even prevented. In this chapter we discuss physiologic explanation. Several theories have been pro

252
Diagnosis and Management of Ophthalmic Complications of Diabetes 253
TABLE 21.1 Ocular complications of diabetes and treatment options.
Ocular complication Treatment options
Diabetic retinopathy
• Non‐proliferative Observation; control of risk factors*
• Proliferative Pan‐retinal photocoagulation and/or intravitreal
• Diabetic macular edema • Anti‐VEGF injections
• Intravitreal anti‐VEGF injections
• Focal laser
• Intravitreal steroid injections and implants
• Pars plana vitrectomy
• Macular ischemia No established treatment available
Complications of proliferative diabetic retinopathy
• Vitreous hemorrhage Observation, pars plana vitrectomy
• Tractional retinal detachment Pars plana vitrectomy if threatening the macula
• Neovascular glaucoma intravitreal anti‐VEGF injections and/or PRP combined with topical glaucoma
medications; commonly requires incisional glaucoma surgery
Cataract Observation, cataract surgery
Corneal complications
• Dry eye syndrome Lubricating agents, topical cyclosporine A, punctal plugs, autologous serum
tears
• Recurrent epithelial erosions Anterior stromal puncture, phototherapeutic keratectomy
• Corneal edema Hypertonic saline drops
Orbital cellulitis/Mucormycosis Aggressive antifungal therapy with tissue debridement; orbital exenteration
*
Recommended for all ocular diabetic complications.
posed including changes in intravascular osmotic pressure of undiagnosed metabolic derangement. Up to one‐fifth of
leading to small vessel damage, insulin‐induced produc individuals with Type 2 DM already have retinopathy at the
tion of reactive oxygen species causing an increase in time of diagnosis [12]. Following the initial screening exami
VEGF levels, or other potential etiologies for rapid changes nation, recommended screening intervals have been out
in local cytokine productions, but all hypotheses remain lined in several national and international guidelines and
tentative at present and lack substantial evidence [9]. depend on retinopathy severity [13, 14].
Screening TABLE 21.2 Initial screening recommendations for DR in

patients with diabetes. Adapted from [11].
Current guidelines for DR screening differentiate between
the types of diabetes and incorporate several risk factors. Examination by
Studies have shown that in patients with Type 1 DM, vision ophthalmologist or
threatening retinopathy rarely develops in the first 5 years Classification optometrist
following diabetes diagnosis, or before puberty [10]. Thus,
Type 1 diabetes Within 5 years of diagnosis
current screening guidelines recommend an initial compre Type 2 diabetes At time of diagnosis
hensive eye exam with dilated funduscopic examination by Women with preexisting Before pregnancy or
an ophthalmologist or optometrist within 5 years of diagno diabetes planning pregnancy or during first trimester
sis [11] (Table 21.2). On the other hand, patients with Type 2 who have become pregnant
Women who develop No screening is necessary
DM should have an initial ophthalmic screening exam at the
gestational diabetes
time of diagnosis [11], as these patients may have had years
254 Management of Disease Complications
Progression of retinopathy is seen in a substantial propor Blood Pressure Control

tion of diabetic women who become pregnant – between In diabetic patients with concomitant hypertension, good
16% and 85%, depending on the initial stage of DR [5]. control of blood pressure (BP) is associated with lower risk
The Diabetes in Early Pregnancy Study (DEPS) has of retinopathy. The UKPDS data suggested that maintaining
found that both poor glycemic control and increasing BP pressure under 150/85 results in 34% and 47% reduction
severity of pre‐existing retinopathy elevate the risk of DR in worsening of retinopathy and visual acuity, respectively,
progression [15]. In addition, elevated systolic blood pres when compared to a goal of blood pressure of lower than
sure and longer duration of diabetes also appear to contrib 180/105 [21]. Subsequent trials, such as Action in Diabetes
ute to the risk of worsening retinopathy during and Vascular Disease: Preterax (Perindopril/indapamide)
pregnancy [5]. The current guidelines recommend for all and Diamicron‐MR (gliclazide) Controlled Evaluation
women with diabetes who plan to become pregnant to have (ADVANCE) and the ACCORD trial did not demonstrate
an ophthalmic examination either prior to pregnancy or the benefit of strict blood pressure control on progression of
during the first trimester [11]. The follow‐up examination DR when comparing systolic blood pressure targets of
timing is then based upon retinopathy severity. Women 120 mmHg vs. 140 mmHg [20, 22]. Current guidelines rec
without pre‐existing diabetes who develop gestational dia ommend optimization of BP in all patients with diabe
betes mellitus do not have an increased risk of developing tes [11], although there is no current level 1 evidence
DR and do not require ophthalmologic screening [16]. supporting a blood pressure target lower than 140 with
respect to slowing DR progression.
Prevention
Lipid‐lowering Strategies
Glycemic Control
Hyperlipidemia is associated with an increased risk
Strict glycemic control has remained the mainstay for pre
of microvascular complications of diabetes. Several
vention and slowing progression of diabetic eye complica
studies have specifically evaluated the role of hyperlipi
tions. The Diabetes Control and Complications Trial
demia and lipid‐lowering medications on diabetic
(DCCT) demonstrated substantially reduced risks of both
retinopathy.
DR development and progression in patients with Type 1
DM with tight glycemic control vs conventional glycemic
control (average hemoglobin A1c (HbA1c) of approxi Fenofibrate
mately 7% vs 9%, respectively) [17]. The EDIC study The results of the Fenofibrate Intervention and Event
(extension of DCCT) has further highlighted the long‐ Lowering in Diabetes (FIELD) study demonstrated that
term benefits of maintaining good glycemic control, with fenofibrate (200 mg daily) versus placebo is effective in
HbA1c levels being the most important predictors for reducing the need for laser photocoagulation [23]. A sub
development of PDR, diabetic macular edema, and pro study of the FIELD participants with fundus photographs
gression to ocular surgery for diabetic eye complica demonstrated a beneficial effect of fenofibrate on slowing
tions [18]. Similar beneficial outcomes of good glycemic progression of DR in patients with retinopathy at baseline,
control on DR were seen in Type 2 DM patients as revealed as well as on the development of macular edema. Similarly,
by The United Kingdom Prospective Diabetes Study the ACCORD Eye Study revealed a reduction in progres
(UKPDS) [19] and a more recent Action to Control sion of retinopathy in patients assigned to intensive ther
Cardiovascular Risk in Diabetes Eye Study (ACCORD) [20]. apy with fenofibrate versus placebo (6.5% versus 10.2%) in
In most patients, an HbA1c level of 7% is a recommended addition to a statin medication [24]. However, fenofibrate
target according to current guidelines [11]. However, this has not been shown to reduce the risk of cardiovascular
target may not be appropriate for certain patients, and mortality (at least in patients with triglycerides < 400 mg/
patient‐specific characteristics such as comorbidities and dl) resulting in reluctance of general medical practitioners
life expectancy need to be taken into account. to prescribe it for patients with DR [11].
Statins low‐up compared to those on a regular diet [30]. Similarly,

In a small prospective trial of 50 diabetic patients who were another recent study concluded that consumption of ω‐3
randomly assigned to simvastatin daily or to placebo over polyunsaturated fatty acids (which have previously been
6 months, none of the patients in the simvastatin group, but shown to be protective in animal models of diabetic retin
a substantial proportion of patients in the placebo group opathy [31]) decreased the risk of incident sight‐threaten
experienced worsening visual acuity [25]. The fluorescein ing DR [32]. Overall, there is a growing, yet limited,
angiography findings and fundus photographs showed evidence base regarding the impact of diet on DR develop
worsening of DR in the placebo group as compared to the ment, and additional research in this area is warranted.
statin group. In addition, there is evidence that use of
statins in diabetic patients may lower the risk of developing
Ocular Complications
DME, while hypertriglyceridemia increases such risk [26].
Many patients with DM have lipid abnormalities and are Diabetic Retinopathy
prescribed a statin to decrease the risk of cardiovascular Diabetic retinopathy is a primarily microvascular retinal
disease. Although the role of fenofibrate in slowing DR has pathology and is classified based on clinical features
been suggested by two large clinical trials, it is not com (Table 21.3). The pathogenesis of DR is complex and
monly used clinically for this purpose due to the lack of involves microvascular damage to pericytes followed be
effect on cardiovascular outcomes. Further research in this endothelial cells loss due to hyperglycemia, formation of
area and collaborations between the ophthalmologists and free radicals and advanced glycosylation end products as
general medical physicians to consider this treatment in well as activation of pro‐angiogenic pathways and factors,
patients with DR is encouraged.
Aspirin and anticoagulation TABLE 21.3 International classification of diabetic retinopathy

and diabetic macular edema. Adapted from [13].
The ETDRS trial assessed potential benefits and side‐
effects of high‐dose daily aspirin (650 mg) in patients with Findings observable on dilated
diabetic retinopathy (mild to severe NPDR or early PDR) Disease ophthalmoscopy
versus placebo. The results indicated no increase in the rate
Diabetic retinopathy (DR)
of development of high‐risk PDR or decrease in the risk of
No visible DR No abnormalities
visual loss with aspirin versus placebo, and the rate of vitre
ous hemorrhage was similar between groups [27]. Similar Mild NPDR Microaneurysms only
findings were observed in a recent systematic review of Moderate NPDR Microaneurysms and other signs (e.g.,
clinical trials looking at high‐dose aspirin therapy in dia dot and blot hemorrhages, hard
betic patients [28]. Anticoagulation also did not appear to exudates, cotton wool spots), but less
alter the rate of peri‐operative vitreous hemorrhage in than severe NPDR
patients undergoing vitrectomy for diabetic eye complica PDR Severe NPDR plus neovascularization
tions as reported by a recent retrospective study [29].
Diabetic macular edema (DME)
Diet No DME No retinal thickening or hard

There is some early evidence suggesting that diet modifica exudates in the macula
tions may be beneficial in preventing DR. Specifically, a Non center‐involving Macular edema that does not involve
nutritional trial assessing effects of a Mediterranean diet DME the central 1 mm subfield
(MedDiet) in patients at high risk for cardiovascular dis Center‐involving DME Macular edema involving the central
ease reported that the subset of diabetic patients receiving 1 mm subfield
MedDiet combined with extra virgin olive had an approxi NPDR: non proliferative diabetic retinopathy; PDR: proliferative
mately 50% reduced risk of developing DR at 6 year fol diabetic retinopathy.
including a prominent role of vascular endothelial growth with enlargement of the foveal avascular zone and may lead
factor (VEGF) [33, 34]. In addition, there is a strong evi to vision loss even in the absence of DME or other pathol
dence that diabetes leads to stimulation of the supportive ogy [36]. Currently, there are no specific treatment known
glial cells in the retina, upregulation of pro‐inflammatory to reverse macular ischemia; management primarily
cytokines (such as interleukin‐8 and tumor necrosis fac focuses on the control of DR risk factors.
tor‐alpha among others), and activation of the complement
system all of which lead to the establishment of chronic low
Treatment of Diabetic Retinopathy
grade neuro‐inflammatory state that contributes to the
microvascular pathology [35]. The early stages of DR are Non‐proliferative Diabetic Retinopathy
referred to as non‐proliferative DR (NPDR) and are char NPDR often does not significantly impact vision unless it is
acterized by a number of intraretinal vascular changes. accompanied by diabetic macular edema (see below) or
NPDR is subdivided into mild, moderate, and severe based macular ischemia. Thus, current management of isolated
on the extent and pattern of retinal changes identified on NPDR primarily involves controlling of systemic risk fac
examination. Retinal ischemia due to capillary obstruction tors to prevent progression to more advanced stages of
may eventually lead to formation of new extraretinal1 retinopathy. Some patients with severe NPDR and features
blood vessels, typically accompanied by fibrous tissue suggesting an increased risk of progression to PDR may
(referred to as neovascularization or fibrovascular prolif benefit from early treatment with laser panretinal photoco
eration), which is the hallmark of the more severe, vision‐ agulation (PRP). The decision for PRP treatment in these
threatening stage of DR, proliferative DR (PDR). PDR is cases is based on consideration of various clinical factors
further subdivided in early, high‐risk, and advanced. including fellow eye status (especially if fellow eye vision
Fluorescein angiography (FA) is a useful adjunct to clinical loss related to sequelae of PDR), inability to or risk for
examination in characterizing DR as it allows for more being lost to follow‐up, poorly controlled diabetes, and
careful assessment of retinal ischemia and degree and loca presence of other comorbid conditions among others.
tion of neovascularization and its regression in response to
the treatment. Proliferative Diabetic Retinopathy
The following abnormalities cause vision decline in DR: Development of PDR places patients at a much higher risk
for visual loss (Figure 21.1). Thus, treatment is initiated for
●● vascular leakage (macular edema) most patients whose retinopathy has progressed to this
●● capillary obstruction (macular ischemia) stage. In selected patients with early PDR and low risk of
●● complications of neovascularization (vitreous hemorrhage, progression, close observation may be considered if the
tractional retinal detachment, neovascular glaucoma). patient is motivated to defer treatment and adhere to strict
Importantly, visual decline in DR may not occur until follow‐up.
very advanced stages of retinopathy, thus highlighting the
importance of screening of all diabetic patients. Diabetic Laser Photocoagulation
macular edema (discussed later in this chapter) can develop Pan‐retinal laser photocoagulation has remained the cor
at any stage of retinopathy, but is more frequently seen in nerstone of treatment for advanced stages of DR for many
more severe stages; the development of visually significant decades. The procedure involves placing 1200 to 1600 laser
DME is usually what causes patients to seek the initial oph burns to the peripheral retina, outside the macula, to
thalmologic evaluation. PDR, when left untreated, often induce thermal damage to ischemic parts of the retina in
leads to significant visual deterioration. order to decrease angiogenic drive (by reducing intravitreal
Macular ischemia is a less common complication of DR VEGF levels [34]) and prevent complications of uncon
and is usually seen in patients with long‐standing disease. trolled neovascularization.
Its features are decreased capillary density in the fovea (the The Diabetic Retinopathy Study (DRS), completed in
center of the macula responsible for the sharpest vision) early 1980s, established that eyes receiving treatment with
800
700
600
Retina Thickness [μm]
500
400
300
200
100
FIG 21.1 Diabetic retinopathy and diabetic macular edema. A. Fundus photograph of the left eye of a diabetic patient showing
features of proliferative diabetic retinopathy and diabetic macular edema. There is significant neovascularization of the disc (NVD)
as well as a very large front of abnormal fibrovascular tissue inferotemporally to the disc (neovascularization elsewhere, NVE) and a
small patch of NVE in the superotemporal macula. The patient received prompt pan‐retinal photocoagulation and multiple laser
burns are visible outside the macula. Patient also had center‐involving DME on presentation with numerous hard exudates and
central macular thickening. Features of non‐proliferative DR could also be appreciated with dot‐blot hemorrhages and microaneu-
rysms (arrow heads) scattered throughout the fundus. B. Fluorescein angiography imaging of the same eye shows significant dye
leakage from the sites of NVD and NVE. In addition, areas of retinal ischemia due to capillary non‐perfusion are clearly seen. It can
also be appreciated that neovascularization develops close to the boundaries of ischemic and non‐ischemic areas. C. Optical
coherence tomography (OCT) generated thickness map of the macula of the same patient shows abnormal retinal thickening
which is especially pronounced temporally to the fovea. D. Cross‐sectional optical coherence tomography scan of the macula
through the fovea reveals retinal thickening due to the presence of intra‐retinal fluid (seen as hypo‐reflective cystic areas). Exudates
are also visible as focal areas of signal hyper‐reflectivity. E. Anterior‐segment photograph showing iris neovascularization (NVI) due
to advanced PDR.
PRP experienced 50% lower rate of severe vision loss (SVL) ETDRS found that eyes with mild and moderate NPDR
compared to the untreated eyes at 5‐year follow‐up [37]. should be observed, while PRP may be considered in eyes
The Early Treatment Diabetic Retinopathy Study (ETDRS) with more severe non‐proliferative diabetic retinopathy
subsequently provided evidence‐based guidelines for the and should not be delayed if retinopathy has advanced to
timing for PRP initiation (also termed scatter laser). The the proliferative stage with high‐risk features related to the
presence, location (optic disc versus elsewhere), and sur Combination Treatment
face area of neovascularization, and whether associated Given the advantages and disadvantages of both therapies,
vitreous hemorrhage is present [38]. treatment of PDR should be individualized based on
The PRP procedure is most commonly performed in an patient‐specific factors including the presence of DME, the
office setting with either topical or local (peri‐ or retrobul overall patient’s health and ability to follow up, and the sta
bar) anesthesia, and may take place over two or more ses tus of the fellow eye among others. At present, the majority
sions. PRP is associated with a number of potential adverse of providers prefer a combination approach utilizing both
effects including pain during the procedure, worsening or PRP and anti‐VEGF in most PDR patients [42]. This prac
development of macular edema with associated reduction tice does have a clinical trial evidence basis [43]. A careful
in visual acuity (usually temporary), decreased peripheral explanation of the options and discussion with the patients
vision, decrease in dark adaptation, and, rarely, subretinal regarding their preferences is essential.
neovascularization or vascular occlusions. PRP induces
regression of neovascular activity and often contraction of Surgical Treatment/Vitrectomy
fibrovascular tissue. Although this effect is desirable, it The current indications for pars plana vitrectomy in eyes
occasionally leads to vitreous hemorrhage and/or develop with DR include the presence of vitreous hemorrhage
ment or worsening tractional retinal detachment within (VH) that has not cleared over a meaningful period of
the first 1–2 months following treatment. Despite those observation, progressive tractional retinal detachment
adverse effects, clinical evidence demonstrates that the (TRD) that involves or threatens the macula, combined
long‐term, durable effects of PRP in preventing progres tractional‐rhegmatogenous retinal detachment, and pro
sion of retinopathy and severe vision loss significantly out gressive neovascularization unresponsive to PRP.
weigh the associated procedural risks. VH arises due to bleeding from the sites of neovascu
larization as those immature vessels lack endothelial tight
Anti‐VEGF junctions, which predispose them to spontaneous bleed
The use of anti‐VEGF agents in diabetic eye disease was ing. The fibrous component that accompanies those fragile
initially focused on treating DME (discussed later in this vessels may also put additional stress on them due to its
chapter). More recently, several clinical trials have revealed contraction [44]. TRDs develop when fibrovascular epi
that intravitreal injections of anti‐VEGF halt the progres centers coalesce into large vitreoretinal tractional bands
sion of advanced stage of DR with efficiency similar to that that contract and elevate the retina. TRDs may progress to
of PRP [39, 40]. However, for most patients with active extensive detachment of the retina, but this is often a slow
PDR, anti‐VEGF treatment requires multiple injections, process which may be halted with prompt PRP; TRDs that
initially with a monthly frequency, which is associated with progress despite medical therapy and encroach on the
significant economic and healthcare system burden. In macula and the fovea require semi‐urgent vitrectomy to
addition, interruptions of anti‐VEGF therapy could lead to preserve central vision. Tractional forces from pre‐retinal
rapid worsening of proliferative retinopathy and significant fibrous membranes may also cause breaks in the retina
vision loss [41]. In comparison, PRP often provides long‐ which could result in combined tractional‐rhegmatoge
term control of the disease with lower costs and much less nous retinal detachment, which require urgent surgical
frequent visits for most patients. Each intravitreal injection intervention due to their threat of rapid progression.
also poses a small risk of endophthalmitis, vitreous hemor The Diabetic Retinopathy Vitrectomy Study (DRVS)
rhage, and retinal detachment (see more detailed discus completed in 1985 found improved visual outcomes in
sion later in the section on DME). A particular advantage Type I DM patients with severe vitreous hemorrhage who
of anti‐VEGF therapy over PRP includes less frequent rate were randomized to either early or delayed (1 year) vitrec
of DME. The initial benefit of greater peripheral visual tomy [37]. No clear benefit was seen in Type II diabetics.
field preservation with anti‐VEGF therapies versus PRP However, it is unclear whether those findings remain appli
was not sustained through 5‐year follow‐up [40]. cable to contemporary patients due to significant interval
changes in surgical technologies (e.g. endolaser) and tech prognosis of NVG is especially poor as severe intraocular
niques, as well as anti‐VEGF pharmacotherapy. Docu pressure elevations may cause irreversible vision loss at the
mented outcomes of vitrectomy in the ETDRS study time of initial presentation, and because pressure control
published in 1992 were significant improvement in visual may be labile or refractory even with prompt surgical glau
acuity in diabetic eyes with either vitreous hemorrhage or coma interventions, leading to irreversible visual decline in
retinal detachment; the total rate of eyes requiring this pro many patients [50]. Hence, early detection of neovasculari
cedure was around 5% among all recruited diabetic zation of the anterior eye structures and the outflow path
patients [45]. ways is a crucial part of each diabetic eye exam. Current
More recent data shows that with modern vitrectomy management of NVG includes lowering of intraocular
tools and techniques, the vision could be stabilized in pressure and aggressive control of neovascularization with
about 80% of patients even with advanced diabetic TRDs a combination of anti‐VEGF agents, PRP, and potentially
and combined tractional‐rhegmatogenous detach incisional glaucoma surgery, all of which are much more
ments [46]. Complications of vitrectomy include forma effective at early stages of the disease.
tion or progression of cataract, recurrent VH, retinal
detachment, and endophthalmitis.
Diabetic Macular Edema
Diabetic macular edema (DME) is one of the main causes
Retinal Vein Occlusion
of visual impartment in patients with diabetes.
While DM mostly causes retinal microangiopathy, studies Epidemiologic studies estimate the overall prevalence of
have also revealed an association of DM with intermediate DME in all patients with diabetes around 7% [1]. DME is
size retinal vascular pathology, specifically, a higher inci characterized by the presence of intraretinal fluid (edema)
dence of branch retinal vein occlusion (BRVO) in diabetic with associated retinal thickening within the macula. The
patients [47]; no such association was seen for central reti macula is the central anatomic zone of the retina that is
nal vein occlusion. BRVO can lead to further retinal responsible for high resolution vision within approximately
ischemia and neovascularization; it can also lead to vision the 20 central degrees of visual field. The ETDRS trial con
loss from macular edema, either due to the BRVO itself or ducted by NIH in 1990s established the first widely‐used
due to exacerbation of pre‐existing DME. The treatment of classification of DME based on clinical funduscopic exami
neovascularization and macular edema in such eyes is sim nation [51]. It defined the term clinically significant macu
ilar to that described for PDR and DME. lar edema (CSME) which was based on the presence and
area of retinal thickening, or hard exudates associated with
thickening, as well as the proximity to the center of the
Neovascular Glaucoma
macula. The term CSME is still used, but has less clinical
PDR and other conditions that result in significant ocular relevance due to advancements in both imaging technol
ischemia (such as retinal venous or arterial occlusions, ogy and pharmacologic treatments. Optical coherence
ocular ischemic syndrome, and radiation retinopathy) tomography (OCT) is a non‐invasive test with micrometer
could lead to a fearsome complication of neovascular glau resolution that provides a high‐resolution “optical biopsy”
coma (NVG). Some earlier studies have estimated that up of the retina in an office setting within a few minutes.
to 43% of patient with PDR may develop NVG [48]; how Modern guidelines recommend classifying DME as either
ever, the incidence of this disease has substantially center‐involving (retinal thickening that affects 1mm cen
decreased after the advent of intravitreal anti‐VEGF ther tral subfield zone) and non‐center involving (retinal thick
apy [49]. The hallmark of this type of glaucoma is neovas ening outside that area) DME based on clinical and OCT
cularization of the iris and of the outflow pathways of the criteria [13].
eye (the iridocorneal angle), resulting in their blockage and Fluorescein angiography (FA) is another useful test for
a subsequent severe increase in intraocular pressure. Visual evaluating DME. It involves intravenous injection of fluo
rescein dye followed by a sequence of fundus photographs throughout the macula and not to microaneurysms. While
depicting flow of the dye through the retinal vasculature. it was not inferior to the focal laser with respect to visual
FA facilitates assessment of vascular and microvascular reti acuity outcomes at 12‐month interval, the reduction of
nal abnormalities and visualization of vascular leakage in retinal edema was less favorable [55]. Subthreshold laser
areas of blood‐retinal barrier breakdown. The pattern of technique (with lower energy delivered via micropulses
leakage seen on FA may help distinguish DME from other such that laser spots are not visible after the application)
etiologies of macular edema. Leakage on FA, however, does has also been evaluated in treating DME and has shown
not always indicate retinal swelling, and correlation of FA promising results [56, 57]; however, the data on its efficacy
with examination findings and OCT is needed. FA can be is currently limited.
helpful in characterizing whether areas of retinal thickening
are predominantly caused by leaking microaneurysms Corticosteroid
(which may benefit from focal laser treatment), or whether Corticosteroids can produce rapid improvement in retinal
there is more diffuse breakdown of the blood/retinal barrier thickening in patients with DME. However, this effect
that may be more effectively treated pharmacologically. comes with the downsides of increased risk of glaucoma
and cataract development. Moreover, reduction in edema
is usually transient and repeated steroid injections are nec
Treatment of DME
essary which increases the risk of adverse effects.
Macular Laser Corticosteroid therapies for DME have been evaluated
Focal laser therapy as a treatment for CSME was first evalu by multiple comparative trials. Current evidence suggests
ated by the ETDRS trial. It demonstrated a 50% or greater that steroid monotherapy is less efficacious than anti‐
reduction in the risk of moderate vision loss in eyes with VEGF pharmacotherapy when factoring in complications
CSME that were treated with focal laser versus eyes that of treatment. A two‐year trial of intravitreal triamcinolone
were observed [51]. However, approximately 40% of eyes acetonide versus focal laser therapy demonstrated better
with CSME in this trial had visual acuity of 20/20 or bet mean visual acuity in the laser treatment group at two
ter [52]. In the era of current pharmacotherapeutics, new years [58] and patients receiving triamcinolone had higher
evidence suggests that observation until vision loss occurs rate of glaucoma and cataract.
may be a reasonable option now that better rescue therapy Several intravitreal biodegradable implants have been
exists [53]. designed offering sustained low‐dose release of a steroid
The current protocol for focal laser treatment of DME is medication. These implants are injected intravitreally in an
based on modified ETDRS guidelines and involves applica office setting similar to other intravitreal injections.
tion of laser burns directly to microaneurysms and micro Dexamethasone 0.7mg implant (Ozurdex) provides a 3‐ to
vascular lesions in the center of the rings of hard exudates, 4‐month interval of sustained dexamethasone release. In
at least 500 microns from the fovea. The procedure usually the 3‐year Macular Edema: Assessment of Implantable
takes only a few minutes with topical anesthesia and causes Dexamethasone in Diabetes (MEAD) trial, repeated injec
minimal discomfort. In most cases, the reduction in retinal tions of this implant were associated with modestly
edema takes place over weeks or months [54]. Compli improved visual acuity as compared to sham injections
cations of focal laser photocoagulation are rare and include (22% of patients gained greater than 15 letters in dexa
laser scars that increase in size over time causing a blind methasone group versus 12% in sham group) [59].
spot near the center of the vision (paracentral scotoma), However, a greater proportion of patients developed cata
inadvertent application of laser to the fovea (which may ract and elevated IOP in the dexamethasone implant group.
decrease visual acuity), and choroidal neovascularization. In another study, addition of Ozurdex to anti‐VEGF ther
An alternative to focal laser treatment, mild macular apy in patients with persistent DME did not produce any
grid (MMG) laser was evaluated in a randomized, prospec additional visual improvement, although anatomical
tive trial [55]. This technique involves application of laser reduction of retinal edema was enhanced [60]. Direct
head‐to‐head comparison of Ozurdex to one of the availa strate that a substantial proportion of patients with DME
ble anti‐VEGF agents (bevacizumab) showed similar visual (up to 50%) may not need additional anti‐VEGF treat
gains between the groups at 24 months, with significantly ments after 3 years of injections [65].
reduced injection burden in the Ozurdex group [61].
Similarly to other studies, patients in the Ozurdex group Choice of the agent
required cataract surgery more frequently and had higher Currently, there are four anti‐VEGF agents on the market
IOP. that have been studied and used for ocular disease: bevaci
Fluocinolone acetonide (FA) 0.19 mg implant (Iluvien) zumab (Avastin; Roche, Switzerland), ranibizumab
was designed to deliver a low‐dose of FA over extended (Lucentis; Novartis, Switzerland), aflibercept (Eylea; Bayer,
period of time, up to 36 months. The Fluocinolone Germany), and brolucizumab (Beovu; Novartis,
Acetonide in Diabetic Macular Edema (FAME) study that Switzerland). Ranibizumab and aflibercept are FDA‐
evaluated this implant in patients with refractory DME approved for the treatment of DME and DR [66, 67].
who had undergone at least one prior macular laser treat Bevacizumab does not carry an FDA label for treatment of
ment found that a significantly larger proportion of ocular diseases, yet it is commonly used off‐label by oph
patients gained more than 15 letters of vision versus sham thalmologists for a variety of conditions including DME
implant at 36 months [62]. In current practice, monother due to its lower cost; when used intravitreally, bevacizumab
apy with corticosteroids for DME is uncommonly is repackaged in a compounding pharmacy into aliquots
employed as initial therapy. It is most commonly utilized containing approximately 1/500th of the dose used in sys
for patients whose DME is refractory to serial anti‐VEGF temic cancer therapy. Brolicuzimab was only recently
injections, or as an adjunctive treatment for incomplete approved by the FDA as a treatment of exudative age‐
responders. related macular degeneration (AMD) and has not yet been
studied for DME.
Anti‐VEGF Therapy There is a growing literature guiding the choice of
Pharmacologic inhibition of VEGF via intravitreal injec anti‐VEGF agent for treatment of DME. The DRCR
tion of anti‐VEGF agents has rapidly become standard of Protocol T clinical trial recently released 2‐year results
care therapy for DME since the 2010 DRCR network comparing efficacy of three anti‐VEGF agents for this
Protocol I trial which demonstrated excellent long‐term indication [68]. All three medications offered signifi
visual outcomes in patients with DME treated with anti‐ cant improvement in visual acuity at 1 and 2 year time
VEGF agents as compared to macular laser photocoagula points. The study suggested that at 2‐year follow‐up all
tion or intravitreal corticosteroid therapy [63]. Soon three agents have similar visual outcomes in eyes with
thereafter, practice patterns shifted toward a significant mild visual impairment (20/32 to 20/40 vision).
increase in intravitreal anti‐VEGF therapy for DME [64]. Aflibercept showed better visual outcomes over bevaci
Current guidelines now recommend anti‐VEGF pharma zumab, but not over ranibizumab in eyes with worse ini
cotherapy as the first line of treatment for DME [13]. tial visual acuity (20/50 or worse) at 2 year time point.
Anti‐VEGF inhibitors are usually administered on a Bevacizumab therapy was found to be not as effective in
monthly basis for the first few injections. Following that, reducing retinal thickening vs two other agents at 2
one of several clinical extension protocols is usually years. The median number of injections was similar
employed. In the treat‐and‐extend protocol, as the disease among the three drugs.
activity is controlled, patients are evaluated and treated at Given a substantial cost difference between FDA‐
increased intervals; if disease activity increases, the treat approved and off‐label anti‐VEGF agents clinicians are
ment intervals are shortened. In the as‐needed or pro re required to make difficult decisions in selecting a medica
nata (PRN) protocol, after retinal edema has subsided, tion for each patient. Such factors as the cost of care for the
patients are monitored without treatment and injections patient or society as a whole, the stage or severity of pathol
are re‐instituted if edema worsens. Recent data demon ogy, prior ophthalmic history and the state of the fellow
eye, as well as published comparative results of treatment DME did not show a significant difference regarding the risk
efficacies all play a role. for arterial thromboembolic events or overall mortality [71].
It is noteworthy, however, that poor response to a par Similarly, a recent retrospective cohort study using a large
ticular anti‐VEGF agent does not indicate a therapeutic U.S. insurance database did not reveal a higher rate of sys
failure of the whole class of the medications and that an temic adverse effects (cerebrovascular disease, myocardial
individual patient may have an improved response to a sec infarction, major bleeding) in patients treated with anti‐
ond or third drug. Therefore, many clinicians use bevaci VEGF agents versus those treated with macular laser for
zumab as an initial treatment (after a discussion of its off‐label DME [72]. Further investigations are needed to assess long‐
status with the patient) and then switch to more expensive term safety of these medications as some of the patients may
FDA‐approved agents if there is suboptimal response. receive ongoing treatment for well over a decade.
Efficacy over other treatments Surgery

There is growing evidence that anti‐VEGF agents are supe Pars plana vitrectomy (PPV) as an alternative to other treat
rior to other forms of treatment of DME. Two head‐to‐ ments for DME has been evaluated in several studies. DRCR
head randomized comparison trials of aflibercept versus Protocol D assessed eyes with at least moderate vision loss
focal laser demonstrated significantly better visual and from DME and vitreomacular traction [73]. The study was
anatomic outcomes in the aflibercept group at designed as a prospective cohort study with 87 patients
52 weeks [67]. In a recent meta‐analysis of 13 clinical trials undergoing PPV as well as additional procedures based on
comparing aflibercept and ranibizumab with macular laser surgeon’s usual routine (those included epiretinal mem
treatment, both of these anti‐VEGF agents showed supe brane (ERM) peeling, internal limiting membrane (ILM)
rior visual outcomes at 12 month timepoint [69]. Patients peeling, panretinal photocoagulation and steroid injection
who have DME and only mild visual impairment (visual at the end of the procedure). At 6‐month follow‐up central
acuity 20/25 or better) may not require treatment for retinal thickening was decreased by 50% in 68% of patients
extended periods of time and could be monitored. As a with improvement in visual acuity by greater than 10 letters
recent prospective randomized trial has suggested, such in 38%. At the same time, visual acuity decreased by more
patients do equally well with initial observation and rescue than 10 letters in 22% of patients. Subsequent subgroup
aflibercept therapy if vision decreases when compared to analysis of the study results revealed that eyes with worse
initial treatment with anti‐VEGF injections or laser ther baseline visual acuity and those that underwent ERM peel
apy over the mean follow‐up of 2 years [53]. ing experienced greater visual acuity improvement [74].
However, the results of other trials that examined the
Side effects role of vitrectomy for DME were variable with some show
Regardless of the drug used, intra‐vitreal injections carry a ing advantage of PPV over macular laser [75], while others
risk of infectious endophthalmitis (approximately 1 in failing to determine significant visual acuity
3000–5000 injections), vitreous hemorrhage, lens damage, improvement [76].
and retinal detachment. In addition, there is risk of sterile A recent systematic review of vitrectomy trials for DME
inflammation with all anti‐VEGF agents that appears to be comparing vitrectomy surgery versus observation or focal
slightly higher for aflibercept versus other agents. photocoagulation has suggested that PPV does not provide
A recent study has also brought to light a concern for significant functional advantage over laser; the safety of
potentially increased risk for death and cerebrovascular acci PPV for DME was comparable to that of PPV for other
dents in those patients with the highest anti‐VEGF injections indications [77]. At present, vitrectomy for DME is primar
exposure (treated with monthly injections for 24 months) [70]. ily to address vitreoretinal interface abnormalities that may
On the other hand, meta‐analysis of trials comparing anti‐ exacerbate DME, or less commonly as a last resort for
VEGF drugs with either sham or laser photocoagulation for patients who have failed all medical therapies.
Future Directions DME – Summary

Sustained intraocular delivery of VEGF‐neutralizing A significant proportion of patients with DM will develop
agents could substantially reduce both the treatment and DME. Current evidence suggests that only center‐involv
financial burden for patients and the health care system. ing DME necessitates treatment. Among patients with CI‐
Current therapies require frequent office visits with associ DME, some patients may retain normal or nearly normal
ated imaging, pharmaceutical, professional, and travel vision, and new data suggests that such patients could be
costs. Several recent Phase 2 and Phase 3 trials have exam observed for at least 2 years as long as visual acuity remains
ined various methodologies for such sustained anti‐VEGF very good (20/25 or better). For patients with significant
release including port delivery system (a permanent, surgi visual decline from DME, intravitreal injections of anti‐
cally implanted, office refillable reservoir for holding a VEGF agents are usually the preferred initial treatment.
large pool of anti‐VEGF agent) [78] and gene therapy (with Injections are started on a monthly basis with frequent
adeno‐associated virus vector encoding aflibercept) [79] (usually monthly) follow‐ups; however, following the first
among others. Phase 3 trials of some of those systems are few months those are typically extended. Many patients
currently underway. may not need treatment after 3 years. A choice of anti‐
Another method of inhibiting VEGF production is VEGF agent is determined by patient‐specific characteris
through the use of small interfering RNA (siRNA). This tics and cost of care.
approach has shown promising results in a mouse model A proportion of patients may not have a full response to
when siRNA was delivered to the RPE cells to inhibit the first anti‐VEGF agent, in such cases a trial of a second
expression of VEFG receptor [80]. Yet another avenue and third agent should be considered. Patients with incom
explored for targeting VEGF pathway is through sequester plete response to anti‐VEGF therapy could be candidates
ing this peptide within the cell. Flt23k intraceptor can bind for focal laser treatment or intra‐vitreal steroid injections
to VEGF and prevent it from leaving the endoplasmic (or intra‐vitreal steroid implants) as adjuncts. Laser treat
reticulum resulting in effective inhibition. Experiments in ment can also be used as primary therapy in patients who
primate and murine models have demonstrated efficient are not good candidates for repeat intravitreal injections
VEGF pathway inhibition with the delivery of recombinant due to comorbidities or social factors.
Flt23K intraceptor plasmid with a single intravenous injec Surgical options for DME include pars plana vitrectomy,
tion [81]. However, these two latter technologies are still in but it is used only rarely and is typically reserved for a select
the research stage and have not yet been translated to number of refractory cases. Overall, with the currently
human studies. available treatment options, many patients who develop
There is also ongoing research in non‐molecular treat DME can retain good vision. Future therapies currently
ment approaches for DR and DME. Photobiomodulation being developed may reduce the patient and healthcare
(PBM) has recently shown promising results in treating burden associated with frequent treatments typically
DME in preclinical trials [82, 83]. PBM involves applica required needed for DME.
tion of far‐read or near‐infrared light (NIR) (in the 600–
1000 μm range) which allows deep tissue penetration and
Cornea
may work by inhibiting production of ROS and attenuat
ing cell death as well as by increasing energy production The cornea is affected by diabetes in a variety of ways col
by mitochondria [84]. Currently, the DRCR Retina lectively known as diabetic keratopathy. Studies have esti
Network is recruiting patients with DME for a pilot trial mated that up 70% of patients with DM may have diabetes
with a PBM device that used twice per day to deliver 90 –associated corneal complications [85]. Diabetic eyes
seconds of NIR light through closed eyelid. If found effi exhibit weakening of corneal epithelial barrier which may
cacious, PBM may become be the first at‐home therapy to lead to higher risks of corneal infections. This risk may be
treat DME. exacerbated by diminished corneal peripheral sensation
that is also seen in patients with DM [86] and is thought to DME” after CE over a 2‐year period reported a rate of 1%
arise from reduced small fiber nerve density in diabetic for eyes with no DR versus 13% for eyes with severe NPDR
cornea [87]. Decreased corneal sensitivity also leads to at baseline; the risk of DME development peaked at
higher risk of dry eye syndrome [88]. Other studies have 3–6 months after surgery [96].
revealed corneal endothelial dysfunction in DM [89] which The effect of CE on DR progression is not clear with
may lead to increase in corneal thickness and edema [90]. some studies suggesting worsening of DR [97], while oth
Diabetes may also diminish the strength of adhesion of the ers showing no such association [98, 99]. Universally, stabi
epithelium to the underling basement membrane which lization of retinopathy and DME should be accomplished
may lead to prolonged healing and recurrent corneal ero prior to CE.
sions and defects [91].
Orbital Disease
Treatment Patients with uncontrolled diabetes are in general more sus
The treatment of dry eye syndrome consists of daily appli ceptible to infections and in particular are at a higher risk for
cation of lubricating eye drops and ointments. Additional acute orbital cellulitis. Most common pathogens in orbital
therapies include punctal occlusions, topical use of low cellulitis are typical sinus bacteria such as Streptococcus
doses of immunosuppressive drugs (Cyclosporine A) that pneumoniae and Haemophilus influenzae. However,
reduce inflammation in the tear glands and improve tear patients with uncontrolled DM, especially those in diabetic
secretion. Refractory cases may require autologous serum ketoacidosis, are prone to mucormycosis – a rapidly spread
drops. ing fungal infection of the orbit and sinuses caused by sapro
Corneal edema can be managed by hypertonic agents, phytic fungi of the order Mucorales [100]. Unless diagnosed
such as sodium chloride 5% solution eye drops and oint early and treated aggressively mucormycosis may lead to
ment. They act by drawing the water out of the cornea pre death in a few days. The management of this severe infection
venting or reducing the buildup of edema. usually requires a combination of aggressive systemic anti
Recurrent corneal erosions could be alleviated by ante fungal therapy (such as Amphotericin B) as well as surgical
rior stromal puncture technique [92]. In this office‐based debridement. In cases of significant orbital invasion, orbital
procedure, a small needle is used to create several micro‐ exenteration (removal of the orbital content) may be
punctures through the epithelium just below the Bowman’s required to control the disease spread. Given its high mortal
layer which leads to a strengthened bond between these ity, mucormycosis needs to be on the differential list in all
corneal layers reducing the risk of erosions. Similar results cases of orbital cellulitis in diabetics, especially in those with
could be achieved using laser phototherapeutic keratec significant metabolic derangements.
tomy, a more invasive surgical procedure which utilizes
excimer laser [93]. Note
1. In this context meaning projecting from the retina into the
Cataract vitreous cavity.
Cataract is an important cause of vision impairment in dia References

betics. In a population‐based study the risk of cataract was
1. Yau JW et al. Global prevalence and major risk factors of
approximately two‐fold higher than in general popula
diabetic retinopathy. Diabetes Care. 2012;35(3):556–564.
tion [94]; this risk increases with duration of diabetes,
2. Cheung N, Mitchell P, Wong TY. Diabetic retinopathy.
severity of retinopathy and higher HbA1c levels [94, 95]. Lancet. 2010;376(9735):124–136.
Visual recovery after uncomplicated cataract extraction 3. Klein R et al. The Wisconsin epidemiologic study of diabetic
(CE) and intra‐ocular lens implantation surgery in general retinopathy: XXII the twenty‐five‐year progression of
depends on presence of DME and stage of DR. A recent retinopathy in persons with type 1 diabetes. Ophthalmology.
study looking at the development of “treatment‐requiring 2008;115(11):1859–1868.
4. Altaf QA et al. Obstructive sleep apnea and retinopathy in tions in insulin‐dependent diabetes mellitus. N Engl J Med.
patients with type 2 diabetes. a longitudinal study. Am J 1993;329(14):977–986.
Respir Crit Care Med. 2017;196(7):892–900. 18. Hainsworth DP et al. Risk factors for retinopathy in type 1
5. Axer‐Siegel R et al. Diabetic retinopathy during pregnancy. diabetes: the DCCT/EDIC Study. Diabetes Care. 2019;42(5):
Ophthalmology. 1996;103(11):1815–1819. 875–882.
6. Zander E et al. Increased prevalence of proliferative retinopathy 19. Stratton IM et al. Association of glycaemia with macrovas
and cardiovascular autonomic dysfunction in IDDM patients cular and microvascular complications of type 2 diabetes
with proteinuria. Exp Clin Endocrinol. 1992;99(2):102–107. (UKPDS 35): prospective observational study. BMJ. 2000;
7. Lecaire T et al. Lower‐than‐expected prevalence and sever 321(7258):405–412.
ity of retinopathy in an incident cohort followed during the 20. Group AS et al. Effects of medical therapies on retinopathy
first 4–14 years of type 1 diabetes: the Wisconsin Diabetes progression in type 2 diabetes. N Engl J Med. 2010;
Registry Study. Am J Epidemiol. 2006;164(2):143–150. 363(3):233–244.
8. Early worsening of diabetic retinopathy in the Diabetes 21. UK Prospective Diabetes Study Group. Tight blood pressure
Control and Complications Trial. Arch Ophthalmol. control and risk of macrovascular and microvascular com
1998;116(7):874–886. plications in type 2 diabetes: UKPDS 38. BMJ. 1998;317(7160):
9. Bain SC et al. Worsening of diabetic retinopathy with rapid 703–713.
improvement in systemic glucose control: A review. Diabetes 22. Beulens JW et al. Effects of blood pressure lowering and
Obes Metab. 2019;21(3):454–466. intensive glucose control on the incidence and progression
10. Klein R et al. The Wisconsin epidemiologic study of diabetic of retinopathy in patients with type 2 diabetes mellitus: a
retinopathy. II. Prevalence and risk of diabetic retinopathy randomised controlled trial. Diabetologia. 2009;52(10):
when age at diagnosis is less than 30 years. Arch Ophthalmol. 2027–2036.
1984;102(4):520–526. 23. Keech AC et al. Effect of fenofibrate on the need for laser
11. Solomon SD et al. Diabetic retinopathy: a position statement treatment for diabetic retinopathy (FIELD study): a
by the American Diabetes Association. Diabetes Care. 2017; randomised controlled trial. Lancet. 2007;370(9600):

40(3):412–418. 1687–1697.
12. Klein R et al. The Wisconsin epidemiologic study of diabetic 24. Chew EY et al. The effects of medical management on the
retinopathy. III. Prevalence and risk of diabetic retinopathy progression of diabetic retinopathy in persons with type 2
when age at diagnosis is 30 or more years. Arch Ophthalmol. diabetes: the Action to Control Cardiovascular Risk in
1984;102(4):527–532. Diabetes (ACCORD) Eye Study. Ophthalmology. 2014;
13. Wong TY et al. Guidelines on Diabetic Eye Care: The 121(12):2443–2451.
International Council of Ophthalmology Recommendations 25. Sen K et al. Simvastatin retards progression of retinopathy in
for Screening, Follow‐up, Referral, and Treatment Based on diabetic patients with hypercholesterolemia. Diabetes Res
Resource Settings. Ophthalmology. 2018;125(10):1608–1622. Clin Pract. 2002;56(1):1–11.
14. International Council of Ophthalmology. Updated 2017 26. Chung YR et al. Association of statin use and hypertriglyc
ICO Guidelines for Diabetic Eye Care. 2017. San Francisco, eridemia with diabetic macular edema in patients with type
CA: ICO. 2 diabetes and diabetic retinopathy. Cardiovasc Diabetol.
15. Chew EY et al. Metabolic control and progression of retin 2017;16(1):4.
opathy. The Diabetes in Early Pregnancy Study. National 27. Early Treatment Diabetic Retinopathy Study Research Group.
Institute of Child Health and Human Development Diabetes Effects of aspirin treatment on diabetic retinopathy. ETDRS
in Early Pregnancy Study. Diabetes Care. 1995;18(5): report number 8. Ophthalmology. 1991;98(5 Suppl):757–765.
631–637. 28. Bergerhoff K, Clar C, Richter B. Aspirin in diabetic retinopa
16. Gunderson EP et al. A 20‐year prospective study of child thy. A systematic review. Endocrinol Metab Clin North Am.
bearing and incidence of diabetes in young women, control 2002;31(3):779–793.
ling for glycemia before conception: the Coronary Artery 29. Brown JS, Mahmoud TH, Anticoagulation and clinically sig
Risk Development in Young Adults (CARDIA) Study. nificant postoperative vitreous hemorrhage in diabetic vit
Diabetes. 2007;56(12):2990–2996. rectomy. Retina. 2011;31(10):1983–1987.
17. Diabetes C et al. The effect of intensive treatment of diabetes 30. Diaz‐Lopez A et al. Mediterranean diet, retinopathy,
on the development and progression of long‐term complica nephropathy, and microvascular diabetes complications: a
post hoc analysis of a randomized trial. Diabetes Care. proliferative diabetic retinopathy (PROTEUS Study).
2015;38(11):2134–2141. Ophthalmology. 2018;125(5):691–700.
31. Sapieha P et al. Omega‐3 polyunsaturated fatty acids pre 44. Garner A, Histopathology of diabetic retinopathy in man.
serve retinal function in type 2 diabetic mice. Nutr Diabetes. Eye (Lond). 1993;7(Pt 2):250–253.
2012;2:e36. 45. Flynn HW, Jr. et al, The Early Treatment Diabetic
32. Sala‐Vila A et al. Dietary marine omega‐3 fatty acids and Retinopathy Study Research Group. Pars plana vitrec
incident sight‐threatening retinopathy in middle‐aged and tomy in the Early Treatment Diabetic Retinopathy Study.
older individuals with type 2 diabetes: prospective investiga ETDRS report number 17. Ophthalmology. 1992;99(9):
tion from the PREDIMED trial. JAMA Ophthalmol. 2016; 1351–1357.
134(10):1142–1149. 46. Storey PP et al. Visual and anatomical outcomes after dia
33. Wang W, Lo ACY. Diabetic retinopathy: pathophysiology betic traction and traction‐rhegmatogenous retinal detach
and treatments. Int J Mol Sci. 2018;19(6). ment repair. Retina. 2018;38(10):1913–1919.
34. Aiello LP et al. Vascular endothelial growth factor in ocular 47. Klein R et al. The epidemiology of retinal vein occlusion: the
fluid of patients with diabetic retinopathy and other retinal Beaver Dam Eye Study. Trans Am Ophthalmol Soc.
disorders. N Engl J Med. 1994;331(22):1480–1487. 2000;98:133–141; discussion 141–143.
35. Rubsam A, Parikh S, Fort PE. Role of inflammation in dia 48. Madsen PH. Rubeosis of the iris and haemorrhagic glau
betic retinopathy. Int J Mol Sci. 2018;19(4). coma in patients with proliferative diabetic retinopathy. Br J
36. Dupas B et al. Association between vessel density and visual Ophthalmol. 1971;55(6):368–371.
acuity in patients with diabetic retinopathy and poorly con 49. Al‐Bahlal A et al. Changing epidemiology of neovascular
trolled type 1 diabetes. JAMA Ophthalmol. 2018;136(7): glaucoma from 2002 to 2012 at King Khaled Eye Specialist
721–728. Hospital, Saudi Arabia. Indian J Ophthalmol. 2017;65(10):
37. The Diabetic Retinopathy Study Research Group. 969–973.
Photocoagulation treatment of proliferative diabetic retin 50. Shazly TA, Latina MA. Neovascular glaucoma: etiology,
opathy. Clinical application of Diabetic Retinopathy Study diagnosis and prognosis. Semin Ophthalmol. 2009;24(2):
(DRS) findings, DRS Report Number 8. Ophthalmology. 113–121.
1981;88(7):583–600. 51. Early Treatment Diabetic Retinopathy Study Research
38. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema. Early
Group. Early photocoagulation for diabetic retinopathy. Treatment Diabetic Retinopathy Study report number 1.
ETDRS report number 9. Ophthalmology. 1991;98(5 Suppl): Arch Ophthalmol. 1985;103(12):1796–1806.
766–785. 52. Early Treatment Diabetic Retinopathy Study design and
39. Sivaprasad S et al. Clinical efficacy of intravitreal aflibercept baseline patient characteristics. ETDRS report number 7.
versus panretinal photocoagulation for best corrected visual Ophthalmology. 1991;98(5 Suppl):741–756.
acuity in patients with proliferative diabetic retinopathy at 53. Baker CW et al. Effect of initial management with
52 weeks (CLARITY): a multicentre, single‐blinded, ran Aflibercept vs laser photocoagulation vs observation on
domised, controlled, phase 2b, non‐inferiority trial. Lancet. vision loss among patients with diabetic macular edema
2017;389(10085):2193–2203. involving the center of the macula and good visual acuity:
40. Gross JG et al. Five‐year outcomes of panretinal photocoag a randomized clinical trial. JAMA. 2019;321(19):1880–
ulation vs intravitreous ranibizumab for proliferative dia 1894.
betic retinopathy: a randomized clinical trial. JAMA 54. Diabetic Retinopathy Clinical Research Network. The
Ophthalmol. 2018;136(10):1138–1148. course of response to focal/grid photocoagulation for dia
41. Wubben TJ, Johnson MW, Anti‐VEGF Treatment betic macular edema. Retina. 2009;29(10):1436–1443.
Interruption Study Group. Anti, anti‐vascular endothelial 55. Writing Committee for the Diabetic Retinopathy Clinical
growth factor therapy for diabetic retinopathy: conse Research Network et al. Comparison of the modified Early
quences of inadvertent treatment interruptions. Am J Treatment Diabetic Retinopathy Study and mild macular
Ophthalmol. 2019;204:13–18. grid laser photocoagulation strategies for diabetic macular
42. American Society of Retina Specialists. Preferences and edema. Arch Ophthalmol. 2007;125(4):469–480.
Trends Survey 2019. 2019. 56. Friberg TR, Infrared micropulsed laser treatment for dia
43. Figueira J et al. Ranibizumab plus panretinal photocoagula betic macular edema – subthreshold versus threshold
tion versus panretinal photocoagulation alone for high‐risk lesions. Semin Ophthalmol. 2001;16(1):19–24.
57. Vujosevic S et al. Microperimetry and fundus autofluores ing individual patient‐level data. BMC Ophthalmol. 2018;
cence in diabetic macular edema: subthreshold micropulse 18(1):340.
diode laser versus modified early treatment diabetic retin 70. Avery RL, Gordon GM. Systemic safety of prolonged
opathy study laser photocoagulation. Retina. 2010;30(6): monthly anti‐vascular endothelial growth factor therapy for
908–916. diabetic macular edema: a systematic review and meta‐anal
58. Diabetic Retinopathy Clinical Research Network. A rand ysis. JAMA Ophthalmol. 2016;134(1):21–29.
omized trial comparing intravitreal triamcinolone acetonide 71. Virgili G et al. Anti‐vascular endothelial growth factor for
and focal/grid photocoagulation for diabetic macular edema. diabetic macular oedema. Cochrane Database Syst Rev.
Ophthalmology. 2008;115(9):1447–1449, 1449 e1–10. 2014(10):CD007419.
59. Boyer DS et al. Three‐year, randomized, sham‐controlled 72. Maloney MH et al. Risk of systemic adverse events associ
trial of dexamethasone intravitreal implant in patients with ated with intravitreal anti‐VEGF therapy for diabetic macu
diabetic macular edema. Ophthalmology. 2014;121(10): lar edema in routine clinical practice. Ophthalmology.
1904–1914. 2019;126(7):1007–1015.
60. Maturi RK et al. Effect of adding dexamethasone to contin 73. Diabetic Retinopathy Clinical Research Network Writing
ued ranibizumab treatment in patients with persistent dia Committee et al. Vitrectomy outcomes in eyes with diabetic
betic macular edema: A DRCR network phase 2 randomized macular edema and vitreomacular traction. Ophthalmology.
clinical trial. JAMA Ophthalmol. 2018;136(1):29–38. 2010;117(6):1087–1093.e3.
61. Fraser‐Bell S et al. Bevacizumab or dexamethasone implants 74. Flaxel CJ et al. Factors associated with visual acuity out
for DME: 2‐year results (the BEVORDEX study). comes after vitrectomy for diabetic macular edema: diabetic
Ophthalmology. 2016;123(6):1399–1401. retinopathy clinical research network. Retina. 2010;30(9):
62. Campochiaro PA et al. Long‐term benefit of sustained‐deliv 1488–1495.
ery fluocinolone acetonide vitreous inserts for diabetic mac 75. Yanyali A et al. Modified grid laser photocoagulation versus
ular edema. Ophthalmology. 2011;118(4):626–635 e2. pars plana vitrectomy with internal limiting membrane removal
63. Diabetic Retinopathy Clinical Research Network et al. in diabetic macular edema. Am J Ophthalmol. 2005;139(5):
Randomized trial evaluating ranibizumab plus prompt or 795–801.
deferred laser or triamcinolone plus prompt laser for dia 76. Cho M, D’Amico DJ. Transconjunctival 25‐gauge pars
betic macular edema. Ophthalmology. 2010;117(6):1064– plana vitrectomy and internal limiting membrane peeling
1077 e35. for chronic macular edema. Clin Ophthalmol. 2012;6:
64. Parikh R et al. Trends of anti‐vascular endothelial growth 981–989.
factor use in ophthalmology among privately insured and 77. Jackson TL et al. Pars plana vitrectomy for diabetic macular
medicare advantage patients. Ophthalmology. 2017;124(3): edema: a systematic review, meta‐analysis, and synthesis of
352–358. safety literature. Retina. 2017;37(5):886–895.
65. Elman MJ et al. Intravitreal ranibizumab for diabetic macu 78. Campochiaro PA et al. The port delivery system with ranibi
lar edema with prompt versus deferred laser treatment: 5‐ zumab for neovascular age‐related macular degeneration:
year randomized trial results. Ophthalmology. 2015;122(2): results from the randomized phase 2 ladder clinical trial.
375–381. Ophthalmology. 2019;126(8):1141–1154.
66. Brown DM et al. Long‐term outcomes of ranibizumab ther 79. Grishanin R et al. Preclinical evaluation of ADVM‐022, a
apy for diabetic macular edema: the 36‐month results from novel gene therapy approach to treating wet age‐related
two phase III trials: RISE and RIDE. Ophthalmology. macular degeneration. Mol Ther. 2019;27(1):118–129.
2013;120(10):2013–2022. 80. Liu HA et al. A lipid nanoparticle system improves
67. Korobelnik JF et al. Intravitreal aflibercept for diabetic mac siRNA efficacy in RPE cells and a laser‐induced murine
ular edema. Ophthalmology. 2014;121(11):2247–2254. CNV model. Invest Ophthalmol Vis Sci. 2011;52(7):4789–
68. Wells JA et al. Aflibercept, bevacizumab, or ranibizumab for 4794.
diabetic macular edema: two‐year results from a compara 81. Luo L et al. Targeted intraceptor nanoparticle therapy
tive effectiveness randomized clinical trial. Ophthalmology. reduces angiogenesis and fibrosis in primate and murine
2016;123(6):1351–1359. macular degeneration. ACS Nano. 2013;7(4):3264–3275.
69. Muston D et al. An efficacy comparison of anti‐vascular 82. Eells J et al. 670 nm photobiomodulation as a therapy for
growth factor agents and laser photocoagulation in diabetic diabetic macular edema: a pilot study. Investigative
macular edema: a network meta‐analysis incorporat Ophthalmology & Visual Science. 2017;58(8):932–932.
83. Tang J, Herda AA, Kern TS. Photobiomodulation in the 93. O’Brart DP, Muir MG, Marshall J. Phototherapeutic kera
treatment of patients with non‐center‐involving diabetic tectomy for recurrent corneal erosions. Eye (Lond). 1994;
macular oedema. Br J Ophthalmol. 2014;98(8):1013–1015. 8(Pt 4):378–383.
84. Geneva, II. Photobiomodulation for the treatment of reti 94. Becker C et al. Cataract in patients with diabetes mellitus‐
nal diseases: a review. Int J Ophthalmol. 2016;9(1): incidence rates in the UK and risk factors. Eye (Lond).
145–152. 2018;32(6):1028–1035.
85. Herse PR. A review of manifestations of diabetes mellitus in 95. Klein BE, Klein R, Moss SE. Prevalence of cataracts in a
the anterior eye and cornea. Am J Optom Physiol Opt. 1988; population‐based study of persons with diabetes mellitus.
65(3):224–230. Ophthalmology. 1985;92(9):1191–1196.
86. Dogru M, Katakami C, Inoue M. Tear function and ocular 96. Denniston AK et al. The UK Diabetic Retinopathy
surface changes in noninsulin‐dependent diabetes mellitus. Electronic Medical Record (UK DR EMR) Users Group,
Ophthalmology. 2001;108(3):586–592. Report 2: real‐world data for the impact of cataract sur
87. Messmer EM et al. in vivo; confocal microscopy of corneal gery on diabetic macular oedema. Br J Ophthalmol.
small fiber damage in diabetes mellitus. Graefes Arch Clin 2017;101(12):1673–1678.
Exp Ophthalmol. 2010;248(9):1307–1312. 97. Chew EY et al. Results after lens extraction in patients with
88. Zhang X et al. Dry eye syndrome in patients with diabetes diabetic retinopathy: early treatment diabetic retinopathy
mellitus: prevalence, etiology, and clinical characteristics. J study report number 25. Arch Ophthalmol. 1999;117(12):
Ophthalmol. 2016; 2016:8201053. 1600–1606.
89. Saini JS, Mittal S. in vivo; assessment of corneal endothelial 98. Squirrell D et al. A prospective, case controlled study of the
function in diabetes mellitus. Arch Ophthalmol. 1996;114(6): natural history of diabetic retinopathy and maculopathy
649–653. after uncomplicated phacoemulsification cataract surgery
90. Su DH et al. Diabetes, hyperglycemia, and central corneal in patients with type 2 diabetes. Br J Ophthalmol.
thickness: the Singapore Malay Eye Study. Ophthalmology. 2002;86(5):565–571.
2008;115(6):964–968 e1. 99. Wagner T et al. Influence of cataract surgery on the dia
91. Foulks GN et al. Factors related to corneal epithelial compli betic eye: a prospective study. Ger J Ophthalmol. 1996;5(2):
cations after closed vitrectomy in diabetics. Arch Ophthalmol. 79–83.
1979;97(6):1076–1078. 100. Dokmetas HS et al. Diabetic ketoacidosis and rhino‐orbital
92. Cormier G et al. Anterior stromal punctures for bullous mucormycosis. Diabetes Res Clin Pract. 2002;57(2):
keratopathy. Arch Ophthalmol. 1996;114(6):654–658. 139–142.
22 Upper Gastrointestinal Manifestations
of Diabetes
Michael Camilleri
Clinical Enteric Neuroscience Translational and Epidemiological Research Program, Division of Gastroenterology and
Hepatology, Mayo Clinic, Rochester, MN, USA
LE A R NI NG P OI NT S testinal dysfunction [4]; however, people with T2DM and

men with T1DM used laxatives more frequently than con
●● Gastrointestinal manifestations of diabetes typically
trols. Moreover, that study and a Finnish population‐based
accompany other microvascular complications. study reported that people with T1DM had a lower preva
●● Gastroparesis is associated with increased mortality in lence of heartburn [5].
diabetes. In contrast, a study from Australia reported the preva
Nutritional support is a key part of the management
●●
lence of several upper and lower gastrointestinal symptoms
of gastroparesis.
was higher in 423 patients with predominantly (95%)
T2DM than in controls [6].
In Olmsted County, Minnesota, the age‐adjusted inci
dence per 100 000 person‐years of definite gastroparesis
Introduction
for the years 1996–2006 was 2.4 (95% confidence interval
Although diabetes can affect the entire gastrointestinal (CI) 1.2–3.8) for men and 9.8 (95% CI 7.5–12.1) for
tract, most of the impact results from upper gastrointesti women. The age‐adjusted prevalence of definite gastropa
nal dysfunction, which may affect quality of life, impair resis per 100 000 persons on January 1, 2007, was 9.6 (95%
nutrition, and affect glycemic control. CI 1.8–17.4) for men and 37.8 (95% CI 23.3–52.4) for
women [7].
Diabetic gastroparesis may cause severe symptoms and
Epidemiology
result in nutritional compromise, impaired glucose control
Studies in selected patient groups, often from tertiary refer and a poor quality of life, independently of other factors
ral centers, suggest that gastrointestinal symptoms are com such as age, tobacco use, alcohol use or type of diabetes [8].
mon in diabetes mellitus [1, 2]. In an epidemiological study Studies of the natural history of gastroparesis have been
from Olmsted County, Minnesota, [3] the prevalence of gas limited by relatively small numbers of patients, potential
trointestinal symptoms (i.e., nausea and/or vomiting, dys referral bias or short follow‐up periods. The data suggest
pepsia, heartburn, irritable bowel syndrome, constipation, that gastric emptying and its symptoms are generally stable
and fecal incontinence) was not significantly different during 12–25 years of follow‐up or more [9, 10]. In a study
between individuals with either type 1 (T1DM) or type 2 of 86 patients with diabetes who were followed for at least 9
(T2DM) diabetes mellitus and age‐matched controls in years, gastroparesis was not associated with mortality after
whom it is known that there is a high background popula adjustment for other disorders [11]. Among patients with
tion prevalence (~20%) of symptoms suggestive of gastroin gastroparesis, overall survival was significantly lower than

269
the age‐ and sex‐specific expected survival computed from unmyelinated fibers, without much neuronal loss [18, 19].
the Minnesota white population [7]. The most common The loss of nerve fibers is often multifocal, suggestive of
causes of death were cardiovascular disease (24.6%), res ischemic injury. Within the enteric nervous system,
piratory failure (23.2%), malignancy (15.9%), chronic renal reduced neuronal staining and, to a lesser extent, neuronal
failure in 11 (15.9%), cerebrovascular accident (10.1%), loss, particularly inhibitory neurons expressing nitric oxide
and other causes (10.1%). synthase (NOS), have been described in several animal
Delayed radionuclide gastric emptying predicts morbid models of diabetes [20]. In theory, this reduction in nitrer
ity in patients with diabetes with gastroparesis symptoms, gic inhibitory functions may contribute to impaired gastric
including hospitalizations and days of hospitalizations, accommodation and accelerated intestinal transit in diabe
emergency department and office visits [12]. Over one tes. Loss of NOS may impair pyloric relaxation and thereby
decade (1996–2006), gastroparesis was associated with retard gastric emptying. Loss of intestinal cells of Cajal
increased mortality [7], although the cause of death was (ICCs), documented in several animal models and case
not gastrointestinal and the gastroparesis was simply a reports of diabetes, may also contribute to gut dysmotil
marker of the severity of the diabetes and the associated ity [15, 20].
complications, the classical triopathy of nephropathy, neu Several mechanisms, including apoptosis, oxidative
ropathy and retinopathy [13]. More recent studies confirm stress (possibly associated with reduced CD 206 positive
that delayed gastric emptying is associated with retinopa M2 macrophages), advanced glycation end products and
thy and a number of complications of diabetes; however, neuroimmune mechanisms may be responsible for neu
with so many patients having type 2 diabetes rather than ronal loss and gut dysmotility [16]. The loss of ICCs has
insulin‐dependent and ketosis‐prone type 1 diabetes, 39% been attributed to a reduction in heme‐oxygenase
of diabetic patients with gastroparesis in the NIH (HO‐1) and other protective mechanisms against hyper
Gastroparesis Consortium database did not have any dia glycemia [15].
betic complications [14].
Pathophysiology of gastric dysfunctions
in humans with diabetes
Mechanisms and pathophysiology
Diabetes is associated with accelerated or delayed gastric
underpinning upper gastrointestinal
emptying, increased and reduced gastric sensation, and
symptoms in diabetes mellitus
impaired gastric accommodation (Figure 22.1). A vagal
Gastrointestinal dysmotility in diabetes is caused by extrin neuropathy can cause antral hypomotility and/or pyloros
sic (i.e., sympathetic and parasympathetic) neural dysfunc pasm, which may delay gastric emptying [21]. Among 108
tion, intrinsic or enteric nervous system (excitatory and patients with diabetes who underwent tests to evaluate
inhibitory neurons) or pacemaker dysfunction (interstitial both gastric emptying of solids and gastric accommodation
cells of Cajal or PDGFRα‐positive fibroblast‐like cells), using SPECT imaging, there were roughly equal propor
and hyperglycemia [15]. Neural dysfunction has been tions of patients with abnormal gastric emptying, abnor
attributed to several mechanisms (e.g., oxidative stress) [16] mal gastric accommodation, and both abnormal or neither
due to loss of anti‐inflammatory macrophages and abnormal; the latter presumably reflecting gastroduodenal
increased expression of genes associated with pro‐inflam hypersensitivity [22]. Gastric emptying may be accelerated
matory macrophages that have been reported in full‐thick in diabetic patients with upper gastrointestinal symp
ness gastric biopsies from patients with gastroparesis [17]. toms [23]. The pathophysiology of rapid gastric emptying
in diabetes may result from impaired gastric accommoda
Pathophysiology of diabetic gastroparesis: tion resulting from a vagal neuropathy [24], thereby
insights from animal studies increasing gastric pressure and accelerating gastric empty
In animal models, extrinsic neural dysfunction has been ing of liquids. Some patients with diabetes and gastropare
primarily implicated to a loss of myelinated and sis also have small intestinal dysmotility, more frequently
Liquids or solids in stomach (% total consumed)
100
Gastroparesis
Gastric emptying
of solids
Gastric
emptying of
50 homogenized
solids
Gastric
emptying
of liquids
0 1 2 3 4 6
Hours
0h 2h 4h
normal
0% 57% 100% emptied
slow 0h 2h 4h
0% 20% 71% emptied
0h 1h 2h
fast
0% 52% 98% emptied
FIG 22.1 Assessment of gastric emptying by scintigraphy. Normally liquids are emptied in an exponential manner while solid
emptying is characterized by an initial lag phase, followed by a more rapid, linear emptying (left panel). The lag phase corresponds
to the time required for antral trituration and gastric accommodation, during which approximately 10% of solids are emptied.
Representative examples of normal, delayed and accelerated gastric emptying respectively of an egg meal, labeled with 99mTc, in
patients with diabetes are shown in the right panel. At time 0 (i.e. first image in each panel), the entire meal was in the stomach.
Thereafter, the normal ranges for gastric emptying are 4.4–35% at 1 hour, 25–78.5% at 2 hours and 76.2–100% at 4 hours.
characterized by reduced than by increased motility [25]. Whereas postprandial symptoms without delayed gas
Small bowel dysmotility may also contribute to gastric tric emptying were previously called diabetic dyspepsia in
stasis. which vomiting was less prevalent, appreciation that the
Whereas acute hyperglycemia delays gastric emptying pathophysiology includes delayed gastric emptying,
in healthy subjects and in patients with T1DM [26–29], impaired gastric accommodation and abnormal sensation,
the clinical significance of hyperglycemia in terms of the term “gastroparesis” has been extended to include
symptoms and morbidity, or the reversibility of patients with normal gastric emptying and a significant
delayed emptying by optimizing control of glycemia are component of upper abdominal pain, which is associated
the topic of continuing research without unequivocal with opioid use and impaired quality of life [34–37].
answers [30, 31]. Severely delayed gastric emptying of less than 65% at 4
hours reflects a significant delay often associated with
nutritional consequences, the need for nutritional supple
Upper gastrointestinal manifestations
mentation, jejunal feeding or gastric decompression [38].
associated with diabetes
Patients with diabetic gastroparesis frequently have
Dysphagia and heartburn long‐standing T1DM with other microvascular complica
Clinical features tions including retinopathy, nephropathy, peripheral neu
Esophageal dysmotility is common, may cause dysphagia, ropathy and other forms of autonomic dysfunction [13].
and may be related to cardiovascular autonomic neuropa These can present as abnormal pupillary responses, anhi
thy in diabetes mellitus [32]. The amplitude of peristaltic drosis, gustatory sweating, orthostatic hypotension, impo
contractions and basal lower esophageal sphincter pres tence, retrograde ejaculation and dysfunction of the
sures are generally normal, and dysphagia is mostly associ urinary bladder. In patients with diabetes, delayed gastric
ated with incoordinated contractions. Symptoms of emptying is associated with retinopathy and a number of
gastroesophageal reflux are also common, particularly in complications of diabetes [14].
patients with impaired gastric emptying who have vomit Clinical clues to the presence of gastroparesis include
ing. Rarely, recurrent vomiting may lead to Mallory–Weiss vomiting of undigested food eaten several hours previ
tears and bleeding. ously, weight loss, gastric succussion splash or features of
an autonomic neuropathy.
Diagnosis
Dysphagia and heartburn should prompt upper gastroin Diagnosis
testinal endoscopy to exclude reflux and other incidental In patients with upper gastrointestinal symptoms, an upper
mucosal diseases, such as candidiasis and neoplasms. gastrointestinal endoscopy is necessary to exclude peptic
Because of the high prevalence of coronary atherosclerosis ulcer disease and neoplasms, either of which can cause gas
in diabetes, testing for coronary artery disease should be tric outlet obstruction. Upper endoscopy may reveal gas
considered when necessary in patients with chest pain. tric bezoars, which suggest antral hypomotility. Metabolic
derangements such as diabetic ketoacidosis or uremia, and
Gastroparesis and other upper gastrointestinal medications, particularly opiates, calcium‐channel block
symptoms ers and anticholinergic agents, may contribute to dysmotil
Clinical features ity. Rarely, patients with gastroparesis present with
Although gastroparesis refers to a syndrome characterized retrosternal or epigastric pain, and cardiac, biliary or pan
by symptoms of nausea, vomiting, early satiation after creatic disease may be considered.
meals, impaired nutrition and objective evidence of mark Upper gastrointestinal endoscopy or barium meal is
edly delayed gastric emptying, gastric retention may be essential to exclude mechanical obstruction of the stomach
asymptomatic [14], perhaps because of the afferent dys outflow. Barium x‐rays of the small intestine or enterogra
function associated with vagal denervation [33]. phy with computed tomography or MRI should be consid
Upper Gastrointestinal Manifestations of Diabetes 273
ered only when the clinical features raise the possibility of Abdominal Pain
small intestinal obstruction. Patients with diabetes are obviously susceptible to the usual
Whereas, autonomic function testing is useful [39] to causes of abdominal pain seen in the general population.
indicate the underlying severity of neuropathy, delayed There is an increased prevalence of gallstones because of
gastric emptying can be documented by scintigraphy or altered gallbladder contractility and of mesenteric ischemia
stable isotope breath tests for solids, by ultrasonography for caused by generalized atherosclerosis; however, there is lit
liquids, and by the presence of a large amount of retained tle evidence that altered gallbladder contractility per se
food in the stomach. The proportion of radioisotope (i.e., in the absence of gallstones) causes symptoms.
retained in the stomach at 2 and 4 hours distinguishes nor Thoracolumbar radiculopathy may result in pain in a gir
mal function from delayed gastric emptying with a sensi dle‐like distribution which does not cross the midline.
tivity of 90% and a specificity of 70% [40]. The importance Specific tests are indicated in the clinical features of pain
of obtaining scans at 4 hours after a meal cannot be over suggest these disorders. It is essential to elicit a careful his
emphasized. Because gastric emptying is slow initially, it is tory. Among 583 patients in the NIH Gastroparesis
not accurate to extrapolate emptying from scans taken for Consortium database, 41% were taking opioids which
a shorter duration. Another useful test for measuring solid included 33% on potent morphine‐like agents, and the
phase gastric emptying utilizes a standardized meal with indication was abdominal pain for 61% of patients taking
biscuit enriched with 13C, a substrate containing the stable opioids. These patients have higher symptoms scores,
isotope. When metabolized, the proteins, carbohydrates greater levels of gastric retention, worse quality of life,
and lipids of the S. platensis or the medium chain triglycer increased hospitalization, and increased use of antiemetic
ide, octanoate, give rise to respiratory CO2 that is enriched and pain modulator medications compared with nonusers
in 13C. Measurement of 13CO2 breath content (a reflection or those using weaker opioids (e.g., tramadol, tapentadol,
of the amount of biscuit remaining in the stomach) by iso codeine, or propoxyphene) [34].
tope ratio mass spectrometry allows an estimation of gas
tric emptying t1/2 [41]. Antropyloroduodenal manometry
Management of gastroparesis
and EndoFLIP are specialized techniques that assess pres
and dyspepsia
sure profiles in the stomach and small bowel, as well as the
dimensions and distensibility of the pylorus. They may The principles of management are to address fluid and
play a greater role in the future to guide management, such nutritional requirements, improve glycemic control and
as selection of prokinetics versus performance of G‐POEM treat symptoms (Table 22.1 [45]).
procedure.
Gastric accommodation in response to meal ingestion Nutritional support
may be impaired in diabetes [24]. This may contribute to Because both liquids and homogenized solids are more
the gastrointestinal symptoms of nausea, bloating and early readily emptied from the stomach than solids, food that is
satiety. Imaging of the stomach wall using SPECT and i.v. liquid or blenderized is better tolerated. A randomized,
99m
Tc pertechnetate allows measurement of gastric volume controlled trial has shown that small particle size meals are
after meal ingestion. Other centers use an intragastric associated with reduced symptoms in gastroparesis [46].
barostat or intraluminal high resolution manometry to While nutritional requirements and symptoms can be
measure gastric accommodation [42, 43]. However, a addressed to a variable extent in patients with mild and
nutrient drink test is a surrogate for gastric accommoda compensated gastroparesis, patients with severe gastropa
tion and increased sensation [44]; thus, a maximum toler resis often require hospitalization for one or more of the
ated volume of less than about 750 kilocalories is following measures: intravenous hydration and correction
significantly correlated with impairment of gastric accom of metabolic derangements (ketoacidosis, uremia, hypo‐/
modation measured by an intragastric balloon and could hyperglycemia), nasoenteric decompression, and/or
serve as a biomarker for accommodation and sensation. enteral nutrition to manage vomiting and nutritional
TABLE 22.1 Treatment strategies for patients with gastroparesis (reproduced from [45]).
Mild gastroparesis (10–15% Moderate gastroparesis (15–35% Severe gastroparesis (> 35%

Management strategies 4 h gastric retention) 4 h gastric retention) 4 h gastric retention)
General measures Review and eliminate medications

inhibiting motility and optimize
glycaemic control in patients with
diabetes
Diet • Small, frequent meals • Small, frequent meals • Blenderized diet
• Low‐fat low‐fibre diet • Low‐fat low‐fibre diet • Routine use of liquid nutrient
• Small‐particle diet when • Small‐particle diet when supplements
symptomatic symptomatic
Nutritional support Rarely needed • Caloric liquids • Caloric liquids
• Orally • Orally
• Rarely requires nutrition by PEJ • May require nutrition by PEJ
feeding tube feeding tube
Pharmacological: prokinetic Metoclopramide Metoclopramide • Metoclopramide or domperidone
• Erythromycin
• Prucalopride
Pharmacological: anti‐emetic Promethazine or prochlorperazine • Promethazine or prochlorperazine • Ondansetron
• Ondansetron • Aprepitant or mirtazapine
Pharmacological: symptom Not needed Not needed Nortriptyline
modulators
Non‐pharmacological Not needed Not needed • Gastrostomy tube decompression
• Laparoscopic and endoscopic
interventions
c22.indd 274 09/24/2021 11:55:48

requirements [38]. Parenteral nutrition may become neces 5–10 mg t.i.d., administered 30 minutes before meals as
sary in cases of malnutrition. Bezoars may be mechanically a liquid formulation, for a short duration; the higher
disrupted during endoscopy, followed by gastric decom dose is 20 mg t.i.d. A systematic review of trials con
pression to drain residual non‐digestible particles. cluded there was limited evidence to support the use of
For patients with severe gastroparesis, it may be neces domperidone, which is another dopaminergic antagonist
sary to bypass the stomach with a jejunal feeding tube. This not approved for use in the USA [51].
procedure should be preceded by a trial of nasojejunal 3 Experimental prokinetics in development Potentially,
feeding for a few days with infusion rates of at least 60 mL effective pharmacological approaches are directed at the
iso‐osmolar nutrient per hour. It is preferable to place jeju "traditional" pathways, specifically, 5‐HT4, dopamine D2
nal feeding tubes directly into the jejunum either by endos and D3, and NK1 receptors, or novel mechanisms (dis
copy or, if necessary, by laparoscopy, rather than via cussed in the next section).
percutaneous endoscopic gastrostomy tubes. Such tubes
i 5‐HT4 receptor agonists In a randomized, placebo‐con
allow restoration of normal nutritional status, but they are
trolled, cross‐over study of 4 weeks duration with a 2‐
associated with adverse effects. There is no evidence to
week washout, prucalopride improved symptom
suggest that gastrectomy relieves symptoms or enhances
control as well as gastric emptying in 28 patients with
quality of life. Patients with gastroparesis often have con
idiopathic and 6 patients with diabetic gastropare
comitant small intestinal denervation which is likely to
sis [52]. Similarly, velusetrag, was efficacious in the
cause persistent symptoms after gastrectomy [24, 47] or
treatment of diabetic and idiopathic gastropare
difficulties with tolerance of jejunal feeding.
sis [53]. Among "new generation" 5‐HT4 agonists, pru
calopride, velusetrag, and naronapride are selective for
Prokinetics
5‐HT4 receptors without hERG effects [54, 55]. Results
1 Erythromycin at a dose of 3 mg/kg body weight intrave of the pharmacodynamics effects of a new specific 5‐
nously every 8 hours can accelerate gastric emptying [48, HT4 receptor agonist, TAK‐954 (TD‐8954), are keenly
49]. When oral intake is resumed, treatment with oral awaited (NCT03281577). Velusetrag and TAK‐954 had
250 mg erythromycin t.i.d. for 1–2 weeks is worthwhile. no significant effects on canine, porcine, and human
Thereafter, the prokinetic effects of erythromycin are coronary artery tone, human platelet aggregation,
limited by tachyphylaxis. Anecdotal findings suggest hERG potassium channel conductance, or off‐target
that erythromycin may be effective, if courses are sepa actions [56].
rated by a drug‐free period (e.g., lasting 2 weeks). ii Targeting sensations with D2/D3 and NK1 antagonists
Hyperglycemia interferes with the prokinetic effect of For increased gastric sensation, the dopamine D2/3
intravenous erythromycin on gastric emptying in antagonist, TAK‐506, significantly increased the vol
healthy subjects and in patients with diabetes [50]. If the ume to fullness compared to baseline with 1 week of
patient remains symptomatic, other prokinetic agents treatment [57]. The NK1 receptor antagonist, aprepi
may be considered as adjuncts. tant, improved multiple symptoms of gastroparesis
2 Metoclopramide is the only available medication in the including nausea [58], and it has been shown to
USA; it is a peripheral cholinergic and antidopaminergic enhance gastric accommodation rather than affect
agent. During acute administration, it initially enhances gastric emptying in healthy controls [59]. Similarly,
gastric emptying of liquids in patients with diabetic gas tradipitant, a novel NK1 receptor antagonist, improved
troparesis, but its symptomatic efficacy is probably nausea and other symptoms of gastroparesis in a 4‐
related to its central antiemetic effects. Its long‐term use week, randomized, controlled trial [60].
is restricted by a decline in efficacy and by a troubling iii Ghrelin receptor agonist Ghrelin is a 28‐amino acid
incidence of central nervous system side effects. orexigenic hormone found mainly in the stomach.
Therefore, the author prefers to prescribe a dose of Administration of a pharmacological dose of ghrelin
induced premature phase III of the migrating motor remained significant [76]. Second, in a large, multicenter,
complex, increased proximal gastric tone through randomized, double‐blind trial with cross‐over from
central and peripheral sites of action [61, 62], and France performed in 172 patients (66% women; mean age
accelerated gastric emptying in some studies of 45±12 years; 133 with gastroparesis) with chronic
patients with gastroparesis (reviewed in [63]). (> 12 months) refractory vomiting [77], the vomiting
Relamorelin, a pentapeptide ghrelin receptor agonist, scores improved significantly when the device was on in
has potent prokinetic effects estimated to be 15‐ to patients with delayed or normal gastric emptying. However,
130‐fold more potent than natural ghrelin [64]. gastric emptying and quality of life were not different.
Relamorelin, 100 mg subcutaneous (s.c.), accelerated Gastric per‐oral endoscopic myotomy (G‐POEM): Pyloric
gastric emptying of solids in patients with prior docu dysfunction was first described in diabetic gastroparesis in
mentation of delayed gastric emptying and either type 1986 [78]. Multiple studies have demonstrated efficacy in
1 or type 2 diabetes mellitus [65, 66], and it increased relief of symptoms and improvement in gastric emptying
the frequency of distal antral contractions without in patients with gastroparesis who undergo G‐POEM, as
inhibiting gastric accommodation or inducing satia summarized in a recent systematic review and meta‐
tion in healthy volunteers [67]. Relamorelin has analysis [79].
proven clinical efficacy and safety in phase 2A and 2B,
randomized, controlled trials in patients with diabetic Conclusion
gastroparesis [68–70].
Diabetic gastroparesis is a significant problem in
Endoscopic and device therapies patients with other complications of diabetes including
Endoscopic injection of botulinum toxin into the pylorus was retinopathy and triopathy. Patient management requires
not effective in controlled studies primarily of patients with a multi‐disciplinary approach, including coordination
idiopathic gastroparesis [71, 72], There have been multiple, between diabetologist and gastroenterologist, as well as
open‐label trials of intra‐pyloric botulinum toxin injec support from nutrition experts. There are promising
tions attesting to the efficacy of this approach [73–75], but new medications for the treatment of upper gastrointes
there have been only two randomized, controlled trials [71, tinal symptoms in patients with diabetes, and potential
72], both of which failed to show any improvement in endoscopic treatment with pyloromyotomy. However,
symptoms. However, at least in the trial performed by clinical trials are still required for the promising new
Friedenberg and colleagues [71], botulinum neurotoxin medications, and sham‐controlled trials for pyloromy
improved gastric emptying compared to placebo, suggest otomy. Future research is required to identify optimal,
ing that the intervention can result in the desired physio individualized therapy in patients with diabetes and
logical effect, but not the clinical outcome. gastroparesis.
Gastric electrical stimulation is approved as a humanitar
ian use device for refractory gastroparesis with efficacy References
possibly greater in diabetic gastroparesis; its use for this
indication is controversial. Two recent studies argue in 1. Feldman M, Schiller ER. Disorders of gastrointestinal
favor of beneficial effects of gastric electrical stimulation. motility associated with diabetes mellitus. Ann Intern Med.
1983;98:378–384.
First, in a study from the Gastroparesis Clinical Research
2. Clouse RE, Lustman PJ. Gastrointestinal symptoms in
Consortium (GpCRC), 238 patients without gastric electri
diabetic patients: lack of association with neuropathy. Am J
cal stimulation and 81 with gastric electrical stimulation Gastroenterol. 1989;84:868–872.
showed improvement by ≥ 1 point and change from enroll 3. Dyck PJ, Karnes JL, O’Brien PC et al. The Rochester
ment with gastric electrical stimulation compared to the Diabetic Neuropathy Study: reassessment of tests and
group without the device. However, when adjusting for criteria for diagnosis and staged severity. Neurology.
patient characteristics, only the symptom of nausea 1992;42:1164–1170.
4. Maleki D, Locke GR 3rd, Camilleri M et al. Gastrointestinal 19. Sinnreich M, Taylor BV, Dyck PJ. Diabetic neuropathies:
tract symptoms among persons with diabetes mellitus in the classification, clinical features, and pathophysiological basis.
community. Arch Intern Med. 2000;160:2808–2816. Neurologist. 2005;11:63–79.
5. Janatuinen E, Pikkarainen P, Laakso M et al. Gastrointestinal 20. Vittal H, Farrugia G, Gomez G et al. Mechanisms of disease:
symptoms in middle‐aged diabetic patients. Scand J the pathological basis of gastroparesis – a review of experi
Gastroenterol. 1993;28:427–432. mental and clinical studies. Nat Clin Pract Gastroenterol
6. Bytzer P, Talley NJ, Leemon M et al. Prevalence of gastroin Hepatol. 2007;4:336–346.
testinal symptoms associated with diabetes mellitus: a popu 21. Feldman M, Corbett DB, Ramsey EJ et al. Abnormal gastric
lation‐based survey of 15000 adults. Arch Intern Med. function in lonstanding, insulin‐dependent diabetic patients.
2001;161:1989–1996. Gastroenterology. 1979;77:12–17.
7. Jung HK, Choung RS, Locke GR 3rd et al. The incidence, 22. Chedid V, Brandler J, Vijayvargiya P et al. Characterization
prevalence, and outcomes of patients with gastroparesis in of upper gastrointestinal symptoms, gastric motor functions
Olmsted County, Minnesota, from 1996 to 2006. and associations in patients with diabetes at a referral center.
Gastroenterology. 2009;136:1225–1233. Am J Gastroenterol. 2019;114:143–154.
8. Talley NJ, Young L, Bytzer P et al. Impact of chronic gastro 23. Bredenoord AJ, Chial HJ, Camilleri M et al. Gastric accom
intestinal symptoms in diabetes mellitus on health‐related modation and emptying in evaluation of patients with upper
quality of life. Am J Gastroenterol. 2001;96:71–76. gastrointestinal symptoms. Clin Gastroenterol Hepatol.
9. Jones KL, Russo A, Berry MK et al. A longitudinal study of 2003;1:264–272.
gastric emptying and upper gastrointestinal symptoms in 24. Samsom M, Salet GA, Roelofs JM et al. Compliance of the
patients with diabetes mellitus [see Comment]. Am J Med. proximal stomach and dyspeptic symptoms in patients with
2002;113:449–455. type 1 diabetes mellitus. Dig Dis Sci. 1995;40:2037–2042.
10. Chang J, Russo A, Bound M et al. A 25‐year longitudinal 25. Camilleri M, Malagelada JR. Abnormal intestinal motility in
evaluation of gastric emptying in diabetes. Diabetes Care. diabetics with gastroparesis. Eur J Clin Invest. 1984;14:
2012;35:2594–2596. 420–427.
11. Kong MF, Horowitz M, Jones KL et al. Natural history of dia 26. MacGregor IL, Gueller R, Watts HD et al. The effect of acute
betic gastroparesis. Diabetes Care. 1999;22:503–507. hyperglycemia on gastric emptying in man. Gastroenterology.
12. Hyett B, Martinez FJ, Gill BM et al. Delayed radionucleotide 1976;70:190–196.
gastric emptying studies predict morbidity in diabetics with 27. Fraser RJ, Horowitz M, Maddox AF et al. Hyperglycaemia
symptoms of gastroparesis. Gastroenterology. 2009;137: slows gastric emptying in type 1 (insulin‐dependent) diabe
445–452. tes mellitus. Diabetologia. 1990;33:675–680.
13. Kassander P. Asymptomatic gastric retention in diabetics 28. Oster‐Jorgensen E, Pedersen SA, Larsen ML. The influence
(gastroparesis diabeticorum). Ann Intern Med. 1958;48: of induced hyperglycaemia on gastric emptying rate in
797–812. healthy humans. Scand J Clin Lab Invest. 1990;50:831–836.
14. Parkman HP, Wilson LA, Farrugia G et al. Delayed gastric 29. Schvarcz E, Palmer M, Aman J et al. Physiological hypergly
emptying associates with diabetic complications in diabetic cemia slows gastric emptying in normal subjects and patients
patients with symptoms of gastroparesis. Am J Gastroenterol. with insulin‐dependent diabetes mellitus. Gastroenterology.
2019;114:1778–1794. 1997;113:60–66.
15. Ordog T. Interstitial cells of Cajal in diabetic gastroenteropa 30. Bharucha AE, Kudva Y, Basu A et al. Relationship between
thy. Neurogastroenterol Motil. 2008;20:8–18. glycemic control and gastric emptying in poorly controlled
16. Chandrasekharan B, Srinivasan S. Diabetes and the enteric type 2 diabetes. Clin Gastroenterol Hepatol. 2015;13:466–476.
nervous system. Neurogastroenterol Motil. 2007;19:951–960. 31. Parthasarathy G, Kudva YC, Low PA et al. Relationship
17. Grover M, Bernard CE, Pasricha PJ et al. Diabetic and idio between gastric emptying and diurnal glycemic control in
pathic gastroparesis is associated with loss of CD206‐posi type 1 diabetes mellitus: A randomized trial. J Clin Endocrinol
tive macrophages in the gastric antrum. Neurogastroenterol Metab. 2017;102:398–406.
Motil. 2017;29(6). doi: 10.1111/nmo.13018. 32. Ascaso JF, Herreros B, Sanchiz V et al. Oesophageal motility
18. Schmidt RE. Neuropathology and pathogenesis of dia disorders in type 1 diabetes mellitus and their relation to car
betic autonomic neuropathy. Int Rev Neurobiol. 2002;50: diovascular autonomic neuropathy. Neurogastroenterol Motil.
257–292. 2006;18:813–822.
33. Rathmann W, Enck P, Frieling T et al. Visceral afferent neu with diabetic gastroparesis: a randomized controlled trial.
ropathy in diabetic gastroparesis. Diabetes Care. 1991;14: Am J Gastroenterol. 2014;109:375–385.
1086–1089. 47. Jones MP, Maganti K. A systematic review of surgical therapy
34. Hasler WL, Wilson LA, Nguyen LA et al. Opioid use and for gastroparesis. Am J Gastroenterol. 2003;98:2122–2129.
potency are associated with clinical features, quality of life, 48. Annese V, Janssens J, Vantrappen G et al. Erythromycin
and use of resources in patients with gastroparesis. Clin accelerates gastric emptying by inducing antral contractions
Gastroenterol Hepatol. 2019;17:1285–1294. and improved gastroduodenal coordination. Gastroenterology.
35. Parkman HP, Wilson LA, Hasler WL et al. Abdominal pain 1992;102:823–828.
in patients with gastroparesis: associations with gastropare 49. Janssens J, Peeters TL, Vantrappen G et al. Improvement of
sis symptoms, etiology of gastroparesis, gastric emptying, gastric emptying in diabetic gastroparesis by erythromycin:
somatization, and quality of life. Dig Dis Sci. 2019;64: preliminary studies [see Comment]. N Engl J Med. 1990;322:
2242–2255. 1028–1031.
36. Mearin F, Malagelada JR. Gastroparesis and dyspepsia in 50. Samsom M, Akkermans LM, Jebbink RJ et al. Gastrointestinal
patients with diabetes mellitus. Eur J Gastroenterol Hepatol. motor mechanisms in hyperglycaemia induced delayed gas
1995;7:717–723. tric emptying in type 1 diabetes mellitus. Gut. 1997;40:
37. Kumar A, Attaluri A, Hashmi S et al. Visceral hypersensitiv 641–646.
ity and impaired accommodation in refractory diabetic gas 51. Sugumar A, Singh A, Pasricha PJ. A systematic review of the
troparesis. Neurogastroenterol Motil. 2008;20:635–642. efficacy of domperidone for the treatment of diabetic gastro
38. Camilleri M. Clinical practice: diabetic gastroparesis. N Engl paresis. Clin Gastroenterol Hepatol. 2008;6:726–733.
J Med. 2007;356:820–829 [erratum appears in N Engl J Med. 52. Carbone F, Van den Houte K, Clevers E et al. Prucalopride in
2007;357:427]. gastroparesis: a randomized placebo‐controlled crossover
39. Bharucha AE, Camilleri M, Low PA et al. Autonomic dys study. Am J Gastroenterol. 2019;114:1265–1274.
function in gastrointestinal motility disorders. Gut. 1993;34(3): 53. Abell T, Kuo B, Esfandyari T et al. Velusetrag improves gastro
397–401. paresis both in symptoms and gastric emptying in patients with
40. Camilleri M, Zinsmeister AR, Greydanus MP et al. Towards diabetic or idiopathic gastroparesis in a 12‐week global phase
a less costly but accurate test of gastric emptying and small 2B study. Gastroenterology. 2019;156(6):S164.
bowel transit. Dig Dis Sci. 1991;36:609–615. 54. De Maeyer JH, Prins NH, Schuurkes JA et al. Differential
41. Szarka LA, Camilleri M, Vella A et al. A. stable isotope effects of 5‐hydroxytryptamine4 receptor agonists at gastric
breath test with a standard meal for abnormal gastric empty versus cardiac receptors: an operational framework to
ing solids in the clinic and in research. Clin Gastroenterol explain and quantify organ specific behavior. J Pharmacol
Hepatol. 2008;6:635–643. Exp Ther. 2006;317:955–964.
42. Tack J, Piessevaux H, Coulie B et al. Role of impaired gastric 55. Tack J, Camilleri M, Chang L et al. Systematic review: car
accommodation to a meal in functional dyspepsia. diovascular safety profile of 5‐HT4 agonists developed for
Gastroenterology. 1998;115:1346–1352. GI disorders. Aliment Pharmacol Ther. 2012;35:745–767.
43. Carbone F, Tack J, Hoffman I. The intragastric pressure 56. Beattie DT, Higgins DL, Ero MP et al. An in vitro; investiga
measurement: a novel method to assess gastric accommoda tion of the cardiovascular effects of the 5‐HT(4) receptor
tion in functional dyspepsia children. J Pediatr Gastroenterol selective agonists, velusetrag and TD‐8954. Vascul Pharmacol.
Nutr. 2017;64:918–924. 2013;58:150–156.
44. Tack J, Caenepeel P, Piessevaux H et al. Assessment of meal 57. Dukes GE, Scimia C, Kuo B et al. Safety, tolerability and
induced gastric accommodation by a satiety drinking test in pharmacodynamics of tak‐906, a dopamine 2,3 antagonist,
health and in severe functional dyspepsia. Gut. 2003;52: in patients with diabetic or idiopathic gastroparesis.
1271–1277. Neurogastroenterol Motil. 2019;31(Suppl 3):e13657.
45. Camilleri M, Chedid V, Ford AC et al. Gastroparesis. Nat Rev 58. Pasricha PJ, Yates KP, Sarosiek I et al. Aprepitant has mixed
Dis Primers. 2018;4:41. effects on nausea and reduces other symptoms in patients
46. Olausson EA, Störsrud S, Grundin H et al. A small particle with gastroparesis and related disorders. Gastroenterology.
size diet reduces upper gastrointestinal symptoms in patients 2018;154:65–76.e11.
59. Jacob D, Busciglio I, Burton D et al. Effects of NK1 receptors 69. Camilleri M, McCallum RW, Tack J et al. Efficacy and safety
on gastric motor functions and satiation in healthy humans: of relamorelin in diabetes with symptoms of gastroparesis:
results from a controlled trial with the NK1 antagonist, a randomized, placebo‐controlled study. Gastroenterology.
aprepitant. Am J Physiol Gastrointest Liver Physiol. 2017;313: 2017;153:1240–1250.e2.
G505–G510. 70. Camilleri M, Lembo A, McCallum R et al. Overall safety and
60. Carlin JL, Lieberman VR, Dahal A et al. Tradipitant, a tolerability of relamorelin in adults with diabetic gastropare
novel NK‐1 receptor antagonist, significantly improved sis: analysis of phase 2a and phase 2b trial data. Neurogastro
nausea and other symptoms of gastroparesis: results of a enterol Motil. 2019;31(S4):S15.
multicenter, randomized, double‐blind, placebo‐controlled 71. Friedenberg FK, Palit A, Parkman HP et al. Botulinum toxin
phase II trial. Gastroenterology. 2019;156(Suppl 1): A for the treatment of delayed gastric emptying [see
S1510–S1511. Comment]. Am J Gastroenterol. 2008;103:416–423.
61. Tack J, Depoortere I, Bisschops R et al. Influence of ghrelin 72. Arts J, Holvoet L, Caenepeel P et al. Clinical trial: a rand
on interdigestive gastrointestinal motility in humans. Gut. omized‐controlled crossover study of intrapyloric injection
2006;55:327–333. of botulinum toxin in gastroparesis. Aliment Pharmacol Ther.
62. Peeters TL. Central and peripheral mechanisms by which 2007;26:1251–1258.
ghrelin regulates gut motility. J Physiol Pharmacol. 2003;54: 73. Bai Y, Xu MJ, Yang X et al. A systematic review on intrapy
95–103. loric botulinum toxin injection for gastroparesis. Digestion.
63. Camilleri M, Papathanasopoulos A, Odunsi ST. Actions and 2010;81:27–34.
therapeutic pathways of ghrelin for gastrointestinal disor 74. Pasricha TS, Pasricha PJ. Botulinum toxin injection for
ders. Nat Rev Gastroenterol Hepatol. 2009;6:343–352. treatment of gastroparesis. Gastrointest Endosc Clin N Am.
64. Van der Ploeg L, Laken H, Sharma S et al. Preclinical gastro 2019;29:97–106.
intestinal prokinetic efficacy and endocrine effects of the 75. Thomas A, de Souza Ribeiro B, Malespin M et al. Botulinum
ghrelin mimetic RM‐131. Life Sci. 2014;109:20–29. toxin as a treatment for refractory gastroparesis: a literature
65. Shin A, Camilleri M, Busciglio I et al. Randomized con review. Curr Treat Options Gastroenterol. 2018;16:479–488.
trolled phase Ib study of ghrelin agonist, RM‐131, in type 2 76. Abell TL, Yamada G, McCallum RW et al. Effectiveness of
diabetic women with delayed gastric emptying: pharmacoki gastric electrical stimulation in gastroparesis: Results from a
netics and pharmacodynamics. Diabetes Care. 2013;36: large prospectively collected database of national gastropa
41–48. resis registries. Neurogastroenterol Motil. 2019;31:e13714.
66. Shin A, Camilleri M, Busciglio I et al. The ghrelin agonist 77. Ducrotte P, Coffin B, Bonaz B et al. Gastric electrical stimu
RM‐131 accelerates gastric emptying of solids and reduces lation reduces refractory vomiting in a randomized cross‐
symptoms in patients with type 1 diabetes mellitus. Clin over trial. Gastroenterology. 2019;Oct 21:ii. S0016–5085(19)
Gastroenterol Hepatol. 2013;11:1453–1459. 41463–41467. doi: 10.1053/j.gastro.2019.10.018. [Epub ahead
67. Nelson AD, Camilleri M, Acosta A et al. Effects of ghrelin of print].
receptor agonist, relamorelin, on gastric motor functions and 78. Mearin F, Camilleri M, Malagelada JR. Pyloric dysfunction
satiation in healthy volunteers. Neurogastroenterol Motil. in diabetics with recurrent nausea and vomiting. Gastro
2016;28:1705–1713. enterology. 1986;90:1919–1925.
68. Lembo A, Camilleri M, McCallum R et al. Relamorelin 79. Aghaie Meybodi M, Qumseya BJ, Shakoor D et al. Efficacy
reduces vomiting frequency and severity and accelerates and feasibility of G‐POEM in management of patients with
gastric emptying in adults with diabetic gastroparesis. refractory gastroparesis: a systematic review and meta‐analysis.
Gastroenterology. 2016;151:87–96.e6. Endosc Int Open. 2019;7:E322–E329.
Index
abdominal pain, 273 incretins effect, 221

acarbose, 177–178 new techniques, 225
A Diabetes Outcome Progression Trial (ADOPT), 177 rescue therapy, 221–223
advanced glycosylation end products (AGES), 94 surgery, weight loss, 225
Advanced Hybrid Closed Loop (AHCL), 119 type 2 diabetes mellitus (T2DM), 221
aggressive risk factor control, 164 vertical sleeve gastrectomy (VSG), 221
alanine amino‐transferase (ALT), 141 weight loss, 221
alpha‐glucosidase inhibitors, 190 Bezafibrate Infarction Prevention (BIP), 205
American College of Gastroenterology (ACG), 50 biguanides/metformin, 133, 190
American College of Obstetricians and Gynecologists combination therapy, 135–137
(ACOG), 77 contraindications, 137
American Diabetes Association (ADA), 12, 45, 102, 128 gastrointestinal, 138
amylin mimetics, 190–191 lactic acidosis, 137–138
Anglo‐Scandinavian Cardiac Outcomes Trial mechanism of action, 134–135
(ASCOT), 200 pharmacokinetics, 135
Antihypertensive and Lipid‐Lowering treatment to polycystic ovary syndrome (PCOS), 137
prevent Heart Attack Trial (ALLHAT), 201 bilateral internal thoracic artery (BITA), 233
antipsychotic drug blood glucose (BG), 91
natural history, 57 bromocriptine, 178
schizophrenia, 56
screening, 57–58 caloric sweeteners, 129
second‐generation antipsychotics, 57 Canagliflozin and Renal Events in Diabetes With
atherosclerotic cardiovascular disease (ASCVD), 152 Established Nephropathy Clinical Evaluation
(CREDENCE), 180
bariatric surgery, 11 candidate gene studies, 18
first‐line therapy, 221–222 carbohydrate, 129
gastric banding, 223–224 cardiovascular disease (CVD)
gastric bypass, 223–224 aggressive risk factor control, 164

280
Index 281
carotid ultrasound, 168 Centers for Medicare and Medicaid Services (CMS), 90
coronary artery calcium score, 166–167 Cholesteryl ester transfer protein (CETP), 199
coronary artery disease (CAD), echocardiogram chronic kidney disease (CKD), 148
for, 168 closed loop control (CLC)
coronary computed tomography angiography Auto‐mode enabled MiniMedTM, 119
(CCTA), 167–168 continuous glucose monitoring (CGM), 116
diabetes‐specific clinical risk predictors, 165–166 continuous subcutaneous insulin infusion (CSII), 116
endothelial function studies, 168 control IQ technology, 119–120
guidelines, 163–164 cost of diabetes technologies, 121
heart failure (HF), screening for, 168–169 FLAIR trial, 119
hyperglycemia, 164–165 future technologies, 121
hypertriglyceridemia, 203–204 impaired hypoglycemia awareness, 116
myocardial perfusion imaging, 168 insulin pump, 119–121
National Health and Nutrition Examination Survey MiniMed 670G, 118–119
(NHANES), 163 multiple daily injections (MDI), 116
nontraditional risk factors, 166 ongoing pivotal studies, 120
percutaneous coronary intervention (PCI), 164 post‐marketing studies, 119–121
proprotein convertase subtilisin/kexin type 9 SmartGuardTM, 119
(PCSK9), 202–203 T1D, 120
risk calculators, 169–170 T1DM, currently approved CLC for, 118
screening tools, 166–170 t:slim X2TM, 119–120
stress testing, 168 Cochrane Collaborative, 89
traditional risk factors, 165 colesevelam, 178
cardiovascular risk reduction Collaborative Atorvastatin Diabetes Study (CARDS), 200
acarbose, 177–178 Congenital rubella, 19
bromocriptine, 178 continuous glucose monitoring (CGM), 95–96, 102
colesevelam, 178 diabetes care, 116
diabetes, 174–175 diabetes management, 107
glucagon‐like peptide‐1 (GLP‐1)‐based therapy, duration of, 107
178–180 glycemic control, 106
metformin, 175 glycemic variability, 108–110
pramlintide, 178 hypoglycemia, 106–107
sodium glucose cotransporter 2 inhibitors (SGLT‐2i), implantable, 117–118
180 pregnancy, 117
sulfonylureas, 175–176 time in range (TIR), 108
thiazolidinediones, 176–177 WISDM Study, 116
Carotid Intimal Medial Thickness (CIMT), 166 continuous subcutaneous insulin infusion (CSII)
carotid intima‐media thickness, 142 therapy, 94
carotid ultrasound, 168 cornea, 263–264
cataract, 264 treatment, 264
282 Index
coronary artery bypass surgery (CABG), 164 diagnostic criteria, 3–4

coronary artery calcium score, 166–167 diet, 11–12
coronary artery disease (CAD), echocardiogram exercise, 11–12
for, 168 glucose control, 229
coronary computed tomography angiography (CCTA), haemochromatosis, natural history of, 49
167–168 hereditary haemochromatosis (HH), 48–49
cow’s milk, 19–20 hormone excess, 54–55
Coxsackievirus B4, 19 human immunodeficiency virus (HIV) infection,
C‐peptide, 39, 69 58–59
Cushing’s syndrome, 47, 54–55 hyperglycemia, 229
cystic fibrosis‐related diabetes (CFRD) immunotherapy, 59–60
clinical features of, 52 medical therapy, 229
definition, 50 optimal medical therapy (OMT), 229
natural history, 50 oxidative stress, 228
other forms of diabetes, 50–52 pancreatic‐cancer‐associated diabetes, 53–54
screening, 52–53 patients, subgroup analysis of, 232–233
cystic fibrosis transmembrane regulator (CFTR), 40 percutaneous coronary intervention, 234
cytotoxic T lymphocyte antigen 4 (CTLA‐4), 18 post‐transplantation diabetes mellitus (PTDM), 55–56
prediabetes, 7–8, 11
Da Qing studies, 10–11 prevention, 11
Detection of Ischemia in Asymptomatic Diabetics reclassifying, novel approaches to, 40–41
(DIAD), 164 risk stratification, biomarkers for, 233
Diabeloop Generation 1 (DBLG1) Hybrid Closed symptom control for stable patients, 231
Loop, 120 types of , 45–46
Diabetes Complication and Control Trial (DCCT), 88, diabetes prediction and prevention study (DIPP), 26
89, 174 diabetes prevention trial‐type 1 (DPT‐1), 24–26
diabetes mellitus (DM) diabetes‐specific clinical risk predictors, 165–166
acute coronary syndromes (ACS), 228, 233 diabetic gastroparesis, 269, 270
antiplatelet therapy, 229 diabetic macular edema (DME), 259–260
antipsychotic drug use, 56–57 anti‐VEGF therapy, 261–262
bariatric surgery, 11 corticosteroid, 260–261
Bypass Angioplasty Revascularization Investigation future directions, 263
(BARI), 229 macular laser, 260
cardiovascular disease (CVD), 232 surgery, 262
cardiovascular risk reduction, 174–175 diabetic retinopathy (DR)
coronary artery bypass grafting, 233–234 anti‐VEGF, 258
coronary artery disease (CAD), 228 blood pressure control, 254
coronary revascularization, 230–231 cataract, 264
cystic fibrosis‐related diabetes, 50 combination treatment, 258
delay of, 11 cornea, 263–264
Index 283
diabetic macular edema (DME) see diabetic macular type 2 diabetes (T2D), 199–202
edema (DME) dyspepsia, 273
glycemic control, 254 dysphagia, 272
laser photocoagulation, 256–258
lipid‐lowering strategies, 254–255 eicosapen taenoic acid (EPA), 205–206
neovascular glaucoma, 259 electrocardiogram (ECG), 92
non‐proliferative diabetic retinopathy, 256 endogenous glucose production (EGP), 146
ocular complications, 255–256 endothelial function studies, 168
orbital disease, 264 end‐stage renal disease (ESRD), 148, 170
prevention, 254–255 Enterovirus, 19
proliferative diabetic retinopathy, 256 Epidemiology of Diabetes Interventions and
retinal vein occlusion, 259 Complications (EDIC), 88, 102
risk factors for, 252–253 erythromycin, 275
screening, 253–254 estimated glomerular filtration rate (eGFR), 148
surgical treatment/vitrectomy, 258–259 European Association for the Study of Diabetes
treatment of, 256–259 (EASD), 57
diet, 11–12 European nicotinamide diabetes intervention trial
American Diabetes Association, 128 (ENDIT), 24
caloric sweeteners, 129 European Psychiatric Association (EPA), 57
carbohydrate, 129 exenatide, 148–149
cultural norms, 125 exercise, 11–12
food availability, 125
intermittent fasting, 126 fasting plasma glucose (FPG), 50, 94
ketogenic diets, 127 fetal monitoring, 80
medical nutrition therapy, 124 first‐phase insulin response (FPIR), 25
Mediterranean diet plan, 127 5‐HT 4 receptor agonists, 275
plant‐based diets, 125 flash glucose monitoring systems, 96–99
regular dietary fiber intake, 129 food availability, 125
sugar substitutes, 129
time‐restricted eating (TRE), 126 Galega officinalis, 133
type 1 diabetes, 129 gastric dysfunctions, 270–272
type 2 diabetes, 124, 129 gastroparesis, 272, 273
dipeptidyl peptidase‐IV (DDP‐4) inhibitors, 93, 191 gestational diabetes mellitus (GDM)
drug‐eluting stents (DES), 230 definition, 75
dulaglutide, 150 diagnosis, 75–77
dyslipidemia fetal monitoring, 80
atherosclerotic cardiovascular disease (ASCVD), future perspectives, 81–82
211–212 gestational weight gain, 79–80
PCSK9 see proprotein convertase subtilisin/kexin glucocorticoid therapy, 80
type 9 (PCSK9) glycemic goals, 77
284 Index
gestational diabetes mellitus (GDM) (cont’d) micro‐ and macrovascular complications, 102–103
insulin therapy, 77–78 post prandial glucose (PPG), 105
intrapartum management, 80–81 self‐monitored blood glucose (SMBG), 104–105
lifestyle modification, 77 vascular risk, 101
oral glucose‐lowering agents, 78–79 hepatitis C infection, 56
postpartum maternal care, 81 hereditary haemochromatosis (HH)
postpartum neonatal care, 81 natural history of, 49
gestational weight gain, 79–80 screening case, 49–50
Ghrelin receptor agonist, 275–276 high‐density lipoprotein (HDL), 141
glucagon‐like peptide‐1 (GLP‐1), 93, 178–180, 191 highly active antiretroviral therapy (HAART), 141
clinical benefits, 147 HNF1A mutations, 70
dulaglutide, 150 HNF4A mutations, 70
early development of, 147 HorizonTM Pivotal study, 120
exenatide, 148–149 human immunodeficiency virus (HIV) infection, 58–59
liraglutide, 149–150 human leukocyte antigen (HLA), 17–18, 38
lixisenatide, 150 hyperglycemia, 69, 105, 164–165
semaglutide, 150 critical illness, 242
side effects, 147 glucose testing, 247
glucagon suppression, 38 interpretation, 244–246
glucocorticoids, 54, 80 mechanisms, 242
glucokinase mutations, 69–70 nutritional support, 247–249
glucose‐dependent insulinotropic polypeptide hypertriglyceridemia, 199
(GIP), 147
glucose variability (GV), 115 immune intervention therapies, 27–32
glutamic acid decarboxylase (GAD‐65), 18, 29, 39 immunotherapy trials, 30–31
glyburide, 78 impaired fasting glucose (IFG), 3,
glycemic goals, 77 5, 9–11, 164
glycemic measurements, 6–7 impaired glucose tolerance (IGT), 3, 5, 164
incretin‐based therapy, 32–33
heartburn, 272 cardiovascular outcomes, 152–153
heart failure (HF), screening for, 168–169 clinical practice, 153–154
Heart Outcomes Prevention Evaluation (HOPE), 165 dipeptidyl peptidase‐4 inhibitors (DPP‐4i) see
hemoglobin A1c (HbA1c) dipeptidyl peptidase‐4 inhibitors (DPP‐4i)
clinical practice, integration into, 102 glucagon‐like peptide‐1 (GLP‐1) see glucagon‐like
clinical utility of, 103 peptide‐1 (GLP‐1)
continuous glucose monitoring (CGM) see incretin effect, 146–147
continuous glucose monitoring (CGM) insulin, 18, 188
discovery, 102 insulin autoantibody (IAA), 25
glucose hypothesis, 101 insulin sensitizers
hyperglycemia, 105 biguanides see biguanides
limitations of, 103–104 thiazolidinediones see thiazolidinediones (TZDs)
Index 285
insulin therapy, 53, 77–78 retinopathy, 8

intermittent fasting, 126 Model Predictive Control (MPC), 119
internal thoracic artery (ITA), 230 monogenic diabetes
International Diabetes Federation (IDF), 95 glucokinase mutations, 69–70
intestinal cells of Cajal (ICCs), 270 HNF1A mutations, 70
intrapartum management, 80–81 HNF4A mutations, 70
islet‐cell antibodies (ICA), 18 maturity onset diabetes of the young (MODY), 68
screen for , 69
ketogenic diets, 127 type 1 diabetes, 68–69
ketosis, 69 type 2 diabetes, 69
mortality, 9–10
lactic acidosis, 138 Multi‐Ethnic Study of Atherosclerosis (MESA), 166
large for gestational age (LGA), 75 myocardial infarction (MI), 143, 229, 231
latent autoimmune diabetes of adults (LADA), 39–40 myocardial perfusion imaging, 168
LDL‐cholesterol levels, 212
lipid lowering, 141 National Health and Nutrition Examination Survey
anticoagulation, 255 (NHANES), 6, 8, 163, 187
aspirin, 255 National Screening Committee (NSC), UK, 47
diet, 255 nephropathy, 8–9
fenofibrate, 254 neuropathy, 9
statins, 255 Niacin/nicotinic acid, 205
lipodystrophies, 141 Nicotinamide treatment, 24
liraglutide, 149–150 nitric oxide synthase (NOS), 270
lixisenatide, 150 nonalcoholic fatty liver disease (NAFLD), 140–141
Long‐term Intervention with Pravastatin in Ischemic nontraditional risk factors, 166
Disease (LIPID), 201
low‐density lipoprotein (LDL), 141, 164 omega‐3 fatty acid, 205
lymphoid phosphatase (LYP), 18 oral glucose‐lowering agents, 78–79
oral glucose tolerance test (OGTT), 3, 6–7, 77
maturity‐onset, 68 orbital disease, 264
maturity onset diabetes of the young (MODY), 60, 68 oxidative post‐translationally modified insulin
mean absolute relative difference (MARD), 117 (oxPTM‐ INS), 19
medical nutrition therapy, 124
Mediterranean diet plan, 127 pancreatic‐cancer‐associated diabetes, 53–54
metabolic memory, 88 Patch Pump Series, 120
metformin, 78, 175, 192 see biguanides percutaneous coronary intervention (PCI), 164
metoclopramide, 275 peripheral artery disease (PAD), 165
microvascular complications, prediabetes peroxisome proliferator activated receptor (PPAR),
mortality, 9–10 138–139, 176
nephropathy, 8–9 mechanism, 139
neuropathy, 9 phenformin, 133
286 Index
pheochromocytoma/paraganglioma (PPGL), 54 diabetes, patients with, 214

pioglitazone, 140 efficacy, 212–213
plant‐based diets, 125 future perspectives, 214–215
polycystic ovary syndrome (PCOS), 137 LDL‐cholesterol levels, 212
postpartum maternal care, 81 novel lipid‐lowering therapies, 215–217
postpartum neonatal care, 81 safety, 212–213
post prandial glucose (PPG), 94, 105 Prospective Pioglitazone Clinical Trial in Macrovascular
post‐transplantation diabetes mellitus (PTDM) Events (PROACTIVE), 140
clinical features, 56 Public Health England (PHE), 47
screening, 56
pramlintide, 178 randomized clinical trials (RCTs), 89, 147
prediabetes randomized control trail (RCT), 116
American diabetes association 2020, 12 regular dietary fiber intake, 129
bariatric surgery, 11 retinopathy, 8
Da Qing studies, 10–11 rosiglitazone, 140
definition, 3 Roux‐en‐Y‐Gastric Bypass (RYGB), 38
diabetes mellitus (DM), 3–4, 7–8, 11–12
diabetes prevention program, 10 secondary prevention, 24
diagnostic criteria, 3–4 self‐monitoring of blood glucose (SMBG), 104–105
diet, 11–12 continuous glucose monitoring systems, 95–96
epidemiology of, 4 definition, 87–88
exercise, 11–12 flash glucose monitoring systems, 96–99
glycemic measurements, 6–7 postprandial, 94–95
impaired fasting glucose, 5 type 1 diabetes (T1D), 88–89
impaired glucose tolerance (IGT), 3, 5 type 2 diabetes (T2D), 89–94
management of, 10 semaglutide, 150
microvascular complications see microvascular sensitivity, 6–7
complications, prediabetes 780G pivotal trial, 119
oral glucose tolerance test (OGTT), 3, 6–7 short‐term intensive insulin therapy, 40
pharmacologic therapy, 11 670G System, 119
progression, risk of, 7–8 small inhibiting RNA (siRNA), 212
reproducibility, 6–7 sodium glucose cotransporter 2 inhibitors (SGLT‐2i), 180
screening recommendations, 5–6 sodium‐glucose cotransporters (SGLT), 191–192
sensitivity, 6–7 sterol regulatory element binding protein (SREBP), 212
primary prevention, 21–24 stress testing, 168
prokinetics, 275–276 sugar substitutes, 129
Proportional Integrated Derivative (PID), 118 sulfonylurea (SU), 91
proprotein convertase subtilisin/kexin type 9 (PCSK9) sulfonylureas, 175–176, 188–190
cardiovascular, 213–214 Synergy between Percutaneous Coronary Intervention
clinical use, 214–215 with Taxus and Cardiac Surgery (SYNTAX), 232
Index 287
tertiary prevention, 26–27 primary prevention, 21–24

The Environmental Determinants of Diabetes in the secondary prevention, 24
Young (TEDDY) study, 23 self‐monitoring of blood glucose (SMBG), 88–89
thiazolidinediones (TZDs), 138–139, 176–177, 190 tertiary prevention, 26–27
adverse effects, 142–143 vitamin D deficiency, 20
carotid intima‐media thickness, 142 type 2 diabetes (T2D), 88, 116
combination therapy, 139–140 autoantibodies, 39–40
contraindications, 142 clinical features of, 52
lipid lowering, 141 cystic fibrosis‐related diabetes, 40
lipodystrophies, 141 diet, 124
mechanism of action, 138–139 environmental influences, disease presentation,
nonalcoholic fatty liver disease (NAFLD), 140–141 38–39
pharmacokinetics, 139 first‐line therapy, 221–222
polycystic ovary syndrome, 141 gastric banding, 223–224
time‐restricted eating (TRE), 126 gastric bypass, 223–224
Treating to New Targets (TNT), 201 genetic influences, disease presentation, 38–39
troglitazone (TGZ), 144 glucose toxicity, 40
type 1 diabetes (T1D) hyperglycemia, 37–38
autoantibodies, 39–40 incretin‐based therapy see incretin‐based therapy
clinical features of, 52 incretins effect, 221
cow’s milk, 19–20 insulin secretion, 39
cystic fibrosis‐related diabetes, 40 monogenic diabetes, 69
diabetes prediction and prevention study (DIPP), 26 new techniques, 225
diabetes prevention trial‐type 1 (DPT‐1), 24–26 pathophysiology, 37–38
diet, 129 rescue therapy, 221–223
environmental factors, 19 self‐monitoring of blood glucose (SMBG), 89–94
environmental influences, disease presentation, 38–39 surgery, weight loss, 225
European nicotinamide diabetes intervention vertical sleeve gastrectomy (VSG), 221
trial (ENDIT), 24 weight loss, 187–188, 221
GAD65, 29 United Kingdom Prospective Diabetes Study (UKPDS),
genetic influences, disease presentation, 38–39 88, 103, 175
glucose toxicity, 40 upper gastrointestinal manifestations
hyperglycemia, 37–38 abdominal pain, 273
immune intervention therapies, 27–32 diabetic gastroparesis, 270
immunotherapy trials, 30–31 diagnosis, 272–273
incretin‐based therapies, 32–33 dyspepsia, 273
insulin secretion, 39 dysphagia, 272
monogenic diabetes, 68–69 endoscopic and device therapies, 276
pathogenesis of, 17–19 epidemiology, 269–270
prevention of, 21–22 gastric dysfunctions, 270–272
288 Index
upper gastrointestinal manifestations (cont’d) diabetic medications for, 192

gastroparesis, 272, 273 DPP‐4 inhibitors, 191
heartburn, 272 GLP‐1 receptor agonists, 192
mechanisms, 269 glucagon‐like peptide (GLP)‐1, 191
nutritional support, 273–275 guide to, 193
pathophysiology, 269 insulin, 188
prokinetics, 275–276 metformin, 192
obesity, 187–188
vitamin D deficiency, 20, 26–27, 223 sodium‐glucose cotransporters (SGLT), 191–192
sulfonylureas, 188–190
weight loss, medication for thiazolidinediones (TZDs), 190
alpha‐glucosidase inhibitors, 190 type 2 diabetes, 187–188
amylin mimetics, 190–191
biguanides/metformin, 190 Zinc Transporter 8 (ZnT8), 18, 39
WILEY END USER LICENSE AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.

Clinical Dilemmas In: Diabetes

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Clinical Dilemmas In: Diabetes

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Clinical Dilemmas In: Diabetes

Uploaded by

Copyright:

Available Formats

14.

Clinical Dilemmas in Diabetes

Cover Design: Wiley

Adrian Vella MD FRCP(Edin.)

17 New Agents for Treatment of Dyslipidemia 211 21 Diagnosis and Management of Ophthalmic

Part IV Management of Disease

Sabreen Ahmed Robert Cuddihy Aoife M. Egan

Kristin Grdinovac Blandine Laferrère Silvia Pieralice

Steve A. Smith Adrian Vella Alexander J. Williams

LE A R NI NG P OI NT S between 140–199 mg/dL [2]. Unlike the ADA, the WHO

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

Bimodal distribution of two-hour glucose levels

for the development of microvascular complications, spe­ Epidemiology of prediabetes

Peripheral Systematic appearance of portal

EGP Portal insulin secretion

T1D have been shown in countries having both high and

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

Beta cell mass (%)

to autoantibody-­ positivity in children at increased European Nicotinamide Diabetes Intervention

Global Immune T cell co- Innate Vitamin Innate

TABLE 2.2 Current TrialNet studies. (Source: www.trialnet.net)

Study Status Description

NCT02354911 A Randomized, Double-­Blind, Placebo-­Controlled, Immunoregulatory Dendritic MMTT C-­peptide at

LE A R NI NG P OI NT S of diabetes would require documenting an immune infil­

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

describe a variety of forms of diabetes for which a cause

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

TABLE 4.1 Types of diabetes.

Type 1 Diabetes Mellitus Immune‐mediated Endocrinopathies Acromegaly

Idiopathic Cushing’s syndrome

TABLE 4.2 Overview of secondary diabetes due to pancreatic disease and hormone excess.

Evidence for benefit Screening for DM

Pancreatic Hereditary Serum ferritin ++ OGTT/FG

Hormone Cushing’s 24hour UFC − OGTT

Adult 1st degree

Fasting transferrin saturation

TS <45% and TS >45% and ferritin

No further iron Genotype

TABLE 4.3 Clinical features of T1DM, T2DM and CFRD. Adapted from [49].

T1DM T2DM CFRD

BMI Usually normal Usually increased Usually decreased

performed on 2 separate days in the absence of classical Pancreatic‐cancer‐associated diabetes

Monogenic forms of diabetes

References 16. Barton JC, Acton RT. Diabetes in HFE Hemochromatosis. J

LE A R N I N G P OI NT S diagnosis may also facilitate genetic counselling to affected

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

When to screen for monogenic diabetes? Common forms of monogenic diabetes

LE A R NI NG P OI NT S In order to preserve carbohydrate for the growing fetus,

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

TABLE 6.1 Options for GDM screening and diagnosis.

Number of abnormal Glucose cut‐offs

FIGURE 6.1 Key factors contributing to

LE A R NI NG P OI NT S Self‐monitoring of blood glucose

Clinical Dilemmas in Diabetes, Second Edition. Edited by Adrian Vella.

BREAKFAST LUNCH DINNER BEDTIME

2/24 179 199 227 9.9-11.0-15.1

2/25 198 177 189 Shopping 11.0-9.8-10.5

for the development of microvascular complications, spe Epidemiology of prediabetes

to autoantibody- positivity in children at increased European Nicotinamide Diabetes Intervention

NCT02354911 A Randomized, Double-Blind, Placebo-Controlled, Immunoregulatory Dendritic MMTT C-peptide at

LE A R NI NG P OI NT S of diabetes would require documenting an immune infil

delivery. Based on deficiency in insulin secretion, three

References with insulin-treated diabetes in the landmark study. Diabetes

LE A R N I N G P OI NT S production (EGP). In the fasting state, reduced plasma glu