4 5800639631472986150

Landmark Papers in
General Surgery
Landmark Papers in . . . series
Titles in the series
Landmark Papers in Neurosurgery
Edited by Reuben D. Johnson and Alexander L. Green
Landmark Papers in Anaesthesia
Edited by Nigel R. Webster and Helen F. Galley
Landmark Papers in Cardiovascular Medicine
Edited by Aung Myat and Tony Gershlick
Landmark Papers in Nephrology
Edited by John Feehally, Christopher McIntyre, and J. Stewart Cameron
Landmark Papers in General Surgery
Edited by Graham J. MacKay, Richard G. Molloy, and Patrick J. O'Dwyer
Landmark Papers in Allergy
Edited by Aziz Sheikh, Tom Platts-Mills, and Allison Worth
Advisory editor Stephen Holgate
Landmark Papers in
General Surgery
Edited by
Graham J. MacKay
Consultant Colorectal Surgeon
Lister Department of Surgery
Glasgow Royal Infirmary
Glasgow, UK
Richard G. Molloy
Consultant Colorectal Surgeon
Department of Surgery
Gartnavel General Hospital
Glasgow, UK
Patrick J. O’Dwyer
Professor of Surgery
Department of Surgery
Gartnavel General Hospital
Glasgow, UK
1
1
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Oxford University Press, 2013
The moral rights of the authors have been asserted
First published in 2013
Impression: 1
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, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
ISBN 978–0–19–964425–4
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Oxford University Press makes no representation, express or implied, that the drug dosages in this
book are correct. Readers must therefore always check the product information and clinical
procedures with the most up-to-date published product information and data sheets provided by the
manufacturers and the most recent codes of conduct and safety regulations. The authors and the
publishers do not accept responsibility or legal liability for any errors in the text or for the misuse or
misapplication of material in this work. Except where otherwise stated, drug dosages and
recommendations are for the non-pregnant adult who is not breast-feeding.
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Foreword
This is a very different and practical book, highlighting the current state of knowledge in
a crisp easy-to-read format which should greatly assist evidence-based decision-making.
Modern landmark papers in each of many subjects are carefully presented and assessed in
an objective manner, with a useful critique then added. It should prove helpful as a quick
update for busy consultant surgeons, as well as being beneficial to surgeons in training
for decision-making in clinical management. Many candidates for postgraduate surgical
examinations could rapidly find valuable information in these pages. Medical students
may also be attracted to this book, as the presentation method of contemporary data and
its relevance is so clear.
The authors and editors are to be congratulated on both their enlightened approach
and concise presentation of such a diverse range of important clinical topics. The uniform
layout of the various sections is a very good feature of this valuable production from the
Glasgow Medical School.
Clem W. Imrie
Emeritus Professor of Surgery
This page intentionally left blank
Contents
Contributors ix
Abbreviations xi
Introduction xv
1 Methodological considerations in surgical trials 1

2 General surgery 25
3 Emergency surgery 55
4 Oesophagogastric surgery 105
5 Hepatopancreaticobiliary surgery 155
6 Colorectal surgery 209
7 Vascular surgery 261
8 Transplant surgery 297
9 Breast surgery 317
10 Endocrine surgery 357
11 Trauma surgery 369
Appendix: Levels of evidence 382

Author Index 384
Subject Index 389
Contributors
Gavin J. Bryce Julie C. Doughty

Surgical Registrar Consultant Breast Surgeon
Lister Department of Surgery University Department of Surgery
Glasgow Royal Infirmary Western Infirmary
UK Glasgow, UK
C. Ross Carter Ronald K.G. Eifell
Consultant Surgeon Specialty Registrar in
West of Scotland Pancreatic Unit General Surgery
Glasgow Royal Infirmary Western Infirmary
UK Glasgow, UK
Lucia Chung Matthew Forshaw
Specialty Registrar in Consultant Upper GI Surgeon
General Surgery Lister Department of Surgery
West of Scotland Deanery Glasgow Royal Infirmary
Glasgow, UK UK
Marc J. Clancy Simon C. Gibson
Consultant Transplant & General Upper GI Fellow
Surgeon, Transplant Unit Concord Hospital
Western Infirmary Sydney
Glasgow, UK Australia
Joseph Crozier Omar Hilmi
Consultant Colorectal & General Consultant Otorhinolaryngologist &
Surgeon Head and Neck Surgeon
Forth Valley Royal Hospital Glasgow Royal Infirmary
Larbert, UK UK
Andrew B. C. Crumley Graham J. MacKay
Specialty Registrar in General Consultant Colorectal Surgeon
Surgery Lister Department of Surgery
Lister Department of Surgery Glasgow Royal Infirmary
Glasgow Royal Infirmary, UK UK
Euan J. Dickson Ken MacKenzie
Consultant Surgeon Consultant Otorhinolaryngologist &
West of Scotland Pancreatic Unit Head and Neck Surgeon
Surgical Lead for Critical Care Glasgow Royal Infirmary
Glasgow Royal Infirmary UK
UK
x CONTRIBUTORS
Iain R. Macpherson Patrick J. O’Dwyer

Clinical Senior Lecturer in Medical Professor of Surgery
Oncology Department of Surgery
University of Glasgow Gartnavel General Hospital
Glasgow, UK Glasgow, UK
Elizabeth A. Mallon Paul N. Rogers
Consultant Pathologist Consultant General &
Department of Pathology Vascular Surgeon
Southern General Hospital Western Infirmary
Glasgow, UK Glasgow, UK
Kathryn McCarthy Tamim Siddiqui
Consultant Colorectal Surgeon Specialty Registrar in
North Bristol NHS Trust General Surgery
UK West of Scotland Deanery
Colin J. McKay Glasgow, UK
Consultant Pancreatic Surgeon Sheila Stallard
West of Scotland Pancreatic Unit Consultant Breast Surgeon
Glasgow Royal Infirmary Department of Surgery
UK Western Infirmary
Richard G. Molloy Glasgow, UK
Consultant Colorectal Surgeon Karen Stevenson
Department of Surgery Specialty Registrar in
Gartnavel General Hospital General Surgery
Glasgow, UK West of Scotland Deanery
Susan J. Moug Glasgow, UK
Specialist Registrar Robert C. Stuart
in General Surgery Consultant Upper GI Surgeon
West of Scotland Deanery Independent Practice
Glasgow, UK
John Norrie
Professor & Director
Centre for Healthcare Randomised
Trials (CHaRT)
Health Services Research Unit
Aberdeen University, UK
Abbreviations
5-FU 5-fluorouracil CHD common hepatic duct

5-HIAA 5-hydroxyindoleacetic acid CHI Community Health Index
AAA abdominal aortic aneurysm CI confidence interval
ABOc ABO compatible CME complete mesocolic excision
ABOi ABO incompatible CMF cyclophosphamide, methotrexate,
ABPI ankle brachial pressure index and 5-fluorouracil
ABx antibiotic(s) CMT combined modality treatment
ACPGBI Association of Coloproctology of CMV cytomegalovirus
Great Britain and Ireland CNS central nervous system
AFS amputation-free survival CP chronic pancreatitis
AI aorto-iliac CRM circumferential resection margin
AIP autoimmune pancreatitis CRP C-reactive protein
APER abdomino-perineal excision of CRT chemoradiotherapy
the rectum CT computed tomography
ARDS adult respiratory distress syndrome CVA cerebrovascular accident
ASA American Society of CVP central venous pressure
Anesthesiologists CVSO2 central venous oxygen saturation
ASCRS American Society of Colon and CXR chest X-ray
Rectal Surgeons
DBD donation after brain death
ASGBI Association of Surgeons of Great
Britain and Ireland DCD donation after cardiac death
DCIS ductal carcinoma in situ
AUC area under curve
DCL damage control laparotomy
AUGIS Association of Upper
Gastrointestinal Surgeons DDFS distant disease-free survival
AXR abdominal X-ray DFS disease-free survival
BASO British Association of Surgical DGF delayed graft function
Oncology DL definitive laparotomy
bd twice daily DLC delayed laparoscopic cholecystectomy
BDI bile duct injury DNA deoxyribonucleic acid
BMI body mass index DSA digital subtraction angiography
BMT best medical therapy DVT deep venous thrombosis
BP blood pressure EAUS endo-anal ultrasound
BSG British Society of Gastroenterology EBD endoscopic balloon dilation
CBD common bile duct EBM evidence-based medicine
CEA carcino-embryonic antigen EBTI endoscopic botulinum
CEA carotid endarterectomy toxin injection
CEPOD confidential enquiry into EBV Epstein–Barr virus
peri-operative deaths ECF epirubicin, cisplatin and
CgA chromogranin-A 5-fluorouracil
xii ABBREVIATIONS
ECMO extra-corporeal membrane HR hazard ratio

oxygenation HRQoL health-related quality of life
ED emergency department HS hand-sewn
eGFR estimated glomerular filtration rate HTA health technology assessment
ELAPE extra-levator abdomino-perineal IA interval appendicectomy
excision
IAP International Association of
ELC early laparoscopic cholecystectomy Pancreatology
EN enteral nutrition IC intermittent claudication
EORTC European Organization for Research ICA internal carotid artery
and Treatment of Cancer
ICU intensive care unit
ER oestrogen receptor
IDMC independent data monitoring
ERAS enhanced recovery after surgery committee
ERCP endoscopic retrograde IM intramuscular
cholangiopancreatography
IONM intra-operative nerve monitoring
ES early surgery
IPMN intraductal papillary mucinous
ESRD end-stage renal disease neoplasm
ESRF endstage renal failure IPN infected pancreatic necrosis
EU European Union ISCVS International Society for
EUS endoscopic ultrasound Cardiovascular Surgery
EVAR endovascular aneurysm repair ISD Information Services Division
EVLT endovenous laser therapy ISS injury severity score
F/U follow-up ITT intention to treat
FAP familial adenomatous polyposis IV intravenous
FDA Food and Drug Administration KAT Knee Arthroscopy Trial
FIT faecal immunological test LA local anaesthetic
FNA fine-needle aspiration LAF laparoscopic anterior fundoplication
FOBt faecal occult blood test Lap laparoscopic
FP femoro-popliteal LBO large bowel obstruction
GA general anaesthetic LC laparoscopic cholecystectomy
GBS Glasgow–Blatchford score LDN laparoscopic donor nephrectomy
GCSF granulocyte colony-stimulating LFN laparoscopic Nissen fundoplication
factor LGD low-grade dysplasia
GERSS Gastroesophageal Reflux Symptom LHM laparoscopic Heller’s myotomy
Score LoS length of stay
GFR glomerular filtration rate LPF laparoscopic posterior fundoplication
GI gastrointestinal LSN least significant number
gFOBt guaiac faecal occult blood test LSV long saphenous vein
GORD gastro-oesophageal reflux disease LUQ left upper quadrant
GP general practitioner MCID minimal clinically important
GSV great saphenous vein difference
GTN glyceryl trinitrate MDT multi-disciplinary team
hCG human chorionic gonadotrophin MEN multiple endocrine neoplasia
HDU high-dependency unit MI myocardial infarction
HGD high-grade dysplasia MMF mycophenolate mofetil
HPF high-power field MOF multi-organ failure
ABBREVIATIONS xiii
MRA magnetic resonance angiography PV portal vein

MRC Medical Research Council QALY quality-adjusted life-year
MRCP magnetic resonance QoL quality of life
cholangiopancreatography RCT randomized controlled trial
MRI magnetic resonance imaging RFA radiofrequency ablation
NCEPOD National Confidential Enquiry into RLN recurrent laryngeal nerve
Patient Outcome and Deaths
RR relative risk
NE nasoenteric
RT radiotherapy
NET neuroendocrine tumour
SAP severe acute pancreatitis
NHS National Health Service
SBO small bowel obstruction
NICE National Institute for Health and
SCPRT short-course pre-operative
Clinical Excellence
radiotherapy
NIH National Institutes of Health
SD standard deviation
NIHR National Institute for Health
SE standard error
Research
SEMS self-expanding metal stents
NJ nasojejunal
SEP supervised exercise programme
NNT number needed to treat
SFA superficial femoral artery
ns not significant
SGV short gastric vessel
NODAT new-onset diabetes after
transplantation SHIP Scottish Health Informatics
Programme
NSAID non-steroidal
anti-inflammatory drug SIRS systemic inflammatory response
syndrome
NSAP non-specific abdominal pain
SLI severe limb ischaemia
OAC oesophageal adenocarcinoma
SMR standardized mortality ratio
OGJ oesophagogastric junction
SMS somatostatin
OR odds ratio
SMV superior mesenteric vein
OS overall survival
SNB sentinel node biopsy
PA postero-anterior
SNS sacral nerve stimulation
PAD peripheral artery disease
SOS Swedish Obese Subjects
PATI penetrating abdominal trauma index
SPN sterile pancreatic necrosis
PBD pre-operative biliary drainage
SSRS somatostatin receptor
PCA patient-controlled analgesia
scintigraphy
PE pulmonary embolism
TA tranexamic acid
PET positron emission tomography
TAPP trans-abdominal pre-peritoneal
PJ pancreatico-jejunostomy
tds three times daily
PP per protocol
TEN total enteral nutrition
PPH procedure for prolapse and
TEP totally extra-peritoneal
haemorrhoids
TFT thyroid function test
PPI proton pump inhibitor
THO trans-hiatal oesophagectomy
PR progesterone receptor
TIA transient ischaemic attack
PTA percutaneous transluminal
angioplasty TME total mesorectal excision
PTC percutaneous trans-hepatic TNF tumour necrosis factor
cholangiography TPN total parenteral nutrition
PTH parathyroid hormone TSC Trial Steering Committee
PTMC papillary thyroid microcarcinoma TTO trans-thoracic oesophagectomy
xiv ABBREVIATIONS
UGFS ultrasound-guided foam VTE venous thromboembolism

sclerotherapy WCC white cell count
UKCRC UK Clinical Research Collaboration WHO World Health Organization
USS ultrasound scan WMD weighted mean difference
UTI urinary tract infection WOSCOPS West of Scotland Coronary
VARD video-assisted retroperitoneal Prevention Study
debridement WSC water-soluble contrast
VAS visual analogue scale
Introduction
Evidence-based medicine (EBM) is one of the cornerstones of effective clinical practice.

While the term ‘evidence-based medicine’ may be relatively new, the concept is not. The
importance of systematic appraisal of outcomes following surgical intervention was
appreciated as far back as the eighteenth century by pioneers such as John Hunter, one of
the fathers of modern surgery. More recently, EBM has been promoted at an institutional
and organizational level and is a significant focus of both undergraduate and postgraduate
medical education programs.
Our aim with Landmark Papers in General Surgery is to give surgeons and other health-
care professionals easy access to the evidence base behind current clinical practice in
General Surgery. While a book of this kind can never fully distil the conclusions of the
many thousands of articles published in the surgical literature every year, we hope that it
will provide a useful resource for busy clinicians and trainees alike. In particular, the text
is aimed at higher surgical trainees preparing for the academic and specialty sections of
the FRCS exit examination. By collating the current evidence in one text and combining
it with guidance on critical appraisal, it is hoped that this will prove a useful revision tool
for surgical trainees. We would also expect it to appeal to clinicians unable to devote the
necessary time to studying multiple sources of medical research and for allied health pro-
fessionals keen to improve the depth of their understanding of surgery.
There could be many definitions of what makes a truly ‘landmark’ paper. In this volume
you will find a range of studies from case series and randomized trials to meta-analyses
and international guidelines. All the papers have been selected by groups of surgeons
active in research within their sub-specialty area. They have sought to bring together
some of the papers which have had a significant impact on the day-to-day clinical prac-
tice of surgery while covering a breadth of topics within each sub-specialty. While you
may not agree with all of their selections, we hope that at the very least they will provide
a starting point to stimulate debate and further reading of the wealth of surgical literature
that is available.
We have tried to summarize each paper using a common template to provide the
reader with a basic structure with which to appraise the medical literature. This approach
also allows ease of access to salient points. The authors have, where appropriate, graded
the evidence using the Oxford Centre for Evidence-Based Medicine Levels of Evidence
(March 2009). There is of course a degree of subjectivity to this grading, and it is included
for guidance rather than as a statement of fact. We expect that readers will use their own
critical faculties to decide whether they agree with the authors’ assessment. In addition to
a summary of each paper, there is a short section describing the impact of the research on
current clinical practice and any relevant criticisms.
Chapter 1
Methodological considerations
in surgical trials
1.1 Introduction 2
1.2 Designs for surgical trials 3
1.3 Specific design issues in surgical trials 7
1.4 Issues in the conduct of randomized controlled trials 16
1.5 Analysis of surgical trials 18
1.6 Other issues in surgical trials 22
1.7 Summary 23
2 METHODOLOGICAL CONSIDERATIONS IN SURGICAL TRIALS
1.1 Introduction
This new volume will provide interested groups with a valuable guide to surgical research
that changed practice—landmark studies that improved professional practice and out-
comes for patients undergoing surgical procedures. It will also be a useful resource
for researchers trying to build the best designs for future surgical studies, giving them
insights into how to overcome the challenges inherent in conducting high-quality surgi-
cal research.
This chapter is intended as a primer for how to get the most out of the subsequent chap-
ters—the evidence for these developments in surgical interventions. However, it is impor-
tant to remember that not all evidence is good evidence. Just because a trial incorporated
randomization at the start does not guarantee that it is high quality. It may provide much
poorer evidence than a well-conducted and appropriately analysed observational study.
Although one would hope that most research published in major high-profile general-
ist journals would be scrutinized carefully to ensure the analyses are appropriate to the
design, and the interpretations are in line with the data, with conclusions that are justified,
this is not always the case. As a consumer of research articles it is important to develop
skills of critical reading, to be able to sense what is credible and what is dubious.
In this chapter we will try to develop those critical faculties by describing the most
common designs and their methodological properties, such as PICO (Populations,
Interventions, Controls, and Outcomes) and randomized designs (parallel groups, cross-
over, and cluster designs), as well as observational studies (cohort and case–control).
We will discuss common challenges to the assumptions behind these designs (issues of
selection bias, lack of blinding, clustering through surgeon effects, missing data, patient-
reported outcomes), and also discuss how trials and other high-quality designs should be
reported (the various CONSORT statements and STROBE guidelines).
There is an excellent overview of methodological challenges in surgical trials from a
recent paper in The Lancet (Ergina et al. 2009). The interested reader should also consult
the Medical Research Council (MRC) guidance on complex interventions for a full dis-
cussion of not just surgical trials but the broader class of complex interventions in which
they sit (MRC 2008).
Ergina, P.L., Cook, J.A., Blazeby, J.A., et al. (2009) Challenges in evaluating surgical innovation. Lancet,
374, 1097–1104.
Medical Research Council (2008) Developing and evaluating complex interventions: new guidance.
<http://www.mrc.ac.uk/Utilities/Documentrecord/index.htm?d=MRC004871>
DESIGNS FOR SURGICAL TRIALS 3
1.2 Designs for surgical trials

Surgeons perform surgical procedures based on evidence garnered from a wide variety of
surgical trials. However, it is not always possible to apply the same methodological rigour
to the assessment of a new surgical intervention compared with non-surgical interven-
tions such as the administration of a new drug. There are a number of reasons for this,
but the two most important methodological issues affecting surgical trials are as follows:
◆ surgical interventions are often impossible to ‘blind’ and hence are conducted in an
open fashion;
◆ surgical interventions are complex, and are delivered according to the surgeon’s skills
and experience and his/her interaction with the surgical team.
As a result, surgical trials, which are amongst the most complex of interventions, may
not compare favourably when judged against the standards and methodology developed
for drug trials. Regulatory drug trials develop in a very logical and ordered fashion, build-
ing evidence for the safety of a drug (phase I) at the right dose (phase II), and for effec-
tiveness in the right population (phase III) at the right price and acceptability (phase IV)
(see Fig.1.1). Surgical trials do not have this same trajectory, nor in certain regard do they
need it.
Although all trials share more in common than they differ by, there are important
issues in surgical trials (such as blinding and standardizing the intervention) that set
them apart from drug trials. For this reason they must be judged on their merits and
not against standards fit for another purpose. This applies not just to journal editors and
reviewers assessing papers for publishing, but also to funding panels deciding what grants
to fund.
Despite these difficulties, the randomized controlled trial (RCT) offers the best oppor-
tunity to assess a surgical intervention. A well-designed RCT which is properly conducted,
Drug trials
Phases
IV
III
II
I
Logical Highly Highly Few

development resourced regulated surprises
Fig. 1.1 Drug trials.

analysed, and reported offers the best chance to prospectively address issues that might
seriously undermine the credibility of a study if left uncontrolled or unaddressed.
However, RCTs are expensive and time-consuming, and even after succeeding in get-
ting funding (often a huge obstacle), the investigator may have to wait years to obtain
the long-term outcome data for a specific intervention. Trials also tend to select subjects
who fit certain inclusion criteria, and patients often self-select by how willing they are
to take part and how motivated they are to comply with the follow-up. Trials are often
time-limited, with a short follow-up that cannot assess longer-term safety issues or the
sustainability of treatment effects.
Alternative designs to RCTs (including electronic disease registries and in particu-
lar retrospective case–control studies) can offer cheap and rapid evidence. There is an
increasing ability to record–link into national outcomes databases giving details at a
population level of deaths, cancers, hospitalizations, and possibly health resources use
(prescriptions for example). These registries are becoming increasingly feasible and trac-
table, offering good-quality data on potentially huge numbers, with linked databases
interleaved securely at a fraction of the cost of a randomized trial. Indeed, randomized
trials themselves are being built at ultra-low cost using this technology of assessing all the
outcomes via routine national data sources. However, as indicated, care is needed to assess
the quality of non-randomized trials in order to make appropriate interpretations.
The most important distinctions in study design are as follows.
Observational versus randomized

The major tool for equalizing underlying risk of the outcome under study (usually at
baseline, before the surgical procedure) is randomization. Studies that are observational
(e.g. comparing two types of surgery) need to try to provide reassurance that those getting
one type of surgery were not at a lower level of underlying risk than the comparison type
of surgery. Otherwise the comparison of these two surgeries will be confounded by this
differential risk.
Pragmatic versus explanatory

An explanatory trial is concerned with estimating efficacy. That is, if we could aspire to be
perfect surgeons (who are adequately skilled and experienced, having reached and main-
tained a plateau on their learning curve) operating on perfect patients (with no comor-
bidities) in perfect operating theatres (with state of the art equipment, and no threat of
contamination or infection) and with perfect delivery of the surgery to which the patient
was randomized (no cancelled operations, meaning missing outcomes, and no crossover
operations because once on the operating table the surgeon just had to follow their own
judgement and switch procedures), what would the effect of the intervention be under
these idealized conditions?
On the other hand, pragmatic trials estimate effectiveness rather than efficacy. That is,
under real-world conditions, what is the effect of the intervention on an unselected group
of patients (with minimal if any exclusions for comorbidities) by an unselected group of
DESIGNS FOR SURGICAL TRIALS 5
surgeons (of all levels of skills and experiences and possibly preferences for one type of
operation over another) operating in a busy NHS hospital (with delays and cancellations
and possibly less than perfect operating conditions), randomized to but not necessarily
receiving the surgery allocated, and with missing data and loss to follow-up?
Identifying where a drug trial is on the exploratory–pragmatic continuum is reason-
ably straightforward. Often complex intervention surgical trials are a heady mixture of
the two, measuring all sorts of explanatory-type measures in a process evaluation to bet-
ter understand the mechanisms as to why one surgical technique might be superior to
another.
Prospective versus retrospective

In general, if data are collected in the future to a specified plan, soundly executed, they
should be less vulnerable to bias. For example, when accessing retrospective data from a
hospital database on surgical performance, you can only go on what is there. For instance,
if all surgeries that resulted in legal action were removed from the database it would create
a potentially major bias.
Cross-sectional versus longitudinal

As the terms suggest, there is a fundamental difference between a survey done at one
point in time and a study that follows up either the same patients over many time points
(a longitudinal study) or a study that follows a new set of subjects having a surgery over
many years (a panel-type study).
In addition, surgical trials have levels of difficulty within themselves. Adrian Grant,
a very experienced trialist and former Director of the Health Services Research Unit in
Aberdeen, which has pioneered many surgical trials and their methodological innova-
tion, created a useful categorization of surgical trials, according to what interventions
were being compared (Fig. 1.2).
• Variation A
• Variation B
SURGICAL TECHNIQUE
Easiest
• Operation A
DIFFERENT OPERATIONS
Harder • Operation B
• Surgery
• Medical management DIFFERENT OPTIONS
Hardest
Fig. 1.2 Levels of complexity of surgical trials.

For further reading on study design consider Pocock (1983) and Piantadosi (2005), and
as a general reference the Encyclopaedia of Biostatistics (Armitage and Colton 2005) has
good chapters on all the major designs (and some more exotic variants).
Pocock, S.J. (1983) Clinical Trials: A Practical Approach. Chichester: Wiley.
Piantadosi, S. (2005) Clinical Trials: A Methodologic Perspective (2nd edn). Hoboken, NJ: Wiley.
Armitage, P. and Colton, T. (2005) Encyclopedia of Biostatistics (2nd edn). Hoboken, NJ: Wiley.
SPECIFIC DESIGN ISSUES 7
1.3 Specific design issues in surgical trials

Ethical issues
Emergency surgery
There are particular ethical issues in trials of surgical interventions when the patient has a
life-threatening condition and needs to undergo emergency rather than elective surgery.
If the patient is unable to give informed consent, provision can be made to gain assent from
a relative or designated responsible person. This can then be converted to full consent if/
when the patient has recovered sufficiently to grant (or withhold) informed consent.
Vulnerable groups
All research must be seen to be observing the rights of the patient, and this can pose par-
ticular challenges for vulnerable groups—by age (babies, infants, children), or those who
are cognitively impaired (e.g. those with dementia), or those with learning difficulties. It is
important not to take the easy route and argue that it is too difficult to include such people
in research trials. For example, the best interventions for children may be quite different
from those that work best for adults.
Placebo surgery trials

The very notion of a placebo surgery generates strong opinions. Unfortunately, placebo
surgery is also inappropriately referred to as ‘sham’ surgery, although it is interesting to
note that this term, which has unwelcome connotations (e.g. fraud or hoax), is not used
to refer to placebos in drug trials.
From a trial design perspective, one can argue that a placebo surgery arm would be
helpful as a comparator for some surgical procedures that have little or no formal high-
quality evidence to support them. On the other hand, there are those who quite reason-
ably argue that all surgery carries a risk (e.g. there is a strictly non-zero chance that a
patient will die on the operating table) and therefore no placebo trial is ethically accept-
able. Placebo surgery trials have been carried out (Moseley et al. 2002). The interested
reader should consult the excellent paper by Campbell et al. (2010) for a thorough discus-
sion of the difficulties in designing and conducting a placebo surgery trial in the NHS.
Withdrawal in surgical trials

It is a fundamental principle of the informed consent process that a patient can withdraw
from the trial at any point without prejudice to their ongoing management and treat-
ment. In practice, withdrawal causes many problems logistically, but these can usually be
avoided if clarity is brought to bear on discussing the options with the patient.
Withdrawal of consent Sometimes a participant does not want anything more to do
with the research. This should be respected, but note that even in this situation one can
usually still use the data captured up to that point (as withdrawal of consent cannot be
retrospective—the exception to this might be a patient on awakening from emergency
surgery, having been assented into a trial by a relative or competent person).
Withdrawal other than withdrawal of consent For example, a patient in ICU may not
want to be contacted again as it is distressing to be reminded of the car crash and subse-
quent surgery, or a patient has suffered a serious adverse event that they blame on their
participation in the trial. Here, it is important to understand what the patient is withdraw-
ing from. Quite often a patient will be happy for their data to be used, and happy that
further access is made to their medical notes, but they do not want to attend the hospital
for any further follow-up visits. It might be that they are happy to be phoned or emailed
to collect a twelve-month outcome. It is unfortunate that the term ‘withdrawal’ is used to
cover these varied and complex situations. Too often it can translate in the study person-
nel mindset that this means ‘don’t get any more follow-up’, when such information is in
reach, and from the scientific point of view is often very valuable (as these may be the
most interesting cases, whose data cannot be assumed to be missing at random). Therefore
it is important to design ‘change of status’ (not ‘withdrawal’) forms to allow a proper
enquiry to be made of the patient as to what their preferences are for the level of further
involvement in the study and to document these.
More complex designs

Equivalence or non-inferiority trials
The most common and conceptually easiest design is a ‘superiority’ study. In this design,
the objective is to show that a new surgical technique has better efficacy or effectiveness
than, perhaps, the standard surgical technique. However, it might be that a new technique
is highly unlikely to be quite as effective as the standard technique but it has some other
advantage, such as being quicker (allowing it to be done as a day case, rather than need-
ing an overnight stay) or less painful, or cheaper (more cost effective), or that it can be
done by less experienced surgeons. In such a situation you may want to show that the
new technique is almost as good as the standard in terms of effectiveness, but that this
slight underperformance is traded off against other desirable dimensions where it is bet-
ter. In this situation you might use a non-inferiority design, in which you specify a non-
inferiority margin—that is, the amount that you and your clinical colleagues would accept
as being the maximum shortfall between the new and the standard techniques in terms
of effectiveness. An equivalence study is just a two-sided version of the non-inferiority
study (in which as well as demonstrating that the new surgery is no worse by a speci-
fied amount, you also show that it is no better than by a specified amount). It should be
realized that non-inferiority studies are statistically complex and fairly controversial, for
several reasons.
◆ The choice of non-inferiority margin is usually difficult. Just as choosing the
‘minimal clinically important difference’ (MCID) in a superiority trial creates
challenges, likewise trying to work out what ‘proportion of the standard treatment
effect’ should be preserved is difficult. The US Food and Drug Administration (FDA),
usually in the context of drug trials, suggests that preserving 50% of an established
treatment effect is a minimum.
◆ The structure of the hypotheses is reversed in a non-inferiority trial—the null

hypothesis is that they are different, and the alternative hypothesis is that they are the
same—and hence the most conservative analysis (that which is least likely to reject
the null hypothesis) is not the usually mandatory intention to treat (ITT) analysis,
but some form of appropriately defined per protocol (PP) analysis. For example, one
might exclude those who did not comply meaningfully with the intervention (e.g.
missed the surgery) on the basis that if half those randomized to what in reality is a
better technique do not actually get it, it must dilute the treatment effect and make
it more, not less, likely to reject the null hypothesis. However, understandably, by
stepping away from the ITT analysis one becomes concerned about bias, and often
both the ITT and PP analyses are presented and any discrepancy between them is
commented on.
Crossover studies
The mainstay of surgical RCTs is the parallel group design, in which the patient is ran-
domized to one, and only one, surgery. There is another design, the crossover design,
in which a subject receives more than one intervention (say, A and B) and the order in
which the interventions is given is randomized (e.g. either A then B, or B then A). This
requires the participant to have stable disease and there must be the prospect of a return
to the original state between treatments; for example, having a hip replacement would not
qualify. However, it would be possible if, say, the surgery was for symptomatic relief and
the symptoms tended to return within a finite period.
Cluster randomized control trials

Sometimes it is not possible to conduct an RCT with the individual as the ‘unit of ran-
domization’. For example, logistically it may be impossible for an operating theatre to be
prepared for two quite different surgical techniques during the same operating list, and
so instead the hospital becomes the unit of randomization, and for a period all patients
within that hospital receive the same operation. This is an example of a cluster RCT, and
special statistical techniques are needed to design and analyse such studies. Note that with
such a cluster design it is quite natural to then use a crossover design with, say, all patients
in the first year getting intervention A, and all those in year 2 getting intervention B,
or vice versa. The advantage of such a crossover design is that each site acts as its own
‘control’ and hence it might be more statistically efficient, since over time the same site
might tend to be more alike than different sites.
Stepped wedge designs
The stepped wedge design is a variant of the cluster–crossover design (you rarely see
clusters of size 1, which would correspond to the individually randomized case). Just as
in individually randomized trials, there is an issue that crossover is impossible (e.g. for
surgery for a broken leg), so you might have the problem that once the surgeons and
operating staff have learned the new technique, they cannot unlearn it to then switch back
to the standard intervention. In such a circumstance you can do better than a completely
uncontrolled study by undertaking a stepped wedge design, in which the new technique
always follows the old technique, but the time of transition from one to the other is
randomized.
Stepped wedge designs are particularly suited to evaluating a change that is going to
happen regardless, and this is an opportunity to obtain reliable information on efficacy,
safety, costs, and acceptability before widespread uptake precludes research.
Observational studies
As indicated previously, much valuable information can be obtained from observational
data which, if carefully collected and sensibly analysed, can provide insights into effective-
ness and safety that can, if feasible, then be confirmed in randomized trials. Indeed, for
safety data (the incidence of perhaps rare complications of surgery), observational data
(or systematic overviews of observational data combined with RCTs) may be the only
way to obtain reliable evidence. The interested reader should consult an epidemiologist
or a good text on epidemiology (Rothman et al. 2008). In general terms, however, obser-
vational studies can be broadly divided into the two following types—cohort studies and
case–control studies.
Cohort studies
Cohort studies look at a defined set of subjects, for example all surgical patients on a ward
during a single list or all surgical procedures in Scotland in a 5-year period. It may be
cross-sectional or it may be longitudinal (outcomes measured over a period of follow-up,
e.g. mortality in the 20 years following hip replacement surgery). The critical questions
that the reader needs to assess are (a) the quality of the data and (b) whether analyses have
been appropriately adjusted for confounding.
The quality of the data This primarily relates to the accuracy and completeness of the
data. Suppose that one is looking at the complication rate for an operation performed
within a specified hospital. However, only two of the four surgeons who perform this
operation actually complete the online web registry. In addition, it may be that most of the
complications only become apparent after discharge home and therefore are missed if the
form is filled in at the time of discharge. Suppose, in addition, that one of the surgeons
only completes the online form for operations that he/she thinks have gone well. Although
this is clearly overstated, one can see how such a scenario might lead to data which are
neither complete nor accurate.
Residual confounding The second issue—whether the analyses have made adequate
adjustment for confounding, or whether there is a worrying level of residual confounding
left which may influence the findings—is more problematic. Confounding arises when
there is a common variant on the causal pathway between exposure and outcome. For
example, suppose that we want to compare two surgical techniques A and B, and we know
that the risk of a poor outcome is heavily influenced by age. In addition, younger people
tend to undergo procedure A, whereas older people tend to be treated with procedure B.
Any comparison of A and B that does not adjust for age will assign a difference in age to a
difference between A and B. There are a number of difficulties with this scenario:
◆ all the relevant confounders may not be known;
◆ even if the relevant confounders are know, it may not be possible to measure their
effect accurately;
◆ it may not be possible to identify the correct functional form relating the confounder
to the underlying risk of outcome (e.g. the relationship may be linear with age, or
more complex such as an exponential, a quadratic, etc.).
National registry data and record linkage Over the last 10 years exciting possibilities
have emerged of undertaking population-level studies to link all aspects of patient data
across various administrative databases holding routine healthcare data. For example, all
Danish residents have a personal identifier which is unique and used across all the nation-
al healthcare databases, such as the Danish National Hospital Registry, which holds data
on all admissions and outpatient visits since 1978 (Andersen et al. 1999), and the Danish
Register of Medicinal Product Statistics, which holds data on all prescriptions filled since
1995 (Gaist et al. 1997). The high quality (in terms of completeness and accuracy) and
duration of the data permit definitive epidemiological investigations of all types of
research questions which would be beyond prospectively collected individual studies.
An advantage of the Danish system is that ethical approval to undertake such record link-
age studies is not required. In Scotland, the Information Services Division (ISD) of the
Scottish Government holds data on all major healthcare events (hospitalizations, cancers,
prescriptions, etc.) linked by a unique Community Health Index (CHI) number, and
researchers are building the infrastructure and processes to access these records securely
and confidentially (see, for example, the Scottish Health Informatics Programme (SHIP)
at http://www.scot-ship.ac.uk>). When mature, the possibilities for surgical research are
tremendous.
Long-term follow-up of clinical trials A useful way of capitalizing on the power of
healthcare administrative databases is to use them to investigate the long-term efficacy
and safety profile of interventions tested in RCTs. For example, the Knee Arthroscopy
Trial (KAT) led by Aberdeen and Oxford Universities and funded by the National Institute
for Health Research (NIHR) Health Technology Assessment (HTA) programme was unu-
sual in that the original primary outcome was at 10 years after randomization and follow-
up for the rest of the participant’s life is underway using record linkage. Although not a
surgical trial, an excellent example of using record linkage of NHS Scotland databases to
discern long-term treatment effects of statins can be found in Ian Ford’s West of Scotland
Coronary Prevention Study (WOSCOPS) long-term follow-up paper (Ford et al. 2007).
Specialist surgical registries As well as national administrative registries, whose advan-
tage is their completeness and coverage of the whole population going back over many
decades, there is still a need for more specialist registries, holding a depth of information
(e.g. on specific risk factors, peri-operative details such as surgical teams, operation
length, and short- and longer-term outcomes of specific interest such as infection, wound
dehiscence, etc.). There are many good examples of these (such as the National Joint
Registry <http://www.njrcentre.org.uk>), and flagship funders such as NIHR HTA are
aware of the benefits of such registries (they have just funded a 17-year registry project for
bariatric surgery for obesity—the SCOTT project).
Case–control studies
The case–control study is a popular design, and if the right analyses are done well on a
high-quality dataset, it can unearth hugely important insights over a range of outcomes.
In contrast with the cohort study, the case–control study specifically identifies cases with
the characteristic of interest (e.g. success in an operation, or a rare side effect in a routine
operation). It then selects several controls that are as similar as possible to the cases (e.g.
matched on gender and within, say, 5 years of age of the case) except that they do not have
the characteristic of interest. In such a way you can then explore what other character-
istics might influence who became a case (e.g. a genetic trait or a lifestyle issue such as
smoking).
Pilot/feasibility studies
Pilot and/or feasibility studies are usually studies on a relatively small scale that are under-
taken to resolve important uncertainties before undertaking a larger-scale definitive
study. A useful review is given by Arain et al. (2010). Although pilot studies are potentially
hugely important in developing the best designs for the right questions, particularly for
complex intervention trials involving surgery which do not have the clear-cut and well-
defined developmental pathway of regulatory drug trials, there is relatively little in the
literature about their purpose, design, conduct, analysis, and reporting. Too often pilots
are poorly thought through and sometimes are only reported if they throw up something
unexpectedly ‘significant’. A common mistake is to extract a ‘treatment effect’ from a pilot
study and then plug that point estimate into the sample-size calculation for the definitive
larger study. Treatment effects from a pilot can be very misleading. Usually the partici-
pants recruited will be atypical and the site will be an expert centre, and the intervention
and outcomes, as well as the control comparator, if applicable, may be poorly developed
and still maturing. Treatment effects from pilots should be expressed with 95% confi-
dence intervals and that interval used, at most, as a ‘plausible range’.
Sample size—the difficulty of underpowered studies

The medical literature is bedevilled by the problem of small studies. That is not to say that
small studies are always underpowered. A study with only a handful of participants can
be adequately powered, depending on its research objective. For example, demonstrating
that a new surgical technique halves the operative time may only need ten subjects in each
randomized group (assuming that an RCT is actually needed at all), whereas demonstrat-
ing that this new technique offers a reduction in some surrogate outcome such as techni-
cal success or complication rate may need several hundred per group, and showing that
it affords an improvement in recurrence or even survival may need several thousand to

be randomized. The problem of ‘small studies’ is when the study is too small (in terms of
statistical information) to provide a sufficiently precise answer to the question of interest.
Note that often the number of participants is a poor proxy for the actual statistical size
of the trial. For example, a vaccine study with 100,000 randomized to each arm might
superficially appear orders of magnitude greater than a trial looking at 40 patients with
advanced brain tumours in a palliative setting. But if the primary outcome in the vaccine
trial is infections and only 20 are expected in 200,000 recruited, whereas in the cancer
trial it is time to death when everyone is likely to die within 12 months, the cancer trial is
actually the larger trial in terms of information gained from the ‘events’—the number of
subjects is a means to an end.
There are many reasons why studies are smaller than they need to be, including the
following:
a) they were designed that way because of over-optimistic evaluation of the size of
the treatment effect supported by the literature, or the complete absence of such a
literature;
b) the general difficulties of recruiting and retaining trial participants in all kinds of
studies, including randomized trials, leading to studies failing to reach their original
target size;
c) the problem of publication bias, wherein a small study with a spectacular finding
is published and the next of similar size which does not find anything positive is
rejected.
The last of these is a particular problem, which then can be exacerbated by meta-analysis
apparently compensating for the lack of adequate size by joining together a group of selec-
tively published small trials to create a possibly utterly false conviction that a treatment
or intervention works, and moreover works well, when in fact there is little evidence of
benefit if all the evidence were available and included.
Internal and external validity

Key questions when reading a report of a clinical study (whether a randomized trial or
an observational design) are to consider carefully the validity of the findings. This is rela-
tively straightforward for a randomized trial, but is often much harder for an observa-
tional study.
Internal validity
Internal validity deals with whether the findings are supported by the data. In a random-
ized trial, randomization should usually deal effectively with the problem of confound-
ing (the major issue in non-randomized studies, with confounding that is not adequately
adjusted for being called residual confounding); hence as long as an appropriate method
of statistical analysis has been used, the major threat to internal validity in an RCT is usu-
ally missing data.
External validity
Assuming that a study has appropriate internal validity, external validity deals with how
reproducible the findings are likely to be in the widest group of patients that they might
be applied to. For example, for understandable reasons all randomized trials have lists of
inclusion and exclusion criteria (e.g. PRECIS (pragmatic–explanatory continuum indica-
tors tool) developed by Thorpe et al. (2009)) to make sure that unnecessary safety risks are
not taken. However, by excluding those at highest risk and/or those with comorbidities or
medical histories, the question naturally arises as to how well both the safety and efficacy
findings translate to these excluded subjects. In addition, it is well known that even with
the eligible participants, those motivated to participate may differ from those who decline
to participate, and that in turn may influence the findings. In addition, in complex inter-
vention studies such as surgery, the surgeons and the centres in which they operate may be
atypical—the surgeons may be expert, and operating in enhanced environments with the
best equipment and policies. On the other hand by being expert centres they may get refer-
rals that are more challenging than the norm. All these and other relevant factors need to be
weighed up when considering how informative the findings are for the most diverse set of
patients in the most diverse healthcare settings to which the findings might be generalized.
Masking randomized treatment (‘blinding’)

A methodological cornerstone of the randomized trial is, wherever possible, keeping
the stakeholders unaware of what they are going to give or receive (before the fact) and
likewise maintaining that ignorance after the intervention has been given until all the
important outcomes are assessed. The obvious worry is that if knowledge of the sequence
of allocation is known beforehand, that might influence the order in which patients
are presented for randomization (the surgeon believes that one technique is superior to
the other and so gives the sickest, or the healthiest, patients what he/she perceives to be
the better intervention). Likewise, after the surgery is done, there may be a tendency to
review the evidence (e.g. imaging scans for recurrence, or adjudicating technical success)
for ascertainment of outcomes in a biased way if they are aware of what has been received.
There is a useful (and amusing) series of papers by Schulz and Grimes (2002) on random-
ization and masking.
Allocation concealment
This refers to the randomization process itself. Nowadays, in most circumstances, there is
little good reason that the allocation of randomized intervention cannot be done remotely
by phoning or accessing a web application set up by an independent third party (e.g.
a clinical trials unit). It has been established for many years that sealed envelopes, for
example, are vulnerable.
Maintaining the blind

Many surgical trials (and, more generally, complex intervention trials of all descrip-
tions) suffer from a prejudice that by naturally not being able to blind the receipt of the
intervention (surgery of a particular type against, say, a no-surgery comparator—the
medical management example in Fig. 1.2), they are somehow ‘second best’ to evidence
generated in a placebo-controlled drug trial. This is indeed a prejudice, since although it
is better if the intervention is unknown (lest it might influence the quality with which the
intervention is delivered, and then the subsequent management of the patient up to the
point post-surgery that the final outcomes are assessed), it is this last issue that is often of
central importance—the assessment of outcomes by someone unaware of the treatment
allocated.
Blinded assessment of outcomes

Even for these surgical interventions that are not possible to blind, the blinded assessment
of the outcomes is very achievable. In addition, the final outcome will often be assessed a
long time after the surgical intervention (e.g. 12 months post-operation) by which time
there may be no visible cues or other information to give a clue as to what they have
received—unlike a drug intervention which, although perfectly blinded in theory with
double dummy placebo tablets, may have signature side effects right up to the assessment
of the final outcomes which may seriously undermine the blind.
Moseley, J.B., O’Malley, K., Petersen, N.J., et al. (2002) A controlled trial of arthroscopic surgery for
osteoarthritis of the knee. New England Journal of Medicine, 347, 81–8.
Campbell, M.K., Skea, Z.C., Sutherland, A.G., et al. (2010) Effectiveness and cost-effectiveness of
arthroscopic lavage in the treatment of osteoarthritis of the knee: a mixed methods study of the
feasibility of conducting a surgical placebo-controlled trial (the KORAL study). Health Technology
Assessment, 14(5), 1–180.
Rothman, K.J., Greenland, S., and Lash, T.L. (2008) Modern Epidemiology (3rd edn). Philadelphia, PA:
Lippincott–Williams & Wilkins.
Andersen, T.F., Madsen, M., Jørgensen, J., Mellemkjoer, L., and Olsen JH. (1999) The Danish National
Hospital Register. A valuable source of data for modern health sciences. Danish Medical Bulletin,
46, 263–8.
Gaist, D., Sørensen, H.T., and Hallas, J. (1997) The Danish prescription registries. Danish Medical
Bulletin, 44, 445–8.
Ford, I., Murray, H., Packard, C.J. et al. (2007) Long-term follow-up of the West of Scotland coronary
prevention study. New England Journal of Medicine, 357, 1477–86.
Arain, M., Campbell, M.J., Cooper, C.L., and Lancaster, G.A. (2010) What is a pilot or feasibility study?
A review of current practice and editorial policy. BMC Medical Research Methodology, 10, 67.
Thorpe, K.E., Zwarenstein, M., Oxman, A.D., et al. (2009) A pragmatic–explanatory continuum indicator
summary (PRECIS): a tool to help trial designers. Journal of Clinical Epidemiology, 62, 464–75.
Schulz, K.F. and Grimes, D.A. (2002) Allocation concealment in randomised trials: defending against
deciphering. Lancet, 359: 614–18.
Schulz, K.F. and Grimes, D.A. (2002) Generation of allocation sequences in randomised trials: chance,
not choice. Lancet, 359, 515–19.
Schulz, K.F. and Grimes, D.A. (2002) Unequal group sizes in randomised trials: guarding against
guessing. Lancet, 359, 966–70.
Medical Research Council (2000) Framework for development and evaluation of RCTs for
complex interventions to improve health. <http://www.mrc.ac.uk/Utilities/Documentrecord/index.
htm?d=MRC003372>
1.4 Issues in the conduct of randomized controlled trials

Trial governance
The ethical, regulatory, and legal framework in which randomized trials are conducted
has become considerably more exacting since the European Union (EU) Clinical Trials
Directive (for Clinical Trials of Investigational Medicinal Products (CTIMPs)) took effect
in 2002. Most funders (e.g. the UK Clinical Research Collaboration (UKCRC) NIHR
HTA programme) mandate the involvement of a registered Clinical Trials Unit to help
with the design, conduct, analysis, and reporting of major trials.
Trial committees
Safety is the fundamental concern in any research project—that by participating in
the research project the subject is not put in harm’s way unnecessarily. Two commit-
tees help in this task—the Trial Steering Committee (TSC), chaired by an independent
individual, and an independent Data Monitoring Committee (iDMC) which, unlike the
TSC, is completely independent and may scrutinize any data emerging from the trial as
it progresses in full knowledge of what the patients have received in terms of randomized
interventions.
Safety reporting
As indicated, safety is paramount. There needs to be rigorous, timely and accurate infor-
mation on any emerging safety issues, be they either unexpected harms or an unaccept-
ably high incidence of expected (known) harms, and whether they are peri-operative or
longer-term (including perhaps psychological distress caused by the surgical procedures,
cosmesis, and quality-of-life issues). Emerging safety issues also incorporate external
information from other contemporaneous trials and/or new meta-analyses of the existing
literature. The real-time monitoring of safety within a randomized trial, and the weighing
up of emerging safety issues alongside potential signals on efficacy by the iDMC and TSC,
allow the risk–benefit trade-off to be determined and the trial to proceed ethically.
Delivering the intervention

Standardization
Researchers have a natural desire to try and standardize the intervention when designing
complex intervention studies. Unless the intervention is delivered to some reasonable
degree of fidelity and integrity, or to an acceptable level of quality and skill, how is it
either ethical (safe) or, beyond that, how would the trial demonstrate the true possibili-
ties of, for example, a new surgical technique? In a drug trial, all the randomized study
medication is manufactured, distributed, and prescribed to an identical specification and
an ultra-high degree of quality control—the only issue is getting patients to comply with
their randomized medication. However, ‘standardizing’ a surgical technique is obviously
more complicated and in some respects not feasible. There are complicated issues about
surgeon skill and experience, and also about creating a ‘standardized’ environment for the
ISSUES IN THE CONDUCT OF RCTS 17
operations to take place in the context of a multi-centre trial. One has to be careful that
in pursuing standardization one does not lose what is real and individual to each surgeon
and their surroundings that enables him/her to deliver the best care to individual patients.
Too much obsession with standardization could lead to ‘standardization of the outcomes’
to possibly a lowest common denominator. What is less controversial is, as far as possible,
standardizing the definition and measurement of the study outcomes.
Learning curves
Probably the most keenly appreciated and methodologically developed issue regard-
ing the variability inherent in delivering surgical interventions is that generated by the
range of performance from the surgeons themselves. This combination of experience and
underlying skill is captured in the concept of the learning curve. It is quite usual to see
trials requiring that a surgeon has undertaken, for example, ‘x procedures within the last
y months’ in an attempt to ensure that he/she is on or near some ‘plateau’ on the learning
curve. The idea is to compare the surgical techniques on a level playing field, when all
have been ‘learned’ to a comparable degree. However, this is complicated terrain. If the
new surgical technique is particularly difficult and/or is constantly innovating, achiev-
ing the ‘plateau’ may be time-consuming, or indeed there may be considerable between-
surgeon variability in what plateau is achieved. This makes it perhaps questionable as
to whether the dynamics of achieving ‘the learning curve’ should be discounted when
considering the relative benefits of the two surgical approaches. In addition, often little
attention is spent on the issue of falling performance as surgeons either temporarily or
permanently deteriorate from the achieved competence.
Expertise-based trials
One way of apparently being able to sidestep issues of learning curves—or more specifi-
cally, the problem that if you are comparing a new technique with a standard technique,
almost certainly a surgeon will be at different points on the learning curve with the two
surgical techniques—is to allow surgeons who are expert in a technique to undertake
only the operations that are within their expertise. These designs have advantages, but are
controversial—see Devereaux et al. (2005) for a useful discussion.
Devereaux, P.J., Bhandari, M., Clarke, M., et al. (2005) Need for expertise based randomised controlled
trials. British Medical Journal, 330, 88.
1.5 Analysis of surgical trials

Although all randomized trials share essential methodological details that dictate similar
analytical strategies, there are some specific challenges in surgical trials.
Randomization
Randomization may pose specific additional challenges within the context of a surgical
trial. Unlike drug trials, in which it is entirely feasible to actually supervise the initiation
of the first dose of randomized medication within moments of randomization, with sur-
gery a good deal of thought has to be given as to when is the best moment to randomize.
Ideally, it should be sufficiently close to the date of surgery to minimize losses to receiv-
ing the randomized allocation, but on the other hand sufficiently far ahead to allow the
patient and the surgery to be properly organized. However, in the present-day NHS, for
example, it may well be that there will be a delay of several weeks, if not months, between
randomization and the date of surgery. During this time patients may change status and/
or operations may be cancelled, or patients may change their mind or get the operation
done privately. This becomes a particular issue if the dynamics of receiving the interven-
tion are different for the comparator (and often most challenging if that comparator is
medical management, not surgery).
Intervention allocated versus intervention received

It can happen in any trial that what the patient is randomized to is not what they receive.
Mis-randomizations can occur due to confusions arising over drug labelling, for example,
with the wrong pack being assigned. However, surgical trials, in addition to being vul-
nerable to administrative blunders, can also have the surgeon decide to deviate from the
randomized allocation either at the outset or possibly during a procedure if difficulties
are encountered (they may ‘lose equipose’ and switch out of the new procedure into safer,
better trod ground). This, of course, complicates the analysis.
‘Crossover’ or rescue surgery

Just as in drug trials, where if a patient shows no signs of efficacy and/or has intolerable
side effects he/she may expect to be put on second-line medication (which may include
one of the other randomized options), so in a surgical RCT if the randomized interven-
tion is not successful or there is a relapse, in time the patient may receive the experimental
or the standard surgery, or other procedures. Again, this type of information may or may
not be taken into account, depending on the purpose of the analysis.
Analysis sets
The above considerations result in the need to define clearly how the outcomes for
groups or ‘analysis sets’ are handled. The mainstay of most analyses of randomized
trial data is the ‘intention-to-treat’ (ITT) grouping, since under randomization it is unbi-
ased, i.e. everyone is analysed according to what they were randomized to, regardless of
ANALYSIS OF SURGICAL TRIALS 19
subsequent behaviour. If a new surgery is superior, but for various reasons some do not
get this despite being randomized to this superior option, but instead get the inferior
comparator, the ITT analysis will underestimate the real effect. This is considered conser-
vative and, as indicated, unbiased.
However, if the new surgical option is not as good or is frankly harmful (and several
patients receive the older and better surgical treatment), an ITT analysis may underesti-
mate the real harm being done by the newer intervention. Therefore, for safety analyses,
it is usually considered more important to analyse according to what was received (as
treated) rather than a strict ITT analysis which might underestimate the real harm being
done. However, it is understood that an ‘as-treated’ analysis may also contain bias, given
that less healthy patients or those with more difficult presentations are more likely to cross
over to the older more established procedure. This is considered better than missing a
safety issue by underestimating it.
A further variant often seen is a PP analysis, which restricts the analysis to those whose
operations followed the protocol is some clearly defined way. It would be usual for subjects
that fell foul of pre-specified protocol violation rules to be excluded as ‘major protocol
violators’. Note that in the exceptional case of non-inferiority or equivalence designs, a
suitably defined PP analysis is usually seen as the primary analysis over the otherwise ubiq-
uitous ITT approach. However, it is common to present both analyses for such designs.
Subgroup analyses
Subgroup analyses are a common feature of randomized trial reports. Having found (or
not found) an overall effect of an intervention such as a new surgical technique versus
standard surgery, it is natural to ask whether the technique works better or worse, or is
safer, or more cost effective, or better tolerated in subsets of patients who share certain
characteristics (e.g. females, or older patients, or those with previous surgery, or more
advanced disease, and so on). There are a number of good articles on the rules for under-
taking and interpreting subgroups (e.g. Rothwell 2005). These should be clearly labelled
as confirmatory—pre-specified in the trial protocol to confirm an association seen in
previous datasets and/or a relationship with a clear biological rationale—or exploratory
(or post-hoc)—interesting but arising from a general search (‘fishing expedition’) in
the dataset at hand. Subgroup analyses, even when pre-specified, usually lack power to
detect anything other than very substantial interactions between the subgroup and the
effect of intervention. Such intervention modifiers (treatment by subgroup interactions)
are sometimes differentiated between quantitative effect modification (in which, say, the
effect on males and females is in the same direction, but is quite different in magnitude)
and qualitative effect modification (in which the direction of effect is different (e.g. benefit
in females and harm in males)).
Causal modelling
A key feature of the above subgroup analyses is that the subgroups must be defined com-
pletely by information known before randomization (e.g. gender, or severity of disease).
However, it is quite natural to wonder how the effect of the intervention might have varied
according to what happened after randomization. In a drug trial, the most obvious ques-
tion might be how good the drug is in people that actually took it, i.e. fully complied with
their randomized medication. However, such analyses are statistically fraught. Once you
define membership of the levels of subgroup by information post-randomization, there
may be serious and subtle confounding with the effect of treatment. For example, suppose
that you wish to estimate the effect of the new surgical technique in those who successfully
underwent the procedure without complications. But who do you compare these select
subjects with? If you compare them with all those who received the standard surgery,
the comparison might be biased by omitting those who had complications of the new
surgery. If you remove those who had complications on the old surgery to try to balance
things up, that might not work either. You would need to assume that the complications
for the two techniques were similar and tended to occur in the same people. However,
there is an emerging methodology to try to deal with these issues in a proper statistical
manner—using causal models. These are complex statistical models which, under strong
assumptions, try to create a counterfactual group in the comparator arm that would have
behaved like the experimental group if they had been randomized to that group instead
of the control group.
Missing data
Missing data is an important issue that may compromise the internal validity of a study.
There are many reasons why important pieces of information are unavailable, although
not infrequently it is impossible to establish a precise reason why certain data items are
missing. It is important to try and establish the likely reasons for missing data items, as
analysis strategies to deal with this issue, such as imputing the missing value, inevitably
make assumptions about the missing data. Without knowing why the data are missing, it is
very difficult to be reassured that these assumptions are tenable. A very readable and com-
prehensive guide to missing data issues in clinical trials is given by Carpenter and Kenward
(2007) and the associated website <http://www.missingdata.org.uk>. For a discussion of
missing data in venous thromboembolism after surgery trials, see Norrie (2007).
Avoiding missing data

The golden rule about missing data is to try as hard as possible to avoid it. This means hav-
ing the best design, rather than hoping that analysis tricks will save the day. Good design
means asking the right question, in the right way, at the right time, and using the best
methodology. For example, many publicly funded surgical trials (e.g. from the UK NIHR
HTA programme) will have patient-reported outcomes such as quality of life as important
outcomes. Such outcomes may be recorded one or two years after the surgery. Retaining
the interest and involvement of the patient is critical. It is unlikely that just contacting
the participants out of the blue two years later will maximize the response rate; there is
strong evidence that providing small gifts (e.g. £10) for participating, unconditional on
returning the forms, improves response rate (Edwards 2002). Sending the participant a
telephone book width of forms to fill in is also likely to meet with a predictable response.
ANALYSIS OF SURGICAL TRIALS 21
Asking sensitive questions (e.g. about sexual dysfunction after prostate surgery) needs
careful handling.
Monitoring missing data

It is important that an unacceptably high level of missing data does not come as a sur-
prise when doing the final analysis. Levels of incompleteness should be continuously
monitored as part of the overall data management and quality control plans within a
randomized trial as the study data accumulate. The study team needs to be able to identify
problems with missing data reliably and be prepared to implement flexible solutions to
overcome this.
Missing completely at random, missing at random,

and informative missingness
Missing completely at random Missing completely at random is when the probability
of missing does not depend on observed or unobserved measurements. Under this
unlikely assumption, the very popular ‘complete case’ analysis (in which the missing data
issue is ignored, and just the data collected are analysed) and the also common ‘last obser-
vation carried forward’ (for serial measurements) are valid.
Missing at random The next level up (and more plausible than missing completely at
random) is ‘missing at random’, in which the probability of missing can depend on the
observed data but not on the unobserved data (e.g. the missing data mechanism can
depend on factors such as age and gender, but not on outcome). Here the usual analysis
technique is multiple imputation.
Informative missingness The final and usually most plausible scenario is ‘not missing
at random’ (also known as ‘informative, or non-ignorable missingness’), in which the
mechanism for missing depends not only on the observed data but also on the unob-
served data. This would be the case, for example, with those in a surgery trial who died or
suffered a debilitating stroke and were so cognitively impaired that they could not com-
plete the forms. Then the missing data would very probably depend on the unobserved
data—the outcome (e.g. quality of life at 1 year). Informative missingness needs the most
complex (and in terms of assumptions, often the most heroic) statistical modelling,
including pattern mixture modelling.
Rothwell, P.M. (2005) Subgroup analysis in randomised controlled trials: importance, indications, and
interpretation. Lancet, 365, 176–86.
Carpenter, J.R. and Kenward, M.G. (2007) Missing Data in Randomized Controlled Trials—A Practical
Guide. Final Report to the National Health Service, United Kingdom. <http://www.hta.nhs.uk/
nihrmethodology/reports/1589.pdf>
Norrie, J. (2007) Trials of venous thromboembolism prevention. Lancet, 370, 915–17.
Edwards, P., Roberts, I., Clarke, M., et al. (2002) Increasing response rates to postal questionnaires: a
systematic review. British Medical Journal, 324, 1183.
Little, R.J.A. and Rubin, D.B. (2002) Statistical Analysis with Missing Data (2nd edn). New York: Wiley.
1.6 Other issues in surgical trials

Reporting
There have been improvements in the way clinical trials have been reported as a result of
the CONSORT (Consolidated Standards of Reporting of Clinical Trials) initiative, recently
updated in 2010. It is now obligatory to follow these guidelines for publication in a top
peer-reviewed journal. Researchers also need to comply with the CONSORT statement
on reporting of harms. There is no specific guidance for surgical trials but Zwarenstein
et al. (2008) have published additional information on pragmatic trials.
Trial registration
As previously discussed, the failure to report findings in clinical trials is a form of publica-
tion bias. It is becoming mandatory to register clinical trials prospectively, ideally before
the first patient is randomized, on an international registry such as <http://www.clinictri-
als.gov> (a US NIH-based site) or <http://www.controlled-trials.com> (European-based).
Protocol publication
Registration should be seen as a minimum requirement. It is much better, and helpful to
fellow researchers, to publish the full protocol (including if possible interesting method-
ological innovations and how problems were solved, as well as details of statistical analysis
plans, data management plans, trial monitoring plans, quality control processes, etc.) in
peer-reviewed open-access journals (e.g. the journal Trials).
Systematic reviews and meta-analysis

It is essential when planning a trial, or when trying to interpret a trial’s findings, that due
weight be given to any rigorous and high-quality summary of the overall evidence. The
reader should consult the Cochrane Collaboration for extensive details. The Cochrane
Handbook is an invaluable resource of methodological detail and insight into all aspects
of conducting and interpreting meta-analyses.
Zwarenstein, M., Treweek, S., Gagnier, J.J., et al. (2008) Improving the reporting of pragmatic trials: an
extension of the CONSORT statement. British Medical Journal, 337, a2390.
CONSORT initiative website <http:// www.consort-statement.org/home>
Cochrane Collaboration website <http://www.cochrane.org>
Cochrane Handbook <http://www.cochrane-handbook.org>
Trials journal website <http://www.trialsjournal.com>
SUMMARY 23
1.7 Summary
There have been very substantial developments in surgical science over the last several
decades. Although randomized trials have played an important part in many of these
advances, it is hoped that the randomized controlled trial will have an even greater part
to play in the future. Surgical trial methodology has itself also developed, rising to the
challenges posed by the huge variety and complexity of research questions in surgery. This
methodology will continue to evolve as researchers learn from trials that worked as much
as from trials that did not work.
So, as you read through this collection of landmark papers, it is important to critically
appraise relevant sections, asking yourself how you would design that trial now, and how
it could be improved. The key to a successful and relevant clinical trial depends on asking
a relevant and important question. Thereafter, a carefully thought out design, coupled
with excellent conduct, analysis, and reporting are the critical elements to a successful
outcome.
Chapter 2
General surgery
2.1 Surgical skin preparation 26
2.2 Venous thromboembolism (VTE) prophylaxis 28
2.3 Intravenous fluid therapy after major abdominal surgery 30
2.4 Abdominal wall closure 32
2.5 Symptomatic groin hernias 34
2.6 Asymptomatic groin hernias 36
2.7 Laparoscopic versus open groin hernia repair 39
2.8 Local or general anaesthesia for groin hernia repair 42
2.9 Chronic groin pain after groin hernia surgery 44
2.10 The Lichtenstein repair 46
2.11 Open versus laparoscopic incisional hernia repair 47
2.12 Antibiotic prophylaxis in inguinal hernia repair 49
2.13 Surgery for ingrown toenails 51
2.14 Surgical volume and patient outcome 53
26 GENERAL SURGERY
2.1 Surgical skin preparation

Details of studies
Surgical wound infection is a common cause of post-operative morbidity. Although
pre-operative skin preparation with antiseptics has been practised since Lister introduced
the concept in the late 1800s, it is surprising that prior to the publication of this paper, no
published randomized trial had adequately investigated whether one antiseptic prepara-
tion was more effective than another in reducing the incidence of post-operative surgical
wound infection. In this study, the authors compared chlorhexidine–alcohol solution with
povidone–iodine skin preparation in patients undergoing clean-contaminated surgery.
Study references
Main study
Darouiche, R.O., Wall, M.J., Jr, Itani, K.M.F., et al. (2010) Chlorhexidine–alcohol versus povidone–
iodine for surgical-site antisepsis. New England Journal of Medicine, 362, 18–26.
Related references
Chaiyakunapruk, N., Veenstra, D.L,. Lipsky, B.A., and Saint, S. (2002) Chlorhexidine compared
with povidone–iodine solution for vascular catheter-site care: a meta-analysis. Annals of Internal
Medicine, 136, 792–801.
Study design
◆ A prospective multi-centre RCT.
Level of evidence 1b
Randomization Chlorhexidine–alcohol scrub versus povidone–iodine skin preparation
Number of centres 6
Number of patients 849
Setting US study between 2004 and 2008
Inclusion criteria Adults undergoing clean-contaminated surgery (i.e. gastrointestinal, biliary,
thoracic, gynaecological, or urological operations without substantial
spillage or unusual contamination)
Exclusion criteria Allergy to any of the skin preparations
Infection at or adjacent to the wound
Inability to follow-up for 30 days
Follow-up 30 days
◆ A special applicator was used to clean using the 2% chlorhexidine gluconate–70%

isopropyl alcohol solution (ChloraPrep, Cardinal Health).
◆ Povidone–iodine as a 10% aqueous solution was used as a scrub and then painted
onto the surgical site.
SURGICAL SKIN PREPARATION 27
Outcome measures
Primary endpoint
◆ Post-operative surgical wound infection within 30 days of surgery.
Secondary endpoints
◆ Individual’s type of infection (superficial, deep incisional, or organ space, which is
defined as infection involving any organ or space other than the incised surgical
wound).
Results
Chlorhexidine–alcohol Povidone–iodine p value
(n = 409) (n = 440)
Overall wound infection rate 9.5% 16.1% 0.004
Superficial wound infection rate 4.2% 8.6% 0.008
Deep wound infection rate 1% 3% 0.05
Organ space infection rate 4.4% 4.5% ns
Adverse events Similar in both groups
◆ The relative risk of developing any post-operative wound infection in the

chlorhexidine–alcohol group compared with the povidone–iodine group was
0.59 (CI 0.41–0.85).
◆ The estimated number needed to treat with the chlorhexidine–alcohol preparation
instead of povidone–iodine in order to prevent one post-operative wound infection
was 17.
Conclusions
Pre-operative skin preparation with chlorhexidine–alcohol is associated with a 41%
reduction in the post-operative wound infection rate after clean-contaminated surgery
compared with one of the most commonly used surgical skin preparations.
Critique
The results of this paper are consistent with an earlier meta-analysis which reported that
chlorhexidine-based solution resulted in a 49% reduction in the risk of vascular catheter-
related infection compared with 10% povidone–iodine solution (Chaiyakunapruk et al.
2002). The authors postulate that the beneficial action of chlorhexidine–alcohol is likely
to be related to its rapid action and persistent activity despite exposure to body fluids,
including blood, which can deactivate iodine containing antiseptics.
Although there are risks associated with alcohol-based skin preparations, the special
applicator used in this study limits the amount of preparation solution that is dispensed at
any one time, thereby reducing the likelihood of pooling.
28 GENERAL SURGERY
2.2 Venous thromboembolism (VTE) prophylaxis

Details of studies
The landmark paper describing the use of low-dose heparin to prevent fatal pulmonary
embolism was written in 1975 by Kakker et al.
Study references
Main study
Kakker, V.V., Corrigan, T.P., Fossard, D.P., et al. (1975) Prevention of fatal postoperative pulmonary
embolism by low doses of heparin: an international multicentre trial. Lancet, 306, 45–64.
Related reference
NICE (2010) Venous thromboembolism—reducing the risk: NICE guideline CG92, January 2010.
Study design
◆ Multi-centre RCT.
Randomization No prophylaxis versus low-dose subcutaneous heparin
Number of patients 4121 (28 centres)
Inclusion criteria Age >40 years
Operations under general anaesthesia ≥30 minutes duration
Hospital stay ≥7 days
Outcome measures Deep venous thrombosis (DVT), pulmonary embolism (PE), and fatal PE
Operative and post-operative haemorrhage
◆ The treatment group received heparin 5000U subcutaneously 2 hours pre-operatively

and every 8 hours until discharge.
Results
◆ The control and treatment groups were well matched:
Control (n = 2076) Low dose heparin p value

(n = 2045)
DVT 24.6% 7.7% <0.005
Operative haemorrhage 126 (6%) 182 (8.9%) ns
Post-operative haemorrhage 117 (5.6%) 158 (7.7%) <0.01
Total mortality 100 (5%) 80 (3.9%) ns
Fatal PE 16% 2% <0.005
VTE PROPHYLAXIS 29
◆ In addition to showing a significant reduction in the incidence of DVT and fatal PE,
it was noted that twice as many patients in the control arm as in the heparin arm
died from acute myocardial infarction (13 versus 7 patients), although this did not
reach statistical significance.
Conclusions
The authors concluded that low-dose subcutaneous heparin prevents fatal post-operative
PE, is well tolerated by the patient, requires no laboratory control, and does not produce
serious bleeding during surgery.
Critique
Venous thromboembolism (VTE) is an important cause of morbidity and mortality in
hospital patients. The cost and morbidity associated with non-fatal VTE is significant.
Kakker et al. (1975) provided the earliest robust evidence that prophylactic heparin can
prevent VTE. At the time there was much resistance to the idea, particularly amongst
surgeons in the USA and UK. There was widespread concern that such a practice would
increase the risk of peri-operative haemorrhage. Following publication, the paper came
under criticism as it was discovered that there were inadequacies in the data from one of
the 28 centres. The results were re-analysed without data from the affected centre and the
same conclusions were still drawn.
NICE guidelines now recommend that all hospital inpatients have their VTE risk
documented upon admission. For general surgical patients, mechanical VTE prophy-
laxis must be started on the day of admission and may include anti-embolism stockings,
foot impulse devices, and intermittent pneumatic compression. Pharmacological VTE
prophylaxis (usually a low-molecular-weight heparin) must also be prescribed for those
deemed to be at a low risk of bleeding during the procedure. Heparin may be continued
for up to 28 days for those with an underlying diagnosis of cancer.
30 GENERAL SURGERY
2.3 Intravenous fluid therapy after major abdominal surgery

Details of studies
Traditionally, surgeons and anaesthetists have tended to err on the side of caution when
prescribing peri-operative fluid regimens for patients undergoing major abdominal sur-
gery. During the 1980s and 1990s a number of publications highlighted the potential mor-
bidity of fluid overload after major surgery. This led to several important studies assessing
the effect of fluid restriction versus conventional fluid therapy during the peri-operative
period.
Study references
Main study
Lobo, D.N., Bostock, K.A., Neal, K.R., et al. (2002) Effect of salt and water balance on recovery of
gastrointestinal function after elective colonic resection: a randomised controlled trial. Lancet, 359,
1812–18.
Related references
Brandstrup, B., Tønnesen, H., Beier-Holgersen, R., et al. (2003) Effects of intravenous fluid restriction
on postoperative complications: comparison of two postoperative fluid regimens. A randomized
assessor-blinded multicentre trial. Annals of Surgery, 238, 641–8.
MacKay, G., Fearon, K., McConnachie, A., et al. (2006) Randomized clinical trial of the effect of
postoperative intravenous fluid restriction on recovery after elective colorectal surgery. British
Journal of Surgery, 93, 1469–74.
Study design
◆ A prospective single-institution RCT.
Randomization Standard post-operative fluids versus fluid restriction
Setting Large teaching hospital in England
Inclusion criteria Patients undergoing colonic resection for cancer
Primary endpoint Solid- and liquid-phase gastric emptying time on day 4
Secondary endpoints Time to first flatus and bowel movement
Time to discontinuation of intravenous fluids
Time to resumption of solid diet
Hospital stay
◆ ‘Standard’ fluids consisted of 3L of water per day (2L 5% dextrose and 1L saline 0.9%
containing 154mmol sodium).
◆ The fluids restriction regime consisted of 2L of water per day (2L 4% dextrose–0.18%
saline containing 77mmol sodium).
IV FLUID THERAPY AFTER MAJOR ABDOMINAL SURGERY 31
◆ All patients received 40–60mmol potassium per day from the second
post-operative day.
◆ Gastric emptying was measured by dual-isotope radionuclide scintigraphy.
◆ Post-operative pain was managed using patient-controlled analgesia (PCA)
morphine (no epidurals were used).
Results
◆ The baseline characteristics between the two groups were similar:
Standard fluids Fluid restriction p value

(n = 10) (n = 10)
Liquid-phase gastric emptying (min) 110 73.5 0.017
Solid-phase gastric emptying (mins) 175 72.5 0.028
Passage of flatus (days) 4 3 0.001
Passage of first bowel movement (days) 6.5 4 0.001
Hospital stay (days) 9 6 0.001
◆ Patients in the standard group were on average 3kg heavier than those in the
restricted group during the post-operative period.
◆ There were significantly more complications in the standard group than in the
restricted group (7 versus 1, p = 0.01).
◆ Complications included peripheral oedema, hyponatraemia, vomiting, confusion,
and wound and respiratory infections.
Conclusions
The authors concluded that in patients undergoing routine colorectal surgery, an excess
of salt and water was associated with a delayed return of gastrointestinal function and
increased hospital stay and complication rate.
Critique
In 1999 a National Confidential Enquiry into Perioperative Deaths (NCEPOD) report
highlighted errors in fluid and electrolyte management as a significant cause of death.
This study was the first to provide evidence challenging the standard practice of intrave-
nous fluid administration during the peri-operative period. The numbers involved in the
study are small (n = 20) and the investigators deliberately looked at a homogenous group
of patients undergoing straightforward colonic resection.
Subsequent larger studies have not found such a dramatic benefit to fluid restriction,
perhaps related to lower volumes of fluid infused during the first 24 hours after surgery.
Fluid optimization with avoidance of fluid overload is now well established and is an
important element of enhanced recovery after surgery programmes.
32 GENERAL SURGERY
2.4 Abdominal wall closure

Details of studies
The development of an incisional hernia after abdominal wall closure can be associ-
ated with significant morbidity, pain, and the risk of incarceration and strangulation
of abdominal contents. This is a common problem, with several observational studies
reporting an incisional hernia rate of 10–20% following midline abdominal incisions.
There is no consensus regarding optimal suture material or surgical technique to close
the abdominal wall.
Study reference
Main studies
Krukowski, Z.H., Cusick, E.L., Engeset, J., and Matheson, N.A. (1987) Polydioxanone and
polypropylene for closure of midline abdominal incisions: a prospective comparative clinical trial.
British Journal of Surgery, 74, 828–30.
Bloemen, A., van Dooren, P., Huizinga, B.F., and Hoofwijk, A.G.M. (2011) Randomised clinical trial
comparing polypropylene or polydioxanone for midline abdominal wall closure. British Journal of
Surgery, 98, 633–9.
Related references
Jenkins, T.P. (1976) The burst abdominal wound: a mechanical approach. British Journal of Surgery,
63, 873–6.
Mudge, M. and Hughes, LE. (1985) Incisional hernia; a 10-year prospective study of incidence and
attitudes. British Journal of Surgery, 72: 70–1.
Study design
Krukowski et al. Bloemen et al.
Level of evidence 1b 1b
Randomization Continuous polydioxanone versus Polydioxanone versus
polypropylene for mass closure polypropylene for closure
of the abdominal wall of the abdominal fascia
Number of patients 757 456
Wound assessment Clinical examination Clinical examination and
ultrasound of wound
Primary outcome Incisional hernia
Secondary outcomes Wound infections
Wound pain
Sinus formation
Follow-up (months) 12 31
ABDOMINAL WALL CLOSURE 33
Results
Krukowski et al.
◆ Wound infection was less likely with polydioxanone than with polypropylene
(3.5% versus 7.0%, p <0.05).
◆ There was one dehiscence in each group.
◆ The incidence of incisional hernia was similar in the two groups (7.7% versus 9.7%).
Bloemen et al
◆ The incisional hernia rate was similar for the two groups (polydioxanone 24.9%
versus polypropylene 20.2%), although there was a trend favouring polypropylene.
◆ Incisional hernias were most likely to develop 13–18 months after surgery.
◆ No significant difference was seen in the development of wound infection and
suture sinus.
Conclusions
◆ Krukowski et al. concluded that absorbable polydioxanone is preferable to
non-absorbable suture because of a decreased wound infection rate.
◆ The most recent study by Bloemen et al. concludes that either type of suture can
be used, as it does not appear to have a bearing on incisional hernia rate.
Critique
The main concern regarding use of non-absorbable suture material (polypropylene) was
the potentially increased risk of deep-seated infection/suture sinus which may contribute
to the development of incisional hernias. However, Bloemen et al. demonstrated that the
type of suture does not influence the wound outcome.
The main concern with the study by Bloemen et al. was the fact that they reported a
high incidence of incisional hernia in both groups. This may have been related to the
longer follow-up, but it is also likely that the use of ultrasound to detect defects may have
picked up an increased number of sub-clinical hernias. However, it now seems clear that
the longer and harder you look for incisional hernias after abdominal surgery, the more
likely you are to find them.
34 GENERAL SURGERY
2.5 Symptomatic groin hernias

Details of studies
Prior to the publication of the paper by Gallegos et al. in 1991, traditional surgical dogma
suggested that groin hernias should be repaired because of the risk of strangulation.
However, the risk of this complication developing was unknown. The authors calculated
the cumulative probability of strangulation in relation to the length of the history.
Study references
Main study
Gallegos, N.C., Dawson, J., Jarvis, M., and Hobsley, M. (1991) Risk of strangulation in groin hernias.
Related references
Neuhauser, D. (1977) Elective inguinal herniorrhaphy versus truss in the elderly. In: Bunker, J.P.,
Barnes, B.A., and Mosteller, F (eds), Costs, Risks and Benefits of Surgery, pp. 223–39. New York:
Oxford University Press.
Study design
◆ A single-centre retrospective cohort study.
Number of patients 561 (data available on 476)
Setting Large UK teaching hospital
Time period 1987–1989
Follow-up 3 years
Inclusion criteria All emergency and elective inguinal and femoral hernias
Age >14 years
◆ The duration of herniation was obtained from the history in order to use the Kaplan–
Meier technique to estimate the probability of strangulation as a function of history.
Outcome measures
◆ Proportion of strangulated hernias.
Results
◆ The median age of patients was 65 years.
◆ 22/439 (5%) of all inguinal hernias repaired were strangulated over the study period.
◆ 12/37 (32%) of femoral hernias repaired were strangulated over the study period.
SYMPTOMATIC GROIN HERNIAS 35
◆ The cumulative probability of strangulation of an inguinal hernia was 2.8% (95% CI,
1.1–4.5%) at 3 months, 4.5% (95% CI, 1.3–7.7%) at 24 months, and 8.6% (95% CI,
2.6–14.6%) at 60 months.
◆ Approximately one-third of the cumulative risk of strangulation was reached within
the first 3 months of a hernia’s presence.
◆ The cumulative probability of strangulation for femoral hernias was 22% (95% CI,
8.3–35.7%) at 1 month and 45% (95% CI, 23–67%) at 21 months.
◆ The greatest proportion (approximately 50%) of the accumulated risk of a femoral
hernia strangulating was seen within 1 month of its appearance.
Conclusions
The cumulative probability of hernia strangulation was greatest in the first 3 months for
both inguinal and femoral hernias. Therefore the authors concluded that patients with a
short history of herniation should be referred urgently to hospital and given priority on
the waiting list.
Critique
This seminal paper on the risks of strangulation of groin hernia was a small retrospective
observational study that is now 21 years old. Prior to this, Neuhauser had shown that the
probability per year of an inguinal hernia strangulating was 0.00291. However, it had been
assumed that the risk was constant per year. Gallegos et al. challenged that evidence and
demonstrated that the risks were greatest in the first few months. The authors noted that
the main drawback of their study was that it was performed in a teaching hospital. This
may have led to under-reporting of some strangulated hernias as patients on the waiting
list for repair may have presented to their local hospitals. The study also failed to separate
the risk of small bowel strangulation from less significant forms of strangulation such as
strangulated omentum. The former may be associated with a significantly higher mortal-
ity compared with elective repair. It is also much more frequent in femoral than inguinal
hernias.
36 GENERAL SURGERY
2.6 Asymptomatic groin hernias

Details of studies
Traditional surgical doctrine has advocated an operative approach to managing both
symptomatic and asymptomatic hernias. Although a number of studies have provided
support for such an approach in symptomatic hernias, until recently convincing evidence
to support such an approach in asymptomatic hernias has been lacking.
The two landmark papers described here are similar in design. Both are prospective
randomized controlled trials comparing outcomes in men with asymptomatic inguinal
hernias who were assigned to surgical intervention or observation alone. The first study is
from Scotland and the second is from the USA.
Study references
Main studies
Chung, L., Norrie, J., and O’Dwyer, P.J. (2011) Long term follow up of patients with a painless inguinal
hernia from a randomized clinical trial. British Journal of Surgery, 98, 596–9.
Fitzgibbons, R.J., Giobbie-Hurder, A., Gibbs, J.O., et al. (2006) Watchful waiting vs repair of inguinal
hernia in minimally symptomatic men: a randomized clinical trial. JAMA, 295, 285–92.
Related references
O’Dwyer, P.J., Norrie, J., Alani, A., et al. (2006) Observation or operation for patients with an
asymptomatic inguinal hernia: a randomized clinical trial. Annals of Surgery, 244, 167–73.
Study design
Chung et al. Fitzgibbons et al.
Randomization Observation versus repair Observation versus repair
Definition of a hernia Visible lump in the groin Cough impulse in the groin
Follow-up 7.5 years 3.2 years
Primary outcomes Pain/crossover to operation Pain/discomfort
General health symptoms Interfering with activities
Change in health status
Secondary outcomes Cost to health service Complication rate
Intervention/complication rate Pain
Functional status
◆ Inclusion criteria were men with asymptomatic inguinal hernias

ASYMPTOMATIC GROIN HERNIAS 37
Results
Chung et al.
◆ 46/80 (57.5%) men allocated to the observation group crossed over to operation
during follow up.
◆ The conversion rate to surgery was 16% (95% CI, 9–26%) at 1 year, 54% (95% CI,
42–66%) at 5 years, and 72% (95% CI, 59–84%) at 7.5 years (Kaplan–Meier method).
◆ Pain was the most common reason for conversion. Less common reasons included a
reduced quality of life and an increase in hernia size.
◆ 2/80 (2.5%) in the observation group presented with acute symptoms over the study
period.
◆ The early follow-up data at 1 year showed no difference in pain scores (44% versus
37%, p = 0.42) or analgesia consumption (12% versus 22%, p = 0.14) seen between
the observation group and the operation group.
◆ There was no significant quality-adjusted life-year (QALY) gain for the surgery group
at 12 months.
Fitzgibbons et al.
◆ At 2 years, 85/366 (23.2%) of men assigned to the observation group crossed over to
the operation group (31% at 4 years).
◆ Pain was the most common reason for conversion to surgery.
◆ 62/358 (17.3%) of patients randomized to the operation group crossed over to the
observation group as they did not want to undergo surgery.
◆ There was no difference in pain that interfered with usual activities between the
observation group and the operation group (5.1% versus 2.2%, p = 0.52).
◆ A single patient (0.3%) in the observation group experienced an acute incarceration
at 2 years.
◆ At 2 years, patients in the operative group experienced a significantly higher
reduction in pain than those in the observation group.
◆ Similarly, the operative group showed a significantly greater improvement in their
ability to perform usual activities than the observation group
Conclusions
Chung et al. concluded from their study that the majority of asymptomatic hernias
develop pain over time, and therefore surgical intervention is recommended.
Fitzgibbons et al. concluded that observation only is a reasonable option for men with
an asymptomatic hernia because of the low risk of acute strangulation.
38 GENERAL SURGERY
Critique
Both studies are well-conducted randomized controlled trials. Although the American
trial recruited more patients, it had a shorter follow-up period. The American cohort
of patients was substantially younger than the British cohort (57 years versus 70 years),
which may have confounded the outcome data as the risk of acute strangulation increases
with age.
The diagnostic criteria for inguinal hernia also differed between the two groups. The
British study required the presence of a visible lump in the groin, whereas 40% of patients
in the American study were diagnosed with a hernia on cough impulse alone. This may
have confounded their results, as the natural history of a hernia diagnosed on cough
impulse alone is not known.
LAPAROSCOPIC VERSUS OPEN GROIN HERNIA 39
2.7 Laparoscopic versus open groin hernia repair

Details of studies
A traditional darn was used for groin hernia surgery for many years until Lichtenstein
introduced the mesh repair. This development also coincided with the advent of the lapa-
roscopic groin hernia repair. The new technique was subject to rigorous assessment in
several high-quality randomized trials from the UK and the USA. Several issues were
addressed, including questions relating to the recurrence rate with laparoscopic repair
and whether a faster recovery time and quicker return to work would offset the extra cost
of laparoscopic surgery.
Study references
Main studies
MRC Laparoscopic Hernia Trial Group (1999) Laparoscopic versus open repair of groin hernia:
a randomised comparison. Lancet, 354, 185–90.
Neumayer, L., Giobbie-Hurder, A., Johansson, O., et al. (2004) Open mesh versus laparoscopic mesh
repair of inguinal hernia. New England Journal of Medicine, 350, 1819–27.
Arvidsson, D., Berndsen, F.H., Larsson, L.G., et al. (2005) Randomized clinical trial comparing 5-year
recurrence rate after laparoscopic versus Shouldice repair of primary inguinal hernia. British Journal
of Surgery, 92, 1085–91.
Related references
Berndsen, F.H., Petersson, U., Arvidsson, D., et al. (2007) Discomfort five years after laparoscopic and
Shouldice inguinal hernia repair: a randomised trial with 867 patients. A report from the SMIL
study group. Hernia, 11, 307–13.
Bernsden, E., Arvidsson, D., Enander, L.K., et al. (2002) Post operative convalescence after inguinal
hernia surgery: prospective randomized multicenter study of laparoscopic versus Shouldice inguinal
hernia repair in 1042 patients. Hernia, 6, 56–61.
Douek, M., Smith, G., Oshowo, A., Stoker, D.L., and Wellwood, J.M. (2003) Prospective randomised
controlled trial of laparoscopic versus open inguinal hernia mesh repair: five year follow up. British
Medical Journal, 326, 1012–13.
Johansson, B., Hallerback, B., Glise, H., et al. (1999) Laparoscopic mesh versus open preperitoneal mesh
versus conventional technique for inguinal hernia repair: a randomized multicenter trial (SCUR
Hernia Repair Study). Annals of Surgery, 230, 225–31.
Wright, D., Paterson, C., Scott, N., Hair, A., and O’Dwyer, P.J. (2002) Five year follow up of patients
undergoing laparoscopic or open hernia repair: a randomized controlled trial. Annals of Surgery,
235, 333–7.
Liem, M., van der Graaf, Y., van Steensel, C., et al. (1997) Comparison of conventional anterior surgery
and laparoscopic surgery for inguinal hernia repair. New England Journal of Medicine, 336, 1541–7.
Eklund, A., Carlsson, P., Rosenblad, A., et al. (2010) Long term cost-minimization analysis comparing
laparoscopic with open (Lichtenstein) inguinal hernia repair. British Journal of Surgery, 97, 765–71.
Hynes, D.M., Stroupe, K.T., Luo, P., et al. (2006) Cost effectiveness of laparoscopic versus open mesh
hernia operation: results of a Department of Veterans Affairs randomized clinical trial. Journal of the
American College of Surgeons, 203, 447–57.
40 GENERAL SURGERY
Study design
MRC study
This was a large British RCT comparing open-mesh repair with laparoscopic repair
(trans-abdominal pre-peritoneal (TAPP) or totally extra-peritoneal (TEP)). More than
half the participants were unemployed at the time of the study, and so ‘return to normal
activities’ was amended to ‘return to usual social activities’.
Neumayer et al.
This study was the largest RCT with 2000 patients randomized to open-mesh or lapa-
roscopic hernia repair (TAPP or TEP). Surgeons involved were specifically trained and
had to undertake at least 25 open and 25 laparoscopic repairs prior to the study. Baseline
characteristics were comparable, although 50% of the study population had a relatively
short history of symptoms of less than a year.
Arvidsson et al.
This large RCT, from the SMIL hernia group in Sweden, randomized patients to open
Shouldice and TAPP laparoscopic repair. The baseline characteristics of participants were
comparable. The unusual feature about this study was that the surgeons involved were
either specialists in laparoscopy or Shouldice experts.
Results
MRC study Neumayer et al. Arvidsson et al.
Open Lap Open Lap Open Lap
Number of patients 406 468 994 989 466 454
Recurrences (%) 0 1.9 4.9 10.1 6.7 6.6
Pain (%) 36.7 28.7 14.3 9.8 11.4 8.5
Return to activities (days) 14 10 5 4
Duration of follow-up 1 year 2 years 5 years
MRC group
◆ At 1 week there were more haematomas in the open group (15.7% versus 7.6%).
◆ Three serious complications occurred following TAPP procedures: injury to the
lateral cutaneous nerve to the thigh, bladder injury, and iliac artery damage.
◆ Laparoscopic surgery took 15 minutes longer than open surgery. However, there was
a shorter length of hospital stay (1 versus 1–2 days, p = 0.008) and a quicker return to
usual activities (28.7% versus 36.7%, p = 0.018).
Neumayer et al.
◆ Complication rates in the first 4 weeks were similar in the two groups.
◆ Complications were mostly haematomas and seromas
LAPAROSCOPIC VERSUS OPEN GROIN HERNIA 41
◆ Life -threatening complications occurred in 1.1% of laparoscopic cases and 0.1% of

open cases.
◆ A further report on the same cohort (Hynes et al. 2006) found that laparoscopic
surgery cost $638 more than open surgery.
Arvidsson et al.
◆ The laparoscopic group had less pain and returned to normal activities more quickly
than the open group.
◆ Recurrences were similar between the groups (6.6% versus 6.7%).
◆ Chronic pain was significantly higher in the open group.
Conclusions
The three trials discussed here confirmed that laparoscopic repair was associated with a
shorter length of hospital stay and quicker return to normal activities. Some studies also
reported less chronic pain with the laparoscopic procedure. However, there are serious
risks associated with the laparoscopic approach which do not occur during open surgery.
The cost of laparoscopic surgery has been shown to be significantly greater than that of
open surgery, even without the use of disposable instruments.
Critique
An overriding theme in relevant studies is that results from experienced and high-volume
specialist centres are better than those obtained in the community. There is clearly a long
learning curve associated with the laparoscopic procedure, and it would take a long time
for the average surgeon to become competent (estimated at up to 5 years, assuming 50
procedures per year). The evidence suggests that there is a higher risk of a serious compli-
cation with the laparoscopic approach; however, this has been associated predominantly
with the TAPP rather than the TEP approach, which stays in the preperitoneal plane.
It is difficult to justify the cost of laparoscopic surgery for unilateral groin hernias, even
with reduced sick leave and a quicker return to work. Despite this, NICE guidelines cur-
rently recommend the laparoscopic approach as a treatment option for inguinal hernias.
42 GENERAL SURGERY
2.8 Local or general anaesthesia for groin hernia repair

Details of studies
Although many surgeons will consider local anaesthesia (LA) when performing a groin
hernia repair in an elderly, frail, or medically unfit patient, specialist hernia centres have
long considered LA to be the most effective and safest method of performing open groin
hernia repair for most patients. Potential benefits include increased patient safety, a
shorter recovery period, better post-operative pain control and reduced cost compared
with general anaesthesia (GA). The two papers discussed here compared outcomes with
LA and GA for groin repair.
Study references
Main studies
O’Dwyer, P.J., Serpell, M.G., Millar, K., et al. (2003) Local or general anesthesia for open hernia repair:
a randomized trial. Annals of Surgery, 237, 574–9.
Nordin, P., Zetterström, H., Gunnarson, U., and Nilsson, E. (2003) Local, regional or general anaesthesia
in groin hernia repair: a multicentre randomised trial. Lancet, 362, 853–8.
Related references
Nordin, P., Zetterstrom, H., Carlsson, P., and Nilsson, E. (2007) Cost-effectiveness analysis of local,
regional and general anaesthesia for inguinal hernia repair using data from a randomized clinical
trial. British Journal of Surgery, 94, 500–5.
Study design
◆ Both studies are prospective RCTs.
◆ Nordin et al. separately reported their cost-effectiveness data in 2007.
O’Dwyer et al. Nordin et al.

Randomization LA vs GA LA vs regional vs GA
Primary outcome Tests of psychomotor recovery Complications
Secondary outcomes Pain Operative time
Return to normal activity Recurrence
Cost Length of stay
Cost
Follow-up 1 year 30 days
Results
O’Dwyer et al.
◆ LA was associated with significantly less pain on movement at 6 hours. There were
no differences in pain scores at rest or on movement at 24 and 72 hours.
LA OR GA FOR GROIN HERNIA REPAIR 43
◆ Both groups required similar analgesia on discharge.

◆ There were no significant differences in vigilance, divided or sustained attention,
memory, and cognitive function between the two groups.
◆ Physical activity was significantly reduced at 24 hours in both groups.
◆ Return to work was 2 weeks for both groups.
◆ Significantly fewer patients in the LA group would recommend the procedure to
others (84% versus 95%, p = 0.011). Despite this, patient satisfaction was similar
between the two groups.
◆ The price of a GA case was quoted as £30.54 (4%) more than an LA case.
Nordin et al.
◆ The LA group had significantly less pain and nausea in the early post-operative
period compared with the GA group.
◆ Early complications were higher in the GA and regional anaesthesia groups than in
the LA group. Catheterization was required in the post-operative period in 29% of
regional cases, compared with 8% in the GA group and none in the LA group.
◆ Duration of surgery was significantly longer in the LA group than in the other two
groups.
◆ LA was associated with a reduced length of stay (LoS) and a lower rate of unplanned
admissions.
◆ There was no difference in return to normal activities between the groups.
◆ The costs associated with an LA repair were significantly less than a GA/regional repair.
Conclusions
◆ O’Dwyer et al. concluded that there are no differences in patient recovery after LA or
GA hernia surgery.
◆ Nordin et al. felt that LA had substantial advantages over GA/regional anaesthesia.
Critique
These two well-designed trials appear to have given conflicting results in terms of cost
and outcome. O’Dwyer et al. used low-dose intravenous midazolam and fentanyl in the
LA group, which may have had an effect on the central nervous and respiratory systems.
Patients were also kept in hospital to assess cognitive function, which may have had an
effect on LoS and cost.
Nordin et al. found a high rate of complications at the 8-day follow-up. Follow-up was
carried out by telephone interview, which may be associated with a high false-positive
rate. In addition, 2% of the regional group were so well sedated that their outcomes had
to be analysed with the GA group.
In summary, it appears that good outcomes can be achieved with LA in non-specialist
hernia centres. Therefore LA should be offered as an option to patients with uncompli-
cated groin hernias.
44 GENERAL SURGERY
2.9 Chronic pain after groin hernia surgery

Details of studies
Prior to the publication of these two papers, chronic pain after groin hernia surgery was
not an area that had received much attention. Although it was known that chronic pain
did occur, there were no data from a large prospective study. Investigators theorized that
the severity and likelihood of developing chronic pain after groin hernia surgery might
be predicted by the intensity of acute post-operative pain. An association between the
intensity of acute post-operative pain and the likelihood of developing chronic pain had
already been noted after leg amputation, breast surgery, and thoracotomy.
Study references
Callesen, T., Bech, K., Kehlet, H. (1999) Prospective study of chronic pain after groin hernia repair.
Cunningham, J., Temple, W., Mitchell, P., et al. (1996) Cooperative hernia study. Pain in the postrepair
patient. Annals of Surgery, 224, 598–602.
Study design
Callesen et al. Cunningham et al.
Study type Prospective cohort Prospective RCT
Setting Danish teaching hospital Two Canadian teaching hospitals
Intervention Elective, day case, open groin Elective open groin hernia repair
hernia repair
Number of patients 500 315 repairs in 276 patients
Primary outcomes Pain Pain
Follow-up Postal questionnaire at Telephone interview at 6 months, and
1 week, 4 weeks, and 1 year clinic review at 12 and 24 months
◆ The trial by Cunningham et al. was part of a larger study evaluating outcomes
following three different open repair techniques (Bassini, McVay, and Shouldice).
◆ Callesen et al. assessed pain levels (none, mild, moderate, severe) at rest, whilst
coughing, and during mobilization. In the Cunningham et al. study, at their 6-month
telephone assessment, patients were asked to rate their peri-operative pain (scale
1–10). At clinic review at 12 and 24 months, patients were asked to rate pain if
present as mild, moderate, or severe.
Results
Callesen et al.
◆ There was a 93% response to the postal questionnaire.
CHRONIC GROIN PAIN AFTER GROIN HERNIA SURGERY 45
◆ At 1 year: 80 (19%) experienced mild pain, 25(6%) had moderate or severe pain, and
24 (6%) experienced pain which restricted normal activity.
◆ The risk of developing chronic pain at 1 year was increased in the group who
experienced pain at 1 week (9% versus 3%, p <0.05) and 4 weeks (24% versus 3%,
p <0.001).
Cunningham et al.
◆ 55.5% of patients complained of moderate/severe pain in the early post-operative
period.
◆ 11.9% had moderate/severe pain at 1 year, and 10.6% (32) at 2 years.
◆ There was no difference in type of hernia and long-term pain
◆ Asymptomatic numbness was ‘commonly found’; however, initial post-operative
numbness declined over the 2-year follow-up period.
◆ The patients with moderate/severe pain at 2 years were recalled for further
assessment. Seventeen re-attended for clinic review: seven had no further symptoms
on review; nine of the remaining ten were assessed by a specialist and diagnosed
with a somatic ligamentous pain syndrome. They were all found to be tender over the
medial insertion of the inguinal ligament.
Conclusions
Callesen et al. concluded that chronic pain was a significant problem after open groin
hernia surgery and may be predicted by the severity of early post-operative pain.
Cunningham et al. noted that standard open hernia repair was associated with long-
term pain and numbness in more than 10% of patients. They also concluded that the most
common type of pain following hernia surgery is somatic pain, which can be reproduced
on pressure.
Critique
These two landmark papers, published more than 20 years ago, were instrumental in
quantifying the severity and nature of pain that can be expected after open groin hernia
surgery. At the time, it was thought that post-operative hernia pain was related to nerve
damage or the development of neuromas which might be amenable to injections, further
surgery, or analgesia. However, Cunningham et al. described a different type of pain at the
medial insertion of the inguinal ligament and attributed this to the insertion of sutures
under tension in this area. They also recommended that the surgeon should avoid placing
sutures directly into the periosteum of the pubic tubercle.
46 GENERAL SURGERY
2.10 The Lichtenstein repair

Details of study
Prior to the publication of this seminal paper, hernia repair involved a sutured repair
under tension (Bassini, Halstead, Shouldice, and McVay). Recurrence rates were high and
patients were often not adequately followed up. In 1989, Lichtenstein et al. described a
totally new approach to groin hernia repair. The technique used a polypropylene mesh to
repair the hernia defect and avoided suture-line tension. It was simple, highly effective,
caused less pain, and rapidly gained acceptance throughout the world.
Study reference
Lichtenstein, I.L., Shulman, A.G., Amid, P.K., Montllor, M.M. (1989) The tension-free hernioplasty.
American Journal of Surgery, 157, 188–93.
Study design
Level of evidence 4
Intervention Primary inguinal mesh repair under local anaesthetic
Number of patients 1000 consecutive patients
Setting Single institution
Follow-up 6 years
◆ A 10 × 15cm polypropylene mesh was used and sutured in place with 0 Novofil.
Results
◆ No recurrences were identified during the follow-up period.
◆ There were two haematomas which settled conservatively, and no wound/mesh
infections.
◆ Normal activities were resumed after 2–3 days.
Conclusions
The authors concluded that the primary cause of recurrence post inguinal hernia repair
is the approximation of normally unopposed tissues. This creates tension on the suture
which is ‘a clear violation of basic surgical principles’. The new technique was simple,
rapid, less painful, safe, and effective.
Critique
This paper revolutionized hernia repair surgical technique, methods of reporting, and
length of follow-up. The only significant criticism to be made is that the authors do not
state whether there was any independent clinical follow-up.
OPEN VERSUS LAPAROSCOPIC INCISIONAL HERNIA REPAIR 47
2.11 Open versus laparoscopic incisional hernia repair

Details of studies
Although there is no standard open approach to repair of incisional hernias, a number
of studies have shown unacceptably high recurrence rates with primary closure and it
is now accepted that mesh should be used to reinforce the majority of repairs. Popular
open techniques to repair with mesh include onlay, inlay, and sublay. Laparoscopic repair
has gained popularity with the introduction of a variety of composite meshes which are
designed to minimize adhesion formation on the peritoneal cavity side but facilitate
fibrous ingrowth on the abdominal wall aspect of the mesh. The following randomized
controlled trials compared open and laparoscopic ventral hernia repair.
Study references
Main studies
Carbajo, M.A., Martin del Olmo, J.C., Blanco, J.L. et al. (1999) Laparoscopic treatment vs open surgery
in the solution of major incisional and abdominal wall hernias with mesh. Surgical Endoscopy, 13,
250–2.
Asencio, F., Aguilo, J., Peiro, S., et al. (2009) Open randomized clinical trial of laparoscopic versus open
incisional hernia repair. Surgical Endoscopy, 23, 1441–8.
Itani, K.M., Hur, K., Kim, L.T., et al. (2010) Comparison of laparoscopic and open repair with mesh for
the treatment of ventral incisional hernias: a randomised trial. Archives of Surgery, 145, 322–8.
Related references
Heniford, B.T., Park, A., Ramshaw, B.J., and Voeller, G. (2003) Laparoscopic repair of ventral hernias:
nine years experience with 850 consecutive hernias. Annals of Surgery, 238, 391–400.
Study design
◆ All three studies were randomized controlled clinical trials.
Carbajo et al. Asencio et al. Itani et al.

Level of evidence 1b 1b 1b
Number of patients 60 84 146
Primary outcome Complications Quality of life Percentage of patients with ≥1
complication at 8 weeks
Secondary outcomes Operative time Complications Complications
Hospital stay Recurrence Recurrence
Hospital stay Pain
Pain Functional status
Operative time Activity levels
Follow-up 27 months 12 months 24 months
48 GENERAL SURGERY
Results
Carbajo et al. Asencio et al. Itani et al.
Open Lap Open Lap Open Lap
Complications (%) 66 6.6 33.3 5.2 47.9 31.5
Recurrences (%) 6 0 7.9 9.7 8.2 12.5
◆ Although less frequent, more serious complications including bowel perforation were
recorded in the laparoscopic surgery groups.
◆ Surgical site infection was less common at 8 weeks in the laparoscopic group—5.6%
versus 23.3% (Itani et al.).
◆ Laparoscopic surgery was associated with less pain and a faster return to normal
activities in one study (Itani et al)
◆ Carbajo et al. and Itani et al. noted that laparoscopic incisional/ventral hernia repair
was quicker and had a shorter length of stay than open surgery
Conclusions
Although Asencio et al. felt that laparoscopic repair did not appear to be superior to open
repair, both Itani et al. and Carbajo et al. advocated the laparoscopic approach because of
reduced complications, operating time, pain, shorter length of hospital stay, and a faster
return to normal activities.
Critique
Laparoscopic incisional hernia repair does seem to offer improvements in some short-
term outcomes, including recovery and wound infection. Despite this, pain has not con-
sistently been found to be reduced by laparoscopy and longer-term outcomes such as
recurrence also appear to be similar to the open approach.
Although there is good evidence that laparoscopic incisional/ventral hernia repair is
associated with a reduced hospital stay and quicker recovery, many remain concerned
that there is a small, but real, risk of unrecognized bowel injury. Heniford et al. and others
have reported on cases of post-operative mortality as a consequence of this complication.
As with laparoscopic inguinal hernia repair, the learning curve is long and so this tech-
nique is perhaps best performed by specialist laparoscopic surgeons.
Sauerland, S., Walgenbach, M., Habermalz, B., Seller, C.M., and Miserez, M. (2011) Laparoscopic versus
open surgical techniques for ventral or incisional hernia repair. Cochrane Database of Systematic
Reviews, March 16 (3), CD007781pub 2 (DOI: 10.1002/14651858).
ANTIBIOTIC PROPHYLAXIS IN INGUINAL HERNIA REPAIR 49
2.12 Antibiotic prophylaxis in inguinal hernia repair

Details of studies
Forty years ago, the use of prophylactic antibiotics to reduce the risk of post-operative
infection in contaminated surgery became increasingly popular. Evidence was lacking for
the use of antibiotics in clean surgery. The routine use of a synthetic mesh in otherwise
clean hernia surgery brought fresh concerns about the potential for post-operative mesh
infection. The meta-analysis by Sanabria et al. (2007) addressed the issue of prophylactic
antibiotic use in mesh hernioplasty.
Study reference
Main study
Sanabria, A., Dominguez, L.C., Valdivieso, E., and Gomez, G. (2007) Prophylactic antibiotics for mesh
inguinal hernioplasty: a meta-analysis. Annals of Surgery, 245, 392–6.
Related references
Morales, R., Carmona, A., Pagán, A., et al. (2000) Utility of antibiotic prophylaxis in reducing wound
infection in inguinal or femoral hernia repair using polypropylene mesh. Cirugía Española, 67, 51–9.
Oteiza, F., Ciga, M.A., and Ortiz, H. (2004) Antibiotic prophylaxis in inguinal hernioplasty. Cirugía
Española, 75, 69–71.
Yerdel, M.A., Akin, E.B., Dolalan, S., et al. (2001) Effect of single dose prophylaxis ampicillin and sul-
bactam on wound infection after tension free inguinal hernia repair with polypropylene mesh: the
randomised double-blind prospective trial. Annals of Surgery, 233, 26–33.
Aufenacker TJ, van Geldere, D., van Mesdaq, T., et al. (2004) The role of antibiotic prophylaxis in
prevention of wound infection after Lichtenstein open mesh repair of primary inguinal hernia: a
multicenter double-blind randomized controlled trial. Annals of Surgery, 240, 955–60.
Perez, A.R., Roxas, M.F., and Hilvano, S.S. (2005) A randomized, double blind, placebo-controlled trial
to determine effectiveness of antibiotic prophylaxis for tension-free mesh herniorrhaphy. Journal of
the American College of Surgeons, 200, 393–7.
Celdrán, A., Frieyro, O., de la Pinta, J.C., et al. (2004) The role of antibiotic prophylaxis on wound infec-
tion after mesh hernia repair under local anaesthesia on an ambulatory basis. Hernia, 8: 20–2.
Study design
◆ A meta-analysis of RCTs.
Level of evidence 1a
Included trials Six RCTs comparing antibiotic prophylaxis versus no prophylaxis/placebo
Time period Published trials between 1985 and 2007
Inclusion criteria Inguinal or groin hernia repair with mesh
Exclusion criteria Contraindication to antibiotic use
Immunosupressed patients
Primary outcome Post-operative wound infection
Follow-up 1–24 months
50 GENERAL SURGERY
◆ The quality of the RCTs was assessed using Cochrane Collaboration criteria.
Quasi-randomized trials were excluded.
Results
◆ Post-operative infection rate was higher in the placebo group than in the antibiotic
group (2.89% versus 1.38%)
Conclusions
The authors conclude that antibiotic prophylaxis reduces post-operative infection by
50% in patients undergoing mesh hernioplasty.
Critique
The main criticism of this meta-analysis is that all the participants were considered to be
low risk. Some of the studies included report an infection rate as low as 1%. It may well
be that cost-effectiveness and potential harm of giving antibiotics in this setting should be
carefully evaluated. While it could be argued that antibiotics are not required because of
the low rate of infection in both groups, rates of wound infection outside RCTs are likely
to be higher.
SURGERY FOR INGROWN TOENAILS 51
2.13 Surgery for ingrown toenails

Details of studies
Until 40 years ago, surgery for ingrown toenail was mutilating and associated with a high
level of recurrence. The following papers addressed the main issues including whether
avulsion should be performed with or without the use of phenol, and whether avulsion
with phenol was better than surgical excision.
Study references
Andrew, T. and Wallace, W.A. (1979) Nail bed ablation—excise or cauterise? A controlled study. British
Medical Journal, 1, 1539.
Grieg, J.D., Anderson, J.H., Ireland, A.J., and Anderson, J.R. (1991) The surgical treatment of ingrowing
toenails. Journal of Bone and Joint Surgery, British Volume, 73, 131–3.
Bos, A.M.C., van Tilburg, A., van Sorge, A.A., and Klinkenbijl, J.H.G. (2007) Randomized clinical trial
of surgical technique and local antibiotics for ingrowing toenail. British Journal of Surgery,
94, 292–6.
Study design
Andrew et al. Grieg et al. Bos et al.
Level of evidence 1b 1b 1b
Randomization Zadik’s nail-bed Avulsion vs nail-edge Excision of matrix vs
ablation vs avulsion excision vs nail-edge phenolization with or
and phenolization excision and without topical
phenolization gentamicin*
Primary outcome Recurrence
Follow-up (months) 6 12 12
*All patients also underwent a partial avulsion
Results
◆ Andrew et al. reported a recurrence rate of 6% in the phenolization group versus 18%
with Zadik’s procedure. They also found less recurrence of nail spikes in the phenol
group (19% versus 32%). There were no cases of osteitis.
◆ Grieg et al. also found that the addition of phenol dramatically reduced the
recurrence rate at 1 year: 73% in the nail avulsion group versus 73% in the nail-
edge excision group versus 9% in the nail-edge excision and phenolization group.
They also found that fewer patients in the phenol group were dissatisfied, despite an
increased risk of post-operative infection.
◆ Bos et al. similarly found that recurrence rates were significantly lower with
phenolization than with excision of the nail-matrix excision alone (p <0.001). They
also found no significant difference in post-operative infection rates. The addition of
topical gentamicin did not have any impact on infection rates.
52 GENERAL SURGERY
Conclusion
All these early studies concluded that phenol cauterization compares favourably with non-
phenol procedures whether the whole nail is avulsed or the nail edge is simply excised.
The addition of topical antibiotics is not justified.
Critique
Andrew et al. and Bos et al. reported a high rate of recurrence in patients undergoing phe-
nol cauterization. This figure was higher (2%) than in other observational studies at that
time. The authors felt that the application of phenol for 3 minutes may have been too brief.
In addition, nail-bed haemorrhage may have ameliorated the action of the phenol. They
suggested that phenol be applied for a longer period of time and that absolute haemostasis
is essential before application.
Grieg et al. reported a significant level of post-operative infection in the phenol cauter-
ization group. Although phenol is itself an antiseptic, this study highlighted the relevance
of tissue injury and the potential for infective complications with this caustic agent.
SURGICAL VOLUME AND PATIENT OUTCOME 53
2.14 Surgical volume and patient outcome

Details of studies
For many decades, local hospitals in the UK provided virtually all secondary care for their
local population. As a result, patients have high expectations about what care they receive
locally and can be fiercely loyal to their local hospital. In 1957, Lee et al. first reported in
the Lancet that there might be a difference in mortality rates between hospitals perform-
ing different volumes of surgical procedures (teaching versus non-teaching hospitals).
Since those early days, there has been a huge quest on both sides of the Atlantic for robust
evidence to support this claim, which has significant implications for the delivery of sur-
gical services.
Study references
Main study
Birkmeyer, J.D., Siewers, A.E., Finlayson, E.V.A., et al. (2007) Hospital volume and surgical mortality in
the United States. New England Journal of Medicine, 346, 1128–37.
Related references
Chowdhury, M.M., Dagash, H., and Pierro, A. (2007) A systematic review of the impact of volume of
surgery and specialization on patient outcome. British Journal of Surgery, 94, 145–61.
Lee, J.A., Morrison, S.L., and Morris, J.N. (1957) Fatality from three common surgical conditions in
teaching and non-teaching hospitals. Lancet, 273, 785–90.
Finks, J.F., Osborne, N.H., and Birkmeyer, J.D. (2011) Trends in hospital volume and operative mortality
for high-risk surgery. New England Journal of Medicine, 364, 2128–37.
Study design
Birkmeyer et al. performed a retrospective US-based observational study (level 2c evi-
dence) assessing the mortality associated with six cardiovascular procedures and eight
types of major cancer resection performed between 1994 and 1999. All data were col-
lected from the National Medicare claims database. Mortality was defined as in-hospital
or within 30 days. Adjustments were made for patient characteristics.
Results
◆ Data from more than 2.5 million procedures were examined.
◆ A reduction in mortality in all 14 procedures was seen in high-volume institutes
compared with low-volume institutes.
◆ The most marked differences in adjusted mortality rates between the highest- and
lowest-volume centres were seen for oesophagectomy (mortality of 8.4% in high-
volume centres versus 20.3% in low-volume centres), pancreatic resection (3.8%
versus 16.3%), and repair of non-ruptured aortic aneurysm (3.9% versus 6.9%).
◆ A less significant difference in mortality was seen for other procedures such as
coronary artery bypass grafting and nephrectomy.
54 GENERAL SURGERY
Conclusions
The authors concluded that the data provided strong evidence that higher-volume
hospitals had lower observed and operative mortality rates for a wide range of operative
procedures. The greatest difference in mortality between hospitals with very low and
very high volumes was seen in complex surgeries including pancreatic resection and
oesophagectomy.
Critique
The main criticism of the Birkmeyer study is the confounding effect of the case mix.
Although the authors stated that they adjusted for risk, the data were obtained from
Medicare patients, which accounts for the majority of patients in the USA who undergo
high-risk surgery and an even larger majority of those who die after surgery. Although
anybody aged more than 65 years who is eligible for social security retirement benefit is
automatically eligible for Medicare insurance, it may not cover the very poorest groups
in society. Similarly, the less affluent population is less likely to be treated in high-volume
American institutes. The study also does not take into account the referral patterns for
high-risk cases, in that they are highly likely to be referred from low-volume to high-
volume institutes. In addition, because the study was limited to Medicare patients aged 65
years or older it may not be generalizable to younger patients, although there is no reason
to suggest that this is not the case.
A subsequent systematic review of the impact of volume of surgery and specialization
assessed 163 studies containing data on almost 10 million patients (Chowdhury et al).
Once again, high-volume hospitals were noted to have significantly better outcomes,
although interestingly this effect was limited in prospective studies. There was also clear
evidence that both high-volume surgeons and specialist surgeons had significantly better
outcomes than general surgeons.
Despite the overwhelming evidence that high-volume specialized units are associated
with improved patient outcomes, surgeons have argued that increased specialization
would reduce general surgical care locally, restrict emergency care, and adversely affect
training for future surgeons. The reality is that with increased emphasis on audit and
accountability, the quality of the service provided, as measured by volume among other
factors, is of prime importance. However, it is not simply a high volume of work that
determines the quality of an individual service. Increasingly, it is recognized that special-
ization and training also play a pivotal role in determining outcome.
Since the publication of this study we have witnessed major changes in surgical practice
with certain procedures, such as oesophagectomy, becoming more concentrated in higher-
volume units and a resulting decrease in mortality. A follow-up study by Finks et al. has
shown that improvements in surgical safety (the surgical checklist), peri-operative care,
and post-operative care have all contributed to improvements in patient outcome.
Chapter 3
Emergency surgery
3.1 Active observation 56
3.2 Opioid analgesia prior to assessment of the acute abdomen 59
3.3 CT assessment of the acute abdomen 62
3.4 Early laparoscopy in the assessment of acute abdominal pain 65
3.5 Laparoscopic versus open appendicectomy 68
3.6 Non-operative management of acute complicated appendicitis 71
3.7 Interval appendicectomy 74
3.8 Non-operative treatment of perforated duodenal ulcer 77
3.9 Eradication therapy after simple closure of perforated duodenal ulcer 80
3.10 Water-soluble contrast in small bowel obstruction 83
3.11 Surgery for left-sided malignant large bowel obstruction 86
3.12 The management of uncomplicated diverticulitis 89
3.13 The management of complicated diverticulitis 92
3.14 Effect of enteral versus parenteral nutrition on post-operative septic
complications 95
3.15 Goal-directed therapy in severe sepsis 99
3.16 Confidential enquiry into peri-operative deaths (CEPOD) 102
56 EMERGENCY SURGERY
3.1 Active observation

Details of studies
These two studies examine the effect of a period of active observation in hospital on out-
comes for patients presenting with acute abdominal pain and suspected appendicitis.
At the time of these studies common surgical teaching was that in cases of diagnostic
uncertainty any delay in surgical intervention was dangerous, and risked perforation in
cases of suspected appendicitis.
Study references
Main studies
Jones, P.F. (1976) Active observation in management of acute abdominal pain in childhood. British
Thomson, H.J. and Jones, P.F. (1986) Active observation in acute abdominal pain. American Journal of
Surgery, 152, 522–5.
Related references
White, J.J., Santillana, M., Haller, JA, Jr. (1975) Intensive in-hospital observation: a safe way to decrease
unnecessary appendectomy. American Surgeon, 41, 793–8.
Study design
Jones 1976
◆ Prospective case series.
◆ Inclusion criteria: children up to the age of 13 admitted to a Scottish hospital with
acute abdominal pain in a 12-month period.
◆ A datasheet containing history, examination, investigations, and examiner’s opinion
was kept.
◆ Patients with uncertain diagnoses underwent ‘active observation’ with review every
2–3 hours by the same examiner.
Thomson and Jones 1986

◆ Inclusion criteria: patients admitted to a Scottish hospital with acute abdominal pain
on every fourth day for a period of 50 weeks.
◆ Exclusion criteria: patients less than 1 year of age.
◆ ‘Active observation’ was carried out in the same manner as in the first study.
Outcome measures
◆ Laparotomy with negative findings.
◆ Evidence of delay in diagnosis leading to harm.
ACTIVE OBSERVATION 57
Results
Jones (n = 363) Thomson and Jones
(n = 220)
Suspected appendicitis 116 48
Immediate operations 96 (83%) 39 (81%)
Operations after first re-examination 19 4
Perforated appendix 1/16 2/4
Operations after second re-examination 1 4
Perforated appendix 0/1 1/3
Operations after third re-examination 0 2
Negative laparotomy (including mesenteric adenitis) 12 (10%) 6 (12%)
Negative laparotomy (excluding mesenteric adenitis) 4 (3%) 4 (8%)
Non-specific abdominal pain 108 (30%) 22 (10%)
◆ The authors reported that no patient who was observed appeared to suffer any
detrimental effects from delay in operation
◆ None of the patients in the later study with non-specific abdominal pain were
readmitted within the year following the study
Critique
At the end of the nineteenth century, when Fitz first described appendicitis, the mortal-
ity associated with the condition was in excess of 25%. With increased recognition of
the pathology and early intervention in dubious cases, the mortality was dramatically
reduced. Until these particular studies this policy of aggressive surgical intervention was
the norm, with rates of negative appendicectomy ranging from 15% to 30%. In fact, some
authors stated that if diagnostic accuracy was above 90% patients were likely to be suffer-
ing a dangerous delay, risking perforation of the appendix and a poor outcome.
Realizing that diagnosis in many cases of acute abdominal pain is uncertain, Jones
proposed a policy of ‘active observation’. This approach formalized what is now regarded
as good clinical practice, combining careful history-taking, examination, and documen-
tation with regular clinical review. In two prospective case series Jones was able to show
a negative appendicectomy rate of less than 10%, with no obvious detrimental effect on
recovery related to diagnostic delay. Based on Jones’s work, a further study was carried out
to compare immediate operation (n = 266) versus active observation (n = 104) over two
consecutive time periods (White et al. 1975). This found that active observation reduced
the rate of negative appendicectomy from 15% to 2% with no difference in the incidence
of perforated appendicitis or mortality. The authors also highlighted that the majority of
appendiceal perforations occur before admission, and may be related more to delay in
seeking medical care.
There are obvious limitations with the case series presented which may simply rep-
resent the results of one particular unit or surgeon. It is also true that the assessment of
acute appendicitis has changed since these studies were reported with the easy availability
of ultrasonography, computed tomography, and laparoscopy. However, it is not clear that
modern investigation has dramatically improved diagnostic accuracy outside the confines
of a research trial. In a population study of 63,707 appendicectomies in Washington State
between 1987 and 1998 the rate of negative appendicectomy was stable at 15% (Flum et al.
2001). In a more recent Swiss study of 7964 appendicectomies between 1995 and 2006 the
negative appendicectomy rate was 6.4% and reduced over the period of the study, sug-
gesting improved diagnosis (Güller et al. 2011). The rate of perforated appendicitis was
unchanged in both studies.
While the addition of new investigations and techniques further improve our assess-
ment of the acute abdomen, it is important to remember that these should still be directed
by careful clinical assessment with active observation in selected cases. As Jones (1990)
puts it, ‘the essentials in caring for a child with an acutely painful abdomen require…a
seat at the bedside, an unhurried approach, careful history-taking, patience and warm
hands…and a willingness to reconsider the doubtful case’. These comments are typical of
his thoughtful and enthusiastic approach to emergency surgery, so evident in his writing
on the subject.
Jones, P.F. (1990) Practicalities in the management of the acute abdomen. British Journal of Surgery, 77,
365–7.
Flum, D.R., Morris, A., Koepsell, T., and Dellinger, E.P. (2001) Has misdiagnosis of appendicitis
decreased over time? A population-based analysis. JAMA, 286, 1748–53.
Thomas, D.R. (1973) Conservative management of the appendix mass. Surgery, 73: 677–80.
Güller, U., Rosella, L., McCall, J., Brügger, L.E., and Candinas, D. (2011) Negative appendicectomy and
perforation rates in patients undergoing laparoscopic surgery for suspected appendicitis. British
OPIOID ANALGESIA PRIOR TO ASSESSMENT OF ACUTE ABDOMEN 59
3.2 Opioid analgesia prior to assessment of the

acute abdomen
Details of studies
Before the publication of the trial by Attard et al. there was little evidence from controlled
trials about the safety of opioid analgesia prior to assessment of acute abdominal pain.
There was a widely held view that opiates could mask physical findings, leading to signifi-
cant delays in diagnosis and treatment. Even after the trial in 1998 a survey found that
80% of emergency department physicians withheld opiate analgesia until after a surgical
review. This was one of the factors prompting the further study by Thomas et al.
Study references
Main studies
Attard, A.R., Corlett, M.J., Kidner, N.J., Leslie, A.P., and Fraser, I.A. (1992) Safety of early pain relief for
acute abdominal pain. British Medical Journal, 305, 554–6.
Thomas, S.H., Silen, W., Cheema, F., et al. (2003) Effects of morphine analgesia on diagnostic accuracy
in emergency department patients with abdominal pain: a prospective, randomized trial. Journal of
the American College of Surgeons, 196, 18–31.
Related references
Zoltie, N. and Cust, M.P. (1986) Analgesia in the acute abdomen. Annals of the Royal College of Surgeons
of England, 68, 209–10.
Study design
◆ Both studies were prospective double-blind single-institution RCTs.
Attard et al.
Randomization 20mg papaveretum IM versus placebo (saline)
Inclusion criteria Emergency patients >16 years admitted with acute abdominal pain
(<48 hours duration) sufficiently severe to warrant opiate analgesia
Outcome measures Pain and tenderness scores
Patient comfort
Accuracy of diagnosis and management decisions
◆ Seen prior to randomization by a house officer who assessed pain and tenderness.
◆ Seen one hour after injection by a surgical registrar who repeated assessments of pain
and tenderness. Confidence of diagnosis was recorded including an assessment of the
group that they thought the patient was in.
Thomas et al.
Randomization Morphine sulphate (up to 15mg) versus placebo
Inclusion criteria Patients >18 years presenting with <72 hours of acute severe abdominal pain
Exclusion criteria Pregnancy
Emergency intervention required for life-threatening condition or systolic
BP <90mmHg
Early opiate analgesia clearly indicated (e.g. renal or biliary colic)
Outcome measures Diagnostic accuracy and physical examination changes
◆ Eligible patients were initially assessed using a standard evaluation form detailing site
of tenderness, specific signs, likely diagnosis, and confidence in the diagnosis. Pain
scores were completed on a visual analogue scale.
◆ One hour after randomization and injection of morphine or placebo the same
clinician repeated the assessment. The effect of medication on follow-up examination
and which group the patient was thought to be in was also assessed.
Results
Attard et al.
Papaveretum Placebo p value
Pain scores (VAS) 3.1 8.3 <0.0001
Tenderness scores (VAS) 5.1 8.1 <0.0001
Patient deemed comfortable 48/50 9/50 <0.01
Incorrect diagnosis/management 2 9
VAS, visual analogue scale
◆ There was no difference in the registrar’s diagnostic confidence between groups.

◆ There was no difference in the accuracy of localization of physical signs.
Thomas et al.
◆ 899 patients with abdominal pain were screened to find 74 eligible patients.
◆ No difference in physical signs (15.1% versus 11.8%) or diagnostic accuracy (64.2%
versus 66.7%) between the morphine and control groups.
◆ Likelihood of change in severity of tenderness and location of pain were similar
between the groups.
OPIOID ANALGESIA PRIOR TO ASSESSMENT OF ACUTE ABDOMEN 61
Conclusions
Attard et al.
Early administration of opiate analgesia to patients with acute abdominal pain can greatly
reduce their pain. This does not interfere with diagnosis, which may even be facilitated
despite a reduction in the severity of physical signs.
Thomas et al.
This study found that early administration of morphine did not have any harmful effects
with respect to diagnostic accuracy or physical examination findings.
Critique
To relieve suffering is a fundamental component of good clinical care. Until these studies
there was concern that opioid analgesia could mask physical signs and potentially lead to
significant diagnostic errors. This view contributed to delay in the administration of effec-
tive analgesia by general practitioners and junior doctors until patients were reviewed by
a senior surgeon.
Both studies suffer from methodological shortcomings. In the study by Attard et al.
the initial assessment and subsequent review were carried out by different clinicians,
introducing inter-observer variability. Although this problem was addressed in the study
by Thomas et al., their patients were non-consecutive, with a very small proportion of
patients approached being felt to be eligible. Both these factors could lead to selection bias.
Despite these problems the studies presented here are a reflection of the consistent
findings of trials examining the effect of opioid administration prior to specialist assess-
ment of acute abdominal pain. These trials provide good evidence that early analgesia in
patients with undifferentiated abdominal pain is both desirable and safe. The findings are
further supported by a recently updated Cochrane review.
Manterola, C., Vial, M., Moraga, J., and Astudillo, P. (2011) Analgesia in patients with acute abdominal
pain. Cochrane Database of Systematic Reviews, 1, CD005660.
3.3 CT assessment of the acute abdomen

Details of studies
Acute abdominal pain is a frequent cause of emergency admission to hospital. Initial
assessment may yield a diagnosis, but this is probably only accurate in around 50–60%
of patients. Delay in diagnosis can result in inappropriate management, which may affect
patient outcome with increased length of hospital stay and healthcare costs.
Both availability and accuracy of computed tomography (CT) have increased dra-
matically. The aim of the following study was to determine whether early CT evaluation
reduces length of hospital stay and improves diagnostic accuracy.
Study reference
Main study
Ng, C.S., Watson, C.J.E., Palmer, C.R., et al. (2002) Evaluation of early abdominopelvic computed
tomography in patients with acute abdominal pain of unknown cause: prospective randomised
study. British Medical Journal, 325, 1387.
Related references
Sala, E., Watson, C.J., Beadsmoore, C., et al. (2007) A randomized, controlled trial of routine early
abdominal computed tomography in patients presenting with non-specific acute abdominal pain.
Clinical Radiology, 62, 961–9.
Laméris, W., van Randen, A., van Es, H.W., et al. (2009) Imaging strategies for detection of urgent
conditions in patients with acute abdominal pain: diagnostic accuracy study. British Medical Journal,
338: b2431.
Study design
Randomization Early CT (<24 hours after admission) versus standard practice
Setting Large teaching hospital in England
Inclusion criteria Patients with acute abdominal pain admitted between 9a.m. Friday and
5p.m. Sunday from October 1999 to September 2000
Exclusion criteria Under 18 years
Pregnant
Patient requiring emergency surgery/CT
Rectal bleeding, perianal abscesses, gynaecological disorders
◆ Standard practice consisted of plain X-rays and, if appropriate, ultrasonography, CT,

and fluoroscopic investigations.
◆ Patients were followed up for 6 months to determine final diagnosis.
CT ASSESSMENT OF ACUTE ABDOMEN 63
Outcome measures
◆ Length of hospital stay and accuracy of diagnosis.
Results
◆ Early CT reduced length of hospital stay by 1.1 days (mean 5.3 versus 6.4 days,
p = 0.17) but this was not significant.
◆ 50% of diagnoses at admission were correct at follow-up at 6 months.
◆ 76% of diagnoses at 24 hours were correct at follow-up at 6 months with no
difference between groups.
◆ Early CT missed significantly fewer serious diagnoses at 24 hours (4% versus 21%,
p = 0.006).
◆ No patient in the early CT group died compared with seven patients in the standard
group (p = 0.014).
Conclusions
Early abdominopelvic CT for acute abdominal pain improves the accuracy of diagnosis
and may reduce length of hospital stay and mortality.
Critique
The early assessment of acute abdominal pain can be difficult, with uncertainty leading to
potential delays in appropriate treatment. Imaging forms an important part of the work-
up of emergency patients presenting to surgery. How different imaging modalities are
employed can vary widely between centres. The study by Ng et al. was one of the first ran-
domized trials to look at the role of early CT in the assessment of acute abdominal pain.
The findings of this study suggest that diagnostic accuracy can be improved for adult
patients with acute abdominal pain by a policy of early CT (within 24 hours of admission).
Although diagnostic accuracy was improved, Ng et al. did not find a significant difference
in length of hospital stay as a result. The same group (Sala et al.) performed another larger
RCT, with CT carried out within 1 hour of admission. This study also found improved
diagnostic certainty, but no difference in length of stay or mortality.
While these studies support the use of CT in the assessment of acute abdominal pain,
they do not provide sufficient evidence to indicate early CT for all emergency patients.
Lameris et al. studied the diagnostic accuracy of different imaging modalities to deter-
mine the strategy with the best sensitivity and lowest radiation dose. They suggested
ultrasound, followed by CT where results were negative or inconclusive, to be the best
approach. The results of this large study of over 1000 patients were based on diagnostic
accuracy rather than final patient outcomes.
Early diagnosis in patients with acute abdominal pain is important from both a clini-
cal and an organizational perspective. Prolonged periods of observation can no longer
be supported, and early imaging is more frequently employed to allow rapid diagnosis
and treatment. In appendicitis, CT has been shown to reduce negative appendicec-

tomy rates, improve patient care, and reduce the use of hospital resources (Rao et al.).
Furthermore the concerns regarding radiation dose have to some extent been addressed
by the use of ‘low-dose CT’ (Kim et al.).
Accurate pre-operative planning also allows patients to be managed by the most appro-
priate surgical team at a time of increased sub-specialization. While the impact of early
imaging is clear, the exact method of stratifying imaging modalities is less so, particularly
as technology improves and the role of newer modalities such as magnetic resonance
imaging (MRI) remains to be determined.
Rao, P.M., Rhea, J.T., Novelline, R.A., Mostafavi, A.A., and McCabe, C.J. (1998) Effect of computed
tomography of the appendix on treatment of patients and use of hospital resources. New England
Journal of Medicine, 338, 141–6.
Kim, K., Kim, Y.H., Kim, S.Y., et al. (2012) Low-dose abdominal CT for evaluating suspected
appendicitis. New England Journal of Medicine, 366, 1596–605.
EARLY LAPAROSCOPY IN THE ASSESSMENT OF ACUTE ABDOMINAL PAIN 65
3.4 Early laparoscopy in the assessment of acute

abdominal pain
Details of studies
It is now four decades since the first studies examining the role of laparoscopy as a diag-
nostic tool in the assessment of acute abdominal pain. In this time the assessment of
the acute abdomen has evolved. The aim of newer modalities has been to improve clinical
outcome by increasing diagnostic accuracy and avoiding potential delay associated with
a traditional ‘wait and see’ approach. Diagnostic laparoscopy has been applied particu-
larly to non-specific abdominal pain (NSAP), defined as acute abdominal pain of less
than 7 days duration, for which there is no diagnosis after examination and baseline
investigations.
Study references
Main studies
Case series
Sugarbaker, P.H., Sanders, J.H., Bloom, B.S., and Wilson, R.E. (1975) Preoperative laparoscopy in
diagnosis of acute abdominal pain. Lancet, 1, 442–5.
Randomized trials
Decadt, B., Sussman, L., Lewis, M.P., et al. (1999) Randomized clinical trial of early laparoscopy in the
management of acute non-specific abdominal pain. British Journal of Surgery, 86, 1383–6.
Morino, M., Pellegrino, L., Castagna, E., Farinella, E., and Mao, P. (2006) Acute nonspecific abdominal
pain: a randomized, controlled trial comparing early laparoscopy versus clinical observation. Annals
of Surgery, 244: 881–8.
Related references
Paterson-Brown, S., Eckersley, J.R., Sim, A.J., and Dudley, H.A. (1986) Laparoscopy as an adjunct to
decision making in the acute abdomen. British Journal of Surgery, 73: 1022–4.
Case series
Study design
◆ Inclusion criteria: patients admitted to a North American hospital with acute
abdominal pain.
◆ Patients thought to have a definite diagnosis proceeded to laparotomy.
◆ All patients with an uncertain diagnosis underwent laparoscopy.
Results
Laparoscopy (n = 29) Laparotomy (n = 27)
Operable lesion 28 21
No operable lesion 1 6
Diagnostic error 3% 22%
Hospital stay 5 days 6 days
Randomized trials
Study design
◆ Two single-centre RCTs (level of evidence 1b).
Decadt et al. Morino et al.

Randomization Early laparoscopy versus active observation
Setting British surgical unit (1995–1998) Italian surgical unit (2001–2004)
Inclusion criteria All patients admitted with NSAP Women aged 13–45 years with NSAP
Outcome measures Diagnosis
Operative procedures, hospital stay, readmission rate, morbidity, mortality,
patient satisfaction, recurrent symptoms
Follow-up 21 months 29.3 months
◆ In the trial by Morino et al., baseline investigations prior to randomization included

ultrasonography in addition to the blood tests, urinalysis, pregnancy test, and
abdominal radiograph carried out in the Decadt study.
Results
Laparoscopy Observation p value
Decadt et al.
Diagnosis 48/59 (81%) 22/61 (36%) <0.001
Hospital stay 2 days 2 days ns
Morbidity 14/59 (24%) 19/61 (31%) 0.362
Readmission rate 17/59 (29%) 20/61 (33%) 0.637
Morino et al.
Diagnosis 44/53 (83%) 23/51 (45%) <0.05
Hospital stay 3.7 days 4.7 days <0.05
Morbidity No difference
Recurrent pain at 1 year 16% 25% ns
EARLY LAPAROSCOPY IN ASSESSMENT OF ACUTE ABDOMINAL PAIN 67
◆ 44/61 (72%) patients in Decadt’s study and 25/51 (49%) patients in Morino’s study
avoided surgery in the active observation arms without increased morbidity.
◆ Decadt et al. reported improved quality of life for patients in the early laparoscopy group.
Conclusion
Early laparoscopy increases diagnostic accuracy and may reduce hospital stay in patients
with NSAP. Morbidity, mortality, and readmission rates are similar when comparing early
laparoscopy with active observation.
Critique
The first study by Sugarbaker et al. looking at the role of diagnostic laparoscopy in the
assessment of the acute abdomen was carried out in an era before ultrasound and CT
were widely available. Laparoscopy was compared against clinical diagnosis and selective
exploratory laparotomy, with Sugarbaker showing that laparoscopy reduced the ‘nega-
tive’ laparotomy rate with an increased diagnostic yield. In this study all patients with an
uncertain diagnosis at admission went on to have laparoscopy performed.
In a similar trial Paterson-Brown found that laparoscopy in patients where the need for
operation was uncertain reduced errors from 19% to zero. The same group also found that
laparoscopy reduced ‘unnecessary appendicectomy’ from 22%, when the clinician was
confident of the diagnosis, to 8% in a group where the need for operation was uncertain.
Non-specific abdominal pain has been reported in up to 40% of patients presenting
with an acute abdomen. Negative appendicectomy rates of 15–30% were also reported,
although this figure can be reduced with active observation (section 3.1). The RCTs by
Decadt and Morino examine a more selective use of laparoscopy in NSAP, with the most
recent trial employing ultrasonography as a baseline examination in all patients. Both
studies show that early laparoscopy can increase diagnostic accuracy. This may be par-
ticularly applicable to women, where the differential diagnosis covers a wider range of
conditions. The majority of the increase in diagnostic yield can be attributed to diagnosis
of acute appendicitis. It is difficult to know whether appendicitis diagnosed purely on
the basis of histology is clinically relevant, as the natural history of ‘early appendicitis’
remains unclear. It may be that some of these patients would have settled with conserva-
tive treatment without the need for operation. It could also be argued that since 50–70%
of patients avoided operation in the active observation arms without increased morbidity
or re-admission rates, this is a safe practice.
The management of patients with NSAP has evolved as new diagnostic tools have been
developed. Active observation is safe and is often combined with radiological investiga-
tions such as ultrasonography and increasing use of CT. Selective laparoscopy continues
to have a role in patients where the diagnosis remains uncertain.
Paterson-Brown, S. (1993) Emergency laparoscopic surgery. British Journal of Surgery, 80, 279–83.
Sheridan, W.G., White, A.T., Havard, T., et al. (1992) Non-specific abdominal pain: the resource
implications. Annals of the Royal College of Surgeons of England, 74, 181–5.
3.5 Laparoscopic versus open appendicectomy

Details of studies
Many papers studying the diagnostic and therapeutic role of laparoscopy in appendicitis
have been published. Before the use of diagnostic laparoscopy the negative appendicec-
tomy rate was around 30% in young women presenting with right iliac fossa pain. In 1988
Paterson-Brown et al. studied the diagnostic role of laparoscopy in suspected appendicitis
(see section 3.4). In addition to the diagnostic role of laparoscopy in the management of
suspected appendicitis, there have been numerous studies comparing the clinical out-
comes of laparoscopic versus open appendicectomy. Two of the largest randomized trials
in the literature are those of Hellberg et al. (1999) and Pedersen et al. (2001).
Study references
Main studies
Hellberg, A., Rudberg, C., Kullman, E., et al. (1999) Prospective randomized multicentre study of lapa-
roscopic versus open appendicectomy. British Journal of Surgery, 86: 48–53.
Pedersen, A.G., Petersen, O.B., Wara, P., Rønning, H., Qvist, N., and Laurberg, S. (2001) Randomized
clinical trial of laparoscopic versus open appendicectomy. British Journal of Surgery, 88, 200–5.
Related references
Sauerland, S., Jaschinski, T., and Neugebauer EAM. (2010) Laparoscopic versus open surgery for sus-
pected appendicitis. Cochrane Database of Systematic Reviews, Issue 10, CD001546.
Paterson-Brown, S., Thompson, J.N., Eckersley, J.R.T., Ponting, G.A., and Dudley, H.A.F. (1988) Which
patients with suspected appendicitis should undergo laparoscopy? British Medical Journal, 296:
1363–4.
Study design
◆ Two multi-centre RCTs amounting to class 1b evidence.
Hellberg et al.
Randomization Laparoscopic versus open appendicectomy
Number of patients 523 patients randomized
Setting One university and four county hospitals in Sweden
Inclusion criteria Patients >15 years with suspected acute appendicitis
Patients randomized if surgeon competent at laparoscopic surgery available
Exclusion criteria Contraindications to laparoscopic surgery
Outcome measures Sick leave as a measure of full recovery
Hospital stay, pain, functional status, complications
Follow-up 28 days from operation by independent observer
LAPAROSCOPIC VERSUS OPEN APPENDICECTOMY 69
Pedersen et al.
Randomization Laparoscopic versus open appendicectomy
Number of patients 583 patients randomized (245 non-randomized)
Setting Three units in two university hospitals in Denmark
Inclusion criteria Clinical diagnosis of acute appendicitis
Patients <15 years were included
Exclusion criteria Contraindications to laparoscopic surgery
Outcome measures Hospital stay
Post-operative morbidity, operating time, convalescence, cosmesis
Follow-up 4 weeks post-discharge
Results
Hellberg et al.
◆ 523 patients randomized, 23 withdrawals (laparoscopic conversion rate 12%)
Laparoscopic group Open group p value

Sick leave 11 days 14 days 0.06
Subjective full recovery 13 days 21 days <0.001
Operating time 60 minutes 35 minutes <0.01
◆ No difference in complication rate or length of hospital stay was seen
Pedersen et al.
◆ 583 patients randomized over 19 months. 245 patients were not randomized
because of lack of consent, suspicion of peritonitis caused by something other than
appendicitis, non-adherence to protocol, etc.
◆ 65 patients randomized to laparoscopic surgery were converted (23%)
◆ Median length of hospital stay was 2 days in both groups
Laparoscopic group Open group p value

Operating time 60 minutes 40 minutes <0.05
Return to normal activity 7 days 10 days <0.05
Return to work 10 days 16 days <0.05
Wound infection 2.8% 7% <0.03
Intra-peritoneal abscess 4.6% 1% <0.03
◆ Cosmesis was improved in the laparoscopic group (p<0.001)

Conclusions
Hellberg et al.
Laparoscopic appendicectomy is as safe as open appendicectomy and has the advantage
of allowing a quicker recovery.
Pedersen et al.
Hospital stay is equally short, but laparoscopic appendicectomy is associated with fewer
wound infections, faster overall recovery, earlier return to work, and improved cosmesis.
Critique
It is clear from these and other similar trials that laparoscopic appendicectomy takes lon-
ger to perform than the open approach and operative costs are higher. With these factors
in mind, laparoscopic surgery must demonstrate clear benefit in comparison with open
appendicectomy to justify its use.
The multi-centre RCT by Hellberg et al. is noteworthy as one of the largest random-
ized studies and suggested that laparoscopic appendicectomy is as safe as the open proce-
dure. They also reported an additional benefit of reduced time to perceived full recovery.
One difficulty with this trial was the availability of a trained laparoscopic surgeon, which
meant they were unable to recruit consecutive patients. The issue of training is pertinent
to the debate, with some arguing that trials of this kind may favour the laparoscopic group
because of the input of a more senior surgeon when compared with open appendicec-
tomy. It is also argued that the open technique is valuable as a training procedure for
junior surgeons.
Pedersen et al. reported a similar large RCT comparing outcomes following laparoscopic
and open appendicectomy. They identified over 800 patients with the clinical diagnosis of
appendicitis over a 19-month period, but failed to randomize approximately 30% of them.
This raises concerns that selection bias may influence the results. They found no differ-
ence in length of hospital stay between the groups, in keeping with Hellberg’s study. They
did find reduced wound infection rates and an earlier return to normal activity, results
borne out by later meta-analyses. Pedersen’s laparoscopic appendicectomy group also had
a higher rate of intra-peritoneal abscesses, which the authors thought were attributed to a
higher number of gangrenous and perforated appendices in the laparoscopic group. The
increased rate of intra-abdominal abscess with laparoscopic surgery has been a consistent
finding in meta-analyses and is one of the principal concerns with this approach.
These two trials are an accurate reflection of the results in the current literature on this
topic. The benefit of laparoscopy in young, female, and obese patients is perhaps clearer,
with data suggesting less wound morbidity and a faster return to normal activity. Despite
this, debate continues over the exact role of laparoscopic appendicectomy, particularly in
light of the increased rate of intra-abdominal abscess and higher cost.
Jørgensen, L.N. and Wille-Jørgensen, P. (2009) Open or laparoscopic appendicectomy? Colorectal
Disease, 11, 795–6.
NON-OPERATIVE MANAGEMENT OF ACUTE COMPLICATED APPENDICITIS 71
3.6 Non-operative management of complicated appendicitis

Details of studies
Around 2–6% of patients with appendicitis present with either an inflammatory mass
or an appendiceal abscess. Conservative or non-operative management was proposed by
Ochsner as early as 1901 and has further developed with the introduction of antibiotics
and percutaneous drainage of intra-abdominal abscesses. Several case series have been
published suggesting improved outcomes for the non-operative approach compared with
the high morbidity reported with immediate appendicectomy/open drainage. Supporting
evidence is from predominantly non-randomized trials, with one recent meta-analysis.
Study references
Main studies
Case series
Foran, B., Berne, T.V., and Rosoff, L. (1978) Management of the appendiceal mass. Archives of Surgery,
113, 1144–5.
Jordan, J.S., Kovalcik, P.J., and Schwab, W. (1981) Appendicitis with a palpable mass. Annals of Surgery,
193, 227–9.
Skoubo-Kristensen, E. and Hvid, I. (1982) The appendiceal mass: results of conservative management.
Annals of Surgery, 196, 584–7.
Paull, D.L. and Bloom, P. (1982) Appendiceal abscess. Archives of Surgery, 117, 1017–19.
Hoffmann, J., Lindhard, A., and Jensen, H.E. (1984) Appendix mass: conservative management without
interval appendicectomy. American Journal of Surgery, 148, 379–82.
Bagi, P. and Dueholm, S. (1987) Nonoperative management of the ultrasonically evaluated appendiceal
mass. Surgery, 101, 602–5.
Oliak, D., Yamini, D., Udani, V.M., et al. (2001) Initial nonoperative management for periappendiceal
abscess. Diseases of the Colon and Rectum, 44, 936–41.
Meta-analysis
Simillis, C., Symeonides, P., Shorthouse, A.J., and Tekkis, P.P. (2010) A meta-analysis comparing con-
servative treatment versus acute appendectomy for complicated appendicitis (abscess or phlegmon).
Surgery, 147, 818–29.
Related references
Bradley, E.L. and Isaacs, J. (1978) Appendiceal abscess revisited. Archives of Surgery, 113: 130–2.
Case series
Study design
◆ Non-randomized case series.
Results
Series No. of patients Diagnosis Findings
Foran 1978 30 non-operative vs Palpable mass ◆ Hospital stay 11 vs 7.5 days
13 operative ◆ 30% failure of non-operative
approach
◆ No difference in morbidity
Jordan 1981 42 operative Palpable mass ◆ 36% morbidity rate including
wound infection, abscesses
and ileus
Skoubo-Kristensen 193 non-operative Palpable mass or ◆ 12% failure rate
1982 intra-operative ◆ 0.5% mortality
Paull 1982 10 non-operative vs Intra-operative ◆ Hospital stay 9.3 vs 15.5 days
51 operative ◆ 18% major morbidity in
operative group
Hoffmann 1984 59 non-operative Palpable mass ◆ 21% failure rate
◆ 20% recurrence
Bagi 1987 40 non-operative Palpable mass + USS ◆ 7% failure rate
percutaneous drainage ◆ 7% recurrence
Oliak 2001 88 non-operative vs Intra-operative or CT ◆ 5.8% failure rate
67 operative ◆ Complication rate 17% vs 36%
◆ No difference in hospital stay
◆ Series varied in their use of interval appendicectomy following initial non-operative

management
Meta-analysis
Study design
◆ Simillis et al. included 16 non-randomized studies comparing conservative treatment
and acute appendicectomy for patients presenting with complicated appendicitis.
◆ Studies included reported at least one of the following outcome measures: hospital
stay, duration of antibiotics, complications, and re-operations.
◆ Odds ratios were used as the summary statistic for comparison of dichotomous
variables, i.e. complications. Weighted mean differences were used to analyse
continuous variables such as hospital stay and duration of antibiotics.
Results
◆ Conservative management was associated with fewer overall complications, wound
infections, intra-abdominal abscess, ileus/obstruction, and re-operation.
◆ No difference was found in duration of hospital stay or duration of antibiotic
treatment.
NON-OPERATIVE MANAGEMENT OF ACUTE COMPLICATED APPENDICITIS 73
◆ No difference in results was seen when analysis of high-quality, recent, or larger-sized

studies was carried out.
◆ No evidence of publication bias was identified.
Critique
Complicated appendicitis describes a group of patients who present with either an inflam-
matory phlegmon or localized perforation of the appendix with abscess formation. The
safety of conservative management for complicated appendicitis has been known for over
a century and favoured by the majority of surgeons. The rationale of such an approach is
to avoid disseminating an already localized septic focus through a potentially challenging
procedure. This was supported by high morbidity rates of 30–35% in historical series of
aggressive operative intervention.
Initial studies comparing conservative and operative management suffered from a lack
of consistency in the use of the terms ‘abscess’ and ‘mass’. Diagnosis was usually based on
either clinical or operative findings, and the rate of missed/alternative diagnoses in early
series was 5–10%. With the subsequent availability of ultrasound scanning (USS) and CT,
more accurate pre-operative diagnosis became possible with the option of percutaneous
guided drainage of abscess collections. Where previously surgeons were concerned about
managing appendiceal abscesses conservatively, later series which included percutaneous
drainage showed high success rates, comparable with those for non-operative treatment
of inflammatory masses. The use of colonoscopy and CT following successful conserva-
tive treatment, particularly in patients aged over 40 years, is advised to reduce the rate of
missed diagnoses, particularly colon cancer.
The evidence for non-operative management of complicated appendicitis comes pri-
marily from retrospective analysis of non-randomized case series. Confounding factors,
which may have biased the selection of patients for operative or non-operative treatment,
are likely to have been present. Conservative management has also changed over the
period covering these series, with the development of new antibiotics and the introduc-
tion of minimally invasive techniques. The physiological status of patients in each group
was rarely reported and criteria for failure of non-operative management were unclear.
Meta-analysis may simply magnify the inherent biases of these series and significant het-
erogeneity between studies was reported. Despite this, the best available evidence con-
tinues to favour the non-operative approach, with high success rates (90–95%), reduced
complications, and no significant difference in hospital stay.
Ochsner, A.J. (1901) The cause of diffuse peritonitis complicating appendicitis and its prevention.
JAMA, 36, 1747–54.
McPherson, A.G. and Kinmonth, J.B. (1945) Acute appendicitis and the appendix mass. British Journal
of Surgery, 32, 365–70.
Thomas, D.R. (1973) Conservative management of the appendix mass. Surgery, 73, 677–80.
3.7 Interval appendicectomy

Details of studies
Around 2–6% of patients with appendicitis present with either an inflammatory mass or
an appendiceal abscess. Non-operative management in this setting has been shown to be
both safe and effective (see section 3.6). The role of interval appendicectomy (IA) follow-
ing non-operative management remains controversial.
This population-based study from Southern California (Kaminski et al. 2005), look-
ing at the rate of recurrent appendicitis in 864 patients after non-operative treatment of
appendix mass without routine IA, is the largest trial of its kind in the literature.
Study references
Main study
Kaminski, A., Liu, I.A., Applebaum, H., Lee, S.L., and Haigh, P.I. (2005) Routine interval appendectomy is
not justified after initial nonoperative treatment of acute appendicitis. Archives of Surgery, 140, 897–901.
Related references
Lai, H.W., Loong, C.C., Chiu, J.H., et al. (2006) Interval appendectomy after conservative treatment of
an appendiceal mass. World Journal of Surgery, 30, 352–7.
Willemsen, P.J., Hoorntje, L.E., Eddes, E.H., and Ploeg, R.J. (2002) The need for interval appendectomy
after resolution of an appendiceal mass questioned. Digestive Surgery, 19, 216–21.
Skoubo-Kristensen, E. and Hvid, I. (1982) The appendiceal mass: results of conservative management.
Study design
◆ Retrospective cohort study using discharge data.
Randomization Non-randomized
Setting 12 regional hospitals in Southern California
Time period 1992–2004
Follow-up* 4 years (6 months—12 years)
*Median (range)
◆ Treatment was decided at the discretion of the individual surgeon.

◆ Non-operative treatment included percutaneous abscess drainage.
◆ Operative records were examined for subsequent appendectomies to determine
whether they were for recurrent appendicitis or elective IA.
Outcome measure
◆ Hospitalization for recurrent appendicitis or IA.
INTERVAL APPENDICECTOMY 75
Results
◆ 1012 patients treated non-operatively represented 3% of all patients admitted with
appendicitis.
◆ 148 (15%) patients had an IA and 864 (85%) did not have an IA.
◆ 39 (5%) patients developed recurrent appendicitis during follow-up.
◆ Mean time to recurrence was 10 months.
◆ Median length of stay for IA was 6 days compared with 4 days for recurrent
appendicitis (p = 0.006).
◆ Age, comorbidity, presence of an appendix abscess, and percutaneous abscess
drainage were not independent risk factors for recurrence.
Critique
The role of interval appendicectomy following non-operative management of an appen-
dix mass remains a matter of controversy. Traditional management has been to carry out
IA about 12 weeks after the index admission to avoid the risk of recurrent appendicitis,
as well as missed diagnoses. In a survey of surgeons in the Mid-Trent region reported in
2005, 75% favoured following this ‘classical approach’. The arguments against routine IA
include the morbidity of the procedure and increased hospital stay and related costs.
Kaminski et al., reporting the largest study of recurrent appendicitis following non-
operative management of appendix mass, found a recurrence rate of 5%. They conclude
that routine IA is not justified with such a low recurrence rate. Other studies report recur-
rence rates ranging from 5% to 25% (Skoubo-Kristensen and Hvid 1982; Lai et al. 2006),
with the majority occurring in the first 6 months following discharge. Up to 60% of those
patients who recur may present before their scheduled interval appendicectomy.
The present study also found that duration of hospital stay for recurrent appendicitis
was less than for interval appendicectomy. This is in keeping with other trials which sug-
gest that recurrent appendicitis follows a milder clinical course and can be treated with
either appendicectomy or non-operative management without increased morbidity or
mortality (Dixon et al. 2003). No accurate predictors for recurrence have been shown in
the surgical literature, consistent with this study.
Hospital stay and morbidity related to routine IA is difficult to determine. While mor-
bidity from IA reported in the literature varies from 9% to 18%, this may not reflect cur-
rent practice (Thomas 1973; Willemsen et al. 2002). This study reports a hospital stay of
6 days for IA, while the surgical technique, recovery pathway, and complication rate are
not reported. The few studies that describe IA by a laparoscopic approach and with more
modern recovery pathways suggest lower hospital stays and morbidity.
Conclusions
Interval appendicectomy does not appear justified by the low rates of recurrent appendi-
citis following non-operative management of an appendix mass. Recurrent appendicitis
is more common within the first 6 months following discharge and can be treated without
an increase in morbidity or mortality. Most authors advise follow-up investigations with
colonoscopy/CT to reduce the risk of missing malignancy of the colon or appendix, par-
ticularly in the age group over 40 years. No clinical features have been shown to accurately
predict recurrence of appendicitis.
Ahmed, I., Deakin, D., and Parsons, S. (2005) Appendix mass: do we know how to treat it? Annals of
the Royal College of Surgeons of England, 87, 191–5.
Dixon, M.R., Haukoos, J.S., Park, I.U., et al. (2003) An assessment of the severity of recurrent
appendicitis. American Journal of Surgery, 186, 718–22.
Thomas, D.R. (1973) Conservative management of the appendix mass. Surgery, 73, 677–80.
Oliak, D., Yamini, D., Udani, V.M., et al. (2001) Initial non-operative management of periappendiceal
abscess. Diseases of the Colon and Rectum, 44, 936–41.
NON-OPERATIVE TREATMENT OF PERFORATED DUODENAL ULCER 77
3.8 Non-operative treatment of perforated duodenal ulcer

Details of studies
For more than a century the standard treatment for perforated peptic ulcer has been sur-
gical repair. The understanding that some perforations will seal spontaneously has also
been around for a long time, with one report by Crisp dating to 1843. Early series of
non-operative treatment for perforated peptic ulcer suggested acceptable outcomes, with
a mortality of 11–14% reported by Taylor. This compared favourably with mortality from
surgery at the time, which approached 20%. The first randomized trial on this subject
presented here was not published until 1989.
Study references
Main study
Crofts, T.J., Park, K.G.M., Steele, R.J.C., Chung, S.S.C., and Li, A.K.C. (1989) A randomized trial of non-
operative treatment for perforated peptic ulcer. New England Journal of Medicine, 320, 970–3.
Related references
Taylor, H. (1946) Perforated peptic ulcer treated without operation. Lancet, 2, 441–4.
Visick, A.H. (1946) Conservative treatment of acute perforated peptic ulcer. British Medical Journal, 2,
941–4.
Berne, T.V. and Donovan, A.J. (1989) Nonoperative treatment of perforated duodenal ulcer. Archives of
Surgery, 124, 830–2.
Study design
◆ Randomized controlled trial.
Randomization Operative versus non-operative treatment of perforated duodenal ulcer
Setting Prince of Wales Hospital, Hong Kong
Time period 13 months (1985–1986)
Follow-up Peri-operative period
◆ Inclusion criteria: patients with a clinical diagnosis of perforated peptic ulcer.

◆ Exclusions: one patient with previous diagnosis of advanced gastric cancer.
◆ Non-operative treatment included close observation, nasogastric suction,
intravenous antibiotics, and intravenous ranitidine.
Outcome measures
◆ Morbidity and mortality.
◆ Hospital stay, avoidance of operation.
Results
Operative (n = 43) Non-operative (n = 40) p value
Morbidity 17 (40%) 20 (50%) ns
Mortality 2 (5%) 2 (5%) ns
Hospital stay 7.8 days 12 days <0.001
◆ 11/40 (28%) patients in the non-operative group showed no sign of improvement at

12 hours and underwent operation.
◆ 6/9 patients over 70 years failed conservative treatment.
◆ The diagnosis was incorrect in 4/83 patients, comprising three gastric cancers and
one sigmoid cancer.
Conclusion
Non-operative treatment of perforated duodenal ulcer is safe in selected patients and with
close observation.
Critique
Non-operative management of perforated peptic ulcer is not a new concept. In 1935
Wangensteen advised conservative management, particularly for gravely ill patients
where the presentation was delayed. Herman Taylor published his initial series on non-
operative management in 1946, eventually reporting on over 200 patients. Using gastric
drainage alone and without antibiotics or anti-secretory medication, Taylor reported a
mortality of 11%.
Donovan et al. published a later series utilizing the non-operative approach in a select
group of patients where a water-soluble contrast study showed no evidence of continu-
ing leakage. In a group of 35 patients there was one (3%) death compared with 16 (6.2%)
deaths in the 259 patients treated surgically over the same period.
The first randomized trial carried out by Crofts et al. confirmed that a non-operative
approach can be followed without increased morbidity or mortality: 72% of patients
treated in this way successfully avoided an operation at the expense of a 35% longer
hospital stay. Despite confirmation of its safety, the non-operative approach has failed
to gain widespread acceptance. The authors highlighted that conservative management
was rarely successful in patients aged over 70 years. This is disappointing in a group that
would perhaps be the most attractive to treat in this way. Close observation by an experi-
enced surgeon is also suggested to be important so that conversion to operative manage-
ment is not delayed to a point where the physiological status of the patient is irreversible.
Current practice in most centres is to reserve conservative management for patients
with a delayed presentation who show signs of improvement and for patients deemed too
high risk for surgery.
NON-OPERATIVE TREATMENT OF PERFORATED DUODENAL ULCER 79
Crisp, E. (1843) Cases of perforation of the stomach, with deductions therefrom relative to the character
and treatment of that lesion. Lancet, 40, 639–49.
Wangensteen, O.H. (1935) Non-operative treatment of localized perforation of the duodenum.
Minnesota Medicine, 18, 477–80.
Taylor, H. (1957) Non-surgical treatment of perforated peptic ulcer. Gastroenterology, 33, 353–8.
Donovan, A.J., Vinson, T.L., Maulsby, G.O., and Gewin, J.R. (1979) Selective treatment of duodenal
ulcer with perforation. Annals of Surgery, 189, 627–36.
3.9 Eradication therapy after simple closure of perforated

duodenal ulcer
Details of studies
The aim of surgery for perforated duodenal ulcer is primarily to save the patient from
the effects of peritonitis. A secondary issue at the time of emergency presentation is the
treatment of the ulcer itself. Primary closure of the perforation has been the most com-
mon operation performed in the emergency setting. The role of definitive acid-reduction
surgery for perforated duodenal ulcers was an area of much controversy in the 1970s and
1980s. The balance of the argument then changed significantly following the discovery of
Helicobacter pylori in 1982 and effective eradication therapy in 1987. The randomized trial
by Ng et al. was the first to determine whether H. pylori eradication in combination with
simple repair led to long-term ulcer remission.
Study references
Main study
Ng, E.K.W., Lam, Y.H., Sung, J.J., et al. (2000) Eradication of Helicobacter pylori prevents recurrence of
ulcer after simple closure of duodenal ulcer perforation: a randomized controlled trial. Annals of
Surgery, 231, 153–8.
Related references
Drury, J.K., McKay, A.J., Hutchison, J.S., and Joffe, S.N. (1978) Natural history of perforated duodenal
ulcers treated by suture closure. Lancet, 2, 749–50.
Bornman, P.C., Theodorou, N.A., Jeffery, P.C., et al. (1990) Simple closure of perforated duodenal ulcer:
a prospective evaluation of a conservative management policy. British Journal of Surgery, 77, 73–5.
Griffin, G.E. and Organ, C.H., Jr (1976) The natural history of the perforated duodenal ulcer treated by
suture plication. Annals of Surgery, 183, 382–5.
Tanphiphat, C., Tanprayoon, T., and Na Thalang, A. (1985) Surgical treatment of perforated duodenal
ulcer: a prospective trial between simple closure and definitive surgery. British Journal of Surgery, 72,
370–2.
Christiansen, J., Andersen, O.B., Bonnesen, T., and Baekgaard, N. (1987) Perforated duodenal ulcer
managed by simple closure versus closure and proximal gastric vagotomy. British Journal of Surgery,
74, 286–7.
Study design
◆ Randomized controlled trial.
Randomization Eradication therapy versus omeprazole (4 weeks) alone following simple
repair of perforated duodenal ulcer
Setting Prince of Wales Hospital, Hong Kong
Inclusion criteria Perforated duodenal ulcer
H. pylori detected at intra-operative endoscopy
ERADICATION THERAPY AFTER CLOSURE OF PERFORATED DUODENAL ULCER 81
Exclusion criteria Age <16 or >75 years

Antibiotics or acid suppression <4 weeks before admission
Previous vagotomy or gastrectomy
Sealed perforation
Perforation during inpatient admission
Follow-up 1 year
◆ Patients were considered for definitive surgery if the perforation was too large for
omental patch repair or in the presence of haemorrhage or obstruction.
◆ Follow-up included endoscopy at 8 weeks, three-monthly interviews, and then final
endoscopy at 1 year.
Outcome measures
◆ Initial ulcer healing.
◆ Relapse rate at 1 year.
Results
◆ Groups were similar in age, sex, smoking habit, use of non-steroidal anti-
inflammatory drugs (NSAIDs), previous ulcer history, size of perforation, and
severity of peritoneal contamination.
◆ Of the 129 patients considered for the study, 104 were shown to have H. pylori (81%).
H. pylori eradication (n = 51) Omeprazole (n = 48) p value

H. pylori eradication rate 43 (84%) 8 (17%) <0.001
Initial ulcer healing 42 (82%) 42 (87%) 0.58
All ulcer recurrence 2 (5%) 16 (38%) <0.001
Symptomatic recurrence 1 (2%) 9 (21%) 0.02
Conclusion
Ulcer recurrence rates are low after simple suture repair of duodenal perforation followed
by eradication therapy. Acid-reducing surgery is not required as part of the emergency
procedure.
Critique
Suture repair of perforated duodenal ulcer has been the predominant emergency opera-
tion since the early 1900s. The procedure is simple and safe, even in the hands of relatively
inexperienced surgeons. While suture repair addresses the primary problem, evidence
from case series demonstrated that in the longer term patients suffered a high rate of ulcer
recurrence and re-operation. Fairly consistent findings between different centres showed
that recurrence ranged from 40% to 70% depending on whether results were based on
symptoms or endoscopic findings. Between 15% and 40% of this group required further
surgery either electively or for re-perforation.
In response to this problem some surgeons proposed definitive surgery at the time
of the initial emergency presentation. Partial gastrectomy was found to carry too high
a complication rate, and most groups focused on acid-reducing surgery in the form of
vagotomy with or without a drainage procedure. Randomized trials suggested that defini-
tive surgery significantly reduced ulcer recurrence and the need for re-operation with no
difference in morbidity or mortality.
Despite this evidence practice was slow to change. Insufficient expertise and resources
for what is often an out-of-hours procedure may have been part of the reason for the
reluctance to change. This evidence also came at the time when antisecretory medication
was becoming popular. As the incidence of peptic ulcer disease reduced, so did experi-
ence of elective ulcer surgery.
The controversy effectively ended with the discovery of H. pylori and the understand-
ing of its role in peptic ulcer disease and recurrence. One of the first studies was the trial
by Ng et al. While not a direct comparison of suture repair and definitive surgery, it is
clear that the results of suture repair combined with eradication of H. pylori compare
favourably with those of definitive surgery in the era before modern medical therapy.
These results were confirmed in similar studies performed later.
Boey, J. and Wong, J. (1987) Perforated duodenal ulcers. World Journal of Surgery, 11, 319–24.
Kate, V., Ananthakrishnan, N., and Badrinath, S. (2001) Effect of Helicobacter pylori eradication on
the ulcer recurrence rate after simple closure of perforated duodenal ulcer: retrospective and
prospective randomized controlled studies. British Journal of Surgery, 88, 1054–8.
Rodriguez-Sanjuán, J.C., Fernández-Santiago, R., García, R.A., et al. (2005) Perforated peptic ulcer
treated by simple closure and Helicobacter pylori eradication. World Journal of Surgery, 29, 849–52.
WATER-SOLUBLE CONTRAST IN SMALL BOWEL OBSTRUCTION 83
3.10 Water-soluble contrast in small bowel obstruction

Details of studies
Adhesive small bowel obstruction is the leading cause of small bowel obstruction (SBO)
in adults. Immediate surgery is carried out when there is suspected strangulation of
the bowel and the remaining patients are given a trial of conservative management.
A Cochrane review and a further more recent meta-analysis have been published to assess
the diagnostic role of water-soluble contrast (WSC) in adhesive SBO. The aim is to predict
which patients are most likely to fail on conservative management, hence avoiding any
delays to surgery. The two main studies have also attempted to identify whether the admin-
istration of WSC may have a therapeutic role in the non-operative management of SBO.
Study references
Main studies
Abbas, S., Bissett, I.P., and Parry, B.R. (2007) Oral water-soluble contrast for the management of
adhesive small bowel obstruction. Cochrane Database of Systematic Reviews, CD004651.pub3
(DOI: 10.1002/14651858).
Branco, B.C., Barmparas, G., Schnuriger, B., Inaba, K., Chan, L.S., and Demetriades, D. (2010) Systematic
review and meta-analysis of the diagnostic and therapeutic role of water-soluble contrast agent in
adhesive small bowel obstruction. British Journal of Surgery, 97, 470–8.
Related references
Chen, S.C., Lin, F.Y., Lee, P.H., Yu, S.C., Wang, S.M., and Chang, K.J. (1998) Water-soluble contrast
study predicts the need for early surgery in adhesive small bowel obstruction. British Journal of
Surgery, 85, 1692–4.
Biondo, S., Parés, D., Mora, L., Martí-Rague, J., Kreisler, E., and Jaurrieta, E. (2003) Randomized
clinical study of Gastrografin administration in patients with adhesive small bowel obstruction.
Di Saverio, S., Catena, F., Ansaloni, L., Gavioli, M., Valentino, M., and Pinna, A.D. (2008) Water-soluble
contrast medium (Gastrografin) value in adhesive small intestine obstruction (ASIO): a prospective,
randomized, controlled, clinical trial. World Journal of Surgery, 32, 2293–304.
Study designs
◆ Both studies comprise a systematic review and a meta-analysis of:
1. Prospective studies assessing the diagnostic potential of WSC in adhesive SBO;
2. RCTs evaluating the therapeutic role of WSC in adhesive SBO.
Number of studies 10 prospective clinical trials: 6 diagnostic; 6 therapeutic (Abbas et al)
14 prospective observational studies or clinical trials: 7 diagnostic; 9
therapeutic (Branco et al)
Primary outcomes Ability of WSC to predict need for surgery
Proportion of patients requiring surgery (Abbas et al)
Resolution of SBO without surgery (Branco et al)
Secondary outcomes Time from admission to resolution, length of hospital stay, complications,
and mortality
Inclusion criteria Patients admitted to hospital with clinical and radiological evidence of SBO
with a previous history of abdominal surgery
Exclusion criteria Surgery ≤6 weeks before obstructive episode; signs of strangulation,
peritonitis, abdominal malignancy, or irreducible hernia
Results
Study Results
Abbas et al. Diagnostic
◆ Pooled sensitivity 97%, specificity 96%

◆ Not possible to determine optimal timing of X-ray
Therapeutic
◆ Gastrografin reduced hospital stay in patients who did not require

surgery but did not cause resolution of SBO
◆ No difference in resolution of SBO without surgery, small bowel
strangulation, complications, or mortality between Gastrografin and
placebo
Branco et al. Diagnostic
◆ Presence of WSC in colon predicted resolution of SBO with 96%

sensitivity and 98% specificity
◆ No difference in sensitivity or specificity between timing of abdominal
X-ray at 4–8 hours or 24 hours
Therapeutic
◆ Significant reduction in the need for surgery was observed with

administration of WSC (p = 0.007)
◆ Significant reduced length of hospital stay (1.87 days)
Conclusions
◆ WSC is recommended as a predictive test for non-operative resolution of adhesive
SBO. WSC in the colon within 4–24 hours is predictive of resolution of SBO.
◆ There is strong evidence that WSC reduces the length of hospital stay and decreases
the need for surgery.
Critique
We know that conservative management is successful in up to 90% of patients who pres-
ent with adhesive SBO. What is more difficult is to predict which patient will resolve and
which will progress to strangulation or require a surgical remedy. These two systematic
reviews and meta-analyses have been discussed together as both set out to answer the
same questions. With regard to the diagnostic role of WSC, the studies come to the same
WATER-SOLUBLE CONTRAST IN SMALL BOWEL OBSTRUCTION 85
conclusion. Both found that the use of WSC can predict the probability of successful
conservative management. These reviews provide important information for the clinical
management of a common surgical emergency. They show that it is possible to decide
whether a patient will respond to conservative management within 24 hours of admission
with a simple, safe, and cost-effective diagnostic tool. Using this method it is possible to
reduce the length of hospital stay and delays in proceeding to surgery.
When determining the therapeutic effect of WSC the studies reach differing conclu-
sions. The Cochrane review concluded that WSC did not cause resolution of an adhesive
SBO, but did reduce hospital stay for patients who did not require surgery. Branco et al.
have included two more recent RCTs by Di Saverio et al. (2008) and Farid et al. (2010)
which led to the conclusion that the hyper-osmolar effects of WSC has a therapeutic role
in adhesive SBO in addition to reducing hospital stay.
One criticism has been that the studies included in these meta-analyses are of vari-
able quality, with some having low Jadad scores. It is also suggested that the variety of
presentations of SBO in these trials make them difficult to compare. In response to these
criticisms, Branco et al. repeated the analysis including only trials with a Jadad score ≥3
and found no difference in results from those previously published.
Farid, M., Fikry, A., El Nakeeb, A., et al. (2010) Clinical impacts of oral Gastrografin follow-through in
adhesive small bowel obstruction (SBO). Journal of Surgical Research, 162, 170–6.
Branco, B.C., Barmparas, G., Schnüriger, B., Inaba, K., Chan, L.S., and Demetriades, D. (2010) Authors’
reply: Systematic review and meta-analysis of the diagnostic and therapeutic role of water-soluble
contrast agent in adhesive small bowel obstruction. British Journal of Surgery, 97, 1311–12.
3.11 Surgery for left-sided malignant large bowel obstruction

Details of studies
Until the latter part of the twentieth century the surgical management of left-sided large
bowel obstruction (LBO) involved a two- or even three-stage technique. Most commonly
the initial operation involved a Hartmann’s procedure, although 30–60% of patients never
had their colostomy reversed. This led to interest in performing a single-stage procedure
with a primary anastomosis. There has been debate as to whether segmental resection
with intra-operative colonic lavage and primary anastomosis or a subtotal colectomy is
the superior operation for a malignant left-sided LBO. The SCOTIA trial was the first
prospective randomized clinical trial comparing these two procedures.
Study references
Main study
SCOTIA Study Group. (1995) Single-stage treatment for malignant left-sided colonic obstruction: a pro-
spective randomized clinical trial comparing subtotal colectomy with segmental resection following
intraoperative irrigation. British Journal of Surgery, 82, 1622–7.
Related references
Murray, J.J., Schoetz, D.J., Coller, J.A., Roberts, P.L., and Veidenheimer, M.C. (1991) Intraoperative
colonic lavage and primary anastomosis in nonelective colon resection. Diseases of the Colon and
Rectum, 34, 527–31.
Finan, P.J., Campbell, S., Verma, R., et al. (2007) The management of malignant large bowel obstruction:
ACPGBI position statement. Colorectal Disease, 9 (Suppl 4), 1–17.
Study design
◆ A multi-centre RCT of subtotal colectomy versus segmental colectomy with on-table
irrigation.
◆ Completion of standard forms for all patients including pre-operative, operative, and
post-operative parameters.
◆ Following hospital discharge, patients were followed up at 8 weeks, six-monthly for 2
years, and annually thereafter.
◆ Inclusion criteria: patients presenting as emergencies with presumed malignant left-
sided LBO who were suitable for both operations.
◆ Powered to detect differences in quantitative functional outcomes.
Outcome measures
◆ Early outcome measures:
● anastomotic leakage rate
● hospital mortality
● length of hospital stay
● rate of stoma formation.
SURGERY FOR LEFT-SIDED MALIGNANT LARGE BOWEL OBSTRUCTION 87
◆ Late outcome measures:

● bowel function
● late complications
● perceived outcome
● tumour recurrence.
Results
Endpoints Subtotal colectomy Segmental resection p value
(n = 47) (n = 44)
Early
Anastomotic leak 4/47 2/44 0.68
Hospital mortality 6/47 5/44 1.00
Length of stay 12 (10–14) days* 11 (9–13) days* 0.30
Permanent stoma 7/47 1/44
Late
Increased bowel frequency 14/35 4/35 0.01
(≥3/24 hour)
Night-time bowel movements 10/35 3/35 0.03
*Values are median (range)
◆ Significantly more patients in the subtotal colectomy group than in the segmental
resection group consulted their general practitioner with bowel problems (15/37
versus 3/35, p = 0.004).
◆ Postal questionnaire on Nottingham Health Profile: patients in the subtotal
colectomy group perceived more restrictions on their mobility because of
their health
Conclusions
Segmental resection following intra-operative irrigation is the preferred treatment for
left-sided malignant LBO because it has a better functional outcome. The safety of a sin-
gle-stage approach, with anastomotic leak <5% and mortality of around 10%, was also
confirmed. Subtotal colectomy is recommended for perforation of the caecum or in the
presence of synchronous tumours.
Critique
It is clear from the surgical literature that, compared with a single-stage procedure,
Hartmann’s procedure for left-sided LBO increases the likelihood of a patient living with
a permanent stoma. The SCOTIA trial investigated the single-stage procedure further
and demonstrated a functional benefit for patients undergoing segmental resection with
irrigation compared with a subtotal colectomy, without increasing the risk of anastomotic
leakage or mortality. The study also confirmed the safety of primary anastomosis with
anastomotic leak (<5%) and mortality at acceptable levels.
Patient numbers appear small, but the SCOTIA trial is the largest prospective RCT
available comparing these two operative groups. One of the main difficulties with stud-
ies comparing one-stage and two-stage procedures or subtotal and segmental colectomy
is the potential for selection bias. The Association of Coloproctology of Great Britain
and Ireland (ACPGBI) position statement in 2007 agreed that a single-stage procedure
with primary anastomosis is preferred. They also concluded that subtotal colectomy and
segmental resection are equally safe where there is a choice of operation, and that each
patient should be assessed individually, with surgeon preference taken into account. It is
accepted that some patients still remain unsuitable for primary anastomosis.
In recent years there has been increasing interest in the use of colonic stenting in the
management of malignant left-sided LBO as a bridge to elective resection. van Hooft et al.
randomized 98 patients to colonic stenting or emergency surgery, but had to suspend the
RCT because of an increased 30-day morbidity in the colonic stenting group. They con-
cluded that colonic stenting had no clinical advantages over emergency surgery, and there
were concerns that there may be an increased risk of tumour spread caused by perforation
as a complication of endoluminal stenting.
A recent meta-analysis included five trials and a total of 207 patients (102 underwent
stenting and 105 had emergency surgery). Stents successfully relieved obstruction in 78%
of patients, with reduced hospital stay but no reduction in mortality or complications.
Some of the trials included in the review did not take into consideration the potential
need for subsequent surgery to resect the colon cancer, which remains in situ following
insertion of a stent. Therefore additional hospitalization, procedure time, morbidity, and
potential mortality may be a part of the stent group when performed as a bridge to sur-
gery rather than as a definitive palliative treatment. Currently, the multi-centre CReST
(colorectal stenting) trial is randomizing patients to stenting or conventional treatment
with an attempt to recruit a minimum of 400 patients. It is hoped that this will give a
clearer picture of the role of stenting in malignant large bowel obstruction.
van Hooft, J.E., Bemelman, W.A., Oldenburg, B., et al. (2011) Colonic stenting versus emergency
surgery for acute left-sided malignant colonic obstruction: a multicentre randomised trial. Lancet
Oncology, 12, 344–52.
Sagar, J. (2011) Colorectal stents for the management of malignant colonic obstructions. Cochrane
Database of Systematic Reviews, 11, CD007378.
Saida, Y., Sumiyama, Y., Nagao, Y., and Uramatsu, M. (2003) Long-term prognosis of preoperative
‘bridge to surgery’ expandable metallic stent insertion for obstructive colorectal cancer: comparison
with emergency operation. Diseases of the Colon and Rectum, 46, S44–9.
Kim, J.S., Hur, H., Min, B.S., et al. (2009) Oncologic outcomes of self-expanding metallic stent insertion
as a bridge to surgery in the management of left-sided colon cancer obstruction: comparison with
nonobstructing elective surgery. World Journal of Surgery, 33, 1281–6.
MANAGEMENT OF UNCOMPLICATED DIVERTICULITIS 89
3.12 The management of uncomplicated diverticulitis

Details of studies
Diverticulosis is a common condition in the Western world with an increasing incidence
in younger patients. Prevalence increases with age, and by 80 years of age almost 80% of
people are affected. While a majority are asymptomatic, one in ten develop symptoms or
complications. The pathogenesis of diverticular disease is unclear and most popular theo-
ries are unable to explain all the different presentations of the disease.
Uncomplicated diverticulitis presents with lower abdominal pain and fever secondary
to inflammation of the colon. Standard treatment has included intravenous fluids and
antibiotics, as a bacterial infection is thought to be the cause. This has recently been chal-
lenged with the first RCT to compare antibiotic versus no antibiotic treatment.
Study references
Main study
Chabok, A., Påhlman, L., Hjern, F., Haapaniemi, S., Smedh, K, and AVOD study group. (2012) Randomized
clinical trial of antibiotics in acute uncomplicated diverticulitis. British Journal of Surgery, 99, 532–9.
Related reference
Hjern, F., Josephson, T., Altman, D., et al. (2007) Conservative treatment of acute colonic diverticulitis.
Are antibiotics always mandatory? Scandinavian Journal of Gastroenterology, 42, 41–7.
Study design
◆ A multi-centre RCT
Randomization Antibiotics versus no antibiotics for uncomplicated diverticulitis
Number of centres 11 centres in Sweden and Iceland
Study period 2003–2010
Acute lower abdominal pain/tenderness with temperature ≥38oC
Raised inflammatory markers (WCC/CRP)
Signs of diverticulitis on CT
Exclusion criteria Signs of complicated diverticulitis on CT (e.g. abscess, fistula, perforation)

High fever, peritonitis, or sepsis
Immunosuppressives, pregnancy, on antibiotics
Stratification By centre
Follow-up 12 months
◆ Imaging of the colon was carried out (colonoscopy, CT colonography, barium

enema) 6–8 weeks after discharge.
◆ A telephone or postal questionnaire was completed at 12 months.
Outcome measures
◆ Complications, need for surgery, hospital stay, abdominal pain, fever.
◆ Recurrence, long-term need for surgery, change in bowel habit, result of colorectal
examinations.
Results
◆ Baseline demographics were similar between the two groups.
Antibiotics (n = 314) No antibiotics (n = 309) p value

Complications 1.0% 1.9% ns
Hospital stay (days) 2.9 2.9 ns
Recurrent diverticulitis 15.8% 16.2% ns
◆ There was no difference in pain scores or normalization of temperature between the

groups during their hospital stay
◆ There were three perforations in each group with three additional abscesses in the
‘no-antibiotic’ group.
◆ In the ‘no-antibiotic’ group one patient required sigmoid resection and other
complications were treated conservatively. In the antibiotic group all three patients
required sigmoid resection.
◆ Six patients in the ‘no-antibiotic’ group had surgery in the follow-up period
(stricture, fistula, recurrent diverticulitis) compared with two patients in the
antibiotic group (strictures)
Conclusion
Antibiotics for uncomplicated diverticulitis do not accelerate recovery, reduce complica-
tions, or prevent recurrence.
Critique
This was a well-designed and reported multi-centre trial which suggests that uncom-
plicated acute diverticulitis does not require treatment with antibiotics. These results
challenge the idea that diverticulitis is due to bacterial infection and lends support to
an emerging view that it may represent a type of inflammatory bowel disease. They also
support the findings of previous observational studies (Hjern et al.).
Recruitment was non-uniform between centres, suggesting that many patients who
would have been suitable were not enrolled. This raises concerns over selection bias or
that sicker patients were not included. The authors address the possibility of selection bias
MANAGEMENT OF UNCOMPLICATED DIVERTICULITIS 91
by highlighting that the results were stratified by centre and that patients enrolled in the
study with findings of more severe disease did not have a worse outcome.
Another important finding of the study is the low risk of developing serious compli-
cations in this group of patients (1.4%). Further randomized trials are awaited to con-
firm the findings of the current study. However, it seems clear that antibiotics should be
reserved for patients with complicated diverticulitis, while patients with uncomplicated
disease can be managed without antibiotics and possibly on an outpatient basis.
3.13 The management of complicated diverticulitis

Details of studies
Traditional surgical dogma has dictated that perforated diverticular disease should
be treated by resecting the perforated segment with formation of an end colostomy
(Hartmann’s procedure). However, reversal of the stoma carries a not insignificant risk
of morbidity and even mortality. As a consequence, many patients and surgeons elect not
to consider reversal, leaving patients with a life-long stoma and its attendant reduction in
quality of life and risk of stoma complications. As a result, primary resection and anasto-
mosis has emerged as an alternative approach, although morbidity and mortality are still
significant, given that anastomotic leak rates of more than 10% have been reported in this
setting.
O’Sullivan et al. advocated a novel approach of laparoscopic washout. This approach
avoids a major laparotomy and formation of end-colostomy, but does not remove the per-
forated colon. An initial paper introduced the technique in a small series of eight patients
who all made a full recovery. Although other groups have adopted the concept of lapa-
roscopy with washout, some have subsequently performed an interval elective resection
(Taylor et al.; Faranda et al). O’Sullivan’s group subsequently reported the results of a
much larger prospective study of laparoscopic peritoneal lavage in perforated diverticuli-
tis, which is the main study reference in this section.
Study references
Main study
Myers, E., Hurley, M., O’Sullivan, G.C., Kavanagh, D., Wilson, I., and Winter, D.C. (2008) Laparoscopic
peritoneal lavage for generalized peritonitis due to perforated diverticulitis. British Journal of
Surgery, 95, 97–101.
Related references
Hinchey, E.J., Schaal, P.G., and Richards, G.K. (1978) Treatment of perforated diverticular disease of the
colon. Advances in Surgery, 12, 85–109.
Krukowski, Z.H. and Matheson, N.A. (1984) Emergency surgery for diverticular disease complicated by
generalized and faecal peritonitis: a review. British Journal of Surgery, 71, 921–7.
O’Sullivan, G.C., Murphy, D., O’Brien, M.G., and Ireland, A. (1996) Laparoscopic management of gen-
eralized peritonitis due to perforated colonic diverticula. American Journal of Surgery, 171, 432–4.
Taylor, C.J., Layani, L., Ghusn, M.A., and White, S.I. (2006) Perforated diverticulitis managed by laparo-
scopic lavage. ANZ Journal of Surgery, 76, 962–5.
Faranda, C., Barrat, C., Catheline, J.M., and Champault, G.G. (2000) Two-stage laparoscopic manage-
ment of generalized peritonitis due to perforated sigmoid diverticula: eighteen cases. Surgical
Laparoscopy, Endoscopy and Percutaneous Techniques, 10, 135–8.
MANAGEMENT OF COMPLICATED DIVERTICULITIS 93
Study design
◆ A prospective case series.
Level of evidence 4
Intervention Laparoscopic washout for complicated diverticulitis
Number of centres 1
Study duration 2000–2007
Inclusion criteria Patients admitted with generalized peritonitis on clinical assessment with
radiological evidence of perforation on CXR or CT
Follow-up (median) 36 months
◆ All recruited patients received intravenous antibiotics and fluid resuscitation

before undergoing laparoscopy. The peritoneal cavity was examined and the stage
of diverticulitis classified according to Hinchey et al. (1978). In patients who were
staged as Hinchey stage IV (faecal peritonitis) Hartmann’s procedure was performed.
All other stages underwent four-quadrant lavage.
◆ Antibiotics were given intravenously for 72 hours and then orally for a
further week.
Outcome measures
◆ Resolution of complicated diverticulitis.
◆ Time to diet, hospital stay, complications, mortality.
Results
◆ The median age of patients was 62.5 (range 39–94) years.
◆ The male-to-female ratio was 2:1 and median American Society of Anesthesiologists
(ASA) grade was III (range II–IV).
◆ Five of 100 patients had history of diverticulosis.
◆ 92 patients were managed with laparoscopic washout and 82 of these recovered
without morbidity (89% of the overall study population).
◆ The rate of post-operative medical (cardiorespiratory) complications
was 4%.
◆ Two patients developed pelvic abscesses. Both were initially drained radiologically;
one failed to resolve and underwent Hartmann’s procedure.
◆ Post-operative mortality was 3%.

◆ Recurrent diverticulitis developed in only two patients during the follow-up period.
Hinchey 2 (n = 25) Hinchey 3 (n = 67) Hinchey 4 (n = 8)

Surgical procedure Lavage only Lavage only Hartmann’s
Time to diet (days) 2 (1–6) 5 (2–19) 14 (6–22)
Time to discharge (days) 8 (7–18) 10 (7–22) 18 (10–44)
Conclusion
Laparoscopic washout is a reasonable alternative to the traditional open resection
for Hinchey grades 2 and 3 perforated diverticulitis with generalized peritonitis. With a
documented low mortality rate, this approach avoids formation of a colostomy.
Critique
This prospective study provided further evidence that laparoscopic lavage in selected
patients avoids resection and colostomy in both the short and long term after acute perfo-
rated diverticulitis. At the time of publication it was the largest prospective study analysing
this new approach to perforated diverticulitis with reasonable follow-up allowing further
insight into the debate on when to operate after an attack of acute diverticulitis. Aside
from being non-randomized, the main limitation of this study is that little is mentioned
on patient selection. Over a seven-year period, only 100 out of 1257 patients admitted
with diverticulitis were found to have generalized peritonitis. Although it is stated that all
patients who were approached consented, the study does not mention if these 100 patients
were consecutive or whether any patient was excluded.
Since this study was published, many more studies have gone on to suggest that laparo-
scopic washout is appropriate for Hinchey 2 and 3 diverticulitis, with a systematic review
from 2010 combining two prospective studies, nine retrospective case series, and two
case reports. That analysis found that lavage controlled abdominal and systemic sepsis in
95.7% of Hinchey 2/3 (Toorenvliet et al.). Results are currently awaited from a large Dutch
multi-centre randomized trial comparing lavage versus resection for purulent peritonitis
and Hartmann’s procedure versus resection with primary anastomosis for purulent or
faecal peritonitis in perforated diverticulitis (The Ladies trial).
Toorenvliet, B.R., Swank, H., Schoones, J.W., Hamming, J.F., and Bemelman, W.A. (2010) Laparoscopic
peritoneal lavage for perforated colonic diverticulitis: a systematic review. Colorectal Disease, 12,
862–7.
Swank, H.A., Vermeulen, J., Lange, J.F., et al. (2010) The Ladies trial: laparoscopic peritoneal lavage or
resection for purulent peritonitis and Hartmann’s procedure or resection with primary anastomosis
for purulent or faecal peritonitis in perforated diverticulitis (NTR2037). BMC Surgery, 10, 29.
EFFECT OF EN VERSUS PN ON POST-OPERATIVE SEPTIC COMPLICATIONS 95
3.14 Effect of enteral versus parenteral nutrition on

post-operative septic complications
Details of studies
This meta-analysis was an important milestone in the evidence base comparing methods
of nutritional support in high-risk surgical patients. Prior to publication the surgical lit-
erature contained few studies with relatively small sample sizes. This study by Moore et al.
compared total enteral nutrition (TEN) with total parenteral nutrition (TPN) and the
effect on septic complications.
Study references
Main study
Moore, F.A., Feliciano, D.V., Andrassy, R.J., et al. (1992) Early enteral feeding, compared with parenteral,
reduces postoperative septic complications. Annals of Surgery, 216, 172–83.
Related reference
Al-Omran, M., Albalawi, Z.H., Tashkandi, M.F., and Al-Ansary, L.A. (2010) Enteral versus parenteral
nutrition for acute pancreatitis. Cochrane Database of Systematic Reviews, Issue 1, CD002837.
Study design
◆ Meta-analysis of eight prospective RCTs.
Randomization TEN (jejunostomy/nasoenteral tube) versus TPN
Number of centres Eight North American centres (four enrolled only trauma patients)
Inclusion criteria Moderately to severely stressed patients
TEN/TPN started <72 hours after operation
Vivonex TEN and nutritionally similar TPN
Prospective recording of complications
Exclusion criteria Pre-existing diseases

Conditions precluding use of TEN/TPN
Severe head injury
Nutritional support/operation or prolonged stay during pre-operative
hospitalization
Non-study nutritional solution
◆ Dropouts were defined by early withdrawal (<72 hours) for non-diet-related issues
(e.g. re-laparotomy, transfer to another hospital, death, etc.).
◆ Treatment failures included access complications, gastrointestinal (GI) intolerance,
metabolic problems.
◆ Phase I was a PP analysis.
◆ Phase II was an ITT analysis comprising five subgroups: all patients, all trauma
patients, penetrating trauma patients, blunt trauma patients, and non-trauma
patients.
◆ The non-trauma group contained a diverse group of cardiac, oesophageal, gastric,
pancreatic, colonic, and vascular procedures
Outcome measures
Primary endpoints
◆ Non-septic and septic complications.
Secondary endpoints
◆ Length of hospital stay, days in the intensive care unit, mortality, time to regular oral
diet, GI intolerance, hospitalization cost.
Results
Phase I (PP analysis)
◆ 194 patients (92 TEN, 102 TPN) with similar baseline demographics (age, sex, race,
injury/surgery type).
TEN TPN p-value

Dropouts 26 10 0.001
Treatment failures 16% (GI intolerance) 3% (metabolic) 0.001
Time to start support 32.5 hours* 32.8 hours* ns
Time to regular diet 11 days 11 days ns
GI intolerance
Abdominal distention 46% 24% 0.003
Diarrhoea 34% 9% 0.001
Septic complications 17% 44% 0.0001
Total complications 38% 59% 0.007
*Mean values
Phase II (ITT analysis)

◆ 230 patients (118 TEN, 112 TPN).
◆ Similar demographics, even when split into the five subgroups.
◆ Fewer septic complications in TEN group (16 versus 35%, p = 0.001) even after
excluding catheter-related sepsis (16% versus 29%, p = 0.03).
◆ Fewer total complications in TEN group, although not significant (41 versus 52%,
p = 0.09).
EFFECT OF EN VERSUS PN ON POST-OPERATIVE SEPTIC COMPLICATIONS 97
◆ No notable difference in septic complications was seen in the non-trauma subgroup

(numbers were very small).
◆ No difference in mortality, or hospital stay.
◆ Hospital costs were too heterogeneous to compare.
Critique
It is well recognized that the stress response to major surgery or injury leads to a
negative nitrogen balance and a catabolic state with the potential for delayed wound-
healing and suppressed immune functions. This study was an important part of
growing evidence that early nutritional support is beneficial in high-risk surgical
patients. Prior to this study the optimal route of delivery for nutritional support was
unclear, and there was a reluctance to use the GI tract in the early post-operative/post-
injury period.
The evidence for early nutritional support suffered from a lack of power because of
the small number of patients who could be randomized at single institutions. This meta-
analysis was one way of addressing the problem by combining trials from multiple centres
with similar methodology. The authors suggest that the homogeneous nature of the trials
included is one of the main strengths of this study.
This meta-analysis shows that post-operative septic complications are reduced with
early enteral nutrition compared with TPN. Sensitivity analysis showing consistency of
results between the phase I and II analyses further supports these findings. The main criti-
cism of this study is with regard to the generalizability of the findings. The findings appear
to be valid for trauma patients who made up the majority of the participants. However, it
is argued that trauma patients are usually younger than the typical surgical patient requir-
ing nutritional support, which may affect the results. No difference in complications was
observed in the non-trauma/elective surgical group. This may represent a type II error,
as numbers in this group were small and a diverse range of procedures was included.
Evidence of the superiority of enteral nutrition in severe pancreatitis would also suggest
that the beneficial effect is wider than for trauma patients alone.
While this study lends support to the view that enteral nutrition reduces infective com-
plications when compared with TPN, the reason for the difference remains a matter of
debate. There is a view that in severely stressed patients the gut barrier function is lost,
leading to increased bacterial translocation—the so-called ‘gut origin of sepsis’. It is sug-
gested that enteral nutrition may reduce villous atrophy and better preserve this barrier
function. Others argue that the likely mechanism is related to overfeeding and hypergly-
caemia in the TPN group, leading to increased complications. Certainly direct complica-
tions from the mode of delivery are an important consideration in both groups and are
regularly highlighted in the literature.
Conclusions
Early nutrition reduces post-operative septic complications and where possible the
GI tract should be preferred as the route of delivery.
Kalfarentzos, F., Kehagias, J., Mead, N., Kokkinis, K., Gogos, C.A. (1997) Enteral nutrition is superior
to parenteral nutrition in severe acute pancreatitis: results of a randomized prospective trial. British
Braga, M., Gianotti, L., Gentilini, O., Liotta, S., and Di Carlo, V. (2002) Feeding the gut early after
digestive surgery: results of a nine-year experience. Clinical Nutrition, 21, 59–65.
Veterans Affairs Total Parenteral Nutrition Cooperative Study Group (1991) Perioperative total
parenteral nutrition in surgical patients. New England Journal of Medicine, 325, 525–32.
GOAL-DIRECTED THERAPY IN SEVERE SEPSIS 99
3.15 Goal-directed therapy in severe sepsis

Details of studies
Sepsis and related syndromes are the principal cause of multi-organ failure and death
in critically ill patients. Infection provokes an inflammatory cascade which can result in
vasodilatation, hypotension, and tissue hypoxia. Goal-directed therapy has been used in
the management of septic patients in an intensive care setting with equivocal results. This
study evaluated the effects of early goal-directed therapy to balance tissue oxygen demand
and oxygen delivery in sepsis prior to admission to intensive care.
Study references
Main study
Rivers, E., Nguyen, B., Havstad, S., et al. (2001) Early goal-directed therapy in the treatment of severe
sepsis and septic shock. New England Journal of Medicine, 345, 1368–77.
Related references
Shoemaker, W.C., Appel, P.L., Kram, H.B., Waxman, K., and Lee, T.S. (1988) Prospective trial of supra-
normal values of survivors as therapeutic goals in high-risk surgical patients. Chest, 94, 1176–86.
Gattinoni, L., Brazzi, L., Pelosi, P., et al. (1995) A trial of goal-oriented hemodynamic therapy in criti-
cally ill patients. New England Journal of Medicine, 333, 1025–32.
Study design
Randomization ‘Goal-directed therapy’ versus standard care
Setting Academic tertiary care hospital in Detroit
Inclusion criteria 2/4 criteria for systemic inflammatory response syndrome (SIRS)
Systolic BP <90mmHg (after crystalloid fluid challenge—20–30ml/kg over
30min) or lactate >4mmol/L
Exclusion criteria Age <18 years or pregnancy
Acute cardiorespiratory/cerebral event or trauma
Active haemorrhage or requirement for immediate surgery
Contraindication to CVP line
Follow-up 60 days
◆ The goal-directed group received fluid resuscitation, vasopressors/dilators, red cell

transfusions, and inotropes in sequential fashion to achieve central venous pressure
(CVP) of 8–12mmHg, mean arterial pressure of 65–90mmHg, central venous oxygen
saturation >70%, and haematocrit >30%.
◆ The standard group was treated at the clinician’s discretion according to the
department protocol for haemodynamic support
◆ All patients had arterial and central lines inserted and were managed in the
emergency department (ED) resuscitation room by ED physicians
◆ Patients in the goal-directed group were managed in the ED for at least the first 6
hours prior to critical care transfer. Patients in the standard group were transferred to
critical care when a bed was available.
Outcome measures
Primary endpoint
◆ In-hospital mortality.
Secondary endpoints
◆ Resuscitation endpoints, organ dysfunction scores, coagulation-related variables,
administered treatment, and consumption of health care resources.
Results
◆ 27 patients did not complete the initial six-hour study period.
◆ No difference between the groups at baseline with respect to characteristics,
adequacy of antibiotics, organ dysfunction scores, or other secondary endpoints
◆ Patients in the standard group stayed in the ED for a significantly shorter period of time
Goal-directed Standard p-value

Time in ED 8 hours 6.3 hours <0.001
Haemodynamic goals achieved <6 hours 99.2% 86% <0.001
Mean CVSO2 (7–72 hours) 70.4% 65.3% ≤0.02
Lactate (7–72 hours) 3.0mmol/L 3.9 mmol/L ≤0.02
Base deficit (7–72 hours) 2.0 mmol/L 5.1 mmol/L ≤0.02
APACHE II score (7–72 hours) 13 15.9 <0.001
Length of stay (discharged patients) 18.4 days 14.6 days 0.04
In-hospital mortality 30.5% 46.5% 0.009
CVSO2, central venous oxygen saturation
◆ The rate of in-hospital death due to sudden cardiovascular collapse was significantly
higher in the standard therapy group than in the early therapy group (p = 0.02). The
rate of death due to multi-organ failure was similar in the two groups (p = 0.27).
◆ Patients in the goal-directed therapy group received more fluid during the first 6
hours (p <0.001) but less fluid in the next 7–72 hours of standard therapy (p = 0.01).
During the overall period (baseline to 72 hours) there was no significant difference
between the two groups in total volume of fluid administered (p = 0.73).
GOAL-DIRECTED THERAPY IN SEVERE SEPSIS 101
Conclusions
Goal-directed therapy provided during the early stages of severe sepsis and septic
shock has significant short- and long-term benefits. These benefits are realized by early
identification of patients at high risk for cardiovascular collapse and early therapeutic
intervention to restore the balance between oxygen delivery and oxygen demand.
Critique
Severe sepsis and septic shock are common presentations to hospital, with a high mor-
bidity and mortality rate. Most outcome studies that have investigated septic shock and
severe sepsis have only looked at interventions once patients have been admitted into
intensive care. Rivers et al. suggested that there was a golden period, analogous to the
golden hour in trauma, prior to admission to intensive care. The aim of this study was to
investigate if goal-directed management in this period would improve outcome.
Because of the nature of the study it was not possible to perform a double-blind trial.
Therefore one of the potential weaknesses is that it was only partially blinded. In the ini-
tial period of the trial it was recognized that patients in the standard group received some
goal-directed therapy because of the direct influence of the investigators. The authors
address this in their conclusions and suggest that it affected only a small interval of the
patients’ stay with a non-significant effect. This was because after 6 hours the physicians
managing the patients were blinded to the initial treatment group. However, it is difficult
to be sure that patients in the goal-directed group did not receive closer attention in the
early stages.
Another criticism of the study was the mortality rate in the standard group. Mortality
in this group was felt to be high in comparison with other trials in the literature with
similar APACHE scores. The method of achieving some of the physiological goals, in
particular red cell transfusion, was also a matter of debate.
Despite these criticisms Rivers’ study does seem to show that early goal-directed ther-
apy improves outcome in patients with sepsis. The beneficial effect of this approach would
appear to be predominantly through early aggressive fluid resuscitation. These findings
would be in keeping with subsequent goal-directed fluid trials in elective surgery.
Gan, T.J., Soppitt, A., Maroof, M., et al. (2002) Goal-directed intraoperative fluid administration reduces
length of hospital stay after major surgery. Anesthesiology, 97, 820–6.
Noblett, S.E., Snowden, C.P., Shenton, B.K., and Horgan, A.F. (2006) Randomized clinical trial assessing
the effect of Doppler-optimized fluid management on outcome after elective colorectal resection.
3.16 Confidential enquiry into peri-operative deaths (CEPOD)

Details of studies
This large-scale audit was unique for being the first audit of peri-operative mortality to
include both anaesthetists and surgeons. The study was a collaboration between the Royal
College of Anaesthetists and the Association of Surgeons of Great Britain and Ireland
with independent funding from the Nuffield Provincial Hospitals Trust and the King’s
Fund for London. The aim was to assess the quality of surgery through an audit of mortal-
ity with the hope of influencing clinical practice in the future.
Study references
Main study
Buck, N., Devlin, H.B., and Lunn, J.N, (1987) The Report of the Confidential Enquiry into Perioperative
Deaths. London: Nuffield Provincial Hospitals Trust, King’s Fund.
Related reference
Lunn, J.N. and Devlin, H.B. (1987) Lessons from the confidential enquiry into perioperative deaths in
three NHS regions. Lancet, 330, 1384–6.
Study design
◆ Large-scale audit.
Level of evidence 2c
Setting Three regional health authorities in England
Duration 12 months (1986)
Inclusion criteria All deaths within 30 days of an operative procedure carried out by an
anaesthetist and/or a surgeon
◆ Questionnaires for each death were completed by the consultant surgeon and
anaesthetist involved.
◆ After return to the CEPOD office, forms were anonymized for patient, clinician, and
hospital.
◆ Trained assessors (surgeons and anaesthetists) scored each questionnaire, assessing
the quality of care, desirability of operation, adequacy of monitoring, and avoidability
of death.
◆ If surgery or anaesthesia was felt to be partly/wholly responsible, the contributions of
surgical condition, comorbidity, and professional deficiencies were assessed
◆ The data were protected by Crown privilege, and following preparation of the report
the original data were destroyed, protecting against medico-legal action
CONFIDENTIAL ENQUIRY INTO PERI-OPERATIVE DEATHS (CEPOD) 103
Results
◆ 4034 deaths were identified in 555,258 operations—a mortality rate of 0.7%.
◆ 78 (11 anaesthetists and 68 surgeons) of 1396 consultants declined to take part—a
participation rate of 94%.
◆ Data for analysis were available for 70% of deaths with data from both surgeon and
anaesthetist available for 59%.
◆ 79% of CEPOD patients were aged over 65 years.
◆ The presenting surgical condition contributed to death in two-thirds of patients and
was the sole cause in one-third.
◆ Medical comorbidity contributed to 50% of deaths.
◆ 75% of deaths occurred within the first 15 days.
◆ CEPOD deaths accounted for 3% of deaths from all causes in the regions involved.
◆ Surgical errors were felt to be the sole cause of 0.04% of deaths.
◆ Anaesthesia was the cause of death in 1 in 185,000 procedures (0.0005%).
◆ Avoidable anaesthetic and surgical factors were felt to contribute to up to 20% of
deaths.
◆ 30–50% of deaths had no consultant contribution to their clinical care.
◆ Consultants were involved in less than 25% of operations at weekends and
overnight.
◆ Almost 50% of institutions did not undertake regular audit or meetings to review deaths.
◆ Only 5% of clinicians took up the invitation to receive feedback.
Conclusions
◆ Post-operative mortality was low at 0.7% and was largely unavoidable because of the
presenting pathology, comorbidity, and advanced age.
◆ Participation was encouragingly high at almost 95%, although there were significant
difficulties with the quality of record-keeping and availability of notes.
◆ There was inadequate supervision of trainees by consultants, and trainees often failed
to involve senior help before, during, or after surgery.
◆ Unnecessary surgery was performed on patients with no hope of recovery. The
decision to operate should be taken at consultant or senior registrar level for all
urgent or emergency cases.
◆ Time to assess medical conditions and for adequate resuscitation was recommended
to avoid rushing under-resuscitated patients to theatre.
◆ General surgeons were found to be operating inappropriately because of a reluctance

to refer to more specialist colleagues.
◆ Many units did not hold regular morbidity and mortality meetings, and combined
meetings with anaesthetics and surgery were very rare
Critique
This ambitious peer audit was an important step forward in understanding the standard
of surgery in the UK, as well as highlighting areas for improvement. The report did not
pull its punches and met with a mixed reception. It was independently published, with
some suggesting that the methodology would not have been accepted for publication in a
scientific journal and that it was politically naive.
Despite these criticisms this first report led to the introduction of the National
Confidential Enquiry into Patient Outcome and Deaths (NCEPOD), which is govern-
ment funded and has reported annually since 1990. During this time the remit has
widened to include all specialties and analyses near-misses as well as deaths. During its
23 years, NCEPOD has become an integral part of the surgical vocabulary and in particu-
lar the NCEPOD classification of the urgency of operation. It has also been consistent in
its findings of incomplete record keeping, inadequate supervision, a need for increased
involvement of senior clinicians, surgeons operating outside their specialist field,
procedures being unnecessarily carried out during the night, poor provision of emer-
gency services, and limited access to critical care facilities. In addition to these specific
areas of clinical practice the report also demonstrated the critical importance of audit to
ongoing standards of care.
Morgan, M. (1988) A confidential enquiry into peri-operative deaths. Anaesthesia, 43, 91–2.
National Confidential Enquiry into Patient Outcome and Deaths <http://www.ncepod.org.uk>
Chapter 4
Oesophagogastric surgery
4.1 Surgery for achalasia 106
4.2 Health benefits from bariatric surgery 109
4.3 Risk scoring for upper gastrointestinal haemorrhage 112
4.4 Single versus dual endotherapy for peptic ulcer bleeding 115
4.5 Proton pump inhibitor treatment following endoscopic therapy for bleeding
peptic ulcer 117
4.6 Laparoscopic fundoplication versus medical therapy for the treatment of
gastro-oesophageal reflux 120
4.7 Technical aspects of laparoscopic anti-reflux surgery and para-oesophageal
hernia repair 123
4.8 Incidence of adenocarcinoma in patients with Barrett’s oesophagus 127
4.9 Relationship of hospital volume to outcome in oesophagectomy 129
4.10 Trans-thoracic versus trans-hiatal oesophagectomy 132
4.11 The role of lymphadenectomy in gastric cancer surgery 136
4.12 Minimally invasive surgery for oesophagogastric cancer 140
4.13 Multimodality treatment for operable gastric cancer 144
4.14 Pre-operative chemotherapy or chemoradiotherapy for locally advanced
oesophageal cancer 148
4.15 Palliation of dysphagia in advanced oesophageal cancer 152
106 OESOPHAGOGASTRIC SURGERY
4.1 Surgery for achalasia

Details of studies
Achalasia is a rare disorder of oesophageal motility characterized by absent oesopha-
geal peristalsis and incomplete relaxation of the lower oesophageal sphincter. Treatment
modalities include endoscopic botulinum toxin injection, pneumatic dilation, and surgi-
cal myotomy by either an open or a laparoscopic approach. There are few high-quality
randomized trials to guide management. The best available evidence comes from meta-
analyses and one recent randomized trial.
Study references
Randomized trial
Boeckxstaens, G.E., Annese, V., des Varannes, S.B. et al. (2011) Pneumatic dilation versus laparoscopic
Heller’s myotomy for idiopathic achalasia. New England Journal of Medicine, 364, 1807–16.
Meta-analyses
Campos, G.M., Vittinghoff, E., Rabl, C., et al. (2009) Endoscopic and surgical treatments for achalasia: a
systematic review and meta-analysis. Annals of Surgery, 249, 45–57.
Wang, L., Li, Y.M., and Li, L. (2009) Meta-analysis of randomized and controlled treatment trials for
achalasia. Digestive Diseases and Sciences, 54, 2303–11.
Randomized trial
Study design
Randomization Pneumatic dilation versus laparoscopic Heller’s myotomy (LHM) + Dor’s
fundoplication for achalasia
Number of patients 201 patients
Setting 14 hospitals in five European countries (2003–2008)
Primary outcome Therapeutic success (drop in Eckardt score ≤3)
Secondary outcomes Need for re-treatment, lower oesophageal sphincter pressure, oesophageal
emptying, quality of life, complications
Follow-up (mean) 43 months
◆ Inclusion criteria: patients aged 18–75 years with achalasia and Eckardt score >3.
◆ Exclusion criteria: unfit for surgery, previous treatment for achalasia or oesophageal
or gastric surgery, pseudoachalasia, mega-oesophagus.
SURGERY FOR ACHALASIA 107
Results
Pneumatic dilation LHM (n = 106) p-value
(n = 95)
Therapeutic success
Year 1 90% 86%
Year 2 93% 90% 0.46
LOS pressure 12mmHg 10mmHg 0.27
Quality of life (QLQ-OES24) 15 13 0.28
◆ There was no difference in oesophageal emptying, exposure to gastric acid, or

oesophagitis between the two groups
Meta-analyses
Study design
Campos et al. (2009)
◆ Treatments reviewed included endoscopic botulinum toxin injection (EBTI), endoscopic
balloon dilation (EBD), and surgical myotomy by an open trans-thoracic or trans-
abdominal approach or a laparoscopic approach with or without an anti-reflux procedure.
◆ Main outcomes were symptom relief, gastro-oesophageal reflux, and complications.
Wang et al. (2009)

◆ Strict inclusion criteria meant that only randomized trials were included.
◆ Evaluation of clinical efficacy included remission rate, relapse rate, complications,
and adverse effects 1 year after therapy
◆ Clinical remission was defined as >50% improvement in total symptom score, and
failure or relapse was defined as <50% improvement in total symptom score
Results
Campos et al.
◆ 105 articles, reporting 7855 patients including 98 cohort studies and seven
randomized trials.
◆ EBD was associated with greater symptom improvement at 12 months compared
with EBTI (68.2% versus 40.6%).
◆ Laparoscopic myotomy demonstrated better symptom improvement at 12 and >36
months when compared with EBD (89% and 89% versus 68% and 56%, p<0.01).
◆ 0.7% of patients developed clinical manifestation of a leak after laparoscopic
myotomy. Endoscopic dilation carried a perforation rate of 1.6%.
Wang et al.
◆ Seventeen articles analysing 761 patients were included.
◆ There were significant differences in remission rate between laparoscopic myotomy

and pneumatic dilation (95% versus 77.8%, p = 0.001)
◆ The relapse rate for laparoscopic myotomy was 5.1% versus 35.7% for pneumatic
dilation (p = 0.007).
Conclusions
Boeckxstaens et al.
Laparoscopic myotomy and pneumatic dilation are both effective treatments for
achalasia.
Campos et al.
Laparoscopic myotomy is the most effective and long-lasting treatment for achalasia.
Wang et al.
Laparoscopic myotomy offers better clinical results than pneumatic dilation.
Critique
The relative rarity of the condition of achalasia has resulted in few high-quality com-
parative studies with adequate numbers of patients. Heterogeneity within retrospective
studies and publication bias all affect the interpretation. Furthermore, when addressing
certain outcome criteria such as presence of reflux after treatment there is an over-reliance
on symptomatology rather than objective analysis with pH studies.
Although these two large meta-analyses have demonstrated that surgery confers the
greatest symptom control compared with endoscopic therapies, there is a relative lack
of data from randomized trials. Indeed, in the analysis performed by Wang et al., only
two small randomized trials comparing surgery with balloon dilation were identified.
Nevertheless, surgery has been shown to be an effective, long-lasting, and safe treatment
for achalasia.
A subsequent large European multi-centre randomized trial, comparing laparoscopic
myotomy and pneumatic dilatation, has recently been reported (Boeckxstaens et al. 2011).
This study demonstrated equivalence between the two groups with regard to symptom
control and safety. One criticism of the study is that the intra-operative surgical perfo-
ration rate was relatively high (12%). Furthermore, the balloon dilatation protocol was
altered early in the study as four of the first 13 patients were perforated at endoscopy. This
group of patients was not included in the final analysis.
Success rates reported in the literature differ depending on the criteria used to define
success. This applies in particular when success is defined as the absence of need for
subsequent interventions. If this criterion is used, the success rate for balloon dilation
is significantly lower than for surgical myotomy. The rate of repeat balloon dilatation
in Boeckxstaens’ randomized study was highest when daily pre-existing chest pain was
reported and in those patients <40 years of age with a non-dilated oesophagus. This sup-
ports the view that surgical myotomy may be most useful in the younger fit patient.
HEALTH BENEFITS FROM BARIATRIC SURGERY 109
4.2 Health benefits from bariatric surgery

Details of studies
The Swedish Obese Subjects (SOS) study is generally regarded as the most important study
to have addressed the issue of the medical health benefits of obesity surgery. Sjöström et al.
(2004) assessed the impact of obesity surgery on diabetes and cardiac risk factors over a
minimum of 2-years follow-up. Subsequently in 2007 this group also analysed mortality
in the same cohort of 4047 subjects followed up for an average of 10.9 years.
Study references
Sjöström, L., Lindroos, A.K., Peltonen, M., et al. (2004) Lifestyle, diabetes, and cardiovascular risk fac-
tors 10 years after bariatric surgery. New England Journal of Medicine, 351, 2683–93.
Sjöström, L., Narbro, K., Sjöström, C.D., et al. (2007) Effects of bariatric surgery on mortality in Swedish
Obese Subjects. New England Journal of Medicine, 357, 741–52.
Study design
Sjöström et al. (2004)
◆ A prospective non-randomized intervention trial.
Inclusion criteria Surgically treated obese patients were compared with matched
conventionally treated patients
Primary outcome Overall mortality
Secondary outcomes Changes in body weight, risk factors, energy intake, and physical activity
Difference in incidence of risk conditions among those unaffected at
baseline (i.e. preventive effect of weight loss)
Difference in rate of recovery from risk conditions among those affected at
baseline
Follow-up Minimum 2 years

◆ Ten-year mortality was reported for the same cohort of patients.
Results
◆ 851 surgically treated patients and 852 matched controls had been enrolled in the
study for 10 years at the time of analysis.
Obesity surgery Obese controls p–value

Weight change
2 years –23.4% + 0.1% <0.001
10 years –16.1% + 1.6% <0.001
Obesity surgery Obese controls p–value

Incidence of diabetes
2 years 1% 8% <0.001
10 years 7% 24% <0.001
Incidence of hypertension
2 years 24% 29% 0.06
10 years 41% 49% 0.13
Recovery from diabetes
2 years 72% 21% <0.001
10 years 36% 13% 0.001
Recovery from hypertension
2 years 34% 21% <0.001
10 years 19% 11% 0.02
◆ In the surgical group the post-operative mortality rate was 0.25%, and 13%
experienced complications.
◆ The mean energy intake was lower in the surgical group while the proportion of
physically active patients was higher compared with the matched controls
◆ Incidence and recovery from hypercholesterolaemia was not significantly different
between the groups at 2 and 10 years.
◆ Gastric bypass led to greater weight loss at 2 and 10 years compared with gastric
banding and vertical banded gastroplasty

◆ 129 (6.3%) patients in the control group died compared with 101 (5%) patients
treated surgically after an average of 10.9 years of follow-up.
◆ The unadjusted overall hazard ratio was 0.76 in the surgery group (p = 0.04)
compared with the control group.
◆ When adjusted for age, sex, and risk factors the hazard ratio was 0.71 (p <0.001).
◆ The most common causes of death were myocardial infarction and cancer.
Critique
Despite the large number of enrolled subjects and the long length of follow-up involved
in these studies, the limitation of their non-randomized nature remains. It is a widely
held view that large randomized studies of surgery versus control obese populations will
never be performed and these studies represent the best alternative likely to be available.
Furthermore, statistical power does not exist for subgroup analysis of the influence of
variables such as extent of weight loss, initial body mass index (BMI), age, or surgical
procedure on survival. Whilst older subjects demonstrated the greatest improvement in
survival, this may simply reflect the need for longer follow-up in younger subjects before
HEALTH BENEFITS FROM BARIATRIC SURGERY 111
similar differences are observed. In studies that look at the negative effect of obesity on
mortality, obesity becomes a significant predictor of mortality only after 26 years. Thus it
is not surprising that it took many years until a favourable treatment effect on mortality
was observed in the second study. Slightly unusual entry criteria were selected for these
studies, with subjects required to be over 37 years of age with a BMI >34 for men and >38
for women. This reflects the fact that the study predates consensus guidelines for bariatric
surgery.
In conjunction with the rest of the surgical literature, these studies show that there is
good evidence that bariatric surgery has a beneficial effect on diabetes, sleep apnoea, joint
pain and mobility, health-related quality of life (HRQoL), and survival.
4.3 Risk scoring for upper gastrointestinal haemorrhage

Details of studies
Upper gastrointestinal bleeding is a common emergency presentation and accounts for
50–172 admissions per 100,000 adults annually in the UK. Bleeding can range from minor
coffee-ground vomiting to exsanguination, with mortality from upper GI bleeding rang-
ing from 4% to 14%. A number of scores have been developed in an attempt to predict
which patients are ‘low risk’ and ‘high risk’ for adverse outcome or need for intervention.
The two most commonly used scores are the Rockall score and the Glasgow–Blatchford
score (GBS). The Rockall score uses clinical and laboratory criteria in addition to endo-
scopic findings (full Rockall), while an abbreviated score can be calculated at admission
(pre-endoscopy Rockall). The GBS uses clinical and laboratory criteria only and can be
calculated at initial presentation.
Study references
Main study
Blatchford, O., Murray, W.R., and Blatchford, M. (2000) A risk score to predict need for treatment for
upper gastrointestinal haemorrhage. Lancet, 356, 1318–21.
Related references
Rockall, T.A., Logan, R.F.A, Devlin, H.B., and Northfield T.C. (1996) Risk assessment after acute upper
gastrointestinal haemorrhage. Gut 1996; 38, 316–21.
Stanley, A.J., Ashley, D., Dalton H.R., et al. (2009) Outpatient management of patients with low-risk upper
gastrointestinal haemorrhage: multicentre validation and prospective evaluation. Lancet, 373, 42–7.
Chen, I.C., Hung, M.S., Chiu, T.F., Chen, J.C., and Hsiao, C.T. (2007) Risk scoring systems to predict
need for clinical intervention for patients with nonvariceal upper gastrointestinal tract bleeding.
American Journal of Emergency Medicine, 25, 774–9.
Study design
Phase 1
◆ Retrospective audit of 1748 patients admitted with upper GI bleeding.
◆ Setting: 19 hospitals in the west of Scotland.
◆ Score derived by logistic regression.
◆ Score is based on the need for treatment as the response is variable.
◆ Need for intervention included blood transfusion, endoscopy, surgery, or death.
Phase 2
◆ Prospective validation of the score.
◆ 197 consecutive patients over a three-month period.
◆ Setting: three hospitals in the west of Scotland
RISK SCORING FOR UPPER GI HAEMORRHAGE 113
Results
Full risk score
Admission risk marker Value Score component value
Blood urea (mmol/L) 6.5–7.9 2
8–9.9 3
10–25 4
≥25 6
Haemoglobin (g/L)(men) 120–129 1
100–119 3
<100 6
Haemoglobin (g/L)(women) 100–119 1
<100 6
Systolic blood pressure (mmHg) 100–109 1
90–99 2
<90 3
Other markers Pulse ≥100bpm 1
Malaena 1
Syncope 2
Hepatic disease 2
Cardiac failure 2
Screening tool
Blood urea <6.5mmol/L
Haemoglobin >130g/L (men) or >120g/L (women)
Systolic blood pressure >100mmHg
Pulse <100bpm
Absence of syncope, melaena, hepatic disease or cardiac failure
◆ The sensitivity of the GBS (0.92 [95% CI 0.88–0.95]) was higher than the pre-
endoscopy Rockall score (0.71 [0.64–0.78]) and the full Rockall score (0.75 [0.67–
0.83]) at predicting need for intervention.
◆ The abbreviated screening tool identified 99% of patients with serious bleeds
requiring intervention.
Conclusion
The Glasgow–Blatchford score is a sensitive tool for predicting need for intervention in
upper gastrointestinal bleeding.
Critique
The majority of patients who present as an emergency with upper gastrointestinal bleed-
ing will resolve spontaneously without the need for blood transfusion, endoscopic haemo-
stasis, or surgery. Previously, a large proportion of patients presenting with this problem
would be admitted to hospital for a period of observation and possibly urgent endoscopy.
The Glasgow–Blatchford score presented here has been shown to have internal validity
and is suggested to be more sensitive than the Rockall score. The advantage of this score is
the ability to calculate it on initial presentation without the need for endoscopy and that
only objective criteria are used.
Further studies have confirmed the external validity of the score in different popula-
tions and conditions including Taiwan, Japan, and Canada. One of the first of these stud-
ies, by Stanley et al., looked at 676 patients in four separate centres in the UK. In the first
part of the trial they found that the GBS was superior to the Rockall score at predicting
either need for intervention or death. They also found that no patients with a GBS of zero
required intervention. Chen et al. and Srirajaskanthan et al. have confirmed the sensitivity
of the score in further studies.
In the trial by Stanley et al. the investigators went on to introduce a policy of outpatient
management for patients in a ‘low-risk’ category (GBS 0). Twenty-two per cent of patients
were deemed ‘low-risk’, and 68% of these avoided admission without any adverse events
recorded. The authors suggest that such a policy could save 60,000–100,000 bed-days per
year across the UK. In a similar study by Stephens et al., GBS <2 and age <70 years was
shown to safely select patients for management in the community.
Blatchford, O., Davidson, L.A., Murray, W.R., Blatchford, M., and Pell, J. (1997) Acute upper gastroin-
testinal haemorrhage in west of Scotland: a case ascertainment study. British Medical Journal, 1997;
315, 510–14.
Stephens, J.R., Hare, N.C., Warshow, U., et al. (2009) Management of minor upper gastrointesti-
nal haemorrhage in the community using the Glasgow Blatchford Score. European Journal of
Gastroenterology and Hepatology, 21, 1340–6.
Srirajaskanthan, R., Conn, R., Bulwer, C., and Irving, P. (2010) The Glasgow Blatchford scoring system
enables accurate risk stratification of patients with upper gastrointestinal haemorrhage. International
Journal of Clinical Practice, 64, 868–74.
Masaoka, T., Suzuki, H., Hori, S., Aikawa, N., and Hibi, T. (2007) Blatchford scoring system is a useful
scoring system for detecting patients with upper gastrointestinal bleeding who do not need endo-
scopic intervention. Journal of Gastroenterology and Hepatology, 22, 1404–8.
SINGLE VERSUS DUAL ENDOTHERAPY FOR PEPTIC ULCER BLEEDING 115
4.4 Single versus dual endotherapy for peptic ulcer bleeding

Details of studies
Endoscopic therapy is known to reduce the rate of recurrent peptic ulcer bleeding and
the need for surgery. Injection of adrenaline is the most common form of endotherapy
used. The Cochrane review by Vergara et al. (2007, updated 2010) addresses the issue of
the need for a second haemostatic procedure immediately after adrenaline injection and
during the index endoscopy in patients with peptic ulcer bleeding.
Study references
Main study
Vergara, M., Calvet, X., and Gisbert, J.P. (2007) Epinephrine injection versus epinephrine injection and
a second endoscopic method in high-risk bleeding ulcers. Cochrane Database of Systematic Reviews,
(2), CD005584.
Related reference
Cook, D.J., Guyatt, G.H., Salena, B.J., and Laine, L.A. (1992) Endoscopic therapy for acute nonvariceal
upper gastrointestinal haemorrhage: a meta-analysis. Gastroenterology, 102, 139–48.
Study design
Number of studies 18 RCTs
Randomization Adrenaline injection versus adrenaline injection + a second endoscopic
method
Inclusion criteria Patients with high-risk bleeding peptic ulcers
Primary outcome Recurrent bleeding
Secondary outcomes Surgery, complications, mortality
Results
Adrenaline only Combined therapy p-value
Re-bleeding 18.5% 10% <0.001
Surgery 10.7% 6.7% 0.021
Complications 0.9% 1.1% 0.77
Mortality 4.7% 2.5% 0.077
◆ Subgroup analysis showed that the risk of re-bleeding was reduced, regardless of
which second modality was used
Conclusions
Endoscopic haemostasis using injection of adrenaline and a second therapeutic agent
reduces the risk of re-bleeding and surgery compared with adrenaline alone.
Critique
The quality of the 18 randomized studies included in this meta-analysis was considered
high and the risk of bias was low. Outcomes did not depend on subjective observations
and the homogeneity between studies was high. As re-bleeding from peptic ulcers is the
main determinant of mortality from peptic ulcer bleeding, it is reasonable to conclude
that combination endoscopic therapy will be superior to injection of adrenaline alone in
high-risk patients.
The main fear about the use of combined endoscopic treatments is the risk of perfora-
tion. This meta-analysis clearly demonstrated that although there may be a slight increase
in risk of perforation, this is more than offset by the benefits of further reducing the rate of
re-bleeding. Therefore fear of perforation is not a reason for avoiding combined therapy.
Comparison of these studies reveals quite considerable variation in the definition of the
primary and secondary endpoints described. Therefore this analysis does not allow use
of a homogenous predetermined criterion of success for treatments. However, because
of the high quality of the individual studies, comparison between treatments is possible.
Similarly, it is not possible to perform subgroup analysis to determine which dose or con-
centration of adrenaline to recommend or which form of combination therapy confers
the greatest benefit.
The authors conclude that combined therapy should be considered as the standard
procedure in high-risk peptic ulcer haemorrhage (Forrest 1a, 1b, 11a). They also noted
a trend towards a decrease in mortality in the dual-therapy group, although this did not
reach statistical significance.
PPI TREATMENT FOLLOWING ENDOSCOPIC THERAPY FOR BLEEDING PEPTIC ULCER 117
4.5 Proton pump inhibitor treatment following endoscopic

therapy for bleeding peptic ulcer
Details of studies
Following endoscopic haemostasis for peptic ulcer bleeding, re-bleeding occurs in
approximately 15% of patients, with significant associated morbidity and mortality. Prior
to the paper by James Lau and colleagues, the evidence for use of H2 antagonists follow-
ing endoscopy had been conflicting and trials using proton pump inhibitors (PPIs) had
significant methodological flaws. A meta-analysis of 27 randomized trials concluded that
the use of H2-receptor antagonists was beneficial in patients with bleeding gastric ulcers
(Collins et al. 1985).
Study references
Main studies
Lau, J.Y.W., Sung, J.J.Y., Lee, K.K.C., et al. (2000) Effect of intravenous omeprazole on recurrent bleeding
after endoscopic treatment of bleeding peptic ulcers. New England Journal of Medicine, 343, 310–16.
Sung, J.J.Y., Barkun, A., Kuipers, E.J., et al. (2009) Intravenous esomeprazole for prevention of recurrent
peptic ulcer bleeding. Annals of Internal Medicine, 150, 455–64.
Related reference
Collins, R,. AND Langman, M. (1985) Treatment with histamine H2 receptor antagonists in acute upper
gastrointestinal haemorrhage—Implications of randomized trials. New England Journal of Medicine,
313, 660–6.
Study design
Lau et al.
◆ A randomized double-blind placebo-controlled single-centre study.
Randomization 72 hour omeprazole infusion versus placebo following endoscopic
haemostasis of upper GI bleeding
Inclusion criteria Ulcers with active haemorrhage or a visible vessel
Age >16 years
Primary outcome Recurrent bleeding
Follow-up 30 days after treatment
◆ All patients admitted with upper GI bleeding underwent endoscopy within 24 hours
of admission. Endoscopic treatment was with adrenaline injection and heater probe
coagulation.
◆ In the treatment arm patients received an 80mg bolus of intravenous (IV) omeprazole
followed by an infusion of 8mg/hour for 72 hours. After 72 hours all patients were
given 20mg oral omeprazole for 8 weeks with or without eradication therapy.
◆ Bleeding was considered to have recurred if any of the following occurred: vomiting
of fresh blood, shock (systolic blood pressure ≤90mmHg or pulse rate ≥110bpm)
with malaena after stabilization, or a drop in haemoglobin of >2g/dl less than 24
hours after transfusion to a level of 10g/dl
Sung et al.
◆ A randomized blinded placebo-controlled parallel group multi-centre study.
Setting 91 hospitals in 16 countries
Randomization 72 hour esomeprazole infusion versus placebo following endoscopic
haemostasis of upper GI bleeding
Inclusion criteria Gastric or duodenal ulcers with high-risk stigmata
Exclusion criteria Multiple ulcers
Other major disease or life expectancy <6 months
NSAID, COX-2 inhibitor, aspirin, or clopidogrel during first 7 days of study
Primary outcome Recurrent bleeding within 72 hours
Secondary outcomes Re-bleeding within 7–30 days, death, surgery, endoscopic re-treatment,
blood transfusion, hospitalization, safety
◆ Quality control of stigmata of haemorrhage findings at initial endoscopy was

achieved using photo documentation and review by an endpoint committee
◆ Endoscopic haemostasis was by adrenaline injection, thermocoagulation, or both, or
by application of haemoclips
Results
Lau et al.
◆ Out of 739 patients admitted with bleeding peptic ulcers, 267 required endoscopic
treatment and 240 were randomized to the study.
Omeprazole infusion Placebo (n = 120) p-value

(n = 120)
Re-bleeding 6.7% 22.5% <0.001
Blood transfused (mean) 2.7 units 3.5 units 0.04
Hospital stay (median) 4 days 5 days 0.006
Surgery 2.5% 7.5% 0.14
Deaths 4.2% 10% 0.13
PPI TREATMENT FOLLOWING ENDOSCOPIC THERAPY FOR BLEEDING PEPTIC ULCER 119
Sung et al.
Esomeprazole infusion Placebo p-value
(n = 375) (n = 389)
Re-bleeding within 72 hours 5.9% 10.3% 0.026
Endoscopic re-treatment 6.4% 11.6% 0.012
Total blood transfused 589 units 935 units 0.034
Surgery 2.7% 5.4% 0.059
Mortality 0.8% 2.1% 0.22
◆ Approximately 50% of the patients in both groups received single endoscopic

therapy.
◆ The protective effect of esomeprazole with respect to re-bleeding was maintained at 7
and 30 days post-endoscopy
Critique
The role of endoscopic haemostasis in the treatment of peptic ulcer bleeding was not
established during the period of studies that examined the role of histamine H2 receptor
therapy in peptic ulcer bleeding. Early trials of PPI therapy had small numbers of patients,
used low dosage regimens, lacked discrete outcome variables, and either had poor-qual-
ity or no endoscopic therapy administered. The study from James Lau and colleagues in
Hong Kong is considered to be the landmark paper in which most of these issues were
addressed. In particular, this study was set within a centre of excellence in terms of pub-
lished clinical outcomes for endoscopic haemostasis and employed a high dose of IV PPI
therapy. Racial differences in parietal cell mass and genetic polymorphisms affecting PPI
metabolism, coupled with regional variations in prevalence of Helicobacter pylori, have
raised doubts about the generalizability of the favourable results from this study.
The subsequent study by Joseph Sung and colleagues addresses these criticisms. The
study includes a larger patient cohort, it is multi-centre and international, and a mix of
racial backgrounds is represented in the patient populations studied. There are significant
improvements in study design, methodology, and quality control of patient entry and
documentation. It sets the gold standard for this type of trial in therapy for peptic ulcer
bleeding.
Interestingly, neither study attempted to use mortality as a primary endpoint. The Lau
study was not sufficiently powered to address this issue, and the Sung study avoided it by
excluding many high-risk patients, only using re-bleeding rates at 72 hours as the pri-
mary endpoint. It remains to be determined whether pre-endoscopy PPI therapy, oral PPI
therapy, or lower-dose PPI infusion may have similar beneficial effects on reducing the
risk from bleeding peptic ulcers.
4.6 Laparoscopic fundoplication versus medical therapy for the

treatment of gastro-oesophageal reflux
Details of studies
Gastro-oesophageal reflux disease (GORD) is most often treated medically with PPIs.
Anti-reflux surgery is an effective alternative but is traditionally reserved for patients with
persistent symptoms on medication. Four randomized trials and a Cochrane review have
examined the role of laparoscopic Nissen fundoplication (LFN) surgery in the treatment
of gastro-oesophageal reflux disease.
Study references
Main studies
Mahon, D., Rhodes, M., Decadt, B., et al. (2005) Randomized clinical trial of laparoscopic Nissen fundo-
plication compared with proton-pump inhibitors for treatment of chronic gastro-oesophageal reflux.
REFLUX study
Grant, A.M., Wileman, S.M., Ramsay, C.R., et al. (2008) Minimal access surgery compared with medi-
cal management for chronic gastro-oesophageal reflux disease: UK collaborative randomised trial.
British Medical Journal, 337, a2664.
Anvari, M., Allen, C., Marshall, J., et al. (2011) A randomised controlled trial of laparoscopic Nissen
fundoplication versus proton pump inhibitors for treatment of patients with chronic gastroesopha-
geal reflux disease (GERD): 3-year outcomes. Surgical Endoscopy, 25, 2547–54.
LOTUS study
Galmiche, J.P., Hatlebakk, J., Attwood, S., et al. (2011) Laparoscopic antireflux surgery vs esomeprazole
treatment for chronic GERD: The LOTUS randomized clinical trial. JAMA, 305, 1969–77.
Related references
Wileman, S.M., McCann, S., Grant, A.M., Krukowski, Z.H., and Bruce, J. (2010) Medical versus surgi-
cal management for gastro-oesophageal reflux disease (GORD) in adults. Cochrane Database of
Systematic Reviews, (3), CD003243.
Study design
◆ All four studies are RCTs.
Mahon et al. REFLUX study Anvari et al. LOTUS study

Level of evidence 1b 1b 1b 1b
Randomization LFN vs PPI Lap fundoplication LFN vs PPI Lap anti-reflux
vs medical therapy surgery vs PPI
Number of patients 217 357 101 554
Number of centres 2 21 1 11
Inclusion criteria >12 months (>6 months Mahon et al.) of symptoms requiring PPI treatment
and endoscopic or pH evidence of reflux
LAPAROSCOPIC FUNDOPLICATION VERSUS MEDICAL THERAPY FOR TREATMENT OF GOR 121
Outcome measures
Mahon et al.
◆ Primary outcome: acid exposure and oesophageal manometry at 3 months.
◆ Secondary outcomes: quality of life at 3 and 12 months assessed by Physiological

General Well-being Index and the Gastrointestinal Symptom Rating Scale.
REFLUX study
◆ Primary outcome: REFLUX questionnaire score.
◆ Secondary outcomes: quality of life as assessed by SF-36 and EQ-5D, serious

morbidity, mortality, NHS and patient costs.
Anvari et al.
◆ Primary outcome: Gastroesophageal Reflux Symptom Score (GERSS).
◆ Secondary outcomes: endoscopy, manometry, oesophageal pH studies, and quality of

life as assessed by SF-36 and EQ-5D.
LOTUS study
◆ Primary outcome: time to treatment failure.
◆ Secondary outcomes: severity of symptoms at 5 years, adverse events, and mortality.
Results
Mahon et al.
◆ Mean DeMeester acid exposure scores were reduced significantly in both groups but
were lower in the surgical group than in the PPI group at 3 months (8.6 versus 17.7, p
<0.001).
◆ At 12 months, improvement in the mean gastrointestinal well-being scores was
greater in the surgical group than in the PPI group (p <0.001). Improvement in
general well being scores were also greater in the surgery group (p = 0.002).
REFLUX study
◆ The recruitment and randomization process was complex and complicated by
significant delays between randomization and surgery. Consequently, the ITT
surgical group (n = 178) was much larger than the PP surgical group (n = 111).
◆ At 1 year, disease-specific symptoms and general quality of life were significantly
better in the surgical group than in the medical group.
◆ At 12 months, 38% of those randomized to surgery were taking reflux medication
compared with 90% in the medical therapy group.
◆ There was no difference in dysphagia scores between the two groups at 12 months.
Anvari et al.
◆ The two groups did not differ significantly in change of GERSS from baseline at 3
years.
◆ The surgical group had significantly more heartburn-free days per week and a higher
pressure in the lower oesophageal sphincter.
◆ There was no significant difference between the groups when 24 hour pH studies
were performed at 3 years.
◆ Quality of life scores were significantly higher after surgery.
◆ At 3 years six surgical patients (11.8%) and eight medical patients (16%) had failed
their primary treatment
LOTUS study
◆ Time to treatment failure, expressed as estimated remission rates at 5 years, was 92%
in the esomeprazole group and 85% in the antireflux surgery group (p <0.05).
◆ Severity of acid regurgitation at 5 years was reduced in the surgery group.
◆ Prevalence and severity of dysphagia, bloating, and flatulence were increased in the
surgery group.
◆ There was no difference in adverse events between surgery and esomeprazole
(29% versus 24%).
Critique
The Cochrane Database analysis of the evidence concluded that there is evidence that
laparoscopic fundoplication surgery is more effective than medical management for the
treatment of GORD, at least in the short to medium term. Surgery does carry some risk,
and it is uncertain whether its benefits are sustained in the long term. Since the publica-
tion of the Cochrane review the longer-term results for the study by Anvari et al. (3 years)
and the LOTUS study (5 years) have been published and confirm similar outcomes to
those of the initial publications.
The slight differences in outcomes between the trials reflect the different inclusion
criteria used. When patients with partial or poor response to PPI therapy are excluded
the beneficial effects of surgery are smallest, as seen in the LOTUS study. Although the
REFLUX study is the largest study, there were methodological and logistical complexities
that complicate the analysis. The fact that the majority of patients approached for ran-
domization in the REFLUX study refused it and entered the parallel patient-preference
study reflects the importance of patient and physician preference in the management of
reflux symptoms.
The data do support the conclusion that both medical and surgical management of
GORD are very effective in controlling symptoms and improving quality of life. Surgery
is more expensive in the short term and, although it can be performed with few adverse
outcomes, there is potential for morbidity when it is widely applied outside academic or
specialist centres.
LAPAROSCOPIC ANTI-REFLUX SURGERY AND PARA-OESOPHAGEAL HERNIA REPAIR 123
4.7 Technical aspects of laparoscopic anti-reflux surgery

and para-oesophageal hernia repair
Details of studies
A laparoscopic approach to anti-reflux surgery and para-oesophageal hernia repair is
now considered standard. Recent long-term data (Salminen et al.) have confirmed greater
patient satisfaction, fewer disrupted wraps, and fewer incisional hernias within the lapa-
roscopic group.
Controversies persist with regard to the type of fundoplication (anterior versus poste-
rior, partial versus total), short gastric vessel division, and the use of prosthetic reinforce-
ment to the hiatus. The meta-analyses by Broeders et al. and Markar et al. have helped to
resolve the first two issues. The use of prosthetic reinforcement remains controversial. A
review by Granderath et al. has shown that prosthetic reinforcement reduces the recur-
rence rate at the expense of increased dysphagia. Serious complications, such as mesh ero-
sion, have been reported and so recent interest has focused on the use of biological mesh.
Oelschlager et al. describe the short- and long-term outcomes from an RCT looking at
biological mesh versus sutured repair.
Study references
Main studies
Broeders, J.A., Roks, D.J., Ahmed Ali, U., et al. (2011) Laparoscopic anterior versus posterior fundo-
plication for gastroesophageal reflux disease: systematic review and meta-analysis of randomized
clinical trials. Annals of Surgery, 254, 39–47.
Markar, S.R., Karthikesalingam, A.P., Wagner, O.J., et al. (2011) Systematic review and meta-analysis
of laparoscopic Nissen fundoplication with or without division of the short gastric vessels. British
Oelschlager, B.K., Pellegrini, C.A., Hunter, J., et al. (2006) Biologic prosthesis reduces recurrence after
laparoscopic paraesophageal hernia repair: a multicenter, prospective, randomized trial. Annals of
Surgery, 244, 481–90.
Oelschlager, B.K., Pellegrini, C.A., Hunter, J.G., et al. (2011) Biologic prosthesis to prevent recurrence
after laparoscopic paraesophageal hernia repair: long-term follow-up from a multicenter, prospec-
tive, randomized trial. Journal of the American College of Surgeons, 213, 461–8.
Related references
Salminen, P., Hurme, S., and Ovaska, J. (2012) Fifteen-year outcome of laparoscopic and open Nissen
fundoplication: a randomized clinical trial. Annals of Thoracic Surgery, 93, 228–33.
Granderath, F.A., Carlson, M.A., Champion, J.K., et al. (2006) Prosthetic closure of the esophageal hia-
tus in large hiatal hernia repair and laparoscopic antireflux surgery. Surgical Endoscopy, 20, 367–79.
Type of fundoplication: Broeders et al.

Study design
◆ A meta-analysis of seven RCTs: two comparing laparoscopic anterior fundoplication
(LAF) and partial laparoscopic posterior fundoplication (LPF), and five comparing
LAF and total LPF.
◆ Primary outcomes: oesophageal acid exposure, heartburn, dysphagia score, and

re-operation rate.
◆ Secondary outcomes: endoscopic oesophagitis, regurgitation, inability to belch, gas
bloating, ability to relieve bloating, satisfaction with intervention, willingness to
undergo surgery again, operating time, conversion rate, in-hospital complications,
and length of hospital stay.
◆ Short-term (6–12 months) and long-term (2–10 years) results were pooled separately.
Results
◆ 683 fundoplications were included in the analysis.
◆ Short term: mean oesophageal acid exposure time was higher (3.3% versus 0.8%, p
<0.001), heartburn was more prevalent (21% versus 8%, p <0.001), and the mean
dysphagia score was lower (2.5 versus 5.7, p <0.001) after LAF.
◆ Surgical re-intervention was more common after LAF than after LPF, although this
difference did not reach statistical significance (8% versus 4%, p = 0.06).
◆ Long term: LAF had a twofold higher rate of heartburn (31% versus 14%, p <0.001)
associated with more PPI use (25% versus 10%, p = 0.002). Dysphagia scores and
ability to relieve bloating became similar in the two groups.
◆ The re-operation rate remained twice as high after LAF in the long term (10% versus
5%, p = 0.03). No differences were noted in other secondary outcomes in the longer
term.
Conclusions
Oesophageal acid exposure time and the prevalence of heartburn are higher after LAF
than after LPF. In the short term this is counterbalanced by less severe dysphagia.
However, dysphagia scores become similar in the long term, with a substantial and per-
sistent increase in prevalence of heartburn and PPI use after LAF. The re-operation rate
is twice as high after LAF, mainly because of re-interventions for recurrent reflux. The
prevalence of gas-related symptoms is similar.
Division of the short gastric vessels: Markar et al.

Study design
◆ A meta-analysis of five RCTs comparing short gastric vessel (SGV) division versus no
SGV division at laparoscopic Nissen fundoplication.
◆ Primary outcomes: requirement for re-operation, post-operative dysphagia, and
presence of post-operative gastro-oesophageal reflux.
◆ Secondary outcomes: duration of operation, length of hospital stay, presence of post-
operative complications, and gas bloat syndrome.
◆ The quality of each trial was assessed using the Jadad scoring system.
LAPAROSCOPIC ANTI-REFLUX SURGERY AND PARA-OESOPHAGEAL HERNIA REPAIR 125
Results
◆ 388 patients were included in the analysis.
◆ There were no significant differences between SGV division and non-division in
re-operation requirement (6% versus 2%), post-operative dysphagia (21% versus
17%), or post-operative GORD (13% versus 18%).
◆ No significant differences were noted in post-operative complications, length of
hospital stay, or symptoms of gas bloat, although there was statistical heterogeneity
between studies. This heterogeneity was also noted in operation time, which
suggested a longer time when SGVs were divided.
Conclusions
Clinical outcomes following laparoscopic Nissen fundoplication appear to be similar
regardless of whether the short gastric vessels are divided.
Biological mesh versus sutured repair: Oelschlager et al.

Study design
Randomization Biological prosthesis versus suture repair
Inclusion criteria Symptomatic para-oesophageal hernia >5cm on contrast swallow
Exclusion criteria Previous oesophagogastric surgery
Emergency presentation
Outcome measure Recurrent (>2cm) hiatus hernia on contrast examination
Quality of life (SF-36)
Results
◆ After 6 months follow-up the recurrence rate was significantly lower in the mesh
group (9% versus 24%, p = 0.04).
◆ After a median follow-up of 58 months there was no significant difference in
recurrence between the groups (54% versus 59%).
◆ There were no significant differences in quality of life between the two groups.
◆ There were no cases of mesh erosion, strictures, or dysphagia due to mesh.
Conclusions
Laparoscopic para-oesophageal hernia repair results in long and durable relief of symp-
toms and improvement in quality of life with either sutured or biological mesh repair.
There does not appear to be a higher rate of complications or side effects with biological
mesh, but its benefit in reducing recurrence diminishes at long-term follow-up.
Critique
The meta-analysis by Broeders et al. included a large number of patients. The results pro-
vide strong guidance to surgeons in the choice of antireflux procedure. This meta-analysis
only included RCTs with a high mean Jadad score (4). However, there was some het-
erogeneity in the quality of the studies and there were reported variations in the imple-
mentation of oesophageal fixation and short gastric division. Not all the trials reported
longer-term results. In a separate meta-analysis by the same authors the two common
posterior fundoplications (Toupet and Nissen) were compared and shown to be equiva-
lent with regard to symptomatic reflux control, although the Toupet repair was associated
with less functional symptoms such as dysphagia. This particular meta-analysis has been
criticized on the basis of its methodology. As a result, there is insufficient evidence to
recommend a particular posterior fundoplication.
The meta-analysis by Markar et al. was also well conducted and included only RCTs,
although these were of varying quality. Heterogeneity was also noted in clinical outcomes
as well as in follow-up period. Because of the relatively small sample size and low volume
of some endpoints the lack of significant difference between groups may be due to a type
2 error. However, this analysis suggests that SGV division is not a requirement for success-
ful Nissen fundoplication, and the decision to undertake this should probably be at the
discretion of the operating surgeon.
The use of prosthetic mesh reduces recurrence rate following laparoscopic para-
oesophageal hernia repair at the potential cost of dysphagia and erosion into the oesopha-
gus. The randomized trial by Oelschlager et al. assessed the use of biological mesh. The
study was blinded and groups were well matched. Whilst an objective follow-up of 90%
was achieved in the shorter term, in the longer term only 56% of patients underwent
repeat contrast swallow. The study confirmed the safety of biological prostheses and sug-
gested a significant reduction in recurrence in the shorter term. However, longer-term
follow-up demonstrated no difference in recurrence rate between groups, suggesting that
the use of this mesh may delay rather than prevent recurrence.
Broeders, J.A., Mauritz, F.A., Ahmed Ali, U., et al. (2010) Systematic review and meta-analysis of laparo-
scopic Nissen (posterior total) versus Toupet (posterior partial) fundoplication for gastro-oesopha-
geal reflux disease. British Journal of Surgery, 97, 1318–30.
ADENOCARCINOMA IN PATIENTS WITH BARRETT’S OESOPHAGUS 127
4.8 Incidence of adenocarcinoma in patients with

Barrett’s oesophagus
Details of studies
There is marked heterogeneity in the reported incidence of adenocarcinoma in patients
with Barrett’s oesophagus. The majority of studies are of small patient cohorts with
poor discrimination of those with prevalent high-grade dysplasia (HGD) or oesopha-
geal adenocarcinoma (OAC) from those truly developing incident HGD/OAC. Studies
have tended to be of selected groups of patients entered into surveillance programmes
in specialist referral centres and are less likely to accurately reflect the wider population
of Barrett’s oesophagus patients, the majority of whom do not even enter surveillance
strategies. The pooled incidence of OAC in two recent reviews has been 5.3 and 6.5 cases
per 1000 person-years and the incidence estimates for the combined endpoint of HGD/
OAC were 9.1 and 10.2 cases per 1000 person-years. Most agree that current surveillance
strategies and cost effectiveness are based on an incidence of five cases per 1000 years of
person follow-up. There remains widespread suspicion that publication bias exists and
that the true incidence of HGD/OAC in Barrett’s oesophagus may be significantly lower
than the published figure of 0.5% per year. Accurate knowledge of the relative risk of
tumour development is imperative in the context of ascertaining the effectiveness and
appropriateness of surveillance and screening programmes. Recently two large national
population based studies have investigated this issue.
Study references
Main studies
de Jonge, P.J.F, van Blankenstein, M., Looman, C.W.N, Casparie, M.K., Meijer, G.A., and Kuipers, E.J.
(2010) Risk of malignant progression in patients with Barrett’s oesophagus: a Dutch nationwide
cohort study. Gut, 59, 1030–6.
Hvid-Jensen, F., Pedersen. L., Drewes, A.M., Sorensen, H.T., Funch-Jensen, P. (2011) Incidence of
adenocarcinoma among patients with Barrett's esophagus. New England Journal of Medicine, 365,
1375–83.
Study design
◆ Both are nationwide population-based cohort studies
de Jonge et al. Hvid-Jensen et al.

Setting Dutch national pathology database (1991– Danish national pathology
2006) database (1992–2009)
Inclusion criteria Patients with Barrett’s oesophagus and either no dysplasia or low-grade
dysplasia (LGD) at index endoscopy
Number of patients 42,207 11,028
Follow-up 2 21
◆ Both studies identified and excluded as prevalent cancer any patient developing an
OAC within 1 year of the initial diagnosis.
◆ Progression to HGD and adenocarcinoma was estimated and risk factors for
malignant progression were evaluated.
Results
de Jonge et al. Hvid-Jensen et al.
Progression to HGD/OAC 5.8/1000 person-years 2.6/1000 person-years
Annual risk of carcinoma 0.4% 0.12%
Risk factors for progression Male sex, increasing age, Male sex, increasing age,
presence of LGD presence of LGD
◆ Only 16,365 patients (39%) in the study by de Jonge et al. underwent a follow-up
endoscopy. When the full population is included the annual risk of carcinoma is
0.14%.
Critique
Prior to the publication of these two studies there had been a reliance on meta-analysis of
pooled patient data to estimate the annual risk of Barrett’s oesophagus patients develop-
ing OAC. These high-quality population-based studies are sufficiently powered in terms
of patient numbers and longitudinal follow-up to improve the reliability in more accu-
rately determining the risk of a patient with Barrett’s oesophagus developing OAC and/or
HGD. Although the Dutch study reports an annual risk of 0.4% for incident OAC, this fig-
ure is derived from the subpopulation of 39% of patients within endoscopic surveillance.
When recalculated based on the complete population-based cohort the annual incidence
falls to 0.14%/year. This figure is remarkably similar to the Danish study which did not
subdivide the population and found an annual cancer risk of 0.12%.
Both studies also found in multivariate analysis that male sex, older age, and presence
of LGD at index endoscopy were independent variables associated with the subsequent
development of HGD/OAC.
Current guidelines and cost-effectiveness models for Barrett’s surveillance are based on
the previously held view that the annual rate of progression to OAC was of the order of
0.5%. Both these studies demonstrate that the actual incidence is significantly lower and
therefore both groups conclude that the risk of incident oesophageal adenocarcinoma
among patients with Barrett’s oesophagus is so minor that, in the absence of dysplasia,
routine surveillance of such patients is of doubtful value. In order to apply this change
in strategy it will be necessary to be accurate in assessing whether prevalent disease is
present, as one-third of cancers are detected within 1 year of index endoscopy. Both stud-
ies have identified independent risk factors for progression to HGD/OAC and future
research will need to concentrate on risk stratification to identify those patients who will
benefit from surveillance.
RELATIONSHIP OF HOSPITAL VOLUME TO OUTCOME IN OESOPHAGECTOMY 129
4.9 Relationship of hospital volume to outcome

in oesophagectomy
Details of studies
The relationship between hospital volume and outcome for oesophagectomy has been
promoted as one of the reasons for centralization of oesophageal cancer services. In 1998
Begg et al. correlated hospital volume with mortality for major cancer surgery, with the
most striking relationship between the two occurring in oesophagectomy patients. Since
then a number of studies have further addressed whether hospital volume impacts on
outcome post-oesophagectomy. Markar et al. summarized the recent data in their system-
atic review and meta-analysis.
Study references
Main studies
Begg, C.B., Cramer, L.D., Hoskins, W.J., et al. (1998) Impact of hospital volume on operative mortality
for major cancer surgery. JAMA, 280, 1747–51.
Finlayson, EV., Goodney, P.P., and Birkmeyer, J.D. (2003) Hospital volume and operative mortality in
cancer surgery: a national study. Archives of Surgery, 138, 721–5.
Markar, S.R., Karthikesalingam, A., Thrumurthy, S., et al. (2011) Volume–outcome relationship in
surgery for esophageal malignancy: systematic review and meta-analysis 2000–2011. Journal of
Gastrointestinal Surgery 16, 1055–63.
Related references
Metzger, R., Bollschweiler, E., Vallböhmer, D., et al. (2004) High volume centers for esophagectomy:
what is the number needed to achieve low postoperative mortality? Diseases of the Esophagus, 17,
310–14.
Lin, H.C., Xirasagar, S., Lee, H.C., and Chai, C.Y. (2006) Hospital volume and inpatient mortality after
cancer-related gastrointestinal resections: the experience of an Asian country. Annals of Surgical
Oncology, 13, 1182–8.
Wouters, M.W., Wijnhoven, B.P., Karim-Kos, H.E. et al. (2008) High-volume versus low-volume for
esophageal resections for cancer: the essential role of case-mix adjustments based on clinical data.
Annals of Surgical Oncology, 15, 80–7.
Rodgers, M., Jobe, B.A., O’Rourke, R.W., et al. (2007) Case volume as a predictor of inpatient mortality
after esophagectomy. Archives of Surgery, 142, 829–39.
Santin, B., Kulwicki, A., and Price, P. (2008) Mortality rate associated with 56 consecutive esophagec-
tomies performed at a ‘low-volume’ hospital. Is procedure volume as important as we are trying to
make it? Journal of Gastrointestinal Surgery, 12, 1346–50.
Study design
Begg et al.
◆ A retrospective cohort study.
◆ Inclusion criteria: all patients aged over 65 years in a national Medicare-linked
database registry who underwent major cancer resection between 1984 and 1993,
with separate analysis of patients undergoing oesophageal resection.
◆ Primary endpoint: 30-day mortality.
Finlayson et al.
◆ A retrospective cohort study.
◆ Inclusion criteria: patients in a national database undergoing major cancer resection
between 1995 and 1997, with separate analysis of patients undergoing oesophageal
resection.
◆ Primary endpoint: in-hospital mortality.
Markar et al.
◆ A meta-analysis of nine cohort studies comparing outcome following
oesophagectomy in high- and low-volume centres between 2000 and 2011.
◆ Endpoints: either in-patient mortality or 30-day mortality.
◆ Pooled odds ratios were calculated for the effect of high-volume centres on discrete
outcomes.
Results
Begg et al.
Oesophagectomy volume 1–5 6–10 >11

30-day mortality 17.3% 3.9% 3.4%
◆ The inverse correlation between volume and outcome was highly significant
(p <0.001).
Finlayson et al.
Oesophagectomy volume <4 4–9 >9
In-hospital mortality 15% 13.8% 6.5%
Markar et al.
◆ 27,843 patients were included in the analysis.
High-volume Low-volume OR 95% CI p-value

30-day mortality 2.09% 0.73% 0.31 0.19–0.51 <0.0001
In-hospital mortality 8.48% 2.82% 0.29 0.16–0.53 <0.0001
◆ No evidence of publication bias was identified

◆ There were insufficient data for conclusive statistical analysis of length of hospital stay
or post-operative complications. There was a trend towards increased complications
and longer hospital stay in the low-volume centres.
RELATIONSHIP OF HOSPITAL VOLUME TO OUTCOME IN OESOPHAGECTOMY 131
Critique
The studies by Begg et al. and Finlayson et al. were among the first to demonstrate the
relationship between volume and outcome following major cancer resection, in particular
oesophagectomy. Begg et al. had sufficient patient numbers to demonstrate a significant
inverse correlation between volume and outcome even in the subgroup of oesophagec-
tomy patients and taking into account other risk factors. The inclusion criteria (>65 years
old) reduced the general applicability of the paper and also meant that the operative vol-
ume, even in the supposed high-volume category, was still relatively small. Finlayson et al.
included all eligible patients undergoing major resection. This meant that the volume
distributions were more realistic and applicable. Again, oesophagectomy was shown to be
one of the procedures which is most sensitive to volume. The inverse relationship between
volume and outcome was most pertinent in patients with higher risk factors.
The recent meta-analysis by Markar et al. pooled sufficient patient numbers to demon-
strate significant differences in 30-day and in-hospital mortality between high- and low-
volume centres. Although a trend towards higher complications and longer hospital stay
was described in the low-volume group, this did not reach significance, perhaps because
of the fewer studies reporting these endpoints.
All studies included in the meta-analysis were observational, and so a direct effect of
volume and outcome can only be inferred. It is highly unlikely that an RCT will ever be
performed, and so the meta-analysis provides best possible evidence. The major criticism
of the paper is the heterogeneity of pooled studies, which was statistically significant. This
was most marked in the definition of high- and low-volume centres used in each publica-
tion (e.g. Lin et al.—low <78, high >346; Wouters et al.—low <9, high >9). Heterogeneity
was also demonstrated in other areas such as the data sources. Given this wide variability,
the study does not allow a cut-off for low- and high-volume centres to be defined. Some
authors have attempted to define such a cut-off; for example, Metzger et al. defined a high-
volume centre as >20 resections per year.
The current evidence suggests that mortality post-oesophagectomy is inversely pro-
portional to hospital volume, although many other factors such as case-mix may be just
as relevant. It is also worth remembering that many low-volume units have demonstrated
excellent results.
4.10 Trans-thoracic versus trans-hiatal oesophagectomy

Details of study
There is little consensus among surgeons about the appropriate surgical approach for
operable oesophageal cancer. This study was the first large RCT that compared the post-
operative risk and long-term outcomes of the more radical trans-thoracic oesophagec-
tomy (TTO) with the less radical trans-hiatal oesophagectomy (THO) in patients with
middle- and lower-third oesophageal adenocarcinoma.
Study references
Main study
Hulscher, J.B., van Sandick, J.W., de Boer, A.G. et al. (2002) Extended transthoracic resection compared
with limited transhiatal resection for adenocarcinoma of the esophagus. New England Journal of
Medicine, 347, 1662–9.
Related reference
Omloo, J.M., Lagarde, S.M., Hulscher, J.B., et al. (2007) Extended transthoracic resection compared with
limited transhiatal resection for adenocarcinoma of the mid/distal esophagus: five-year survival of a
randomized clinical trial. Annals of Surgery, 246, 992–1001.
Study design
◆ A prospective two centre RCT.
Randomization THO versus TTO with extended en bloc lymphadenectomy
Setting Two academic medical centres in The Netherlands, each performing >50
operations per year
Inclusion criteria Histologically confirmed adenocarcinoma of the mid to distal oesophagus
or adenocarcinoma of the gastric cardia involving the distal oesophagus
>18 years old and ASA I–II
Exclusion criteria Metastatic disease including positive cervical or coeliac lymph nodes
Previous or coexisting cancer
Previous gastric or oesophageal surgery
Neoadjuvant chemotherapy or radiation therapy
Recurrent laryngeal nerve palsy
Extension of the tumour which made it impossible for the surgeon to
construct a gastric tube
Follow-up Median 4.7 years
◆ Pre-operative work up included upper GI endoscopy, endoscopic ultrasound,

ultrasound of the neck and abdomen, chest X-ray, indirect laryngoscopy, and
bronchoscopy (as indicated). CT was not routinely performed.
TRANS-THORACIC VERSUS TRANS-HIATAL OESOPHAGECTOMY 133
◆ Lymphadenectomy in the THO group included peritumoral and left gastric artery
lymph nodes only. Coeliac lymph nodes were only dissected if felt to be involved.
◆ Lymphadenectomy in the TTO group included lower and middle mediastinal,
subcarinal, right-sided paratracheal, and aortopulmonary window lymph nodes in
the right chest and paracardial, lesser curvature, left gastric artery, coeliac trunk,
common hepatic artery, and splenic artery lymph nodes in the abdomen
◆ Patients in both groups had a cervical anastomosis. Lymph nodes in the neck were
not dissected.
◆ Standardized pathological assessment (TNM 1997), recording of post-operative
complications, and follow-up protocol.
Outcome measures
Primary endpoints
◆ Overall and disease-free survival
Secondary endpoints
◆ Early morbidity and mortality, number of quality-adjusted life-years gained and cost
effectiveness
Results
◆ No differences were observed between the two groups with respect to baseline
demographics.
◆ 93 of the 106 patients in the THO group (88%) and 109 of the 114 patients in the
TTO group (96%) underwent the planned procedure (p = 0.08).
◆ Patients in the standard group stayed in the ED for a significantly shorter period of
time.
THO (n = 106) TTO (n = 114) p-value

Operation time 3.5 hours 6.0 hours <0.001
Blood loss 1.0L 1.9L <0.001
Pulmonary complications 27% 57% <0.001
Chyle leakage 2% 10% 0.02
Ventilation time* 1 day 2 days <0.001
ICU stay* 2 days 6 days <0.001
Hospital stay* 15 days 19 days 0.04
In-hospital mortality 2% 4% 0.45
Lymph nodes excised 16 31 <0.001
Overall survival* 1.8 years 2.0 years 0.38
Disease-free survival* 1.4 years 1.7 years 0.15
Quality-adjusted life years* 1.5 years 1.8 years 0.26
*Values are medians
◆ There were no significant differences in tumour staging, radicality of surgery (R0

resections), and patterns of tumour recurrence between the two groups.
◆ The mean total direct and indirect costs of the procedure were 56% higher for the
TTO group than for the THO group (€37,099 versus €23,809).
Conclusions
TTO with extended en bloc lymphadenectomy is associated with increased morbidity (but
not mortality) and hospital costs when compared with THO. Although median overall,
disease-free, and quality-adjusted survival did not differ statistically between the groups,
there was a trend towards improved long-term survival at 5 years with the extended TTO
approach.
Critique
The principle of oesophagectomy is to remove all neoplastic tissue in the hope of provid-
ing a worthwhile period of survival, and hopefully cure, whilst minimizing morbidity and
mortality. The choice of a particular surgical approach depends on the size and location
of the tumour, planned lymphadenectomy, patient comorbidities, likely tolerance of tho-
racotomy, and surgeon preference.
There are two schools of thought with regard to the radicality of oesophagectomy, in
particular lymphadenectomy. On the one hand, lymph node metastases are regarded as
a marker for systemic disease and radical resection of these lymph nodes is unlikely to
confer significant benefit. In this case, a less radical resection, such as THO, is advocated.
On the other hand, there is a belief that a more radical, often TTO, approach may confer
a survival benefit in addition to improved staging and better locoregional control. This
study directly addressed these two approaches to oesophageal cancer surgery.
Prior to this study only three small RCTs had been performed. All these trials involved
small patient numbers (32–67 patients) and failed to demonstrate any difference in post-
operative morbidity or mortality and long-term survival in patients undergoing either
THO or TTO. The study by Hulscher et al. was adequately powered to demonstrate an
increase in the 2-year survival rate from 30% to 45% among patients undergoing trans-
thoracic resection with extended en bloc lymphadenectomy. The study was well con-
ducted in two high-volume centres with good surgical quality assurance, as evidenced by
almost 90% of the TTO group’s having 15 or more lymph nodes removed and identified
by the pathologist.
Whilst it was clear that post-operative morbidity, ICU and hospital stay, and hospital
costs were increased in the TTO group, the authors could only suggest that there was a
trend towards improved overall and disease-free survival in the TTO group. The Kaplan–
Meier curves for disease-free and overall survival were similar early after surgery in both
groups but diverged (without the difference reaching statistical significance) after 3 years,
with the difference favouring the extended resection. The authors recommended further
follow-up of the patients in this study to clarify whether the long-term benefits of the
TRANS-THORACIC VERSUS TRANS-HIATAL OESOPHAGECTOMY 135
extended approach, in terms of survival, outweigh the increase in early morbidity and
associated costs. A subsequent report by Omloo et al. demonstrated 5-year survival rates
of 34% and 36% in the THO and TTO groups respectively (p = 0.71). Subsequent sub-
group analysis demonstrated a significant survival benefit towards TTO in patients with
Siewert type 1 tumours and in patients with between one and eight positive lymph nodes.
The main criticism of this study is that the patients treated now represent a historical
cohort. Modern staging investigations including CT and positron emission tomography–
computed tomography (PET–CT) were not utilized. Eight per cent of patients did not
undergo the planned procedure. Multi-modality treatments such as chemotherapy and/
or radiotherapy were not used for these patients. Furthermore, there have been improve-
ments in peri-operative care and surgical techniques, such as minimally invasive surgery,
since this study was performed. However, it remains an important study in the discussion
of surgical approach for operable oesophageal cancer.
Goldminc, M., Maddern, G., Le Prise, E., et al. (1993) Oesophagectomy by a transhiatal approach or
thoracotomy: a prospective randomized trial. British Journal of Surgery, 80, 367–70.
Chu, K.M., Law, S.Y., Fok, M., and Wong, J. (1997) A prospective randomized comparison of transhiatal
and transthoracic resection for lower-third esophageal carcinoma. American Journal of Surgery, 174,
320–4.
Jacobi, C.A., Zieren, H.U., Müller, J.M., and Pichlmaier, H. (1997) Surgical therapy of esophageal
carcinoma: the influence of surgical approach and esophageal resection on cardiopulmonary
function. European Journal of Cardiothoracic Surgery, 11, 32–7.
4.11 The role of lymphadenectomy in gastric cancer surgery

Details of studies
The aim of surgery for gastric cancer is to excise the primary lesion with clear margins.
The type of resection is determined by the position and pre-operative stage of the cancer
and the planned lymphadenectomy. A systematic or D2 lymphadenectomy involves the
excision of the primary lesion together with the omentum and the first two tiers of lymph
nodes (N1 and N2) that drain the affected area of the stomach. This approach is now
regarded as routine in Japan with 5-year survival rates >50% having been reported from
non-randomized studies. The results of D2 lymphadenectomy have proved to be difficult
to replicate in the West. Cuschieri et al. published the short- and long-term results of a UK
MRC trial comparing D1 and D2 resection for gastric cancer. A number of randomized
and non-randomized studies have been published since this trial. Memon et al. recently
brought these results together in a meta-analysis.
Study references
Main studies
Cuschieri, A., Fayers, P., Fielding, J., et al. (1996) Postoperative morbidity and mortality after D1 and D2
resections for gastric cancer. Preliminary results of the MRC randomised controlled surgical trial.
The Surgical Co-operative Group. Lancet, 347, 995–9.
Cuschieri, A., Weeden, S., Fielding, J., et al. (1999) Patient survival after D1 and D2 resections for gastric
cancer: long-term results of the MRC randomized surgical trial. Surgical Co-operative Group.
British Journal of Cancer, 79, 1522–30.
Memon, M.A., Subramanya, M.S., Khan, S., et al. (2011) Meta-analysis of D1 versus D2 gastrectomy for
gastric adenocarcinoma. Annals of Surgery, 253, 900–11.
Related references
Bonenkamp, J.J., Hermans, J., Sasako, M., et al. (1999) Extended lymph-node dissection for gastric
cancer. New England Journal of Medicine, 340, 908–14.
Songun, I., Putter, H., Kranenbarg, E.M., Sasako, M., and van de Velde, C.J. (2010) Surgical treatment
of gastric cancer: 15-year follow-up results of the randomised nationwide Dutch D1D2 trial. Lancet
Oncology, 11, 439–49.
Study design
Cuschieri et al.
◆ A prospective multi-centre randomized trial.
Randomization D1 versus D2 gastrectomy
Inclusion criteria Histologically proven, potentially curable gastric cancer confirmed at
laparotomy
Exclusion criteria <20 years old
Previous gastric surgery
Coexisting cancer or serious comorbidities
Outcome measures Early mortality, morbidity, hospital stay, and 5-year survival
LYMPHADENECTOMY IN GASTRIC CANCER SURGERY 137
Memon et al.
◆ Six randomized trials, comprising 1876 patients, comparing D1 and D2 gastrectomy
for histologically proven adenocarcinoma were analysed.
◆ Outcomes: hospital stay, post-operative complications, anastomotic breakdown,
re-operation rate, 30-day mortality, and 5-year survival
◆ Meta-analyses were performed using odds ratios (ORs) for binary outcomes
and weighted mean differences (WMDs) for continuous outcome measures.
Heterogeneity among studies was assessed using the Q statistic and I2 index.
◆ Patients were included only if a curative resection was being undertaken (no distant
metastases, peritoneal involvement, or adjacent organ involvement)
◆ The dissection was carried out in accordance with the guidelines of the Japanese
Research Society for Gastric Cancer
Results
Cuschieri et al.
D1 (n = 200) D2 (n = 200) p-value
Lymph nodes (median) 13 17 <0.001
Hospital mortality 6.5% 13% 0.04
Overall morbidity 28% 46% <0.001
Hospital stay (median) 14 days 14 days ns
5-year survival 35% 33% ns
◆ Logistic regression analysis showed that the difference in morbidity and mortality
between D1 and D2 approaches became non-significant after allowance for
pancreatico-splenectomy.
◆ Five-year survival was not related to either the number of lymph nodes removed
or the described extent of lymph node resection, but was significantly lower if a
pancreato-splenectomy had been performed.
◆ In multivariate analysis, poorer survival was evident in patients who were older,
male, had advanced stage of disease, or who underwent pancreatico-splenectomy.
Memon et al.
◆ There was almost a perfect agreement (κ = 0.99) between the authors regarding the
inclusion and exclusion of various RCTs.
◆ Publication bias and significant heterogeneity were evident for two outcomes: length
of hospital stay and post-operative complications.
◆ The mean Jadad score was 2, suggesting poor methodological quality of included
trials.
◆ The D1 group was significantly favoured over the D2 group for the following
outcomes: hospital stay, post-operative complications, anastomotic breakdown,
re-operation rate, 30-day mortality rate, and 5-year survival
Conclusions
Cuschieri et al.
D2 gastric resections are associated with higher morbidity and mortality than D1 resec-
tions and with similar 5-year survival. The higher morbidity and mortality is mainly
related to performing a pancreato-splenectomy. These findings indicate that the classical
Japanese D2 resection offers no survival advantage over D1 surgery. However, the pos-
sibility that D2 resection without pancreato-splenectomy may be better than standard D1
resection cannot be dismissed on the basis of the results of this trial.
Memon et al.
D1 gastrectomy is associated with fewer complications and similar 5-year survival to D2
gastrectomy.
Critique
A systematic or D2 lymphadenectomy for gastric cancer allows more accurate staging
and better treatment of nodal metastasis. Japanese and Western studies have shown that
a significant proportion of patients with N2 disease survive for more than 5 years after a
D2 resection. However, it has been difficult to replicate these results in randomized trials
performed in the West. Only small studies from specialized European centres have been
able to achieve equivalent results to the Far East, with operative mortality rates well under
5% and corresponding improvements in survival.
The MRC Gastric Cancer Surgical Trial by Cuschieri et al. was a well-conducted study
with an appropriate sample size. The pre-operative staging laparotomy ensured few proto-
col violations. Uniformity of surgical technique and quality assurance was attempted. The
main criticism of the trial is contamination of extent of lymphadenectomy in both groups.
Although significantly more nodes were removed in D2 patients, there was evidence of
non-compliance. In the D2 arm the requisite level 8 lymph nodes were removed in only
95 patients. It was accepted that many of the surgeons in the D2 arm were still in their
‘learning curve’ phase. In subgroup analysis it was clear that routine pancreato-splenec-
tomy had significantly contributed to the increased morbidity and mortality following D2
gastrectomy. The routine use of pancreato-splenectomy may have been partially respon-
sible for the lack of survival benefit following D2 gastrectomy.
The Dutch Gastric Cancer Trial (Bonenkamp et al.) was a similar multi-centre prospec-
tive RCT comparing D1 and D2 lymphadenectomy. A total of 711 patients were recruited
and surgical quality control was more rigorous, with direct input from a Japanese sur-
geon. The results were very similar to the MRC Gastric Cancer Surgical Trial. Post-
operative morbidity, mortality, and hospital stay were significantly worse in the D2 group,
LYMPHADENECTOMY IN GASTRIC CANCER SURGERY 139
but 5-year survival was similar to the D1 group. Criticisms of this trial were also similar.
Songun et al. recently reported the long-term outcomes from this trial. After a median
follow-up of 15 years, D2 lymphadenectomy was associated with lower locoregional
recurrence and gastric-cancer-related death rates than D1 surgery. It should be noted
that much of the stage-specific improvement in survival after D2 resection is likely to be
the result of better pathological staging (stage migration factor), particularly in stages
II and IIIa. Twenty per cent of the D2 group with N2 nodes were still alive at 11 years.
The authors concluded that, because a safer spleen-preserving D2 resection technique is
currently available in high-volume centres, D2 lymphadenectomy should be the recom-
mended surgical approach for patients with resectable gastric cancer.
The meta-analysis by Memon et al. found that D2 gastrectomy was associated with an
increased complication rate and no improvement in 5-year survival compared with D1
gastrectomy. However, methodological issues with the studies included and the meta-
analysis itself mean that extended lymphadenectomy should not be abandoned, particu-
larly in the light of results from the Japanese literature. Studies included were historical,
dating from 1988, and most pre-dated the current best practice of neoadjuvant treatment.
There was also a great deal of heterogeneity between trials evidenced by differing prior
surgical experience of the D2 technique, variable use of routine pancreato-splenectomy,
and post-operative complication rates. One study compared D1 subtotal gastrectomy
with D2 total gastrectomy. Significant protocol breaches in at least two of the trials have
been reported and this is likely to have affected other studies. The overall quality of the
studies was low (mean Jadad score 2), and statistical heterogeneity and publication bias
were demonstrated. Nonetheless, this review shows that without sufficient experience D2
gastrectomy (particularly with pancreas/spleen resection) is likely to be associated with
a greater complication rate, which will mitigate any long-term benefit of the extended
dissection.
4.12 Minimally invasive surgery for oesophagogastric cancer

Details of studies
Minimally invasive oesophagogastric cancer resections aim to achieve the same results as
open resections, but with less morbidity and a quicker return to normal function. There
were initial concerns about the introduction of such techniques relating to the learning
curve of these procedures, the applicability in Western populations with more advanced
disease and obesity, and the ability to perform a radical lymphadenectomy. The study by
Luketich et al. was one of the first and largest to suggest that minimally invasive oesopha-
gectomy could be safely performed. Much of the data for laparoscopic gastrectomy came
from Asian populations. Husker et al.’s small RCT confirmed that laparoscopic gastrec-
tomy could be safely performed in Western populations.
Study references
Main studies
Luketich, J.D., Alvelo-Rivera, M., Buenaventura, P.O., et al. (2003) Minimally invasive esophagectomy:
outcomes in 222 patients. Annals of Surgery, 238, 486–95.
Husker, C.G., Mingoli, A., Sgarzini, G., et al. (2005) Laparoscopic versus open subtotal gastrectomy for
distal gastric cancer: five-year results of a randomized prospective trial. Annals of Surgery, 241, 232–7.
Related references
Smithers, B.M., Gotley, D.C., McEwan, D., Martin, I., Bessell, J., and Doyle, L. (2001) Thoracoscopic
mobilization of the esophagus. A 6-year experience. Surgical Endoscopy, 15, 176–82.
Palanivelu, C., Prakash, A., Senthilkumar, R., et al. (2006) Minimally invasive esophagectomy: thoraco-
scopic mobilization of the esophagus and mediastinal lymphadenectomy in prone position—experi-
ence of 130 patients. Journal of the American College of Surgeons, 203, 7–16.
Tanimura, S., Higashino, M., Fukunaga, Y., et al. (2007) Laparoscopic gastrectomy with regional lymph
node dissection for upper gastric cancer. British Journal of Surgery, 94, 204–7.
Berrisford, R.G., Wajed, S.A., Sanders, D., and Rucklidge, M.W. (2008) Short-term outcomes following
total minimally invasive oesophagectomy. British Journal of Surgery, 95, 602–10.
Minimally invasive oesophagectomy: Luketich et al.

Study design
◆ A prospective case series of 222 patients in a single high-volume centre in the USA.
Randomization Open versus laparoscopic subtotal gastrectomy
Inclusion criteria Pre-operative diagnosis of localized distal gastric cancer
Outcome measures Short-term outcomes; 5-year overall and disease-free survival
◆ Inclusion criteria: patients with either HGD or oesophageal cancer who were fit for
operation with a resectable lesion after evaluation with endoscopic ultrasound (EUS)
and CT.
MINIMALLY INVASIVE SURGERY FOR OESOPHAGOGASTRIC CANCER 141
◆ Laparoscopic trans-hiatal oesophagectomy (n = 8); three-stage minimally invasive

oesophagectomy (n = 214).
Results
◆ The conversion rate was 7.2%.
◆ The median ICU stay was 1 day (range 1–30 days); the median hospital stay was 7
days (range 3–75 days).
◆ Time to oral intake was 4 days (range 1–40 days).
◆ The 30-day mortality was 1.4% (n = 3).
◆ Morbidity included anastomotic leak (11.7%), atrial fibrillation (11.7%), pneumonia
(7.7%), pleural effusion (6.3%), chylothorax (3.2%), gastric tip necrosis (3.2%), and
vocal cord palsy (3.6%).
◆ SF-36, HRQoL, and dysphagia scores were similar to age-matched US norms.
◆ Stage-specific cancer survival was similar to open surgery series.
Conclusions
This single-centre series suggests that minimally invasive oesophagectomy offers results
as good as or better than an open operation. The mortality rate and hospital stay were
shorter than most open series. These results provide justification for a multi-centre RCT.
Laparoscopic gastrectomy: Husker et al.

Study design
◆ A prospective single-centre (single-surgeon) RCT
Results
◆ The two groups were well matched in terms of demographic characteristics. A similar
proportion of D1 and D2 resections were performed in each group.
Laparoscopic Open p-value

Operation time 196min 168min ns
Blood loss 229mL 391mL <0.001
Lymph nodes excised (mean) 30.0 33.4 ns
Post-operative mortality 3.3% 6.7% ns
Post-operative morbidity 26.7% 27.6% ns
Time to oral intake (mean) 5.1 days 7.4 days <0.001
Hospital stay (mean) 10.3 days 14.5 days <0.001
◆ Five-year overall and disease-free survival rates were similar.
Conclusions
Laparoscopic radical subtotal gastrectomy for distal gastric cancer is a feasible and safe
oncological procedure with short- and long-term results similar to those obtained with an
open approach. Additional benefits for laparoscopic gastrectomy are reduced blood loss,
shorter time to resumption of oral intake, and earlier discharge from hospital.
Critique
Minimally invasive surgery is generally associated with reduced post-operative pain, ear-
lier ambulation, and a quicker return to normal activities. Conversely, open resections for
oesophagogastric cancer are associated with significant morbidity, mortality, and delay
in regaining pre-operative quality of life. As a result, many oesophagogastric surgeons
have now adopted minimally invasive techniques. By 2009, 30% of all oesophagectomies
and 13% of all gastrectomies in the UK were performed by minimally invasive surgery
(National Oesophago-Gastric Cancer Audit).
A number of different approaches are available for minimally invasive oesophagectomy
including hybrid and hand-assisted techniques. Luketich ’s prospective case series showed
that a totally minimally invasive approach was safe and feasible in a single high-volume
centre. However, caution should be applied in the generalization of these results since they
come from a single very experienced unit. Other case series, such as those reported by
Smithers et al. and Palanivelu et al., have confirmed good short-term clinical outcomes.
It has also been shown that the lymph node yield from laparoscopic lymphadenectomy is
similar to that obtained with open surgery. However, some series have reported concerns
relating to thermal injuries to the airway and a higher than expected incidence of gastric
tube necrosis (Berrisford et al.).
A recent meta-analysis by Nagpal et al. included 12 studies with 672 patients undergo-
ing either minimally invasive oesophagectomy or hybrid minimally invasive oesophagec-
tomy, and 612 patients undergoing open oesophagectomy. The minimally invasive group
had lower blood loss, shorter hospital stay, and reduced respiratory complications and
total morbidity. There was no significant difference in 30-day mortality.
The evidence for laparoscopic or laparoscopic-assisted gastrectomy is predominantly
based upon T1 and T2 distal gastric cancer in Asian populations. The small RCT con-
ducted in Italy by Husker et al. clearly demonstrated many of the advantages of minimally
invasive surgery already seen in the RCTs from Asia, namely quicker return of gut func-
tion and shorter hospital stay. This trial also showed that it was possible to perform a D2
lymphadenectomy without any increase in morbidity and mortality compared with open
surgery. However, the trial had small numbers with no power calculation performed, and
all the surgeries (both laparoscopic and open) were performed by a single surgeon in a
single centre. Nevertheless, this is an important trial on laparoscopic gastrectomy in a
Western population.
A recent meta-analysis by Ohtani et al. included five RCTs comparing laparoscopic
and open distal gastrectomy for gastric cancer. Most of these trials had small sample sizes
and limited follow-up, with a total of 326 patients included in the analysis. There was a
significant difference in the volume of intra-operative blood loss, length of hospital stay,
frequency of analgesic administration, and rate of complications favouring the laparo-
scopic group. There was no difference in the resumption of oral intake, rate of tumour
MINIMALLY INVASIVE SURGERY FOR OESOPHAGOGASTRIC CANCER 143
recurrence, and mortality. The operative time was significantly longer and the number of
lymph nodes harvested was significantly smaller in the laparoscopic group.
Whilst it is clear that minimally invasive surgery for oesophagogastric cancer is both
safe and feasible, the evidence for the benefit of such surgery requires further develop-
ment. Data from well-designed RCTs is required to demonstrate the optimal approach.
National Oesophago-Gastric Audit, Third Annual Report, 2010. NHS Information Centre.<http://www.
ic.nhs.uk/og>
Nagpal, K., Ahmed, K., Vats, A., et al. (2010) Is minimally invasive surgery beneficial in the manage-
ment of esophageal cancer? A meta-analysis. Surgical Endoscopy, 24, 1621–9.
Ohtani, H., Tamamori, Y., Noguchi, K., et al. (2011) Meta-analysis of laparoscopy-assisted and open
distal gastrectomy for gastric cancer. Journal of Surgical Research, 171, 479–85.
4.13 Multimodality treatment for operable gastric cancer

Details of studies
Radical gastric cancer surgery in Western populations is associated with a significant risk
of disease relapse. These two studies form the main evidence base for the different treat-
ment approaches to multimodality treatment of gastric cancer in Western populations.
Study references
Main studies
MAGIC study
Cunningham, D., Allum, W.H., Stenning, S.P., MAGIC Trial participants, et al. (2006) Perioperative
chemotherapy versus surgery alone for resectable gastroesophageal cancer. New England Journal of
Medicine, 355, 11–20.
Macdonald, J.S., Smalley, S.R., Benedetti, J., et al. (2001) Chemoradiotherapy after surgery compared
with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. New England
Study design
◆ Both studies were prospective multi-centre RCTs.
MAGIC study
Randomization Six cycles of peri-operative ECF chemotherapy (three cycles pre-operative–
three cycles post-operative) and surgery versus surgery alone (CS vs S)
Inclusion criteria WHO performance status 0–1
Adenocarcinoma of the stomach or lower third of the oesophagus at least
through the submucosa but with no evidence of distant metastases or
locally advanced inoperable disease
Exclusion criteria Previous chemotherapy or radiotherapy
Uncontrolled cardiac disease
Creatinine clearance <60mL/min
Outcome measures Primary: overall survival
Secondary: progression-free survival, surgical and pathological assessments
of downstaging, surgical assessment of cure and quality of life
◆ Three 3-week cycles of combination epirubicin, cisplatin, and 5-fluorouracil (ECF)

chemotherapy were followed by a 3–6-week break prior to surgery. A further three
cycles of chemotherapy were given 6–12 weeks after surgery.
MacDonald et al.
◆ The treatment protocol consisted of combination 5-fluorouracil and leucovorin on
day 1 with chemoradiotherapy starting on day 28 (4500cGy in five fractions with
MULTIMODALITY TREATMENT FOR OPERABLE GASTRIC CANCER 145
Randomization Post-operative combination of fluorouracil + leucovorin + locoregional
radiation versus post-operative observation alone
Inclusion criteria Completely resected adenocarcinoma of the stomach or gastro-
oesophageal junction (stage IB-IVM0)
Performance status 2 or less
Adequate function of major organs
Adequate caloric intake
Registered for study within 41 days of surgery
Outcome measures Overall and relapse-free survival
concurrent chemotherapy) followed by two more cycles of chemotherapy 1 month

later
◆ Radiotherapy was delivered to the tumour bed, regional nodes, and 2cm beyond the
proximal and distal margins of resection. Quality assurance of radiotherapy fields
was performed.
Results
MAGIC study
◆ Only 104 patients (41.6%) completed all six cycles of chemotherapy.
◆ Among patients treated by radical surgery, resection was considered curative by the
operating surgeon in 169 of 213 patients (79.3%) in the CS group compared with 166
of 236 patients (70.3%) in the S group (p = 0.03).
◆ Post-operative morbidity, 30-day mortality, and hospital stay were similar in the two
groups.
◆ Pathological assessment revealed that the tumours in the CS group were significantly
smaller (p <0.001) with a higher proportion of T1/2 tumours (p = 0.002) and N0/1
staging (p = 0.01), all indicating a downstaging effect.
◆ The CS group had a significantly higher likelihood of progression-free survival
(hazard ratio (HR) for progression, 0.66; 95% CI, 0.53–0.81; p <0.001) and overall
survival (HR for death, 0.75; 95% CI, 0.60–0.93; p = 0.009).
◆ Five-year survival rates were 36.3% for patients in the CS group and 23.0% for those
in the S group.
MacDonald et al.
◆ The two groups were well matched in terms of demographic characteristics.
◆ A review of surgery performed revealed that 54% of patients had undergone a D0
resection and only 10% had undergone a D2 resection.
◆ 64% of patients in the chemoradiotherapy arm completed the post-operative

treatment. Haematological and gastrointestinal toxic effects occurred in 54% and
33% of patients, respectively.
◆ With a median follow-up period of 5 years, the median duration of survival was 36
months in the chemoradiotherapy group and 27 months in the surgery-only group.
HR for death in the surgery-only group, compared with the chemoradiotherapy
group, was 1.35 (95% CI, 1.09–1.66; p = 0.005).
◆ The median duration of relapse-free survival was 30 months in the
chemoradiotherapy group and 19 months in the surgery-only group. HR for relapse
in the surgery-only group, as compared with the chemoradiotherapy group, was 1.52
(95% CI, 1.23–1.86; p <0.001).
Conclusions
MAGIC study
In patients with operable gastric or lower oesophageal adenocarcinomas, a peri-oper-
ative regimen of ECF chemotherapy decreased tumour size and stage and significantly
improved progression-free and overall survival.
MacDonald et al.
Post-operative chemoradiotherapy should be considered for all patients at high risk for
recurrence of adenocarcinoma of the stomach or gastro-oesophageal junction who have
undergone curative resection.
Critique
The aim of systemic therapy in gastric cancer is to reduce the late patterns of failure
including nodal and liver metastases and peritoneal disease. Initial approaches centred
upon adjuvant chemotherapy, but the results from several trials using different chemo-
therapy regimes in both Western and Asian populations showed limited survival benefit,
limiting its widespread practice. A recent meta-analysis by the GASTRIC group of 3838
patients in 17 trials demonstrated an approximate 5% survival benefit.
The two trials described form the basis of current multimodality treatment for gastric
cancer. In Europe, the results of the MAGIC study (Cunningham et al.) have dictated prac-
tice. This was a well-conducted trial using a chemotherapy regime that had been shown
to be effective in advanced gastric cancer. The main criticism is that only 42% of patients
in the peri-operative chemotherapy group completed all protocol treatment. Despite this,
there was good evidence for a tumour downstaging effect, which translated into an esti-
mated improvement of 13 percentage points in the 5-year survival rate, corresponding to
a 25% reduction in the risk of death. A recent RCT from France in 223 patients using a
peri-operative chemotherapy based upon cisplatin and 5-fluorouracil (Ychou et al.) has
shown similar improvements in overall and disease-free survival.
MULTIMODALITY TREATMENT FOR OPERABLE GASTRIC CANCER 147
In North America, current practice is based upon the results of the trial by MacDonald
et al. The apparently toxic post-operative treatment regime was reasonably well tolerated.
Although haematological and gastrointestinal toxicity was commonly reported, treat-
ment-related mortality was low (1%). The main criticism of this trial is that a significant
proportion of patients (54%) had a D0 resection and the 3-year survival in the surgery-
alone group was relatively low (41%). It is conceivable that the adjuvant chemoradiother-
apy is making up for less than adequate surgery and these results may not be reproducible
when more extensive resections are performed. A follow-up paper by the same group
(Hundahl et al.) determined the risk of unresected regional nodal disease in each trial
patient and showed that surgical under-treatment had undermined survival in the trial.
There are still unanswered questions regarding optimal systemic therapy for gastric
cancer. Ongoing and proposed trials include the Intergroup study (Cancer and Leukemia
Group B 80101), which is assessing the role of a potentially more active post-operative
chemoradiotherapy regimen. The MRC STO3 study is examining the role of adding beva-
cizumab to peri-operative chemotherapy, and the planned CRITICS study by the Dutch
Gastric Cancer Group will evaluate the role of post-operative chemoradiation in combi-
nation with pre-operative chemotherapy.
GASTRIC (Global Advanced/Adjuvant Stomach Tumor Research International Collaboration) Group.
(2010) Benefit of adjuvant chemotherapy for resectable gastric cancer: a meta-analysis. JAMA, 303,
1729–37.
Ychou, M., Boige, V., Pignon, J.P., et al. (2011) Perioperative chemotherapy compared with surgery alone
for resectable gastroesophageal adenocarcinoma: an FNCLCC and FFCD multicenter phase III trial.
Journal of Clinical Oncology, 29, 1715–21.
Hundahl, S.A., Macdonald, J.S., Benedetti, J., et al. (2002) Surgical treatment variation in a prospective,
randomized trial of chemoradiotherapy in gastric cancer: the effect of undertreatment. Annals of
Surgical Oncology, 9, 278–86.
4.14 Pre-operative chemotherapy or chemoradiotherapy

for locally advanced oesophageal cancer
Details of studies
Long-term survival following surgery alone for locally advanced oesophageal cancer may
be as low as 15–20%. Several different strategies depending upon tumour location and
histology have evolved in an effort to improve these survival rates. These include pre-
operative (neoadjuvant), pre-operative and post-operative, and definitive non-surgical
treatments. The role of both pre-operative chemotherapy and chemoradiotherapy against
surgery alone has been examined in a number of RCTs.
Study references
Main studies
OEO2 trial
Medical Research Council Oesophageal Cancer Working Group. (2002) Surgical resection with or
without preoperative chemotherapy in oesophageal cancer: a randomised controlled trial. Lancet,
359, 1727–33.
Sjoquist, K.M., Burmeister, B.H., Smithers, B.M., et al. (2011) Survival after neoadjuvant chemotherapy
or chemoradiotherapy for resectable oesophageal carcinoma: an updated meta-analysis. Lancet
Oncology, 12, 681–92.
Related references
Kelsen, D., Ginsberg, R.J., Pajak, T.F., et al. (1998) Chemotherapy followed by surgery compared with
surgery alone for localized oesophageal cancer. New England Journal of Medicine, 339, 1979–84.
Allum, W.H., Stenning, S.P., Bancewicz, J., et al. (2009) Long-term results of a randomized trial of sur-
gery with or without preoperative chemotherapy in esophageal cancer. Journal of Clinical Oncology,
27, 5062–7.
Study design
OEO2 trial
◆ A multi-centre prospective RCT.
Randomization Pre-operative chemotherapy and surgical resection (CS) versus surgery
alone (S)
Inclusion criteria Histologically proven resectable cancer of the oesophagus or cardia
Exclusion criteria Cervical lymph node involvement/metastases
Previous or concomitant malignant disease
No contraindication to surgery or chemotherapy
Normal renal function
White cell count >3.5 x 109/L and platelets >100 x 109/L
Outcome measures Overall and disease-free survival
Dysphagia and performance status
PRE-OPERATIVE CHEMO- OR CHEMORADIOTHERAPY FOR LOCALLY ADVANCED OESOPHAGEAL CANCER 149
◆ Chemotherapy: two 4-day cycles of cisplatin 80mg/m2 by infusion 3 weeks apart plus
fluorouracil 1000mg/m2 daily by continuous infusion for 4 days.
◆ Participating centres could also decide to administer pre-operative radiotherapy to
all their trial patients.
◆ Local surgeons determined the appropriate surgical procedure.
Sjoquist et al.
◆ A meta-analysis of all RCTs comparing survival after surgery alone or in combination
with neoadjuvant chemoradiotherapy/chemotherapy in the initial management of
resectable oesophageal or oesophagogastric junction (OGJ) carcinoma.
◆ The primary outcome was all-cause mortality. The secondary endpoint was the effect
on all-cause mortality of treatment for each histological subtype.
◆ A secondary analysis compared survival benefits of neoadjuvant chemotherapy with
neoadjuvant chemoradiotherapy.
Results
OEO2
◆ The two groups were well matched with respect to baseline demographics.
CS group (n = 400) S group (n = 402) p-value

Complete resection 212 (60%) 187 (53%) <0.0001
Post-operative complications 42% 41% ns
30-day mortality 10% 10% ns
Median survival 512 days 405 days
2-year survival 43% 34%
◆ Tumours in the CS group were smaller (p = 0·0001), extended into surrounding

tissue less frequently, and showed less lymph node involvement.
◆ Overall survival was better in the CS group than in the S group (p = 0·004; HR 0·79;
95% CI 0·67–0·93); estimated reduction in risk of death was 21%.
◆ Disease free survival was also better in the CS group than in the S group (p = 0·0014;
HR 0·75; 95% CI 0·63–0·89).
◆ There were no differences in dysphagia scores or World Health Organization (WHO)
performance status between the two groups at 1 year.
Sjoquist et al.
◆ Twenty-four RCTs with a total of 4188 patients were included in the analysis.
◆ There was no evidence of publication bias or significant heterogeneity.
◆ Thirteen studies were included in the comparison of neoadjuvant chemoradiotherapy

with surgery alone (n = 1932).
HR 95% CI p-value
All-cause mortality 0.78 0.70–0.88 <0.0001
Squamous carcinoma 0.80 0.68–0.93 0.004
Adenocarcinoma 0.75 0.59–0.95 0.02
◆ Absolute survival benefit at 2 years was 8.7% (NNT = 11).

◆ Ten studies were included in the comparison of neoadjuvant chemotherapy with
surgery alone (n = 2062).
HR 95% CI p-value
All-cause mortality 0.87 0.79–0.96 0.005
Squamous carcinoma 0.92 0.81–1.04 0.18
Adenocarcinoma 0.83 0.71–0.95 0.01
◆ Absolute survival benefit at 2 years was 5.1% (NNT = 19).

◆ The survival benefit of neoadjuvant chemoradiotherapy versus neoadjuvant
chemotherapy was calculated in 2220 patients. The overall HR was 0.88 (0.76–1.01;
p=0.07) in favour of neoadjuvant chemoradiotherapy. There was no evidence to
suggest that the potential benefit of the neoadjuvant treatment regimens was offset by
a higher post-operative mortality rate.
Conclusions
OEO2 trial
Two cycles of pre-operative cisplatin and fluorouracil improve survival without additional
serious adverse events in the treatment of patients with resectable oesophageal cancer.
Sjoquist et al.
This meta-analysis provides strong evidence for a survival benefit of neoadjuvant chemo-
radiotherapy or chemotherapy over surgery alone in patients with oesophageal carcinoma.
A clear advantage of neoadjuvant chemoradiotherapy over neoadjuvant chemotherapy
could not be established.
Critique
Most patients who undergo radical resection for oesophageal cancer will eventually
relapse and die as a result of their disease. Because of difficulties in administering chemo-
therapy or radiotherapy soon after a surgical procedure, high peri-operative morbidity,
and the disappointing results from studies of adjuvant treatment, most trials have focused
on pre-operative treatment. Whilst Sjoquist et al. provide the most up-to-date analysis
PRE-OPERATIVE CHEMO- OR CHEMORADIOTHERAPY FOR LOCALLY ADVANCED OESOPHAGEAL CANCER 151
of pre-operative treatments, the OEO2 trial, together with the MAGIC study which also
included adenocarcinoma of the oesophagogastric junction (see section 4.12), is com-
monly used in many centres as a justification for pre-operative treatment, with the exclu-
sion of radiotherapy. A good overview of all multimodality approaches can be found in
the latest AUGIS/BSG/BASO guidelines (Allum et al.).
Until the late 1990s, most of the RCTs examining pre-operative chemotherapy had
small sample sizes and were insufficiently powered to show significant survival differ-
ences. The OEO2 trial was one of the first large adequately powered RCTs conducted for
pre-operative chemotherapy. This was a pragmatic multi-centre trial and it demonstrated
a survival benefit for pre-operative chemotherapy. Updated results confirmed an ongo-
ing overall and disease-free survival benefit. Five-year survival with surgery alone was
17%, compared with 23% with pre-operative therapy (Allum et al.). The OEO2 trial can
be criticized because the initial staging investigations would be regarded as suboptimal
by modern standards, radiotherapy was allowed as part of the pre-operative treatment,
and there was little surgical quality assurance. The results of the OEO2 trial also con-
flicted with the results of the Intergroup-0113 trial (Kelsen et al.) which reported at about
the same time. The Intergroup-0113 study randomized 467 patients to surgery alone or
three cycles of chemotherapy, followed by surgery and post-operative chemotherapy in
responders. This trial did not report a significant difference in median overall survival.
Reasons for this include the excessive toxicity of the chemotherapy, which fewer patients
completed, as well as a delay in definitive surgery in patients not responding to treatment.
Sjoquist et al. provide the most up-to-date analysis of pre-operative chemotherapy and
chemoradiotherapy. The evidence supported the use of either pre-operative chemora-
diotherapy or chemotherapy for oesophageal adenocarcinoma. The survival benefit for
oesophageal squamous cell carcinoma was only significant for pre-operative chemoradio-
therapy. The indirect comparison of pre-operative chemoradiotherapy with pre-operative
chemotherapy showed weak evidence in favour of pre-operative chemoradiotherapy. The
meta-analysis included over 4000 patients and publication bias was minimized. However,
many of the trials were small sample sizes from over 20 years ago when the methods for
diagnosis, staging, treatment delivery, and outcome measurement were less advanced.
Combining studies with different doses and scheduling potentially weakens the results,
although there was no evidence of significant heterogeneity. There were only two direct
comparisons of pre-operative chemotherapy and chemoradiotherapy, so most of the data
were obtained from indirect comparison of pooled outcomes. Furthermore, assumptions
were made to estimate hazard ratios in some of the trial results.
Allum, W.H., Blazeby, J.M., Griffin, S.M., et al. (2011) Guidelines for the management of oesophageal
and gastric cancer. Gut, 60, 1449–72.
4.15 Palliation of dysphagia in advanced oesophageal cancer

Details of studies
The majority of patients diagnosed with oesophageal cancer receive only palliative treat-
ment. A variety of methods are available to palliate dysphagia: self-expanding metal stents
(SEMS), external beam radiotherapy or intraluminal brachytherapy, and endoscopic
techniques such as Nd:Yag laser, argon plasma coagulation, or photodynamic therapy.
The evidence base comparing these methods is limited. In a study by Dallal et al., laser
treatment is directly compared with SEMS in terms of dysphagia, quality of life, and cost.
In a more recent trial, Homs et al. chose comparison of SEMS with brachytherapy.
Study references
Main studies
Dallal, H.J., Smith, G.D., Grieve, D.C., et al. (2001) A randomized trial of thermal ablative therapy
versus expandable metal stents in the palliative treatment of patients with esophageal carcinoma.
Gastrointestinal Endoscopy, 54, 549–57.
Homs, M.Y., Steyerberg, E.W., Eijkenboom, W.M., et al. (2004) Single-dose brachytherapy versus metal
stent placement for the palliation of dysphagia from oesophageal cancer: multicentre randomised
trial. Lancet, 364, 1497–504.
Related reference
Adam, A., Ellul, J., Watkinson, A.F., et al. (1997) Palliation of inoperable esophageal carcinoma: a pro-
spective randomized trial of laser therapy and stent placement. Radiology, 202, 344–8.
Study design
Dallal et al.
◆ A prospective single-centre RCT.
Randomization SEMS versus laser therapy
Inclusion criteria Patients with inoperable oesophageal/oesophagogastric cancer and
dysphagia
Exclusion criteria Tumour too close to upper oesophageal sphincter
Extremely short life expectancy
Tracheal involvement
Outcome measures Dysphagia score, complications, re-intervention, HRQoL, and cost
◆ A median of 2.5 (0-15) sessions of Nd:Yag laser therapy were conducted.

◆ Ultraflex, Streker, or Wallstent self-expanding stents were used (covered if a fistula
was suspected).
PALLIATION OF DYSPHAGIA IN ADVANCED OESOPHAGEAL CANCER 153
Homs et al.
Randomization SEMS versus brachytherapy
Inclusion criteria Patients with dysphagia secondary to oesophageal (including OGJ) cancer,
not suitable for radical treatment becausee of fitness or stage
Exclusion criteria Tumour length >12cm or tumour growth <3cm from the upper
oesophageal sphincter
Previous radiation treatment or stent placement
Tracheal involvement
Outcome measures Dysphagia score, complications, treatment for persistent or recurrent
dysphagia, HRQoL, and cost
Results
Dallal et al.
◆ The median dysphagia score at 1 month was unchanged in both groups.
◆ 32% of patients in the laser group reported normal swallowing at 1 month compared
with none of the patients in the stent group.
◆ There was no significant difference in re-interventions or post-procedure
complications between the stent and laser groups
◆ Median survival was increased in the laser group (125 versus 68 days, p <0.05)
◆ Median hospital stay, hospital admissions, and overall cost were significantly higher
in the laser group.
◆ Patients in the stent group had a poorer quality of life and pain had deteriorated
more at 1 month.
Homs et al.
◆ The dysphagia score improved more rapidly after stent placement than after
brachytherapy. By 30 days there was no difference between the two groups.
◆ Stent placement had more complications than brachytherapy (33% versus 21%,
p = 0.02), mainly due to an increased incidence of late haemorrhage.
◆ 43% of brachytherapy patients and 40% of stent patients had persistent or recurrent
dysphagia, with 47% of brachytherapy patients subsequently requiring a stent.
◆ Total medical costs and survival were similar between the groups.
◆ Quality-of-life scores favoured brachytherapy compared with stent placement.
Conclusions
Dallal et al.
Laser therapy provides a more satisfactory swallow and better quality of life than SEMS,
but at increased cost and patient stay.
Homs et al.
Brachytherapy should be used as first-line palliation in patients with oesophageal cancer.
Critique
The variety of techniques available for palliating dysphagia in advanced oesophageal can-
cer suggests that no single treatment is superior or applicable in all clinical situations.
SEMS produce a more rapid palliation of dysphagia and are cost-effective, but patients
do not regain a normal swallow and complication rates are high. Brachytherapy is still a
limited resource and does not produce effective palliation in the short term (almost half
the patients receiving brachytherapy required later stenting in the trial by Homs et al.).
This makes brachytherapy less applicable for patients with severe dysphagia or short life
expectancy. Dysphagia and quality of life are improved in the longer term with lower
complication rates. These results were supported by a subsequent trial by Berquist et al.
The paper by Dallal et al. emphasizes the poor outcome in palliation of oesophageal
cancer. Although neither group demonstrated an improvement in dysphagia score, more
patients reported a normal swallow, and QOL and pain improved more in those receiving
laser. The additional cost of laser in this study could have been reduced by day-case treat-
ment in line with current practice. Survival was increased in the laser group although this
may have been an anomaly due to the small patient numbers or differences in adjuvant
therapies between the groups.
A more recent pragmatic trial by Shenfine et al. examined the clinical and cost-effec-
tiveness of SEMS with a range of other palliative therapies. This confirmed that plastic
stents were associated with a worse quality of swallowing and increased late morbidity.
Small-diameter (18mm) SEMS were as effective as large (24mm) SEMS and induced less
pain. There was no difference in the cost or effectiveness between stent and non-stent
treatments.
What can be concluded from these studies? In patients with trachea-oesophageal fis-
tula, a short life expectancy, or severe dysphagia secondary to circumferential malignancy,
stenting would seem to be appropriate. In other patients, laser therapy or brachytherapy
may be a more appropriate first-line therapy, accepting that further treatments and stent-
ing may be required.
Shenfine, J., McNamee, P., Steen, N et al. (2009) A randomized controlled clinical trial of palliative
therapies for patients with inoperable esophageal cancer. American Journal of Gastroenterology, 104,
1674–85.
Bergquist, H., Wenger, U., Johnsson, E., et al. (2005) Stent insertion or endoluminal brachytherapy
as palliation of patients with advanced cancer of the esophagus and gastroesophageal junction.
Diseases of the Esophagus, 18, 131–9.
Chapter 5
Hepatopancreaticobiliary surgery
5.1 The evolution of laparoscopic cholecystectomy 156
5.2 Clinical trials of laparoscopic cholecystectomy 160
5.3 Biliary injury during laparoscopic cholecystectomy 163
5.4 Early versus late laparoscopic cholecystectomy in acute cholecystitis 168
5.5 Enteral versus parenteral nutrition in acute pancreatitis 171
5.6 Prophylactic antibiotics for severe acute necrotizing pancreatitis 175
5.7 Early versus late necrosectomy in severe necrotizing pancreatitis 178
5.8 Minimally invasive versus open necrosectomy for necrotizing pancreatitis 181
5.9 Endoscopic versus surgical therapy for chronic pancreatitis 184
5.10 Diagnosis of autoimmune pancreatitis 188
5.11 Pre-operative biliary drainage for head of pancreas cancer 191
5.12 Adjuvant therapy for pancreatic cancer 195
5.13 Management of intraductal papillary mucinous neoplasm 198
5.14 Management of neuroendocrine tumours 201
5.15 Surgical resection of colorectal liver metastases 205
156 HEPATOPANCREATICOBILIARY SURGERY
5.1 The evolution of laparoscopic cholecystectomy

Details of studies
The first acknowledged open cholecystectomy was performed in Germany in 1882 by
Carl Langenbuch. Over 100 years later the first ‘laparoscopic cholecystectomy’ was per-
formed by Professor Dr Erich Mühe in Boblingen, Germany, in 1985. The procedure had
only a small number of technical elements that would constitute modern laparoscopic
cholecystectomy (LC), as later Mühe stopped using a pneumoperitoneum or optical guid-
ance and only a single incision was made in the subcostal region.
The first LC in its current form can be reasonably attributed to Phillipe Mouret, a gyn-
aecologist from Lyon. In 1987 he operated on a 50-year-old female suffering from pelvic
adhesions and symptomatic gallstones. Having performed over 8000 previous laparosco-
pies he was able to perform an LC that was later detailed in the first published description
of the technique. This publication was submitted as a case series by François Dubois, who
copied Mouret’s procedure after being shown a videotape of the procedure during a meet-
ing with Mouret.
Study references
Main study
Dubois, F., Icard, P., Berthelot, G., and Levard, H. (1990) Coelioscopic cholecystectomy. Preliminary
report of 36 cases. Annals of Surgery, 211, 60–2.
Related references
Mühe, E. (1992) Long-term follow-up after laparoscopic cholecystectomy. Endoscopy, 24, 754–8.
Traverso, L.W. (1976) Carl Langenbuch and the first cholecystectomy. American Journal of Surgery, 132,
81–2.
Study design
◆ Case series
Level of evidence 4
Setting University hospital in Paris between May and December 1988
◆ A pre-operative IV cholangiogram was performed in all patients. No common bile

duct (CBD) stones were identified, but three patients had cystic duct stones.
Technique
◆ General anaesthesia.
◆ A nasogastric tube was inserted to prevent gastric distension.
◆ The patient was ‘horizontally’ placed in a double-access position and could be
brought into Trendelenburg’s position.
EVOLUTION OF LAPAROSCOPIC CHOLECYSTECTOMY 157
Equipment
◆ 10mm cold–light optic and video set.
◆ CO2 insufflator (set to 14mmHg).
◆ Suction/irrigation.
◆ 2 10mm trocars with ‘trumpet sheaths’ and 2 5mm trocars.
◆ Scissors with diathermy, atraumatic grasping forceps, ‘craw’ grasping forceps (to
grasp and extract gall bladder), coagulating forceps, and a clip applicator (8mm
titane clips).
Procedure
◆ Port placements included a 10mm umbilical port for the ‘endoscope’, a 5mm
subcostal port for an atraumatic grasper to catch the gall bladder, a 5mm epigastric
port for suction/irrigation (also used for a liver retractor), and a 10mm left upper
quadrant (LUQ) port for scissors/hook/clipper (see Fig. 5.1).
◆ A needle was used to empty the gall bladder if necessary.
◆ Retrograde cholecystectomy was performed with clips applied to the cystic artery
and duct (see Fig. 5.2) and removal through the umbilicus (with extended incision
for large stones).
◆ Irrigation was routinely performed and the skin was closed with clips.
Results
◆ Two complications were reported.
Fig. 5.1 Abdominal insertion

areas: endoscope; atraumatic
3 grasping forceps; irrigation/
2 suction; scissors/hook/clip.
4 Reproduced with permission
from F. Dubois, P. Icard, G.
1 Berthelot et al, (1990)
Coelioscopic Cholecystectomy:
A Preliminary Report of 36 cases,
Annals of Surgery 211, 1,
© Wolters Kluwer Health.
Aquapurator as a retractor
on the liver Clip on the artery
Forceps on
the gallbladder
Clip on the cystic duct
Scissors
Fig. 5.2 Endoscopic aspect of the biliary pedicle.

Reproduced with permission from F. Dubois, P. Icard, G. Berthelot et al. (1990) Coelioscopic
cholecystectomy: a preliminary report of 36 cases, Annals of Surgery, 211, 1. © Wolters Kluwer
Health.
● Intra-operative cystic artery bleed requiring ‘mini-laparotomy’ for control.

● Post-operative pain at day 48 due to a sub-hepatic collection which required
drainage. The collection contained bile and was related to the gall bladder bed
(no cystic duct leak) with no further collections/drainage required after initial
aspiration.
◆ 34 patients were pain free at 48 hours and all were discharged between days 3 and
7 post-operatively.
Critique
Although the technique had been described previously (see Details of studies), this was
the first case series demonstrating that laparoscopic cholecystectomy was both feasible
and safe. The surgical technique has been refined over the intervening period and there
are several points worthy of discussion.
The process of establishing a pneumoperitoneum with a Veress needle has been largely
replaced with the open cut-down, or Hasson, technique in an effort to avoid iatrogenic
injury to underlying structures. Port site placement has also evolved, with the right lateral
port for grasping the gall bladder fundus replacing the left abdominal port. This allows
two grasping forceps to be applied to the gall bladder, an important part of displaying the
anatomy during dissection.
This leads us to perhaps the most significant modification of the original description
of this procedure. This paper does not mention bile duct injury (BDI) or the steps which
may be taken to prevent it. The footnote in the paper comments on a total of 220 LCs
being performed with no complications in the last 180, although the follow-up period and
EVOLUTION OF LAPAROSCOPIC CHOLECYSTECTOMY 159
process are not described. Accepted safe practice would now involve establishing a large
‘posterior window’ or ‘critical view of safety’ as described by Strasberg in 1995.
The length of hospital stay at 3–7 days is clearly longer than would be expected in the
modern era, but the authors explain this with reference to their institutional policy, not-
ing that all patients could have been potentially discharged prior to this. Nevertheless, this
is a groundbreaking paper describing a surgical technique that, rightly or otherwise, has
emerged as the standard of care, often being performed as a day-case procedure. Both the
conversion rate and the incidence of immediate post-operative complications in this case
series are acceptable for an operation very much in its infancy.
Strasberg, S.M., Hertl, M., and Soper, N.J. (1995) An analysis of the problem of biliary injury during
laparoscopic cholecystectomy. Journal of the American College of Surgeons, 180, 101–25.
5.2 Clinical trials of laparoscopic cholecystectomy

Details of studies
Laparoscopic cholecystectomy rapidly replaced open cholecystectomy as the standard
treatment for symptomatic gallstones following publication of the first case report in
1987. Although it was widely felt that the new technique was associated with a reduced
length of stay and faster return to normal activity, it also quickly became apparent that
there appeared to be an increased incidence of more serious complications including BDI,
vascular injury, and bowel perforation. As a consequence, a large number of studies were
conducted comparing laparoscopic cholecystectomy with standard open cholecystectomy
and also mini-laparotomy cholecystectomy. The later procedure was performed through
a small (<8cm) transverse subcostal incision and also appeared to result in a faster post-
operative recovery compared with the standard open operation.
The Cochrane review discussed here summarizes the findings of three separate system-
atic reviews comparing laparoscopic versus small-incision (mini-laparotomy) cholecys-
tectomy and small-incision versus open cholecystectomy.
Study references
Main study
Keus, F., Gooszen, H.G., and van Laarhoven, C.J. (2010) Open, small-incision, or laparoscopic cholecys-
tectomy for patients with symptomatic cholecystolithiasis. An overview of Cochrane Hepato-Biliary
Group reviews. Cochrane Database of Systematic Reviews, 1, CD008318. DOI: 10.1002/14651858.
Related references
Reynolds, W. (2001) The first laparoscopic cholecystectomy. Journal of the Society of Laparoendoscopic
Surgeons, 5, 89–94.
McMahon, A.J., Russell, I.T., Baxter, J.N., et al. (1994) Laparoscopic versus minilaparotomy cholecystec-
tomy: a randomised trial. Lancet, 343, 135–8.
Study design
◆ Systematic review with meta-analysis
Randomization Laparoscopic vs small-incision vs open cholecystectomy
Inclusion criteria Adults presenting with symptomatic gallstones
Exclusion criteria Acute cholecystitis
◆ The design and quality of each study was analysed and documented
CLINICAL TRIALS OF LAPAROSCOPIC CHOLECYSTECTOMY 161
Outcome measures
Primary endpoints
◆ Mortality.
◆ Complications.
◆ Symptom relief.
Secondary endpoints
◆ Operative time, hospital stay, convalescence.
Results
Laparoscopic versus small-incision cholecystectomy
◆ The meta-analysis included 13 trials which had recruited 2337 patients.
◆ The complication rate for both laparoscopic and small-incision cholecystectomy was
relatively high (17%), although not statistically different.
◆ The overall risk of bias was felt to be low.
◆ There was no significant difference in length of hospital stay and convalescence.
◆ Small-incision cholecystectomy resulted in a shorter operation time in trials where
there was a low risk of bias, whereas there was no difference in trials where there was
a high risk of bias.
Laparoscopic versus open cholecystectomy

◆ The meta-analysis included 38 trials which had recruited 2338 patients.
◆ The risk of bias was high.
◆ Laparoscopic cholecystectomy patients had a shorter hospital stay (mean difference,
random effects model −3 days [95% CI, −3.9 to −2.3]) and shorter convalescence
(mean difference, random effects model –22.5 days [95% CI, –36.9 to –8.1])
compared with open cholecystectomy.
◆ There was no significant difference in mortality, complications, or procedure time
Small-incision versus open cholecystectomy

◆ The meta-analysis included seven trials which had recruited 571 patients.
◆ The bias risk was high.
◆ Small-incision cholecystectomy patients had a shorter hospital stay (mean difference,
random effects model –2.8 days [95% CI, –4.9 to –0.6) compared with open
cholecystectomy.
◆ There was no significant difference in mortality, complications, or procedure time.
Conclusions
This meta-analysis showed no difference in mortality or complications between open,
laparoscopic, and small-incision cholecystectomy. However, laparoscopic and small-
incision cholecystectomy are associated with a shorter hospital stay and faster recovery.
Small-incision cholecystectomy is less expensive (McMahon et al).
Critique
This Cochrane review provides good evidence that standard-incision, small-incision, and
laparoscopic cholecystectomy have similar outcomes in terms of mortality and complica-
tions. It is interesting to note that although small-incision cholecystectomy has similar
outcomes in terms of length of stay and recovery time, for the most part it has been aban-
doned in favour of the more expensive laparoscopic procedure. The precise reasons for
this are unclear, although patient preference and surgeon satisfaction with performing the
laparoscopic procedure are likely to be important factors.
BILIARY INJURY DURING LAPAROSCOPIC CHOLECYSTECTOMY 163
5.3 Biliary injury during laparoscopic cholecystectomy

Details of studies
Compared with open cholecystectomy, the incidence of injuries to the bile duct seemed to
be increased with laparoscopic cholecystectomy. Whilst this was not initially predictable,
in order to prevent a potential enduring problem a review of the literature to classify and
prevent BDI was performed by Strasberg et al. and published in January 1995. This review
has become the cornerstone of safe teaching in LC, with the figured critical view of safety
widely accepted in surgical practice.
Study references
Main study
Strasberg, S.M., Hertl, M., and Soper, N.J. (1995) An analysis of the problem of biliary injury during
laparoscopic cholecystectomy. Journal of the American College of Surgeons, 180, 101–25.
Related references
Strasberg, S.M. (2002) Avoidance of biliary injury during laparoscopic cholecystectomy. Journal of
Hepato-Biliary-Pancreatic Surgery, 9, 543–7.
Study design
Literature review with personal perspectives and a plan of answering the following eight
specific questions.
What is the incidence of biliary injury in LC, and how does it compare to
open cholecystectomy?
◆ Open cholecystectomy: numerous reports quote the rate of biliary injury as 0.3%
or less (with some reporting 0%), but a single State-wide database study (c.1993) of
approximately 4000 cholecystectomies over six months revealed a rate of 0.125%
(0.075% major injury).
◆ LC: numerous institutional case series report rates from 0 to 2.4%, but when
considering those with complete data collection the rates were at the higher end with
reports of 0.7% (0.5% major) and 1.3% (0.53% major). State-wide database trawls
revealed rates of 0.55% bile duct injuries (0.4% major) in New York (c.1993) and
0.32% major bile duct injuries in Orlando (c.1993).
What types of injury occur and how should they be classified?

◆ Strasberg proposed a classification to include bilomas, bile leaks, and the Bismuth
classification of major injuries.
● Type A: Bile leak from a minor duct still in continuity with the common bile duct.
Usually cystic duct or liver bed leaks and usually resolved by biliary stenting.
● Type B: Occlusion of part of biliary tree. Commonly injury to aberrant right
hepatic duct during misidentification. Often asymptomatic with liver atrophy and
contralateral compensatory hyperplasia.
● Type C: Bile leak from duct not in communication with CBD. Almost always
transection of aberrant right hepatic duct resulting in biliary peritonitis.
● Type D: Lateral injury to extra-hepatic bile ducts. The biliary tree remains in
continuity but, unlike type A, type D often requires laparotomy with possible later
stenosis. The role of intentional choledochotomy for operative cholangiogram is
not considered an injury unless other injury is noted or later stricturing occurs.
● Type E (Bismuth 1–5): Circumferential injury of major bile ducts. Further
subclassification into loss of a ductal segment or stenoses helps further guide
management where non-operative options exist.
◆ The majority of reported injuries are types A (33–65%) and D (up to 77%).
What are the risk factors for biliary injury?

◆ Training and experience: a ‘learning curve’ effect was thought to exist, with many
reports stating that most injuries occurred in the first 13 procedures. Subsequent
data have shown that more injuries occurred after 50 LCs were performed. Therefore
more data are required and other factors are felt to be of more importance.
◆ Local operative factors: injuries are more likely during difficult procedures and in
acute cholecystitis (Russell et al. reported a 0.51% injury rate). Other factors include
chronic inflammation with dense scarring, operative bleeding or fat in the portal
area.
◆ Aberrant anatomy: the aberrant right hepatic duct anomaly is the most common, and
may be under-reported in the often asymptomatic type B injury
◆ Equipment: whilst few surgeons now use lasers, thermal injuries are a concern and
vigilance should be maintained to focal loss of instrument insulation.
What are the direct causes of laparoscopic biliary injury?

◆ Misidentification of bile ducts as cystic ducts: either the CBD is mistaken for a cystic
duct or an aberrant right duct as a cystic duct. This can results in the ‘classical’ injury
(Davidoff et al.) where the CBD receives all three clips. Further excision of the gall
bladder requires division of the common hepatic duct (CHD). Associated right
hepatic arteries are not uncommon.
◆ Technical causes: failure to occlude the cystic duct securely, too deep a plane of
dissection on the liver bed, a tenting injury of the cystic duct, injudicious use of clips
to control bleeding, and injuries due to improper techniques of ductal exploration.
How to prevent biliary injuries during laparoscopic cholecystectomy?

◆ General: LC should be performed by trained surgeons, avoiding difficult cases early
in training.
◆ Avoidance of misidentification of ducts: the key issue in misidentification is failure
to conclusively identify the cystic structures before clipping. Principles are: 1) Calot’s
triangle must be dissected free of fat, fibrous tissue, and areolar tissue, and the lower
Fig. 5.3 The critical view of

safety.
Reproduced with permission
from Strasberg, S., Eagon, C.,
and Drebin, J. (2000) The
‘hidden cyctis duct’ syndrome
and the infundibular tech-
nique of laparoscopic chole-
cystectomy—the danger of
the false infundibulum.
Journal of the American
College of Surgeons, 191, 6.
© Elsevier.
end of gall bladder dissected off the liver (essential to preclude abberant injury);
2) there should only be two structures entering the gall bladder; 3) the bottom of
the liver bed should be visible. Identification of the CBD is not necessary. At this
point you would have the ‘critical view of safety’ (see Fig. 5.3). Failure to achieve
this should warrant an operative cholangiogram or open conversion to define
ductal anatomy. The only exception is identification of a clearly pulsatile structure,
in which case only one other structure should be identified. Routine operative
cholangiography has not been shown to prevent biliary injury, although it may
increase the chance of intra-operative recognition.
◆ Prevention of ductal injury as a result of technical causes includes failure of
occlusion of the cystic duct by clips (ensure three clips, no further manipulation
after application, and ligature use for large thick ducts), entering a duct in the liver
bed (stay within the correct tissue plane using meticulous dissection, cautery, and
irrigation), cautery injuries (use low cautery in portal dissections of 25W or less,
do not use diathermy when dividing the cystic duct, and brisk bleeding requires
conversion).
How do biliary injuries present and how should they be investigated?
◆ Clinical presentation: type A injuries tend to present in the first week with 1)
symptoms of general or localized biliary peritonitis, 2) external bile leakage (i.e.
through an incision), or 3) vague symptoms of anorexia and failure to thrive. Type E
injuries are usually noted intra-operatively (25%) or otherwise with painful/painless
jaundice. Types C and D often present like type A patients.
◆ Investigation: in the event of pain, fever, and sepsis (but not jaundice) proceed to CT
or USS to establish the presence of bile. If bile is found, hepatobiliary scintigraphy,
or recently magnetic resonance cholangiopancreatography (MRCP), should help
identify the injury. Endoscopic retrograde cholangiopancreatography (ERCP) and
percutaneous trans-hepatic cholangiography (PTC) are therapeutic.
What is the correct management of biliary injuries?

◆ Management of injuries recognized at the initial operation: commonly type C, D,
or E injuries where open conversion is indicated; most would advocate early repair.
This is often difficult requiring dissection of the hepatic ducts up to at least the
bifurcation and then formation of a hepatico-jejunostomy, or occasionally a duct-to-
duct anastomosis. Given the technical requirements, simple drainage and referral to
a hepatobiliary unit is considered safe, although lateral bile duct injuries can often be
repaired over a T-tube.
◆ Management of biliary injuries in the post-operative period and management of
patients referred with biliary injuries.
● Type A: bile drainage and ERCP (sphincterotomy and/or stenting) with PTC if
failed ERCP or high injury.
● Type B: symptomatic patients often require hepatico-jejunostomy or segmental
hepatic resection. If asymptomatic, a conservative approach should be considered.
● Type C: drainage of bile with either biliary-enteric anastomosis or ligation of the
transected duct
● Type D: ERCP and stenting have been successful. T-tube drainage (through the
same or a second choledochotomy) can be considered if injury noted soon after.
Smaller injuries can be considered for primary closure.
● Type E: Strictures and clip occlusions can be treated by non-operative means
(balloon dilatation and stents). In stent failure or ductal discontinuity, an operation
is required. Hepatico-jejunostomy is used in almost all cases.
◆ The optimal timing of operative repair is uncertain. However, a period of
‘stabilization’ is suggested to give the best chance of success.
What is the long-term outcome of treatment?

◆ Most surgical series report good short-term results. There is a progressive re-stenosis
rate of 5–28%. At the time of publication, there was limited available evidence on the
subject.
Conclusions
Biliary injury is the most significant problem in LC. The key is prevention, where meticu-
lous dissection allows only structures that have been unequivocally and conclusively iden-
tified to be divided.
Critique
This is an excellent review of bile duct injury during laparoscopic cholecystectomy. It
provides a comprehensive and detailed description of the incidence, classification, recog-
nition, management, and prognosis of these injuries. More importantly, this was the first
paper to describe the risk factors for BDI and, crucially, how to avoid it. The ‘critical view
of safety’ remains standard teaching for safe laparoscopic cholecystectomy. Management

algorithms are described for each type of injury, and this paper brings clarity and focus to
this relatively uncommon, but potentially devastating, iatrogenic surgical complication.
A more recent review of BDI from The Netherlands in 2010 recommends the continued
use of the critical view of safety. In addition, it suggests that intra-operative cholangiog-
raphy or laparoscopic ultrasound assessment of the biliary anatomy should be performed
routinely. The data to support these additional techniques to prevent BDI are currently
lacking. Strasberg’s technique remains the gold standard.
Buddingh, K.T., Nieuwenhuijs, V.B., van Buuren, L., Hulscher, J.B., de Jong, J.S., and van Dam, G.M.
(2011) Intraoperative assessment of biliary anatomy for prevention of bile duct injury: a review of
current and future patient safety interventions. Surgical Endoscopy, 25, 2449–61.
Russell, J.C., Walsh, S.J., Mattie, A.S., and Lynch, J.T. (1996) Bile duct injuries, 1989–1993: A statewide
experience. Archives of Surgery, 131, 382–8.
Davidoff, A.M., Pappas, T.N., Murray, E.A., et al. (1992) Mechanisms of major biliary injury during
laparoscopic cholecystectomy. Annals of Surgery, 215, 196–202.
5.4 Early versus late laparoscopic cholecystectomy in

acute cholecystitis
Details of studies
LC in the management of early acute cholecystitis has caused concern over disseminating
infection, increased incidence of BDI, and increased conversion to an open procedure. A
number of RCTs have been published comparing early versus late LC. Three meta-analy-
ses have now been published, one of which is an update of a Cochrane review (Gurusamy
2006) and is described in this section.
Study references
Main study
Gurusamy, K., Samraj, K., Gluud, C., Wilson, E., and Davidson, B.R. (2010) Meta-analysis of random-
ized controlled trials on the safety and effectiveness of early versus delayed laparoscopic cholecystec-
tomy for acute cholecystitis. British Journal of Surgery, 97, 141–50.
Included trials
Lo, C., Liu, C., Fan, S.T., Lai, E.C., and Wong, J. (1998) Prospective randomized study of early versus
delayed laparoscopic cholecystectomy for acute cholecystitis. Annals of Surgery, 227, 461–7.
Dávila, D., Manzanares, C., Pichó, M.L., et al. (1999) Experience in the treatment (early vs. delayed) of
acute cholecystitis via laparoscopy. Cirugía Española, 66, 233.
Johansson, M., Thune, A., Blomqvist, A., Nelvin, L., and Lundel, L. (2003) Management of acute
cholecystitis in the laparoscopic era: results of a prospective, randomized clinical trial. Journal of
Gastrointestinal Surgery, 7, 642–5.
Kolla, S.B., Aggarwal, S., Kumar, A., et al. (2004) Early versus delayed laparoscopic cholecystectomy for
acute cholecystitis: a prospective randomized trial. Surgical Endoscopy, 18, 1323–7.
Lai, P.B., Kwong, K.H., Leung, K.L., et al. (1998) Randomized trial of early versus delayed laparoscopic
cholecystectomy for acute cholecystitis. British Journal of Surgery, 85, 764–7.
Related references
Gurusamy, K.S. and Samraj, K. (2006) Early versus delayed laparoscopic cholecystectomy for acute
cholecystitis. Cochrane Database of Systematic Reviews, 4, CD005440.
Lau, H., Lo, C.Y., Patil, N.G., and Yuen, W.K. (2006) Early versus delayed-interval laparoscopic chole-
cystectomy for acute cholecystitis: a meta analysis. Surgical Endoscopy, 20, 82–7.
Study design
◆ 535 RCTs were identified by searching electronic databases and controlled trials
registers.
◆ Selection criteria: RCTs comparing early laparoscopic cholecystectomy (ELC) (<7
days of symptom onset) with delayed laparoscopic cholecystectomy (DLC) (<6 weeks
after index admission).
◆ Excluded trials: 366 were ‘irrelevant’ from titles and abstracts, 158 were duplicates,
three did not fit within the definitions of the intervention, two were multiple reports,
and one was non-randomized.
EARLY VERSUS LATE LAPAROSCOPIC CHOLECYSTECTOMY IN ACUTE CHOLECYSTITIS 169
Outcomes
Primary outcomes
◆ Mortality.
◆ Surgical morbidity (bile leak, re-operation, infection, bleeding).
◆ Complications during wait (pancreatitis, recurrent cholecystitis, obstructive
jaundice).
◆ Conversion to open operation.
Secondary outcomes
◆ Operating time, incidence of CBD stones, hospital stay, number of working days lost,
quality of life.
Included trials
Author Timing Patients ELC in DLC Surgical
(days vs weeks) (ELC:DLC) group experience
Lo et al. 1998 <7 vs 8–12 49:50 9 (18%) >300 LCs
Lai et al. 1998 <7 vs 6–8 53:51 8 (16%) >50 LCs
Dávila et al. 1999 <4 vs 8 27:36 5 (14%) Not stated
Johansson et al. 2003 <7 vs 6–8 74:71 18 (25%) >25 LCs
Kolla et al. 2004 <4 vs 6–12 20:20 0 (0%) Consultant
Results
◆ There were no deaths in any included RCT.
◆ There was no significant difference in BDI (1/222 (0.5%) ELC versus 3/216 (1.4%)
DLC, p = 0.54).
◆ A trend towards bile leak requiring ERCP was noted in ELC but was not significant
(7/222 (3.2%) ELC versus 0/216 DLC, p = 0.05).
◆ There was no significant difference in the conversion rate (45/222 (20.3%) ELC
versus 51/216 (23.6%) DLC, p = 0.47 by ITT analysis).
◆ There was no significant difference in collections requiring intervention (p = 0.31),
superficial wound infection (p = 0.48), or deep wound infections (p = 0.28).
◆ 40/228 (17.5%) did not resolve or had recurrent attacks while awaiting surgery.
18/40 (45%) required open conversion and two developed cholangitis, but no patient
developed pancreatitis.
◆ Hospital stay was significantly reduced in the ELC group (ELC mean stay 4.1–7.6
days versus DLC 8.0–11.6 days, p<0.001).
◆ Significantly fewer working days were lost following ELC (15 versus 26 days, p = 0.01
(Lo et al.)).
◆ There were no significant differences in median operating time, incidence of CBD

stones, or quality of life at 3 and 6 months.
◆ There was no difference in outcome when stratified for surgical experience.
◆ Risk of bias was considered to be high, as blinding was not performed in any RCT.
Conclusions
ELC during acute cholecystitis appears to be safe and shortens total hospital stay. Surgery
is more complex and conversion rates are higher in acute cholecystitis than in uncompli-
cated symptomatic gallstone disease. Therefore although this meta-analysis showed no
effect of surgeon’s experience between early and delayed surgery on any of the outcome
measures, surgeons with adequate experience are likely to perform better when dealing
with acute cholecystitis.
Critique
Symptomatic gallstone disease is a common surgical pathology and may present electively
or in the acute setting. Whilst there are reasonable data to support early open cholecystec-
tomy with no increase in morbidity or mortality, the evidence base for the laparoscopic
approach was lacking. Laparoscopic cholecystectomy has evolved to become the standard
of care (70–90% of UK cases) and therefore it was important to determine whether this
procedure could be performed safely at the time of acute presentation. Concerns regard-
ing early cholecystectomy centred on the acute inflammatory process involving Calot’s
triangle, with an anticipated higher conversion rate to open surgery and increase in mor-
bidity, particularly BDI.
This meta-analysis of data from five randomized clinical trials provided clear evidence
to support early laparoscopic cholecystectomy. The initial concerns regarding open con-
version, BDI, or other morbidity were not borne out by these data, although there was a
non-significant trend towards ERCP for post-operative bile leak in the ELC group.
A previous meta-analysis by Lau et al. in 2006 demonstrated similar findings in a study
population of 504 patients. Not only is ELC as safe as DLC, but it is also more cost-effec-
tive. The total length of hospital stay is shorter with early operation and it also avoids the
17.5% re-admission rate whilst awaiting surgery on a delayed basis. There may be implica-
tions for seniority of operating surgeon and institutional capability, but the evidence base
supports the role of early laparoscopic cholecystectomy in the acute setting.
ENTERAL VERSUS PARENTERAL NUTRITION IN ACUTE PANCREATITIS 171
5.5 Enteral versus parenteral nutrition in acute pancreatitis

Details of studies
Severe acute pancreatitis is a catabolic and protracted illness. Therefore nutritional sup-
port is critical in this patient population. Before sufficient evidence was available the most
common mode of delivery had been total parenteral nutrition (TPN). The proposed ben-
efits of TPN were avoidance of pancreatic stimulation and of problems with gastric stasis
or outlet obstruction when feeding via the enteral route. The first study to change this was
a randomized trial of 38 patients comparing TPN and enteral nutrition via the nasoenteral
route. This study has been followed by eight further RCTs, all reporting similar results.
Study references
Main study
Kalfarentzos, F., Kehagias, J., Mead, N., Kokkinis, K., and Gogos, C.A. (1997) Enteral nutrition is supe-
rior to parenteral nutrition in severe acute pancreatitis: results of a randomized prospective trial.
Related references
McClave, S.A., Greene, L.M., Snider, H.L., et al. (1997) Comparison of the safety of early enteral versus
parenteral nutrition in mild acute pancreatitis. Journal of Parenteral and Enteral Nutrition, 21,
14–20.
Abou-Assi, S., Craig, K., and O’Keefe S.J. (2002) Hypocaloric jejunal feeding is better than total paren-
teral nutrition in acute pancreatitis: results of a randomized comparative study. American Journal of
Gastroenterology, 97, 2255–62.
Oláh, A., Pardavi, G., Belágyi, T., Nagy, A., Issekutz, A., and Mohamed, G.E. (2002) Early nasojejunal
feeding in acute pancreatitis is associated with a lower complication rate. Nutrition, 18, 259–62.
Gupta, R., Patel, K., Calder, P.C., Yaqoob, P., Primrose, J.N., and Johnson, C.D. (2003) A randomised
clinical trial to assess the effect of total enteral and total parenteral nutritional support on metabolic,
inflammatory and oxidative markers in patients with predicted severe acute pancreatitis (APACHE
II > or = 6). Pancreatology, 3, 406–13.
Louie, B.E., Noseworthy, T., Hailey, D., Gramlich, L.M., Jacobs, P., and Warnock, G.L. (2005) Enteral
or parenteral nutrition for severe pancreatitis: a randomized controlled trial and health technology
assessment. Canadian Journal of Surgery, 48, 298–306.
Petrov, M.S., Kukosh, M.V., and Emelyanov, N.V. (2006) A randomized controlled trial of enteral versus
parenteral feeding in patients with predicted severe acute pancreatitis shows a significant reduction
in mortality and in infected pancreatic complications with total enteral nutrition. Digestive Surgery,
23, 336–45.
Casas, M., Mora, J., Fort, E., et al. (2007) Total enteral nutrition vs. total parenteral nutrition in patients
with severe acute pancreatitis. Revista Española de Enfermedades Digestivas, 99, 264–9.
Study design
Randomization TPN versus nasoenteric (NE) feeding
Number of patients 38 of 40 (two failed placements of NE tube)
Follow-up Continuous until hospital discharge

Nutritional support monitoring (nitrogen balance calculation)
Monitoring of infectious complications (CRP 3/week, blood and urine
cultures 2/week, central lines changed weekly)
Patient records reviewed on discharge
◆ Inclusion criteria: three or more Imrie criteria or APACHE II score ≥8, C-reactive
protein (CRP) concentration >120mg/L within 48 hours of admission, and CT grade
D or E according to the Balthazar criteria.
◆ Exclusion criteria: patients treated elsewhere for >2 days before admission and those
with failed placement of feeding tube/line.
◆ NE group: patients received enteral nutrition of a peptide-based feed through an NE
feeding tube placed fluoroscopically distal to ligament of Treitz (Flexiflo 10fr, 115cm
within 48 hour of admission).
◆ TPN group: patients received parenteral nutrition containing crystalline l-amino
acid, carbohydrates in the form of dextrose, fat emulsion, vitamins, and minerals
through a subclavian central venous line.
Outcome measures
Primary endpoint
◆ Comparison of nutritional maintenance via EN and TPN routes.
Secondary endpoints
◆ Complications during the course of the disease including death.
◆ Length of intensive care unit (ICU) stay.
◆ Length of hospital stay.
◆ Number of days of artificial ventilation.
◆ Number of days receiving tube feeding or TPN.
◆ Number of antibiotics used and number of days on antibiotics.
◆ Mean cost of EN and TPN per day.
Results
◆ There was no difference between the groups with regard to demographics or
aetiology, duration, or prognostic signs of pancreatitis.
◆ There was no difference in need for ICU support, use of antibiotics, or length of
hospital stay
◆ There was no difference in the number of days of nutritional support, caloric intake,
or nitrogen balance (which improved progressively in both groups).
ENTERAL VERSUS PARENTERAL NUTRITION IN ACUTE PANCREATITIS 173
Complications EN (n = 18) TPN (n = 20)

Nutrition-related
Catheter-related sepsis 0 2
Hyperglycaemia 4 9
Pancreatitis-related
Infected peri-pancreatic necrosis 1 4
Abscess formation 0 0
Pancreatic fistula formation 0 2
Non-pancreatitis-related
Blood: –ve culture + ve sepsis 1 3
UTI 1 2
Pneumonia/ARDS 2 4
Patients with septic complications 5 10
Septic complication 6 15
Patients with any complication 8 15
Total number with any complication 10 27
UTI, urinary tract infection; ARDS, adult respiratory distress syndrome.
◆ There was significantly less morbidity associated with the EN group (p <0.05).
◆ The mean number of infections per patient was lower in the EN group (0.56 versus
1.35, p <0.01).
◆ The risk of developing a complication was 3.47 times greater with TPN.
◆ In the EN group, two patients underwent a total of three operations. In the TPN
group, four patients underwent a total of eleven operations.
◆ TPN cost more than EN (£100 versus £30 per patient per day).
Conclusions
Early EN is well tolerated in severe acute pancreatitis (SAP) and is of comparable efficacy
to TPN. EN was accompanied by significantly fewer complications and cost less than
TPN. The results suggest that the enteral route should be used preferentially for nutri-
tional support in patients with SAP.
Critique
Severe acute pancreatitis results in a profound physiological insult to the host and requires
early nutritional support. Historically, the concept of avoiding pancreatic stimulation by
‘resting the gut’ has mandated parenteral nutrition. This has been challenged by several
studies over the last 15 years.
The study by Kalfarentzos et al. demonstrated that not only is enteral nutrition by the
nasojejunal route tolerated in these patients, but also it is safer and less costly. However,
there are several criticisms. Firstly this trial, and those that followed, are in relatively small
patient groups (n = 38). Secondly, prophylactic antibiotics were used until the restoration
of a normal CRP. This would not be considered standard care in most tertiary referral
centres, as elevation of this inflammatory marker is a normal function of SIRS. Thirdly,
one of the indications for surgical intervention was multi-organ failure (MOF) after 5
days of intensive care. This would probably not have a place in the modern management
of this condition. Despite these criticisms, this trial is significant in that it effectively chal-
lenged the dogma regarding enteral nutrition in severe acute pancreatitis.
A more recent randomized study by Eatock et al. pushed this envelope further by
demonstrating that nasogastric feeding is as effective as nasojejunal feeding in SAP. The
remaining indication for TPN is failure to establish adequate nutritional support by
the enteral route, for example in patients with prolonged ileus, high output fistulae, or
increased intestinal motility, or in those requiring a period of dual (EN and PN) feeding.
Eatock, F.C., Chong, P., Menezes, N., et al. (2005) A randomized study of early nasogastric versus
nasojejunal feeding in severe acute pancreatitis. American Journal of Gastroenterology, 100, 432–9.
PROPHYLACTIC ANTIBIOTICS FOR SEVERE ACUTE NECROTIZING PANCREATITIS 175
5.6 Prophylactic antibiotics for severe acute

necrotizing pancreatitis
Details of studies
The use of antibiotic (Abx) prophylaxis to prevent infected pancreatic necrosis (IPN) and
reduce mortality has remained controversial for decades. Earlier RCTs favoured the use
of Abx prophylaxis. However, the publication of a well-conducted trial by Dellinger et al.
contradicted this. The first meta-analysis to include this trial, published in 2008, changed
clinical practice and is described here. A number of further systematic reviews and meta-
analyses have been published with similar conclusions.
Study references
Main study
Bai, Y., Gao, J., Zou, D., and Li, Z.S. (2008) Prophylactic antibiotics cannot reduce infected pancreatic
necrosis and mortality in acute necrotizing pancreatitis: evidence from a meta-analysis of ran-
domised controlled trials. American Journal of Gastroenterology, 103, 104–10.
Included trials
Pederzoli, P., Bassi, C., Vesentini, S., and Campedelli, A. (1993) A randomized multicenter clinical trial
of antibiotic prophylaxis of septic complications in acute necrotizing pancreatitis with imipenem.
Surgery, Gynecology and Obstetrics, 176, 480–3.
Nordback, I., Sand, J., Saaristo, R., et al. (2001) Early treatment with antibiotics reduces the need for sur-
gery in acute necrotizing pancreatitis—a single-center randomized study. Journal of Gastrointestinal
Surgery, 5, 113–20.
Isenmann, R., Rünzi, M., Kron, M., et al. (2004) Prophylactic antibiotic treatment in patients with
predicted severe acute pancreatitis: a placebo-controlled, double-blind trial. Gastroenterology, 126,
997–1004.
Sainio, V., Kemppainen, E., Puolakkainen, P., et al. (1995) Early antibiotic treatment in acute necrotising
pancreatitis. Lancet, 346, 663–7.
Schwarz, M., Isenmann, R., Meyer, H., and Beger, H.G. (1997) Antibiotic use in necrotizing pancreatitis.
Results of a controlled study. Deutsche medizinische Wochenschrift, 122, 356–61.
Dellinger, E.P., Tellado, J.M., Soto, N.E., et al. (2007) Early antibiotic treatment for severe acute necrotiz-
ing pancreatitis: A randomized, double-blind, placebo-controlled study. Annals of Surgery, 245,
674–83.
RØkke, O., Harbitz, T.B., Liljedal, J., et al. (2007) Early treatment of severe pancreatitis with imipenem: a
prospective randomized clinical trial. Scandinavian Journal of Gastroenterology, 42, 771–6.
Related references
Wittau, M., Mayer, B., Scheele, J., Henne-Bruns, D., Dellinger, E.P., and Isenmann, R. (2011) Systematic
review and meta-analysis of antibiotic prophylaxis in severe acute pancreatitis. Scandinavian Journal
of Gastroenterology, 46, 261–70.
Villatoro, E., Mulla, M., and Larvin, M. (2010) Antibiotic therapy for prophylaxis against infection of
pancreatic necrosis in acute pancreatitis. Cochrane Database of Systematic Reviews, 12, CD002941.
Segarra-Newnham, M. and Hough, A. (2009) Antibiotic prophylaxis in acute necrotizing pancreatitis
revisited. Annals of Pharmacotherapy, 43, 1486–95.
Jafri, N.S., Mahid, S.S., Idstein, S.R., Hornung, C.A., and Galandiuk, S. (2009) Antibiotic prophylaxis is
not protective in severe acute pancreatitis: a systematic review and meta-analysis. American Journal
of Surgery, 197, 806–13.
Study design
◆ 359 trials were identified by searching electronic databases, reference listings of these
trials, and major conference abstract books.
◆ Selection criteria: RCTs of prophylactic antibiotics versus placebo for CT-proven
severe acute pancreatitis with necrosis.
◆ Outcome: infected pancreatic necrosis and mortality.
◆ Excluded trials: 344 were non-randomized, four had no data on pancreatic necrosis,
three were early trials and so CT was not available, and one involved the use of
antibiotics and selective gut decontamination.
Included trials
Author Setting Patients Blinding Dosage and duration (IV)
Pederzoli 1993 Multi-centre 74 Single Imipenem 0.5g tds
Sainio 1995 Single centre 60 Single Cefuroxime 1.5g tds
Schwarz 1997 Single centre 26 Single Ofloxacin 0.2g bd +
metronidazole 0.5g bd
Nordback 2001 Single centre 39 Single Imipenem 1g tds
Isenmann 2004 Multi-centre 76 Double Ciprofloxacin 0.4g bd +
metronidazole 0.5g bd
Dellinger 2007 Multi-centre 100 Double Meropenem 0.5g tds
Røkke 2007 Multi-centre 73 None Imipenem 0.5 g tds
◆ The risk of bias was high for all trials except for Isenmann et al. and Dellinger et al.
Results
◆ A total of 467 patients were included in the seven trials.
◆ There was no significant reduction in IPN with ABx use versus placebo (18% versus
23%, p = 0.32).
◆ There was no significant reduction in mortality with ABx use (9% ABx versus 15%
controls, p = 0.17).
◆ Subgroup analyses did not show any significant reduction of IPN when considering
single-centre trials (p = 0.55), multi-centre trials (p = 0.46), single-blinded trials
(p = 0.21), double-blinded trials (p = 0.40), trials with a high risk of bias (p = 0.11),
trials with a low risk of bias (p = 0.40), and trials with different ABx agents (beta-
lactam, p = 0.13; quinolones p = 0.61).
◆ Subgroup analyses did not show any significant reduction of mortality when
PROPHYLACTIC ANTIBIOTICS FOR SEVERE ACUTE NECROTIZING PANCREATITIS 177
considering multi-centre trials (p = 0.62), double-blinded trials (p = 0.94), trials with

a high risk of bias (p = 0.06), trials with a low risk of bias (p = 0.94), and trials with
different ABx agents (beta-lactam, p = 0.29; quinolones, p = 0.31).
◆ There was a significant difference in mortality in subgroup analysis of single-centre
trials (RR 0.30, 95% CI 0.10–0.95, p = 0.04) and single-blinded trials (RR 0.40, 95%
CI 0.16–0.96, p = 0.04).
Conclusions
Prophylactic antibiotics do not reduce IPN and mortality in patients with acute necrotiz-
ing pancreatitis. The mortality rate was reduced only in low-quality settings (single-centre
and single-blinded), and the significance was marginal (p = 0.04).
Critique
There are two mortality peaks in patients with severe acute pancreatitis. The first occurs
early and is intimately related to the development of systemic complications, including
multi-organ failure. The second peak is associated with secondary infection of pancreatic
necrosis. As a result of this second phenomenon there has been considerable interest in
the use of prophylactic antibiotics in an attempt to reduce the incidence of infected pan-
creatic necrosis and thereby mortality. A Cochrane review in 2003 reported ‘strong evi-
dence’ to support the use of intravenous antibiotic prophylactic therapy for 10–14 days to
reduce the risk of super-infection of necrosis and mortality in patients with proven pan-
creatic necrosis on CT. A further Cochrane review in 2006 was more guarded, noting that
antibiotic prophylaxis appeared to be associated with significantly decreased mortality
but not infected pancreatic necrosis. Only beta-lactams were associated with a significant
reduction in infected pancreatic necrosis. The authors commented upon the heteroge-
neous nature of the various trials as a confounding factor in interpreting these data.
The meta-analysis by Bai et al. was the first review to include the Dellinger et al. ran-
domized trial. It concluded that prophylactic antibiotics do not reduce the incidence of
infected pancreatic necrosis or mortality in patients with acute necrotizing pancreatic
necrosis. This conclusion regarding mortality was supported by the subsequent 2010
Cochrane review. In summary, therefore, there is no proven role for prophylactic antibi-
otics in the context of pancreatic necrosis.
Bassi, C., Larvin, M., and Villatoro, E. (2003) Antibiotic therapy for prophylaxis against infection of
pancreatic necrosis in acute pancreatitis. Cochrane Database of Systematic Reviews, (4), CD002941.
Villatoro, E., Bassi, C., Larvin, M. (2006) Antibiotic therapy for prophylaxis against infection of pancre-
atic necrosis in acute pancreatitis. Cochrane Database of Systematic Reviews, (4), CD002941.
Villatoro, E., Mulla, M., and Larvin, M. (2010) Antibiotic therapy for prophylaxis against infection of
pancreatic necrosis in acute pancreatitis. Cochrane Database of Systematic Reviews, (5), CD002941.
5.7 Early versus late necrosectomy in severe

Details of study
Many aspects of the surgical management of severe acute necrotizing pancreatitis remain
under debate. However, the question over timing of debridement has been answered in
the only prospective RCT published in 1995. This study was terminated prior to comple-
tion because of the high mortality in one of the study groups.
Study references
Main study
Mier, J., LeÓn, E.L., Castillo, A., Robledo, F., and Bianco, R. (1997) Early versus late necrosectomy in
severe necrotizing pancreatitis. American Journal of Surgery, 173, 71–5.
Related references
Takeda, K., Matsuno, S., Sunamura, M., and Kobari, M. (1998) Surgical aspects and management of
acute necrotizing pancreatitis: recent results of a cooperative national survey in Japan. Pancreas, 16,
316–22.
Hungness, E.S., Robb, B.W., Seeskin, C., Hasselgren, P., and Luchette, F.A. (2002) Early debridement for
necrotizing pancreatitis. Is it worthwhile? Journal of the American College of Surgeons, 194, 740–5.
Study design
Randomization Early (48–72 hours after onset of illness) versus late (≥12 days after onset)
necrosectomy
Primary outcome Hospital all-cause mortality
Secondary outcome Bacteriological findings to establish the presence of infective necrosis
◆ Inclusion criteria: diagnosis of acute pancreatitis by typical clinical features,

laboratory findings (serum amylase twice normal, leucocytosis, and characteristic
findings on abdominal ultrasound or CT), and patient deemed a surgical candidate
based on early fulminant course and clinical and laboratory signs of multi-organ
failure with deterioration despite maximal ICU input. Severity was substantiated with
a high Ranson’s score and/or extensive necrosis on contrast CT.
◆ Exclusion criteria: patients who had been treated elsewhere prior to first contact or
had a period of relative benign disease progression during their initial treatment
◆ Surgical technique: open debridement and packing approach as described by Bradley
with multiple re-explorations and repeat debridements. Pre-operative CT-guided
EARLY VERSUS LATE NECROSECTOMY IN SEVERE NECROTIZING PANCREATITIS 179
fine-needle aspiration (FNA) was not performed. Three separate aspirates or

specimens of necrotic tissue were collected at operation for bacteriology. The same
group of surgeons performed all operations.
Outcome measures
Primary endpoint
◆ In-hospital all-cause mortality.
Secondary endpoints
◆ Presence of infected necrosis in necrosectomy specimens.
Results
◆ There was no difference between the groups in demographics or disease
characteristics including number of necrosectomies, percentage of IPN and sterile
pancreatic necrosis (SPN), and extent of pancreatic necrosis.
Early necrosectomy Late necrosectomy

(n = 25) (n = 11)
Overall mortality 14/25 (56%) 3/11 (27%)
Mortality in IPN 11/15 (73%) 2/7 (29%)
Mortality in SPN 3/10 (30%) 1/4 (25%)
◆ IPN was present in 47% (9/19) of survivors versus 76% (13/17) of non-survivors (not
statistically significant).
◆ The study was terminated as the odds ratio was found to be 3.4 times higher for
patients undergoing early necrosectomy, and that three patients in the late group
responded to non-operative management with no clear indication for necrosectomy
by the time of planned intervention.
Conclusions
Early intensive conservative management, with late necrosectomy in selected cases, should
be the current approach in managing patients with severe acute necrotizing pancreatitis.
Critique
There has been controversy regarding the management of severe acute pancreatitis for
more than a hundred years. In particular, there has been debate over the last two decades
regarding the mode and timing of intervention in the 15–20% of patients with pancre-
atic necrosis. Although the evolution has moved on to embrace minimally invasive tech-
niques to deal with this potentially devastating complication, this paper from 1997 is still
relevant today. It provides randomized evidence for avoiding intervention in the first two
weeks of the disease where possible.
There are several criticisms: it is a small single-institution study (n = 36), the mortality
rate between the early and delayed open necrosectomy groups did not reach statistical
significance, and there was a clear disparity in the size of the two groups. However, the
authors’ explanation regarding early termination of the study prior to completing recruit-
ment is reasonable. The odds ratio for mortality in the early surgery group was 3.4, despite
both groups being otherwise adequately matched.
One of the reasons for the increase in mortality with early surgery is probably the
nature of the operation performed. In 1963, Watts noted that early surgery often resulted
in a ‘resective’ rather than a ‘debridement’ approach as separation of the necrotic tissue
from viable pancreas had not yet been established. This resulted in total pancreatectomy
rather than limited removal of devitalized tissue and, as a consequence, a greater surgical
and physiological insult. Mier et al. provide the only prospective RCT on timing of open
necrosectomy. Their data support the role of conservative critical care management in the
first two weeks with delayed rather than early surgical intervention.
Bradley, E.L. (1993) A fifteen year experience with open drainage for infected pancreatic necrosis.
Watts, G.T. (1963) Total pancreatectomy for fulminant pancreatitis. Lancet, 2, 384.
MINIMALLY INVASIVE VERSUS OPEN NECROSECTOMY FOR NECROTIZING PANCREATITIS 181
5.8 Minimally invasive versus open necrosectomy for

Details of studies
The optimal strategy for dealing with infected necrotizing pancreatitis has been debated
for some time. Percutaneous drainage is often unsuccessful, whilst the traditional open
approach is associated with high rates of morbidity (34–95%) and mortality (11–39%). A
number of other less invasive approaches to necrosectomy have been reported with varied
success. The first paper to consider these techniques in a ‘step-up’ approach is reported in
this Dutch RCT published in 2010.
Study references
Main study
van Santvoort, H.C., Besselink, M.G., Bakker, O.J., et al. (2010) A step-up approach or open
necrosectomy for necrotizing pancreatitis. New England Journal of Medicine, 362, 1491–1502.
Besselink, M.G., van Santvoort, H.C., Nieuwenhuijs, V.B., et al. (2006) Minimally invasive ‘step-up
approach’ versus maximal necrosectomy in patients with necrotising pancreatitis (PANTER trial):
design and rationale of a randomised controlled multicenter trial. BMC Surgery, 6, 6.
Related references
Carter, C.R., McKay, C.J., and Imrie C,W. (2000) Percutaneous necrosectomy and sinus tract endoscopy
in the management of infected pancreatic necrosis: an initial experience. Annals of Surgery, 232,
175–80.
van Santvoort, H.C., Besselink, M.G., Horvath, K.D., et al. (2007) Videoscopic-assisted retroperitoneal
debridement in infected necrotizing pancreatitis. HPB (Oxford), 9, 156–9.
Horvath, K.D., Kao, L.S., Ali, A., Wherry, K.L., Pellegrini, C.A., and Sinanan, M.N. (2001) Laparoscopic
assisted percutaneous drainage of infected pancreatic necrosis. Surgical Endoscopy, 15, 677–82.
Windsor, J.A. (2007) Minimally invasive pancreatic necrosectomy. British Journal of Surgery, 94, 132–3.
Study design
Randomization Minimally invasive ‘step-up approach’ versus maximal
necrosectomy by laparotomy
Stratification Based on possibility of percutaneous drainage
Peri-pancreatic necrosis on CT
Decision for surgical intervention made because of suspected
peri-pancreatic necrosis
Drain placement possible
Exclusion criteria Previous drainage or necrosectomy

Diagnosis of pancreatitis during previous exploratory laparotomy
Acute exacerbation of chronic pancreatitis
Bleeding
Perforation of visceral organ
Abdominal compartment syndrome
Post-abdominal surgery necrotizing pancreatitis
Visits at 3 and 6 months
◆ All patients were treated by a pre-established treatment protocol consisting of

nasojejunal (NJ) enteral nutrition, early ERCP with sphincterotomy in cases of
predicted severe biliary pancreatitis, and no antibiotic prophylaxis.
◆ Step-up technique: a percutaneous (or endoscopic) drain was placed, with a second
drainage procedure after 72 hours if no improvement/inadequate drain placement. If
not possible or no improvement after further 72 hours, video-assisted retroperitoneal
debridement (VARD) with lavage was performed.
◆ Open necrosectomy: a maximal necrosectomy at laparotomy was performed in the
form of radical debridement combined with post-operative lavage
Outcome measures
Primary endpoint
◆ Mortality.
◆ New-onset multi-organ failure or systemic complications.
◆ Enteric fistula (either small or large bowel).
◆ Perforation of a visceral organ requiring intervention.
◆ Bleeding requiring intervention.
Secondary endpoints
◆ ‘Minor’ pancreatic complications (such as pancreatic fistula etc.), new-onset sepsis or
SIRS, total number of interventions, hospital and ICU stay, quality of life, total (direct
and indirect) cost.
Results
Step-up (n = 43) Open (n = 45) p-value
Mortality 8 (19%)) 7 (16%) 0.70
MOF or systemic complication 5 (12%) 19 (42%) 0.001
Major complication or death 17 (40%) 31 (69%) 0.006
Other complication
Incisional hernia 3 (7%) 11 (24%) 0.03
New onset diabetes 7 (16%) 17 (38%) 0.02
Use of pancreatic enzymes 3 (7%) 15 (33%) 0.002
MINIMALLY INVASIVE VERSUS OPEN NECROSECTOMY FOR NECROTIZING PANCREATITIS 183
Step-up (n = 43) Open (n = 45) p-value

Necrosectomies (Lap or VARD)
0 17 (40%) 0
1 19 (44%) 31 (69%)
2 6 (14%) 8 (18%)
≥3 1 (2%) 6 (13%) <0.001
Total number of operations 53 91 0.004
Total drainage procedures 82 32 <0.001
New ICU admissions after intervention 7 (16%) 18 (40%) 0.01
◆ The step-up approach reduced costs by 12%.
Conclusions
The preferred treatment strategy for patients with necrotizing pancreatitis and secondary
infection, from both a clinical and an economic point of view, is a minimally invasive
step-up approach consisting of percutaneous drainage followed, if necessary, by mini-
mally invasive retroperitoneal necrosectomy.
Critique
The management of pancreatic necrosis has evolved dramatically over the past two
decades. The previous gold standard was early aggressive open resectional surgery. This
was associated with a high mortality. Modern management now revolves around critical
care support, delaying intervention for as long as possible, and intervening with a mini-
mally invasive modality if and when it is required.
The PANTER study provided good-quality randomized data to challenge the tradi-
tional surgical dogma surrounding necrosectomy. Patients with documented or suspected
infected pancreatic necrosis were randomized to either primary open necrosectomy or a
step-up approach involving endoscopic ultrasound (EUS) or percutaneous drainage as
the initial manoeuvre. The number of patients per institution was low (88 patients from
19 hospitals), which in part accounted for the lack of a unified approach to the initial,
and subsequent, interventional management. Interestingly the rate of death did not differ
between the groups, but by using death or a major pre-defined complication as the com-
posite primary endpoint, the authors were able to demonstrate a significant difference in
the outcome between the groups (69% in the open necrosectomy group versus 40% in the
step-up group achieved this endpoint).
One of the most clinically relevant and significant findings of this study is that 35%
of patients with infected pancreatic necrosis in the step-up group were managed with
primary drainage alone (EUS or percutaneous) and did not require necrosectomy. This
has major implications for the future management of this disease, and again challenges
the preconception that early or delayed necrosectomy is necessary for all patients with
infected necrosis. This landmark study provides randomized evidence, on clinical and
economic grounds, for a step-up approach to the management of patients with severe
acute pancreatitis and infected pancreatic necrosis.
5.9 Endoscopic versus surgical therapy for chronic pancreatitis

Details of studies
Endoscopic and surgical pancreatic duct drainage procedures remain options in patients
with painful chronic pancreatitis (CP). A number of retrospective studies and a large ret-
rospective multi-centre analysis have reported good results with endoscopic treatment.
Two prospective RCTs from Brno, (Czech Republic) and Amsterdam directly compare
these modalities for treating painful CP.
Study references
Main studies
Díte, P., Ruzicka, M., Zboril, V., and Novotný, I. (2003) A prospective, randomized trial comparing
endoscopic and surgical therapy for chronic pancreatitis. Endoscopy, 35, 553–8.
Cahen, D.L., Gouma, D.J., Nio, Y., et al. (2007) Endoscopic versus surgical drainage of the pancreatic
duct in chronic pancreatitis. New England Journal of Medicine, 356, 676–84.
Related references
Rösch, T., Daniel, S., Scholz, M., et al. (2002) Endoscopic treatment of chronic pancreatitis: a multi-
center study of 1000 patients with long-term follow-up. Endoscopy, 34, 765–71.
Study designs
◆ Both studies are prospective single-centre RCTs.
Díte et al.
Randomization Endoscopic versus operative intervention for CP
Number of patients 140 (72 strictly randomized)
Primary outcome Clinical response
Further intervention necessary
Inclusion criteria CP diagnosed by established imaging methods (US, ERCP, CT, EUS
Age 18–70 years
Obstructive form of disease by same methods
Painful form of CP (Melzack’s score >3)
Failure of conservative treatment during previous 3 years
5-year duration of CP
Indication for intervention agreed by surgeon and gastroenterologist
Previous interventional therapy
Suspected pancreatic malignancy
Follow-up 6-monthly examination
Results assessed at 1, 3, and 5 years
ENDOSCOPIC VERSUS SURGICAL THERAPY FOR CHRONIC PANCREATITIS 185
◆ Endoscopic techniques were as required to treat stones or strictures, with technical

success defined as complete stone extraction or successful drainage of strictures by
repeated stenting every 2–4 months. Stent changes took place over 12–24 months
regardless of endoscopic appearances.
◆ Surgery was tailored to the patient’s situation and included drainage and resectional
procedures.
Cahen et al.
Randomization Endoscopic versus operative intervention for CP
Number of patients 39 (study terminated early)
Primary outcome Pain (average Izbicki pain scores over follow-up period)
Secondary outcomes Pain relief, physical and mental health, morbidity, mortality, hospital stay,
number of procedures, changes in endocrine and exocrine pancreatic
function
Inclusion criteria CP diagnosed from symptoms and changes on imaging, pancreatic
functional insufficiency or both
Obstruction of the pancreatic duct (MRCP/CT and severe recurrent pain
requiring opiates
Age 18–70 years
Exclusion criteria Pancreatic head >4cm
Contraindication to surgery
Contraindication to endoscopic treatment
Previous pancreatic surgery
Suspected pancreatic cancer
Life expectancy <2 years
Pregnancy
Follow-up Six weeks, then 3, 6, 12, 18 and 24 months
◆ In Amsterdam, varied endoscopic procedures were used with repeat pancreatic duct
stenting every 3 months. Stenting was discontinued if complete run-off of contrast
could be seen, and if a balloon could be easily withdrawn.
◆ All surgical procedures consisted of a pancreaticojejunostomy as described by
Partington and Rochelle.
Results
Díte et al.
◆ 64 patients received endoscopic treatment with technical success in 62 (requiring a
mean of two sessions of initial treatment). No patient required surgical intervention
and there were no deaths.
◆ 76 patients received surgical treatment with re-operation required in two cases

(because of ileus and anastomotic leakage). There were no deaths.
Total group (n = 140) Randomized group (n = 72)

Endoscopic Surgery p-value Endoscopic Surgery p-value
(n = 64) (n = 76) (n = 36) (n = 36)
Pain relief
Complete 14.3% 36.9% 0.002 15% 33.8% 0.002
Partial 50.8% 49.3% ns 46.4% 52.1% ns
No change 34.9% 13.8% ns 38.2% 14.1% ns
Body weight
Increased 26.9% 52.1% 0.002 28.6% 47.2% 0.003
Unchanged 23.9% 19.1% ns 25.7% 25.0% ns
Decreased 49.2% 28.8% ns 45.7% 27.8% ns
Cahen et al.
◆ There was one death due to duodenal ulcer perforation (5%).
◆ Twenty patients underwent surgical treatment. One patient required re-laparotomy
because of anastomotic leak. There were no deaths.
Endoscopic Surgery p-value

(n = 19) (n = 20)
Pain score 51 ± 23 25 ± 13 <0.001
Pain relief 6 (32%) 15 (75%) 0.007
Complete 3 (16%) 8 (40%)
Partial 3 (16%) 7 (35%)
None 13 (68%) 5 (25%)
Technical success 10 (53%) 20 (100%) <0.001
Procedures* 8 3 <0.001
Physical health (SF-36) 38 ± 9 47 ± 7 0.003
*Median values
◆ Four patients in the endoscopic group were converted to surgery for intractable pain.
◆ There was one death from duodenal ulcer perforation in the endoscopic group and
no deaths in the surgery group.
◆ The safety committee terminated the study early after an unscheduled interim
analysis showed a significant difference in outcome (p <0.001) favouring the surgical
group. At the date of termination, seven patients had not completed the follow-up
period.
ENDOSCOPIC VERSUS SURGICAL THERAPY FOR CHRONIC PANCREATITIS 187
Conclusions
Surgery is superior to endoscopic intervention for long-term pain control in patients with
painful obstructive CP. Better selection of patients for endoscopy might maximize results,
and because of its low degree of invasiveness it could be offered as a first-line treatment,
reserving surgery for failure or recurrence of symptoms.
Critique
The incidence of chronic pancreatitis appears to be increasing. The predominant feature
is pain, which may be disabling, refractory to opiate analgesia, and result in markedly
diminished quality of life. After lifestyle modification and analgesia, there are two prin-
ciple interventions—surgery or endotherapy. Prior to these papers the published data
comparing the outcomes of these interventions were limited to retrospective reviews. The
studies cited here are well-designed prospective RCTs, which support surgery over endo-
therapy for long-term pain control. The interim analysis of the Amsterdam data resulted
in early termination of the trial because the results of surgery were markedly superior.
Adequate follow-up was obtained in each trial (2 years and 5 years) and the entry criteria
were clearly defined.
Neither trial addresses the role of stenting as a bridge to surgery to allow a period of
optimization, or to determine whether pancreatic duct decompression results in symp-
tomatic relief. The other potential role for endotherapy is in the patient who is not fit
for major pancreatic surgery. Given the aetiological risk factors for chronic pancreatitis,
affected patients may have other significant comorbidity. In these cases endotherapy may
be the only salvage option. Despite these limitations, both papers provide clear random-
ized evidence for the role of surgery over endotherapy in treating pain associated with
chronic pancreatitis in patients for whom both options are feasible.
Partington, P.F. and Rochelle, R.E. (1960) Modified Puestow procedure for retrograde drainage of the
pancreatic duct. Annals of Surgery, 152, 1037–43.
5.10 Diagnosis of autoimmune pancreatitis

Details of studies
Prior to the study by Chari et al. in 2006, the diagnosis of autoimmune pancreatitis
(AIP) relied on fulfilling the Japan Pancreas Society minimal consensus criteria. These
criteria require the presence of characteristic imaging features (a diffusely enlarged sau-
sage-shaped gland and a diffusely irregular attenuated pancreatic duct) and either posi-
tive serology or histology. However, a number of additional features not included in the
Japanese criteria, including IgG4 serology, other organ involvement, and response to ste-
roid therapy, were described in the literature. Chari et al. reworked the diagnosis of AIP
with the important addition of steroid response.
Study references
Main study
Chari, S.T., Smyrk, T.C., Levy, M.J., et al. (2006) Diagnosis of autoimmune pancreatitis: the Mayo Clinic
experience. Clinical Gastroenterology and Hepatology, 4, 1010–16.
Related references
Sugumar, A. and Chari, S.T. (2011) Autoimmune pancreatitis. Journal of Gastroenterology and
Hepatology, 26, 1368–73.
Members of the Criteria Committee for Autoimmune Pancreatitis of the Japan Pancreas Society.
(2002) Diagnostic criteria for autoimmune pancreatitis by the Japan Pancreas Society. Journal of the
Japanese Pancreas Society, 17, 585–7.
Study design
◆ A review of Mayo Clinic data to develop expanded diagnostic criteria for AIP.
◆ Five criteria for AIP were analysed in 29 consecutive patients (HISORt criteria).
Category Criteria
Histology At least one of the following:
(n = 29)
1. Periductal lymphoplasmacytic infiltrate with obliterative phlebitis
and storiform fibrosis (LPSP)
2. Lymphoplasmacytic infiltrate with storiform fibrosis showing
abundant (≥10 cells/HPF) IgG4-positive cells
Pancreatic imaging Typical: diffusely enlarged gland with delayed (rim) enhancement;
diffusely irregular attenuated main pancreatic duct
(CT = 22, pancreatogram = 7)
Others: focal pancreatic mass/enlargement, focal pancreatic duct
stricture, pancreatic atrophy, pancreatic calcification, or pancreatitis
Serology Elevated serum IgG4 level
Other organ involvement Hilar/intrahepatic biliary strictures, persistent distal biliary stricture,
parotid/lacrimal gland involvement, mediastinal lymphadenopathy,
retroperitoneal fibrosis
Response to steroids Resolution/marked improvement of pancreatic/extra-pancreatic
manifestation with steroid therapy
HPF, high-power field.

DIAGNOSIS OF AUTOIMMUNE PANCREATITIS 189
Results
◆ The mean age of patients was 63 years with 79% >50 years.
◆ Presentation was with obstructive jaundice in 23 patients, pancreatic mass in three,
and steatorrhoea, diabetes, and pancreatitis in one each.
Histology
◆ Thirteen pancreatic resection specimens (eight with IgG4 immunostaining
performed) and 16 core biopsies were available, of which 12/13 and 7/16 showed
LPSP, with 7/8 and 15/16 patients IgG4-positive. The resection without LPSP was
IgG4-positive. The biopsy patient without IgG4 cells had full-spectrum LPSP.
Imaging
◆ Six patients had diffuse pancreatic enlargement on CT. There were a variety of
other findings on CT including focal enlargement (ten), distinct mass (three), acute
pancreatitis (one), calcification (four), and normal glands (two).
◆ There was no correlation between CT and pancreatogram, although one patient with
a normal CT had a diffusely irregular duct.
Serology
◆ IgG4 was more frequently elevated than total IgG (15/21 versus 8/21, p = 0.03). All 6
patients with normal IgG4 serology were IgG4-positive on histology.
Other organ involvement

◆ Other organ involvement was seen in 11/29 (38%) including the CHD, intra-hepatic
bile duct, retroperitoneum, salivary gland, and mediastinal nodes
Response to steroids
◆ Seventeen patients received steroids and all showed complete resolution or marked
improvement.
Proposed diagnostic criteria

Group A: One or both of the following:
Diagnostic pancreatic ◆ resection specimen/core biopsy showing the full spectrum of changes of
histology LPSP
◆ ≥10 IgG4-positive cells/HPF IgG4 on immunostain of pancreatic
lymphoplasmacytic infiltrate
Group B: Presence of all of the following:
Typical imaging + serology ◆ CT or MRI showing diffusely enlarged pancreas with delayed and ‘rim’
enhancement
◆ pancreatogram showing diffusely irregular pancreatic duct elevated

serum IgG4 levels
Group C: Presence of all of the following:
Response to steroids ◆ unexplained pancreatic disease after negative work-up for known
aetiologies including cancer
◆ elevated serum IgG4 levels and/or other organ involvement confirmed

by presence of abundant IgG4-positive cells
◆ resolution/marked improvement in pancreatic and/or extra-pancreatic

manifestations with steroid therapy
◆ Twenty further patients were identified based on these new diagnostic criteria. There
was no difference in age distribution, gender, proportion with diffuse pancreatic
enlargement, proportion with elevated serum IgG4, or response to steroids among
the three groups.
◆ The proportion with other organ involvement was significantly greater in patients
in group C, all of whom were steroid responsive. These patients frequently had no
pancreatic involvement.
Conclusions
The proposed criteria reflect the current understanding of AIP as a systemic steroid-
responsive disorder characterized by tissue infiltration with IgG4-positive cells. They
identify a wider spectrum of manifestations than the Japanese criteria.
Critique
This paper documented the Mayo Clinic experience of autoimmune pancreatitis and fun-
damentally revised our approach to the disease. Prior to this publication the diagnosis of
AIP was based on the criteria of the Japan Pancreas Society. These were heavily biased
towards pancreatic imaging and did not recognize the importance of response to ste-
roid therapy as a diagnostic marker. The Mayo criteria (HISORt—Histology, pancreatic
Imaging, Serology, Other organ involvement, Response to steroid therapy) were the first
to incorporate steroid response and other additional features to provide a comprehensive
means of identifying the full spectrum of clinical presentations.
AIP is a relatively rare disease, although increasingly recognized, and this is reflected
in the small number of patients described in this paper. Only 17 patients were treated
with steroids, and all of these exhibited either resolution or marked improvement in pan-
creatic/extra-pancreatic manifestations. However, the follow-up period after treatment
was relatively short (median 6.5 months). AIP is clinically significant, in part because of
the constellation of local and systemic abnormalities with which it is associated, but also
because of the diagnostic dilemma it poses. The authors note that steroids should not be
given until a comprehensive evaluation of other potentially treatable aetiologies has been
performed. In particular, it may be difficult to differentiate focal AIP from pancreatic
malignancy. Despite the small study size and limited follow-up, this paper has revolution-
ized the diagnosis and management of AIP.
PRE-OPERATIVE BILIARY DRAINAGE FOR HEAD OF PANCREAS CANCER 191
5.11 Pre-operative biliary drainage for head of

pancreas cancer
Details of studies
There have been conflicting views on the need for pre-operative biliary drainage (PBD) in
patients presenting with resectable head of pancreas or peri-ampullary tumours. This has
been fuelled by retrospective and experimental studies suggesting reduced post-operative
morbidity and mortality after drainage. The findings of meta-analyses are contradictory,
partly because of the complications associated with pre-operative drainage itself. The only
randomized controlled trial, reported here, has been responsible for unifying opinion
against pre-operative drainage.
Study references
Main study
van der Gaag, N.A., Rauws, E.A., van Eijick, C.H., et al. (2010) Preoperative biliary drainage for cancer
of the head of the pancreas. New England Journal of Medicine, 362, 129–37.
Related references
Wang, Q., Gurusamy, K.S., Lin, H., Xie, X., and Wang, C.P. (2008) Preoperative biliary drainage for
obstructive jaundice. Cochrane Database of Systematic Reviews, 3, CD005444.
Sewnath, M.E., Karsten, T.M., Prins, M.H., Rauws, E.J., Obertop, H., and Gouma, D.J. (2002)
A meta-analysis on the efficacy of preoperative biliary drainage for tumors causing obstructive
jaundice. Annals of Surgery, 236, 17–27.
Sewnath, M.E., Birjmohun, R.S., Rauws, E.A., Huibregtse, K., Obertop, H., and Gouma, D.J. (2001) The
effect of preoperative biliary drainage on postoperative complications after pancreaticoduodenec-
tomy. Journal of the American College of Surgeons, 192, 726–34.
Pisters, P.W., Hudec, W.A., Hess, K.R., et al. (2001) Effect of preoperative biliary decompression on
pancreaticoduodenectomy-associated morbidity in 300 consecutive patients. Annals of Surgery, 234,
47–55.
Lillemoe, K.D. (1999) Preoperative biliary drainage and surgical outcome. Annals of Surgery,
230, 143.
Study design
Randomization Pre-operative biliary drainage (PBD) versus early surgery (ES)
Inclusion criteria Age 18–85 years
Serum bilirubin 40–250µmol/L
No evidence of metastases/local vascular involvement
Interval between CT and randomization <4 days
Exclusion criteria
Serious coexisting illness
Contraindication to surgery
Ongoing cholangitis
Previous pre-operative biliary drainage
Receiving neoadjuvant chemotherapy
Presence of gastric outlet obstruction
Follow-up Up to 12 weeks post-discharge
◆ Intervention: Patients underwent ERCP and plastic biliary stent placement, with a
second attempt if required, followed by a percutaneous trans-hepatic cholangiogram
if ERCP was unsuccessful. A drop in bilirubin of ≥50% within 2 weeks was
considered successful, with a new stent placed if drainage was inadequate. After 4–6
weeks surgery was performed.
◆ Control: patients underwent surgery within 1 week of diagnosis with a standard peri-
operative protocol including prophylactic antibiotics. Pylorus-preserving resection
was standard. Classic Whipple’s procedure was performed if spread involved the
pylorus/duodenum. Wedge resection of the superior mesenteric vein/portal vein
(SMV/PV) was included if indicated. Surgeons from 13 hospitals performing at least
10 resections per year were included.
Outcome measures
Primary endpoint
◆ Rate of serious complications up to 120 days after randomization.
Secondary endpoints
◆ Mortality.
Results
◆ There were no differences with regard to demographics or clinical characteristics,
although patients were more likely to be male (p = 0.01) or have lower BMI (p = 0.03)
in the early-surgery (ES) group
◆ Five patients in ES group required PBD.
◆ Adequate drainage in PBD was achieved in 77/102 (75%) at the first attempt and
96/102 (94%) at the second attempt.
◆ 63/94 (67% ES) versus 57/102 (56%) PBD patients underwent surgical resection
(p = 0.11).
PRE-OPERATIVE BILIARY DRAINAGE FOR HEAD OF PANCREAS CANCER 193
Complications Early surgery (n = 94) PBD (n = 102)

PBD-related 2 (2%) 47 (46%)
Pancreatitis 0 7 (7%)
Cholangitis 2 27 (26%)
Stent complications 3 (3%) 46 (45%)
Related to surgery 35 (37%) 48 (47%)
PJ leakage 11 (12%) 8 (8%)
Delayed gastric emptying 9 (10%) 18 (18%)
Wound infection 7 (7%) 13 (13%)
Pneumonia 5 (5%) 9 (9%)
Need for repeat laparotomy 13 (14%) 12 (12%)
◆ The rate of serious complications was greater in the PBD group than in the ES group
(75/102 (74%) versus 37/94 (39%), p <0.001).
◆ PBD patients had significantly more admissions and on average stayed 2 days longer
in hospital.
◆ The rate of complications in those undergoing bypass was greater in those that had
PBD than with ES (18/33 (55%) versus 5/28 (18%), p = 0.003).
◆ There was no significant difference in outcome between academic and community
hospitals (p = 0.21).
Conclusions
Routine pre-operative biliary drainage in patients undergoing surgery for cancer of the
pancreatic head increases the rate of complications.
Critique
Nearly 60% of patients with pancreatic head cancer present with obstructive jaundice.
Pre-operative biliary decompression was introduced in those patients with resectable dis-
ease in an attempt to mitigate the effects of jaundice and improve post-operative outcome.
This approach dates from an era when operating on jaundiced patients was frequently
associated with a negative outcome. With advances in peri-operative care and centraliza-
tion of resectional surgery, concerns have been raised regarding the risk–benefit analysis
of PBD. These concerns are well documented in the related references. Earlier small ran-
domized trials had a mean duration of drainage before surgery of 7–18 days. The current
study had a mean duration of pre-operative drainage of 4–6 weeks, which may have con-
ferred benefit with respect to restoration of the entero-hepatic cycle and normal major
synthetic function.
The rate of serious complications was significantly different between the early surgery
and the pre-operative drainage groups (39% and 74%, respectively). Adequate drain-
age was only achieved in 75% of patients at the first attempt, which is rather low. The
overall mortality for the study population was 7%, higher than currently accepted in high-
volume centres (<5%). These data may be partly explained by the large number of par-
ticipating units.
Despite these limitations, this paper demonstrates that PBD prior to pancreatic head
resection in patients with serum bilirubin up to 250µmol/L increases the rate of complica-
tions. In the absence of an absolute indication to decompress the biliary tree, it provides
good evidence to proceed directly to surgical resection in the jaundiced patient.
ADJUVANT THERAPY IN PANCREATIC CANCER 195
5.12 Adjuvant therapy in pancreatic cancer

Details of studies
The role of adjuvant therapy in resectable pancreatic cancer was unclear until the results
of recently published European studies (ESPAC 1 and CONKO-001). The previously
reported Gastrointestinal Tumour Study Group (GITSG) trial from 1985/87 reported sig-
nificantly increased median survival with adjuvant therapy. The small numbers included,
the absence of a control arm in the second paper, inclusion of patients with head/body/
tail cancers, and the failure of subsequent RCTs to confirm these findings questioned the
results. The two later European papers give clearer evidence regarding the use of adjuvant
therapy in pancreatic cancer.
Study references
Main study
Neoptolemos, J.P., Dunn, J.A., Stocken, D.D., et al. (2001) Adjuvant chemoradiotherapy and chemother-
apy in resectable pancreatic cancer: a randomised controlled trial. Lancet, 358, 1576–85.
Neoptolemos, J.P., Stocken, D.D., Friess, H., et al. (2004) A randomized trial of chemoradiotherapy and
chemotherapy after resection of pancreatic cancer. New England Journal of Medicine, 350, 1200–10.
Related references
Boeck, S., Ankerst, D.P., and Heinemann, V. (2007) The role of adjuvant chemotherapy for patients with
resected pancreatic cancer: systematic review of randomized controlled trials and meta-analysis.
Oncology, 72, 314–21.
Oettle, H., Post, S., Neuhaus, P., et al. (2007) Adjuvant chemotherapy with gemcitabine vs observation
in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial.
JAMA, 297, 267–77.
Study design
◆ A large prospective multi-centre RCT with a two-by-two factorial design.
Number of patients 289 patients following resection for histology-proven pancreatic
adenocarcinoma
Randomization Chemoradiotherapy alone vs chemotherapy alone vs chemoradiotherapy +
chemotherapy vs observation
Inclusion criteria Following complete macroscopic resection of histology-proven pancreatic
adenocarcinoma with no local or distant spread
Stratification According to randomization centre and the status of the resection margin
(positive or negative)
Follow-up Three-monthly until death
◆ Chemoradiotherapy: 20Gy radiotherapy over 2 weeks + 5-fluorouracil.

◆ Chemotherapy: 5-fluorouracil + folinic acid.
Outcome measures
Primary endpoint
◆ Two-year survival rate (from date of resection to death).
Secondary endpoints
◆ Incidence of adverse effects.
◆ Incidence of recurrence.
◆ Measures of the quality of life.
Results
◆ 289 patients from 53 hospitals and 11 countries were included.
◆ Median time to treatment was 46 days (34–67 days for chemotherapy) and 61 days
(47–80 days for chemoradiotherapy).
◆ There were 237/289 (82%) deaths, with median follow-up of 47 months (33–62).
◆ With chemoradiotherapy (n = 145) the median survival was 15.9 months compared
with 17.9 months in patients who received chemotherapy alone or underwent
observation (n = 144) (HR 1.28; 95% CI, 0.99–1.66; p = 0.05).
◆ The median survival with chemotherapy (n = 147) was 20.1 months compared with
15.5 months in patients who underwent chemoradiotherapy or observation (n = 142)
(HR 0.71; 95% CI, 0.55–0.92; p = 0.009).
◆ 158 patients were reported to have tumour recurrence of which 56 (35%) had local
recurrence, 53 (34%) had distant metastases, and 43 (27%) had both.
◆ Disease recurrence occurred in 84/102 (82%) undergoing chemoradiotherapy and
74/106 (70%) who did not. Median time to recurrence was 10.7 months versus 15.2
months (p = 0.04).
◆ Disease recurrence occurred in 79/110 (72%) undergoing chemotherapy and 79/98
(81%) who did not. Median time to recurrence was 15.3 months versus 9.4 months
(p = 0.02).
◆ There were no differences between chemotherapy and chemoradiotherapy in QoL
measures within 12 months in the 53% of patients who completed questionnaires.
Conclusions
Adjuvant chemotherapy has a significant survival benefit in patients with resected pan-
creatic cancer, whereas adjuvant chemoradiotherapy has a deleterious effect on survival.
Critique
Pancreatic adenocarcinoma remains one of the most lethal solid tumours, even in
the 20% of patients with surgically resectable disease. Prior to the publication of the
ESPAC-1 trial the role of adjuvant treatment was not clear. This study provided the first
ADJUVANT THERAPY IN PANCREATIC CANCER 197
randomized evidence that post-operative chemotherapy conferred a definite survival

advantage. Perhaps more controversially, it demonstrated that the addition of adjuvant
chemoradiotherapy was in fact harmful. This was a carefully planned, well-conducted,
appropriately powered multi-institutional trial involving 289 patients from European cen-
tres. In the group assigned to chemoradiotherapy the median survival was 15.9 months,
significantly less than the group who did not receive it (median survival 17.9 months).
In contrast, the median survival in the group receiving chemotherapy was 20.1 months
compared with 15.5 months in those who did not. However, the trial was not sufficiently
powered to compare these four groups directly.
Despite these data, adjuvant radiotherapy is used as part of the multimodal manage-
ment of patients with resectable and irresectable pancreatic cancer, particularly in North
American centres. The rationale for its use following surgical resection is to reduce the
rate of locoregional recurrence, with chemotherapy employed as a systemic agent in an
attempt to control metastatic disease. There are no robust phase III trials at present to
support the use of adjuvant radiotherapy outside a clinical trial. The ESPAC-1 trial is a
landmark study confirming the role of adjuvant chemotherapy, whilst nearly a decade
later the role of chemo-radiation remains controversial.
5.13 Management of intraductal papillary mucinous neoplasm

Details of study
The International Association of Pancreatology (IAP) guidelines published in 2006 rec-
ommended that all patients with main-duct intraductal papillary mucinous neoplasms
(IPMNs) should undergo resection. The primary reason was the frequency of malignancy
(in situ and invasive) in main-duct IPMNs, which ranged from 60% to 92% (mean 70%).
Approximately two-thirds of these malignant neoplasms were invasive. The ability to
exclude malignancy on clinical and radiological criteria was limited.
The largest published series on main-duct IPMNs, by Salvia et al., comprised 140
patients and found that 29% of patients with malignant IPMNs involving the main duct
were asymptomatic. Therefore reliance on symptoms could not exclude malignancy.
Furthermore, the long-term follow-up of resected patients showed excellent survival
for benign and non-invasive neoplasms, with 5-year survival for invasive carcinomas
between 36% and 60%. Based on this, the recommendation was to resect all main-duct
and mixed-variant IPMNs as long as the patient was a good surgical candidate with rea-
sonable life expectancy.
However, the management of branch-duct IPMNs was more controversial prior to
the publication of the paper by Rodriguez et al. The IAP guidelines recommended resec-
tion of all symptomatic branch-duct IPMNs and, in asymptomatic patients, resection of
tumours ≥30mm or where mural nodes were present. Conversely, Nagai et al. advocated
aggressive resection for all patients. The publication of the first large dual-centre study,
described here, helped to clarify the situation.
Study references
Main study
Rodriguez, J.R., Salvia, R., Crippa, S., et al. (2007) Branch-duct intraductal papillary mucinous neo-
plasms: observations in 145 patients who underwent resection. Gastroenterology, 133, 72–9.
Tanaka, M., Cahri, S., Adsay, V., et al. (2006) International consensus guidelines for management
of intraductal papillary mucinous neoplasms and mucinous cystic neoplasms of the pancreas.
Pancreatology, 6, 17–32.
Related references
Salvia, R., Fernández-del Castillo, C., Bassi, C., et al. (2004) Main duct intraductal papillary mucinous
neoplasms of the pancreas: clinical predictors of malignancy and long-term survival following resec-
tion. Annals of Surgery, 239, 678–87.
Nagai, K., Doi, R., Kida, A., et al. (2008) Intraductal papillary mucinous neoplasms of the pancreas:
clinicopathologic characteristics and long-term follow-up after resection. World Journal of Surgery,
32, 271–80.
MANAGEMENT OF INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM 199
Study design
Level of evidence 4
Setting Two tertiary referral centres in Boston, USA, and Verona, Italy, between
1990 and 2005
Inclusion criteria Resected branch-duct IMPN confirmed on histology
Follow-up Clinical examination, fasting glucose and tumour markers, imaging (US, CT,
or MRI)
Benign IMPN Annually for 5 years and then on a symptomatic basis
Malignant IPMN Six-monthly for 2 years and then annually
Results
◆ 113 (78%) of 145 patients had a benign IPMN and 32 (22%) were malignant.
Clinical characteristics
◆ 58 patients (40%) were asymptomatic, with no difference between benign and
malignant neoplasms (42% versus 28%, p = 0.12).
◆ Patients were more likely to be jaundiced in malignancy (p = 0.022), or have
abdominal pain if benign (p = 0.025).
◆ Eight patients (5.5%) (six adenoma, two malignant) presented with synchronous or
metachronous neoplasms.
Tumour location/size
◆ Two-thirds of branch-duct IPMNs were located in the head or uncinate process.
◆ Larger lesions on pre-operative radiology were more likely to be malignant.
Pathology
◆ Malignant IPMNs were larger (40mm versus 23.8mm, p = 0.003), with a cyst wall
thickness >3mm (p = 0.001), and associated with the presence of nodules (p = 0.001).
◆ Malignancy was less likely in tumours <30mm (p = 0.001).
◆ There was a statistically significant difference between radiological and pathological
size (median radiological size 30mm versus pathological size 20mm, p = 0.001).
Long-term follow-up and survival

◆ Median follow-up was 46 months.
◆ Overall 5- and 10-year survivals for the entire cohort were 95% and 70%,
respectively.
◆ The 5-year disease-specific survival was 100% in benign IMPN or carcinoma in situ,
and 63% in invasive carcinoma.
◆ 6/16 patients (38%) developed metastases (local, lung, or liver) from invasive disease.
All these patients died within 24 months of diagnosis of recurrence.
Conclusions
Although branch-duct IPMNs exhibit reduced frequency of dysplastic change and cancer
than main-duct IPMNs, the ability to progress to invasive and metastatic cancer remains.
As stated in the IAP guidelines, identifying asymptomatic individuals with tumours
<30mm and without mural nodes can accurately exclude malignancy and select patients
suitable for non-operative observation.
Critique
There has been a marked increase in the incidence of non-inflammatory cystic lesions of
the pancreas. This has largely been driven by the improvements in quality and availability
of abdominal imaging. This has resulted in previously unrecognized incidental lesions
now being detected earlier in the disease process. This phenomenon has posed diagnostic
and management dilemmas, particularly in the asymptomatic individual.
There have been opposing views regarding the role of surgery, ranging from serial
observation to an aggressive resection policy. The Sendai Guidelines published in 2006
by Tanaka et al. were the first attempt to provide an international consensus view on the
diagnosis, imaging, and management of these lesions. This paper brought clarity to the
distinction between branch-duct and main-duct IPMN. In particular it recommended
surgical resection for branch-duct IPMNs that are symptomatic, are larger than 3cm,
and have intramural nodules or malignant cytology on cyst fluid analysis. Branch-duct
IPMNs not satisfying any of these criteria can be safely observed with serial imaging. In
contrast, all main-duct IPMNs should be resected and patients should have annual imag-
ing post-operatively in an effort to detect early recurrence.
A follow-up paper from the University of Verona and Massachusetts General Hospital
validated the Sendai Guidelines in a large cohort of patients with resected branch-duct
IPMN. Cancer was present in 22% of cases, but there was no evidence of malignancy in
patients with lesions less than 3cm and those that were asymptomatic or without mural
nodules. The management algorithms for non-inflammatory pancreatic cystic neoplasms
will continue to be refined, but these papers have provided a solid cornerstone for an
evidence-based practice.
MANAGEMENT OF NEUROENDOCRINE TUMOURS 201
5.14 Management of neuroendocrine tumours

Details of studies
Although not exclusively a hepatobiliary topic, neuroendocrine tumour (NET) manage-
ment forms an important part of working within the specialty. With few RCTs in the
management of NETs there has previously been no consensus on the management of
the condition. A multidisciplinary team devised guidelines on behalf of joint medical
and surgical committees (including the ASGBI, related surgical specialty associations
and the UK NET group) published in Gut in 2005. The following year the European
Neuroendocrine Tumour Society published a number of even more detailed guidelines
in Neuroendocrinology in 2006. Although detailed information on this subject is outwith
the scope of this chapter, the important recommendations from the 2005 guidelines are
summarized below.
Study references
Main study
Ramage, J.K., Davies, A.H.G., Ardill, J. et al. (2005). Guidelines for the management of gastroenteropan-
creatic neuroendocrine (including carcinoid) tumours. Gut, 54, iv1–16.
Related references
De Herder, W.W., O’Toole, D., Rindi, G., and Wiedenmann, B. (eds) (2008) ENETS consen-
sus guidelines for the diagnosis and treatment of neuroendocrine gastrointestinal tumours.
Neuroendocrinology, 87(1).
Study design
◆ A search of Medline was made returning 41,533 citations, of which 212 were
referenced in the current guidelines
◆ The guidelines apply to carcinoids and NETs arising from the gut, including the
pancreas and liver
◆ Categories of evidence are based on the Oxford Centre for Evidence-based Medicine
grades of recommendation
Aetiology, epidemiology, genetics, and clinical features

◆ Most NETs are sporadic but are associated with an increased familial risk (one first-
degree relative affected confers a 4x increased risk, two first-degree relatives confer a
12x increased risk).
◆ NETs may occur as part of familial syndromes such as multiple endocrine neoplasia
1 (MEN1) (25–40% association in gastrinomas alone).
Recommendations (grade C)
1. Clinical examination to exclude complex cancer syndromes (including MEN1 and
MEN2, and neurofibromatosis 1) should be performed in all cases of NETs and a
family history taken.
2. In all cases where there is a family history of carcinoid or NET, or a second endocrine
tumour, a familial syndrome should be suspected.
3. Testing for germline MEN1 mutations should be considered in individuals with
sporadic or familial bronchial or gastric carcinoid. Management of MEN1 families
includes screening for endocrine parathyroid and enteropancreatic tumours from
late childhood, with predictive testing for first-degree relatives of known mutation
carriers.
4. All patients should be evaluated for second endocrine tumours and possibly for other
gut cancers (20% develop other cancers).
Diagnosis
◆ Diagnosis is based on clinical symptoms, hormone concentrations, radiology and
nuclear medicine imaging, and histological confirmation.
◆ Plasma chromogranin A (CgA) is produced in very significant quantities by NET
cells and may be useful in diagnosis.
◆ Pancreatic polypeptide is found in high concentrations in 80% of pancreatic tumours
and 50% of carcinoids.
Recommendations
1. Baseline tests should include chromogranin-A (CgA) and 5-hydroxyindoleacetic acid
angiography (5-HIAA) (grade C). Others that may be appropriate include thyroid
function tests (TFTs), parathyroid hormone (PTH), calcium, calcitonin, prolactin,
α-fetoprotein, carcino-embryonic antigen (CEA), and beta human chorionic
gonadotrophin (β-hCG) (grade D).
2. Specific biochemical tests should be requested, depending on which syndrome is
suspected.
Imaging
Recommendations
1. A multimodality approach is best for detecting the primary tumour and may include
CT, MRI, somatostatin receptor scintigraphy (SSRS), EUS, endoscopy, digital
subtraction angiography (DSA), and venous sampling (grades B/C).
2. SSRS is the most sensitive modality for detecting metastases (grade B).
3. SSRS may be indicated for follow-up when the primary tumour has been resected
(grade D).
MANAGEMENT OF NEUROENDOCRINE TUMOURS 203
Sensitivities of the various imaging modalities for locating specific NETs:
Carcinoid Gastrinoma Insulinoma

Primary Liver Primary Liver Primary
metastases metastases
Ultrasound 46% 83% 23% 50% 27%
CT 64% 88% 38–75% 54–88% 30%
MRI 56% 85% 22–90% 63–90% 10%
SSRS 80% 90% 72% 97% 25%
EUS 80% 90–100% 88%
Angiography + Ca2+ 93% 95%
stimulation
Pathology
Recommendations
◆ Tumours should be classified according to the recent WHO classification which
places each NET into one of four categories:
● well-differentiated endocrine tumour of probable benign behaviour;
● well-differentiated endocrine tumour of uncertain behaviour;
● well-differentiated endocrine carcinoma;
● poorly differentiated endocrine carcinoma.
Therapy
Recommendations
◆ The extent of the tumour, its metastases, and its secretory profile should be
determined as far as possible before planning treatment (grade C).
◆ Surgery should be offered to patients who are fit and have limited disease (grade C)
or in those with liver metastases and potentially resectable disease (grade D).
◆ Where abdominal surgery is undertaken and long-term treatment with somatostatin
(SMS) analogues is likely, cholecystectomy should be considered.
◆ For patients who are not fit for surgery, the aim of treatment is to improve and
maintain an optimal quality of life (grade D).
◆ The choice of treatment depends on the symptoms, stage of disease, degree of uptake
of radionuclide, and histological features of the tumour (grade C).
◆ Treatment choices for non-resectable disease include SMS analogues, biotherapy,
radionuclides, ablation therapies, and chemotherapy (grade C).
◆ External beam radiotherapy may relieve bone pain from metastases (grade C).
◆ Chemotherapy may be used for inoperable or metastatic pancreatic and bronchial
tumours or poorly differentiated NETs (grade B).
Prognosis
◆ Overall 5-year survival for all NETs is 83% following complete resection. Five-year
survival for pancreatic NETs is 50–80%, with insulinoma and gastrinoma having a
5-year survival of up to 94%.
Critique
Neuroendocrine tumours are rare and may pose interdisciplinary management chal-
lenges. Although progress has been made in the characterization and diagnosis of these
lesions, there is a relative paucity of randomized data upon which to base management
algorithms. These guidelines are significant in that they represent the first collaborative
attempt to provide a comprehensive and cohesive framework to guide and inform the
decision-making process. Given the lack of evidence regarding NETs when compared
with other solid tumours, these guidelines are intended to offer a multidisciplinary team
approach to the genetic evaluation, diagnosis, imaging, and treatment modalities avail-
able. They were followed by European guidelines a year later and a series of consensus
publications by the North American NET Society in 2010.
Kulke, M.H., Anthony, L.B., Bushnell, D.L., et al. (2010) NANETS treatment guidelines: well-differenti-
ated neuroendocrine tumors of the stomach and pancreas. Pancreas, 39, 735–52.
SURGICAL RESECTION OF COLORECTAL LIVER METASTASES 205
5.15 Surgical resection of colorectal liver metastases

Details of studies
Approximately 25% of patients with colorectal cancer will have liver metastases at pre-
sentation, and ultimately 50% of patients will be diagnosed with metastatic liver disease.
Historically, we know that if left untreated, such patients have an extremely poor out-
come, with a median survival of 5–21 months, and few are likely to be alive at 5 years (Jaffe
et al). Previously, hepatic resection was associated with a significant risk of intra-operative
haemorrhage and associated morbidity and mortality. However, over the last 20 years,
both surgical and anaesthetic techniques have evolved, such that most hepatic resections
can now be performed with minimal blood loss and morbidity. As a consequence, the
indications for resection have been considerably expanded, and many more patients with
liver metastases from colorectal cancer now undergo hepatic resection.
Currently, surgical resection is the primary treatment option for patients with colorec-
tal liver metastases and, depending on the extent of liver involvement/resection, 5-year
survival figures of 30% and upwards have been reported. The main study reference in
this section is one of the largest patient series from an internationally recognized centre,
performing hepatectomy for colorectal liver metastases.
Study references
Main study
Fong, Y., Fortner, J., Sun, R., Brennan, M., and Blumgart, L.H. (1999) Clinical score for predicting recur-
rence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases.
Related references
Jaffe, B.M., Donegan, W.L., Watson, F., and Spratt, J.S., Jr. (1968) Factors influencing survival in patients
with untreated hepatic metastases. Surgery, Gynecology and Obstetrics, 127, 1–11.
Study design
◆ A single-centre case series
Level of evidence 4
Patients Colorectal cancer with liver metastases
Intervention Hepatic resection
Inclusion criteria Medically fit
No disseminated disease
Hepatic metastases when resected would leave adequate functioning liver
parenchyma
Resectable or resected colorectal primary
Study duration 1985–1998
Outcome measures
◆ To determine prognostic factors for the recurrence of metastatic disease.
Results
◆ 581 men and 420 women, with a median age of 61 years (range 27–87 years),
underwent resection.
◆ 741 patients had a colonic primary, and in 260 the primary site was the rectum.
◆ Number of hepatic tumours: solitary in 517 patients; 2–3 in 330 patients; ≥4 in 154
patients.
◆ 63% of hepatic resections involved removal of at least one lobe of the liver.
◆ The median hospital stay was 11 days (range 1–70 days).
◆ 28 patients died within 30 days of surgery (2.8%).
◆ Median survival from hepatic resection was 42 months.
◆ Actuarial survival at 1, 3, 5, and 10 years was 89%, 57%, 37%, and 22%, respectively
◆ Multivariate analysis allowed the development of a clinical risk score. Factors
associated with a poorer outcome included the following:
● positive nodal status from the primary tumour;
● disease-free interval <12 months (from colorectal to hepatic metastases);
● number of hepatic tumours >1;
● pre-operative CEA >200ng/mL;
● size of largest hepatic tumour >5cm.
Conclusions
This large series provided solid evidence that resection of hepatic colorectal liver metasta-
ses represents safe and effective therapy with the possibility of long-term cure.
Critique
Although no single paper on hepatic resections for colorectal cancer has dramatically
changed practice, this case series was chosen as it is from a well-established, highly
regarded, and widely published group. It gives an insight into the range of patients who
can be considered for hepatic resection. The results of surgery in this large group of
patients show that hepatic resection is a safe and oncologically effective treatment that
can provide long-term cure in some patients (Jarnagin et al.) This study also highlights
poor prognostic factors in the development of its clinical risk score.
As hepatic resection for colorectal liver metastases has become widely accepted, its
role has been extended. Bilobar disease or multiple metastases are no longer viewed as a
contraindication, and tumours occupying up to 75% of the total hepatic volume can be
excised safely. Hepatic resection can also be performed safely in the elderly, and resection
SURGICAL RESECTION OF COLORECTAL LIVER METASTASES 207
can be repeated if recurrence occurs in the liver. Synchronous resection of the primary
colorectal cancer and metastatic liver disease is now also considered in selected patients
(Reddy et al.) In addition, patients with resectable hepatic and extra-hepatic metastatic
disease should be considered for synchronous or metachronous resection.
Although surgical resection remains the mainstay of management of resectable meta-
static liver disease, a number of novel local ablative therapies have been introduced over
the last 20 years. Of these, radiofrequency ablation (RFA) has received the most attention.
Although long-term survival after RFA may be inferior to surgical resection, it may have
a role in patients not medically fit for surgery. A recent systematic review reported a 24%
5-year survival (Pathak et al.).
Jarnagin, W., Gonen, M., Fong, Y., et al. (2002) Improvement in perioperative outcome after hepatic
resection: analysis of 1,803 consecutive cases over the past decade. Annals of Surgery, 236(4),
397–407.
Reddy, S., Pawlik, T.M., Zorzi, D., et al. (2007) Simultaneous resections of colorectal cancer and syn-
chronous liver metastases: a multi-institutional analysis. Annals of Surgical Oncology, 14, 3481–91.
Pathak, S., Jones, R., Tang, J.M., et al. (2011) Ablative therapies for colorectal liver metastases: a system-
atic review. Colorectal Disease, 13, e252–65.
Chapter 6
Colorectal surgery
6.1 Mechanical bowel preparation in colorectal surgery 210
6.2 Stapled versus handsewn anastomosis in colorectal surgery 212
6.3 Laparoscopic colorectal surgery 214
6.4 Enhanced recovery after colorectal surgery 218
6.5 Infliximab in the management of fistulating Crohn’s disease 221
6.6 Non-surgical treatment of chronic anal fissure 224
6.7 Stapled haemorrhoidectomy 226
6.8 Laparoscopic ventral mesh rectopexy for rectal prolapse 229
6.9 Sacral neuromodulation for faecal incontinence 231
6.10 Surgery for pilonidal sinus disease 234
6.11 Bowel cancer screening 236
6.12 Total mesorectal excision for rectal cancer 240
6.13 The circumferential resection margin in rectal cancer 244
6.14 Extralevator abdomino-perineal resection of the rectum 247
6.15 Neoadjuvant therapy in primary rectal cancer 250
6.16 Adjuvant therapy in colon cancer 255
6.17 Primary chemoradiotherapy for anal cancer 258
210 COLORECTAL SURGERY
6.1 Mechanical bowel preparation in colorectal surgery

Details of studies
Traditionally bowel preparation was given to patients prior to undergoing elective
colorectal surgery in the belief that it reduced the risk of infection and anastomotic leak-
age (Chung et al.). In addition to being unpleasant and uncomfortable, mechanical bowel
preparations may also be associated with side effects including dehydration, abdomi-
nal discomfort, and even renal failure. The paper discussed here is the third update of a
Cochrane review which brings together data from a number of older trials and five new
trials in order to determine the impact of bowel preparation on morbidity and mortality
in patients undergoing colorectal surgery.
Study references
Main study
Guenaga, K.F., Matos, D., and Wille-JØrgensen, P. (2011) Mechanical bowel preparation for elective
colorectal surgery. Cochrane Database of Systematic Reviews, CD001544.
Related references
Chung, R.S., Gurll, N.J., Berglund, E.M. (1979) A controlled trial of whole gut lavage as a method of
bowel preparation for colonic operations. American Journal of Surgery, 137, 75–81.
Study design
Randomization Bowel preparation (all types) versus no preparation
Number of trials 18 RCTs
Number of participants 5805
Inclusion criteria Patients undergoing elective colorectal surgery
◆ The methodological quality of each trial was assessed (including randomization,

blinding, analysis type, and numbers lost to follow-up).
Outcome measures
Primary endpoint
◆ Anastomotic leak.
Secondary endpoints
◆ Mortality and morbidity (peritonitis, re-operation, wound infection, extra-abdominal
complications, and overall surgical site infections).
MECHANICAL BOWEL PREPARATION 211
Results
Bowel preparation No preparation Odds ratio (95% CI)
(n = 2275) (n = 2258)
Anastomotic leak (overall) 4.4% 8.5% 0.99 (0.74, 1.31)
Low anterior resection 8.8% 10.3% 0.88 (0.55, 1.40)
Colonic resection 3.0% 3.5% 0.85 (0.58, 1.26)
Wound infection 9.6% 8.5% 1.16 (0.95, 1.42)
◆ A further analysis subdivided groups into mechanical bowel preparation versus

rectal enema.
Bowel preparation Rectal enema Odds ratio (95% CI)

(n = 601) (n = 609)
Anastomotic leak (overall) 4.4% 3.4% 1.32 (0.74, 2.36)
Anterior resection 7.4% 7.9% 0.93 (0.34,2.52)
Colonic resection 4.0% 2.0% 2.15 (0.79, 5.84)
Wound infection 9.9% 8.0% 1.26 (0.85, 1.88)
Conclusion
Despite the inclusion of more studies with nearly 6000 participants, there is no statisti-
cally significant evidence that patients benefit from mechanical bowel preparation or the
use of rectal enemas.
Critique
Although many have questioned the use of mechanical bowel preparation over the last
40 years, it continues to be an ongoing area of research. This is the third update of a
Cochrane review that again finds no advantage in patients receiving mechanical bowel
preparation or rectal enemas. Specifically, there were no differences in anastomotic leak-
age. Although the authors state that more trials are required on bowel preparation in
rectal surgery and laparoscopic surgery before conclusions are drawn, this review strongly
suggests that there is no clinical benefit.
In contrast, many surgeons still appear to be using mechanical bowel preparation.
A recent questionnaire survey of 398 members of the Association of Coloproctology of
Great Britain and Ireland found that, of the 195 surgeons who replied, full bowel prepara-
tion was routinely used for many patients undergoing open surgery (Drummond et al.).
This varied depending on the operation: 9.5% for right hemicolectomy, 43.4% for left hemi-
colectomy, 20.5% for abdomino-perineal resection and 72.2% for low anterior resection.
Drummond, R.J., McKenna, R.M., and Wright D.M. (2011) Current practice in bowel preparation for
colorectal surgery: a survey of the members of the Association of Coloproctology and Great Britain
and Ireland. Colorectal Disease, 13, 708–10.
6.2 Stapled versus handsewn anastomosis in colorectal surgery

Details of studies
The issue of whether to hand sew or staple is a common conundrum when performing a
colorectal resection. Although the indication and circumstances of performing the resec-
tion may have a bearing on what type of anastomosis is fashioned, for the most part the
decision is down to the individual surgeon’s preference as there is no consensus as to the
superiority of one technique over the other. A number of studies have addressed this
issue, with the two studies discussed being among the first.
Study references
Main studies
Kracht, M., Hay, J.-M., Fagniez, P.-L., and Fingerhut, A. (1993) Ileocolic anastomosis after right
hemicolectomy for carcinoma: stapled or handsewn? A prospective, multicenter, randomized trial.
International Journal of Colorectal Disease, 8, 29–33.
Docherty, J.G., McGregor, J.R., Akyol, A.M., Murray, G.D., and Galloway, D.J. (1995) Comparison of
manually constructed and stapled anastomoses in colorectal surgery. Annals of Surgery, 221, 176–84.
Related references
Choy, P.Y., Bissett, I.P., Docherty, J.G., Parry, B.R., and Merrie, A. (2011) Stapled versus handsewn methods
for ileocolic anastomoses. Cochrane Database of Systematic Reviews, CD004320pub3.
Moran, B.J. (1996) Stapling instruments for intestinal anastomosis in colorectal surgery. British Journal
of Surgery, 83, 902–9.
Muñoz-Juárez, M., Yamamoto, T., Wolff, B.G., and Keighley, M.R.B. (2001) Wide-lumen stapled
anastomosis vs. conventional end-to-end anastomosis in the treatment of Crohn’s disease. Diseases
of the Colon and Rectum, 44, 20–6.
Neutzling, C.B., Lustosa, S.A.S, Proneca, I.M., da Silva, E.M., and Matos, D. (2012) Stapled versus
handsewn methods for colorectal anastomosis surgery. Cochrane Database of Systematic Reviews,
CD003144.pub2.
Study design
Kracht et al. Docherty et al.
Randomization Stapled versus HS end-to-end Stapled versus HS anastomosis
interrupted or continuous/HS end to
side interrupted or continuous
Type of anastomosis Ileocolic anastomoses for cancer Ileocolic, colo-colic, colorectal, and
colostomy closure for benign and
malignant disease
Setting Multicentre—France Multicentre—Scotland
Follow-up In-hospital (day 8–10) 45–52 months
Outcomes Clinical/radiological anastomotic leak Clinical/radiological anastomotic leak
Morbidity and mortality
Long-term cancer outcome
STAPLED VERSUS HANDSEWN ANASTOMOSIS 213
Results
Kracht et al.
◆ The stapled group had a significantly lower overall anastomotic leak rate than the
handsewn (HS) group (2.8% versus 8.3%).
◆ The stapled technique was approximately 10 times more expensive than the
handsewn group.
Docherty et al.
◆ The radiological leak rate was significantly higher in the handsewn group than in the
stapled group (14.4% versus 5.2%, p <0.05).
◆ No difference was seen in the clinical leak rate, morbidity rate, or mortality rate.
◆ In participants who underwent potentially curative resections for bowel cancer
(177/418), tumour recurrence and cancer-specific mortality was found to be higher
in the handsewn than in the stapled groups (7.5% versus 6.7%) and higher in those
patients who had suffered an anastomotic leak
Conclusions
Kracht et al. concluded that a stapled side-to-side anastomosis is associated with fewer
anastomotic leaks than the traditional handsewn technique. Docherty et al. concluded
that handsewn and stapled anastomoses are equally safe. Recent meta-analysis has failed
to show any superiority of one technique over the other.
Critique
Further studies regarding anastomotic technique have reached conflicting results, and
the most recent meta-analysis on the topic (Neutzling et al.) seems to suggest no definite
difference in terms of anastomotic leak rate or mortality. In general, handsewn anastomo-
ses take longer to perform and stapled anastomoses may have a higher rate of stenosis.
Handsewn anastomoses may be more operator dependent, although technical factors in
stapled anastomoses are also important. The use of adjuncts, such as fibrin sealants, to
improve anastomotic integrity is currently under investigation.
6.3 Laparoscopic colorectal surgery

Details of studies
It was readily apparent soon after the introduction of laparoscopic cholecystectomy
that this technique was going to gain rapid and widespread acceptance. However, the
development of laparoscopic colorectal surgery had a much more difficult birth and
evolution. Although the early pioneers started to perform laparoscopic colorectal sur-
gery in the early 1990s, it has only really gained widespread acceptance over the last
10–15 years.
There were a number of reasons for its slow acceptance. These relate in part to con-
cerns regarding the technical feasibility outside specialist centres, combined with early
concerns regarding the oncological safety in colorectal cancer resection and the risk of
port site recurrence.
The randomized trials discussed here, along with subsequent long-term follow-up
studies provided clear evidence that laparoscopic colorectal cancer surgery can be per-
formed safely and with equivalent cancer outcomes compared with open surgery.
Study references
Main studies
Lacy, A.M., García-Valdecasas, J.C., Delgado, S., et al. (2002) Laparoscopy-assisted colectomy versus
open colectomy for treatment of non-metastatic colon cancer: a randomised trial. Lancet, 359,
2224–9.
Veldkamp, R., Kuhry, E., Hop, W.C. et al. (2005) Colon Cancer Laparoscopic or Open Resection Study
Group (COLOR). Laparoscopic surgery versus open surgery for colon cancer: short term outcomes
of a randomised trial. Lancet Oncology, 6, 477–84.
Guillou, P.J., Quirke, P., Thorpe, H., et al. (2005) Short-term endpoints of conventional versus
laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre
randomised controlled trial. Lancet, 365, 1718–26.
The National Cancer Institute–sponsored Clinical Outcomes of Surgical Therapy (COST) Study Group.
(2004) A comparison of laparoscopically assisted and open colectomy for colon cancer. New England
Journal of Medicine 2004; 350, 2050–9.
Related references
Bonjer, H.J., Hop, W.C.J., Nelson, H., et al. (2007) Laparoscopically assisted versus open colectomy for
colon cancer: a meta-analysis. Archives of Surgery, 142, 298–303.
Jayne, D.G., Thorpe, H.C., Copeland J, et al. (2020) Five-year follow-up of the Medical Research Council
CLASICC trial of laparoscopically assisted versus open surgery for colorectal cancer. British Journal
of Surgery, 97, 1638–45.
Association of Coloproctology of Great Britain and Ireland (2007) Guidelines for the
Management of Colorectal Cancer. (3rd edn) <http://www.mccn.nhs.uk/userfiles/documents/
Nat%20Ass%20of%20Coloproctology%20Guidelines%281%29.pdf>.
LAPAROSCOPIC COLORECTAL SURGERY 215
Study design
◆ All were prospective randomized trials.
Lacy et al. COST Color CLASICC

Setting Barcelona USA, Canada European UK
Number of centres 1 48 29 27
Number of patients 442 872 1248 794
Study period 1993–1998 1994–2001 1997–2003 1996–2002
Tumour site Colon Colon Colon Colon/rectum
Exclusions Obstruction Obstruction Obstruction Obstruction
Advanced primary Advanced primary Advanced primary Cancer <5 years
previously
Metastatic disease Metastatic disease Metastatic disease Synchronous lesion
Previous colon Transverse colon Transverse colon Transverse colon
resection cancers cancers cancers
IBD/FAP Previous colon
resection
Synchronous lesion
BMI>30
Follow-up (median) 43 months 4.4 years 53 months 38.8 months
Primary endpoints Cancer survival Time to recurrence Cancer-free survival Resection margins,
3 years post- % Dukes’ C2
surgery tumours, mortality
Secondary Disease-free Overall survival, Overall survival,
endpoints survival, overall resection margins, recurrence,
survival, local and distant complications,
complications, recurrence, transfusion rate,
recovery, QoL morbidity, mortality, QoL
blood loss
◆ The initial short-term results from the four studies were published, followed by
longer-term data (presented here) on outcomes after colonic resection
Lacy et al. COST Color CLASICC

Conversion rate 11 21 17 25
Morbidity
Laparoscopic 10.8% 21% 21% 26%
Open 28.7% 20% 20% 27%
Mortality
Laparoscopic 0.9% 0.5% 1.0% 4.0%
Open 2.9% 1.0% 2.0% 5.0%
Disease-free survival at 3 years (%)

Laparoscopic 83% 68.4% 74.2% 66.3%
Open 73% 69.2% 76.2% 67.7%
Overall survival at 3 years
Laparoscopic 82% 86% 81.8% 68.4%
Open 74% 85% 84.2% 66.7%
◆ No significant differences were seen in overall survival between laparoscopic and

open groups in all trials except the Lacy study (cancer-specific survival 91% in the
laparoscopic group versus 79% in the open group, p = 0.006).
◆ Local and distant recurrence rates were similar in the laparoscopic and open groups
in all four studies.
◆ No increased risk of port site recurrence was seen with laparoscopic surgery.
◆ No significant differences in resection margins were seen between the two groups.
◆ The CLASICC trial assessed rectal cancer outcomes. There was no difference in the
number of positive resection margins, operative mortality, complication rates, or
QoL scores. There was a higher rate (non-significant) of circumferential margin
positivity with laparoscopic rectal surgery. On long-term follow-up, there was no
significant difference in local recurrence rates (7.8% versus 7.0%, p = 0.70) or overall
survival (p = 0.17).
◆ 34% of rectal cancer resections were converted from laparoscopic to open surgery.
Converted patients were significantly more likely to develop a complication.
Conclusion
These four trials concluded that laparoscopic colon and rectal surgery was safe and pro-
vided effective oncological surgery.
Critique
The main criticism of these four trials lies in the tight patient-selection criteria. By exclud-
ing locally advanced cancers or cancers in difficult positions (splenic flexure and trans-
verse colon), these trials may not fully reflect everyday surgical practice. However, these
were the first randomized laparoscopic trials and the authors have been pragmatic by
showing that this novel approach is possible in ‘straightforward’ resections before assess-
ing laparoscopic surgery’s role in more difficult cases. Indeed, as laparoscopic surgery
has become more widely practised, it can be seen from case reports in the literature that
patient inclusion criteria are widening.
The authors from these four trials went on to combine their results in a meta-analysis
(Bonjer et al.). Their analysis confirmed that there were no statistical differences in
overall survival at 3 years (82.2% for laparoscopic versus 83.5% for open surgery) or
LAPAROSCOPIC COLORECTAL SURGERY 217
in disease-free survival (75.8% versus 75.3%). The authors concluded that laparoscopic
colectomy for cancer is safe.
Further evidence for equivalent long-term outcomes in both colon and rectal cancer
was provided by the CLASICC trial group who published their 5-year follow-up results in
2010 (Jayne et al.). Again, no significant differences were found between laparoscopic and
open surgery in overall survival, disease-free survival, and local and distant recurrence.
To date, laparoscopic surgery appears to provide effective oncological surgery, leading
professional bodies to support laparoscopic resection in appropriately trained surgeons
(Association of Coloproctology 2007).
6.4 Enhanced recovery after colorectal surgery

Details of studies
Traditionally, recovery after colonic surgery was cautious, which meant that oral nutri-
tion and mobilization were not immediately encouraged in the belief that such measures
would minimize complications. In the 1990s this approach, which invariably prolonged
hospital stay, was challenged by a small number of studies which championed the use of
epidural analgesia, early nutrition and early mobilization (Bardram et al.). Integration of
these factors into a multimodal approach aimed to attenuate the surgical stress response,
reducing end-organ dysfunction. The paper discussed here reported on outcomes in a
unit that routinely used such a multimodal enhanced recovery programme after open
colonic surgery to reduce hospital stay and complication rates.
Study references
Main study
Basse, L., Jakobsen, D.H., Billesbolle, P., Werner, M., and Kehlet, H. (2000) A clinical pathway to acceler-
ate recovery after colonic resection. Annals of Surgery, 232, 51–7.
Related references
Bardram, L., Funch-Jensen, P., Jensen, P., Crawford, M.E., and Kehlet, H. (1995) Recovery after laparoscopic
colonic surgery with epidural analgesia, and early oral nutrition and mobilisation. Lancet, 345, 763–4.
Anderson, A.D.G., McNaught, C.E., MacFie, J., Tring, I., Barker, P., and Mitchell, C.J. (2003)
Randomized clinical trial of multimodal optimization and standard perioperative surgical care.
Delaney, C.P., Zutshi, M., Senagore, A.J., Remzi, F.H., Hammel, J., and Fazio, V.W. (2003) Prospective,
randomized, controlled trial between a pathway of controlled rehabilitation with early ambulation
and diet and traditional postoperative care after laparotomy and intestinal resection. Diseases of the
Colon and Rectum, 46, 851–9.
Varadhan, K.K., Neal, K.R., Dejong, C.H.C., Fearon, K.C.H., Ljungqvist, O., and Lobo, D.N. (2010)
The enhanced recovery after surgery (ERAS) pathway for patients undergoing major elective open
colorectal surgery: a meta-analysis of randomized controlled trials. Clinical Nutrition, 29, 434–40.
Study design
◆ A prospective case series.
Level of evidence 4
Intervention Multimodal enhanced recovery after surgery (ERAS)
Number of patients 60 consecutive
Number of centres 1
Inclusion criteria Patients undergoing open colonic resection
Exclusion criteria Low anterior/AP resection of the rectum
Surgery for inflammatory bowel disease
Follow-up Discharged 48 hours after surgery
Review at day 8 for removal of stitches
Final post-operative check at day 30
ENHANCED RECOVERY AFTER COLORECTAL SURGERY 219
◆ Pre-operatively, four patients (group A) were classified as unsuitable for discharge

at 48 hours (previous cerebrovascular accident (CVA); inpatient in psychiatry;
wheelchair bound; DVT in last 10 days requiring anticoagulation).
◆ In theatre, three patients (group B) were found to have locally advanced disease
requiring a more extensive resection.
◆ The remaining 53 patients (group C) were deemed suitable for potential discharge
after 48 hours.
Multimodal enhanced recovery programme

◆ Pre-operative patient education.
◆ No pre-medication, standardized intra-operative fluid regime, intra-operative
normothermia.
◆ Thoracic epidural and opioid-sparing analgesia.
◆ Transverse surgical incisions.
◆ Standardized post-operative recovery with no nasogastric tubes, early diet and
mobilization, removal of urinary catheter on day 1 and epidural catheter on day 2.
Outcome measures
◆ Hospital stay.
◆ Return of normal gastrointestinal function, complications, readmission rate.
Results
◆ The median age of patients was 74 years, with a third of patients ASA grade III-IV
◆ Return of gastrointestinal function (defecation) was <48 hours in 57 patients (95%).
◆ The median hospital stay was 2 days (range 2–62) with 32 patients discharged within
48 hours.
◆ There were two deaths (3.3%) and five patients (8.3%) had complications.
◆ The readmission rate was 15%, including two patients readmitted with an
anastomotic leak.
◆ 73% of patients were satisfied with their post-operative care; the remaining 27%
thought that they were discharged too early.
Conclusion
A multimodal rehabilitation programme after elective open colonic surgery may reduce
inpatient hospital stay, even in high-risk patients.
Critique
Enhanced recovery programmes were a response to changes in the evidence base around
peri-operative care at the time, and the reduced hospital stay beginning to be seen follow-
ing laparoscopic colorectal surgery. The paper presented here followed on from a smaller
series published by Kehlet’s group in 1995 (Bardram et al.). This larger series of patients
was remarkable at the time for the hospital stay of 2 days following open colonic resection
compared with the average hospital stay of 7–10 days. In fact, very few have managed to
achieve such short hospital stays despite more widespread acceptance of the principles.
Initial criticisms of ERAS were that patients were highly selected, readmission rates
were unacceptably high, and the evidence was limited to case series and single-institution
experiences. However, in this paper the patient population was older than previously doc-
umented, with a third of patients deemed high risk (ASA ≥III). This study was followed
by several RCTs and meta-analyses (Delaney et al., Anderson et al., Varadhan et al.) con-
firming reduced ileus, reduced post-operative morbidity and shorter hospital stay with no
evidence of a significant increase in readmissions.
INFLIXIMAB IN THE MANAGEMENT OF FISTULATING CROHN’S DISEASE 221
6.5 Infliximab in the management of fistulating Crohn’s disease

Details of studies
Work in the early 1990s showed that tumour necrosis factor alpha (TNF-α) played an
important role in mediating the inflammatory response in Crohn’s disease. Mucosal levels
and inflammatory cell production of TNF-α were found to be elevated in this condition,
leading researchers to develop monoclonal antibodies to TNF-α. The first mass-produced
antibody to be introduced into clinical practice was infliximab. This was shown to reduce
inflammation and induce rapid healing in many patients. Not unsurprisingly, researchers
wondered if this new drug might also have a beneficial effect in Crohn’s-related fistulae,
which can affect up to one-third of patients with Crohn’s disease. Prior to this paper, the
therapeutic options for Crohn’s-related fistulae were limited and surgery was often viewed
as being the only viable option.
Study references
Main study
Present, D.H., Rutgeerts, P., Targan, S., et al. (1999) Infliximab for the treatment of fistulas in patients
with Crohn’s disease. New England Journal of Medicine, 340, 1398–405.
Related references
Lichtiger, S., Binion, D.G., Wolf, D.C. et al. (2010) The CHOICE trial: adalimumab demonstrates safety,
fistula healing, improved quality of life and increased work productivity in patients with Crohn’s
disease who failed prior infliximab therapy. Alimentary Pharmacology and Therapeutics, 32, 1228–39.
Haveran, L.A., Sehgal, R., Poritz, L.S., et al. (2011) Infliximab and/or azathioprine in the treatment of
Crohn’s disease-like complications after IPAA. Diseases of the Colon and Rectum, 54, 15–20.
Study design
◆ A randomized multi-centre double-blind placebo-controlled trial.
Randomization Infliximab 5mg/kg vs infliximab 10mg/kg vs placebo
Inclusion criteria ≥1 draining abdominal or peri-anal fistula for ≥3 months as a complication
of Crohn’s disease
Aged 18–65 years
Stable on current therapy for ≥4 months prior to the trial; current therapy
was continued during trial
Exclusion criteria Concurrent ciclosporin therapy
Recent therapy stopped <4 weeks prior to trial
Stricture or abscess
Previous anti-TNF treatment
Refusal to take birth control
Stratification By treatment, site, and number of fistulae
Follow-up 34 months
◆ Infliximab was given intravenously at 0, 2, and 6 weeks

◆ Fistula closure was defined as no discharge/closed fistula for at least two consecutive
follow-up visits (i.e. more than 21 days since open).
Outcome measures
Primary endpoint
◆ ≥50% decrease in number of draining fistulae at two or more consecutive
follow-ups.
Secondary endpoint
◆ Closure of all fistulae.
Results
Placebo Infliximab 5 mg/kg Infliximab 10 mg/kg
(n = 31) (n = 31) (n = 32)
Primary endpoint achieved 8 21* 18*
Secondary endpoint achieved 4 17* 12*
* p <0.05 when compared with placebo group
◆ Of the 29 patients that achieved closure of all fistulae, 15 had a single fistula and
14 had multiple fistulae.
◆ The median duration of response was 3 months.
◆ A total of six patients were withdrawn from the study: three in the placebo group
because of lack of efficacy, one in the placebo group for administration reasons, one
in the low-dose infliximab group because of withdrawal of consent, and one in the
high-dose infliximab group because of an adverse event.
◆ Overall, 60% of patients reported adverse effects, mostly minor (headache, abscess,
upper respiratory tract infection, and fatigue) with a trend to increased side effects in
the high-dose versus low-dose infliximab group.
Conclusion
Infliximab is an effective treatment for fistulating Crohn’s disease.
Critique
This was one of the first studies to support the use of infliximab for the treatment of
Crohn’s fistulae. The lower dose (5mg/kg) was found to be effective in addition to having
a better side-effect profile. This landmark paper has gone on to be cited in over 1800 pub-
lications including supporting statements from gastroenterology and colorectal profes-
sional organizations in the USA, Europe, and the UK.
INFLIXIMAB IN THE MANAGEMENT OF FISTULATING CROHN’S DISEASE 223
Although this treatment is clearly efficacious in the short term, clinical experience has
shown that fistulae tend to re-open if long-term therapy is not continued. Unfortunately,
patients may also become tolerant to long-term therapy. As a result, second-line treat-
ments with other anti-TNF monoclonal antibodies are now under investigation. One of
the most promising of these new anti-TNF antibodies is adalimumab, which has been
shown to be a safe and effective treatment in infliximab refractory or intolerant patients
(Lichtiger et al.). In addition, these treatments are now being considered for patients with
ulcerative colitis who have developed Crohn’s disease-like complications after undergoing
proctocolectomy with pouch formation (Haveran et al.).
6.6 Non-surgical treatment of chronic anal fissure

Details of studies
Chronic anal fissure is a common problem. Most occur in the posterior midline and are
associated with intense sphincter spasm. Traditionally, the mainstay of treatment was sur-
gical, using the lateral internal sphincterotomy. This technique reduces sphincter spasm
but can be associated with impaired continence.
The introduction of topical glyceryl trinitrate (GTN) ointment led to a paradigm shift
in the management of chronic anal fissure. Healing rates of up to 68% compared with pla-
cebo were described (Lund and Scholefield). As the risk of incontinence was negligible,
GTN ointment rapidly became first-line treatment. Other non-surgical treatments for
anal fissure have subsequently been introduced, including local injection of the sphincter
with botulinum toxin (Botox™) (Gui et al.). The landmark paper discussed in this section
was one of the first studies to compare non-surgical treatments of chronic anal fissure.
Study references
Main study
Brisinda, G., Maria, G., Bentivoglio, A.R., Cassetta, E., Gui, D., and Albanese, A. (1999) A comparison
of injections of botulinum toxin and topical nitroglycerin ointment for the treatment of chronic anal
fissure. New England Journal of Medicine, 341, 65–9.
Related references
Lund, J.N. and Scholefield, J.H. (1997) A randomised, prospective, double-blind, placebo-controlled
trial of glyceryl trinitrate ointment in treatment of anal fissure. Lancet, 349, 11–14.
Gui, D., Cassetta, E., Anastasio, G., et al. (1994) Botulinum toxin for chronic anal fissure. Lancet, 344,
1127–8.
Kocher, H.M., Steward, M., Leather, A.J., and Cullen, P.T. (2002) Randomized clinical trial assessing
the side-effects of glyceryl trinitrate and diltiazem hydrochloride in the treatment of chronic anal
fissure. British Journal of Surgery, 89, 413–17.
Maria, G., Brisinda, G., Bentivoglio, A.R. et al. (1998) Botulinum toxin injections in the internal anal
sphincter for the treatment of chronic anal fissure: long-term results after two different dosage regi-
mens. Annals of Surgery, 228, 644–9.
Study design
◆ A prospective single-centre randomized blinded trial.
Randomization Botox injection versus 0.2% GTN ointment
Number of centres 1
Inclusion criteria Chronic (≥2 months) symptomatic posterior anal fissures
Follow-up 1 and 2 months in all patients
11 months (in healed patients)
NON-SURGICAL TREATMENT OF CHRONIC ANAL FISSURE 225
◆ 20 units of Botox was injected anteriorly into the internal sphincter; 0.2% GTN
ointment was applied topically twice daily for 6 weeks.
Outcome measures
Primary endpoint
◆ Complete healing of fissure.
Secondary endpoints
◆ Incidence of side effects, change in anal pressures (at baseline, and 1 and 2 months).
Results
Botox (n = 25) GTN (n = 25) p-value

Healed at 1 month 22 10
Healed at 2 months 24 15 0.005
Faecal incontinence 0 0
Side effects 0 5
Failure to heal 1 10
◆ Healed fissures were followed up ≥11 months with no relapses recorded.

◆ Side effects in the GTN group were all related to transient headaches.
Conclusions
Both Botox and GTN are effective treatments for chronic anal fissure, although better
results were seen with Botox.
Critique
This paper is one of the first randomized trials comparing non-surgical treatments of
chronic anal fissure. Although not powered, it makes a strong case for supporting both
Botox and GTN as effective treatments for anal fissures. The same authors had previously
compared different dosages of Botox (20 versus 15 units) and found the higher dose to be
more effective (Maria et al.).
This paper also noted that many patients suffered headaches as a side effect of GTN
ointment. Diltiazem is a calcium-channel blocker that can also be used as a topical treat-
ment for chronic anal fissure. It has been shown to have similar efficacy to GTN, but with
fewer headaches. In 2002, a prospective randomized trial confirmed the superiority of
diltiazem, leading many surgeons to adopt this drug as their first-line for chronic anal fis-
sures (Kocher et al.). To date, there is no RCT comparing diltiazem with Botox injection
in the treatment of chronic anal fissure.
6.7 Stapled haemorrhoidectomy

Details of studies
Haemorrhoids are a common condition accounting for a significant proportion of refer-
rals to colorectal clinics. Many interventions have been described, ranging from dietary
advice and rubber band ligation (for grade 1–2) to excisional haemorrhoidectomy (for
grade 3–4) (Shanmugam et al.). One of the most popular haemorrhoidectomy opera-
tions is the Milligan–Morgan technique, originally described in 1937. Although a suc-
cessful treatment, this procedure, like other forms of traditional haemorrhoidectomy,
can be associated with significant pain. This has led to the development of a number
of novel approaches to the management of haemorrhoids. Perhaps the most notable of
these is the operation called stapled haemorrhoidectomy or the procedure for prolapse
and haemorrhoids (PPH). Initially described in the late 1990s, the technique avoids the
need for wounds in the sensitive anal mucosa or skin and was reported as reducing post-
operative pain (Longo 1998). The trial discussed here (Mehigan et al. 2000) is one of
the first randomized trials assessing post-operative pain in patients undergoing stapled
haemorrhoidectomy versus the older excisional approach.
Study references
Main study
Mehigan, B.J., Monson, J.R., and Hartley, J.E. (2000) Stapling procedure for haemorrhoids versus
Milligan–Morgan haemorrhoidectomy: randomised controlled trial. Lancet, 355, 782–5.
Related references
Shanmugam, V., Thaha M.A., Rabindranath K.S., et al. (2005) Rubber band ligation versus excisional
haemorrhoidectomy for haemorrhoids. Cochrane Database of Systematic Reviews, CD005034.
Milligan, E.T.C. and Morgan, C.N. (1937) Surgical anatomy of the anal canal and the operative
treatment of haemorrhoids. Lancet, 2, 1119–24.
Longo, A. (1998) Treatment of haemorrhoidal disease by reduction of mucosa and haemorrhoidal
prolapsed with a circular stapling device: a new procedure. Proceedings of the 6th World Congress
of Endoscopic Surgery. Bologna: Monduzzi, 77784.
Study design
◆ A prospective randomized trial
Randomization Excisional versus stapled haemorrhoidectomy
Number of centres Single-centre (two participating surgeons)
Inclusion criteria Symptomatic prolapsing haemorrhoids
Exclusion criteria ASA grade 3 or 4
Patients on anticoagulants
Follow-up 6–10 weeks post-procedure
STAPLED HAEMORRHOIDECTOMY 227
◆ Excisional haemorrhoidectomy: Milligan–Morgan approach using diathermy.

◆ Stapled haemorrhoidectomy: circular stapling device applied at least 2cm above the
dentate line.
◆ Anaesthesia and post-operative analgesia was standardized between the two
approaches.
Outcome measures
Primary endpoints
◆ Post-operative pain (assessed by VAS).
◆ Presence or absence of faecal incontinence
Secondary endpoints
◆ Anaesthetic time, use of analgesia, time to first bowel motion, time to return to
normal activity.
Results
◆ Demographics were similar between the two groups
Stapled (n = 20) Excisional (n = 20) p-value

Anaesthesia time (min) 18 (9–25) 22 (15–25) 0.007
Average pain score 2.1 (0.2–7.6) 6.5 (3.1–8.5) 0.0001
Time to bowel motion (days) 1 (0–6) 2 (0–5) ns
Length of hospital stay (days) 1 (0–4) 1 (0–3) 0.376
Return to normal activity (days) 17 (3–60) 34 (14–90) 0.0002
All values documented as median (range)
◆ Pain scores were consistently lower on the first 10 days post-operatively for stapled
haemorrhoidectomy.
◆ Early complications: one stapled (urinary retention); two excisional (urinary
retention; secondary haemorrhage requiring transfusion and suture ligation of
pedicle).
◆ Late complications of severe pain: one stapled (fissure); two excisional (fissures and
delayed healing of excisional site).
◆ No long-term incontinence was reported in either group.
◆ Five patients in the stapled group complained of persistent skin tags, but only one
requested excision.
◆ At long-term follow-up, 85% of the stapled group rated their symptom control as
good or excellent compared with 75% for the excisional haemorrhoidectomy group;
5% in each group reported an unsatisfactory result.
Conclusions
Stapled haemorrhoidectomy is associated with a significant improvement in post-
operative pain control with an earlier return to work than the traditional approach of
excisional haemorrhoidectomy.
Critique
This study found improved short-term outcomes with stapled haemorrhoidectomy.
Other small-volume studies have subsequently reported similar findings (Ng et al.). They
note that external skin tags may persist after stapled haemorrhoidectomy, highlighting
that patients should be made aware of this. However, the main criticism of performing
stapled haemorrhoidectomy is that the surgeon performs the stapling almost blind. This
has resulted in documented complications such as rectal perforation, sepsis, and steno-
sis, leading working parties to state that stapled haemorrhoidectomy should only be per-
formed by appropriately trained individuals.
Although confident about short-term results, Mehigan et al. were cautious about com-
menting on long-term outcomes and highlighted that their study used pain as its pri-
mary endpoint with the study powered accordingly. They felt that recurrent symptoms or
impaired incontinence would require long-term follow-up before conclusions could be
drawn. Indeed, there is little long-term evidence in the published literature, and to date
there are no randomized prospective trials with more than 5 years follow-up.
Kam et al. performed a seven-year review of 7302 patients who had undergone stapled
haemorrhoidectomy. Of the 1834 patients who completed the follow-up question-
naire, 95% reported complete resolution or improvement of their symptoms, with only
26 reporting severe bleeding that required medical attention. Although no definitive con-
clusions can be drawn from this study, encouragingly it failed to identify any significant
longer-term problems with the procedure.
Ng, K.H., Ho, K.S., Ooi, B.S., Tang, C.L., and Eu, K.W. (2006) Experience of 3711 stapled
haemorrhoidectomy operations. British Journal of Surgery, 93, 226–30.
Kam, M.H., Ng, K.H., Lim, J.F., et al. (2011) Results of 7302 stapled haemorrhoidectomy operations
in a single centre: a seven-year review and follow-up questionnaire survey. ANZ Journal of Surgery,
81, 253–6.
LAPAROSCOPIC VENTRAL MESH RECTOPEXY FOR RECTAL PROLAPSE 229
6.8 Laparoscopic ventral mesh rectopexy for rectal prolapse

Details of studies
It is generally accepted that abdominal rectopexy has a lower recurrence rate than the
perineal procedures of Delorme’s mucosectomy (1990) or Altemeier’s perineal rectosig-
moidectomy (1965). After mobilization of the rectum, a rectopexy using sutures, mesh, or
a sponge is performed. Unfortunately, the main post-operative risk is constipation which
is estimated to occur in up to 50% of patients and may be a result of autonomic nerve
injury during mobilization (Brown et al.).
Laparoscopic ventral mesh rectopexy is based on the principle that the major portion
of an external rectal prolapse is the anterior rectal wall. This approach combines the ben-
efits of laparoscopic surgery (decreased post-operative pain, early recovery) with a poten-
tial reduction in nerve damage by avoiding posterolateral rectal mobilization. Originally
described by D’Hoore et al. in 2004, the paper analysed here is the same group’s long-term
follow-up paper with a larger group of patients.
Study references
Main study
D’Hoore, A. and Penninckx, F. (2006) Laparoscopic ventral recto(colpo)pexy for rectal prolapse:
surgical technique and outcome for 109 patients. Surgical Endoscopy, 20, 1919–23.
Related references
D’Hoore, A., Cadoni, R., and Penninckx, F. (2004) Longterm outcome of laparoscopic ventral rectopexy
for total rectal prolapse. British Journal of Surgery, 91, 1500–5.
Brown, A.J., Anderson, J.H., McKee, R.F., and Finlay, I.G. (2004) Strategy for selection of type of
operation for rectal prolapse based on clinical criteria. Diseases of the Colon and Rectum, 47, 103–7.
Delorme, E. (1990) Communication sur le traitement des prolapsus du rectum totaux par l’excision
de la muqueuse rectal ou recto-colique. Bulletin, Société de Chirurgie de Paris, 26, 498–518 (transl.
Diseases of the Colon and Rectum 28, 544–53, 1985).
Altemeier, W.A. and Culbertson, W.R. (1965) Technique for perineal repair of rectal prolapse. Surgery,
58, 758–64.
Study design
◆ A single-centre case series.
Level of evidence 4
Intervention Laparoscopic ventral mesh rectopexy
Number of patients 109 patients with full-thickness rectal prolapse
Follow-up Not reported
◆ 33 patients had prior pelvic surgery, of which 18 were for rectal prolapse.
◆ Operative technique: peritoneum opened over sacral promontory and extended
caudally over rectum; rectovaginal septum opened down to pelvic floor; mesh fixed
to anterior wall of distal rectum and vagina, and then attached to sacral promontory;
peritoneum closed to exclude the mesh for the peritoneal cavity.
Outcome measures
◆ Conversion rate and peri-operative outcomes.
Results
◆ 4/109 patients (3.7%) required conversion to open surgery.
◆ 8/109 patients (7%) experienced morbidity (five UTIs, one neuralgia at a port site,
one prolonged ileus, one pyrexia of unknown origin).
◆ There were no complications relating to the mesh.
◆ The median hospital stay was 5.1 days (range 2–10).
◆ Recurrent rectal prolapse occurred in four patients (3.66%). This was due to
detachment of the mesh from the sacral promontory.
◆ One patient developed an enterocele due to dehiscence of the colpopexy.
◆ One patient had incomplete reduction of the rectal prolapse at the time of surgery.
Conclusions
Laparoscopic ventral mesh rectopexy is an acceptable approach to full-thickness rectal
prolapse and may have fewer functional side effects than other techniques.
Critique
The initial paper published by D’Hoore et al. in 2004 introduced the technique of laparo-
scopic ventral rectopexy in 42 patients. After a median follow-up of 61 months, only two
recurrences were noted with functional improvements in incontinence (28/31 patients)
and obstructed defecation (16/19 patients). No patient developed severe constipation
during follow-up.
It would be expected that D’Hoore’s next paper would build on this and focus on longer-
term follow-up, in particular, recurrence rate and functional outcomes. With increased
patient numbers the second paper does confirm an acceptably low recurrence rate, but
there was no mention of the duration of follow-up in the paper. With the study duration
including some patients who underwent surgery in 2004 (2 years prior to publication), the
follow-up for some patients must be shorter than that documented in the original paper.
The main attraction of this new approach is that it may have a reduced risk of post-
operative functional upset. However, a randomized multi-centre trial comparing laparo-
scopic rectopexy with laparoscopic ventral rectopexy with long-term follow-up is needed
to answer the questions of long-term recurrence and functional outcome fully. Such a
trial is currently underway in Denmark where Professor Laurberg of Aarhus University
is the lead investigator. The trial is comparing laparoscopic posterior rectopexy with lapa-
roscopic anterior mesh rectopexy and is expected to finish recruiting in December 2012.
SACRAL NEUROMODULATION FOR FAECAL INCONTINENCE 231
6.9 Sacral neuromodulation for faecal incontinence

Details of studies
Faecal incontinence is a common presenting symptom at the colorectal clinic, affecting
up to 5% of the population. The management of faecal incontinence is often multifacto-
rial, ranging from conservative measures including dietary modification, medication, and
pelvic floor exercises to surgical intervention. Three Cochrane reviews separately ana-
lysed the evidence for surgery (levatorplasty, post-anal repair, total pelvic floor repair),
peri-anal bulking agents (PTQ®, carbon-coated beads), and electrical stimulation (with
anal biofeedback) in the treatment of faecal incontinence, and concluded that there was
only limited evidence to support any specific treatment option. A satisfactory review of
the evidence is hampered by the small numbers of patients enrolled in a small number
of trials. Of all of the techniques that have been described over the last 30 years, sacral
nerve stimulation (SNS) appears to offer the most promise in dealing with this difficult
and incapacitating problem.
Although SNS had been used in urinary voiding problems for some years, Matzel et al.
introduced the concept of SNS in the management of faecal incontinence in 1995 with the
publication of a small case series of three patients. (Matzel et al. 1995). They subsequently
published the results of the first multi-centre prospective trial assessing the efficacy of
SNS in faecal incontinence (Matzel et al. 2004).
Study references
Main study
Matzel, K.E., Kamm, M.A., Stosser, M., et al. (2004) Sacral spinal nerve stimulation for faecal
incontinence: multicentre study. Lancet, 363, 1270–6.
Related references
Deutekom, M. and Dobben, A.C. (2012) Plugs for containing faecal incontinence. Cochrane Database
of Systematic Reviews, CD005086.
Brown, S.R., Wadhawan, H., and Nelson, R.L. (2010) Surgery for faecal incontinence in adults. Cochrane
Database of Systematic Reviews, CD001757.
Hosker, G., Cody, J.D., and Norton, C.C. (2007) Electrical stimulation for faecal incontinence in adults.
Cochrane Database of Systematic Reviews, CD001310.
Matzel, K.E., Stadelmaier, U., Hohenfellner, M., and Gall, F.P. (1995) Electrical stimulation of sacral
spinal nerves for treatment of faecal incontinence. Lancet, 346, 1124–7.
Study design
◆ A prospective non-randomized multi-centre trial.
Number of centres 8
Intervention Initial trial of SNS for ≥10 days; patients with ≥50% reduction in incontinent
episodes went forward for permanent implantation of a neurostimulator
Number of patients 37 underwent a trial stimulation; 34 had permanent implantation

Inclusion criteria Faecal incontinence (≥1/week)
Intact external sphincter for ≥50% of its length (endo-anal ultrasound (EAUS))
Incontinence refractory to medical treatment/biofeedback
Aged 18–75 years
Exclusion criteria Congenital anorectal malformations
Previous rectal surgery
Previous or current rectal prolapse
Chronic diarrhoea
Neurological disease
Follow-up (median) 23.9 months
◆ Function was assessed by bowel habit diary, American Society of Colon and Rectal
Surgeons (ASCRS) QoL questionnaire for faecal incontinence and SF-36 at baseline
and 3, 6, 12, 24, and 36 months
Outcome measures
Primary endpoint
◆ Reduction in number of incontinence episodes each week and/or reduction in
number of days with incontinence. Patients served as their own control.
Secondary endpoints
◆ Functional outcomes
Results
◆ Eight of the 37 patients recruited had undergone previous sphincter repair. Patients
who had a previous sphincter repair had similar QoL results at baseline to those who
had not had a previous repair.
◆ Frequency of incontinence episodes decreased from a baseline mean of 16.4 per
week to 3.1 per week at 12 months (p <0.001), 2.0 at 24 months (p <0.001), and 1.8 at
36 months (p = 0.034).
◆ There was a reduction in the mean number of days/week with incontinent episodes
(4.5 versus 1.4 at 12 months and 1.2 at 24 months, p<0.0001).
◆ Staining and pad usage was significantly reduced following SNS.
◆ Substantial improvements in QoL scores on SF-36 and ASCRS questionnaires were
seen following SNS.
◆ Adverse events were reported in 12 patients. The majority were resolved.
◆ One patient had significant deterioration of their symptoms and required a stoma.
SACRAL NEUROMODULATION FOR FAECAL INCONTINENCE 233
Conclusions
Sacral neuromodulation results in immediate and sustained improvements in faecal
incontinence with associated improvement in quality of life.
Critique
This paper was the first prospective multi-centre trial assessing SNS in faecal inconti-
nence. Although patient selection was narrow and the study was limited by small patient
numbers, significant improvements in faecal incontinence were found. The study sug-
gested that SNS was an effective treatment and most patients appeared to tolerate the
procedure well. It was also apparent that the majority of adverse effects (primarily pain)
could be managed by adjustment or reprogramming of the device.
Since the publication of this landmark paper, many authors have gone on to confirm
the benefits of SNS in faecal incontinence, and a recent multi-centre prospective trial
involving 120 patients found that 83% of patients achieved therapeutic success (≥50%
reduction in incontinent episodes per week) after 12 months, leading the authors to con-
clude that patient selection should be expanded to include patients who do not have intact
anal sphincters (Wexner et al.). In addition, they have proposed a randomized trial com-
paring SNS with anal sphincter repair.
Although the early results of SNS are very encouraging, little is known about its long-
term efficacy or if there are side effects associated with long-term stimulation of the sacral
nerves. The same group that introduced the concept of SNS for faecal incontinence have
recently reported their long-term results (Matzel et al. 2009). They reviewed 12 patients
who had undergone implantation of a permanent neurostimulator between 1994 and
1999. The device had been removed in three patients because of pain or neurological
disease. Mean follow-up in the remaining nine patients was 9.8 years (range 7–14). There
was no deterioration in efficacy during this period, although the pulse generator had to be
exchanged because of battery fatigue in eight of the nine patients.
The results of SNS have stimulated interest in percutaneous posterior tibial nerve stim-
ulation as a treatment for faecal incontinence. This new, non-invasive outpatient treat-
ment for faecal incontinence may be more cost-effective, with promising early results
from a small case series (Govaert et al.).
Wexner, S.D., Coller, J.A., Devroede, G., et al. (2010) Sacral nerve stimulation for fecal incontinence.
Results of a 120-patient prospective multicenter study. Annals of Surgery, 251, 441–9.
Govaert, B., Parea, D., Delgado-Aros, S., et al. (2010) A prospective multicentre study to investigate
percutaneous tibial nerve stimulation for the treatment of faecal incontinence. Colorectal Disease,
12, 1236–41.
Matzel, K.E., Lux. P., Heuer, S., Besendörfer, M., and Zhang, W. (2009) Sacral nerve stimulation for
faecal incontinence: long-term outcome. Colorectal Disease, 11, 636–41.
6.10 Surgery for pilonidal sinus disease

Details of studies
Although drainage with curettage of hair and debris is viewed as standard practice in the
management of an acute pilonidal abscess, the best method of managing chronic piloni-
dal sinus disease is less certain. Multiple procedures have been described, which is per-
haps less a testament to the ingenuity of surgeons than a reflection on the fact that no
single operation has met with universal success. Although many case series have been
published, only a handful of comparative trials have been carried out. The paper discussed
here is a Cochrane review comparing open procedures (healing with secondary intention)
versus primary closure for pilonidal sinus and midline closure versus off-midline closure.
Study references
Main study
Al-Khamis A, McCallum, I., King, P.M., and Bruce, J. (2010) Healing by primary versus secondary
intention after surgical treatment for pilonidal sinus. Cochrane Database of Systematic Reviews,
CD006213.
Related references
Bascom, J.U. (1982) Pilonidal disease: correcting overtreatment and undertreatment. Contemporary
Surgery, 18, 13–18.
Bascom, J. and Bascom, T. (2002) Failed pilonidal surgery: new paradigm and new operation leading to
cures. Archives of Surgery, 137, 1146–50.
Karydakis, G.E. (1973) New approach to the problem of pilonidal sinus. Lancet, 2, 1414–15.
Daphan, C., Tekelioglu, M.H., and Sayilgan, C. (2004) Limberg flap repair for pilonidal sinus disease.
Diseases of the Colon and Rectum, 47, 233–7.
McDermott, F.T. (1967) Pilonidal sinus treated by Z-plasty. ANZ Journal of Surgery, 37, 64–9.
Yilmaz, S., Kirimlioglu, V., and Katz, D. (2000) Role of simple V–Y advancement flap in the treatment
of complicated pilonidal sinus. European Journal of Surgery, 166, 269.
Study design
Intervention Open-wound healing versus surgical closure
Midline versus off-midline closure
Total number of trials 26
Number of centres 56 centres in four countries
Inclusion criteria Chronic pilonidal sinus disease
◆ Databases searched: Medline, EMBASE, EBSCO, and CINAHL. Bias assessment was
completed for each study included.
SURGERY FOR PILONIDAL SINUS DISEASE 235
Outcome measures
Primary endpoint
◆ Healing times, surgical site infection, and recurrence.
Secondary endpoints
◆ Time to return to work, length of hospital stay, QoL, cost, morbidity.
Results
◆ Seventeen studies compared open-wound healing with surgical closure and six
studies compared midline with off-midline closure.
◆ Nine of 13 studies reported faster healing times with primary closure.
◆ No significant difference in surgical site infections was seen between the open-wound
and surgical closure groups (RR 1.31, 95% CI 0.93–1.85).
◆ Recurrence of pilonidal sinus occurred in 44/828 patients (5.3% ) after open healing
compared with 73/838 patients (8.7%) after closed (all) techniques.
◆ Midline closure was found to take significantly longer to heal than off-midline
closure (MD 5.4 days, 95% CI 2.3-8.5), had an increased risk of infection (RR 3.72,
95% CI 1.86–7.42), and a higher chance of recurrence (OR 4.54, 95% CI 2.30–8.96).
◆ No conclusions were reached with regard to cost, pain, and QoL because of study
bias and data heterogeneity.
Conclusions
This review showed a clear benefit in favour of off-midline closure. Overall, open-wound
healing results in a lower risk of recurrence than primary closure, but that must be bal-
anced against a longer healing time.
Critique
This Cochrane review is limited by the quality of data from the included studies. Th is is a
reflection of the many differing surgical approaches to pilonidal sinus treatment, result-
ing in the majority of the published literature commenting on case series or comparative
studies between only two approaches. The authors acknowledge potential bias in their
review, although their methodology does try to minimize it, allowing some conclusions
to be drawn.
Many surgeons now support off-midline closure, believing that removal of the cleft
minimizes recurrence and improves wound healing. As a result, most of the aforemen-
tioned techniques are based on off-midline closure. This review notes the interesting
result that open surgery with healing by secondary intention had the lowest risk of recur-
rence. This approach is perhaps seen as the more traditional approach and is likely not to
be considered first line because of the prolonged healing time and time off work, particu-
larly in the younger population.
6.11 Bowel cancer screening

Details of studies
In 1968 Wilson and Jungner described the principles of a screening programme that
forms the basis of many of the current programmes in the UK. With respect to colorectal
cancer, the main barriers have been the acceptability, sensitivity, and specificity of the
screening test.
Although the faecal occult blood test (FOBt) is simple, non-invasive, and well toler-
ated, it has a high proportion of false positives, leading to colonoscopy in patients with no
significant pathology. The first paper in this section (Towler et al.) is a systematic review
analysing the evidence for FOBt. The second study (Atkin et al.) assesses the screening
value of a one-off flexible sigmoidoscopy in patients aged 55–65 years, potentially remov-
ing the need for colonoscopy.
Study references
Main studies
Towler, B., Irwig, L., Glasziou, P., Kewenter, J., Weller, D., and Silagy, C. (1998) A systematic review of
the effects of screening for colorectal cancer using the faecal occult blood test, hemocult. British
Atkin, W.S., Edwards, R., Krajl-Hans, I., et al. (2010) Once-only flexible sigmoidoscopy in prevention of
colorectal cancer: a multicentre randomised controlled trial. Lancet, 375, 1624–33.
Related references
Wilson, J.M.G. and Jungner, G. (1968) Principles and practice of screening for disease. WHO Chronicle,
22, 473.
Bowel Cancer Screening programme England <http://www.cancerscreening.nhs.uk/bowel/>
Bowel Cancer Screening programme Scotland <http://www.bowelscreening.scot.nhs.uk/>
Tower et al.
Study design
◆ Systematic review with meta-analysis.
Intervention Bowel cancer screening using the faecal occult blood test
Studies included Four RCTs and two non-randomized trials
Number of patients 443,000
Patient population ≥40 years of age from five different countries
◆ Authors were contacted for unpublished data and/or clarification of results.

◆ The design and quality of each study were analysed and documented.
◆ A comprehensive search was carried out of the health literature for colorectal cancer
screening with FOBts.
BOWEL CANCER SCREENING 237
Outcome measures
Primary endpoint
◆ Effectiveness of screening with FOBt on mortality from colorectal cancer.
Secondary endpoints
◆ Benefits and harms of colorectal cancer screening.
Results
◆ Hemoccult® testing was used in all studies (some hydrated, some rehydrated).
◆ A positive test was reported in 2–6% of the screened populations. These patients
underwent either a colonoscopy or flexible sigmoidoscopy + barium enema.
◆ The sensitivity of the Hemoccult® test for colorectal cancer varied from 46% to 92%.
◆ Attendance for screening was higher in US than in European studies (90%
versus 60%).
◆ Meta-analysis found an overall reduction of 16% in the risk of death from colorectal
cancer with screening (RR 0.84, 95% CI 0.77–0.93).
◆ The number needed to screen to prevent one death from colorectal cancer over
10 years is 1173.
◆ Complications of bleeding and perforation from colonoscopy and sigmoidoscopy
varied from 12 to 30 per 10,000 procedures (sigmoidoscopy had a lower risk).
Conclusions
The estimate of the reduction in mortality associated with screening for colorectal cancer
with Hemoccult® is now well quantified.
Atkin et al.
Study design
◆ A multi-centre RCT.
Randomization Flexible sigmoidoscopy screening versus no intervention
Patient population 55–64 years of age
Number of centres 14
Follow-up Median 11.2 years
◆ Two previous pilot studies refined the study protocol

Inclusion criteria
◆ Men and women in the age range registered with participating general practices.
◆ Patients were vetted by their GP who excluded ineligible patients.

◆ Suitable patients were sent a questionnaire asking if they would be interested in

participating in screening.
Exclusion criteria
◆ Inability to provide informed consent.
◆ History of colorectal cancer, adenoma, or inflammatory bowel disease.

◆ Severe or terminal disease or life expectancy <5 years.
◆ Sigmoidoscopy/colonoscopy in the previous 5 years.
◆ People with a strong family history of colorectal cancer.
◆ Symptomatic patients.
Outcome measures
Primary endpoint
◆ Colorectal cancer incidence and mortality.
Secondary endpoints
◆ Incidence of distal and proximal cancer.
◆ All-cause mortality and mortality due to non-colorectal cancer causes
Results
◆ The mean age was 60 years (SD 2.9) in both groups.
◆ 71% of the intervention group attended for flexible sigmoidoscopy.
◆ 95% were discharged after screening; the remaining 5% were referred for colonoscopy.
Control (n = 112,939) Screening (n = 57,099) HR (p-value)

Total number of cancers 1818 706 0.77 (<0.0001)
Distal cancers 1192 386 0.64 (<0.0001)
Proximal cancers 628 311 0.98 (0.75)
All-cause mortality 13,768 6775 0.97 (0.0519)
Cancer-specific mortality 637 221 0.68 (<0.0001)
Non-cancer mortality 13131 6554 0.99 (0.33)
◆ The number needed to screen to prevent one cancer diagnosis is 191.

◆ The number needed to screen to prevent one colorectal cancer death is 489.
Conclusions
Flexible sigmoidoscopy is a safe and practical test which confers a substantial and long-
lasting protection from colorectal cancer.
Critique
These two studies provide high-level evidence that screening reduces colorectal cancer
mortality, with FOBt and sigmoidoscopy showing a 16% and 31% reduction in mortality,
respectively. In response to the growing evidence for screening in colorectal cancer, a pilot
BOWEL CANCER SCREENING 239
UK bowel cancer screening programme was set up in 2000 (West Midlands, and Tayside,
Grampian, and Fife). In this pilot 478,250 residents were invited to submit an FOBt. Those
with a positive test were invited to undergo colonoscopy. Of those invited to participate,
56.8% submitted an FOBt and 1.9% had a positive result. A carcinoma was detected in
1.62/1000 people who underwent screening. The positive predictive value was 10.9% for
cancer and 35% for adenomas.
Following the pilot, a bowel cancer screening programme was rolled out across the
UK in 2006. In Scotland, people aged 50–74 years are sent the FOBt kit every 2 years.
In England, the screened population is aged 60–69 years. Concern over the sensitivity
(11–37%) of the current guaiac FOBt (gFOBt) has led to further developments of the
technology. The newer faecal immunochemical test (FIT) has greater sensitivity and spec-
ificity (Smith et al.), and positivity rates can be adjusted by quantifying faecal haemoglo-
bin concentration. FIT is now used in the Scottish screening programme for patients with
equivocal results on gFOBt.
The main criticism of flexible sigmoidoscopy as a screening test is that it only assesses
the distal colon. Atkin et al. address this in their discussion by quoting publications show-
ing that proximal colon cancers can be predicted from the characteristics of adenomas
in the rectum and sigmoid colon (Levin et al.). As a result, patients with high-risk pol-
yps in this study underwent colonoscopy, while those with low-risk polyps did not. This
tight protocol led to fewer patients being referred for colonoscopy than in other studies
(Weissfeld et al.).
Both gFOBt and FIT were compared with flexible sigmoidoscopy in a recent random-
ized trial of 15,000 patients (Hol et al.). The participation rate was highest for FIT (61.5%),
followed by gFOBt (49.5%), and finally flexible sigmoidoscopy (32.4%). When the three
tests were compared for their detection rate of neoplasia, flexible sigmoidoscopy per-
formed best (2.4/100 invitees compared with 1.5 for FIT and 0.6 for gFOBt). The authors
concluded that FIT was superior to gFOBt, and that long-term follow-up was required to
compare mortality with FIT versus flexible sigmoidoscopy.
Flexible sigmoidoscopy is now being offered as part of the bowel screening programme
in England. Patients will be invited for flexible sigmoidoscopy at age 55, followed by FOBt
from the age of 60.
Hol, L., van Leerdam, M.E., van Ballegooijen, M., et al. (2010) Screening for colorectal cancer:
randomised trial comparing guaiac-based and immunochemical faecal occult blood testing and
flexible sigmoidoscopy. Gut, 59, 62–8.
Smith, A., Young, G.P., Cole, S.R., et al. (2006) Comparison of a brush-sampling fecal immunochemical
test for haemoglobin with a sensitive guaiac-based fecal occult blood test in detection of colorectal
neoplasia. Cancer, 107, 2152–9.
Levin, T.R., Palitz, A., Grossman, S., et al. (1999) Predicting advanced proximal colonic neoplasia with
screening colonoscopy. Journal of the American Medical Association, 281, 1611–17.
Weissfeld, J., Schoen, R., Pinsky, P., et al. (2005) Flexible sigmoidoscopy on the PLCO cancer screening
trial: results from the baseline screening examination of a randomised controlled trial. Journal of the
National Cancer Institute, 97, 989–97.
6.12 Total mesorectal excision for rectal cancer

Details of studies
The introduction of circular surgical staplers into surgical practice in the late 1970s
allowed surgeons to fashion very low anastomoses. This led to an increase in the number
of low anterior resections performed for low rectal cancers, with a concomitant reduction
in the rate of abdomino-perineal resection.
In tandem with this change, Heald focused attention on the importance of performing
a total mesorectal resection (TME). The concept behind the operation is that rectal can-
cer spreads within the mesorectum rather than within the rectal muscle itself. Therefore
Heald felt that it was critical to remove an intact mesorectal envelope along with the rec-
tum in middle- and lower-third cancers in order to reduce the risk of developing recur-
rent disease. It was also felt that as long as the mesorectum had been completely removed,
it was not necessary to have an additional 5cm clearance below the lower limit of the
tumour as distal spread within the bare muscle tube where the mesorectum peters out is
usually less than 2cm. Therefore, in order to validate the TME operation, long-term data
on recurrence and survival were required.
The paper discussed here provided validated long-term follow-up for the original paper
that introduced the concept of TME for middle and low rectal cancer, and this procedure
quickly became the standard of care.
Study references
Main study
Heald, R.J. and Ryall, R.D.H. (1986) Recurrence and survival after total mesorectal excision for rectal
cancer. Lancet, i, 1479–82.
Related references
Heald, R.J., Husband, E.M., and Ryall, R.D. (1982) The mesorectum in rectal cancer surgery—the clue to
pelvic recurrence? British Journal of Surgery, 69, 613–16.
Heald, R.J. (1980) Towards fewer colostomies—the impact of circular stapling devices on the surgery of
rectal cancer in a district hospital. British Journal of Surgery, 60, 198–200.
Heald, R.J. and Leicester, R.J. (1981) The low stapled anastomosis. British Journal of Surgery,
68, 333–7.
Heald, R.J., Moran, B.J., Ryall, R.D., Sexton, R., and MacFarlane, J.K. (1998) Rectal cancer:
the Basingstoke experience of total mesorectal excision, 1978–1997. Archives of Surgery,
133, 894–9.
Quirke, P., Durdey, P., Dixon, M.F., and Williams, N.S. (1986) Local recurrence of rectal
adenocarcinoma due to inadequate surgical resection. Histopathological study of lateral tumour
spread and surgical excision. Lancet, ii, 996–9.
MacFarlane, J.K., Ryall, R.D.H., and Heald, R.J. (1993) Mesorectal excision for rectal cancer. Lancet,
341, 457–60.
Adam, I.J., Mohamdee, M.O., Martin, I.G., et al. (1994) Role of circumferential margin involvement in
the local recurrence of rectal cancer. Lancet, 334, 707–11.
TOTAL MESORECTAL EXCISION FOR RECTAL CANCER 241
Study design
◆ A 7-year prospective case series.
Level of evidence 4
Intervention Anterior resection with TME (curative)
Number of centres A single surgeon in an English district general hospital
Number of patients 115 (consecutive)
Inclusion criteria Patients undergoing ‘curative’ anterior resection for rectal cancer defined
as adenocarcinoma within 15cm of the anal verge
Exclusion criteria Recto-sigmoid cancers, polyp cancers, FAP or colitis
Follow-up Every 3–4 months for the first 2 years
Every 6 months for a further 3 years
Annually thereafter
Clinical review by two consultants + CEA levels
FAP, familial adenomatous polyposis; CEA, carcino-embryonic antigen.
◆ Of 188 patients referred to the firm with rectal cancer, 30 did not have an anterior
resection (abdomino-perineal excision of the rectum (APER) n = 21) and 35
underwent a palliative procedure for either metastatic (19) or residual local (16)
disease.
◆ Patients undergoing curative resection had high ligation of the inferior mesenteric
vessels and total mesorectal excision, with sharp dissection in the avascular plane.
◆ No patient included in the curative group of the study received adjuvant radiotherapy
or chemotherapy.
Outcome measures
◆ Local recurrence.
◆ Tumour-free survival.
Results
◆ 30-day mortality was 2.6% (3/115).
◆ 69 patients had a distal resection margin of <5cm and 39 had a resection margin of
<2.5cm.
◆ Local recurrence at an average of 4.2 years follow-up occurred in three cases (2.6%)
(3.7% cumulative risk of local recurrence at 5 years).
◆ None of these three recurrences was attributable to a reduced distal margin.
◆ In the palliative group 4/35 patients died within 30 days (11%); local recurrence or
residual disease was present in 12/35 of patients (34%).
◆ Survival at 5 years was 87.5% and overall disease-free survival was 81.7% (94% for
Dukes’ stage A, 87% for Dukes’ stage B, and 58% for Dukes’ stage C).
◆ There was no significant difference in survival related to the tumour height (<6cm,
7–9cm, 10–15cm) or the length of distal resection margin (<1.5cm, 1.5–3cm, >3cm).
Conclusions
Very low rates of local recurrence are achievable using the total mesorectal excision tech-
nique for middle to low rectal cancer. This improvement can be achieved with a reduced
distal resection margin, avoiding abdomino-perineal resection and permanent colostomy
in many patients with lower-third tumours (APER accounted for 11% of all resections in
this study).
Critique
This study is a good example of how a case series can make a significant impact on surgi-
cal practice. Prior to the publication of Heald’s work, local recurrence rates after rectal
cancer resection were high and the reasons for this remained unclear. This series showed
that using the technique of total mesorectal excision, very low rates of local recurrence
were achievable. In an updated report published in 1993 local recurrence remained
unchanged at 4% after 10 years of follow-up, suggesting that the results of the technique
are durable.
This case series alone is insufficient evidence to prove that the low rate of local recur-
rence is directly attributable to TME when other confounding factors may have played a
role. However, since its publication supporting evidence has come both from pathological
studies and through the application of the technique in other surgical centres.
Quirke reported a study of 52 patients undergoing rectal cancer surgery, followed by
a more detailed histopathological examination of the resection specimen involving serial
transverse slices taken through the whole tumour and mesorectum. He was able to show
a strong correlation between lateral spread of tumour (circumferential margin) and local
recurrence. Of the 14 patients (27%) in the study who had an involved circumferential
margin, 12 developed local recurrence. Involvement of the circumferential margin was
also found to be an independent determinant of overall survival. These findings were
entirely in keeping with Heald’s hypothesis that local recurrence was due to incomplete
surgical resection (see 6.13).
Further validation of the results of TME come from studies carried out following its
implementation through national training initiatives in Norway, Sweden, and Holland
which found similar reductions in local recurrence and improvement in survival. These
studies also add weight to the argument for specialization in rectal cancer surgery.
The TME technique, attributed to Professor Heald, is now regarded as the gold stan-
dard in rectal cancer surgery. The focus on the circumferential margin and the acceptance
of closer distal margins has meant that more patients are considered for sphincter-pre-
serving surgery. Consequently, the rate of APER and permanent colostomy has reduced.
TOTAL MESORECTAL EXCISION FOR RECTAL CANCER 243
Understanding of the relative importance of the circumferential resection margin has

revolutionized both the staging of rectal cancer and the selection of patients for neoadju-
vant therapy.
More recently, Hohenberger et al. have applied the concept of ensuring an intact pack-
age of the tumour and its main lymphatic drainage to colonic resection. They describe
a complete mesocolic excision (CME) for colonic cancer, with the aim of resecting the
colon and mesocolic tissue from the parietal plane. In addition they describe a ‘true cen-
tral ligation of the supplying arteries and draining veins right at their roots’. In a prospec-
tive assessment of their results they noted a reduction in 5-year local recurrence rates in
colon cancer from 6.5% to 3.6% after adopting this technique for colon resection. They
also noted a concomitant improvement in 5-year survival from 82.1% to 89.1%.
Wibe, A., Møller, B., Norstein, J., et al. (2002) A national strategic change in treatment policy for rectal
cancer—implementation of total mesorectal excision as routine treatment in Norway. A national
audit. Diseases of the Colon and Rectum, 45, 857–66.
Martling, A.L., Holm, T., Rutqvist, L.E., et al. (2005) Impact of a surgical training programme on rectal
cancer outcomes in Stockholm. British Journal of Surgery, 92, 225–9.
Kapiteijn, E., Putter, H., and van de Velde, C.J. (2002) Impact of the introduction and training of total
mesorectal excision on recurrence and survival in rectal cancer in The Netherlands. British Journal
of Surgery, 89, 1142–9.
Nagtegaal, I.D., van de Velde, C.J., van der Worp, E., et al. (2002) Macroscopic evaluation of rectal
cancer resection specimen: clinical significance of the pathologist in quality control. Journal of
Clinical Oncology, 20, 1729–34.
Hohenberger, W., Weber, K., Matzel, K., Papadopoulos, T., and Merkel, S. (2009) Standardized surgery
for colonic cancer: complete mesocolic excision and central ligation—technical notes and outcome.
Colorectal Disease, 11, 354–65.
6.13 The circumferential resection margin in rectal cancer

Details of studies
In TME surgery, the plane of dissection is formed by the mesorectal fascia covering the
rectum. It is this fascia that forms the circumferential resection margin (CRM), also
known as the lateral resection margin. The first paper discussed here by Quirke et al.
focused attention on the quality of rectal cancer surgery in terms of maintaining an intact
mesorectal fascial envelope. The paper demonstrated a clear relationship between the
presence of tumour at the CRM and the risk of developing recurrent disease.
The quality of MRI advanced dramatically in the years following the publication of
Quirke’s paper, so that by the late 1990s radiologists became increasingly confident that
they could accurately stage local spread of rectal cancer and its relationship to the meso-
rectal fascia. This implied that it might be possible to predict which tumours were likely
to result in a positive CRM, assuming that the surgery was in the mesorectal fascial
plane and the rectal cancer was not downstaged with pre-operative radiotherapy. The
MERCURY Study Group analysed the capability of MRI to predict involvement of the
CRM in patients who subsequently underwent total mesorectal resection.
Study references
Main studies
Quirke, P., Dixon, M.F., Durdey, P., and Williams, N.S. (1986) Local recurrence of rectal
adenocarcinoma due to inadequate surgical resection. Histopathological study of lateral tumour
spread and surgical excision. Lancet, 2, 996–9.
MERCURY Study Group. (2006) Diagnostic accuracy of preoperative magnetic imaging in predicting
curative resection of rectal cancer: prospective observational study. British Medical Journal, 333, 779.
Related reference
Birbeck, K.F., Macklin, C.P., Tiffin, N.J., et al. (2002) Rates of circumferential resection margin
involvement vary between surgeons and predict outcomes in rectal cancer surgery. Annals of
Surgery, 235, 449–57.
Quirke et al.
Study design
◆ A case series.
Patients Resected rectal cancer
Analysis Pathological assessment of the CRM
Outcome Local recurrence
◆ Patients were matched retrospectively in stage and grade to a control group for
comparison.
CIRCUMFERENTIAL RESECTION MARGIN IN RECTAL CANCER 245
Outcome measures
◆ The relationship between a positive CRM and local recurrence.
Results
◆ The CRM was positive in 14/52 specimens.
◆ 12/14 went on to develop local recurrence.
◆ The specificity of CRM positivity was 92%, sensitivity was 95%, and positive
predictive value was 85%.
Conclusions
Local recurrence is mainly due to lateral spread in rectal adenocarcinoma.
MERCURY study
Study design
◆ A prospective observational study.
Patients Rectal adenocarcinoma undergoing surgical resection
Number of patients 408 (originally 679)
Number of centres 11 (across four European countries)
Intervention Pre-operative MRI assessing CRM involvement
Outcome Comparison with pathological CRM post-operatively
◆ Inclusion criteria: >18 years of age

◆ Exclusion criteria: pregnant, previous pelvic malignancy/radiotherapy, pelvic floor
surgery, unable to undergo MRI because of implanted metal
◆ High resolution MRI was used pre-operatively (also after neoadjuvant therapy).
MRIs were analysed by radiologists who were blinded to the pathological CRM
outcome. The CRM was deemed positive on MRI if the tumour was ≤1mm from the
mesorectal fascia.
◆ The CRM was reported as negative on pathological analysis if there was ≥1mm
between the tumour and the CRM
Outcome measures
◆ Accuracy of MRI in predicting CRM involvement compared to post-operative
pathological analysis.
Results
◆ Of the 408 patients, 311 underwent primary surgery/short-course radiotherapy (RT)
and 97 underwent long-course radiotherapy/chemoradiotherapy (RT/CRT).
Surgery/short-course RT (n = 311) Long-course RT/CRT (n = 97)

Accuracy 91% 77%
Sensitivity 42% 94%
Positive predictive value 71% 45%
Specificity 98% 73%
Negative predictive value 93% 98%
Conclusions
High-resolution MRI accurately predicts surgical resection margin involvement. The
technique was reproduced accurately in multiple centres, allowing selection of patients
for neoadjuvant treatment.
Critique
These two papers have had an important influence on the management of rectal cancer.
The ability to predict pre-operative involvement of the CRM facilitates informed mul-
tidisciplinary decision-making on the need for neoadjuvant therapy prior to surgical
resection. Resection then focuses on achieving a negative CRM by performing a total
mesorectal excision.
One area of recent debate has been the definition of a positive CRM, with 1mm being
challenged and greater margins proposed. A recent paper by Taylor et al. (2011) compared
different CRMs on pre-operative MRI to determine the best predictor of local recurrence
in patients undergoing surgery (≤1mm, >1mm–2mm, >2mm–5mm, and >5mm). Only
the margin of ≤1mm on MRI remained significant for local recurrence regardless of pre-
operative treatment, leaving the cut-off for CRM to remain at ≤1mm. More recently,
attention has also focused on the predictive value of an involved lymph node accounting
for a positive margin compared with direct tumour infiltration. It seems likely that direct
tumour infiltration is a much more relevant predictor.
Taylor, F.G., Quirke, P., Heald, R.J. et al. and the MERCURY study group (2011) One millimetre is the
safe cut-off for magnetic resonance imaging prediction of surgical margin status in rectal cancer.
EXTRALEVATOR ABDOMINO-PERINEAL EXCISION OF THE RECTUM 247
6.14 Extralevator abdomino-perineal excision of the rectum

Details of studies
Local recurrence is greater in patients with low rectal cancer who undergo abdomino-
perineal excision of the rectum (APER) than in patients undergoing anterior resection
(with TME). Most case series have documented a much higher positive circumferential
margin (CRM) rate (12–30%) and an increased frequency of bowel perforation (13.7–
15.4%) in APER, both of which are risk factors for local recurrence (Stelzner et al.). Many
of these perforations and positive CRMs occur in the upper anal canal/distal rectum
where the mesorectum peters out. If the surgeon follows this plane down to the upper
anal canal, ‘waisting’ may be seen in the specimen with traditional APER resection tech-
niques. This can give rise to a positive CRM in T3 and even T2 tumours.
In an attempt to avoid this, Holm and others have advocated a more radical APER
which involves en bloc removal of the levator ani muscles. This extended extra-levator
abdomino-perineal excision (ELAPE) aims to deliver a cylindrical specimen. Typically,
the perineal part of the dissection is performed in the prone position. Although this has
been described as a new technique, it is not dissimilar to the original description of APER
by Sir Ernest Miles in 1908.
Study references
Main study
Holm, T., Ljung, A., Häggmark, T., Jurell, G., and Lagergren, J. (2007) Extended abdominoperineal
resection with gluteus maximus flap reconstruction of the pelvic floor for rectal cancer. British
Related reference
Stelzner, S., Holm, T., Moran, B., et al. (2011) Deep Pelvic Anatomy Revisited for a Description of
crucial Steps in extralevator abdominoperineal excision for rectal cancer. Diseases of the Colon and
Rectum, 54, 947–57.
Study design
◆ A case series.
Level of evidence 4
Intervention Extended APER for low rectal adenocarcinoma
Number of centres 1
Follow-up (median) 16 months (range 1–45)
◆ Inclusion criteria: rectal adenocarcinoma within 6cm of the anal verge; T3 or T4

tumour on MRI; tumour fixed or tethered at rectal examination.
◆ Neoadjuvant treatment was given to all patients (six patients had intra-operative
radiotherapy).
◆ The perineal defect was closed with unilateral or bilateral gluteus maximus flaps.
Outcome measures
◆ Reporting of technique and complications.
◆ Local recurrence.
Results
◆ The median age was 66 years (range 49–86 years), with 19 men and 9 women.
◆ Pathological staging of patients: two ypT0, twenty ypT3, and six ypT4.
◆ The CRM was positive in two patients (7.1%).
◆ Bowel perforation occurred in 1 patient (3.6%).
◆ Local perineal wound complications were encountered in four patients (three wound
infections, one bleed from a flap requiring re-operation).
◆ Local recurrence occurred in two patients (both pT4).
◆ Eight patients died during follow-up: four had no recurrent disease, three had distant
metastases, and one had local and distant metastases.
Conclusions
The data show that the rate of perforation and CRM involvement can be kept low with
properly performed APER. In addition, the rate of perineal wound complications was low
after gluteus maximus musculocutaneous flap reconstruction of the pelvic floor.
Critique
This paper has drawn attention to the importance of gaining a clear circumferential
margin by performing a wide excision. Holm et al. focus attention on the importance of
taking the levators in order to avoid ‘coning in’ as the distal mesorectum thins out onto
rectal muscle wall. By taking more surrounding tissue that lies close to the tumour (and
subsequently dissecting further away from the tumour) there is less likelihood that the
bowel will be perforated and that the CRM will be positive, reducing local recurrence.
This case series reported a reduction in the positive CRM rate for APERs and included
only a single instance of bowel perforation. The follow-up in this small series is short
(median 16 months), and it is yet to be seen if this will translate into a long-term reduc-
tion in local recurrence, which is the ultimate goal.
Since the publication of this paper, a multi-centre study has been published (eleven
surgeons in nine institutions) where all 176 cases were retrospectively reviewed to con-
firm that extra-levator surgery had taken place (West et al.). Results were then compared
with a series of standard APERs from one centre (n = 124). CRM positivity and intra-
operative bowel perforation were significantly reduced in the ELAPE group: 20% versus
EXTRALEVATOR ABDOMINO-PERINEAL EXCISION OF THE RECTUM 249
49% for CRM positivity; 8% versus 28% for perforation; p <0.001 in both analyses).
Indeed, the bowel perforation rate was significantly lower when the perineal dissection
was performed in the prone jack-knife position compared with lithotomy. Functionally,
there were no significant differences between the groups, but the ELAPE group did expe-
rience increased wound complications, irrespective of whether a reconstructive flap had
been used, whether neo-adjuvant therapy had been administered, or whether surgery was
laparoscopic or open. Some argue that for this reason ELAPE should be used selectively
for more advanced tumours to avoid increasing wound morbidity.
The authors concluded that, although local recurrence and survival rates are not cur-
rently available, as they are closely related to CRM status and intra-operative bowel per-
foration, it is expected that these outcomes will improve by an order of magnitude similar
to that seen following TME by anterior resection.
In a recent non-randomized trial from the Cleveland Clinic, de Campos-Lobato et al.
compared outcomes of APER with the perineal part of the operation in either the prone or
lithotomy position. They reported a bowel perforation rate of 2.4% for lithotomy and 4.6%
for the prone position. The positive CRM rate was 14.6% and 9%, and local recurrence at
5 years was 5.7% and 14.4%, with no significant difference between the two groups. These
impressive results set a new standard for outcomes in this difficult patient group and also
suggest that the prone position is not a prerequisite for a good outcome.
Miles, W.E. (1908) A method of performing abdominoperineal excision for carcinoma of the rectum
and the terminal portion of the pelvic colon. Lancet, 2, 1812–13.
West, N.P., Anderin, C., Smith, K.J., et al. (2010) Multicentre experience with extralevator
abdominoperineal excision for low rectal cancer. British Journal of Surgery, 97, 588–99.
de Campos-Lobato, L.F., Stocchi, L., Dietz, D.W., Lavery, I.C., Fazio, V.W., and Kalady, M.F. (2011)
Prone or lithotomy positioning during an abdominoperineal resection for rectal cancer results in
comparable oncologic outcomes. Diseases of the Colon and Rectum, 54, 939–46.
6.15 Neoadjuvant therapy in primary rectal cancer

Details of studies
Decisions regarding neoadjuvant therapy in primary rectal cancer are usually made after
multidisciplinary review of the clinical, pathology, and radiology findings. An MRI of
the rectum is critical to assess the proximity of the tumour to the circumferential resec-
tion margin (CRM). Although total mesorectal excision (TME) is widely accepted as the
surgical standard in the resection of rectal cancer, the optimal neoadjuvant therapy for
tumours where the CRM is either threatened or clear is debatable.
Approaches include proceeding directly to surgery, short course pre-operative radio-
therapy (SCPRT) for one week followed by TME surgery one week later, SCPRT for
one week followed by TME surgery 6–12 weeks later; and long-course chemoradio-
therapy (CRT) for 5–6 weeks with TME surgery 6–12 weeks after completion of the last
cycle. Numerous randomized trials have assessed these approaches. In this section we
address each of these options, discussing the relevant papers that have shaped current
management.
Pre-operative versus post-operative chemoradiotherapy

Pre-operative or neoadjuvant radiotherapy has two main advantages over post-operative
treatment: decreased tumour seeding at operation and increased patient compliance with
treatment. This must be balanced with the potential over-staging that may occur in some
patients, leading to unnecessary neoadjuvant radiotherapy.
Study references
Sauer, R., Becker, H., Hohenberger, W., et al. (2004) Preoperative versus postoperative
chemoradiotherapy for rectal cancer. New England Journal of Medicine, 351, 1731–40.
Roh, M.S., Colangelo, L.H., O’Connell, M.J. et al. (2009) Preoperative multimodality therapy improves
disease-free survival in patients with carcinoma of the rectum: NSABP R-03. Journal of Clinical
Oncology, 27, 5124–30.
Study design
◆ Both are prospective RCTs.
Sauer et al. NSABP R-03

Randomization Pre-operative versus post-operative CRT
Tumour stage All patients T3/T4
TME surgery Yes Not all
◆ Pre-operative radiotherapy (RT) consisted of 50Gy for all groups, except for the
post-operative CRT group in Sauer et al. who received 55Gy.
NEOADJUVANT THERAPY IN PRIMARY RECTAL CANCER 251
Results
Sauer et al. NSABP R-03
Pre-op CRT Post-op CRT Pre-op CRT Post-op CRT
(n = 421) (n = 402) (n = 130) (n = 137)
5-year local recurrence 6% 13%* 10.7% 10.7%
5-year overall survival 76% 74% 74.5% 65.6%
*p = 0.006
◆ In the Sauer et al. study, some patients thought on initial assessment to require an
APER subsequently had a restorative procedure following pre-operative CRT.
◆ Local control was improved in the pre-operative CRT group in the Sauer et al. study.
◆ Five-year disease-free survival was significantly better in the pre-operative CRT
group in the NSABP R-03 study.
◆ Neither study showed a significant difference in overall survival.
Conclusions
Pre-operative CRT results in improved local control in patients with T3/T4 rectal cancer
compared with post-operative CRT.
Critique
The validity of the results from the NSABP R-03 study was limited, as TME was not per-
formed in all patients. In addition, this study closed early because of low patient recruit-
ment. In contrast, all patients underwent a TME resection in the Sauer et al. study,
providing level I evidence supporting the use of neoadjuvant CRT in patients with T3/T4
rectal cancer. It is also notable that no survival improvements were documented despite
significant benefits in local recurrence.
Neoadjuvant combined chemoradiotherapy versus

radiotherapy alone
Given that chemotherapy potentially sensitizes tissues to the actions of radiotherapy,
these two trials compared long-course neoadjuvant radiotherapy with long-course radio-
therapy plus chemotherapy to determine if the addition of chemotherapy improved
long-term outcomes.
Study references
Gérard, J.P., Conroy, T., Bonnetain, F., et al. (2006) Preoperative radiotherapy with or without
concurrent fluorouracil and leucovorin in T3–4 rectal cancers: results of FFCD 9203. Journal of
Bosset, J.F., Collette, L., Calais, G., et al. (2006) Chemotherapy with preoperative radiotherapy in rectal
cancer: EORTC 22921. New England Journal of Medicine, 355, 1114–23.
Study design
◆ Both were RCTs.
◆ The FFCD 9203 trial compared pre-operative RT with or without chemotherapy.
◆ The EORTC 22921 trial compared pre-operative RT ± post-operative chemotherapy
and pre-operative CRT ± post-operative chemotherapy.
◆ Chemotherapy consisted of 5-fluorouracil–leucovorin and pre-operative RT was 45Gy.
Results
FFCD 9203 EORTC 22921

Pre-op RT Pre-op CRT Pre-op RT Pre-op RCT
(n = 367) (n = 375) (n = 505) (n = 506)
T3/T4 tumours 85.6%/11.2% 88.5%/9.9% 90%/10% 90%/10%
5-year local recurrence 16.5% 8.1%* 17% (9.6% with 8.7%* (7.6% with
post-op CT) post-op CT)
5-year overall survival 67.9% 67.4% 64.8% 65.8%
*p <0.05
Conclusions
The administration of neoadjuvant radiotherapy and chemotherapy in patients with
T3/T4 rectal cancer significantly reduces local recurrence, but does not significantly affect
overall survival.
Critique
Both studies suffer from the same limitation, namely that not all patients underwent TME
surgery (36.8% of patients in the EORTC trial). However, a significant benefit in terms of
local control with neoadjuvant CRT was apparent, particularly when the four arms of the
EORTC group were analysed.
The evidence for short course pre-operative radiotherapy (SCPRT)

Three trials from four countries provide the current evidence for SCPRT. The first study
came from Sweden in 1997, followed by the Dutch study in 2001 and the UK–Canadian
trial published most recently in 2006.
Study references
Swedish Rectal Cancer Trial (1997) Improved survival with preoperative radiotherapy in resectable
rectal cancer. New England Journal of Medicine, 336, 980–7.
Dutch Colorectal Cancer Group (2001) Preoperative radiotherapy combined with total mesorectal
excision for resectable rectal cancer. New England Journal of Medicine, 345, 638–46.
Sebag-Montefiore, D., Stephens, R.J., Steele, R., et al. (2009) Preoperative radiotherapy versus
postoperative chemoradiotherapy in patients with rectal cancer (MRC CR07 and NCIC-CTG C016):
a multicentre, randomised trial. Lancet, 373, 811–20.
NEOADJUVANT THERAPY IN PRIMARY RECTAL CANCER 253
Study design
◆ All three studies are RCTs.
Swedish trial Dutch trial MRC CR07

Randomization SCPRT versus surgery alone
Number of centres 70 108 80
Stage I:II:III/IV (%) 38:35:27 32:33:35 30:28:40 27:27:43 31:28:41 22:34:44
TME No 924 (49.7%) Yes
Follow-up Minimum 5 years Median 2 years Median 4 years
◆ A total dose of 25Gy was administered in five fractions in the SCPRT groups.
◆ In the Swedish study, patients underwent barium enemas for pre-operative staging
with no pre-operative MRI or endoanal ultrasound.
◆ In the MRC CR07 trial, selected patients in the surgery-only group received post-
operative CRT (45Gy in 25 fractions) if the CRM was positive.
Results
Swedish trial Dutch trial MRC CR07
SCPRT Surgery SCPRT Surgery SCPRT Surgery
(n = 583) (n = 585) (n = 924) (n = 937) (n = 674) (n = 676)
Local recurrence
2-year (%) 9* 24 2.4* 8.2 3.4* 8.3
5-year (%) 11* 27 5.6* 10.9 4.7* 11.5
Overall survival
2-year (%) 82.0 81.8 86.1 84.8
5-year (%) 58* 48 64.2 63.5 70.3 67.9
* p <0.01
◆ Subgroup analysis of the three trials investigated the relationship of local recurrence
to tumour stage:
● the MRC-CR07 showed no statistical differences between stages;
● the Dutch trial showed a significant difference in stage III patients;
● the Swedish trial demonstrated a significant reduction in local recurrence in stages
I, II, and III.
Conclusions
SCPRT reduces local recurrence in patients with resectable rectal cancer but does not
appear to have an impact on long-term survival.
Critique
All three randomized trials found that SCPRT reduced local recurrence in resectable rec-
tal cancer. One criticism of the earlier Swedish and Dutch trials is that not all patients
underwent TME surgery. The MRC CR07 trial, which was conducted in the era of TME,
found reduced local recurrence with SCPRT translating into an increase in disease-free
survival from 74.9% to 79.5%. The quality of TME surgery performed in this trial was
documented in a further publication by Quirke et al. Overall, 52% of TMEs were graded
as good (mesorectal) and 13% as poor (muscularis propria plane). This resulted in the
latter group having a significantly higher local recurrence rate (4% versus 13% at 3 years).
The Swedish study was the only one to find an improvement in survival in patients who
had received SCPRT, a result that was sustained on longer follow-up (median 13 years)
(Folkesson et al.). This may be explained by the higher local recurrence in this study
which did not involve TME surgery.
The Association of Coloproctology of Great Britain and Ireland (ASGBI) has produced
a guideline which summarizes the results of these and other trials to facilitate clinical
decision-making within a multi-disciplinary team (MDT). Although decisions must be
individualized for each patient, the following summarizes current thinking in terms of
the management of rectal cancer.
◆ Long-course pre-operative CRT rather than RT alone should be used to downstage
rectal cancers and reduce the risk of local recurrence in rectal tumours where the
CRM is threatened or involved by tumour extension or if metastatic internal iliac/
obturator nodes are present (FFCD 9203; EORTC 22,921).
◆ Patients with early rectal cancers (stage I, T1–T2) should go straight to surgery as
SCPRT offers no benefit in terms of survival or risk of local recurrence (Dutch and
MRC CR07 trials).
The area of debate lies between these two extremes. For stage III patients with a resect-
able primary, the Dutch trial supports the administration of SCPRT (confirmed in a sub-
sequent study with follow-up of median 12 years (van Gijn et al.)). In contrast, the MRC
CR07 trial did not find a difference, although the study may have lacked the power to show
an interaction. The ACPGBI advocates discussion with the patient in these circumstances
and the potential side effects of SCPRT must be taken into consideration. These include
delayed wound-healing and long-term bowel dysfunction. An up-to-date Cochrane review
is expected which may provide answers for patients with stage II and stage III rectal cancers.
Quirke, P., Steele, R., Monson, J., et al. (2009) Effect of the plane of surgery achieved on local recurrence
in patients with operable rectal cancer: a prospective study using data from the MRC CR07 and
NCIC-CTG C016 randomised clinical trial. Lancet, 373, 821–8.
Folkesson, J., Birgisson, H., Påhlman, L., et al. (2005) Swedish Rectal Trial: long lasting benefits from
radiotherapy on survival and local recurrence rate. Journal of Clinical Oncology, 23, 5644–50.
Association of Coloproctology of Great Britain and Ireland (2007) Guidelines for the Management of
Colorectal Cancer (3rd edn). London: ACPGBI.
van Gijn, W., Marijnen, C.A., Nagtegaal, I.D., et al. (2011) Preoperative radiotherapy combined with
total mesorectal excision for resectable rectal cancer: a 12-year follow-up of the multicentre
randomised controlled TME trial. Lancet Oncology, 12, 575–82.
ADJUVANT THERAPY IN COLON CANCER 255
6.16 Adjuvant therapy in colon cancer

Details of study
Prior to the publication of this study, there was no clear long-term benefit for post-
operative adjuvant chemotherapy in patients who had undergone ‘curative’ resection of
colon cancer. Indeed, a meta-analysis had found an increased mortality risk in patients
who had received chemotherapy (Buyse et al.). Initial support for adjuvant therapy came
from a small study assessing adjuvant chemotherapy with levamisole and fluorouracil
(5-FU), which reported a reduction in cancer recurrence (Laurie et al.). The same group
subsequently published the early results of a much larger study which showed that patients
with stage III resected colon cancer treated with 5-FU plus levamisole had a 41% reduc-
tion in recurrence and a 33% reduction in mortality compared with patients who had not
received chemotherapy (Moertel et al.). The long-term results from the same study form
the basis of this seminal paper, which led to the widespread use of post-operative adjuvant
chemotherapy in patients with stage III cancer of the colon.
Study references
Main study
Moertel, C.G., Fleming, T.R., MacDonald, J.S., et al. (1995) Fluorouracil plus levamisole as effective
adjuvant therapy after resection of stage III colon carcinoma: a final report. Annals of Internal
Medicine, 122, 321–6.
Related references
Buyse, M., Zeleniuch-Jacquotte, A., and Chalmers, T.C. (1988) Adjuvant therapy of colorectal cancer.
Why we still don’t know. JAMA, 259, 3571–8.
Laurie, J.A., Moertel, C.G., Fleming, T.R., et al. (1989) Surgical adjuvant therapy of large-bowel
carcinoma: an evaluation of levamisole and the combination of levamisole and fluorouracil. The
North Central Cancer Treatment Group and the Mayo Clinic. Journal of Clinical Oncology, 7,
1447–56.
Moertel, C.G., Fleming, T.R., Macdonald, J.S., et al. (1990) Levamisole and fluorouracil for adjuvant
therapy of resected colon carcinoma. New England Journal of Medicine, 322, 352–8.
Andre, T., Bensmaine, M., Louvet, C., et al. (1999) Multicenter phase II study of bimonthly high-dose
leucovorin, fluorouracil infusion and oxaliplatin for metastatic colorectal cancer resistant to the
same leucovorin and fluorouracil regimen. Journal of Clinical Oncology, 17, 3560–8.
Twelves, C., Wong, A., Nowacki, M.P., et al. (2005) Capecitabine as adjuvant treatment for stage III
colon cancer. New England Journal of Medicine, 352, 2696–704.
Study design
Randomization No adjuvant therapy vs adjuvant levamisole vs 5-FU + levamisole
Number of centres 1
Inclusion criteria Stage III resected colon cancer
Follow-up 6.5 years
◆ Patients were not blinded to the type of adjuvant treatment they received.
◆ Levamisole (50mg orally three times daily for 3 days) was started 7–35 days after
surgery and repeated every 2 weeks for 1 year.
◆ Patients randomized to combination chemotherapy received the same regimen of
levamisole and 5-FU 450mg/m2 intravenously for 5 days in weekly cycles for
48 weeks.
Outcome measures
Primary endpoint
◆ Time to recurrence
Secondary endpoint
◆ Overall survival
Results
No adjuvant treatment Levamisole only 5-FU + levamisole p-value
(n = 315) (n = 310) (n = 304)
Recurrence 177 172 119 <0.0001*
Died 168 158 121 0.007*
* log rank test: significant reduction in recurrence and increased survival in patients receiving 5-FU + levamisole com-
pared with other groups.
◆ The three groups displayed similar demographics and pathology.

◆ After adjustment for other prognostic variables, the 5-FU + levamisole group showed
a 40% reduction in recurrence and a 33% reduction in mortality compared with the
no adjuvant treatment group.
◆ There was no significant reduction in recurrence or mortality in the levamisole-only
group.
◆ In patients who developed a recurrence, those who had received 5-FU + levamisole
had shorter survival than untreated controls.
◆ No long-term side effects of adjuvant therapy were recorded.
Conclusions
The results show that adjuvant therapy with 5-FU plus levamisole may increase cure rates
for patients with resected high-risk colon cancer, with potential reductions in overall
mortality.
Critique
The results reported in this paper encouraged the development of further adjuvant che-
motherapy agents in colon and, subsequently, rectal cancer. It is now widely accepted that
ADJUVANT THERAPY IN COLON CANCER 257
stage III colon cancer patients should receive adjuvant chemotherapy after surgical resec-
tion to improve long-term survival. Newer agents have been introduced which may confer
increased survival benefits and potentially even downstage metastatic disease. Oxaliplatin
is one of these agents, and when it is used in combination with 5-FU, additional survival
advantages may be achieved but at the expense of increased side effects (e.g. peripheral
neuropathy) (Andre et al.). Oral agents have been introduced (capecitabine) and bio-
logical agents are currently being assessed (e.g. cetuximab, bevacizumab). Areas requiring
further research include adjuvant therapy in patients with rectal cancer and patients with
colon cancer who have ‘high-risk’ features (e.g. pT4 disease, vascular invasion).
6.17 Primary chemoradiotherapy for anal cancer

Details of studies
The standard treatment for anal cancer until this trial was abdomino-perineal resection,
or local excision for small tumours. Evidence from case series suggested that high-dose
external beam radiotherapy could also achieve acceptable rates of local control. This
trial was designed to compare the newer regimens of radiotherapy (RT) with combined
modality treatment (CMT), consisting of RT and chemotherapy, in the setting of a ran-
domized trial.
Study references
Main study
UKCCCR Anal Cancer Trial Working Party (1996) Epidermoid anal cancer: results from the UKCCCR
randomised trial of radiotherapy alone versus radiotherapy, 5-fluorouracil and mitomycin (ACT I).
Lancet, 348, 1049–54.
Northover, J., Glynne-Jones, R., Sebag-Montefiore, D., et al. (2010) Chemoradiation for the treatment
of epidermoid anal cancer: 13-year follow-up of the first randomised UKCCCR Anal Cancer Trial
(ACT I). British Journal of Cancer, 102, 1123–8.
Related references
Bartelink, H., Roelofsen, F., Eschwege, R., et al. (1997) Concomitant radiotherapy and chemotherapy
superior to radiotherapy alone in treatment of locally advanced anal cancer: results of a phase III
randomized trial of the EORTC radiotherapy and chemotherapy cooperative groups. Journal of
Flam, M., John, M., Pajak, T.F., et al. (1996) Role of mitomycin in combination with fluorouracil and
radiotherapy, and of salvage chemoradiation in the definitive nonsurgical treatment of epidermoid
carcinoma of the anal canal: results of a phase III randomized intergroup study. Journal of Clinical
Oncology, 14, 2527–39.
Study design
◆ A prospective multi-institutional RCT.
Randomization RT alone versus RT + 5-FU and mitomycin (CMT)
Number of centres 56 centres in four countries
Stratification Radiotherapy centre
Follow-up Every 2 months for the first year
Every 3 months for the second year
Every 6 months until the fifth year
Annually thereafter
PRIMARY CHEMORADIOTHERAPY FOR ANAL CANCER 259
◆ Inclusion criteria: epidermoid cancer of the anal canal or margin; no major hepatic
or renal dysfunction; no previous treatment for cancer at another site; no patients
suitable for local excision (T1, N0).
◆ The RT dose was 45Gy in 20/25 fractions over 4/5 weeks by external beam
irradiation.
◆ The chemotherapy regimen was 5-FU (1000mg/m2 for four days) by continuous
infusion on the first and last week of the radiotherapy. Mitomycin 12mg/m2 was
given as a bolus on day 1 only.
◆ Patients with <50% response at 6 weeks went for radical surgery, while patients with
>50% response had a 15Gy boost of RT or an iridium implant.
Outcome measures
Primary endpoint
◆ Local failure rate >6 weeks after treatment (because of disease or complications of
treatment).
Secondary endpoints
◆ Cause-specific mortality and overall survival.
Results
◆ More patients in the combined modality treatment (CMT) group had
T4 disease ± palpable nodes.
◆ Greater than 90% of patients in both groups had >50% clinical response at 6 weeks,
with around one-third having a complete response
RT CMT p-value
Local failure 164 (61%) 101 (39%) <0.0001
Deaths: anal cancer 105 (39%) 77 (28%) 0.02
Non-anal cancer 20 (7%) 34 (12%)
Overall 3-year survival 155 (58%) 168 (65%) ns
◆ The addition of chemotherapy did not impact on the radiotherapy schedule.

◆ There was significantly more early morbidity with CMT, but late morbidity
was similar in the two groups. Morbidity was primarily skin, gastrointestinal,
genitourinary, and haematological in origin.
◆ Of 228 local failures due to residual/recurrent disease, 133 (58%) underwent salvage
abdomino-perineal resection.
Conclusions
The addition of 5-fluorouracil and mitomycin chemotherapy to radiotherapy reduces the
rate of local failure and requirement for radical surgery. There is also an improvement in
cancer-specific survival with combined modality therapy.
Critique
Until the late 1980s, radical surgical excision was the first-line treatment for anal cancer
in the UK. Nigro et al. introduced the idea of combined chemotherapy and radiotherapy
as early as 1974 and published impressive 5-year survival rates of 80%. The ACT I trial
was the largest randomized trial to look at this area and established CMT as the treatment
of choice for the majority of patients with anal cancer. The trial clearly demonstrated an
improvement in local control with CMT compared with RT alone. As a result, radical
surgery in the form of abdomino-perineal resection is now reserved for salvage purposes,
with a corresponding reduction in the rate of permanent colostomy.
The UK trial did show an increase in the rate of early morbidity and more non-anal
cancer deaths in the CMT arm. There were also six deaths due to toxicity from com-
bined treatment. This led to criticisms that CMT may represent over-treatment for earlier
tumours. Current guidelines suggest that T1 tumours within 2cm of the anal margin can
be treated by local excision with high rates of local control and survival. Late complica-
tions were not increased and this was confirmed by subsequent trials.
The results of the UK trial were further supported by similar trials in Europe (EORTC
trial) and the USA (RTOG trial). The EORTC trial used a similar regimen to the UK
trial but was smaller, with 110 patients randomized. It showed an 18% reduction in local
failure and a 32% increase in the colostomy-free rate at 5 years. The RTOG trial looked
specifically at the addition of mitomycin to radiotherapy and 5-FU in 310 patients, with
a significant improvement in disease-free survival shown. Despite an improvement in
disease-free survival, none of the trials showed a significant improvement in overall sur-
vival during their follow-up.
The results of the ACT I trial were updated after a median follow-up of 13 years. The
results showed a sustained benefit for CMT, which the authors suggest outweighs the early
excess risk of non-anal cancer death. After 12 years the reduction in loco-regional relapse
was 25%, with 12% fewer anal cancer deaths.
An ACT II trial was set up to investigate the use of cisplatin as an alternative to mito-
mycin, but early results suggest no significant improvement in outcome.
Nigro, N.D., Vaitkevicius, V.K., and Considine, B., Jr (1974) Combined therapy for cancer of the anal
canal: a preliminary report. Diseases of the Colon and Rectum, 17, 354–6.
Association of Coloproctology of Great Britain and Ireland (2007) Guidelines for the Management of
Colorectal Cancer (3rd edn). London: ACGBI.
James, R., Wan, S., Glynne-Jones, D., Sebag-Montefiore, L., et al. (2009) A randomized trial of
chemoradiation using mitomycin or cisplatin, with or without maintenance cisplatin/5FU in
squamous cell carcinoma of the anus (ACT II). Journal of Clinical Oncology, 27, 18s.
Chapter 7
Vascular surgery
7.1 Surgery for symptomatic carotid atherosclerosis 262
7.2 Surgery for asymptomatic carotid stenosis 266
7.3 General or local anaesthesia for carotid endarterectomy 270
7.4 The UK Small Aneurysm Trial 272
7.5 The EVAR 1 trial 275
7.6 The EVAR 2 trial 278
7.7 Bypass or angioplasty for severe limb ischaemia (BASIL) 281
7.8 Non-surgical management of intermittent claudication 285
7.9 Endovenous therapy for varicose veins 290
7.10 Superficial venous surgery in leg ulceration 294
262 VASCULAR SURGERY
7.1 Surgery for symptomatic carotid atherosclerosis

Details of studies
The practice of carotid endarterectomy (CEA), although widely practised previously,
was only put on a firm scientific footing for the first time by the publication in 1991 of
two multi-centre prospective randomized studies comparing surgery with best medical
therapy (BMT). One was based in North America and organized by the North American
Symptomatic Carotid Endarterectomy Trial (NASCET) Collaborators and the other,
the European Carotid Surgery Trial (ECST), was based in Europe under the auspices of
the MRC. Following the publication of the initial results in 1991, each of these research
groups subsequently published additional information addressing the question of surgery
for moderate degrees of stenosis and also long-term follow-up of the original cohorts.
Study references
Initial reports
NASCET
North American Symptomatic Carotid Endarterectomy Trial Collaborators (1991) Beneficial effect of
carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. New England
MRC ECST
European Carotid Surgery Trialists’ Collaborative Group (1991) MRC European Carotid Surgery Trial:
interim results for symptomatic patients with severe (70–99%) or with mild (0–29%) carotid steno-
sis. Lancet, 337, 1235–43.
Subsequent publications
NASCET
Barnett, H.J., Taylor, D.W., Eliasziw, M., et al. (1998) Benefit of carotid endarterectomy in patients with
symptomatic moderate or severe stenosis. North American Symptomatic Carotid Endarterectomy
Trial Collaborators. New England Journal of Medicine, 339, 1415–25.
MRC ECST
European Carotid Surgery Trialists’ Collaborative Group (ECST) (1996) Endarterectomy for moder-
ate symptomatic carotid stenosis: interim results from the MRC European Carotid Surgery Trial.
Lancet, 347, 1591–3.
European Carotid Surgery Trialists’ collaborative group (1998) Randomised trial of endarterectomy
for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial
(ECST). Lancet, 351, 1379–87.
Study designs
◆ Both studies were prospective RCTs.
NASCET MRC ECST

Randomization CEA + BMT versus BMT alone (50:50) CEA + BMT versus BMT
alone (60:40)
SYMPTOMATIC CAROTID ATHEROSCLEROSIS 263
Number of centres 106 97

Primary outcome Ipsilateral stroke Fatal/disabling ipsilateral
ischaemic stroke
Secondary outcomes Death, all strokes, major/fatal stroke Death (surgical/any cause),
major/fatal stroke
Follow-up By neurologists at 1, 3, 6, 9, 12 months At 4 and 12 months and
and every 4 months thereafter Duplex US then annually
at 1 month and then annually
◆ Inclusion criteria: internal carotid artery (ICA) stenosis in association with ipsilateral
retinal or hemispheric transient ischaemic attack (TIA) or non-disabling stroke
within 6 months before study entry
◆ Exclusion criteria: life expectancy <5 years and mild (<30%) stenosis (NASCET); the
presence of other possible sources of embolism (MRC ECST).
◆ In the NASCET trial carotid stenosis was estimated by comparing the narrowest
point of stenosis with the normal distal ICA. In the MRC ECST study the narrowest
point of the stenosis was compared with an estimate of the normal diameter of the
ICA at that point. Measurement by the NASCET criteria produced a higher figure for
any given lesion.
Results
Both studies published their initial findings in 1991 and their results were similar
◆ ECST reported outcomes based on three grades of stenosis: mild (<30%), moderate
(30–69%), and severe (70–99%). It showed that surgery for high-grade stenosis
was of benefit, but that patients whose stenoses were <30% should not have
surgery.
◆ NASCET similarly declared that surgery for stenoses >70% produced
benefit.
◆ Both studies found that the position of patients with moderate stenoses was
undecided and the researchers continued to recruit patients in this group to both
studies.
Similarities between the design and results of these long and elaborate studies meant
that it was possible to consider the results in aggregated fashion. An excellent review with
this aim was conducted by the Cochrane Collaboration and has been published in the
Cochrane Library (Rerkasem and Rockwell 2011). In this analysis the authors reviewed
original trial data and reassessed individual carotid angiograms to ensure consistency of
measurement of stenoses and enable accurate pooling of data. Data from a third, smaller,
trial conducted by the Veterans Affairs (VACSP) were included in some subgroups in this
paper. The data presented here are drawn from this analysis.
Study/subgroup Surgery No surgery

(events/total patients) (events/total patients)
Near occlusion
ECST 6/78 1/47
NASCET 6/79 6/67
Total 12/157 7/114
Severe stenosis (70–99%)
ECST 12/257 20/172
NASCET 10/261 33/264
VACSP 3/71 0/70
Total 25/589 53/506
Moderate stenosis (50–69%)
ECST 24/380 18/266
NASCET 11/428 21/428
Total 35/808 39/694
Minor stenosis (30–49%)
ECST 18/302 10/185
NASCET 19/465 22/477
Total 37/767 32/662
Minimal disease (<30%)
ECST 36/783 14/538
*Modified from Rerkasem and Rothwell (2011).
◆ Events are disabling or fatal ipsilateral ischaemic stroke or operative stroke and
operative death at 5 years follow-up.
◆ Post hoc analysis of the pooled data revealed three statistically significant and
clinically important factors which had an effect on the treatment outcome in the
previously defined groups: sex, age, and time to surgery.
◆ The benefit from surgery was greater in men and the elderly; benefit declined with
increasing time from the index neurological event.
Conclusions
The main conclusions drawn from these data are that CEA is useful for stenoses of 70% or
more and that surgery is harmful in patients with stenoses less than 50% and also those
with near-occlusion.
In the group with moderate stenosis, CEA is beneficial in men but not women.
This finding is accounted for by the fact that, in this group, the risk of stroke in women
on medical treatment is less than in men. In addition, the surgical risk is greater in
women.
SYMPTOMATIC CAROTID ATHEROSCLEROSIS 265
Critique
Few studies in the history of surgery have come close to the scale of NASCET and ECST.
It is difficult to overestimate their importance in determining the modern management of
carotid disease. Not surprisingly, they have been subjected to intense scrutiny and exten-
sive re-analysis, as the Cochrane paper cited above shows. In general terms they have
withstood this assault because of their solid design and meticulous follow-up and record-
ing procedures. However, as with all such studies, the guidance they provide in manage-
ment of the individual patient is incomplete and several questions remain unanswered.
Both papers make the point that the benefits of CEA for symptomatic disease can
be realized only if surgery is conducted safely, i.e. with an acceptable peri-operative
stroke and mortality rate. The 30-day combined stroke and mortality rate was 7% in
ECST and 6.7% in NASCET. The point is repeatedly made in the literature that audit of
peri-operative risk produces larger figures when carried out independently than
non-independent audits. In both NASCET and ECST the ratio of fatal to non-fatal peri-
operative stroke was 1:10; any audit that produces a ratio greater than this should be ques-
tioned for completeness. As peri-operative death is a much harder endpoint than stroke,
a higher ratio points towards ‘missed’ strokes in the audit.
In efforts to stratify risks and benefits more accurately, extensive post hoc subgroup
analysis of these studies has been undertaken. Several potentially useful trends have
been identified. Benefits are greater in men, perhaps because the arteries are narrower in
women. The observation that elderly patients(>75 years) benefit disproportionately from
surgery is weakened by the fact that few patients over the age of 80 had surgery. However,
the point is made that, in a surgically fit patient, age is no bar to CEA.
Although these studies are highly regarded they are not without weaknesses. A recur-
ring contention is that BMT at the time of initiation of these trials was much less rigorous
than it is today. Modern medical therapy, including tight control of hypertension, pre-
scription of statins, and maintenance of antiplatelet therapy might be expected to produce
better results than those achieved in the control groups in the trial. On the other hand,
proponents of surgery contend that the benefits of medical therapy are also available to
surgical patients. Furthermore, it is clear that the benefits of surgery appear early, while
there is a clear delay in the accrual of benefits from BMT. This is obviously important
in patients with TIA in whom the period of maximum risk of subsequent stroke occurs
immediately after the index event.
Rerkasem, K. and Rothwell, P.M. (2011) Carotid endarterectomy for symptomatic carotid stenosis
(review). Cochrane Library, Issue 4, CD001081.
7.2 Surgery for asymptomatic carotid stenosis

Details of studies
In common with patients with symptomatic disease, this problem was addressed by
two major studies, one each from North America and Europe. The American study, the
Asymptomatic Carotid Atherosclerosis Study (ACAS), reported results in 1995. The
European study was the MRC-sponsored Asymptomatic Carotid Surgery Trial (ACST)
which was published in the Lancet in 2004. Although the trials had several major differ-
ences, the overall message from each study was remarkably consistent.
Study references
Main studies
Executive Committee for the Asymptomatic Carotid Atherosclerosis Study (1995) Endarterectomy for
asymptomatic carotid artery stenosis. JAMA, 273, 1421–8.
MRC Asymptomatic Carotid Surgery Trial (ACST) Collaborative Group (2004) Prevention of disabling
and fatal strokes by successful carotid endarterectomy in patients without recent neurological symp-
toms: randomised controlled trial. Lancet, 363, 1491–502.
Related references
Chambers, B.R. and Donnan, G.A. (2005) Carotid endarterectomy for asymptomatic carotid stenosis.
Cochrane Database of Systematic Reviews, Issue 4, CD001923.
Study design
ACAS
◆ Multi-centre randomized comparison of surgery versus no surgery in patients with
asymptomatic carotid artery stenosis of at least 60% determined by non-invasive
imaging or angiography.
Randomization CEA versus medical therapy
Number of centres 39 centres in North America
Follow-up 2.7 years (median)
Primary outcome Ipsilateral stroke or any peri-operative stroke or death
Secondary outcomes Ipsilateral TIA, any stroke, major ipsilateral stroke, any major stroke, death
Interventions Surgical group: angiography if not previously done, followed by carotid
endarterectomy within 14 days of randomization and aspirin 325mg daily
Medical group: aspirin 325mg daily
◆ Carotid endarterectomy was carried out within 14 days of randomization.

◆ Medical therapy included 325mg aspirin and risk factor modification.
◆ Inclusion criteria: age between 40 and 79; a carotid stenosis of at least 60% reduction
in diameter; willingness to be followed for 5 years.
ASYMPTOMATIC CAROTID STENOSIS 267
◆ Exclusion criteria: cerebrovascular events in the distribution of the study artery

or the vertebro-basilar system; symptoms in the contralateral system within the
preceding 45 days; contraindication to aspirin therapy; any comorbidity that would
seriously complicate surgery or preclude follow-up for 5 years.
ACST
◆ Multi-centre randomized comparison of immediate versus delayed surgery for
patients with at least 60% stenosis of an internal carotid artery.
Randomization Immediate versus deferred surgery
Number of centres 126 international, mainly European
Primary outcome Peri-operative stroke or death; non-perioperative stroke
Secondary outcomes Non-perioperative fatal or disabling stroke; non-perioperative carotid
territory ischaemic stroke
Follow-up 3.4 years (mean)
◆ In the deferred surgery group carotid endarterectomy was only carried out if
symptoms developed.
◆ Inclusion criteria: unilateral or bilateral severe (>60%) carotid artery stenosis; no
ipsilateral stroke or TIA within the previous 6 months; equipoise on the part of both
doctor and patient about the need for immediate or deferred CEA; no barriers to
long-term follow-up.
◆ Exclusion criteria: previous ipsilateral CEA; the presence of a possible source of
cardiac embolism; significant coincident illness limiting surgical options.
Results
ACAS
◆ The rate of peri-operative stroke or death was 2.3%. The risk for the medical group in
an equivalent period was 0.4%, yielding an excess risk of 1.9% for surgery.
◆ The absolute risk reduction for peri-operative stroke or death or subsequent ipsilateral
stroke was 3.0% at 2.7 years follow-up.
◆ The estimated 5-year risk of ipsilateral stroke and any peri-operative stroke or
death in the surgical group was 5.1%. The equivalent figure in medical patients
was 11.0%.
◆ The estimated risk of major ipsilateral stroke and any peri-operative stroke or death
in the surgical group was 3.4% compared with 6.0% in medical patients.
◆ Subgroup analysis shows that the peri-operative risk was greater in women, and
largely, but not entirely, as a consequence of this the risk reduction due to surgery in
men was greater.
ACST
◆ In the immediate surgery group the peri-operative stroke or death rate was 2.8%.
◆ The absolute risk reduction for peri-operative stroke or death or subsequent
ipsilateral stroke was 3.1% at 3.4 years follow-up.
◆ The estimated 5-year risk of ipsilateral stroke and any peri-operative stroke or death
in the surgical group was 6.4%. The equivalent figure in medical patients was 11.8%.
◆ The estimated risk of major ipsilateral stroke and any peri-operative stroke or death
in the surgical group was 3.5% compared with 6.1% in medical patients.
◆ Subgroup analysis suggests that men benefit more than women.
Conclusions
Surgery is effective in reducing the incidence of stroke in patients with asymptomatic
severe carotid stenosis. The stroke rate at 5 years is approximately halved.
Critique
A study of surgery for asymptomatic carotid stenosis was a logical sequel to the symp-
tomatic studies. To this day in the USA more endarterectomies are performed for asymp-
tomatic than symptomatic carotid disease. These two studies, ACAS and ACST, have
produced remarkably similar results in terms of the benefits to be gained in stroke risk
reduction from carotid surgery. In essence, each trial demonstrated roughly a halving of
the annual stroke risk.
Uncritical acceptance of these benefits would not be appropriate for two reasons. The
first is that the risk of stroke in asymptomatic patients is already very low and therefore
the NNT to prevent one event is relatively high. This means that the cost of therapy may
be unacceptable. The second caveat relates to the historical nature of the trial and the
developments in BMT that have naturally occurred since the trial was conducted. Some
commentators believe that the small benefits of CEA demonstrated by these studies have
already been obliterated by improvements in the medical care of patients with arterial dis-
ease. On the other hand, proponents of surgery point out that the benefits of medical care
apply equally to both groups and that there should still be an added benefit of surgery.
However, this ignores the fact that any reduction in the overall risk of stroke increases the
NNT and makes any benefit from surgery not only more difficult to obtain but also more
expensive.
Finally, it is acknowledged that surgical results obtained in properly conducted trials
rarely reflect those obtained in real life. This observation is of particular relevance when
dealing with prophylactic surgery. The authors emphasize that the benefits of endarter-
ectomy can only be realized if the morbidity of surgery is low. A small increase in peri-
operative stroke risk effectively removes any benefit of CEA in asymptomatic patients.
Therefore any individual surgeon undertaking CEA in asymptomatic patients must be
able to demonstrate an ability to carry out the surgery with a very low morbidity rate.
ASYMPTOMATIC CAROTID STENOSIS 269
Recently, some authors have firmly decided that surgery now has no role to play in the
prevention of stroke in patients with asymptomatic carotid artery stenosis. In support of
this they cite several factors, some discussed already: improvements in best medical care,
a reduction in the incidence of stroke, and the fact that surgical results ‘in the community’
rarely match the high standards of those obtained in trials.
7.3 General or local anaesthesia for carotid endarterectomy

Details of studies
Proponents of local anaesthesia (LA) for carotid endarterectomy (CEA) postulated
that an awake patient can be monitored reliably for signs of cerebral ischaemia intra-
operatively and if necessary have a shunt inserted to augment cerebral perfusion. This strat-
egy seeks to avoid the risks of placement of unnecessary shunts. In addition, they believed
that LA carried a lower risk of myocardial ischaemia and therefore might be expected to
induce fewer acute coronary events peri-operatively. As a result of these two putative ben-
efits, they thought that the operative risk of CEA could be reduced. These hypotheses were
eventually put to the test in a prospective randomized study called the GALA trial.
Study reference
GALA Trial Collaborative Group (2008) General anaesthesia versus local anaesthesia for carotid surgery
(GALA): a multicentre, randomised controlled trial. Lancet, 372, 2132–42.
Study design
Randomization LA versus GA CEA
Number of centres 95 centres from 24 countries
Primary outcomes Stroke, MI, or death between randomization and 30 days following surgery
Secondary outcomes Stroke-free survival, MI, death within 1 year of anaesthesia; length of stay
in recovery, HDU, ICU, and hospital; HRQoL
Follow-up 1 year (for >75% of patients)
Results
◆ Between randomization and 30 days following surgery 4.0% of patients having CEA
under GA experienced a stroke while 3.7% of those performed under LA did so. This
difference was not statistically significant.
◆ The proportions of patients suffering a primary outcome event were 4.8% for GA and
4.5% for LA, again not a statistically significant difference.
◆ Unexpectedly, there was an excess incidence of myocardial infarction (MI) among
those who received surgery under LA (0.5%) compared with those who had GA
(0.2%).
◆ The incidences of cranial nerve injury and wound haematoma in the two groups were
very similar.
◆ No significant difference in primary outcome was observed for the predefined
subgroups of age, contralateral carotid occlusion, and baseline surgical risk. There
was a trend towards reduced complications in those with contralateral occlusion
treated under LA.
GENERAL OR LOCAL ANAESTHESIA FOR CAROTID ENDARTERECTOMY 271
◆ No differences were observed in any of the secondary outcome measures.

◆ Use of regional anaesthesia is associated with a much reduced use of intra-operative
shunts.
Conclusions
This study did not support the hypothesis that carotid surgery under LA carried a lower
risk of morbidity.
Critique
This study did not demonstrate the expected reduction in adverse events associated with
CEA that previous studies had indicated would occur with the use of regional anaesthesia.
Commentators have pointed out that, despite the huge effort invested in the planning
and execution of this trial, it is underpowered to exclude the possibility of a reduction in
mortality from 1.5% to 1.0% that is suggested by previous observations. Any trial designed
to do this would require around 25,000 patients, an unfeasibly large number.
The apparent increase in the incidence of MI associated with LA may be a real phenom-
enon. It has been postulated that the epinephrine in the local anaesthetic solution may be
partly responsible for this by causing an increase in heart rate. It is also true that lightly
sedated patients undergoing this procedure may experience a greater stress response,
leading to increased circulating norepinephrine.
Critics of this study point out that one of the big differences between the groups was the
incidence of shunt usage. This is not surprising, since one of the objects of LA is to limit
shunting to those who demonstrate signs of reduced cerebral perfusion. However, shunt-
ing is not an uncomplicated procedure, and some measure of the effect of shunting itself
on the stroke rate would have been interesting. Large differences in the use of shunts were
also noted between UK and non-UK surgeons. The type of shunt used was not recorded,
and this may also be relevant because some stiffer shunts may be more traumatic than
others. Finally, statin use was not recorded and many believe that these drugs have a use-
ful function in the peri-operative period in patients undergoing CEA.
So where does this leave the debate about GA versus LA? The non-significant benefit of
LA demonstrated here may encourage those who note that the possibility of a real differ-
ence has not been excluded by this trial. However, three fewer adverse outcomes per thou-
sand patients is a small difference. Ultimately, most commentators believe that the choice
of anaesthesia is a matter for discussion involving patient, surgeon, and anaesthetist.
Although GALA is a landmark paper, it is remarkable for the fact that it has probably
not influenced practice at all!
7.4 The UK Small Aneurysm Trial

Details of studies
Abdominal aortic aneurysms (AAAs) often remain symptomless until they rupture or
rapidly expand, and therefore they are an important cause of sudden death. It is estab-
lished that the risk of rupture increases with aortic diameter. However, uncertainty
existed as to whether prophylactic surgery is the best management for smaller aneurysms
of 4.0–5.5cm in diameter. This RCT was designed to ascertain whether early prophylactic
open surgery for small AAAs reduced long-term mortality.
Study references
Main study
UK Small Aneurysm Trial Participants (1998) Mortality results for randomised controlled trial of early
elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. Lancet, 352,
1649–55.
Study design
◆ A multi-centre RCT
Randomization Ultrasound surveillance versus early elective open surgery for small AAA
(4.0–5.5cm)
Number of centres 93 UK centres
Follow-up (F/U) USS group 4.0–4.9cm—F/U every 6 months, 5.0–5.5cm—F/U every
3 months, >5.5cm or >1cm/year expansion → surgery
Surgery group F/U 1 month post-operatively, survival check at end of trial
◆ Inclusion criteria: patients aged 60–76 years deemed fit for elective surgery and
identified as having symptomless infra-renal abdominal aortic aneurysms.
◆ Exclusion criteria: patients unfit for elective surgery or those with symptoms related
to their aneurysm. Patients were also excluded if unable to give consent for surgery
or to attend follow-up.
◆ Deaths occurring within 2 weeks of elective surgery were attributed to abdominal
aortic aneurysm.
Outcome measures
Primary endpoint
◆ Death.
Secondary endpoints
◆ Aneurysm rupture and death from surgical repair.
UK SMALL ANEURYSM TRIAL 273
Results
◆ 517 of 563 patients randomized to early surgery underwent open repair, with 87%
performed within 5 months of randomization. Forty-three patients did not undergo
repair—nine died prior to admission, nineteen became unfit for elective surgery,
thirteen refused surgery after randomization, and two were found to have other
pathology at laparotomy. Therefore no AAA repair was performed.
◆ 321 of 527 patients randomized to ultrasound surveillance eventually underwent surgery
because of rapid expansion or growth of the aneurysm above the 5.5cm threshold.
◆ 283 patients were operated upon according to protocol and 38 were not; 25 patients
requested surgery, 12 patients underwent imaging by alternative means, showing
diameter in excess of 5.5cm, and one patient had surgery performed by a surgeon not
involved in the trial.
◆ The overall hazard ratio for all-cause mortality in the early elective surgery group
compared with the ultrasound surveillance group was 0.94.
◆ The 30-day operative mortality in the early surgery group was 5.8%.
◆ Mortality did not differ significantly between the groups at 2, 4, or 6 years.
◆ Age, sex, or initial aneurysm diameter did not modify the hazard ratio.
Conclusions
This study shows that early surgery for small abdominal aortic aneurysms does not pro-
vide a long-term survival advantage. Ultrasound surveillance is safe for small AAAs, and
the results from this trial do not support a policy of open surgical repair for AAAs of
diameter 4.0–5.5cm.
Critique
The UK Small Aneurysm Trial has led to the widely held view that abdominal aortic aneu-
rysms of diameter <5.5cm should not be surgically repaired and that it is safe to perform
ultrasound surveillance on these patients. The large numbers and good design of this ran-
domized trial make the results statistically robust and extremely important. Nevertheless,
there are some important discussion points.
The 30-day mortality rate in the early surgery group was 5.8%, which is greater than
in many published series and almost three times higher than the expected mortality rate
used for the power calculation in the study design. This finding adds support to the con-
tention that published mortality figures are often an example of publication bias arising
from the tendency to publicize good results and are not indicative of outcomes in the ‘real’
surgical community. Vascular surgeons who are consistently able to achieve low mortal-
ity rates for open aneurysm repair may feel that early surgery for smaller aneurysms may
confer a long-term benefit in their hands.
There was a statistically insignificant benefit from early surgery for younger patients
(60–66 years) and those with larger aneurysms (4.9–5.5cm). This may well lead to some
units recommending early surgery when aneurysms reach 5.0cm, particularly in young
low-risk patients. However, some suggest that the threshold for intervention in small
AAAs might be even larger than the 5.5cm studied here. Post-mortem studies have dem-
onstrated that more than half the patients with an AAA die with the aneurysm unrup-
tured. Early surgery in these patients does not prolong their lives but exposes them to the
risk of premature mortality from elective surgery. Furthermore, other studies have shown
that 80% of ruptured aneurysms measure at least 6.0cm. The low rate of rupture in this
study is consistent with these previous observations.
There remains uncertainty with regard to the management of patients with AAAs out-
with the inclusion criteria of this trial. Young patients (aged <60 years), in particular, pose
management dilemmas in that the results cannot be extrapolated to them, but in practice
decisions with regard to their management are often based on the recommendations from
this study.
The advent of endovascular aneurysm repair (EVAR) has led to significant change in the
management of AAAs. The lower 30-day mortality from EVAR means that it is difficult to
extrapolate the results from the UK Small Aneurysm Trial to a population now undergo-
ing endovascular repair. Smaller aneurysms are more likely to be anatomically suitable
for EVAR, and therefore further research is needed to clarify the role of this increasingly
popular technique. This will be discussed in more detail in the sections 7.5 and 7.6.
This large high-quality multi-centre RCT has gone a long way towards shaping the way
vascular surgeons manage AAAs. Controversies remain, in particular with the advent of
EVAR clouding the waters in terms of indications for surveillance versus operative endo-
vascular repair. Future studies will aim to clarify this further.
EVAR 1 TRIAL 275
7.5 The EVAR 1 trial

Details of studies
Endovascular aneurysm repair (EVAR) was developed in the early 1990s and initial expe-
rience suggested that it carried a lower operative mortality than open AAA repair. The
long-term results of EVAR were less certain and, given the apparently high early incidence
of graft-related complications, the need for close surveillance was clear. This apparent
trade-off between lower early mortality and early and late graft failure was examined in
the UK Endovascular Aneurysm Repair 1 (EVAR 1) trial. This study was designed to
compare EVAR with open repair in terms of mortality, durability, health-related quality
of life (HRQoL), and cost.
Study references
Main study
United Kingdom EVAR Trial Investigators (2010) Endovascular versus open repair of abdominal aortic
aneurysm. New England Journal of Medicine, 362, 1863–71.
EVAR Trial Participants (2005) Endovascular aneurysm repair versus open repair in patients with
abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet, 365, 2179–86.
EVAR Trial Participants (2004) Comparison of endovascular aneurysm repair with open repair in
patients with abdominal aortic aneurysm (EVAR Trial 1), 30-day operative mortality results: ran-
domised controlled trial. Lancet, 364, 843–8.
Related references
DREAM Study Group (2010) Long-term outcomes of open or endovascular repair of abdominal aortic
aneurysm. New England Journal of Medicine, 362, 1881–9.
Study design
Class of evidence 1b
Randomization EVAR versus open surgical repair of AAA
Inclusion criteria Patients >60 years with AAA >5.5cm anatomically suitable for EVAR and fit
for either EVAR or open surgery
Follow-up 1, 3, and 12 months; then annually
CT performed annually in both groups
◆ If unfit for open repair, patients were offered entry into the EVAR 2 trial (see section 7.6).
◆ Randomization via one-to-one ratio in randomly sized permuted blocks stratified by
centre.
◆ Aneurysm-related mortality was defined as death within 30 days of AAA repair
unless overruled by subsequent post-mortem.
Outcome measures
Primary endpoint
◆ All-cause mortality.
Secondary endpoints
◆ Aneurysm-related mortality, incidence of post-operative complications related to
aneurysm repair and secondary interventions, HRQoL, and hospital costs.
Results
EVAR Open surgery p-value
(n = 626) (n = 626)
30-day operative mortality 1.8% 4.3% 0.02
Aneurysm-related mortality 5.7% 6.4% 0.73
All-cause mortality 41% 42% 0.72
Graft-related complications 45% 12% <0.001
Cost (mean)
Aneurysm repair £13,019 £11,842
Aneurysm-related readmissions £2,283 £442
◆ Median follow-up of the study was 6 years.

◆ In the open repair group, 567 underwent open repair, 31 underwent EVAR, 4
underwent emergency open repair, and 24 had no surgery.
◆ In the EVAR group, 598 underwent EVAR, 16 had open repair, and 12 did not
undergo repair.
◆ The lower 30-day mortality gave the EVAR group an early benefit in aneurysm-
related mortality, but this was lost by the end of the study, in part due to fatal
endograft ruptures.
◆ Additional complications still appeared up to 8 years after randomization.
Conclusions
The EVAR 1 trial demonstrates that EVAR has one-third of the operative mortality
of open AAA repair. With increasing follow-up EVAR offers no significant advantage
with respect to all-cause mortality or aneurysm-related mortality when compared with
open repair. EVAR is more expensive and results in more complications and a greater
need for re-intervention. These findings support the notion that long-term follow-up and
surveillance of EVAR patients is essential.
Critique
The advent of endovascular aneurysm repair has resulted in a huge culture shift in
the management of abdominal aortic aneurysms. This well-designed large multi-centre
EVAR 1 TRIAL 277
RCT aimed to compare the two treatments. It demonstrated that EVAR results in sig-
nificantly lower operative mortality at the expense of more complications and subse-
quent re-interventions, and higher cost. There was no significant difference in terms
of aneurysm-related and all-cause mortality. These results were mirrored by the Dutch
Randomized Endovascular Aneurysm Repair (DREAM) trial where 351 patients with
aneurysms of at least 5cm were randomized to EVAR or open surgery. Six years after
randomization, 69.9% of the open repair patients were still alive compared with 68.9%
of the EVAR group (p = 0.97). During the follow-up period, 29.6% of the EVAR group
underwent at least one re-intervention compared with 18.1% of the open group (p = 0.03).
The DREAM study showed that despite the short-term benefits of reduced operative mor-
tality of EVAR, there was no difference in later overall mortality. Furthermore, EVAR was
associated with long-term problems of durability.
The authors of the EVAR 1 trial acknowledged the limitations of their study and it is
useful to discuss these here. The endografts utilized were second- and third-generation
devices and technology has progressed significantly since the start of this trial. Modern
stent grafts have attempted to address the issues of migration, device fragmentation, and
poor durability with new low profile endografts incorporating suprarenal fixation, fen-
estration, and branched elements amongst other innovations. This has led to increasing
numbers of EVARs being performed for more complex anatomy. There is now much
greater procedural experience, with improved operator skills and better case-selection
strategies leading to enhanced success rates. It is anticipated that future studies will show
that modern devices have lower complication rates, and thus it is difficult to apply these
results unreflectively in the context of modern EVAR.
It is likely that in the early stages of the trial, a proportion of EVAR-related complica-
tions were not appropriately recorded since the study began three years before standard-
ized reporting of endovascular complications was introduced. It is also possible that there
has been an underestimation of re-interventions in both groups as data on outpatient
procedures and readmissions related to complications from open repair (e.g. incisional
hernia) were not collected.
The continued occurrence of graft-related complications, ranging from endoleaks to
graft rupture, has led to robust post-EVAR surveillance programmes. These invariably
lead to increased costs attributed to EVAR, along with radiation exposure and its inherent
risk. At present, there are numerous variations of surveillance protocols aiming to provide
early detection of endograft-related complications in a safe and cost-effective manner.
In the context of this study, incidences of fatal endograft rupture should be significantly
reduced nowadays.
The decision on whether to perform EVAR or open repair on anatomically suitable
patients is based on a number of factors, including patient preference, patient physiol-
ogy, local practice, and cost. As endograft technology becomes more sophisticated, with
reduced complications and increased durability backed up by clinical trials, it is likely that
the role of EVAR in the management of AAA will continue to grow.
7.6 The EVAR 2 trial

Details of studies
Endovascular repair of abdominal aortic aneurysm (EVAR) was originally developed by
Dr Juan Parodi in 1990. He sought an alternative means of repair for abdominal aortic
aneurysm (AAA) with lower morbidity and mortality than open repair and therefore
potentially more suited for those deemed unfit for major surgery. Initial results from EVAR
suggested that peri-operative mortality and morbidity were indeed lower. As a consequence
the United Kingdom Endovascular Aneurysm Repair 2 (EVAR 2) trial was designed to test
the hypothesis that EVAR could be carried out safely and beneficially in such patients.
Study references
Main study
United Kingdom EVAR Trial Investigators (2010) Endovascular repair of aortic aneurysm in patients
physically ineligible for open repair. New England Journal of Medicine, 362, 1872–80.
EVAR Trial Participants (2005) Endovascular aneurysm repair and outcome in patients unfit for open
repair of abdominal aortic aneurysm (EVAR Trial 2): randomised controlled trial. Lancet, 365, 2187–92.
Study design
Randomization EVAR versus BMT
Inclusion criteria Patients aged >60 years with AAA >5cm anatomically suitable for EVAR but
unfit for open repair
Exclusion criteria Fit for open repair but hostile abdomen prevented open surgery
Follow-up 1 and 3 months initially in EVAR group; annually in both groups
CT performed annually in both groups
◆ Randomization was via 1:1 ratio in randomly sized permuted blocks stratified by
centre.
◆ Aneurysm-related mortality was defined as death within 30 days of AAA repair.
Outcome measures
Primary endpoint
◆ All-cause mortality
Secondary endpoints
◆ Aneurysm-related mortality.
◆ Graft-related complications and re-interventions.
EVAR 2 TRIAL 279
◆ Health-related quality of life.

◆ Hospital costs.
Results
◆ Median follow-up until death or the end of the study was 3.1 years.
◆ In the EVAR group, 179 patients underwent elective EVAR; 18 patients died prior to
AAA repair.
EVAR (n = 197) BMT (n = 207) p-value

30-day operative mortality 7.3% n/a
Aneurysm rupture rate* n/a 12.4
Aneurysm-related mortality* 3.6 7.3 0.02
All-cause mortality* 21.0 22.1 0.97
Graft-related complications 48% n/a
Cost (mean) £14,995 £5169
*Per 100 person-years.
◆ In the no-intervention group, 137 patients had not undergone AAA repair by the end
of the follow-up period. Seventy patients underwent AAA repair: 52 elective EVAR;
12 elective open repairs; 3 emergency EVAR; 3 emergency open repairs.
Conclusions
This RCT comparing EVAR with best medical therapy for patients unfit for open repair
demonstrates that EVAR is associated with a significantly lower rate of aneurysm-related
mortality compared with BMT. There was no difference in all-cause mortality, and EVAR
was associated with increased complications and re-interventions at a higher cost.
Critique
The EVAR 2 trial is the only RCT addressing the use of EVAR in patients medically unfit
for open repair but anatomically suitable for EVAR. The mid-term results of the trial
offered little support for EVAR in these patients, with a 30-day mortality of 9%. The sub-
sequent addition of 66 patients to the study produced a final 30-day operative mortality
of 7.3% for the EVAR group. This was still substantially more than the operative mortality
observed in the EVAR 1 trial. Despite this, with longer follow-up a benefit in terms of
aneurysm-related mortality has become evident, although these patients have a limited
life expectancy irrespective of whether the aneurysm is repaired.
Regardless of the high-quality nature of the trial in terms of methodology, design, and
the research question being answered, there were well-documented problems with imple-
mentation. There was a median delay of 57 days between randomization and subsequent
procedure in the EVAR arm. Eighteen patients died prior to undergoing EVAR, including
nine patients as a result of AAA rupture. It seems possible that this delay in randomization
may well have adversely influenced the mortality results in the group assigned to EVAR.
A total of 70 patients assigned to the best medical treatment group had surgical repair of
their AAA, including 12 patients who underwent elective open repair despite previously
having been deemed unfit. Remarkably, of those in the no-intervention group who had
elective repair (n = 64) only one died within 30 days of the procedure—a mortality rate
of 1.6%. Post hoc analysis of the fitness of this group, using baseline data, suggested that
they were fitter than the cohort assigned initially to EVAR. This may account for the lower
operative mortality. These patients were censored from the trial at the time of repair.
Each of these factors supports those who believe that EVAR 2 has not completely answered
the question addressed by the trial. This notion is supported by the fact that, at the end of
the trial, 99 patients remained alive of whom only 14 had not had an AAA repair. A post
hoc analysis of all patients who had repair would be interesting, if scientifically unsound.
As previously discussed with respect to the EVAR 1 study (section 7.5), both clinical
experience and stent graft technology have improved considerably since the EVAR 2 trial
started in 1999. Better-quality devices are expected to lead to a lower incidence of endo-
graft problems. This will possibly have an effect on aneurysm-related mortality and graft-
related complications, particularly relevant to an unfit patient population.
The EVAR 2 trial showed that EVAR was associated with a significant reduction in
the rate of aneurysm-related mortality in the long-term compared with the best medical
treatment. However, there was no difference in all-cause mortality between the groups,
although the substantial proportion of patients (34%) in the no-intervention group who
eventually had repair confounds things somewhat. This affects the accuracy of the overall
results and confuses their application to clinical practice.
The rate of all-cause death was high in the EVAR group, predominantly due to car-
diovascular events including MI and stroke, although this finding was not statistically
significant. The mid-term analysis of the EVAR 2 trial reported higher operative mortal-
ity than the long-term analysis (9.0% versus 7.3%), and the subsequent reduction was
due to the recruitment of an additional 66 patients to the study. The reduced mortality
in the long-term analysis may also be attributed to the increased use of statins from 2003
onwards, along with an increased emphasis on pre-operative optimization. Nevertheless,
this figure is significantly higher than the operative mortality demonstrated in the EVAR 1
trial (1.8%). In a population group as medically unfit as those in the EVAR 2 trial, it is
hardly surprising that mortality is high.
Controversy remains within the vascular surgery community as to whether the find-
ings of the EVAR 2 trial justify not offering EVAR to a high-risk population. The prob-
lems cited above make it clear that further research is required, with an emphasis on
interventions and other measures that may reduce operative mortality in vulnerable AAA
patients. It may also help identify subsets of ‘unfit’ patients who may benefit from repair.
BYPASS OR ANGIOPLASTY FOR SEVERE LIMB ISCHAEMIA 281
7.7 Bypass or angioplasty for severe limb ischaemia (BASIL)

Details of study
The BASIL trial investigators sought to compare outcomes of bypass surgery first versus
balloon angioplasty first in severe lower limb ischaemia (SLI) (rest pain, ulceration, and
gangrene of the leg) in a multi-centred RCT.
Study references
Main study
Adam, D.J., Beard, J.D., Cleveland, T., et al. (2005) Bypass versus angioplasty in severe limb ischaemia of
the leg (BASIL): multi-centre randomised controlled trial. Lancet, 366, 1925–34.
Related references
Bell, J., Papp, L., Bradbury, A.W. (2000) Bypass or angioplasty for severe ischaemia of the leg: the BASIL
trial. In: Powell, J.T., Greenhalgh, R.M., and Mitchell, A.W (eds), Vascular and endovascular oppor-
tunities. London: WB Saunders, 485–94.
Bradbury, A.W., Bell, J., Lee, A.J., et al. (2002) Bypass or angioplasty for severe limb ischaemia? A Delphi
Consensus Study. European Journal of Vascular and Endovascular Surgery, 24, 411–16.
Bradbury, A.W., Adam, D.J., Bell, J., et al. (2010) Bypass versus angioplasty in severe ischaemia of the
leg (BASIL) trial: Analysis of amputation free and overall survival by treatment received. Journal of
Vascular Surgery, 51: 18S–31S.
Bradbury, A.W., Adam, D.J., Bell, J., et al. (2010) Bypass versus angioplasty in severe ischaemia of the leg
(BASIL) trial: a survival prediction model to facilitate clinical decision-making. Journal of Vascular
Surgery, 51: 52S–68S.
Bradbury, A.W., Adam, D.J., Bell, J., et al. (2010) Bypass versus angioplasty in severe ischaemia of the
leg (BASIL) trial: an intention-to-treat analysis of amputation-free and overall survival in patients
randomized to a bypass surgery-first or a balloon angioplasty-first revascularization strategy. Journal
of Vascular Surgery, 51, 5S–17S.
Forbes, J.F., Adam, D.J., Bell, J., et al. (2010) Bypass versus angioplasty in severe ischaemia of the leg
(BASIL) trial: health-related quality of life outcomes, resource utilization, and cost effectiveness
analysis. Journal of Vascular Surgery, 51, 43S–51S.
Study design
◆ Multi-centre RCT.
Randomization Bypass-first versus angioplasty-first for infra-inguinal occlusive disease in SLI
Number of centres 27 UK hospitals
Inclusion criteria SLI defined as rest pain or tissue loss (ulceration or gangrene) of presumed
arterial aetiology for >2 weeks
Arterial imaging confirming disease equally treatable by bypass or
percutaneous angioplasty (PTA)
Exclusion criteria Intermittent claudication only with no symptoms of SLI for >2 weeks
Arterial disease not treatable by either bypass or PTA
Comorbid conditions contraindicating bypass or PTA
Stratification By centre, then into four groups by clinical presentation: rest pain only vs
tissue loss ± rest pain and ankle pressure ≥50mmHg vs ankle pressure
<50 mmHg
Follow-up 5 years (1999–2004)
◆ An audit was performed at the six highest recruitment centres (by volume) mid-
way through the recruitment over a 6-month period to assess reasons for non-
interventional treatment and non-randomization in potentially eligible patients.
Outcome measures
Primary endpoint
◆ Amputation-free survival (AFS) measured as time to amputation of trial leg above
the ankle or death from any cause, whichever occurred first.
Secondary endpoints
◆ All-cause mortality, 30-day morbidity and mortality, re-interventions, health-related
quality of life, and cost.
◆ Data for all interventions and hospital stays during follow-up were collected. Total
length of hospital stay was collected for 1 year from the date of randomization, and
cost was estimated using the Scottish system of hospital cost statistics for the average
cost per inpatient day.
Results
◆ Follow-up at end of study: 1 year—449 (99%), 2 years—336 (74%), 3 years—216
(48%), 4 years—99 (22%), 5 years—37 (8%). Follow-up ended once the primary
endpoint was reached.
Bypass first (n = 228) PTA first (n = 224)

Assigned intervention attempted 86% 96%
Immediate failure 3% 20%
Re-intervention rate 18% 26%
Morbidity 56% 41%
Mortality 5% 3%
Amputation free survival
1 year 68% 71%
3 years 57% 52%
Cost in first 12 months (mean) £23,322 £17,419
BYPASS OR ANGIOPLASTY FOR SEVERE LIMB ISCHAEMIA 283
◆ Four patients in the PTA group crossed over to surgery and 21 in the surgery group
crossed over to PTA.
◆ At the end of follow-up (5.5 years), 248 patients (55%) were alive with the trial
leg intact, 38 (8%) were alive with the trial leg amputated, 36 (8%) had died after
amputation of the trial leg, and 130 (29%) had died with the trial leg intact.
◆ Post-hoc analysis suggests improved AFS and all-cause mortality in the surgery
group in the period after 2 years from randomization.
◆ Twenty of the PTA patients had complications following surgery after failed PTA.
◆ There were no differences in long-term quality of life between the two groups.
◆ There was no difference in the number of admissions, but the surgery group had
longer hospital stays and more high dependency unit (HDU) admissions.
◆ The BASIL audit: of 585 consecutive patients, 129 needed supra-inguinal
re-vascularization and 220/456 (48%) were deemed unsuitable for revascularization
(unfit, unsalvageable, or improved with medical treatment). Seventy patients (15%
of original cohort) were entered in the BASIL trial of whom 48/456 (10%) agreed to
randomization.
Conclusions
Overall and amputation-free survivals are similar up to 2 years following either PTA or
surgery for infra-inguinal occlusive disease causing SLI when assessed by an ITT analy-
sis. The complication rate is lower following PTA, and the short-term cost of treatment
appears to be greater following surgery.
Post hoc analyses suggest that if patients survive beyond 2 years (70% of patients) and
are initially randomized to surgery, they have a better clinical outcome than those initially
assigned to angioplasty.
Critique
The BASIL trial remains the seminal comparator of balloon angioplasty versus bypass
surgery for infra-inguinal occlusive arterial disease in severe limb ischaemia. This is the
only published RCT on this topic, about which there was equipoise among surgical and
radiological communities on the preferred method of revascularization in this patient
group. The BASIL trial collaborators confirmed a lack of consensus following a survey of
surgeons and radiologists regarding their opinions on best medical practice.
The overall results do not show clear superiority of one technique over the other, and
management of these patients should be determined by local expertise. The findings from
the 2005 publication, suggesting an improved survival benefit in patients randomized
to a surgery-first strategy, prompted further funding of an extension study and analyses
of long-term outcomes, cost, and quality of life. The results of this study were published
in four papers in 2010 and offer insight into a number of clinical questions. Prolonged
overall survival (OS) and a trend towards prolonged AFS were confirmed. Assessment of
outcomes by treatment received (to account for early crossover) suggests that prosthetic
bypass graft performed poorly compared with vein bypass. Survival was also diminished
in patients who underwent angioplasty first and then required bypass after PTA failure
compared with patients who underwent bypass first. As these results were based on out-
comes by treatment received, the analyses lose the power of randomization.
The BASIL trial was set up as a pragmatic study; however, generalization of the trial
results should be done with caution, as the BASIL trial audit (run over a 6-month period
in the top six recruiting centres) revealed that only 10% of patients presenting and initially
randomized were eventually included in the study. This may have been because of the
wide clinical and anatomical variety that was permitted in different centres. The inves-
tigators included patients with SLI as defined by the presence of rest pain with or with-
out tissue loss, but did not characterize and stratify patients according to the anatomical
extent of disease. The investigators later stratified patients according to the presence of
tissue loss, rest pain, and ankle pressure (<50 or ≥50mmHg) to ensure that randomiza-
tion was equal across subgroups, but the study was not powered for these subgroups. The
success of revascularization and the endpoints of OS and AFS would be expected to differ
between these groups.
OS and AFS (duration to amputation or death) are important endpoints in assessing
outcomes following interventions but are not identical to the clinical aim of infra-inguinal
revascularization, which is symptom relief and maintenance of quality of life through
limb preservation. The primary endpoints in comparing two techniques of revasculariza-
tion should be revascularization failure, and peri-operative deaths and morbidity. The
causes of amputation and overall mortality are multifactorial, and without stratification
of the anatomical and clinical severity of initial disease, patient comorbidities, and patient
lifestyle, the interpretation of amputation and mortality rates becomes difficult. A useful
clinical outcome would be clinical events and re-interventions that can be immediately
related to the initial procedure. These would be difficult to measure, as the timing and
decision for re-interventions would be expected to differ between clinicians, making the
eventual outcomes difficult to compare. In the BASIL trial there was an absence of inter-
vention data and a lack of any consensus on graft and angioplasty surveillance.
The finding that prosthetic bypass grafts performed worse than vein bypass should have
been expected, and although the investigators explained their rationale for including pros-
thetic bypass in the open surgical group, the results are presented together. Stratification
would shed light on the efficacy of vein bypass versus prosthetic in SLI.
Unlike vein bypass, there is no consensus gold standard on angioplasty technique and
post-procedure surveillance. Angioplasty techniques can be expected to vary between
different radiologists and units. The study allowed for heterogeneity in angioplasty tech-
niques. The finding that failed angioplasty-first approach led to higher morbidity and
mortality following subsequent bypass surgery may be a reflection of the selection of
patients and technique of angioplasty, but there was insufficient detail in the study to
explain this finding.
NON-SURGICAL MANAGEMENT OF INTERMITTENT CLAUDICATION 285
7.8 Non-surgical management of intermittent claudication

Details of studies
A review of the literature on vascular surgery from the 1970s and 1980s yields a plethora
of studies of various surgical techniques and comparative examinations of different
types of conduit utilized in the surgical treatment of intermittent claudication (IC).
Following the development of percutaneous transluminal angioplasty (PTA) a similar
avalanche of studies appeared, cataloguing its benefits and comparing this new technique
with surgery.
Interestingly, there is a relative dearth of studies questioning whether any interven-
tion at all for intermittent claudication is in fact the correct approach to this common
symptom. A few early studies suggested that exercise therapy had a role to play in IC
patients, but these were generally overlooked in the enthusiasm for intervention. Only
slowly did the vascular community begin to clarify the precise role that this might play in
the treatment of IC.
It is now accepted that exercise is beneficial in improving walking distances and qual-
ity of life. It is also agreed that supervised exercise is superior to unsupervised exercise
in this regard. In this chapter we examine attempts to evaluate the roles of exercise and
angioplasty in the treatment of IC.
Study references
Perkins, M.T., Collin, J., Creasy, T.S., Fletcher, E.W., and Morris, P.J. (1996) Exercise training versus
angioplasty for stable claudication: long and medium term results of a prospective, randomised trial.
European Journal of Vascular and Endovascular Surgery, 11, 409–13.
Whyman, M.R., Fowkes, F.G., Kerracher, E.M., et al. (1997) Is intermittent claudication improved by
percutaneous transluminal angioplasty? A randomized controlled trial. Journal of Vascular Surgery,
26, 551–7.
Greenhalgh, R.M., Belch, J.J., Brown, L.C., et al. (2008) The adjuvant benefit of angioplasty in patients
with mild to moderate intermittent claudication (MIMIC) managed by supervised exercise, smoking
cessation advice and best medical therapy: results from two randomised trials for stenotic femoro-
popliteal and aortoiliac arterial disease. European Journal of Vascular and Endovascular Surgery,
36, 680–8.
Mazari, F.A.K., Khan, J.A., Carradice, D., et al. (2012) Randomized clinical trial of percutaneous translu-
minal angioplasty, supervised exercise and combined treatment for intermittent claudication due to
femoropopliteal arterial disease. British Journal of Surgery, 99, 39–48.
Study designs
Oxford Edinburgh MIMIC Mazari et al.
Level of evidence 2b 2b 2b 2b
Randomization PTA vs supervised PTA + medical PTA + supervised PTA vs supervised
exercise therapy vs medical exercise + BMT exercise vs PTA +
therapy alone vs supervised supervised exercise
exercise + BMT
Patients 56 62 127 178
Centres 1 1 9 1
Interventions PTA of iliac or SFA PTA of iliac or SFA PTA of iliac or PTA of femoro-
SFA popliteal segment
30min exercise 30min exercise 30–60min
class, 2/week for 6 class, ≥1/week for exercises, 3/week
months 6 months for 12 weeks
Inclusion criteria Stable unilateral IC + iliac/femoral arterial lesions suitable for PTA
Exclusion criteria Not stated Previous PTA/ Symptoms too PTA/surgery to
surgery to mild for PTA symptomatic leg
symptomatic leg within 6 months
Recent MI Critical ischaemia Critical ischaemia
Inability to Inability to exercise Inability to
complete treadmill due to other complete treadmill
test disease test
Primary outcome Not clarified Ability to walk Absolute walking Maximum walking
10min on treadmill distance at 24 distance and QoL
pain free months at 12 months
Secondary ABPI, fall in ankle ABPI, maximum ABPI, walking/ ABPI, maximum
outcomes pressure after walking/ claudication walking/
exercise, maximum claudication distances at 6 and claudication
walking/ distances, QoL, 12 months, QoL distances, QoL,
claudication extent of occlusive re-stenosis/
distances disease on duplex intervention rates
Follow-up 5 years 24 months 24 months 12 months
Results
Oxford
◆ In the exercise group the claudication distance was significantly greater than pre-
treatment levels at all time intervals up to 15 months post-treatment. The PTA patients
showed no significant improvement in symptoms at any of these time intervals.
◆ Subdivision of the exercise patients into those with superficial femoral artery (SFA)
and iliac artery disease revealed that those with SFA disease experienced the greatest
increase in walking distance and claudication distance at 15 months.
◆ PTA patients in both SFA and iliac artery groups showed significant increases in
ABPI at various time points up to 15 months. ABPI in the exercise group was not
significantly increased.
◆ At a mean follow-up of 70 months there were no differences between the PTA and
exercise groups with regard to any of the endpoints.
Edinburgh
◆ Approximately 600 consecutive patients were screened to identify 62 suitable for
randomization.
◆ At 6 months the PTA group experienced improved walking, increased ankle brachial
pressure index (ABPI), and reduced pain measured in a QoL score.
◆ At 2 years follow-up there were no significant differences between the PTA and
control patients with regard to patient-reported maximum walking distance,
treadmill onset of claudication or walking distance, ABPI, or QoL.
◆ Both groups of patients reported substantial increases in walking distances at
2 years.
◆ Duplex assessment at 2 years showed a greater number of arterial occlusions and
longer and more complex stenoses in the control group than in the PTA group.
MIMIC
◆ 127 patients in total were recruited from a screened clinic population of 1401. The
intention was to recruit 170 with aorto-iliac (AI) disease and the same number with
femoro-popliteal (FP) disease.
◆ Femoro-popliteal study
● PTA was attempted in 44 of 48 patients randomized. The procedure ‘failed’ in 11.
Of the remaining 33, five had PTA of an AI lesion rather than the target FP lesion
and two had AI stenting in combination with femoro-popliteal PTA.
● Four of 45 control patients went on to have PTA of the target FP lesion during
follow-up.
● Both PTA and control groups attended a similar proportion of the available
exercise sessions—approximately 60% in each case.
● The FP study showed significant evidence of benefit of PTA for all outcome
measures except QoL. Although gains in walking distance were statistically
significant, they were clinically modest and may not have been sufficient to
change the patient’s perception of their disability.
◆ Aorto-iliac study
● Nineteen patients were randomized to PTA and 17 received the treatment.
● Four of 15 patients in the control group eventually had PTA during the study.
● Compliance with exercise therapy was around 50% for both groups.
● Despite the small numbers there was statistically significant evidence of benefit in
the PTA group for all outcome measures.
◆ In both AI and FP studies, exercise therapy as experienced by the control groups
appeared almost ineffective.
Mazari et al.
◆ 178 (15.4%) were recruited out of the 1157 patients assessed over a 6-year interval.
◆ All groups showed significant improvements in all clinical indicators at all time
points up to and including 1 year, with the exception of post-exercise ABPI in the
supervised exercise programme (SEP) only group
◆ Clinical outcomes at 12 months assessed by the International Society for
Cardiovascular Surgery (ISCVS) scores showed that 71% of PTA patients, 70% of SEP
patients, and 80% of PTA + SEP patients had improved.
◆ By the same criteria 12% of PTA patients and 17% of SEP patients had deteriorated.
No patient in the PTA + SEP group deteriorated.
◆ Comparisons between groups showed that PTA and PTA + SEP groups had
significantly higher post-exercise median ABPI than SEP-only patients. No other
outcome measures demonstrated statistically significant differences at 12 months.
◆ Duplex scanning of the PTA and PTA + SEP patients at 12 months revealed
re-stenosis rates of approximately 70% in both groups. 15% of patients in the PTA
group had re-intervention while 10% of SEP-only patients had an intervention.
No patients in the PTA + SEP group received re-intervention.
Conclusions
The disparate results obtained from these four trials make it difficult to reach definite
conclusions.
The Oxford study concluded that exercise training would benefit many patients and
spare them the small risks associated with PTA. This is supported by the near-identical
outcomes at 6 years follow-up.
The Edinburgh study observed that the differences at 2 years in their groups were insuf-
ficient to guide therapy and suggested that much larger trials would be required to eluci-
date which subgroups of patients might benefit from PTA.
The brief conclusion of the MIMIC study was ‘these findings lend weight to the value of
angioplasty as a treatment for intermittent claudication for up to 24 months’.
Mazari et al. conclude that, as there were no significant differences between the treat-
ments at one year, supervised exercise was an effective, safe, and cheap alternative to PTA.
Critique
Interpretation of the results from the Oxford and Edinburgh studies is difficult. The num-
bers are small, which may explain some of the inconsistent findings. Oxford suggest that
PTA produces no benefits, while Edinburgh showed a clear advantage of PTA over medi-
cal care (including self-directed exercise) at 6 months. Both PTA and medical care groups
showed large improvements at 2 years. The supervised exercise patients from Oxford
show significant improvements compared with baseline and with PTA patients up to 15
months after starting treatment. These improvements are essentially lost by 6 years, with
all patients roughly back where they started. The Oxford investigators discovered that
very few of the original SEP patients continued to exercise following cessation of the first
phase of the trial when the exercise classes stopped. They attribute the long-term loss of
benefit to this failure to maintain exercise.
The effects of the various interventions on ABPI are also confusing. In the Oxford study
the PTA group had an immediate rise in ABPI following treatment which was maintained
for 15 months, although this was not matched by an improvement in symptoms. At 6
years ABPI had returned to pre-treatment levels. In the SEP group the improvement in
symptoms over the first 15 months was not matched by any apparent change in ABPI.
This is easier to understand in that many of the benefits of exercise in claudication are
thought to be related to alterations in muscle structure and function rather than predomi-
nantly to augmentation of the collateral blood supply.
In some incidental ways the studies were in agreement and provided useful information.
Firstly, few patients are suitable for this type of study, representing just 10% of all patients
presenting to secondary care with intermittent claudication. The need to have patients
with isolated disease suitable for angioplasty disqualified the vast majority from partici-
pation in the studies, and the conclusion is that most patients have diffuse or multi-focal
disease that is not readily amenable to invasive treatment. Secondly, the amputation rate
was low in both studies, indicating the relatively benign nature of IC in terms of limb-
threatening ischaemia. Therefore pre-emptive revascularization is not necessary in these
patients.
The ambitious aims of the MIMIC study to clarify the uncertainty that followed publi-
cation of the Oxford and Edinburgh studies were not clearly fulfilled, as the investigators
encountered the same problems with recruitment that had bedevilled the earlier studies.
This time fewer than 10% of assessed patients were randomized, so that the numbers
entered and available for eventual analysis were small. Once again this places the appli-
cability of any results to the general peripheral artery disease (PAD) population in some
doubt.
The results of the MIMIC trial are the most bewildering of all of these studies. The
apparently clear benefit of aorto-iliac PTA agrees with other studies and supports the
generally held view in the vascular community. However, the observation that supervised
exercise therapy was almost completely ineffective is at odds with the other studies. The
modest benefits of PTA in the femoro-popliteal group (only just reaching statistical sig-
nificance for the primary outcome measure) may have been exaggerated by the fact that
almost a quarter of the subjects in this group had PTA of the aorto-iliac segment.
Mazari’s single-centre study from Hull succeeded where the multi-centre MIMIC study
failed in that they achieved their target recruitment, albeit over a 6-year period. Their
strike rate in recruiting 15.4% of 1157 is superior in some regards.
If broad conclusions can be reached from the current evidence, these are that PTA for
aorto-iliac disease is generally accepted as effective while supervised exercise is effective
for femoro-popliteal disease and is not clearly inferior to PTA.
7.9 Endovenous therapy for varicose veins

Details of studies
The use of minimal access endovenous treatments for varicose veins by cannulation and
ablation or sclerotherapy has increased in popularity over the past 10 years. There is no
consensus on the benefits of different endovenous treatments over each other or over
conventional surgery, but five RCTs (Darwood et al. 2008; EVOLVeS 2005; Rasmussen
et al. 2007; Rautio et al. 2002; Subramonia et al. 2010) have produced ten publications
comparing these techniques with conventional surgery. Alhough it may seem premature
to regard these publications as ‘landmark’ papers, it is not entirely improbable that they
will eventually come to be so regarded. The topic deserves mention in any event.
Study references
Main studies
Darwood, R.J., Theivacumar, N., Dellagrammaticas, D., Mavor, A.I., and Gough, M.J. (2008)
Randomised clinical trial comparing endovenous laser ablation with surgery for the treatment of
primary great saphenous varicose veins. British Journal of Surgery, 95, 294–301.
Lurie, F., Creton, D., Eklof, B., et al. (2005) Prospective randomised study of endovenous radiofrequency
obliteration (closure) versus ligation and vein stripping (EVOLVeS): two-year follow-up. European
Journal of Vascular and Endovascular Surgery, 29, 67–73.
Lurie, F., Creton, D., Eklof, B., et al. (2003) Prospective randomised study of endovenous radiofrequency
obliteration (closure procedure) versus ligation and stripping in a selected patient population
(EVOLVeS study). Journal of Vascular Surgery, 38, 207–14.
Rasmussen, L.H., Bjoern, L., Lawaetz, B., Blemings, A., and Eklof, B. (2010) Randomised clinical trial
comparing endovenous laser ablation with stripping of the great saphenous vein: clinical outcome
and recurrence after 2 years. European Journal of Vascular and Endovascular Surgery, 39, 630–5.
Rasmussen, L.H., Bjoern, L., Lawaetz, M., Blemings, A., Lawaetz, B., and Eklof, B. (2007) Randomized
trial comparing endovenous laser ablation of the great saphenous vein with high ligation and
stripping in patients with varicose veins: short-term results. Journal of Vascular Surgery, 46, 308–15.
Rasmussen, L.H., Lawaetz, M., Bjoern, L., Lawaetz, B., Blemings, A., and Eklof, B. (2009) Medium-term
follow-up of a randomised trial comparing laser ablation with stripping of the great saphenous vein.
Recurrence rate and pattern after two years. Phlebology, 24, 231.
Perälä, J., Rautio, T., Biancari, F., et al. (2005) Radiofrequency endovenous obliteration versus stripping
of the long saphenous vein in the management of primary varicose veins: three-year outcome of a
randomized study. Annals of Vascular Surgery, 19, 669–72.
Rautio, T., Ohinmaa, A., Perälä, J., et al. (2002) Endovenous obliteration versus conventional stripping
operation in the treatment of primary varicose veins: a randomized controlled trial with comparison
of the costs. Journal of Vascular Surgery, 35, 958–65.
Balakrishnan, A., Mylankal, K., Nalachandran, S., Subramonia, S., and Lees, T. (2008) A randomized
controlled trial of radiofrequency ablation and conventional surgery for primary long saphenous
varicose veins. Phlebology, 23, 198.
Subramonia, S. and Lees, T. (2010) Randomized clinical trial of radiofrequency ablation or conventional
high ligation and stripping for great saphenous varicose veins. British Journal of Surgery, 97, 328–36.
Almeida, J.I., Kaufman, J., Göckeritz, O., et al. (2009) Radiofrequency endovenous ClosureFAST® versus
laser ablation for the treatment of great saphenous reflux: a multicenter, single-blinded, randomized
study (RECOVERY study). Journal of Vascular and Interventional Radiology, 20, 752–9.
ENDOVENOUS THERAPY FOR VARICOSE VEINS 291
Cochrane review
Nesbitt, C., Eifell, R., Coyne, P., Badri, H., Bhattacharya, V., and Stansby, G. (2011) Endovenous ablation
(radiofrequency and laser) and foam sclerotherapy versus conventional surgery for great saphenous
vein varices. Cochrane Database of Systematic Reviews, Issue 10, CD005624.
Randomized trials
Study design
◆ All were RCTs. Darwood et al. (2008), Rautio et al. (2002), and Subramonia et al.
(2008) were single-centre RCTs, Rasmussen et al. (2007) was a two-centre RCT, and
EVOLVeS was a multi-centre study.
Results
Darwood et al.
◆ 95 patients were randomized to endovenous laser therapy (EVLT) or surgery.
◆ There was no neovascularization in the EVLT group compared with 8% in the

surgery group.
◆ The great saphenous vein (GSV) was occluded in 88–97% of patients. There was no
difference between EVLT with a stepwise or continuous laser catheter withdrawal
technique and surgery.
◆ Clinical improvement was similar in EVLT and surgery at 3 months, but return to
activities and work was quicker following EVLT.
Rasmussen et al.
◆ 121 patients randomized to EVLT or surgery.
◆ There was no difference in technical failure, recurrence, or recanalization at 3-, 6-, or

24-month follow-up.
◆ Major and minor complications were similar (e.g. phlebitis, wound infection,
haematoma, and paraesthesia).
◆ Time to return to work was similar (approximately 7 days).
◆ Overall cost of EVLT was slightly greater than surgery (€3084.50 versus
€3396.40).
Lurie et al.
◆ 79 patients were randomized to radiofrequency ablation (RFA) or surgery.
◆ General anaesthesia was used in 27% of the RFA group and 53% of the surgery
group.
◆ Occlusion rates (85–90% versus 100%), recurrence rates (14% versus 21%), and
complications were similar in RFA and surgery after 2 years follow-up.
◆ Return to work was faster after RFA than after surgery (4.7 versus 12.4 days).
◆ The observed early improvement in QoL scores persisted after 2 years.
Rautio et al.
◆ 28 patients were randomized to RFA or surgery.
◆ There was similar efficacy between RFA and surgery in treating GSV reflux
(33% recurrence post-RFA versus 21% post-surgery at 3 years).
◆ There was a faster return to activity and work in the RFA group.
◆ Immediate post-operative pain scores were better in the RFA group but complication
rates were similar.
Subramonia et al.
◆ 88 patients were randomized to RFA or surgery.
◆ 100% ablation of the GSV was achieved in the RFA group versus 81% complete
stripping in the surgery group.
◆ RFA took longer to perform than surgery (76 versus 48 minutes).
◆ There were lower pain scores and a faster return to activities in the RFA group.
Meta-analysis
Nesbitt et al. examined the results of publications comparing endovenous treatment with
surgery in their Cochrane publication. No study of ultrasound-guided foam sclerotherapy
(UGFS) met the inclusion criteria for that review.
Study design
◆ Analysis was performed in accordance with the Cochrane Peripheral Vascular
Diseases Group guidelines.
◆ Subgroup analyses were performed.
◆ Data heterogeneity was noted and explored using the I2 index.
◆ Odds ratio was used as a measure of effect for each dichotomous outcome.
Outcomes measures
Primary outcomes
◆ Recurrence or recanalization determined by ultrasound imaging at set time intervals
post-procedure.
◆ Ultrasound evidence of neovascularization.
◆ Need for re-intervention due to failure of a procedure (technical failure).
◆ Patient satisfaction determined by QoL scores pre- and post-procedure.
◆ Post-operative complications
Secondary outcomes
◆ Type of anaesthetic required for the procedures.
◆ Length of the procedures.

◆ Hospital stay.
◆ Operative costs.
ENDOVENOUS THERAPY FOR VARICOSE VEINS 293
Results
◆ Owing to differences in reporting between authors, not all outcomes could be
subjected to meta-analysis.
◆ There were no differences in recanalization or technical failure when comparing
EVLT with surgery.
◆ There were no differences in recurrence, recanalization, neovascularization, or
technical failure when comparing RFA with surgery
Conclusions
◆ Early recanalization and recurrence are similar in endovenous and conventional
techniques.
◆ Neovascularization can occur following both techniques, but is less common after
endovenous procedures
◆ Return to work after RFA may be quicker than after surgery.
◆ The risks of complications appear to be greater following conventional surgery.
◆ Both endovenous and conventional surgery can be performed as day-cases, but
endovenous procedures can be performed in an outpatient setting.
Critique
Conventional great saphenous vein surgery consisting of high ligation, stripping, and
multiple avulsions is the gold standard for GSV reflux causing symptomatic varices.
Endovenous surgery is potentially an attractive alternative to conventional techniques
because it may reduce discomfort, avoid the need for general anaesthetic, reduce the
complications of open surgery, and eliminate neovascularization as a potential source of
varicose vein recurrence. These putative advantages must be balanced against the net cost
of introducing a new technique that has not yet been demonstrated to improve treatment
efficacy. The major drivers for the introduction of endovenous techniques are patient pref-
erence and the ability to perform them in an outpatient setting using local anaesthesia.
Individually, the trials discussed above were well designed, but variation in the report-
ing of results made meta-analysis difficult. The studies were performed over a 10-year
period during which there have been major advances in endovenous techniques. This
must be taken into consideration in interpreting the earlier papers.
7.10 Superficial venous surgery in leg ulceration

Details of study
An association between leg ulceration and venous disease has long been established.
The role of venous surgery in the treatment of ulceration has been the subject of debate
ever since the association was noted. The ESCHAR study (the Effect of Surgery and
Compression on Healing And Recurrence) is an important trial in the struggle to clarify
this uncertainty.
Study references
Barwell, J.R., Davies, C.E., Deacon, J., et al. (2004) Comparison of surgery and compression with
compression alone in chronic venous ulceration (ESCHAR study): randomised controlled trial.
Lancet, 363, 1854–9.
Gohel, M.S., Barwell, J.R., Taylor, M., et al. (2007) Long term results of compression therapy alone ver-
sus compression plus surgery in chronic venous ulceration (ESCHAR study): randomised controlled
trial. British Medical Journal, 335, 83–9.
Study design
Randomization Compression bandaging alone or in combination with superficial venous
surgery
Stratification Presence or absence of deep venous reflux in addition to superficial reflux
Inclusion criteria Persistent or recently healed lower leg ulceration
ABPI >0.85
Superficial or combined superficial and deep venous reflux
Exclusion criteria Absence of complete duplex imaging
Occluded deep veins
Compression therapy not acceptable
Surgery not possible (even under LA)
Malignant ulcer
Primary endpoints 24-week ulcer healing and 12-month ulcer recurrence rates
Secondary endpoint Ulcer-free time
Follow-up 1 year initially and then a further 3 years if recurrence evident
◆ Patients with unhealed ulcers at the start of the study were treated with layered
compression bandaging until healing and then with class 2 support hosiery. Those
with healed ulcers were managed in class 2 support hosiery.
◆ All study participants were given advice sheets and instructed in the importance of
appropriate exercise and leg elevation whenever possible.
SUPERFICIAL VENOUS SURGERY IN LEG ULCERATION 295
◆ Surgery for long saphenous vein (LSV) incompetence consisted of sapheno-femoral

ligation, stripping of the LSV in the thigh, and avulsion of calf varicosities. Short
saphenous disease was managed by sapheno-popliteal disconnection and avulsion
of varicosities. Patients unfit for GA were treated by junction disconnection alone
under LA.
Results
◆ 1418 consecutive patients were assessed for inclusion in the study. 265 suitable
individuals declined to consent to the study because of reluctance to undergo surgery
of uncertain benefit. 500 patients were randomized.
◆ 300 of the randomized patients had isolated superficial venous reflux, 126 had superficial
and segmental deep reflux, and 74 had mixed superficial and total deep reflux.
◆ There were no differences between the two groups with regard to accepted risk
factors for ulcer healing (age, ulcer size or duration, rheumatoid arthritis, and
popliteal vein reflux).
Compression alone Compression + surgery

24-week healing rate 65% 65%
Ulcer healing at 3 years 89% 93%
12-month ulcer recurrence 28% 12%
Recurrence rate at 4 years 56% 31%
Ulcer-free time at 3 years 85 weeks 100 weeks
◆ Analysis of the subgroups according to deep venous reflux status showed that all
groups had lower recurrence rates with surgery, but this finding was not statistically
significant for those with total deep reflux.
Conclusions
The addition of superficial venous surgery to conservative management of venous ulcer-
ation does not improve ulcer healing rates at 24 weeks. In patients with healed chronic
venous ulceration and isolated superficial reflux, or superficial reflux with segmental deep
venous reflux, surgery reduces ulcer recurrence rates at 12 months by about 20%. If super-
ficial venous reflux is accompanied by total deep venous reflux, the addition of surgery to
conservative management of venous insufficiency does not appear to significantly reduce
ulcer recurrence rates. The addition of superficial venous surgery to treatment with com-
pression in those with healed leg ulcers increases ulcer-free time.
Critique
This was a pragmatic study. It was carried out in an established outpatient setting for
the management of venous ulcers. Consecutive patients were considered for enrolment
and therefore the results should reflect real practice rather than an artificially selected
subgroup of patients. Appropriately trained individuals, both consultants and trainees,

performed the surgery, a further reflection of normal practice.
The study analysis was carried out on an intention-to-treat basis rather than per-pro-
tocol. There was a 24% default from surgery among those who had previously agreed to
randomization, despite appropriate advice and counselling. Therefore it might be con-
cluded that the results underestimate the benefits from surgery in this group of patients.
Although the results confirm that surgery does not influence healing rates, the find-
ing that the rate of ulcer recurrence is reduced means that the amount of ulcer-free time
is increased. This brings benefits to the patient and the healthcare system in reducing
the burden and cost of ulcer disease. The mortality rate in the study was 17% at 3 years,
emphasizing that many of these patients are old and infirm. It is to be expected that relief
from ulcers and associated symptoms, and freedom from the inconvenience and discom-
fort of dressings, might bestow QoL benefits in this vulnerable group.
New methods of treating superficial venous incompetence, including endovenous laser
ablation and radiofrequency ablation of the LSV, have become commonplace since this
study was conceived. It is not unreasonable to suppose that these methods might be an
equally effective adjunct to compression in the management of leg ulcers, and indeed
might improve the poor compliance rate with surgery.
Chapter 8
Transplant surgery
8.1 Reduced exposure to calcineurin inhibitors in renal transplantation 298
8.2 Machine preservation versus static cold storage in deceased donor
transplantation 302
8.3 The mortality benefit of renal transplantation 307
8.4 Comparison of laparoscopic and mini-incision open donor nephrectomy 310
8.5 ABO-incompatible renal transplantation 313
8.6 Liver transplantation 316
298 TRANSPLANT SURGERY
8.1 Reduced exposure to calcineurin inhibitors

in renal transplantation
Details of studies
The standard immunosuppression regime for renal transplant recipients until this
trial was a ciclosporin-based regimen because of its clinical success in reducing acute
rejection episodes. However, longer-term studies demonstrated that the standard doses
of ciclosporin were associated with nephrotoxicity and cardiovascular side effects such as
hypertension and hyperlipidaemia
Tacrolimus, an alternative calcineurin inhibitor, had demonstrated improved acute
rejection rates and allograft survival at 1 year. However, most regimens used doses
targeted to trough levels between 8 and 15ng/dl.
Immunosuppressive regimens that could allow reduced exposure to calcineurin inhibi-
tors were considered to be desirable to reduce long-term nephrotoxicity and cardiovas-
cular morbidity. This strategy would depend on adequate immunosuppression being
maintained with comparable rates of acute rejection.
This trial was designed to assess the use of lower-dose immunosuppressive agents
(ciclosporin, tacrolimus, and sirolimus) in conjunction with a mycophenolate mofetil
(MMF) based regimen. This trial used low-dose maintenance levels of the study drugs
from the day of transplantation.
This study is the largest prospective RCT in renal transplantation to date and defines
the current gold standard for immunosuppression treatment.
Study references
Main study
Ekberg, H., Tedesco-Silva, H., Demirbas, A. et al. (2007) Reduced exposure to calcineurin inhibitors in
renal transplantation. New England Journal of Medicine, 357, 2562–75.
Study design
◆ Prospective multi-institutional RCT.
Randomization Standard dose ciclosporin vs low-dose ciclosporin vs low-dose tacrolimus vs
low-dose sirolimus (1:1:1:1)
Number of patients 1589 randomized and transplanted
Centres 83 sites in 15 countries
Stratification Local participating centre
Presence of donor with expanded criteria
Inclusion criteria Single-organ renal transplant recipient
Live or deceased donor
18–75 years
REDUCED EXPOSURE TO CALCINEURIN INHIBITORS 299
Exclusion criteria Second renal transplant if first graft lost to rejection at <1 year
Need for treatment with azathioprine, methotrexate, cyclophosphamide,
antilymphocyte antibodies, basiliximab, or any investigational drug
Previous treatment with daclizumab or basiliximab
Current or historic panel-reactive antibody >20%
Positive cross-match
Cold ischaemic time >30 hours
Donor after circulatory death
Disorder which may hinder absorption of oral medication
History of cancer
Evidence of active liver disease
Severe anaemia, leucopenia, thrombocytopenia
Follow-up 12 months
◆ Standard-dose ciclosporin group: ciclosporin (target trough level 150–300ng/ml for

first 3 months, then 100–200ng/ml), MMF, and corticosteroids.
◆ Low-dose ciclosporin group: daclizumab (during first two months), ciclosporin
(target trough level 50–100ng/ml), MMF, and corticosteroids
◆ Low-dose tacrolimus group: daclizumab (during first two months), tacrolimus (target
trough level 3–7ng/ml), MMF, and corticosteroids.
◆ Low-dose sirolimus group: daclizumab (during first two months), sirolimus (target
trough level 4–8 ng/ml), MMF, and corticosteroids.
Outcome measures
Primary endpoint
◆ Estimated glomerular filtration rate (eGFR) (Cockcroft–Gault formula) 12 months
after transplantation.
Secondary endpoints
◆ Renal function during the course of the study.
◆ Measured GFR at 12 months.
◆ Acute rejection.
◆ Overall survival and allograft survival at 6 and 12 months.
◆ Time to first episode of rejection.
◆ Frequency of treatment failure during first 12 months (defined as use of additional
immunosuppressive medication, discontinuation of study medication for >14
consecutive days or >30 cumulative days, allograft loss or death).
Results
◆ Mean calculated GFR was higher in the tacrolimus group than in the other three groups.
◆ The rate of biopsy-proven acute rejection in the tacrolimus group was approximately
half that observed in the other three groups.
◆ Allograft survival was significantly higher in the tacrolimus group than in the
standard-dose ciclosporin and sirolimus groups.
◆ Treatment failure was lowest in the tacrolimus group.
◆ The incidence of delayed allograft function was significantly lower in the low-dose
sirolimus group.
Standard Low-dose Low-dose Low-dose p-value

ciclosporin ciclosporin tacrolimus sirolimus
(n = 390) (n = 399) (n = 401) (n = 399)
Primary outcome
eGFR at 12 months 57.1 ± 25.1 59.4 ± 25.1 65.4 ± 27.0 56.7 ± 26.9 <0.001
Secondary outcomes
GFR at 12 months 63.5 ± 25.4 65.3 ± 26.6 69.6 ± 27.9 64.4 ± 28.5 0.04
Biopsy proven acute 24/30.1% 22/27% 11/15% 35/40% <0.001
rejection (6/12 months)
Overall survival 96.5% 98.2% 97.2% 96.8% 0.53
Allograft survival 91.9% 94.3% 96.4% 91.7% 0.02
Treatment failure in first 22.8% 20.1% 12.2% 35.8% <0.001
12 months
Delayed graft function 33.6% 32.4% 35.7% 21.1% 0.004
Adverse events
New-onset diabetes 6.4% 4.7% 10.6% 7.8% 0.02
Opportunistic infections 33.0% 28.1% 26.3% 26.6% 0.03
Malignancy 0.5% 0.7% 1.0% 2.1%
Early wound problems 10.8% 11.0% 9.4% 16.6% 0.006
Lymphoceles (requiring 7.0% 6.8% 4.0% 15.8% <0.001
intervention)
CMV infection 15.3% 11.5% 10.2% 6.5% 0.003
CMV, cytomegalovirus.
◆ Serious adverse events were reported by 52% of patients in the low-dose sirolimus
group compared with 43.3–44.3% in the other study groups.
◆ The most common adverse events for all groups were disorders of blood/lymphatic
systems, gastrointestinal disturbances, infection/infestation, and metabolic disorders.
◆ Patients in the low-dose sirolimus group had significantly higher rates of lymphocele
formation and delayed wound-healing.
REDUCED EXPOSURE TO CALCINEURIN INHIBITORS 301
Conclusions
A regimen of low-dose tacrolimus in combination with daclizumab, MMF, and corti-
costeroids results in adequate immunosuppression with better renal function, fewer
episodes of acute rejection, and improved allograft survival compared with standard-dose
or low-dose ciclosporin or low-dose sirolimus regimens.
Rates of new-onset diabetes are highest in patients treated with tacrolimus.
A low-dose sirolimus regime is associated with a significantly increased risk of delayed
wound-healing and lymphocele formation.
Critique
This trial was designed to compare the use of lower-dose immunosuppressive agents
(ciclosporin, tacrolimus, and sirolimus) in conjunction with a MMF-based regimen.
The study demonstrated a clear benefit with respect to allograft function at 1 year and
acute rejection episodes in the low-dose tacrolimus arm compared with the alternative
regimes studied.
Given the superior rates of biopsy-proven acute rejection in the tacrolimus group, it
may be feasible to consider reduction of the corticosteroid dose to minimize the risk of
development of new-onset diabetes after transplantation (NODAT) or further reduce the
exposure to tacrolimus whilst maintaining efficacy.
This paper established the low-dose tacrolimus regimen as the gold standard and the
basis for comparison for future regulatory studies.
8.2 Machine preservation versus static cold storage in

deceased donor transplantation
Details of studies
The two main forms of storage for organ preservation prior to transplantation of renal
allografts are static cold storage and hypothermic machine perfusion. In static cold stor-
age, after initial perfusion the kidney is flushed with cold preservation solution and then
placed on ice. In hypothermic machine perfusion, after initial perfusion with cold pres-
ervation solution the kidney is connected to a perfusion device. It is then perfused with
preservation solution continuously at hypothermic temperatures until transplantation.
Retrospective studies suggested that machine perfusion could improve short-term out-
come, with lower rates of delayed graft function (DGF) after transplantation of kidneys
from deceased donors.
Study references
Main studies
Moers, C., Smits, J.M., Maathuis, M.H., et al. (2009) Machine perfusion or cold storage in deceased
donor kidney transplantation. New England Journal of Medicine, 360, 7–19.
Watson, C.J.E., Wells, A.C., Roberts, R.J., et al. (2010) Cold machine perfusion versus static cold storage
of kidneys donated after cardiac death: a UK multicentre randomised controlled trial. American
Journal of Transplantation, 10, 1991–9.
Related references
Wright, J.P., Chilcott, J.B., Holmes, M.W., and Brewer, N. (2003) Pulsatile machine perfusion vs. cold
storage of kidneys for transplantation: a rapid and systematic review. Clinical Transplantation, 17,
293–307.
Wright, J., Chilcott, J., Holmes, M., and Brewer, N. (2003) The clinical and cost effectiveness of pulsatile
machine perfusion versus cold storage of kidneys for transplantation retrieved from heart beating
and non-heart beating donors. Health Technology Assessment, 7, 1–94.
Study design
Moers et al.
◆ European multicentre RCT
Randomization One kidney from each donor to cold storage, the other to machine perfusion
Number of patients 336 consecutive donors
Number of centres 59
Inclusion criteria Donor >16 years of age
Donation after brain death (DBD)
Maastricht category III/IV for donation after cardiac death (DCD)
Both kidneys transplanted into different recipients
MACHINE PRESERVATION VERSUS STATIC COLD STORAGE 303
Exclusion criteria Multi-organ transplants

Death of patient <1 week after transplant
Primary outcomes Delayed graft function
Secondary outcomes Duration of DGF, primary non-function, AUC of serum creatinine days 1–14,
creatinine clearance day 14, biopsy-proven acute rejection, calcineurin
inhibitor toxicity, hospital stay, 1-year graft, and patient survival
Follow-up 1 year
AUC, area under curve.
◆ Exclusions after randomization were due to cancelled procedures, use of a kidney

as part of multi-organ transplants, and rejection at the transplant centre. Seven had
technical failure of machine perfusion.
◆ Twenty-five kidneys were switched post-randomization because of anatomical
differences that precluded use of machine perfusion
◆ The cold ischaemic times in each group were 15.0 hours, but no information was
available for the DCD subgroup on warm ischaemic times or anastomosis time.
◆ An amendment to the protocol was made near the end of enrolment to extend the
inclusion of donors after circulatory death until a total of 82 were enrolled to allow
meaningful subgroup analysis. These extra 40 donors were only analysed in the
subgroup analysis of DBD versus DCD.
◆ The European study randomized a majority of brainstem-dead donors rather than
donors after circulatory death, whereas the UK study recruited recipients of DCD
kidneys.
Watson et al.
◆ UK multicentre RCT.
Randomization Cold storage or machine perfusion (+ which kidney to be transplanted first)
Number of patients 90 donors
Number of centres Five
Inclusion criteria Recipients of DCD kidneys >18 years
Negative cross-match
First transplant
Exclusion criteria DBD
Previous transplant
Primary outcomes Delayed graft function
Secondary outcomes Primary non-function, creatinine reduction ratio day 2/day 5, biopsy-proven
acute rejection, graft function (GFR, serum creatinine), patient and graft
survival
Follow-up 1 year
◆ The study had a sequential design whereby analysis of the primary endpoint was
undertaken after the recruitment of 60 patients and then after every 20 patients.
◆ Where organ retrieval occurred away from the base transplant hospital, the kidney
randomized to machine perfusion could first undergo cold storage and commence
machine perfusion on return.
◆ In contrast with the European study, the UK study recruited recipients of DCD
kidneys.
Results
Moers et al.
Machine perfusion Cold storage p-value

(n = 336) (n = 336)
Primary endpoint
Delayed graft function 20.8% 26.5% 0.05
Secondary endpoints
Functional DGF 22.9% 30.1% 0.03
Primary non function 2.1% 4.8% 0.08
Duration of DGF (days) 10 13 0.04
Creatinine clearance (day 14) 42 40 0.25
CNI toxicity 6.3% 5.7% 0.86
Acute rejection 13.1% 13.7% 0.91
Hospital stay (days) 19 18 0.78
◆ There were no significant differences in donor or recipient characteristics between

the machine perfusion and cold storage groups
◆ In the subgroup analysis there was no significant difference in the effect of machine
perfusion versus cold storage on DGF in standard criteria donors or extended criteria
donors (p = 0.75) or in DBD versus DCD (p = 0.42), even after inclusion of the
extended dataset (p = 0.26)
◆ No differences in patient survival at 1 year were observed between the two study
groups.
◆ Cox regression analysis demonstrated that machine perfusion significantly reduced
the risk of graft failure at 1 year (HR 0.52, p = 0.03). However, post hoc analysis
addition of DGF as a covariate to the Cox model indicated that recipients with DGF
had a significantly increased risk of graft failure (HR 1.67, p = 0.001). Once applied,
the hazard ratio for machine perfusion became non-significant.
MACHINE PRESERVATION VERSUS STATIC COLD STORAGE 305
Watson et al.
Machine perfusion Cold storage p-value
(n = 45) (n = 45)
Primary endpoint
Delayed graft function 57.8% 55.6% 0.99
Secondary endpoints
Primary non-function 1 0 –
Creatinine reduction ratio day 2 26/34 29/34 0.48
Creatinine reduction ration day 5 33/45 31/45 0.6
Acute rejection (3 months) 7% 22% 0.06
Patient survival 93.3% 100% 0.08
Graft survival 93.3% 97.8% 0.3
Graft function (GFR [ml/min]) 46.6 46.2 0.64
◆ The mean duration of machine perfusion in this study group was 10.1 hours. This
represented a mean of 69% (29–97%) of the cold ischaemic period.
Conclusions
Moers et al.
Machine perfusion significantly reduces the risk of DGF in deceased donor kidney trans-
plantation and the duration of DGF when it occurred. The reduction in DGF associated
with machine perfusion is thought to be a significant contributory factor to the improve-
ment in graft survival.
Watson et al.
No significant difference was observed in DGF between the machine perfusion and cold
storage groups
Critique
One of the difficulties in assessing the impact of an intervention to improve outcome in
renal transplantation is that there is a spectrum from immediate to slow to DGF requir-
ing dialysis. It is known that DGF negatively impacts on longer-term graft survival. This
is an area of significant interest, as increasing use of extended criteria donors and DCD
allografts carries an increased risk of DGF. Moers et al. demonstrate a benefit in reduced
DGF in deceased donor kidneys that undergo machine perfusion.
The impact of machine perfusion in DCD allografts has more conflicting evidence. The
multi-centre UK study was a sequential design with adequate power and demonstrated
no benefit of machine perfusion over static cold storage. One methodological factor
may have been that kidneys were not immediately machine perfused but transferred
onto machine perfusion only after initial cold static perfusion and transfer from the
donor centre. The trial was stopped early because of futility, and whilst cessation on the
basis of futility rules can reduce the time and cost of trials, it makes interpretation of
a negative study more difficult as it reduces precision. Differences in the nature of the
ischaemic injury accumulated by DBD and DCD kidneys may explain why machine per-
fusion is superior to cold storage for DBD kidneys. This does not explain the differences
observed between the two studies for DCD donors. It is difficult to make a direct com-
parison of the DCD populations between the two studies. The European study reported
only cold ischaemic times and not the more detailed information of warm ischaemia and
anastomosis times, as in the UK study.
Cost effectiveness of machine cold perfusion also remains controversial. No robust
models exist and the variability of consumable component price frequently makes
estimates rapidly out of date.
Normothermic perfusion methods based on extra-corporeal membrane oxygenation
(ECMO) are increasingly establishing a role in organ preservation and may overtake the
methods discussed here in future clinical practice.
MORTALITY BENEFIT OF RENAL TRANSPLANTATION 307
8.3 The mortality benefit of renal transplantation

Details of studies
The extent to which renal transplantation improves survival, whilst suspected, could
never ethically be tested in randomized trials with its benefits discussed in terms of qual-
ity of life rather than improved survival. The potential for selection bias in studies is
significant, as transplant recipients are taken from a highly selected subgroup of patients
on dialysis who are deemed suitable candidates for transplantation. This study was
designed to compare survival of patients undergoing transplantation with survival of
those awaiting transplantation.
Study references
Main study
Wolfe, R.A., Ashby, V.B., Milford, E.L., et al. (1999) Comparison of mortality in all patients on dialysis,
patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. New
England Journal of Medicine, 341, 1725–30.
Related references
Ojo, A.O., Hanson, J.A., Meier-Kriesche, H., et al. (2001) Survival in recipients of marginal cadaveric
donor kidneys compared with other recipients and wait-listed transplant candidates. Journal of the
American Society of Nephrology, 12, 589–97.
Study design
◆ Prospective national registry observational study
Setting US Renal Data System/Transplant Scientific Registry (UNOS)
Inclusion criteria End-stage renal disease (ESRD) patients starting dialysis (1991–1996)
ESRD patients on waiting list for first cadaveric renal transplant
Exclusion criteria Age >70 years on starting dialysis
No cause for ESRD reported
Pre-emptive transplant recipients
Follow-up 7 years
Outcome measures
Primary endpoints
◆ Survival.
Secondary endpoints
◆ Standardized mortality ratios (SMRs).
◆ Time at which risk of death/likelihood of survival equals reference group (all ESRD
patients or wait-listed ESRD patients).
◆ Projected years of life.
Results
All ESRD patients Wait-listed ESRD Cadaveric transplant

(n = 228,552) patients recipients
(n = 46,164) (n = 23,275)
Annual death rate/ 16.1 6.3 3.8
100 patient-years
Deaths 84,713 4353 2436
Adjusted SMR (compared 1.0 0.51 0.31
with all ESRD patients)
◆ All subgroups of patients on the waiting list (male, female, White, Black, Asian,
Native American, diabetes-related ESRD, other-cause ESRD) had a significantly
lower SMR compared with their subgroup of all patients on dialysis.
◆ Risk of death 2 weeks after transplantation is 2.8 times higher than in patients on the
waiting list and remains higher until 106 days after surgery.
Relative risk Time at which Time at which Projected years Projected

of death risk of death survival of life without years of
18 months equal to likelihood transplant life with
post-transplant reference equal to transplant
group (days) reference
group (days)
Overall 0.32 106 244 10 20
White 0.28 100 220 9 14
Black 0.52 109 305 13 23
Asian 0.43 161 673 15 19
Native 0.5 123 304 9 19
American
Male 0.34 110 255 10 19
Female 0.3 94 220 11 23
Diabetes 0.27 57 146 8 19
Other cause 0.38 137 353 12 20
Age
0–19 years 0.33 3 5 26 39
20–39 years 0.24 11 57 14 31
40–59 years 0.33 95 251 11 22
60–74 years 0.39 148 369 6 10
MORTALITY BENEFIT OF RENAL TRANSPLANTATION 309
◆ Transplantation reduced long-term risk of death in both sexes and all races.
◆ The greatest improvement in long-term survival was among patients aged 20–39
years when placed on the waiting list.
◆ Patients aged 60–74 years had improved survival after the first year with an increase
in lifespan of 4 years
Conclusions
Much of the large risk reduction for mortality is due to selection of healthier patients for
the waiting list.
Renal transplantation is associated with significant long-term survival benefits, despite
a short-term increase in mortality shortly after transplantation. These benefits appear
most marked in the diabetic population.
Critique
This is a non-randomized study and therefore can only demonstrate an association. The
study relies on an underlying assumption that the timing of transplantation after being
placed on the waiting list is unrelated to the patient’s underlying chance of survival. Even
if patients do become ill soon after going on the list, for this to be the explanation for the
difference in survival, two-thirds of the patients in that group would have to have died in
that manner. The alternative proposition by Wolfe et al. of a reduction in mortality due to
transplantation (with a rise in the peri-operative period) seems much more likely.
These retrospective results assume a standard ‘quality’ of deceased donor organ.
Increasingly, the range of acceptable organs is being broadened, with organs associated
with poorer function and shorter survival being transplanted. These changes have impor-
tant implications for the survival benefit of renal transplantation and are reviewed else-
where (Ojo et al.).
Furthermore, this study is based on US registry data. The prognosis for maintenance
dialysis patients varies widely between populations. In particular, the UK typically reports
lower mortality rates in this patient group. The benefit of renal or any solid organ trans-
plant must always be weighed against the prevalent best medical treatment in the popula-
tion in question.
8.4 Comparison of laparoscopic and mini-incision

open donor nephrectomy
Details of studies
There has been a significant increase in living kidney donation in recent years. This partly
reflects increased waiting time from the reduction of cadaveric donors combined with
recognition that long-term graft outcomes are superior. Donor nephrectomy remains a
major operation that carries both morbidity and, infrequently, mortality. Perhaps as the
basis of live donor nephrectomy is purely altruistic, it is unsurprising that all efforts that
minimize the morbidity of the procedure have been studied in balance with ensuring
optimal graft function. This study aimed to assess in an RCT the potential benefits of
laparoscopic versus mini incision open donor nephrectomy.
Study references
Main study
Kok, N.F., Lind, M.Y., Hansson, B.M., et al. (2006) Comparison of laparoscopic and mini incision open
donor nephrectomy: single blind, randomised controlled clinical trial. British Medical Journal, 333, 221.
Related references
Wolf, J.S., Jr, Merion, R.M., Leicthman, A.B., et al. (2001) Randomized controlled trial of hand-assisted
laparoscopic versus open surgical live donor nephrectomy. Transplantation, 72, 284–90.
Simforoosh, N., Basiri, A., Tabibi, A., Shakhssalim, N., Hosseini, and Moghaddam, S.M. (2005)
Comparison of laparoscopic and open donor nephrectomy: a randomized controlled trial. British
Journal of Urology International, 95, 851–5.
Øyen, O., Andersen, M., Mathisen, L., et al. (2005) Laparoscopic versus open living donor nephrec-
tomy: experiences from a prospective randomized single centre study focusing on donor safety.
Transplantation, 79, 1236–40.
Nicholson, M.L., Kaushik, M., Lewis, G.R., et al. (2010) Randomized clinical trial of laparoscopic versus
open donor nephrectomy. British Journal of Surgery, 97, 21–8.
Study design
◆ Single blind multi-centre RCT
Randomization Laparoscopic donor nephrectomy versus mini incision muscle splitting donor
nephrectomy
Setting Two university hospitals in the Netherlands
Inclusion criteria Living kidney donors
Exclusion criteria Bilateral abnormalities of renal artery (origin stenosis)
Previous renal or adrenal surgery
Radiological abnormalities necessitating a modified approach (e.g. lesion
requiring frozen section)
Inability to read Dutch
Follow-up 1 year
LAPAROSCOPIC AND MINI-INCISION OPEN DONOR NEPHRECTOMY 311
◆ Living kidney donors were screened by a nephrologist and underwent renal USS and
magnetic resonance/CT angiography to evaluate arterial and venous anatomy.
Outcome measures
Primary endpoint
◆ Physical fatigue (Multidimensional Fatigue Inventory 20 (MFI-20)).
Secondary endpoints
◆ Physical function (Short Form 36 (SF-36)), hospital stay, pain, operating time,
recipient graft function, graft survival, complications.
Results
◆ 105 participants were randomized (four cancelled/postponed for clinical/radiological
reasons, one recipient received a cadaveric kidney)
◆ One-year follow-up was 90% (n = 45) in the laparoscopic group and 88% (n = 44) in
the mini-incision open nephrectomy group
◆ There were no significant differences in donor characteristics, vascular anatomy, or
comorbidity
Laparoscopic donor Mini-incision open p-value

nephrectomy donor nephrectomy
(n = 50) (n = 50)
Hospital stay (days) 3 (1–6) 3 (2–8) 0.002
Pain
Day 1 2.7 3.5 0.04
Day 3 1.4 1.8 0.12
Day 7 0.4 1.7 0.03
Day 14 0 0.4 0.008
Operating time (min) 289.5 226 <0.001
Recipient graft function (µmol/L) 107 114 0.17
Graft survival at 1 year 48 48 1.0
Complications
Intra-operative 6 3 0.23
Post-operative 3 3 1.00
◆ Mean physical fatigue was reduced (–1.3, 95%CI –2.4 to –0.1) and physical function
was improved (6.2, 95% CI 2.0–10.3) over 1 year of follow-up following laparoscopic
nephrectomy.
Conclusions
This study provides the first long-term QoL assessment after laparoscopic versus open
live donor nephrectomy. Several QoL indicators, although not all, were better after
laparoscopic donor nephrectomy versus a mini open incision. Recovery was faster and
pain scores and post-operative morphine requirements were less after laparoscopic
donor nephrectomy.
Critique
Quality of life issues are considered of greater significance in the context of living organ
donation as the donor has no personal underlying need for surgery. The positive findings
in this study have been confirmed in subsequent follow-up studies (Øyen et al.).
Whilst this study has been the only study to attempt blinding, it is likely that this only
impacted during the hospital admission whilst wounds were covered. This effect could
not be maintained after discharge. No information was provided regarding any intra-
operative complications during the recipient operation.
In the UK, the Leicester group compared the safety and efficacy of 28 open donor
nephrectomies with 56 laparoscopic live donor nephrectomies with no mortality or
vascular thrombosis (Nicholson et al.). They showed no difference in renal function or
allograft survival at a median 74 months follow-up. A meta-analysis (of largely non-
randomized studies) has confirmed the shorter operative and warm ischaemic times
of open donor nephrectomy. It also confirmed the finding that laporoscopic donor
nephrectomy was associated with lower analgesic requirements, shorter hospital stay,
and faster return to work.
Overall, outcomes are equivalent in terms of renal function, graft survival, and compli-
cations. Early post-operative pain, hospital stay, and return to work are evidently slightly
better using minimal-access techniques. Adoption of the laparoscopic/hand-assisted
laparoscopic technique has been largely patient driven, and, whatever the objective
clinical merit, minimal-access techniques have become the gold standard for living
kidney donation.
Nanidis, T.G., Antcliffe, D., Kokkinos, C. et al. (2008) Laparoscopic versus open live donor nephrectomy
in renal transplantation: a meta-analysis. Annals of Surgery, 247, 58–70.
ABO-INCOMPATIBLE RENAL TRANSPLANTATION 313
8.5 ABO-incompatible renal transplantation

Details of studies
The increasing need for kidney transplantation has led to significant interest in ABO-
incompatible (ABOi) transplantation. ABOi transplantation could significantly increase
the number of living donor transplantations, as this is a common reason why patients in
the UK enter a paired exchange programme. The length of time on dialysis has a signifi-
cant impact on long-term survival within the transplant population. Therefore facilitating
live donor transplantation rather than awaiting a cadaveric kidney should be considered.
Study references
Main study
Genberg, H., Kumlien, G., Wennberg, L., Berg, U., and Tydén, G. (2008) ABO-incompatible kidney
transplantation using antigen-specific immunoadsorption and rituximab: a 3-year follow-up.
Transplantation, 85, 1745–54.
Study design
◆ A comparative retrospective cross-sectional study.
Randomization Non-randomized
Number of patients 45 (ABOi n = 15 versus ABOc n = 30) adult, five paediatric
Setting Karolinska University, Sweden, 2001–2005
Follow-up 3 years
◆ The protocol involved removal of A/B antibody by repeated antigen-specific

immunoadsorption. Single-dose rituximab (375mg/m2) on day 30 and oral
immunosuppression (tacrolimus–MMF–prednisolone) commenced on day 10.
Intravenous immunoglobulin (0.5g/kg) was given on day 1.
◆ Three pre-emptive immunoadsorptions were performed post-transplantation.
Outcome measures
Primary endpoints
◆ Patient survival.
◆ Graft survival.
Secondary endpoints
◆ Acute rejection, graft function (serum creatinine/GFR/changes in serum
creatinine), proteinuria, infections requiring hospital care, opportunistic
infection (cytomegalovirus (CMV), Epstein–Barr virus (EBV), BK virus, fungal),
hospitalization, B-cell measurements, analysis of cost effectiveness.
Results
◆ There were differences in age (mean 35.1 versus 42.4 years) and in percentage of
patients transplanted pre-emptively (47% versus 10%) in the ABOi versus ABO
compatible (ABOc) adult groups
Primary endpoints ABOi ABOc p-value

Patient survival 15 (100%) 29 (96.7%) ns
Graft survival 13 (86.7%) 26 (86.7%) ns
Secondary endpoints
Biopsy proven acute rejection 1 (6.7%) 4 (13.3%) ns
Antibody mediated rejection 0 1 (3.3%) ns
Cellular rejection 1 (6.7%) 3 (10%) ns
Proteinuria 2/14 (14%) 6/27 (22%) ns
Infections requiring hospital care 6/15 (40%) 19/30 (63.3%) ns
Opportunistic infection 1/0/0/0 7/1/0/0 ns
(CMV/EBV/BK/fungal)
Hospitalization (days)*
Pre-transplantation 6 2 0.001
Hospital stay after transplantation 34.6 27.4 0.036
Cost-effectiveness (additional cost) €31,948 (~9 months
on dialysis)
*Values are means.
◆ At 24 months there were no significant differences between eGFR, serum creatinine,

or change in serum creatinine for the 12/15 patients with available data.
◆ Complete B-cell depletion was demonstrated in all but two patients. The B-cell
depletion was prolonged, taking 1.5 years for the median to recover to >5cells/µL.
◆ A/B antibody titres performed at all time points were significantly reduced from
pre-operative levels. No rebound was observed in the follow-up period.
◆ 10/15 ABOi recipients received low molecular weight heparin. This group
experienced significantly more bleeding complications, including blood transfusions
and re-operations for bleeding.
Conclusions
There does not appear to be any significant difference in patient survival or graft survival
at 3 years after ABOi kidney transplantation compared with ABOc transplantation.
Paediatric ABOi transplantation is feasible, but the small numbers in this study pre-
clude any valid comparison with an ABOc cohort. All the paediatric patients had compli-
cated post-operative courses (urological complications/BK/DGF).
ABO-INCOMPATIBLE RENAL TRANSPLANTATION 315
Critique
Because of the illegality of brainstem death in Japan, protocols for ABOi transplantation
have been developed with excellent short-term results, although these invariably incor-
porate splenectomy. This study is noteworthy for the demonstration that ABOi transplan-
tation can be performed without a significantly higher incidence of immunological or
infective complications and with comparable function to ABOc transplants at 2–3 years.
These results were achieved without a requirement for routine surgical splenectomy.
The statistical analysis bears comment as the authors performed a retrospective power
analysis to determine the probability of rejecting the null hypothesis when it is actually
false. In this study the calculated power was low. The least significant number (LSN) is the
number of observations needed to reduce the variance of the estimates enough to achieve
a significant result—how many observations would make the reported difference become
significant? In this case, whilst the numbers are very small, the finding that the LSN is
considerably higher than the sample size suggests that there is no significant effect on
outcome between ABOi and ABOc.
Further studies have subsequently shown that these short-term results translate to
very acceptable medium-term results, with little suggestion of problems relating to the
increased intensity of immunosuppression. ABOi kidney transplantation is increasingly
regarded as a part of the mainstream treatment options for ESRF, although paired dona-
tion, with primarily logistical barriers, is a more attractive approach.
8.6 Liver transplantation

Details of studies
It was recognized that replacement of the liver with a homograft could potentially treat
liver diseases such as primary liver cancer, cirrhosis, and congenital atresia of the bile
ducts. The feasibility of liver transplantation was explored following the development of
anti-rejection therapy in renal transplantation and canine studies suggesting that the pro-
cedure was technically possible. The first liver transplant in humans was performed by
T.E. Starzl in Denver, Colorado, in 1963.
Study reference
Main study
Starzl, T.E., Marchioro, T.L., von Kaulla, K.N., et al. (1963) Homotransplantation of the liver in humans.
Starzl, T.E., Klintmalm, G.B., Porter, K.A., Iwatsuki, S., and Schröter, G.P. (1981) Liver transplantation
with use of cyclosporin A and prednisone. New England Journal of Medicine, 305, 266–9.
Study design
◆ An experimental study in three terminally ill patients to assess the feasibility of liver
transplantation.
◆ Two of the patients had inoperable liver metastases and the other was a 3-year old
with congenital biliary atresia, hepatosplenomegaly, jaundice, and ascites.
◆ The 3-year old died from haemorrhage on the operating table, but the other patients
lived for 7.5 and 22 days.
◆ The donors were 3, 55, and 69 years old, and all had terminal conditions
Critique
Starzl initially developed the innovative techniques used in this paper in an animal model
before using the technique in human patients. This paper heralded the start of liver trans-
plant surgery, which has saved thousands of lives in the decades following publication of
this initial series of three patients. Although the paper set out the fundamental surgical
approaches to donor and recipient operations, it took some 20 years before improvements
in preservation of the donor liver and better immunosuppression led to the procedure
becoming routine in major transplant centres across the world. Liver transplantation is
currently undertaken with low operative mortality rates and 15-year post-transplant sur-
vival approaching 60%.
Live donor liver transplant, first performed in children and then in adults, has a 5-year sur-
vival rate of 92% for children. This compares with 81% 5-year survival rate for children receiv-
ing livers from deceased donors. The risk of post-operative death for a live donor is 0.2%, not
surprisingly higher with right-lobe (0.23–0.5%) than with left-lobe donation (0.05–0.21%).
NICE (2011) Living-donor liver transplantation. <http://guidance.nice.org.uk/IPG194>
Chapter 9
Breast surgery
9.1 Simple versus radical mastectomy 318
9.2 Breast conservation versus mastectomy for breast cancer 321
9.3 Sentinel node biopsy in breast cancer 324
9.4 Axillary dissection versus no axillary dissection in women with invasive cancer
and sentinel node metastases 327
9.5 Management of localized DCIS 329
9.6 The role of oestrogen receptors in breast cancer 332
9.7 Histological grade and prognosis in breast cancer 334
9.8 The role of the HER-2/neu oncogene in breast cancer 335
9.9 Patterns of gene expression in breast cancer 336
9.10 The origins of adjuvant chemotherapy for breast cancer 337
9.11 Adjuvant anthracyclines 339
9.12 Adjuvant taxane chemotherapy 341
9.13 Tamoxifen as adjuvant therapy in breast cancer 343
9.14 The effect of tamoxifen and adjuvant chemotherapy in breast cancer 345
9.15 Aromatase inhibitors and breast cancer 347
9.16 Extended adjuvant endocrine therapy 349
9.17 Use of Herceptin in breast cancer 352
9.18 Neoadjuvant chemotherapy for operable breast cancer 355
318 BREAST SURGERY
9.1 Simple versus radical mastectomy

Details of studies
In the late nineteenth century, William Halsted at Johns Hopkins University, Baltimore,
Maryland, USA, described the operation of radical mastectomy for breast cancer. The
surgery involved removal of the whole breast, the axillary lymph nodes, and the pectoralis
major and minor muscles. Halsted was also an innovator in terms of his general manage-
ment, as he followed his patients up in clinic and was one of the first surgeons to publish
outcomes after surgery. The Halsted radical mastectomy was quickly adopted as standard
practice in the management of breast cancer until the 1940s.
In the landmark paper discussed here, Patey described a modification of the Halsted
operation. In this less radical operation, Patey preserved the pectoralis major muscle
although the pectoralis minor was still removed along with the axillary contents.
Study reference
Main study
Patey, D.H. and Dyson, W.H. (1948) The prognosis of carcinoma of the breast in relation to the type of
operation performed. British Journal of Cancer, 2, 7–13.
Related references
Halsted, W.S. (1894) The results of operations for the cure of cancer of the breast performed at the Johns
Hopkins hospital from June 1889 to January 1894. Annals of Surgery, 20, 497–555.
Patey, D.H. (1967) A review of 146 cases of carcinoma of the breast operated on between 1930 and 1943.
British Journal of Cancer, 21, 260–9.
Study design
◆ Retrospective review
Number of patients 118 (the results from 146 patients were reported in the 1967 paper)
Setting Single centre (Middlesex Hospital, England)
Inclusion criteria All patients undergoing surgery for breast cancer during the study period
Follow-up The original reported only very short term follow-up whereas the 1967
paper reported on a 20-year follow-up
◆ 49 patients had a standard Halsted radical mastectomy, 69 had a modified radical

mastectomy, 17 had a Patey mastectomy and irradiation of the axilla, and five
underwent Patey mastectomy and irradiation of the chest wall.
Outcome measures
◆ Not clearly defined. The primary aim was to compare safety and outcome in
comparison with the Halsted procedure.
◆ Outcomes in terms of death from surgery, recurrence of breast cancer, and death
from breast cancer were described.
SIMPLE VERSUS RADICAL MASTECTOMY 319
Results
◆ Four (3%) patients died during or soon after surgery.
◆ 37 (25%) patients died from breast cancer within the first 2 years after surgery,
18 (12%) died from breast cancer between years 2 and 5 post-operatively, a further
15 (10%) died from breast cancer between years 5 and 10, and five (3%) died between
years 10 and 20.
◆ 29 patients were lost to follow-up and 22 died from unrelated causes.
◆ 16 patients (11%) were alive and well at 20 years.
◆ 40 of the 49 patients who underwent a Halsted radical mastectomy were followed up,
and 25 (62.5%) died from breast cancer.
◆ 54 of the 69 patients who had the Patey mastectomy were followed up, and 36 (67%)
died of breast cancer by 20 years, i.e. after 20 years of follow-up there was only
a 4.4% difference in mortality between the Halsted and Patey mastectomies.
◆ The average survival was 3.5 years for patients with involved nodes and 7.6 years for
those with no involved nodes.
◆ A single patient in the Halsted radical mastectomy group also had bilateral
oophorectomy.
◆ No systemic chemotherapy was available or administered.
◆ Radiotherapy, which was in its infancy, was administered to some patients for
uncertain indications.
Conclusion
It is not necessary to remove the pectoralis major as part of a standard mastectomy.
Critique
These remarkable papers were outstanding in that Patey carefully recorded the short-
term mortality and long-term follow-up of this group of patients with breast cancer. He
questioned the theory of lymphatic permeation as the method of spread of breast cancer.
This theory underpinned the technique of radical mastectomy, as it was thought that lym-
phatic spread was via chest wall muscles. Patey showed that this theory did not explain
the spread of breast cancer.
The patients in this review were not randomized but were selected for surgery accord-
ing to what seemed to be the best procedure for them. This means that it is difficult to
compare the two types of mastectomy from these data. Patey did try to calculate the
likelihood that the difference between the groups was due to chance. He also noted that
more of the patients in the radical mastectomy group surprisingly had ‘good pathology
results’ compared with the Patey mastectomy patients, i.e. that there may have been bias
in patient selection.
The fact that one of the patients who did very well (in fact led a normal life into her
seventies) was an 11-year-old girl when she underwent mastectomy and axillary clearance
320 BREAST SURGERY
for ‘breast cancer’ suggests that pathological diagnoses may have included patients with
benign disease. This again makes long-term follow-up information difficult to interpret.
Four patients died from their surgery. A 3% mortality rate from mastectomy would be
unacceptable today, but was probably not unremarkable at the time.
Patey’s series reveals remarkable insights into his understanding of breast cancer which,
although accepted as conventional thinking today, at the time of publication provided a
novel insight into the pathophysiology of the condition. These include his observation
that women could still die from breast cancer up to 20 years after mastectomy ‘despite
their treatment’. He also observed that mortality was higher in patients with involved axil-
lary nodes. The modified radical mastectomy as described by Patey is still the standard
operation of mastectomy today, apart from the automatic removal of all the ipsilateral
axillary nodes.
BREAST CONSERVATION VERSUS MASTECTOMY 321
9.2 Breast conservation versus mastectomy for breast cancer

Details of studies
From 1940 until 1973 the Halsted mastectomy was the standard operation carried out for
all patients with breast cancer. Patey modified the operation in 1948 by leaving the chest
wall muscles intact, but otherwise it was the same procedure. No clinicians had consid-
ered preserving the breast during this period. In this study from the National Cancer
Institute in Milan, patients with small cancers of the breast were offered breast-conserving
surgery for the first time. The technique used in this study was described as a quadrantec-
tomy in which a large volume of breast tissue was removed with the cancer. An American
study of similar design was started 3 years later (Fisher et al.). This also examined the role
of breast-conserving surgery with or without radiotherapy to the breast.
Study references
Main study
Veronesi, U., Saccozzi, R., Del Vecchio, M., et al. (1981) Comparing radical mastectomy with quadran-
tectomy, axillary dissection, and radiotherapy in patients with small cancers of the breast. New
England Journal of Medicine, 305, 6–11.
Related references
Veronesi, U., Cascinelli, D., Mariani, L., et al. (2002) Twenty-year follow-up of a randomized study
comparing breast-conserving surgery with radical mastectomy for early breast cancer. New England
Fisher, B., Bauer M., Margolese, R., et al. (1985) Five-year results of a randomized clinical trial
comparing mastectomy and segmental mastectomy with or without radiation in the treatment of
breast cancer. New England Journal of Medicine, 312, 665–73.
Fisher, B., Anderson, S., Bryant, J., et al. (2002) Twenty-year follow-up of a randomized trial comparing
total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast
Study design
Randomization Radical mastectomy versus quadrantectomy, axillary dissection and
radiotherapy to the ipsilateral residual breast tissue
Setting National Cancer Institute, Milan, Italy
Inclusion criteria Pre- and post-menopausal women with clinically node-negative breast cancer
≤2cm on physical examination
Exclusion criteria Patients >70 years of age
Follow-up Three-monthly clinical examination and annual CXR, liver USS, and
mammogram for 10 years
Annual review after 10 years
322 BREAST SURGERY
◆ Patients were stratified according to their menopausal status.

◆ The quadrantectomy group received 50Gy of radiotherapy to the ipsilateral residual
breast tissue with a boost of 10Gy.
◆ After 1976, patients in both groups who had positive nodes received chemotherapy
with cyclophosphamide, methotrexate, and 5-fluorouracil (5-FU).
Outcomes measures
◆ Local recurrence in the breast and distant recurrence.
◆ Death from breast cancer.
Results
◆ 349 patients underwent a radical mastectomy and 352 had a quadrantectomy.
◆ Median follow-up at the time of the original study was 7 years.
◆ There were three local recurrences in the mastectomy group and one local recurrence
in the quadrantectomy group at 7 years of follow-up.
◆ There were no differences in distant or overall breast-specific survival
Conclusions
Breast-conserving surgery and radiotherapy are as safe as mastectomy in patients with
small clinically node-negative breast cancer.
Critique
This publication was the first randomized trial of breast-conserving surgery in breast can-
cer. Although some data were reported in 1977 before the trial had finished recruiting,
the main paper from 1981 led to a major change in the management of patients with small
clinically node-negative breast cancer.
It showed that outcomes in terms of local recurrence, distant recurrence, and overall
survival were as good with breast conservation and radiotherapy as they were after mas-
tectomy after a longer follow-up period of 7 years.
In 2002 high-quality follow-up data obtained over a much longer period was published,
with a median follow-up of 20 years. Over this period only three of 701 trial patients were
lost to follow-up. After a median of 20 years follow-up, 30/352 patients in the quadrantec-
tomy group developed a local recurrence compared with 8/349 in the mastectomy group.
Therefore the mean (±SE) crude cumulative incidence of recurrent tumour in the same
breast was 8.8 ± 3.2% after 20 years in patients treated with breast-conserving surgery and
2.3 ± 0.8% in the radical mastectomy group. Of the 30 patients who developed recurrent
tumour in the same breast, 29 underwent total mastectomy and one underwent a second
local resection. Of these 30 cases of recurrent tumour, 10 appeared in the scar and thus
were defined as true recurrences, whereas 20 occurred in other quadrants of the breast
and therefore were classified as second ipsilateral carcinomas.
BREAST CONSERVATION VERSUS MASTECTOMY 323
The Fisher study published a 20-year follow-up in the same issue of the same journal,
showing a 14.3% ipsilateral recurrence rate in the lumpectomy and radiotherapy group
versus 39.2% in the lumpectomy without radiotherapy group, confirming the importance
of radiotherapy with breast-conserving surgery.
One of the main issues with the original Veronesi study was the actual nature of the
breast-conserving surgery. The operation of quadrantectomy removed a large amount
of breast tissue, and resulted in a relatively poor cosmetic outcome for some patients.
There was no formal assessment of cosmesis of the breast in this trial, but the procedure
of quadrantectomy never became widely used because of the resulting distortion of the
breast after this type of surgery. The need for concomitant post-operative radiotherapy
may have also added to distortion of the remaining breast tissue. Despite the poor cos-
metic outcomes for some patients, the strength of this trial is the very low local recurrence
rates from Milan, which remain the gold standard even today. This trial ended the doubts
about the safety and efficacy of breast-conserving surgery.
324 BREAST SURGERY
9.3 Sentinel node biopsy in breast cancer

Details of studies
Axillary clearance for all patients with breast cancer was standard practice until 2000.
With the introduction of screening, which tended to detect much earlier cancers,
many more patients had node-negative disease (around 75% of screen-detected cases).
This meant that the morbidity of axillary clearance was even less acceptable, given that
most screen-detected cancers would not have involved nodes. Random four-node axil-
lary sampling was also associated with morbidity and gave less prognostic information
than axillary clearance. Therefore sentinel node biopsy in breast cancer was an attractive
concept, given that it was potentially a much smaller and yet accurate staging axillary
procedure.
The technique of sentinel node biopsy was established in melanoma as a way to detect
cancer spread to regional lymph nodes without having to perform a complete node clear-
ance. The technique is based on the premise that lymph node involvement occurs in a
stepwise fashion, extending away from the tumour site. Identification of the first draining
nodes might permit a selective assessment and, if clear, imply that the remaining regional
nodes are also likely to be clear of tumour. Although an attractive concept, it was unclear
whether the same techniques could be transferred to the management of breast cancer.
It was also unclear whether breast cancer spread to the axillary nodes occurred in a step-
wise fashion, or if ‘skip metastases’ occurred (higher nodes involved when lower nodes
were clear).
The first of these two papers described the use of blue dye alone to detect the sentinel
node in breast cancer. The second paper established the technique of combined blue dye
and radio-isotope with technetium 99m (99mTc) albumin colloid for sentinel node biopsy.
Study references
Main studies
Giuliano, A.E., Kirgan, D.M., Guenther, J.M., and Morton, D.L. (1994) Lymphatic mapping and sentinel
lymphadenectomy for breast cancer. Annals of Surgery, 220, 391–401.
Albertini, J.J., Lyman, G.H., Cox, C., et al. (1996) Lymphatic mapping and sentinel node biopsy in the
patient with breast cancer. JAMA, 276, 1818–22.
Related references
Mansel, R.E., Fallowfield, L., Kissen, M., et al. (2006) Randomized multicenter trial of sentinel node
biopsy versus standard axillary treatment in operable breast cancer: the ALMANAC trial. Journal of
the National Cancer Institute, 98, 599–609.
Krag, D.N., Anderson, S.J., Julian, T.B., et al. (2010) Sentinel-lymph-node resection compared
with conventional lymph-node dissection in clinically node-negative patients with breast
cancer: overall survival findings from NSABP P-32 randomised phase 3 trial. Lancet Oncology,
11, 927–33.
SENTINEL NODE BIOPSY IN BREAST CANCER 325
Study designs
◆ Both studies were prospective cohort studies.
Giuliano et al. Albertini et al.

Number of patients 174 consecutive patients 62 consecutive patients
Setting Single centre, California, USA Single centre, Florida, USA
Inclusion criteria Patients undergoing axillary clearance as part of surgery for breast cancer
Intervention Blue dye node biopsy followed Combined blue dye and radio-isotope
by an axillary clearance node biopsy followed by axillary clearance
◆ In the Giuliano study, isosulphan blue dye was injected near the tumour
5–20 minutes before an axillary incision; 0.5–10mL of dye was used in the first
10 patients. Thereafter 3–5mL was used in the remaining patients.
◆ In the Albertini study, a combination of a vital blue dye and filtered technetium-labelled
sulphur colloid was injected intra-operatively. An intra-operative detector was
used to identify positive nodes, in addition to the visual confirmation of the blue
coloration.
Outcome measures
◆ Sentinel node detection rate.
◆ Optimal dose of blue dye for breast cancer node detection.
◆ Accuracy of sentinel node biopsy compared with axillary clearance to detect node-
positive patients.
Results
Giuliano et al. Albertini et al.
(blue dye only) (blue dye and radio-isotope)
Number of cases 174 62
Sentinel node detected 114 (65.5%) 57 (92%)
Patients with positive nodes 62/174 (35.6%) 18/62 (32%)
Positive sentinel nodes 37/114 (32.4%) 18/62 (32%)
False-negative sentinel nodes 5 (4.3%) 0
Concordance 109 (95.6%) 18 (100%)
Both studies showed that the sentinel node biopsy technique identified a sentinel node
in the majority of patients during axillary surgery. They also showed that sentinel node
biopsy accurately predicted the presence or absence of metastatic nodal disease in
the axilla.
326 BREAST SURGERY
Conclusions
Sentinel node biopsy is technically feasible in breast cancer patients and is an accurate
way of detecting node-positive patients. The combination of blue dye and radio-isotope
is the best method to use.
Critique
Giuliano’s team already had experience of sentinel node biopsy using blue dye in mela-
noma and were the first to describe its use in breast cancer. The detection rate of sentinel
nodes as well as the dose of blue dye was assessed in their study.
Albertini et al. showed that the combined technique of blue dye and radio-isotope
improved the detection rate of sentinel nodes and also improved the detection of positive
sentinel nodes. This combined technique is now established as standard practice in breast
cancer surgery.
In both these two early trials, the patients also underwent axillary clearance. In subse-
quent randomized trials patients were randomized to axillary clearance alone or sentinel
node biopsy followed by completion axillary clearance for node-positive women, with no
further surgery for node-negative women. These and other trials found that patients in
the sentinel node arm had reduced morbidity and no detrimental effect on either disease-
free or overall survival. Therefore practice changed, and sentinel node biopsy is now the
standard of care with axillary treatment (either clearance or radiotherapy) as a second
treatment if the sentinel node is positive.
AXILLARY DISSECTION VERSUS NO AXILLARY DISSECTION 327
9.4 Axillary dissection versus no axillary dissection in women

with invasive cancer and sentinel node metastases
Details of study
Sentinel node biopsy (SNB) has become standard practice in the management of breast
cancer. If the SNB is negative, no further axillary treatment is required. If the sentinel
node is positive, traditional thinking has been to consider further treatment of the axilla.
This usually means either a completion axillary clearance or occasionally axillary radio-
therapy. The American College of Surgeons Oncology group challenged this thinking and
designed a randomized trial to assess the impact of no further intervention in patients
with one or two involved nodes.
Study references
Main study
Giuliano, A.E., Hunt, K.K., Ballman, K.V., et al. (2011) Axillary dissection vs no axillary dissection in women
with invasive breast cancer and sentinel node metastases: A randomized clinical trial. JAMA, 305, 569–75.
Related references
Giuliano, A.E., McCall, L., Beitsch, P., et al. (2010) Locoregional recurrence after sentinel lymph node
dissection with or without axillary dissection in patients with sentinel lymph node metastases: the
American College of Surgeons Oncology Group Z0011 randomized trial. Annals of Surgery, 252, 426–33.
Study design
Randomization Completion axillary dissection versus no further axillary treatment
following sentinel node biopsy
Number of centres 115 centres in US and Canada
Number of patients 891*
Recruitment period 1999–2004
Inclusion criteria T1 or T2 cancers undergoing breast-conserving surgery
1 or 2 positive sentinel nodes
Exclusion criteria 3 or more positive sentinel nodes
Primary outcomes Axillary and local recurrence rates
Secondary outcomes Overall survival and disease-free survival
Follow-up (median) 6.3 years
* The study was designed to include 1900 patients and the intention was to perform an analysis after 500 deaths.
The study closed early.
◆ All patients had tangential field radiotherapy to the breast alone and standard
systemic therapy.
◆ Local recurrence was defined as recurrent disease in the ipsilateral breast.
328 BREAST SURGERY
Results
◆ The clinical and tumour characteristics were well matched between the groups
SNB only SNB and axillary dissection

Number of deaths 42 52
Disease-free survival 82.2% 83.9%
Local recurrence rates 8 (1.8%) 15 (3.6%)
Axillary recurrence rate 4 (0.9%) 2(0.5%)
◆ Locoregional recurrence was seen in 29 (3.4%) patients overall.

◆ The SNB only group did not have a statistically inferior overall survival (p = 0.008 for
non-inferiority).
◆ Systemic therapy given was similar between the groups (whether chemotherapy,
hormone therapy, or Herceptin).
Conclusion
An axillary clearance is not necessary for patients with early (T1/T2) cancer who have
only one or two positive sentinel nodes.
Critique
The Z0011 trial was closed early by the independent data and safety monitoring com-
mittee because of a lower than expected recruitment and mortality in both groups. The
calculations to power the study were based on a predicted 80% 5-year survival, but as
the trial progressed, it became apparent that 5-year survival was more than 90%. As a
consequence, it would have taken more than 20 years to observe the planned 500 deaths.
It is interesting to note that additional positive nodes were identified in 26% of patients
who underwent an axillary dissection. Given that both the SNB-only and the SNB and
axillary dissection groups had very low axillary recurrence rates (0.9% and 0.5% respec-
tively), one can surmise that many patients in the SNB only group had residual ‘untreated’
positive nodes that did not result in recurrent axillary disease (within a median of 6.3 years
of follow-up).
Closer examination of the two groups revealed that a higher proportion of patients in
the SNB only group had only micrometastases (<2mm in size) in the involved sentinel
node. In the axillary dissection arm 137/363 patients (37.5%) had only micrometasta-
ses, whereas in the SNB-only arm 164/366 patients (44.8%) had only micrometastases.
However, these differences still do not fully explain why so few patients in the SNB
only group developed nodal recurrence. The most likely explanation remains that those
patients with residual node positive disease were effectively ‘treated’ by systemic therapy.
This study shows that excellent outcomes can be achieved with modern multimodality
treatments for node-positive patients even without axillary dissection.
MANAGEMENT OF LOCALIZED DCIS 329
9.5 Management of localized DCIS

Details of studies
Before the introduction of the breast screening service in the UK in 1988, the incidence of
ductal carcinoma in situ (DCIS) was relatively small. The standard treatment was mastec-
tomy. The papers discussed here challenged this orthodoxy and sought to establish new
and less invasive ways of managing this condition.
The National Surgical Adjuvant Breast and Bowel Project (NSABP) B17 study was a
landmark trial that investigated the role of radiotherapy for localized DCIS. The same
investigative team subsequently reported the results of the NSAPB-24 trial, which
was designed to measure whether tamoxifen had any additional benefit to surgery and
radiotherapy.
Study references
Main studies
Fisher, B., Dignam, J, Wolmark, N., et al. (1998) Lumpectomy and radiation for the treatment of
intraduct breast cancer: findings from the National Surgical Adjuvant Breast and Bowel
Project B-17. Journal of Clinical Oncology, 16, 441–52.
Fisher, B., Dignam, J., Wolmark, N., et al. (1999) Tamoxifen in treatment of intraduct breast cancer.
National Surgical Adjuvant Breast and Bowel Project B-24 randomised controlled trial. Lancet,
353, 1993–2000.
Related references
Goodwin, A., Parker, S., Ghersi, D., and Wilcken, N. (2009) Post operative radiotherapy for ductal
carcinoma in situ of the breast. Cochrane Database of Systematic Reviews, 3, CD000563.
Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) (2010) Overview of the randomized trials
of radiotherapy in ductal carcinoma in situ of the breast. Journal of the National Cancer Institute
Monographs, 41, 162–77.
Cusick, J., Sestak, I., Pinder, S.E., et al. (2011) Effect of tamoxifen and radiotherapy in women with
locally excised ductal carcinoma in situ: long term results from the UK/ANZ DCIS trial. Lancet
Oncology, 12, 21–9.
Wapnir, L., Dignam, J., Fisher, B., et al. (2011) Long term outcomes of invasive ipsilateral breast tumour
recurrences after lumpectomy in B-17 and B-24 randomized clinical trials for DCIS. Journal of the
National Cancer Institute, 103, 478–88.
Study designs
◆ Both studies were prospective RCTs.
NSABP B-17 NSABP B-24

Randomization Lumpectomy versus lumpectomy and Tamoxifen versus no tamoxifen (after
radiotherapy to the breast lumpectomy and breast radiotherapy)
330 BREAST SURGERY
Setting Multi-centre trials in the USA and Canada

Time period 1985–1990 May 1991 to April 2003
Inclusion criteria Pre- and post-menopausal women with Patients who had undergone
localized DCIS suitable for local excision lumpectomy and radiotherapy for DCIS
Follow-up (months) 207 163
◆ In the NSABP B-24 study, tamoxifen was given for 5 years.
Outcome measures
Primary endpoints
◆ Recurrence of DCIS or invasive carcinoma in the ipsilateral/contralateral breast.
Secondary endpoints
◆ Overall and breast cancer-specific survival for both studies.
Results
NSAPB B-17 study
◆ After 8 years of follow-up, the NSAPB B-17 study demonstrated a reduction in
local recurrence from 27% in the surgery alone group to 12% for the surgery and
radiotherapy group.
◆ The addition of radiotherapy reduced the absolute 10-year ipsilateral recurrence
rate (either invasive cancer or DCIS) by 15.2% (SE 1.6%, 28.1% versus 12.9%,
p <0.00001).
◆ At 10 years follow-up there was no difference in breast cancer-specific or overall
survival with radiotherapy.
◆ Of the ipsilateral recurrences after treatment of DCIS, 53% were invasive cancer.
◆ Radiotherapy appeared to have little effect on contralateral or distant events.
NSABP B-24
◆ Tamoxifen reduced the risk of any breast cancer recurrence by 44% (HR 6.53).
◆ The main effect of tamoxifen was seen in the contralateral breast and there was no
effect on overall or breast cancer-specific survival
◆ Tamoxifen did not add any benefit over radiotherapy alone for ipsilateral breast
recurrence
Conclusions
Radiotherapy halves the risk of developing recurrent DCIS and also reduces the risk
of developing invasive cancer in the breast following local surgery. Tamoxifen also has
a beneficial effect in this setting by reducing the risk of breast cancer, primarily in the
contralateral breast.
MANAGEMENT OF LOCALIZED DCIS 331
Critique
Before these trials the standard treatment for DCIS was mastectomy, which not unsurpris-
ingly has a very low rate of recurrence. Between 1985 and 1990 there were four random-
ized trials of radiotherapy after breast-conserving surgery and 2 trials that looked at the
role of tamoxifen. These studies have consistently shown that radiotherapy reduces ipsi-
lateral breast recurrence of DCIS and invasive cancer by about half (HR 0.48, 95%CI 0.41
to 0.58, P<0.00001). There was excellent survival after treatment for DCIS (86% control
vs 87% radiotherapy) in NSAPB B-17 after 12 years of follow-up and little toxicity from
radiotherapy. In NSAPB B-24, clear margins were not required for trial entry (although
75% of patients had clear margins) but the radiotherapy effect was similar.
Despite the clear evidence of benefit for radiotherapy in localized DCIS, the Sloane
Project audit of DCIS cases has shown wide variation in the current use of radiotherapy
across the UK, ranging from almost 100% of cases receiving radiotherapy in Scotland to
some areas in England where it is rarely used (Morrow et al.). Despite the evidence from
the studies presented here, it appears that tamoxifen is not widely used after surgery for
DCIS. Some centres now measure the oestrogen receptor (ER) status of DCIS and use
tamoxifen in ER-positive cases.
Morrow, M., Strom, E.A., Bassett, L.W., et al. (2002) Standard for the management of ductal carcinoma
in situ of the breast (DCIS). CA: Cancer Journal for Clinicians, 52, 256–76.
The Sloane Project website (<http://www.sloaneproject.co.uk>) contains all Sloane Project publications.
332 BREAST SURGERY
9.6 The role of oestrogen receptors in breast cancer

Details of studies
In 1896 Beatson noted a relationship between ovarian function and the behaviour of some
breast cancers. Subsequently, a variety of hormonal manipulations have been used in the
treatment of breast cancer. This study analysed the oestrogen receptor status of breast can-
cers in 286 women undergoing mastectomy and correlated this with rates of recurrence.
Study references
Main study
Cooke, T., George, D., Shields, R., Maynard, P., and Griffiths, K. (1979) Oestrogen receptors and
prognosis in early breast cancer. Lancet, 1, 995–7.
Related reference
Blamey, R.W., Bishop, H.M., Blake, J.R., et al. (1980) Relationship between primary breast tumor recep-
tor status and patient survival. Cancer, 46, 2765–9.
Study design
◆ Single-centre prospective cohort series.
Setting Single-centre consecutive series
Inclusion criteria Women undergoing mastectomy and axillary biopsy or clearance for
breast cancer
Exclusion criteria Patients who received cytotoxic or hormone therapy
Intervention Tumour oestradiol was measured biochemically using the dextran-coated
charcoal method
◆ Tumours staged by TNM classification using clinical and pathological data.

◆ Recurrent disease was noted in relation to the presence or absence of oestrogen
receptors.
Results
◆ 75 women were pre-menopausal. Of these, 36% had tumours containing oestrogen
receptors.
◆ Of 115 post-menopausal women, 55% had tumours containing oestrogen.
◆ Recurrence rate was related to the presence or absence of oestrogen.
◆ Recurrence rates were significantly higher in patients whose tumours were
ER-negative (p <0.001), irrespective of whether they were pre- or post-menopausal.
ROLE OF OESTROGEN RECEPTORS IN BREAST CANCER 333
◆ The highest recurrence rates were seen in patients with axillary metastatic nodal
disease who were ER-negative.
◆ Women who had ER-negative primary tumours and no axillary nodal disease showed
the same high rate of recurrence as all women with axillary nodal disease.
Critique
This was one of the first studies analysing oestrogen receptor status in relation to stage of
disease and outcome. It demonstrates the importance of ER status as a prognostic factor.
Compared with current studies, the numbers are small, the follow-up is short, and the
percentage of ER-positive cases is smaller than we would expect with current methods of
assessment.
334 BREAST SURGERY
9.7 Histological grade and prognosis in breast cancer

Details of studies
Although a number of studies had shown that morphological assessment of the degree
of differentiation in breast cancer could provide useful prognostic information in breast
cancer, tumour morphology and grade were not routinely used in many centres because of
perceived problems with reproducibility and consistency. In this study the authors modi-
fied the Bloom and Richardson classification to make the diagnostic criteria more objective.
Study references
Main study
Elston, C.W. and Ellis, I.O. (1991) Pathological prognostic factors in breast cancer. I. The value of
histological grade in breast cancer: experience from a large study with long-term follow-up.
Histopathology, 19, 403–10.
Related reference
Bloom, H.J. and Richardson, W.W. (1957) Histological grading and prognosis in breast cancer: a study
of 1409 cases of which 359 have been followed for 15 years. British Journal of Cancer, 11, 359–77.
Study design
◆ This study examined patients recruited to the Nottingham Tenovus Primary Breast
Cancer Study.
◆ Tumours were graded by two histopathologists using a modified Bloom and
Richardson system which incorporated more objective criteria in assessment of three
parameters (tubule formation, nuclear pleomorphism, and mitotic count).
◆ The overall grade was derived from a summation of the scores for each of three variables.
◆ The relationship between tumour grade and prognosis was analysed in 1830 patients
where long-term follow-up was available.
Results
◆ There was a highly significant correlation between grade and prognosis.
◆ Patients with grade I tumours had a significantly better survival than those with
grade II and III tumours.
◆ Both recurrence-free survival and overall survival were worse in patients with high-
grade cancers.
Critique
This was the first study to objectively modify the grading of invasive breast cancer and
correlate this grading system with prognosis in a large group of patients. As a result of
this, the modified Bloom and Richardson grading system has been widely used through-
out the world to provide useful information which guides management in patients with
breast cancer.
ROLE OF THE HER-2/NEU ONCOGENE IN BREAST CANCER 335
9.8 The role of the HER-2/neu oncogene in breast cancer

Details of studies
In the 1980s evidence emerged linking proto-oncogenes with the induction and main-
tenance of human malignancies. The Her-2/neu (HER2) gene, which encodes a tyrosine
kinase receptor, was identified on chromosome 17 (band q21). This gene was shown to
be amplified in human breast cancer cell lines. In this seminal paper, the HER-2/neu gene
was examined in a series of 189 patients with breast cancer. Alterations in the gene were
correlated with overall survival and time to relapse.
Study references
Main study
Slamon, D.J., Clark, G.M., Wong, S.G., et al. (1987) Human breast cancer: correlation of relapse and
survival with amplification of the HER-2/neu oncogene. Science, 235, 177–82.
Related reference
Seshadri, R., Firgaira, F.A., Horsfall, D.J., McCaul, K., Setlur, V., and Kitchen, P. (1993) Clinical signifi-
cance of HER-2/neu oncogene amplification in primary breast cancer. The South Australian Breast
Cancer Study Group. Journal of Clinical Oncology, 11, 1936–42.
Study design
◆ 189 primary breast cancer specimens were examined by Southern blot analyses to
determine if there was evidence of HER-2/neu gene amplification.
◆ Gene amplification was correlated with several disease parameters.
Results
◆ HER-2/neu was found to be amplified from twofold to more than 20-fold in 30% of
tumours. A strong statistically significant correlation was found between the degree
of gene amplification and time to relapse (p <0.001) and survival (p = 0.0011).
◆ On univariate analyses gene amplification was superior to all other prognostic factors
with the exception of lymph-node-positive status.
◆ In the 86 node-positive cases, gene amplification continued to be a strong prognostic
factor providing additional and independent predictive information on both time to
relapse and overall survival
Critique
This was the first study providing evidence that HER-2/neu gene amplification was associ-
ated with poor outcome in breast cancer. As a result of this landmark paper and subse-
quent studies, the monoclonal antibody trastuzumab was developed and used as targeted
treatment in HER2-positive breast cancer. This has led to significant improvements in
survival in this poor prognostic group. The HER2 status is now considered essential when
making adjuvant treatment decisions in all breast cancer patients.
336 BREAST SURGERY
9.9 Patterns of gene expression in breast cancer

Details of studies
It is known that breast cancer is a diverse heterogeneous group of diseases with varying
patterns of biological behaviour and response to treatment. Treatment decisions are made
depending on the stage of disease and a variety of pathological parameters. There is a
continuing search to target treatment and choose therapies that will benefit individual
patients. This study examined variation in transcriptional programmes in 65 samples of
human breast tumours from 42 different individuals. The tumours were classified into
different subtypes distinguished by variations in gene expression profiling using comple-
mentary DNA (cDNA) micro-array techniques.
Study references
Main study
Perou, C.M., Sørlie, T., Eisen, M.B., et al. (2000) Molecular portraits of human breast tumours. Nature,
406, 747–52.
Related reference
Sørlie, T., Perou, C.M., Tibshirani, A., et al. (2001) Gene expression patterns of breast carcinomas
distinguish tumour subclasses with clinical implications. Proceedings of the National Academy of
Sciences of the USA, 98, 10,869–74.
Study design
◆ 65 surgical specimens from 42 individuals were examined using cDNA microarrays
representing 8102 human genes.
Results
Hierarchical clustering was used to group genes on the basis of similarity in the pattern
with which their expression varied. These variations were depicted graphically using den-
drograms. The tumours were classified into subtypes distinguished by differences in pat-
terns of gene expression. The ‘molecular portraits’ showed variation in proliferation and
activity of signalling pathways. The tumours were classified as luminal like, basal like,
HER-2/neu positive, and normal breast like.
Critique
This was the first study to examine the molecular portraits of individual tumours and
provided a deeper and more complete understanding of human breast cancer. As a con-
sequence, prognostic and predictive gene expression tests such as Oncotype DX and
Mammoprint have been developed to help improve decision-making around adjuvant
therapy in the individual with breast cancer. It has also encouraged an examination of
classical prognostic factors (ER/PR, HER2, and Ki-67), helping us to look at these factors
more critically. This may ultimately lead to the development of better prognostic indices
which will facilitate tailored individual treatment for each patient with breast cancer.
ORIGINS OF ADJUVANT CHEMOTHERAPY FOR BREAST CANCER 337
9.10 The origins of adjuvant chemotherapy for breast cancer

Details of studies
An appreciation that distant dissemination of cancer cells could occur long before diag-
nosis, and thereby predispose to later metastatic relapse regardless of local control, repre-
sented a paradigm shift in the understanding of breast cancer. This provided a rationale
for both less radical surgery and the development of adjuvant systemic therapies. By the
early 1970s a combination of three cytotoxic drugs—cyclophosphamide, methotrexate,
and 5-FU (CMF)—had shown significant activity in advanced breast cancer and therefore
was selected for testing in the adjuvant setting.
Study references
Main study
Bonadonna, G., Brusamolino, E., Valagussa, P., et al. (1976) Combination chemotherapy as an adjuvant
treatment in operable breast cancer. New England Journal of Medicine, 294, 405–10.
Related references
Fisher, B., Carbone, P., Economou, S.G., et al. (1975) 1-phenylalanine mustard (L-PAM) in the management
of primary breast cancer. A report of early findings. New England Journal of Medicine, 292, 117–22.
Bonadonna, G., Valagussa, P., Moliterni, A., et al. (1995) Adjuvant cyclophosphamide, methotrexate,
and fluorouracil in node-positive breast cancer: the results of 20 years of follow-up. New England
Tancini, G., Bonadonna, G., Valagussa, P., Marchini, S., and Veronesi, U. (1983) Adjuvant CMF in breast
cancer: comparative 5-year results of 12 versus 6 cycles. Journal of Clinical Oncology, 1, 2–10.
Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) (2012) Comparisons between different
polychemotherapy regimens for early breast cancer: meta-analyses of long-term outcome among
100,000 women in 123 randomised trials. Lancet, 379, 432–44.
Study design
◆ Prospective single-centre RCT.
Randomization Adjuvant CMF chemotherapy versus no adjuvant treatment
Setting Single-centre study conducted at the Instituto Nazionale Tumori, Milan, Italy
(1973–1975)
Inclusion criteria Women <75 years
Recent radical mastectomy
≥1 pathologically involved lymph node
No evidence of distant disease
Stratification Age (≤49, 50–75)
Type of surgery (conventional, extended radical mastectomy)
Number of positive nodes (1–3, >4)
338 BREAST SURGERY
◆ The CMF group received cyclophosphamide 100mg/m2 orally on days 1–14,

methotrexate 40mg/m2 IV on days 1 and 8, and 5-FU 600mg/m2 IV on days 1 and 8,
repeated every 28 days. The treatment lasted for twelve 28-day cycles.
◆ Neither the CMF nor the control group received adjuvant radiotherapy or endocrine
therapy.
Outcome measures
◆ Treatment failure (defined as local, regional, or distant recurrence).
Results
◆ The rate of treatment failure was significantly lower in the CMF arm than in the
control arm (5.3% versus 24%, p <10–6).
◆ Distant recurrences accounted for more the 81% of treatment failures.
◆ CMF was associated with nausea, vomiting, and anorexia in almost all patients, with
about one-third failing to comply fully with oral cyclophosphamide for this reason.
◆ Significant bone marrow suppression was rare, and the incidence of infection was not
increased in the CMF group.
◆ Partial alopecia occurred in 55%. Less common toxicities included mucositis,
conjunctivitis, cystitis, and amenorrhoea.
Conclusion
This paper provided one of the first demonstrations that adjuvant systemic chemotherapy
could improve relapse-free survival in woman with node-positive breast cancer. With
longer follow-up the improvements in disease-free survival (DFS) were shown to trans-
late into improved overall survival (OS).
Critique
A later study by the same group demonstrated equivalent efficacy for a shorter course
of CMF (6 months rather than 12 months), and this regimen became the standard of
care for pre-menopausal woman with node-positive breast cancer and the reference arm
for future chemotherapy studies. The benefits of adjuvant CMF versus no chemotherapy
have been confirmed in the Oxford overview meta-analysis, which reported a relative
reduction of 24% in breast cancer-specific mortality. Although the benefits of CMF in
the Bonadonna study appeared to be restricted to the pre-menopausal group, this may
have reflected inadequate chemotherapy delivery in older women for whom the protocol
stipulated lower doses. Indeed, the benefit of adjuvant chemotherapy across age groups
up to at least 70 years of age has subsequently been confirmed in the Oxford overviews.
ADJUVANT ANTHRACYCLINES 339
9.11 Adjuvant anthracyclines

Details of study
The anthracyclines doxorubicin and epirubicin are among the most active single agents in
the treatment of metastatic breast cancer. One of the first trials to incorporate an anthra-
cycline into the adjuvant setting was NSABP-B15, which compared 6 months of CMF
with four cycles of doxorubicin and cyclophosphamide (AC). This study demonstrated
equivalent DFS for the much shorter anthracycline-based regimen. Subsequent trials
aimed to improve outcomes above those achieved with CMF or AC by either replacing
the methotrexate in CMF with an anthracycline (FAC or FEC) or by administering an
anthracycline and CMF sequentially.
Study references
Main study
Levine, M.N., Bramwell, V.H., Pritchard, K.I., et al. (1998) Randomized trial of intensive
cyclophosphamide, epirubicin, and fluorouracil chemotherapy compared with cyclophosphamide,
methotrexate, and fluorouracil in premenopausal women with node-positive breast cancer. National
Cancer Institute of Canada Clinical Trials Group. Journal of Clinical Oncology, 16, 2651–8.
Related references
Fisher, B., Brown, A.M., Dimitrov, N.V., et al. (1990) Two months of doxorubicin-cyclophosphamide
with and without interval reinduction therapy compared with 6 months of cyclophosphamide,
methotrexate, and fluorouracil in positive-node breast cancer patients with tamoxifen-nonresponsive
tumors: results from the Nation Surgical Adjuvant Breast and Bowel Project B-15. Journal of Clinical
Oncology, 8, 1483–96.
French Adjuvant Study Group (2001) Benefit of a high-dose epirubicin regimen in adjuvant chemotherapy
for node-positive breast cancer patients with poor prognostic factors: 5-year follow-up results of
French Adjuvant Study Group 05 randomized trial. Journal of Clinical Oncology, 19, 602–11.
Poole, C.J., Earl, H.M., Hiller, L., et al. (2006) Epirubicin and cyclophosphamide, methotrexate, and
fluorouracil as adjuvant therapy for early breast cancer. New England Journal of Medicine, 355,
1851–62.
Study design
Randomization CMF versus CEF adjuvant chemotherapy
Setting National Cancer Institute of Canada (1989–1993)
Inclusion criteria Pre-menopausal women with node-positive breast cancer following modified
radical mastectomy/lumpectomy and axillary clearance
Stratification Type of surgery
ER status
Number of positive nodes (1–3, 4–10, or >10)
340 BREAST SURGERY
◆ The CMF group received cyclophosphamide 100mg/m2 orally on days 1–14,

methotrexate 40mg/m2 IV on days 1 and 8, and 5-FU 600mg/m2 IV on days 1 and 8.
◆ The CEF group received cyclophosphamide 75mg/m2 orally on days 1–14, epirubicin
60mg/m2 IV on days 1 and 8, and 5-FU 500mg/m2 IV on days 1 and 8.
◆ Each cycle was administered monthly for 6 months.
Outcome measures
◆ Relapse-free survival defined as the time from randomization until the patient
developed local, regional, or distant recurrence.
Results
◆ Five-year relapse-free survival was superior in the CEF arm (63% versus 53%,
p = 0.009).
◆ Five-year actuarial survival rates were 77% (CEF) versus 70% (CMF) (p = 0.03).
◆ CEF was associated with more alopecia, vomiting, and myelosuppression.
◆ The rate of hospitalization for febrile neutropenia was 1.1% in the CMF arm versus
8.5% in the CEF arm.
◆ Five patients in the CEF arm developed acute leukaemia.
Conclusion
Anthracycline-based chemotherapy (CEF) is superior to CMF in terms of relapse-free
and overall survival in pre-menopausal women with axillary node-positive breast cancer.
Critique
This was the first trial to demonstrate a significant advantage for an anthracycline-
containing regimen in node-positive breast cancer. These findings were subsequently
confirmed by several other large studies and by the Oxford overview which described
a further 15–20% proportional reduction in breast cancer mortality rates for anthracy-
cline-based regimens in comparison with CMF. For this reason anthracyclines remain the
mainstay of almost all adjuvant chemotherapy regimens, although concerns regarding
the late toxicities of cardiomyopathy and secondary leukaemia have fuelled interest in
anthracycline-free adjuvant regimens.
ADJUVANT TAXANE CHEMOTHERAPY 341
9.12 Adjuvant taxane chemotherapy

Details of studies
During the 1990s the taxanes (docetaxel and paclitaxel) were shown to be highly active
drugs in advanced breast cancer with similar efficacy to anthracyclines. Their distinct
mechanism of action and the observation of at least partial non-cross-resistance with
anthracyclines provided the rationale for a series of trials in which taxanes were added to
standard anthracycline-based adjuvant regimens.
Study references
Main study
Martin, M., Pienkowski, T., Mackey, J., et al. (2005) Adjuvant docetaxel for node-positive breast cancer.
New England Journal of Medicine, 352, 2302–13.
Related references
Roché, H., Fumoleau, P., Spielmann, M., et al. (2006) Sequential adjuvant epirubicin-based and
docetaxel chemotherapy for node-positive breast cancer patients: the FNCLCC PACS 01 trial.
Journal of Clinical Oncology, 24, 5664–71.
Henderson, C., Berry, D.A., Demetri, G.D., et al. (2003) Improved outcomes from adding sequential
paclitaxel but not from escalating doxorubicin dose in an adjuvant chemotherapy regimen for
patients with node-positive primary breast cancer. Journal of Clinical Oncology, 21, 976–83.
Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) (2012) Comparisons between different
polychemotherapy regimens for early breast cancer: meta-analyses of long-term outcome among
100 000 women in 123 randomised trials. Lancet, 379, 432–44.
Study design
Randomization TAC versus FAC adjuvant chemotherapy
Setting International RCT involving 20 countries (1997–1999)
Inclusion criteria Women aged 18–70 years
Mastectomy/breast-conserving surgery for cancer
At least one histologically positive lymph node
Stratification Institution
Number of positive nodes (1–3, >4)
◆ TAC: docetaxel 75mg/m2, doxorubicin 50mg/m2, and cyclophosphamide 500mg/m2

given IV on day 1 and repeated every 21 days (six cycles).
◆ FAC: 5-FU 500mg/m2, doxorubicin 50mg/m2, and cyclophosphamide 500mg/m2
given IV on day 1 and repeated every 21 days (six cycles).
342 BREAST SURGERY
Outcome measures
◆ Primary endpoint: DFS (defined as the time from randomization to the date of
a clinical relapse, a second cancer, or death).
◆ Secondary endpoints: OS, toxicity, and quality of life (QoL).
Results
◆ Disease-free survival was superior in the TAC group (75% versus 68%, p = 0.001,
HR 0.72, 95%CI 0.59–0.88).
◆ Overall survival was also superior in the TAC group with an estimated 5-year OS rate
of 87% versus 81% in the FAC group (HR 0.70, 95%CI 0.53–0.91, p = 0.008).
◆ TAC was associated with more toxicity, in particular a substantially higher rate of
febrile neutropenia (24.7% versus 2.5%, p <0.001).
◆ TAC was associated with a greater reduction in QoL scores during treatment, but
both groups had returned to baseline after completion of chemotherapy.
Conclusion
The addition of a taxane to anthracycline-based chemotherapy can significantly improve
disease-free and OS.
Critique
Overall, the incorporation of taxanes into adjuvant regimens has resulted in further
incremental improvements in breast cancer mortality, as demonstrated by the most
recent Oxford overview. However, individual clinical trials have produced discordant
results, which may in part be explained by differences in the ‘strength’ of chemotherapy
used in the control arm, in particular the anthracycline dose. Therefore the doxorubicin
dose (50mg/m2) used in this trial could be criticized as being lower than the standard of
60mg/m2, and this may have contributed to the magnitude of benefit seen. On the other
hand, the experimental and control arms were matched with regard to duration and dos-
ages of cytotoxic drugs.
As reported in this trial, the addition of docetaxel to anthracycline-based chemother-
apy is associated with increased toxicity and generally necessitates additional supportive
care, such as the use of granulocyte colony-stimulating factor (GCSF) for the prevention
of febrile neutropenia. For this reason there has been much interest in the discovery of
biomarkers which could identify those patients who will accrue the greatest benefit from
the addition of a taxane. This effort is ongoing, and to date biomarkers predictive of selec-
tive taxane benefit remain elusive.
TAMOXIFEN AS ADJUVANT THERAPY 343
9.13 Tamoxifen as adjuvant therapy in breast cancer

Details of studies
Research performed by ICI Pharmaceuticals (now known as AstraZeneca) in the 1950s
laid the groundwork for the development of tamoxifen. Early work into oral contracep-
tives showed that these could inhibit oestrogen production. It was quickly appreciated
that tamoxifen, a potent anti-oestrogen which was developed out of this work, might
have anti-cancer activity, given that many breast cancers appeared to be dependent on
oestrogen. Early in vitro and animal studies confirmed that it was metabolically active
and appeared to work, at least in part, by blocking binding of oestradiol to the oestrogen
receptor of human breast cancer.
Early clinical studies in patients with metastatic disease demonstrated that the drug was
most beneficial when administered over extended periods of time. Attention was subse-
quently directed to its use in the adjuvant setting. This paper was the first trial to assess the
efficacy of tamoxifen in the adjuvant setting in both pre- and post-menopausal women.
Study references
Main study
Nolvadex Adjuvant Trial Organisation (NATO). (1983) Controlled trial of tamoxifen as adjuvant agent
in management of early breast cancer. Interim analysis at four years. Lancet, 1, 257–61.
Related reference
Scottish Breast Cancer Trials Committee, Scottish Cancer Trials Office (MRC) Edinburgh (1987)
Adjuvant tamoxifen in the management of operable breast cancer: the Scottish trial. Lancet, 2, 171–5.
Study design
◆ A prospective multi-centre randomized trial.
Randomisation Adjuvant tamoxifen versus no further treatment
Inclusion criteria Mastectomy and axillary surgery (sample/clearance)
Pre-menopausal women: pathological stage II
Post-menopausal women: pathological stage I and II
Exclusion criteria Post-operative adjuvant chemotherapy
Follow-up 21 months
◆ Tamoxifen was given 10mg twice daily for 2 years

344 BREAST SURGERY
Outcome measures
◆ Treatment failure (defined as development of recurrent disease or death).
Results
◆ There were significantly fewer treatment failures in patients receiving tamoxifen
compared with the control group (14.2% versus 20.5%, p = 0.01)
◆ Although the tamoxifen group demonstrated fewer treatment failures, there was no
significant difference in mortality between the two groups.
◆ Minimal toxicity was reported and only 14 patients (2.2%) stopped tamoxifen
because of side effects.
Critique
This was the first RCT which demonstrated a clear benefit for post-operative adjuvant
tamoxifen in early breast cancer. It was viewed as being a very large trial for its time
and was also important as it included both pre- and post-menopausal women and both
ER-positive and ER-negative patients.
Despite its large size, the trial was underpowered to look at subgroups and this, com-
bined with the relatively inaccurate method of measuring ER status (ligand binding
assay), meant that the authors came to the incorrect conclusion that tamoxifen was of
equal benefit in both ER-positive and ER-negative patients. In this trial, tamoxifen was
administered for 2 years, although subsequent studies have shown more benefit if it is
administered for 5 years.
EFFECT OF TAMOXIFEN AND ADJUVANT CHEMOTHERAPY 345
9.14 The effect of tamoxifen and chemotherapy in

breast cancer
Details of studies
Prior to this overview there was uncertainty about the benefits of tamoxifen in the adju-
vant setting. Some important trials did show a statistically significant benefit (NATO) but
other smaller underpowered studies failed to demonstrate a benefit, leading to confusion.
The unique collaboration and robustness of the statistical methodology used in this paper
meant that the results of this paper were widely accepted and tamoxifen became standard
therapy in ER-positive post-menopausal patients.
Study references
Main study
Early Breast Cancer Trialists’ Collaborative Group (1988) Effects of adjuvant tamoxifen and of cytotoxic
therapy on mortality in early breast cancer: an overview of 61 randomised trials among 28,896
women. New England Journal of Medicine, 319, 1681–92.
Related references
Nolvadex Adjuvant Trial Organisation (NATO). (1983) Controlled trial of tamoxifen as adjuvant agent
in management of early breast cancer: interim analysis at four years. Lancet, 1, 257–61.
Cuzick, J., Stewart, H., Peto, R., et al. (1987) Overview of randomized trials of postoperative adjuvant
radiotherapy in breast cancer. Cancer Treatment Reports, 71, 15–29.
Early Breast Cancer Trialists’ Collaborative Group (1998) Tamoxifen for early breast cancer; an
overview of the randomised trials. Lancet, 351, 1451–67.
Study design
The Early Breast Cancer Trialists’ Collaborative Group (EBTCG) brought together physi-
cians from around the world to pool trial data for all breast cancer adjuvant trials using
tamoxifen and chemotherapy which began recruiting before 1985. They developed pio-
neering methodology where all collaborators updated their databases and formed a uni-
fied central database for analysis of individual patient data and collection of mortality.
Results
◆ The overview reported the results of a meta-analysis of 28 trials which assessed the
benefit of tamoxifen.
◆ 16,513 women were recruited and almost 4000 deaths were recorded after 5 years of
follow-up.
◆ The systematic overview reported that tamoxifen therapy was responsible for a
significant reduction in mortality in women aged ≥50 years. In this group, the drug
reduced the annual odds of death by about one-fifth over the first 5 years.
◆ 13,442 women were recruited to 40 chemotherapy trials. Chemotherapy was
associated with a clear reduction in mortality in women aged <50 years.
◆ Polychemotherapy in women aged <50 years reduced the annual odds of death
346 BREAST SURGERY
during the first 5 years by around 25%.

◆ Combination chemotherapy was more effective than single-agent regimens.
◆ There appeared to be no survival advantage for longer duration of chemotherapy
(8–24 months) compared with the same regimen over 4–6 months.
◆ Given the large numbers involved, this overview provided compelling evidence that
tamoxifen could reduce 5-year mortality in the over-50s and cytotoxic therapy could
reduce mortality in those less than 50 years.
Critique
The strength of this overview was in developing the methodology and establishing the
worldwide collaboration. The EBCTCG has continued to meet and publish overviews
every 5 years.
It took another 10 years to show that tamoxifen was also of benefit in pre-menopausal
women. This was due mainly to the influence of American trials that kept chemotherapy
in the control arms, which seemed to dilute the effects of tamoxifen.
AROMATASE INHIBITORS AND BREAST CANCER 347
9.15 Aromatase inhibitors and breast cancer

Details of studies
This was the first study to show a benefit in terms of disease-free survival and increased
tolerability of an aromatase inhibitor compared with tamoxifen in the adjuvant setting in
post-menopausal women.
Study references
Main study
ATAC Trialists’ Group (2002) Anastrozole alone or in combination with tamoxifen versus tamoxifen
alone for adjuvant treatment of postmenopausal women with early breast cancer: first results of the
ATAC randomised trial. Lancet, 359, 2131–9.
Related references
Forbes, J.F., Cuzick, J., Buzdar, A., et al. (2008) Effect of anastrozole and tamoxifen as adjuvant treatment
for early-stage breast cancer: 100-month analysis of the ATAC trial. Lancet Oncology, 9, 45–53.
Cuzick, J., Sestak, I., Baum, M., et al. (2010) Effect of anastrozole and tamoxifen as adjuvant treatment
for early-stage breast cancer: 10-year analysis of the ATAC trial. Lancet Oncology, 11, 1135–41.
Coombes, R.C, Kilburn, L.S., Snowdon, C.F. et al. (2007) Survival and safety of exemestane versus
tamoxifen after 2-3 years’ tamoxifen treatment (Intergroup Exemestane Study): a randomised
controlled trial. Lancet, 369, 559–70.
BIG 1-98 Collaborative Group, Mouridsen, H., Giobbie-Hurder, A., Goldhirsch, A., et al. (2009)
Letrozole therapy alone or in sequence with tamoxifen in women with breast cancer. New England
Study design
◆ A prospective RCT.
Randomization Tamoxifen 20mg alone versus anastrozole 1mg alone versus tamoxifen 20mg
and anastrozole 1mg
Inclusion criteria Post-menopausal women
Invasive early breast cancer
Completed primary therapy
Follow-up 33 months
Outcome measures
Primary endpoint
◆ Disease-free survival (DFS).
Secondary endpoint
◆ Overall survival (OS).
348 BREAST SURGERY
Results
◆ DFS was improved with anastrozole compared with tamoxifen (89.7% versus 87.4%,
p = 0.013). There was no difference between the tamoxifen and the combined
tamoxifen and anastrozole groups.
◆ The improved DFS was only seen in ER-positive patients.
◆ The incidence of contralateral breast cancer was significantly lower in the anastrozole
group than in the tamoxifen group (OR 0·42, 95% CI 0·22–0·79, p = 0·007).
◆ Anastrozole was better tolerated, with less toxicity than tamoxifen.
◆ Anastrozole was associated with an increased risk of musculoskeletal disorders and
fractures.
Critique
Although this was a large trial and the first to show a benefit in terms of DFS for an aroma-
tase inhibitor compared with tamoxifen, there were a number of issues which may have
partly compromised the results. Perhaps the most important criticism is the fact that in
this study of endocrine therapy, only 84% were known to be hormone receptor positive.
Although the overall numbers in the study were quite large, the early closure of the com-
bined arm reduced the total number of patients in the trial available for analysis.
Improvements in breast cancer treatment meant that the expected event rates were
actually lower than predicted when the trial was designed. As a consequence, the trial was
ultimately underpowered to show a benefit, especially in terms of survival. The patient
population of post-menopausal women also meant that OS outcomes were likely to be
diluted by non-breast cancer deaths in an ageing population. This effect was also evident
in the 100-month and 10-year follow-up papers, as a survival benefit was still not seen
(Forbes et al. 2008; Cuzick et al. 2010). In fact, although a number of aromatase inhibi-
tor adjuvant trials have now shown a significant improvement in DFS, to date none have
shown a significant benefit in OS in the ITT populations.
EXTENDED ADJUVANT ENDOCRINE THERAPY 349
9.16 Extended adjuvant endocrine therapy

Details of studies
Tamoxifen, an ER antagonist, is well established in the management of ER-positive breast
cancer. Traditionally, it has been administered for a period of 5 years after primary treat-
ment of ER-positive breast cancer, although the beneficial effect in terms of improving
DFS and OS is seen for up to 15 years from diagnosis.
Women who have been treated for 5 years with tamoxifen do develop both new tumours
and recurrent disease after cessation of treatment. Unfortunately, continuing tamoxifen
beyond 5 years does not appear to have any additional benefit and may result in a poorer
DFS, perhaps related to an increase in endometrial cancer and ischaemic events.
The introduction of the aromatase inhibitor letrozole into clinical practice raised the
possibility that it might be used after tamoxifen therapy to further reduce the risk of late
relapse and mortality. Because tamoxifen has a partial agonist–antagonist action, it was
hypothesized that letrozole might be especially effective against micrometastatic cells
that had become resistant to, or dependent on, tamoxifen. Prior to the publication of
this paper, letrozole had been shown to be beneficial in women with metastatic breast
cancer, including those whose disease had progressed despite tamoxifen therapy.
Study references
Main study
Goss, P.E., Ingle, J.N., Martino, S., et al. (2003) A randomized trial of letrozole in postmenopausal
women after five years of tamoxifen therapy for early-stage breast cancer. New England Journal of
Medicine, 349, 1793–802.
Related reference
Goss, P.E., Ingle, J.N., Martino, S. et al. (2005) Randomized trial of letrozole following tamoxifen as
extended adjuvant therapy in receptor-positive breast cancer: updated findings from NCIC CTG
MA.17. Journal of the National Cancer Institute, 97, 1262–71.
Study design
◆ Double-blind placebo-controlled trial. Also known as the (NCIC CTG) MA.17 trial.
Randomization Letrozole 2.5mg for 5 years versus placebo
Setting Multi-centre Canadian and American trial (1998–2002)
Inclusion criteria ≥50 years at start of adjuvant tamoxifen therapy
Any age if post-menopausal or had bilateral oophorectomy at initiation of
tamoxifen therapy
Pre-menopausal and <50 years at start of tamoxifen therapy but became
amenorrhoeic during chemotherapy or treatment with tamoxifen
Completed 5 years of treatment with tamoxifen
ER or progesterone-positive breast cancer
350 BREAST SURGERY
Exclusion criteria Cessation of tamoxifen >3 months prior to enrolment

Receiving systemic hormone replacement therapy
Follow-up 2.4 years
◆ The study was powered to calculate a 2.5% difference in DFS with 80% power at a
two-sided alpha level of 0.05. This assumed a 4-year DFS rate of 88% in the placebo
group and would require enrolment of 4800 women over a 4-year period with 2 years
of follow-up, accounting for 515 events. Two interim analyses were to be conducted,
after 171 and 342 events had occurred. The first interim analysis was performed
after 207 events (40% of the events required for the final analysis). At this point, the
independent data monitoring committee recommended termination of the trial
because of the early benefits of letrozole.
Outcome measures
Primary endpoints
◆ Disease-free survival.
Secondary endpoints
◆ Overall survival.
◆ Quality of life.
◆ Long-term safety.
Results
Placebo (n = 2594) Letrozole (n = 2593)
Recurrences 132 75*
Deaths 42 31
Estimated 4-year OS 94 96
Estimated 4-year DFS 87 93
◆ Letrozole significantly reduced the risk of local or metastatic recurrence or new

contralateral breast cancer compared with placebo (HR 0.76, 95% CI 0.43–0.75,
p >0.001).
◆ Letrozole was at least as effective among women with node-negative disease
(HR for recurrence or contralateral breast cancer, 0.47; p = 0.005) as among those
with node-positive disease (HR 0.60, p = 0.003).
◆ Toxic effects, mainly seen in the letrozole group, were generally mild and consisted
of hot flushes, arthritis, arthralgia, and myalgia.
◆ 4.5% of women in the letrozole group discontinued the study treatment because of
toxic effects compared with 3.6% of the women in the placebo group.
EXTENDED ADJUVANT ENDOCRINE THERAPY 351
◆ A subsequent publication with a median of 30 months of follow-up revealed similar

results. However, in lymph-node-positive patients, OS was statistically significantly
improved with letrozole (HR 0.61, 95% CI 0.38–0.98, p = 0.04)(Goss et al. 2005).
◆ The later study also showed a statistically non-significant reduction in the incidence
of cancer in the contralateral breast.
Conclusions
Extended adjuvant endocrine therapy with letrozole had the following outcomes.
◆ A significant improvement in DFS (related at least in part to a substantial reduction
in the rate of distant metastasis in the letrozole group).
◆ A reduction in deaths from breast cancer.
◆ Letrozole was equally effective in node-negative and node-positive disease.
◆ The study did not show an improvement in OS, but subsequent analysis showed
improved OS in node-positive patients.
◆ The efficacy of letrozole suggested that ER-positive breast cancer remained
dependent on oestrogen despite the development of ‘resistance’ to tamoxifen.
Critique
This is the largest trial of extended endocrine therapy. Unfortunately, even though the
study showed a clear benefit in terms of an improvement in DFS, it did not provide as
much information as had been hoped. After the independent data safety and monitoring
committee recommended early termination after the first interim analysis, patients in the
placebo arm were given the option of taking letrozole, and many patients accepted this
offer. This has resulted in significant difficulties in analysing the trial. A number of unan-
swered questions remain regarding long-term adjuvant endocrine therapy with letrozole.
In particular, information is lacking on the maximum duration of extended letrozole ther-
apy, the long-term toxicity, and the survival benefit, if any, of extended adjuvant therapy.
352 BREAST SURGERY
9.17 Use of Herceptin in breast cancer

Details of studies
Approximately 15% of breast cancers over-express HER2 and have a significantly higher
risk of relapse than comparable HER2-negative cancers. As a result, the monoclonal anti-
body trastuzumab, a humanized anti-HER2 monoclonal antibody, was developed in the
late 1990s and used as targeted therapy in HER2-positive breast cancer patients. Initial
studies suggested that it significantly improved survival in patients with HER2-positive
metastatic breast cancer, prompting the initiation of several large adjuvant trials. The first
of these was known as the HERA trial, which reported in 2005 and is the subject of this
landmark paper. This and subsequent studies have resulted in the introduction of trastu-
zumab into the standard management of this subtype of breast cancer.
Study references
Main study
Piccart-Gebhart, M.J., Procter, M., Leyland-Jones, B., et al. (2005) Trastuzumab after adjuvant
chemotherapy in HER2-positive breast cancer. New England Journal of Medicine, 353, 1659–72.
Related references
Slamon, D.J., Clark, G.M., Wong, S.G., et al. (1987) Human breast cancer: correlation of relapse and
survival with amplification of the HER-2/neu oncogene. Science, 235, 177–82.
Smith, I., Procter, M., Gelber, R.D., et al. (2007) 2-year follow-up of trastuzumab after adjuvant
chemotherapy in HER2-positive breast cancer: a randomised controlled trial. Lancet, 369, 29–36.
Gianni, L., Dafni, U., Gelber, R.D., et al. (2011) Treatment with trastuzumab for 1 year after adjuvant
chemotherapy in patients with HER2-positive early breast cancer: a 4-year follow-up of
a randomised controlled trial. Lancet Oncology, 12, 236–44.
Romond, E.H., Perez, E.A., Bryant, J., et al. (2005) Trastuzumab plus adjuvant chemotherapy for
operable HER2-positive breast cancer. New England Journal of Medicine, 353, 1673–84.
Slamon, D., Eiermann, W., Robert, N., et al. (2011) Adjuvant trastuzumab in HER2-positive breast
Joensuu, H., Kellokumpu-Lehtinen, P.L., Bono, P., et al. (2006) Adjuvant docetaxel or vinorelbine with
or without trastuzumab for breast cancer. New England Journal of Medicine, 354, 809–20.
Study design
◆ A prospective multi-centre RCT
Randomization Observation versus trastuzumab for 1 year versus trastuzumab for 2 years
Setting 478 institutions in 39 countries (2001–2005)
Inclusion criteria Women with HER2-positive early breast cancer
Node-positive or tumours >1cm if node negative
Patients had completed at least four cycles of neoadjuvant or adjuvant chemotherapy
Normal left ventricular ejection fraction
Follow-up 4 years
USE OF HERCEPTIN IN BREAST CANCER 353
◆ Trastuzumab was administered as an 8mg/kg loading dose and then a 6mg/kg IV

maintenance dose every 21 days.
Outcome measures
◆ Primary endpoint: DFS (defined as time from randomization to recurrence of breast
cancer at any site, ipsilateral or contralateral new breast cancer, second non-breast
malignant disease, or death from any cause).
◆ Secondary endpoints: cardiac toxicity, overall survival, and site of first DFS event
Results
◆ The first interim analysis, the subject of this landmark paper, showed that with a
median follow-up of 1 year, DFS events had occurred in 127/1694 patients in the
1-year trastuzumab arm and 220/1693 patients in the observation arm (HR 0.54, 95%
CI 0.43–0.67, p <0.0001).
◆ Superiority of 2 years of trastuzumab versus observation was also reported by the
interim data monitoring committee but no detailed information on this arm has been
released.
◆ A second ITT analysis, with median follow-up of 2 years, showed the DFS benefit to
be maintained (HR 0·64, 95% CI 0·54–0·76, p <0·0001) and also showed improved
overall survival (HR 0·66, 95% CI 0·47–0·91, p = 0·0115) (Smith et al. 2007).
◆ The third and most recent analysis, with median follow-up of 4 years, again
confirmed a DFS benefit (HR 0·76, 95% CI 0.66–0·87, p <0·0001), but the significant
improvement in overall survival present in the earlier analysis was no longer evident
(Gianni et al. 2011).
◆ The incidence of cardiac toxicity was low, but there was a significant drop in left
ventricular ejection fraction (5% versus 1%) and symptomatic congestive cardiac
failure (2% versus 0%) in the trastuzumab group compared with control patients
Conclusion
The addition of trastuzumab to adjuvant chemotherapy significantly improves outcomes
in women with HER2-positive early breast cancer.
Critique
The HERA trial discussed here and the NSABP-B31 and NCCTG-N9831 trials which
were published back-to-back in the New England Journal of Medicine clearly showed that
breast cancer recurrence could be significantly reduced by the addition of trastuzumab to
standard adjuvant chemotherapy.
The initial HERA publication resulted from the identification of a highly significant
improvement in DFS in favour of 1 or 2 years of trastuzumab versus observation in a pre-
planned interim analysis carried out after 475 DFS events had occurred. This prompted
the interim data monitoring committee to recommended early release of the results of
the observation and 1-year trastuzumab arms, although no detailed information on the
354 BREAST SURGERY
2-year arm was released. In addition, patients in the observation arm who had remained
disease free were permitted to cross over and receive trastuzumab. Subsequently, 52% of
patients from the observation arm crossed over to receive trastuzumab and this makes
the interpretation of overall survival in the HERA trial difficult. Nevertheless, an overall
survival benefit for adjuvant trastuzumab has been demonstrated in other large trials such
as NSABP-B31/N9831 and BCIRG006.
Although trastuzumab is now a standard component of adjuvant therapy for women
with HER2-positive disease the optimal regimen is not yet resolved. For example, in
HERA, trastuzumab was delivered after completion of chemotherapy, whereas the N9831
study prospectively randomized patients to receive trastuzumab either after or concur-
rently with chemotherapy. This study showed a trend towards better DFS with concurrent
administration.
The optimum duration of trastuzumab is also unknown. The results of 2 years ver-
sus 1 year of trastuzumab in HERA are not yet reported. In the meantime several trials
are investigating shorter treatment periods (e.g. 6 months versus 1 year). The smaller
FinHER study delivered only 9 weeks of trastuzumab, which was administered concur-
rently at the start of chemotherapy, an approach which is now under investigation in a
large phase 3 trial.
NEOADJUVANT CHEMOTHERAPY FOR OPERABLE BREAST CANCER 355
9.18 Neoadjuvant chemotherapy for operable breast cancer

Details of studies
Pre-operative chemotherapy was initially developed in the 1970s with the intention of
improving outcomes and operability in locally advanced, inoperable, and inflammatory
breast cancer. Clinicians subsequently speculated that pre-operative treatment might also
have a role in patients with operable breast cancer where downsizing large tumours might
enable conservative surgery rather than mastectomy. However, there was concern that
such an approach might compromise long-term or disease-free survival by delaying sur-
gery which was also likely to be less radical than would otherwise have been performed.
The paper discussed here was the first large-scale study to address these issues.
Study references
Main study
Fisher, B., Bryant, J., Wolmark, N., et al. (1998) Effect of preoperative chemotherapy on the outcome of
women with operable breast cancer. Journal of Clinical Oncology, 16, 2672–85.
Related reference
Bear, H.G., Anderson, S., Smith, R., Geyer, C.E., et al. (2006) Sequential pre operative or post operative
docetaxel added to postoperative doxorubicin and cyclophosphamide for operable breast cancer
NSABP B-27. Journal of Clinical Oncology, 24, 2019–27.
Study design
◆ A prospective RCT, known as the National Surgical Adjuvant Breast and Bowel
Project (NSABP) B-18 trial.
Randomization Pre-operative versus post-operative adriamycin and cyclophosphamide
Setting Multi-centre trial based in the USA and Canada
Inclusion criteria Women with primary operable breast cancer
Follow-up 5 years
◆ A clinical assessment of the tumour was performed after completion of

chemotherapy in the pre-operative treatment group. This was graded as complete,
partial, or no clinical response.
◆ A separate pathological assessment was performed in tumours where a complete
clinical response had been recorded. In these patients, the pathological response was
categorized as either a complete or incomplete pathological response, depending on
the presence of viable invasive tumour cells. A proportional hazards model was then
used to investigate the relationship between outcome and tumour response.
356 BREAST SURGERY
Outcome measures
Primary outcome
◆ Disease-free survival (DFS).
Secondary outcomes
◆ Distant disease-free survival (DDFS).
◆ Overall survival (OS).
◆ Complete pathological response.
◆ Breast conservation rates (lumpectomy and radiotherapy versus mastectomy).
◆ Local recurrence rates.
Results
◆ There was no difference in DFS, DDFS, or OS between the groups.
◆ Women who underwent pre-operative chemotherapy were more likely to undergo
breast conservation than mastectomy (67.8% versus 59%).
◆ Ipsilateral breast recurrence was similar in the groups (7.9% versus 5.8%).
◆ Outcome was better in women who showed a pathological complete response with
relapse-free survival rates of 85.7% versus 76.9%.
Critique
This large randomized trial demonstrated the safety of using pre-operative chemotherapy
in patients with operable breast cancer. A later study in which docetaxel was added to
adriamycin and cyclophosphamide (AC), either before or after surgery, resulted in a sig-
nificant reduction in local recurrence as a first event (Bear et al. 2006). With this regimen
the complete clinical response rate doubled, which was also associated with an improve-
ment in overall survival. The addition of the taxane to AC did not lead to an improvement
in DFS or OS, but this may be because some patients received tamoxifen concurrently
with docetaxel.
Neoadjuvant chemotherapy regimens have become established as part of mainstream
practice, and polychemotherapy regimens which include a taxane can achieve clinical
response rates of 60–90% and pathological complete response rates ranging from 10%
to 30%.
Chapter 10
Endocrine surgery
10.1 Laparoscopic adrenalectomy 358
10.2 Operative monitoring of parathyroid gland hyperfunction 360
10.3 Recurrent laryngeal nerve monitoring in thyroid surgery 362
10.4 Management of the N0 neck in T1c papillary thyroid carcinoma 365
358 ENDOCRINE SURGERY
10.1 Laparoscopic adrenalectomy

Details of studies
The first laparoscopic adrenalectomy for benign disease was performed in 1992 by Gagner
et al. Since then it has become widely adopted for benign and malignant endocrine condi-
tions (Conn’s syndrome, Cushing’s syndrome, phaeochromocytoma, incidental adenoma,
and adrenal malignancy).
Study references
Main study
Gagner, M., Lacroix, A., and Bolté, E. (1992) Laparoscopic adrenalectomy in Cushing’s syndrome and
phaeochromocytoma. New England Journal of Medicine, 327, 1033.
Related references
Gagner, M., Pomp, A., Heniford, B.T., Pharand. D., and Lacroix, A. (1997) Laparoscopic
adrenalectomy: lessons learnt from 100 consecutive procedures. Annals of Surgery, 226, 238–47.
Brix, D., Allolio, B., Fenske, W., et al. (2010) Laparoscopic versus open adrenalectomy for
adrenocortical carcinoma: surgical and oncologic outcome in 152 patients. European Urology, 58(4):
609–615.
Study design
The original paper by Gagner et al. (1992) described three patients undergoing
laparoscopic adrenalectomy for Cushing’s disease/syndrome and a right-sided pha-
eochromocytoma (level 4 evidence). The follow-up period was short (up to 6 days,
post-operatively).
The same group subsequently published a much larger series of 100 patients. The
overwhelming majority (97%) were for benign endocrine conditions and the remainder
for carcinoma. The follow-up period ranged from 1 to 44 months. Outcome measures
included conversion to open surgery, length of stay, morbidity, mortality, and hormonal
recurrence.
Results
◆ The original paper was primarily a feasibility study. The laparoscopic technique was
described with an operating time of 2–3 hours.
◆ The three patients described in the paper recovered well from operation and were
discharged 3–6 days after operation.
◆ The subsequent larger series reported a mean operating time of 123 minutes (range
80–360 minutes) with no mortalities.
◆ The morbidity rate was 12% and complications included DVT, haematoma, anaemia,
UTI, subdural haematoma, and acute cholecystitis.
◆ There were three conversions to open surgery.
LAPAROSCOPIC ADRENALECTOMY 359
◆ At the beginning of their experience, the length of stay was 3 days but this fell to
2.4 days by the end of the study.
◆ In the study follow-up period, two patients developed renovascular hypertension and
none experienced a hormonal recurrence.
Conclusions
Gagner et al. concluded that laparoscopic adrenalectomy is safe, effective, and associated
with a rapid recovery and low complication rate. Venous thromboprophylaxis is strongly
recommended to avoid potential post-operative morbidity.
Critique
At the time, it was difficult for other groups to be convinced of the benefits of laparoscopic
surgery for removing adrenal glands up to 15cm in size as Gagner did. Since Gagner’s
work, many case–control studies have demonstrated the benefits of laparoscopic adrenal-
ectomy versus open surgery in terms of blood loss, analgesic requirements, post-operative
complications, hospital stay, and earlier return to normal activity. The original study
primarily focused on adrenalectomy for benign endocrine conditions. More recently, Brix
et al. have demonstrated no difference in oncological outcome between open and laparo-
scopic adrenalectomy for carcinoma.
10.2 Operative monitoring of parathyroid gland hyperfunction

Details of studies
Successful treatment of primary hyperparathyroidism traditionally relied upon the skill
and experience of the operating surgeon. Surgical management consisted of bilateral neck
exploration, identifying all four parathyroid glands and removing the abnormal gland
or glands. Frozen section to distinguish between an adenoma and hyperplasia was not
accurate, and a successful outcome could not be assured until the post-operative calcium
was obtained. Using this traditional approach, persistent hyperparathyroidism occurred
in up to 10% of patients.
Since parathyroid hormone (PTH) has a half-life of minutes, an intra-operative assay
to confirm a drop in serum PTH might offer an attractive means of checking that the
abnormal gland had been identified. Prior to the publication of this paper, the time taken
to perform a PTH assay was too long to be of any practical intra-operative use.
Study references
Main study
Irvin, G.L., Dembrow, V.D., and Prudhomme, D.L. (1991) Operative monitoring of parathyroid gland
hyperfunction. American Journal of Surgery, 162, 299–302.
Related references
Nussbaum, S.R., Thompson, A.R., Hutcheson, K.A., Gaz, R.D., and Wang, C.A. (1988) Intraoperative
measurement of parathyroid hormone in the surgical management of hyperparathyroidism. Surgery,
104, 1121–7.
Flentje, D., Schmidt-Gayk, H., Fischer, S., et al. (1990) Intact parathyroid hormone in primary
hyperparathyroidism. British Journal of Surgery, 77, 168–72.
Westerdahl, J. and Bergenfelz, A. (2007) Unilateral versus bilateral neck exploration for primary
hyperparathyroidism: 5-year follow-up of a randomized controlled trial. Annals of Surgery, 246,
976–81.
Russell, C.F., Dolan, S.J., and Laird, J.D. (2006) Randomized clinical trial comparing a scan-directed
unilateral versus bilateral cervical exploration for primary hyperparathyroidism due to solitary
adenoma. British Journal of Surgery, 93, 418–21.
Study design
◆ A prospective single-institution consecutive case series.
◆ The basis for this paper was the modification of a commercially available
immuno-radiometric assay for intact PTH. The modification enabled a short 15min
turnaround time, thus allowing it to be used intra-operatively.
OPERATIVE MONITORING OF PARATHYROID GLAND HYPERFUNCTION 361
Class of evidence 4
Intervention Quick PTH assay and the standard immuno-radiometric assay performed at:
(i) induction
(ii) after mobilization of the enlarged parathyroid gland
(iii) 10min after excision of the enlarged gland
(iv) 20min after excision of the enlarged gland
(v) every 10min after excision of further enlarged glands
◆ Eighteen patients with primary hyperparathyroidism and three with recurrent

hypercalcaemia following previous neck exploration.
Outcome measures
◆ Correlation between the quick PTH assay and the standard 24-hour
immunoradiometric assay.
◆ The ability of the quick PTH assay to determine that all hyperfunctioning
parathyroid tissue had been removed at the time of surgery.
Results
◆ There was good correlation between the quick PTH and the standard 24-hour
immuno-radiometric assay (correlation 0.967), although the quick test appeared to
lose sensitivity at lower concentrations.
◆ During surgery fewer than four glands were identified in 53% of patients.
◆ The intra-operative quick PTH assay provided evidence that all overactive
parathyroid tissue had been removed in 19 of 21 patients, and the assay predicted
persistent hypercalcaemia in the remaining two patients.
Critique
Although this is a small patient series, this paper introduced a ‘quick’ intra-operative PTH
assay that could be used to determine that all hyperfunctioning parathyroid tissue had
been removed at the time of surgery and identify those patients who do not have a solitary
parathyroid adenoma.
Intra-operative PTH measurement combined with pre-operative localization has led
to the abandonment of bilateral neck exploration and the adoption of a more focused
approach to parathyroidectomy.
10.3 Recurrent laryngeal nerve monitoring in thyroid surgery

Details of studies
The standard method of preserving the recurrent laryngeal nerve (RLN) during thyroid
surgery is by direct visualization and careful dissection. Case series have suggested that
intra-operative nerve monitoring (IONM) may reduce the incidence of post-operative
recurrent laryngeal nerve paralysis. This paper by Barczyński et al. was designed to test
this hypothesis.
Study references
Main study
Barczyński, M., Konturek, A., and Cichoń, A. (2009) Randomized clinical trial of visualization versus
neuro-monitoring of recurrent laryngeal nerves during thyroidectomy. British Journal of Surgery, 96,
240–6.
Related references
Dralle, H., Sekulla, C., Lorenz, K., et al. (2008) Intraoperative monitoring of the recurrent laryngeal
nerve in thyroid surgery. World Journal of Surgery, 32, 1358–66.
Randolph, G.W., Dralle, H., International Intraoperative Monitoring Study Group, et al. (2011)
Electrophysiologic recurrent laryngeal nerve monitoring during thyroid and parathyroid surgery:
international standards guideline statement. Laryngoscope, 121, S1–16.
Study design
Randomization Neuro-monitoring of RLNs versus visualization alone
Number of surgeons 3
Inclusion criteria Bilateral neck surgery for thyroid disorders
Exclusion criteria Previous thyroid surgery
Unilateral thyroid pathology eligible for a minimally invasive approach
Mediastinal goitre
Pre-operative RLN palsy
Pregnancy or lactation
Age <18 years
ASA grade 4
Follow-up Assessment on post-operative day 2
If RLN paresis, an additional examination at 2 weeks and 1, 2, 4, 6, and
12 months post-operatively, or until vocal cord function had recovered
RECURRENT LARYNGEAL NERVE MONITORING IN THYROID SURGERY 363
◆ Two anaesthetists followed a strict protocol with regard to pre-medication,

intubation techniques, and muscle relaxants
◆ The surgical technique was carried out by three endocrine surgeons using a standard
Kocher’s skin incision followed by a standard dissection technique.
◆ In half the patients the RLN was identified by direct visualization only. The
nerve-monitoring group similarly underwent direct visualization of the nerve.
In addition, nerve function was monitored using the Neurosign™ 100 system
(Magstim, Carmarthenshire, Wales). RLN function was assessed by ‘impaling’ the
vocalis muscle (one of the intrinsic muscles in the larynx) through the cricothyroid
membrane, with the neutral electrode placed in the sternocleidomastoid muscle. The
circuit was assessed by dissection of the vagus nerve and subsequent stimulation.
Outcome measures
Primary endpoint
◆ RLN injury events—temporary paresis, permanent paralysis (paralysis present more
than 1 year post-operatively).
Secondary endpoint
◆ Assessment of the utility of neuro-monitoring in identifying RLN anatomical
variations predicting post-operative vocal cord dysfunction.
Results
◆ There were no significant differences between the two groups with regard to age,
sex, disease, surgical procedure, and thyroid size or volume. The duration of the
procedure was 9 minutes longer in the group undergoing RLN monitoring.
◆ Transient and permanent RLN injuries were found in 3.8% and 1.2% respectively
of nerves without RLN monitoring (p = 0·011) and 1.9% and 0.8% respectively of
nerves with RLN monitoring (p = 0·368).
◆ The incidence of RLN injury was 2.3% (p = 0.007) lower in the RLN monitoring group.
◆ Significantly more branches of the RLN were identified in the nerve-monitoring
group. Additional branches were mainly identified near Berry’s ligament (p <0.001)
and the inferior thyroid artery.
◆ The positive predictive value of stimulating the RLN and the RLN functioning
post-operatively was low; 37.8% for indirect stimulation via the vagus and 25.5% for
direct stimulation of the RLN.
Conclusions
Intra-operative nerve monitoring reduces the incidence of temporary post-operative
recurrent laryngeal nerve paralysis.
Critique
Although direct visualization and careful dissection is critical in order to preserve the
recurrent laryngeal nerve during thyroid surgery, there has been increasing interest in the
use of intra-operative nerve monitoring. The authors of this landmark paper attempted
to quantify whether monitoring provided any additional measurable benefit during rou-
tine thyroid surgery. Given the relatively rare incidence of temporary paralysis, and the
even rarer incidence of permanent RLN paralysis, it was always going to be difficult to
carry out a study with sufficient power to demonstrate a significant difference between
the two treatment strategies. Although a multi-centre trial is the ideal study format to
recruit large patient numbers, standardization of the surgical and anaesthetic approach
in such studies can be difficult. Barczyński et al. limited surgery to three surgeons and
two anaesthetists, and used a standard surgical and anaesthetic approach. Although the
study did not show a significant difference in permanent RLN palsy, it did demonstrate a
reduction in the temporary RLN palsy rate when nerve monitoring was used. It is likely
that they were not able to demonstrate a difference in permanent palsy rates because of a
much lower incidence of permanent injury.
Although the authors of the paper are to be complimented for performing this large-
scale single-institution RCT, the study is not without its weaknesses. The most important
of these relates to the failure to document what criteria were used to assess cord function
on indirect laryngoscopy. Although one assumes that patients were recorded as having
either full vocal cord function or complete vocal cord palsy, such an assessment can be
subjective and potentially open to observer error. The second issue relates to the tech-
nique of intra-operative nerve monitoring. In this study the technique involved ‘impaling’
the intrinsic muscles of the larynx through the cricothyroid membrane. This is a par-
ticularly invasive pre-operative technique compared with current techniques where the
RLN monitoring electrodes are incorporated into the endotracheal tube. This technique
appears to be more readily acceptable to surgeons and anaesthetists involved in thyroid
surgery as it is easily positioned under direct visualization once the patient is in the appro-
priate position for the procedure.
Although the current study assessed the benefit of RLN monitoring during routine
thyroid surgery, it is likely that the technique offers the greatest benefit in more challeng-
ing cases for advanced malignancy or recurrent disease. However, it is only through the
routine use of nerve monitoring that familiarity with the rigours of the technique will be
achieved. Therefore it would seem appropriate that RLN monitoring should be carried
out in all cases of thyroid surgery irrespective of the degree of difficulty to ensure that the
best chance of preservation of RLN function is achieved.
MANAGEMENT OF THE N0 NECK IN T1C PAPILLARY THYROID CARCINOMA 365
10.4 Management of the N0 neck in T1c papillary

thyroid carcinoma
Details of studies
Papillary thyroid carcinoma (PTC) is the most common thyroid malignancy. Papillary
thyroid microcarcinoma (PTMC) (primary tumour <10mm) was thought to have a
favourable prognosis when compared with larger primary tumours. Therefore the role of
central-compartment neck dissection as part of the primary surgical treatment of PTMC
was controversial. The landmark papers discussed in this section identified potential stag-
ing and therapeutic benefits to primary surgical staging of the central compartment, in
combination with total thyroidectomy, in these hitherto ‘low-risk patients’.
Study references
Roh, J.L., Kim, J., and Park, C.I. (2008) Central cervical nodal metastasis from papillary thyroid
microcarcinoma: pattern and factors predictive of nodal metastasis. Annals of Surgical Oncology, 15,
2482–6.
Pellegriti, G., Scollo, C., Lumera, G., et al. (2004) Clinical behavior and outcome of papillary thyroid
cancers smaller than 1.5cm in diameter: study of 299 cases. Journal of Clinical Endocrinology and
Metabolism, 89, 3713–20.
Study design
Roh et al.
◆ A single-institution retrospective cohort study.
Intervention Total thyroidectomy and central neck dissection in patients
with PTMC
Setting Large teaching hospital, South Korea
Inclusion criteria Patients with a PTMC
Lesion detected on ultrasound and confirmed on FNA
cytology
Parameters used to assess primary Size
tumour Multifocality
Bilaterality
Extracapsular or lymphovascular invasion
Distribution of nodal metastases Ipsilateral
Para-tracheal and pre-tracheal
Superior mediastinal
Contralateral para-tracheal
Pellegriti et al.
◆ A retrospective single-institution cohort study.
◆ The aim was to determine the recurrence rate of small PTCs following
combined-modality treatment.
Intervention Near-total or total thyroidectomy in 292 patients;
lobectomy in 7 patients
Patients with clinical signs of a PTC also underwent
a central-compartment neck dissection
Setting Large Italian teaching hospital
Inclusion criteria Tumours ≤1.5cm
Parameters used to assess primary tumour Size (≤1cm versus 1.1–1.5cm)
Multifocality
Bilaterality
Extracapsular or lymphovascular invasion
Outcome measures Factors associated with the development of persistent or
recurrent disease (multivariate analysis)
◆ Clinical signs were present, i.e. non-incidental presentation in 148 cases. In the
remaining 151 patients the diagnosis was made at incidental thyroidectomy.
Results
Roh et al.
Characteristics of primary tumour Size 6.9 ± 2.2mm
Bilaterality 15 (20.8%)
Extracapsular invasion 28 (38.9%)
Multifocality 26 (36.1%)
Lymphovascular invasion 6 (8.3%)
Metastasis to central compartment Para-tracheal 27 (37.5%)
Pre-tracheal 8 (11.1%)
Superior mediastinal 4 (5.6%)
Contralateral para-tracheal 1 (1.4%)
MANAGEMENT OF THE N0 NECK IN T1C PAPILLARY THYROID CARCINOMA 367
Pellegriti et al.
Characteristics of primary tumour Size ≤1cm 187
Size 1.1–1.5cm 112
Multifocal disease 95 (31.7%)
Bilateral disease 55 (18.4%)
Extracapsular invasion 60 (20%)
Lymphovascular invasion 90 (30.1%)
Persisting or recurrent disease 77 (25.7%)
Risk factors for recurrent or Non-incidental PTC
persistent disease Lymph node metastases at presentation
Bilateral tumours
◆ The development of distant metastases was associated with the presence of nodal
metastases at presentation.
◆ Tumour size (≤1.0cm versus 1.1–1.5cm) was not a risk factor for developing
recurrent disease.
Conclusions
◆ PTMC is associated with a high rate of central lymph node metastasis to ipsilateral
and pre-tracheal nodes.
◆ The presence of lymph node metastases and bilateral disease are the strongest
predictors for recurrent or persisting disease.
Critique
Both these papers reported the outcomes of patients who on pre-operative assessment
were thought to have ‘very-low-risk’ disease. Although high-resolution ultrasound and
FNA biopsy are increasingly used to diagnose and stage thyroid carcinomas, these papers
clearly demonstrate problems with pre-operative staging of PTC, given the frequency of
multifocal disease and its propensity to metastasize early to the regional lymph nodes. It is
clear that, even in expert hands, multifocal disease and central-compartment lymph node
metastases can be missed.
Although it was previously felt that small PTCs (T1 <1cm) could be satisfactorily
treated with hemithyroidectomy alone, these two papers clearly demonstrated that even
patients with these low-volume tumours have a >20% chance of having bilateral thyroid
disease.
Given that the presence of lymph node metastasis carries significant staging and treat-
ment implications, many surgeons now consider performing a concomitant staging
central-compartment neck dissection at the time of thyroidectomy for small PTCs. Such
a strategy has the advantage that both the thyroid gland and the central compartment
can be accurately staged, which in turn allows better targeting of post-operative radio-
iodine therapy. This has been shown to reduce local recurrence rates. Although some have
advocated a unilateral central-compartment clearance for small PTCs, the two papers
discussed here demonstrate that multifocal and bilateral involvement are not uncommon.
Although central-compartment dissection improves staging, there is no clear evidence
that it has an impact on disease-specific mortality in this very treatable disease. Therefore
some have also argued that the increased risk of complications (hypocalcaemia and vocal
cord paresis or palsy) following central-compartment neck dissection far outweighs any
oncological benefit.
Chapter 11
Trauma surgery
11.1 Systems of trauma care 370
11.2 Damage control surgery 373
11.3 Immediate versus delayed fluid resuscitation for hypotensive patients
with penetrating torso injuries 376
11.4 The effect of tranexamic acid in the bleeding trauma patient 379
370 TRAUMA SURGERY
11.1 Systems of trauma care

Details of studies
During the 1960s the outcome following major trauma was judged to be poor, even in
developed regions such as the North America. Data collected by the National Research
Council in 1966 formed the basis for the report Accidental Death and Disability: The
Neglected Disease of Modern Society. There was a growing recognition that a significant
number of trauma-related deaths were potentially preventable. It was suggested that
a key factor in reducing preventable mortality was the development of a trauma system:
a regionalized and specialized process of care. This paper studied the outcome of trauma
patients managed within different systems in two adjacent counties in California.
Study references
Main study
West, J.G., Trunkey, D.D., and Lim, R.C. (1979) Systems of trauma care. A study of two counties.
Archives of Surgery, 114, 455–60.
Related reference
West, J.G., Cales, R.H., and Gazzaniga, A.B. (1983) Impact of regionalization. The Orange County
experience. Archives of Surgery, 118, 740–4.
Study design
◆ A retrospective review of the results of trauma care in two different healthcare
systems.
◆ The review included death certificates, coroner’s reports, and autopsy data.
◆ Records from the San Francisco deaths were available for review. Trauma patients in
San Francisco County were admitted to a single trauma centre.
◆ Trauma patients in Orange County were admitted to the nearest of 39 hospitals.
◆ Inclusion criteria: 100 consecutive motor vehicle deaths from each county were
analysed if the injured had survived to reach a medical facility.
◆ Exclusion criteria: deaths in the field and deaths during transport were excluded.
◆ The Injury Severity Score (ISS) was calculated for each patient in order to provide
a numerical comparison of severity and to allow prediction of mortality.
◆ Deaths were subdivided into those that were primarily due to central nervous system
(CNS) injury, and those that were not.
Outcome measures
◆ Deaths in the non-CNS-related group were classified as clearly preventable,
potentially preventable, or not preventable
◆ Deaths in the CNS-related group were analysed with respect to neurosurgical
intervention and missed injury
SYSTEMS OF TRAUMA CARE 371
Results
◆ After excluding deaths in those patients who had been referred to the counties after
initial care elsewhere, 90 cases from Orange County and 92 cases from San Francisco
County were analysed.
◆ There were 30 non-CNS-related deaths in Orange County and 16 in San Francisco
County. There were 60 CNS-related deaths in Orange County and 76 in San
Francisco County.
◆ The average ISS for both non-CNS-related deaths and CNS-related deaths was
significantly higher in San Francisco County. The patients in San Francisco County
were also significantly older.
Orange County San Francisco County

Non CNS-related deaths 30 patients 16 patients
Clearly preventable 11 0
Potentially preventable 11 1
Not preventable 8 15
CNS-related deaths 60 patients 76 patients
Neurosurgical procedure 12 55
Undiagnosed space-occupying lesion 8 0
Other untreated injury 9 0
Conclusions
Patients managed within a trauma centre had a lower preventable mortality, received
more aggressive neurosurgical input, and had fewer missed or untreated injuries than
those managed outwith a dedicated trauma system.
Critique
This paper added evidence to support a regionalized system of trauma care. Orange
County medical facilities received younger and less severely injured patients than those
in San Francisco County. Despite this, the preventable mortality in the non-CNS-related
deaths was significantly higher in Orange County. In addition, the patients with CNS-
related deaths in Orange County were much less likely to receive neurosurgical interven-
tion and had a higher rate of missed CNS injuries.
There are limitations to this study. It is a retrospective review with inherent issues
regarding accuracy and quality of data, although the authors attempted to minimize the
impact of this by reviewing several sources of information. Of note, however, they did
not appear to have access to the case records of the fatalities in the Orange County cohort
and relied on post-mortem data such as autopsy reports. There were also important dif-
ferences in the geography of the counties. Orange County covered 2033km2 and had a
population of 1.7 million. San Francisco County covered 127 km2 and had a population
372 TRAUMA SURGERY
of 667,000, which reached 1.6 million during the working day, therefore San Francisco
County was significantly smaller and had one well-staffed and centrally located trauma
unit with senior medical staff either resident or immediately available. Rather than
detracting from the author’s case for a trauma system, these differences in service provi-
sion only serve to enhance it.
It is now well established that an effective trauma system includes appropriate pre-
hospital care, and rapid transport and delivery to a hospital with the capability for
definitive management of injuries, which is not necessarily the closest institution. The dif-
ferences in outcome illustrated in this paper are partly a function of volume, experience,
and pattern recognition all of which facilitate rapid and appropriate decision-making.
However, at the time of publication trauma care in many parts of the developed world was
at best haphazard, and trauma systems were very much in their infancy.
This groundbreaking paper provided a comprehensive report on the deficiencies of
care for the injured patient managed outwith a trauma system. It led to changes in practice
by offering irrefutable data demonstrating a reduction in preventable mortality for those
managed in a dedicated regional trauma centre. These data continue to inform trauma
care more than 40 years after this paper was first published.
A subsequent paper (West et al. 1983) documented the impact of regionalization of
trauma care in Orange County 5 years after it was introduced. The data demonstrated a
marked reduction in preventable mortality and underlined the significance of the original
work by Trunkey and colleagues.
DAMAGE CONTROL SURGERY 373
11.2 Damage control surgery

Details of studies
The traditional approach to the patient with penetrating abdominal injury has included
definitive laparotomy with all injuries repaired during the initial surgical encounter.
Whilst this management strategy worked well in patients with preserved physiological sta-
bility, it was apparent that it was more challenging to apply to the exsanguinating patient
with combined major vascular and visceral injury. Ten years prior to this paper, Stone
et al. described their experience in abbreviated laparotomy in patients who developed
coagulopathy during the initial surgery. The aim of this study was to compare the newly
described three-phase approach of damage control surgery to definitive laparotomy.
Study references
Main study
Rotondo, M.F., Schwab, C.W., McGonigal, M.D., et al. (1993) ‘Damage control’: an approach for
improved survival in exsanguinating penetrating abdominal injury. Journal of Trauma, 35, 375–82.
Related references
Stone, H.H., Strom, P.R., and Mullins, R.J. (1983) Management of the major coagulopathy with onset
during laparotomy. Annals of Surgery, 197, 532–5.
Study design
◆ A retrospective review comparing the outcome following ‘definitive laparotomy’
(DL, n = 22) with ‘damage control laparotomy’ (DCL, n = 24) for penetrating
abdominal trauma.
◆ The pre-operative resuscitation phase was similar in both groups.
◆ During the operative phase the DL group had control of bleeding and contamination,
and an attempt to complete all technical aspects of the surgery
◆ During the operative phase the DCL group had a three-stage surgical approach.
Phase I comprised control of bleeding and contamination using non-standard and
often temporary techniques such as clamping vascular structures and ligation of
hollow visceral injuries. When signs of intra-operative coagulopathy developed,
the abdomen was packed and closed, and the patient was transferred to ICU for
phase II which involved ongoing resuscitation. When physiological stability had been
restored, the patient returned to the operating theatre for definitive management of
their injuries (phase III).
◆ The Penetrating Abdominal Trauma Index (PATI) was used to define injury severity.
◆ The data were retrospectively reviewed and comprised the PATI for each injury, the
probability of survival and the actual survival, time spent in resuscitation, operating
room, and intensive care, and physiological parameters during the laparotomy and
post-operative complications.
374 TRAUMA SURGERY
Outcome measures
◆ Mortality.
◆ Physiological and resuscitation parameters.
Results
◆ The data were first analysed with respect to the overall group, and then a subset
analysis of the 22 patients in the Maximum Injury Subset was performed based on
their Revised Trauma Score and Injury Severity Score.
◆ Survival between the DL and DCL groups was similar (55% versus 58%).
◆ There were no significant differences in physiological parameters between the DL
and DCL groups during the resuscitation phase.
◆ There were no differences in blood/fluid requirements between the groups during the
operative phase.
◆ In the Maximum Injury Subset patients, 1/9 DL patients survived (actual survival
11%, predicted survival 67%) compared with 10/13 DCL (actual survival
77%, predicted survival 81%).
Conclusions
‘Damage control laparotomy’ may offer a survival benefit over conventional surgery in
the exsanguinating patient with multiple penetrating abdominal vascular and visceral
injuries.
Critique
This was the first paper to use the term ‘damage control’ in a surgical context and, more
importantly, the first to describe it as a focused three-phase strategy in the management
of the exsanguinating patient with penetrating abdominal trauma. The association of the
trauma triad of coagulopathy, hypothermia, and acidosis with poor outcome was well
documented, but surgical enthusiasm for definitive repair of every injury at the initial
laparotomy continued to result in potentially avoidable mortality. Prior to this study
there had been reports of abbreviated surgery in the context of hepatic injury, but little to
support this strategy outwith this setting. This paper introduced the concept of ‘damage
control surgery’ as a phased approach comprising rapid control of bleeding and contami-
nation at the first laparotomy, restoration of physiology in the ICU, and return to theatre
for definitive management of abdominal injury.
There are limitations inherent in this study which merit discussion. This was a retro-
spective case-note review comparing ‘definitive laparotomy’ with ‘damage control lapa-
rotomy’ in a non-randomized setting. In the overall analysis this approach did not confer
any survival advantage. The study population was small (n = 46) and, as is frequently
the case in the trauma literature, relatively heterogeneous in terms of injury patterns,
particularly with respect to injured vascular structures. There were ten deaths in the
DAMAGE CONTROL SURGERY 375
DL group, all of which occurred prior to leaving the operating room and were attributed
to failure to control haemorrhage or refractory shock. Therefore the cause of death in this
subgroup can be stated with reasonable confidence. However, in the DCL group, seven of
the ten deaths occurred in the ICU and the cause of death in these patients is less clear.
A major criticism of this paper, and therefore the surgical approach, is the failure to mea-
sure intra-abdominal pressure at any time. The authors describe closing the abdomen as
part of the abbreviated laparotomy, and it may be that this has contributed to the refrac-
tory shock, and subsequent mortality, described in this subgroup of patients. This concept
has now evolved from the author’s original description to the more accepted practice of
laparostomy in an effort to avoid abdominal compartment syndrome. It is also important
to recognize that this work refers only to penetrating trauma and did not include patients
with blunt injury.
Despite these limitations, this is a key paper in the trauma literature. It coined the
term ‘damage control’ and brought together various critical concepts as part of an overall
resuscitation strategy. It offered a phased approach to the management of this challenging
group of patients and recognized the role of limited surgery as an adjunct to restore physi-
ology, in place of lengthy surgery to definitively restore anatomy. As a result of this work,
‘damage control surgery’ has been successfully extrapolated into the emergency general
surgery population and continues to be refined in the trauma patient in established or
impending physiological crisis.
376 TRAUMA SURGERY
11.3 Immediate versus delayed fluid resuscitation for

hypotensive patients with penetrating torso injuries
Details of studies
For several decades the standard management of hypotensive trauma patients has included
aggressive pre-operative intravenous infusion of isotonic fluids. There have been increas-
ing concerns based on experimental work that this approach may be detrimental in the
setting of uncontrolled bleeding. The objective of this study was to test the hypothesis that
the outcome of hypotensive patients with penetrating torso trauma would be improved if
fluid resuscitation was restricted until the time of operative intervention.
Study references
Main study
Bickell, W.H., Wall, M.J., Jr, Pepe, P.E., et al. (1994) Immediate versus delayed fluid resuscitation for
hypotensive patients with penetrating torso injuries. New England Journal of Medicine, 331, 1105–9.
Related references
Martin, R.R., Bickell, W.H., Pepe, P.E., Burch, J.M., and Mattox, K.L. (1992) Prospective evaluation
of preoperative fluid resuscitation in hypotensive patients with penetrating truncal injury:
a preliminary report. Journal of Trauma, 33, 354–61.
Study design
◆ A prospective single-centre randomized trial.
Randomization Immediate versus delayed fluid resuscitation
Inclusion criteria ≥16 years
Gunshot or stab wounds to the torso
Systolic BP <90mmHg at time of on-scene assessment
◆ Only those patients in whom fluid resuscitation might influence outcome were
included in the final study analysis. Therefore the study effectively excluded patients
with a Revised Trauma Score of zero at the scene, patients with a fatal gunshot
wound to the head, and patients with minor injuries not requiring operative
intervention.
◆ The immediate-resuscitation group was given IV fluid before surgical intervention.
The delayed-resuscitation group received no IV fluid until surgery.
◆ Surgical intervention included thoracotomy at the trauma centre, thoracotomy in the
operating room, laparotomy, neck exploration, and groin exploration.
FLUID RESUSCITATION FOR HYPOTENSIVE PATIENTS WITH PENETRATING TORSO INJURIES 377
Outcome measures
Primary endpoint
◆ Mortality.
Secondary endpoints
◆ Intra-operative blood loss.
◆ Post-operative complications.
Results
Immediate resuscitation Delayed resuscitation p-value
(n = 309) (n = 289)
Pre-hospital IV fluids 870mL 92mL
In-hospital IV fluids 1608mL 283mL
Hospital stay 14 days 11 days 0.006
Post-operative complications 30% 23% 0.08
Survival 60% 70% 0.04
◆ The small volumes of fluid administered to the delayed-resuscitation group were

accounted for by IV antibiotics, contrast medium, and fluids used to maintain line
patency.
◆ There was a trend toward increased intra-operative blood loss in the
immediate-resuscitation group (p = 0.11)
◆ There was no significant difference between the groups in the length of stay in
the ICU.
Conclusions
Delayed IV fluid resuscitation until surgical intervention improves outcomes, including
mortality, for hypotensive patients with penetrating torso injuries.
Critique
This seminal paper revolutionized the care of the bleeding trauma patient. Prior to this
work by Mattox and colleagues there had been at least five decades of research on the
subject. Many of these studies were limited by small numbers, heterogeneous patient
populations, and poor methodology, or had been conducted in animal models. Others
did not focus on the crucial question: what is the influence of withholding volume
resuscitation until control of haemorrhage? This was the first randomized trial to test
the hypothesis that aggressive IV fluid therapy in the actively bleeding patient would be
detrimental to outcome. There are three proposed mechanisms for this based on experi-
mental data, but which may be applicable to the exsanguinating human trauma subject.
378 TRAUMA SURGERY
Firstly, hypotension in the context of bleeding probably has a beneficial effect by reducing
the rate of blood loss—raising the blood pressure by volume resuscitation will negate this.
Secondly, expanding the intravascular volume may result in hydraulic disruption of the
immature thrombus—the ‘popping the clot’ phenomenon. Thirdly, crystalloid infusion
will dilute remaining coagulation factors and also lower blood viscosity, and therefore it
will reduce the resistance to flow around an incomplete thrombus.
It was beyond the scope and intent of this work to elucidate the mechanism by which
delayed fluid resuscitation may exert a positive influence. However, it did convincingly
demonstrate a reduction in morbidity and mortality with this approach.
There are several caveats to adopting this approach in all injured patients. This work
demonstrated an outcome benefit in the context of penetrating torso trauma with hypo-
tension, but to date there are no data to support this approach in the patient with blunt
injury. Patients with an associated head injury also represent a subgroup in which peri-
ods of hypotension, however brief, are known to be associated with adverse outcomes.
The transport times, and consequently time to control of bleeding, in this mature trauma
system were very short, and it may not be possible to extrapolate these results to other
environments.
Regardless of these limitations, this key paper established an evidence base for the con-
cept of permissive hypotension in the bleeding trauma patient and dramatically altered
practice which had previously been based on poor-quality research and surgical dogma.
EFFECT OF TRANEXAMIC ACID IN BLEEDING TRAUMA PATIENT 379
11.4 The effect of tranexamic acid in the bleeding

trauma patient
Details of studies
Tranexamic acid (TA), a fibrinolytic inhibitor, has been shown to reduce bleeding and
thereby promote blood conservation in several fields of elective surgery. One consequence
of the mode of action of TA is the potentially increased risk of intravascular thrombosis.
The CRASH-2 study was an international multi-centre randomized placebo-controlled
trial of TA in the bleeding trauma patient. It was designed to assess the effects of TA on
death, vascular occlusive events, and blood transfusion in trauma patients with significant
haemorrhage.
Study references
Main study
Shakur, H., Roberts, I., Bautista, R., et al. (2010) Effects of tranexamic acid on death, vascular occlusive
events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2):
a randomised, placebo-controlled trial. Lancet, 376, 23–32.
Related references
Roberts, I., Shakur, H., Afolabi, A., et al. (2011) The importance of early treatment with tranexamic acid
in bleeding trauma patients: an exploratory analysis of the CRASH-2 randomised controlled trial.
Lancet, 377, 1096–1101.
Study design
◆ A prospective multi-centre randomized placebo-controlled trial.
Randomization Tranexamic acid versus placebo
Inclusion criteria Adult trauma patients (age ≥16 years)
Systolic BP <90mmHg or heart rate >110bpm
Considered at risk of haemorrhage and within 8 hours of injury
Exclusion criteria Clear indication or contraindication for TA
Follow-up In-hospital up to 28 days after randomization
Outcome measures
Primary endpoint
◆ In-hospital mortality within 4 weeks of injury.
◆ Cause of death was described in the following categories: bleeding, vascular
occlusion (MI, stroke, and PE), multi-organ failure, head injury, and other.
380 TRAUMA SURGERY
Secondary endpoints
◆ Vascular occlusive events (MI, stroke, PE, and DVT).
◆ Surgical intervention (neurosurgery, thoracic, abdominal, and pelvic surgery).
◆ Receipt of blood transfusion and units of blood products transfused.
◆ The five-point Modified Oxford Handicap Scale was used to measure dependency
either at the time of discharge or on day 28 if the patient was still in hospital. The
scale was dichotomized into dead/dependent or independent.
Results
Tranexamic acid Placebo RR (95% CI) p-value
(n = 10,060) (n = 10,067)
Any cause of death 1463 (14.5%) 1613 (16%) 0.91 (0.85–0.97) 0.0035
Bleeding 489 (4.9%) 574 (5.7%) 0.85 (0.76–0.96) 0.0077
Vascular occlusion 33 (0.3%) 48 (0.5%) 0·69 (0·44–1·07) 0.096
Multi-organ failure 209 (2.1%) 233 (2.3%) 0·90 (0·75–1·08) 0.25
Head injury 603 (6%) 621 (6.2%) 0·97 (0·87–1·08) 0.60
Other causes 129 (1.3%) 137 (1.4%) 0·94 (0·74–1·20) 0.63
◆ All-cause mortality was significantly reduced with TA (RR 0.91).

◆ The risk of death due to bleeding was significantly reduced with TA (RR 0.85).
◆ Blood product transfusions were given to 5067 patients (50.4%) allocated to TA
versus 5160 of patients (51.3%) allocated to placebo.
◆ 4814 patients (47.9%) in the TA group received one or more surgical intervention
(neurosurgery, or chest, abdominal, or pelvic surgery) versus 4836 (48.0%) in the
placebo group.
Conclusions
A short course of tranexamic acid given early after injury reduced the risk of death in
bleeding trauma patients. There was no increase in vascular occlusive events associated
with the use of TA in this patient population.
Critique
Injury is one of the leading causes of death worldwide, with at least a third of in-hospital
trauma mortality due to bleeding. In recent years there has been a paradigm shift to
embrace the concept of damage control resuscitation, which includes a focused approach
to the management of traumatic coagulopathy. Antifibrinolytic agents, such as TA, have
been shown to reduce blood loss in elective surgery. TA is a synthetic derivative of the
amino acid lysine that exerts its influence on plasminogen. The use of these agents is
limited to a certain extent by the risk–benefit ratio of the haemostatic properties balanced
against the potential for adverse thrombotic events.
EFFECT OF TRANEXAMIC ACID IN BLEEDING TRAUMA PATIENT 381
The CRASH-2 trial was a major international randomized placebo-controlled clinical

trial to determine the effects of TA on death, vascular occlusive events, and blood transfu-
sion in trauma patients with significant haemorrhage. There were 274 participating hos-
pitals in 40 countries with a total study population of 20,211 patients. Trauma patients
with, or perceived to be at risk of, significant bleeding were randomized within 8 hours
of injury to receive either TA (loading dose 1g over 10 minutes, then infusion of 1g over
8 hours) or a matched placebo.
The investigators demonstrated a reduction in all-cause mortality within 28 days of
injury, and in particular a reduction in deaths considered secondary to haemorrhage.
There were no increased thrombotic events or related mortality associated with the use of
TA in this population.
One of the major strengths of this trial, in addition to the robust methodology and
impressive scale, was the inclusion criteria. These were based on clinical judgement
rather than laboratory values, i.e. patients were included if there was clinical evidence
of ongoing haemorrhage or if they were considered to be at significant risk of bleeding.
This has two pragmatic advantages in dealing with this challenging group of patients.
Firstly, it facilitates rapid treatment of the bleeding trauma patient without the inherent
delays associated with laboratory tests. Secondly, it is a strategy which can be employed
in resource-challenged circumstances where access to laboratory values may be limited,
such as developing countries or the combat environment.
A published follow-up subgroup analysis has demonstrated the importance of early
administration of TA in the bleeding trauma patient (Roberts et al.). Treatment given
within 1 hour of injury significantly reduced the risk of death due to bleeding. Treatment
given 1–3 hours post-injury also reduced this risk. Paradoxically, TA given after 3 hours
seemed to increase the risk of death due to bleeding.
Whilst further work is required to clarify this and other issues raised by the CRASH-2
trial, the evidence provided by this study supports the use of TA in the bleeding trauma
patient and therefore has the potential to influence practice and improve outcome.
Appendix
Levels of evidence
APPENDIX 383
Therapy/Prevention/Aetiology/Harm
1a Systematic review (with homogeneity) of RCTs
1b Individual RCT (with narrow CI)
1c All-or-none RCTs
2a Systematic review (with homogeneity) of cohort studies
2b Individual cohort study or low-quality RCT (e.g. <80% follow-up)
2c ‘Outcomes’ research, ecological studies
3a Systematic review (with homogeneity) of case–control studies
3b Individual case–control study
4 Case series (and poor-quality cohort/case–control studies)
5 Expert opinion without explicit critical appraisal, or based on physiology,
bench research, or ‘first principles’
Diagnosis
1a Systematic review (with homogeneity) of level 1 diagnostic studies; or a clinical
decision rule with 1b studies from different clinical centres
1b Validating cohort study with good reference standards; or a clinical decision rule
tested within one clinical centre
1c Absolute SpPins (diagnostic finding whose specificity is so high that a positive
result rules in the diagnosis) and SnNouts (diagnostic finding whose sensitivity
is so high that a negative result rules out the diagnosis)
2a Systematic review (with homogeneity) of level >2 diagnostic studies
2b Exploratory cohort study with good reference standards, clinical decision rule
after derivation, or validated only on split-sample or databases
3a Systematic review (with homogeneity) of 3b and better studies
3b Non-consecutive study, or without consistently applied reference standards
4 Case-control study, poor or non-independent reference standard
Prognosis
1a Systematic review (with homogeneity) of inception cohort studies; or a clinical
decision rule validated in different populations
1b Individual inception cohort study with >80% follow-up, or a clinical decision
rule validated on a single population
1c All-or-none case-series
2a Systematic review (with homogeneity) of either retrospective cohort studies or
untreated control groups in RCTs
2b Retrospective cohort study or follow-up of untreated control patients in an RCT,
or derivation of a clinical decision rule or validated on split-sample only
2c ‘Outcomes’ research
4 Case series (and poor-quality prognostic cohort studies)
Adapted from ‘Levels of Evidence’ <http://www.cebm.net/index.aspx?o=1025>, March 2009, with kind permission of
The Centre for Evidence-Based Medicine.
Author index
Abbas, S. 83–5 Bloom, H.J. 334

Abou-Assi, S. 171 Boeck, S. 195
Adam, A. 152 Boeckxstaens, G.E. 106–7, 108
Adam, D.J. 281–4 Bonadonna, G. 337–8
Adam, I.J. 240 Bonenkamp, J.J. 136, 138–9
Al-Khamis, A. 234–5 Bonjer, H.J. 214, 216–17
Al-Omran, M. 95 Bornman, P.C. 80
Albertini, J.J. 324–6 Bos, A.M.C. 51–2
Allum, W.H. 148, 151 Bosset, J.F. 251–2
Almeida, J.I. 290–1, 292–3 Bowel Cancer Screening programme England 236
Altemeier, W.A. 229 Bowel Cancer Screening programme Scotland 236
Andersen, T.F. 11 Bradbury, A.W. 281
Anderson, A.D.G. 218, 220 Bradley, E.L. 71, 178
Andre, T. 255, 257 Branco, B.C. 83–5
Andrew, T. 51–2 Brandstrup, B. 30
Anvari, M. 120, 121, 122 Brisinda, G. 224–5
Arain, M. 12 Brix, D. 358, 359
Armitage, P. 6 Broeders, J.A. 123, 126
Arvidsson, D. 39, 40, 41 Brown, A.J. 229
Asencio, F. 47–8 Brown, S.R. 231
Association of Coloproctology of Great Britain and Buck, N. 102–4
Ireland (ACPGBI) 214, 217, 254 Buyse, M. 255
ATAC Trialists’ Group 347–8
Atkin, W.S. 236, 237–9 Cahen, D.L. 184–5, 186–7
Attard, A.R. 59–60, 61 Callesen, T. 44–5
Aufenacker, T.J. 49 Campbell, M.K. 7
Campos, G.M. 106, 107, 108
Bagi, P. 71, 72–3 Carbajo, M.A. 47–8
Bai, Y. 175–7 Carpenter, J.R. 20
Balakrishnan, A. 290–1, 292–3 Carter, C.R. 181
Barczynki, M. 362–4 Casas, M. 171
Bardram, L. 218, 220 Celdren, A. 49
Barnett, H.J.M. 262 Chabok, A. 89–91
Bartelink, H. 258 Chaiyakunapruk, N. 26, 27
Barwell, J.R. 294–6 Chambers, B.R. 266
Bascom, J.U. 234 Chari, S.T. 188–90
Basse, L. 218–20 Chen, I.C. 112, 114
Bear, H.G. 355 Chen, S.C. 83
Begg, C.B. 129, 130, 131 Chowdhury, M.M. 53, 54
Bell, J. 281 Choy, P.Y.G. 212
Berndsen, F.H. 39 Christiansen, J. 80
Berne, T.V. 77 Chung, L. 36–8
Bernsden, E. 39 Chung, R.S. 210
Berrisford, R.G. 140, 142 Collins, R. 117
Besselink, M.G. 181–3 Colton, T. 6
Bickell, W.H. 376–8 Cook, D.J. 115
BIG 1-98 Collaborative Group 347 Cooke, T. 332–3
Biondo, S. 83 Coombes, R.C. 347
Birbeck, K.F. 244 Crisp, E. 77
Birkmeyer, J.D. 53–4 Crofts, T.J. 77–8
Blamey, R.W. 332 Cunningham, D. 144, 145, 146
Blatchford, O. 112–14 Cunningham, J. 44, 45
Bloemen, A. 32, 33 Cuschieri, A. 136, 137, 138
AUTHOR INDEX 385
Cusick, J. 329 Gagner, M. 358–9

Cuzick, J. 345, 347, 348 Gaist, D. 11
GALA Trial Collaborative
D’Hoore, A. 229–30 Group 270–1
Dallal, H.J. 152, 153, 154 Gallegos, 34–5
Daphan, C. 234 Galmiche, J.P. 120, 121, 122
Darouiche, R.O. 26–7 Gattinoni, L. 99
Darwood, R.J. 290–1, 292–3 Genberg, H. 313–15
Dávila, D. 168, 169 Gérard, J.P. 251–2
de Campos-Lobato, L.F. 249 Gianni, L. 352, 353
De Herder, W.W. 201 Giuliano, A.E. 324–8
de Jonge, P.J.F. 127–8 Gohel, M.S. 294–6
Decadt, B. 65, 66–7 Goodwin, A. 329
Delaney, C.P. 218, 220 Goss, P.E. 349–51
Dellinger, E.P. 175, 176, 177 Govaert, B. 233
Delorme, E. 229 Granderath, F.A. 123
Deutekom, M. 231 Grant, A.M. 120, 121
Devereaux, P.J. 17 Greenhalgh, R.M. 285–6, 287, 288, 289
Di Saverio, S. 83, 85 Grieg, J.D. 51–2
Díte, P. 184–6, 187 Griffin, G.E. 80
Dixon, M.R. 75 Grimes, D.A. 14
Docherty, J.G. 212, 213 Guenaga, K.F. 210–11
Donovan, A.J. 78 Gui, D. 224
Douek, M. 39 Guillou, P.J. 214–17
Dralle, H. 362 Güller, U. 58
DREAM Study Group 275, 277 Gupta, R. 171
Drummond, R.J. 211 Gurusamy, K. 168–70
Drury, J.K. 80
Dubois, F. 156–9 Halsted, W.S. 318
Dutch Colorectal Cancer Group 252–4 Haveran, L.A. 221, 223
Heald, R.J. 240–3
Early Breast Cancer Trialists’ Collaborative Group Hellberg, A. 68, 69, 70
(EBCTCG) 329, 337, 341, 345–6 Henderson, C. 341
Eatock, F.C. 174 Heniford, B.T. 47, 48
Edwards, P. 20 Hinchey, E.J. 92, 93, 94
Ekberg, H. 298–301 Hjern, F. 89, 90
Eklund, A. 39 Hoffmann, J. 71, 72–3
Elston, C.W. 334 Hohenberger, W. 243
Ergina, L. 2 Hol, L. 239
European Carotid Surgery Trialists’ Collaborative Holm, T. 247–9
Group (ECST) 262–5 Homs, M.Y. 152, 153, 154
EVAR Trial Participants 275–80 Horvath, K.D. 181
Executive Committee for the Asymptomatic Carotid Hosker, G. 231
Atherosclerosis Study (ACAS) 266–7, 268–9 Hulscher, J.B. 132–5
Hundahl, S.A. 147
Faranda, C. 92 Hungness, E.S. 178
Farid, M. 85 Husker, C.G. 140, 141–2
Finan, P.J. 86 Hvid, I. 74, 75
Finks, J.F. 53, 54 Hvid-Jensen, F. 127–8
Finlayson, E.V. 129, 130, 131 Hynes, D.M. 39, 41
Fisher, B. 321, 323, 329–31, 337, 339, 355–6
Fitzgibbons, R.J. 36, 37–8 Irvin, G.L. 360–1
Flam, M. 258 Isenmann, R. 175, 176
Flentje, D. 360 Itani, K.M. 47–8
Flum, D.R. 58
Folkesson, J. 254 Jaffe, B.M. 205
Fong, Y. 205–7 Jafri, N.S. 176
Foran, B. 71, 72–3 Jarnagin, W. 206
Forbes, J.F. 281, 347, 348 Jayne, D.G. 214, 217
Ford, I. 11 Jenkins, T.P. 32
French Adjuvant Study Group 339 Joensuu, H. 352
386 AUTHOR INDEX
Johansson, B. 39 Medical Research Council (MRC) 2

Johansson, M. 168, 169 Asymptomatic Carotid Surgery Trial (ACST)
Jones, P.F. 56–8 Collaborative Group 266, 267, 268–9
Jordan, J.S. 71, 72–3 Laparoscopic Hernia Trial Group 39, 40
Oesophageal Cancer Working Group 148–9,
Kakker, V.V. 28–9 150, 151
Kalfarentzos, F. 171–4 Mehigan, B.J. 226–8
Kam, M.H. 228 Members of the Criteria Committee for Autoimmune
Kaminski, A. 74–5 Pancreatitis of the Japan Pancreas Society 188, 190
Karydakis, G.E. 234 Memon, M.A. 136, 137–8, 139
Kelsen, D. 148, 151 MERCURY Study Group 244, 245–6
Kenward, M. 20 Metzger, R. 129, 131
Keus, F. 160–2 Mier, J. 178–80
Kim, K. 64 Milligan, E.T.C. 226, 227
Kocher, H.M. 224, 225 Moers, C. 302–3, 304, 305–6
Kok, N.F. 310–12 Moertel, C.G. 255–7
Kolla, S.B. 168, 169 Moore, F.A. 95–8
Kracht, M. 212–13 Morales, R. 49
Krag, D.N. 324 Moran, B.J. 212
Krukowski, Z.H. 32–3, 92 Morgan, C.N. 226, 227
Morino, M. 65, 66–7
Lacy, A.M. 214–17 Morrow, M. 331
Lai, H.W. 74, 75 Moseley, J.B. 7
Lai, P.B. 168, 169 Mudge, M. 32
Laméris, W. 62 Mühe, E. 156
Lau, H. 168 Muñoz-Juárez, M. 212
Lau, J.Y.W. 117–18, 119 Murray, J.J. 86
Laurie, J.A. 255 Myers, E. 92–4
Lee, J.A. 53
Levin, T.R. 239 Nagai, K. 198
Levine, M.N. 339–40 Nagpal, K. 142
Lichtenstein, I.L. 46 National Cancer Institute-sponsored Clinical
Lichtiger, S. 221, 223 Outcomes of Surgical Therapy (COST) Study
Liem, M. 39 Group 214–17
Lillemoe, K.D. 191 National Confidential Enquiry into Patient Outcome
Lin, H.C. 129, 131 and Deaths 104
Lo, C. 168, 169 Neoptolemos, J.P. 195–7
Lobo, D.N. 30–1 Nesbitt, C. 291, 292–3
Longo, A. 226 Neuhauser, D. 34, 35
Louie, B.E. 171 Neumayer, L. 39, 40–1
Luketich, J.D. 140–1 Neutzling, C.B. 212, 213
Lund, J.N. 224 Ng, C.S. 62–4
Lunn, J.N. 102 Ng, E.K.W. 80–2
Lurie, F. 290–1, 292–3 Ng, K.H. 228
NICE 28, 29
McClave, S.A. 171 Nicholson, M.L. 310, 312
McDermott, F.T. 234 Nigro, N.D. 260
MacDonald, J.S. 144–6, 147 Nolvadex Adjuvant Trial Organisation
MacFarlane, J.K. 240 (NATO) 343–4, 345
MacKay, G. 30 Nordback, I. 175, 176
McMahon, A.J. 160, 162 Nordin, P. 42, 43
Mahon, D. 120, 121 Norrie, J. 20
Mansel, R.E. 324 North American Symptomatic Carotid
Maria, G. 224, 225 Endarterectomy Trial Collaborators
Markar, S.R. 123, 124–5, 126, 129, (NASCET) 262–5
130, 131 Northover, J. 258–60
Martin, M. 341–2 Nussbaum, S.R. 360
Martin, R.R. 376
Mattox, K.L. 377 O’Dwyer, P.J. 36, 42–3
Matzel, K.E. 231–3 O’Sullivan, G.C. 92
Mazari, F.A.K. 285–6, 288, 289 Oelschlager, B.K. 123, 125, 126
AUTHOR INDEX 387
Oettle, H. 195 Salminen, P. 123

Ohtani, H. 142–3 Salvia, R. 198
Ojo, A.O. 307, 309 Sanabria, A. 49–50
Oláh, A. 171 Santin, B. 129
Oliak, D. 71, 72–3 Sauer, R. 250–1
Omloo, J.M. 132 Sauerland, S. 48
Oteiza, F. 49 Schulz, K.F. 14
Øyen, O. 310, 312 Schwarz, M. 175, 176
SCOTIA Study Group 86–8
Palanivelu, C. 140, 142 Scottish Breast Cancer Trials Committee,
Partington, P.F. 185 Scottish Cancer Trials
Paterson-Brown, S. 65, 67, 68 Office 343
Patey, D.H. 318–20 Sebag-Montefiore, D. 252–4
Pathak, S. 207 Segarra-Newnham, M. 175
Paull, D.L. 71, 72–3 Seshadri, R. 335
Pedersen, A.G. 68, 69, 70 Sewnath, M.E. 191
Pederzoli, P. 175, 176 Shakur, H. 379–81
Pellegriti, G. 365, 366, 367–8 Shanmugam, V. 226
Perälä, J. 290–1, 292–3 Shenfine, J. 154
Perez, A.R. 49 Shoemaker, W.C. 99
Perkins, M.T. 285–7, 288–9 Simforoosh, N. 310
Perou, C.M. 336 Simillis, C. 71, 72
Petrov, M.S. 171 Sjoquist, K.M. 148, 149–51
Piantadosi, S. 6 Sjöström, L. 109–11
Piccart-Gebhart, M.J. 352–4 Skoubo-Kristensen, E. 71, 72–3,
Pisters, P.W. 191 74, 75
Pocock, S.J. 6 Slamon, D.J. 335, 352
Poole, C.J. 339 Sloane Project 331
Present, D.H. 221–3 Smith, A. 239
Smith, I. 352, 353
Quirke, P. 240, 242, 244–5, 246, 254 Smithers, B.M. 140, 142
Songun, I. 136, 139
Ramage, J.K. 201–4 Sørlie, T. 336
Randolph, G.W. 362 Srirajaskanthan, R. 114
Rao, P.M. 64 Stanley, A.J. 112, 114
Rasmussen, L.H. 290–1, 292–3 Starzl, T.E. 316
Rautio, T. 290–1, 292–3 Stelzner, S. 247
Reddy, S. 207 Stephens, J.R. 114
Rerkasem, K. 263 Stone, H.H. 373
Reynolds, W. 160 Strasberg, S.M. 159, 163–7
Rivers, E. 99–101 Subramonia, S. 290–1, 292–3
Roberts, I. 379, 381 Sugarbaker, P.H. 65–6, 67
Roché, H. 341 Sung, J.J.Y. 117, 118, 119
Rochelle, R.E. 185 Swedish Rectal Cancer Trial 252–4
Rockall, T.A. 112
Rockwell, P.M. 263 Takeda, K. 178
Rodgers, M. 129 Tanaka, M. 198–200
Rodriguez, J.R. 198–200 Tancini, G. 337
Roh, J.L. 365, 366, 367–8 Tanimura, S. 140
Roh, M.S. 250–1 Tanphiphat, C. 80
Røkke, O. 175, 176 Taylor, C.J. 92
Romond, E.H. 352 Taylor, F.G. 246
Rösch, T. 184 Taylor, H. 77, 78
Rothman, K.J. 10 Thomas, D.R. 75
Rothwell, P.M. 19 Thomas, S.H. 59, 60, 61
Rotondo, M.F. 373–5 Thomson, H.J. 56–8
Russell, C.F. 360 Thorpe, K.E. 14
Russell, J.C. 164 Toorenvliet, B.R. 94
Towler, B. 236–7, 238–9
Sainio, V. 175, 176 Traverso, L.W. 156
Sala, E. 62, 63 Twelves, C. 255
388 AUTHOR INDEX
UK EVAR Trial Investigators 275–80 West, N.P. 248–9

UK Small Aneurysm Trial Participants 272–4 Westerdahl, J. 360
UKCCCR Anal Cancer Trial Working Wexner, S.D. 233
Party 258–60 White, J.J. 56, 57
Whyman, M.R. 285–6, 287, 288, 289
van der Gaag, N.A. 191–4 Wileman, S.M. 120
van Gijn, W. 254 Willemsen, P.H. 74, 75
van Santvoort, H.C. 181–3 Wilson, J.M.G. 236
Varadhan, K.K. 218, 220 Windsor, J.A. 181
Veldkamp, R. 214–17 Wittau, M. 175
Vergara, M. 115–16 Wolf, J.S., Jr 310
Veronesi, U. 321–3 Wolfe, R.A. 307–9
Villatoro, E. 175 Wouters, M.W. 129, 131
Visick, AH. 77 Wright, D. 39
Wright, J.P. 302
Wang, L. 106, 107–8
Wang, Q. 191 Ychou, M. 146
Wapnir, L. 329 Yerdel, M.A. 49
Watson, C.J.E. 302, 303–4, 305–6 Yilmaz, S. 234
Watts, G.T. 180
Weissfeld, J. 239 Zoltie, N. 59
West, J.G. 370–2 Zwarenstein, M. 22
Subject index
abdominal aortic aneurysms (AAAs) angioplasty

EVAR 1 trial 275–7 balloon angioplasty, for severe limb
EVAR 2 trial 278–80 ischaemia 281–4
UK Small Aneurysm Trial 272–4 percutaneous transluminal angioplasty (PTA)
abdominal injury, damage control surgery 373–5 intermittent claudication 285–9
abdominal pain severe limb ischaemia 281–4
acute see acute abdomen anthracyclines in breast cancer 339–40
non-specific (NSAP) 65–7 and taxanes 341–2
abdominal surgery antibiotic prophylaxis
abdominal wall closure 32–3 inguinal hernia repair 49–50
post-operative intravenous fluid therapy 30–1 necrotizing pancreatitis, severe acute 175–7
abdomino-perineal excision of the rectum uncomplicated diverticulitis 89–91
(APER) 241, 242, 247–9 aortic aneurysms
ABO-incompatible (ABOi) renal repair of non-ruptured 53
transplantation 313–15 see also abdominal aortic aneurysms
achalasia, surgery for 106–8 aorto-iliac (AI) disease 287, 289
acid-reducing surgery, perforated duodenal appendicectomy
ulcers 80, 82 interval 74–6
ACT I trial 258–60 laparoscopic versus open 68–70
ACT II trial 260 negative appendicectomy rate 57, 58
active observation CT 64
acute abdomen 67 diagnostic laparoscopy 67, 68
appendicitis, suspected 56–8 appendicitis
emergency surgery 56–8 complicated, non-operative management 71–3
acute abdomen CT 64
active observation 56–8 diagnostic laparoscopy 67
assessment recurrent 74–6
CT 62–4 suspected, and active observation 56–8
early laparoscopic 65–7 see also appendicectomy
opioid analgesia prior to 59–61 aromatase inhibitors and breast cancer 347–8
adalimumab 223 Association of Coloproctology of Great Britain and
adenocarcinoma Ireland (ACPGBI)
in Barrett’s oesophagus 127–8 left-sided malignant large bowel obstruction,
trans-thoracic versus trans-hiatal surgery for 88
oesophagectomy 132–5 mechanical bowel preparation 211
adenoma, incidental 358 rectal cancer management 254
adrenalectomy, laparoscopic 358–9 Association of Surgeons of Great Britain and
adrenaline (epinephrine) Ireland 102
carotid endarterectomy 271 AstraZeneca 343
peptic ulcer bleeding 115–16, 117, 118 Asymptomatic Carotid Atherosclerosis Study
adriamycin 355–6 (ACAS) 266–7
age factors, ethical issues 7 Asymptomatic Carotid Surgery Trial (ACST) 266,
allocation concealment 14 267, 268
alternative hypothesis in non-inferiority trials 9 autoimmune pancreatitis (AIP), diagnosis 188–90
American College of Surgeons Oncology group 327 axillary dissection versus no axillary dissection in
anal cancer, primary chemoradiotherapy for 258–60 breast cancer 327–8
anal fissure, non-surgical treatment of chronic 224–5
analgesia prior to assessment of acute balloon angioplasty, severe limb ischaemia 281–4
abdomen 59–61 balloon dilation, endoscopic (EBD) in
analysis of surgical trials 18–21 achalasia 107, 108
analysis sets 18–19 bariatric surgery
anastomosis, stapled versus handsewn 212–13 health benefits 109–11
anastrozole 347–8 SCOTT project 12
390 SUBJECT INDEX
Barrett’s oesophagus 127–8 symptomatic carotid atherosclerosis 262–5

BASIL trial 281–4 carotid stenosis, asymptomatic 266–9
BCIRG006 trial 354 case-control studies 12
Beatson, George 332 retrospective 4
bevacizumab causal modelling 19–20
colon cancer 257 cetuximab in colon cancer 257
gastric cancer 147 change of status forms 8
bias chemoradiotherapy
analysis sets 19 anal cancer 258–60
non-inferiority trials 9 gastric cancer 145, 146, 147
publication see publication bias oesophageal cancer 148–51
bile duct injury (BDI), laparoscopic rectal cancer 250–2, 254
cholecystectomy 158, 163–7, 169, 170 pancreatic cancer 195–7
biological mesh repair of para-oesophageal chemotherapy
hernia 123, 125, 126 breast cancer 322, 328, 339
blinding 14–15 anthracyclines 339–42
botulinum toxin (Botox) and Herceptin 352–4
chronic anal fissure 224–5 neoadjuvant 355–6
endoscopic botulinum toxin injection (EBTI) in origins 337–8
achalasia 106, 107 and tamoxifen 345–6
bowel cancer taxanes 341–2
adjuvant therapy in colon cancer 255–7 gastric cancer 144–7
screening 236–9 oesophageal cancer 148–51
see also rectal cancer child patients
brachytherapy in oesophageal cancer 152, 153, 154 ABO-incompatible renal transplantation 314
breast cancer ethical issues 7
aromatase inhibitors 347–8 liver transplantation 316
axillary dissection versus no axillary chlorhexidine–alcohol solution, surgical skin
dissection 327–8 preparation 26–7
chemotherapy cholecystectomy
anthracyclines 339–40 laparoscopic (LC)
and Herceptin 352–4 biliary injury 163–7
neoadjuvant 355–6 clinical trials 160–2
origins 337–8 early versus late in acute cholecystitis 168–70
and tamoxifen 345–6 evolution 156–9
taxanes 341–2 neuroendocrine tumours 203
ductal carcinoma in situ 329–31 cholecystitis, early versus late laparoscopic
endocrine therapy 349–51 cholecystectomy in acute 168–70
genetics ciclosporin in renal transplantation 298, 299,
HER-2/neu oncogene 335 300, 301
patterns of gene expression 336 circumferential resection margin (CRM) in rectal
histological grade and prognosis 334 cancer 244–6, 247, 248–9
mastectomy cisplatin
breast conservation versus 321–3 gastric cancer 144, 146
simple versus radical 318–20 oesophageal cancer 149, 150
oestrogen receptors 332–3 CLASICC trial 214, 215–16, 217
sentinel node biopsy 324–6 Clinical Trials of Investigational Medical Products
tamoxifen 343–6 (CTIMPs) 16
breast conservation versus mastectomy 321–3 cluster randomized controlled trials 9
bypass or angioplasty for severe limb ischaemia Cochrane Collaboration 22
(BASIL) trial 281–4 cognitive impairment, patients with 7
cohort studies 10–12
calcineurin inhibitors, reduced exposure in renal colectomy, subtotal versus segmental 86–8
transplantation 298–301 colon cancer, adjuvant therapy in 255–7
Californian trauma care systems 370–2 colonic stenting in left-sided malignant large bowel
capecitabine in colon cancer 257 obstruction 88
cardiac risk factors, and bariatric surgery 109, 110 colonoscopy
carotid atherosclerosis, symptomatic 262–5 following conservative treatment of complicated
carotid endarterectomy appendicitis 73
asymptomatic carotid stenosis 266–9 recurrent appendicitis 76
general versus local anaesthesia 270–1 COLOR trial 214, 215–16
SUBJECT INDEX 391
colorectal cancer CReST (colorectal stenting) trial 88

screening 236–9 critical view of safety, laparoscopic
see also rectal cancer cholecystectomy 165, 166–7
colorectal liver metastases, surgical resection CRITICS study 147
of 205–7 Crohn’s disease 221–3
colorectal surgery cross-sectional studies 5
anal cancer, chemoradiotherapy 258–60 crossover studies 9
anal fissure, non-surgical treatment 224–5 crossover surgery 18
bowel cancer screening 236–9 Cushing’s syndrome/disease 358
colon cancer, adjuvant therapy 255–7 cyclophosphamide in breast cancer 322, 337–8,
enhanced recovery 218–20 339–40, 341–2, 355–6
faecal incontinence, sacral neuromodulation
for 231–3 daclizumab in renal
fistulating Crohn’s disease, infliximab in 221–3 transplantation 299, 301
laparoscopic 214–17 damage control surgery 373–5
mechanical bowel preparation 210–11 Danish National Hospital Registry 11
pilonidal sinus disease 234–5 Danish Register of Medicinal Product
rectal cancer Statistics 11
circumferential resection margin 244–6 Data Monitoring Committees 16
extralevator abdomino-perineal excision of the deceased donor transplantation, machine
rectum 247–9 preservation versus static cold
neoadjuvant therapy 250–4 storage in 302–6
total mesorectal excision 240–3 Denmark, national registry data 11
rectal prolapse, laparoscopic ventral mesh design issues for surgical trials 3–15
rectopexy 229–30 diabetes and bariatric surgery 109, 110
stapled haemorrhoidectomy 226–8 diltiazem for chronic anal fissure 225
stapled versus handsewn anastomosis 212–13 disease registries 4
colostomy (Hartmann’s procedure) observational studies 11
complicated diverticulitis 92, 93, 94 diverticulitis
left-sided malignant large bowel complicated 92–4
obstruction 86, 87 uncomplicated 89–91
complementary DNA (cDNA) microarrays 336 docetaxel in breast cancer 341–2, 356
complete case analysis 21 donor nephrectomy, laparoscopic versus
complete mesocolic excision (CME) for colonic mini-incision open 310–12
cancer 243 doxorubicin in breast cancer 339, 341–2
complexity levels in surgical trials 5 drug trials
computed tomography (CT) blinding 15
acute abdomen 62–4 causal modelling 20
biliary injuries 165 explanatory–pragmatic continuum 5
complicated appendicitis 73 mis-randomization 18
recurrent appendicitis 76 phases 3
Confidential Enquiry into Peri-Operative Deaths randomization 18
(CEPOD) 102–4 second-line medication 18
confirmatory subgroup analyses 19 standardization 16
confounding subgroup analyses 20
causal modelling 20 ductal carcinoma in situ (DCIS) 329–31
internal validity 13 duodenal ulcer, perforated
observational studies 10–11 eradication therapy after closure 80–2
subgroup analyses 20 non-operative treatment 77–9
CONKO-001 study 195 dysphagia
Conn’s syndrome 358 following laparoscopic anti-reflux
consent, informed 7 surgery 124, 126
conservative management palliation, in advanced oesophageal
complicated appendicitis 71–3 cancer 152–4
perforated duodenal ulcer 77–9
severe acute necrotizing pancreatitis 179, 180 Edinburgh study, intermitted
small bowel obstruction 83–5 claudication 286, 287, 288–9
CONSORT initiative 22 efficacy, explanatory trials 4
coronary artery bypass grafting 53 efficiency, explanatory trials 4–5
COST trial 214, 215–16 electronic disease registries 4
CRASH-2 study 379–81 observational studies 11
392 SUBJECT INDEX
emergency surgery epinephrine (adrenaline)

active observation 56–8 carotid endarterectomy 271
acute abdomen peptic ulcer bleeding 115–16, 117, 118
CT assessment 62–4 epirubicin
early laparoscopic assessment 65–7 breast cancer 339–40
opioid analgesia prior to assessment 59–61 gastric cancer 144
appendicitis equivalence trials 8
interval appendicectomy 74–6 analysis sets 19
laparoscopic versus open eradication therapy after simple closure of perforated
appendicectomy 68–70 duodenal ulcer 80–2
non-operative management 71–3 ESCHAR study 294–6
Confidential Enquiry into Peri-Operative esomeprazole
Deaths 102–4 following endoscopic therapy for bleeding peptic
diverticulitis management ulcer 118, 119
complicated diverticulitis 92–4 gastro-oesophageal reflux disease 122
uncomplicated diverticulitis 89–91 ESPAC-1 study 195, 196
ethical issues 7 ethical issues 7–8
large bowel obstruction, left-sided malignant 86–8 Danish registry data 11
perforated duodenal ulcer safety in RCTs 16
eradication therapy after simple closure 80–2 European Neuroendocrine Tumour Society 201
non-operative treatment 77–9 European Organization for Research and Treatment
sepsis of Cancer (EORTC) trial, anal cancer 260
enteral versus parenteral nutrition 95–8 European Union (EU) Clinical Trials Directive 16
goal-directed therapy 99–101 evidence, levels of 384
small bowel obstruction, water-soluble EVOLVeS study 290, 291
contrast in 83–5 excisional haemorrhoidectomy 226–8
endocrine surgery exercise therapy, intermittent claudication 285–9
adrenalectomy, laparoscopic 358–9 expertise-based trials 17
papillary thyroid carcinoma, management of the explanatory trials 4, 5
N0 neck 365–8 exploratory subgroup analyses 19
parathyroid gland hyperfunction, operative external beam radiotherapy
monitoring of 360–1 anal cancer 258, 259
recurrent laryngeal nerve monitoring in thyroid neuroendocrine tumours 203
surgery 362–4 external validity 14
endocrine therapy in breast cancer 348, 349–51 extra-corporeal membrane oxygenation
endoscopic balloon dilation (EBD) in (ECMO) 306
achalasia 107, 108 extralevator abdomino-perineal excision (ELAPE) of
endoscopic botulinum toxin injection (EBTI) in the rectum 247–9
achalasia 106, 107
endoscopic retrograde cholangiopancreatography faecal immunochemical test (FIT) 239
(ERCP) 165, 166 faecal incontinence, sacral neuromodulation
endoscopic therapy for 231–3
chronic pancreatitis 184–7 faecal occult blood test (FOBt) 236–7, 238–9
peptic ulcer bleeding feasibility studies 12
proton pump inhibitor treatment femoral hernias, symptomatic 34–5
following 117–19 femoro-popliteal (FP) disease, intermittent
single versus dual endotherapy 115–16 claudication 287, 289
endoscopic ultrasound (EUS) in necrotizing fentanyl in groin hernia repair 43
pancreatitis 183 FinHER study 354
endovascular aneurysm repair (EVAR) 274 fistulating Crohn’s disease 221–3
EVAR 1 trial 275–7 Fitz, Reginald 57
EVAR 2 trial 278–80 flexible sigmoidoscopy in bowel cancer
endovenous laser therapy (EVLT) screening 236, 237–8, 239
superficial venous incompetence 296 fluid restriction after major abdominal surgery 30–1
varicose veins 291, 293 fluid resuscitation, immediate versus delayed in
endovenous therapy for varicose veins 290–3 hypotensive patients with penetrating torso
enhanced recovery after colorectal injuries 376–8
surgery 218–20 5-fluorouracil (5-FU)
enteral nutrition anal cancer 258, 259, 260
in acute pancreatitis 171–4 breast cancer 322, 337–8, 339–40, 341–2
post-operative septic complications 95–8 colon cancer 255–6, 257
SUBJECT INDEX 393
gastric cancer 144–6 Helicobacter pylori

oesophageal cancer 149, 150 eradication therapy after closure of perforated
pancreatic cancer 195 duodenal ulcer 80–2
rectal cancer 252 peptic ulcer bleeding 119
folinic acid in pancreatic cancer 195 Hemoccult® test for bowel cancer 237
follow-up of clinical trials, long-term 11 heparin in venous thromboembolism
Food and Drug Administration (FDA) 8 prophylaxis 28–9
full Rockall score in upper GI haemorrhage 112–14 hepatobiliary scintigraphy 165
fundoplication, laparoscopic, in gastro-oesophageal hepatopancreaticobiliary surgery
reflux disease colorectal liver metastases 205–7
medical therapy versus 120–1 intraductal papillary mucinous neoplasm
technical aspects 123–5, 126 management 198–200
laparoscopic cholecystectomy
gallstones see laparoscopy: cholecystectomy biliary injury 163–7
gastrectomy 136–9 clinical trials 160–2
laparoscopic 140, 141–3 early versus late in acute cholecystitis 168–70
partial, for perforated duodenal ulcers 82 evolution 156–9
gastric cancer neuroendocrine tumour management 201–4
lymphadenectomy 136–9 pancreatic cancer
multimodality treatment 144–7 adjuvant therapy 195–7
GASTRIC study 146 pre-operative biliary drainage 191–4
Gastrointestinal Tumour Study Group (GITSG) pancreatitis
trial 195 antibiotic prophylaxis in severe acute
gastro-oesophageal reflux disease (GORD) necrotizing 175–7
laparoscopic fundoplication versus medical autoimmune, diagnosis 188–90
therapy 120–2 early versus late necrosectomy in severe
technical aspects of laparoscopic surgery 123–6 necrotizing 178–80
general anaesthesia (GA) endoscopic versus surgical therapy in
carotid endarterectomy 270–1 chronic 184–7
groin hernia repair 42–3 enteral versus parenteral nutrition in
genetics in breast cancer acute 171–4
HER-2/neu oncogene 335 minimally invasive versus open
patterns of gene expression 336 necrosectomy 181–3
gentamicin in surgery for ingrown toenails 51 HER-2/neu oncogene in breast cancer 335
Glasgow–Blatchford score (GBS) 112–14 Herceptin 352–4
glyceryl trinitrate (GTN) in chronic anal HERA trial 352–4
fissure 224–5 Herceptin in breast cancer 328, 352–4
goal-directed therapy in severe sepsis 99–101 high-grade dysplasia (HGD), incidence in Barrett’s
governance of RCTs 16 oesophagus 127–8
great saphenous vein (GSV), varicose veins 291, HISORt criteria, autoimmune pancreatitis
292, 293 diagnosis 188–90
groin hernias hospital volume and patient outcome 53–4
asymptomatic 36–8 oesophagectomy 53, 54, 129–31
chronic pain after repair 44–5 hyperparathyroidism 360–1
laparoscopic versus open repair 39–41 hypertension, and bariatric surgery 110
Lichtenstein repair 46 hypotensive trauma patients with penetrating torso
local versus general anaesthesia for repair 42–3 injuries 376–8
symptomatic 34–5 hypothermic machine perfusion in deceased donor
guaiac faecal occult blood test (gFOBt) 239 transplantation 302–6
‘gut origin of sepsis’ 97
ICI Pharmaceuticals 343
haemorrhoidectomy, stapled 226–8 iliac artery disease 286
handsewn anastomosis in colorectal surgery 212–13 incisional hernias
Hartmann’s procedure following abdominal wall closure 32–3
complicated diverticulitis 92, 93, 94 open versus laparoscopic repair 47–8
left-sided malignant large bowel obstruction 86, 87 independent Data Monitoring Committees
Hasson technique, laparoscopic cholecystectomy 158 (iDMCs) 16
head of pancreas cancer, pre-operative biliary infection, post-operative
drainage 191–4 antibiotic prophylaxis in inguinal hernia
heartburn following laparoscopic anti-reflux repair 49–50
surgery 124 ingrown toenails surgery 51, 52
394 SUBJECT INDEX
infliximab in fistulating Crohn’s disease 221–3 para-oesophageal hernia repair, technical

Information Services Division (ISD), Scottish aspects 123, 125, 126
Government 11 peritoneal lavage in complicated
informative missingness 21 diverticulitis 92–4
informed consent 7 rectopexy 229–30
ingrown toenails surgery 51–2 large bowel obstruction (LBO), left-sided
inguinal hernias malignant 86–8
antibiotic prophylaxis 49–50 laser therapy
asymptomatic 36–8 dysphagia palliation in oesophageal cancer 152,
Lichtenstein repair 46 153, 154
symptomatic 34–5 laparoscopic cholecystectomy 164
intention to treat (ITT) analysis last observation carried forward analysis 21
analysis sets 18–19 learning curves 17
non-inferiority trials 9, 19 cholecystectomy, laparoscopic 164
Intergroup studies 147, 151 gastric cancer surgery, lymphadenectomy in 138
intermittent claudication (IC), non-surgical groin hernia repair, laparoscopic 41
management 285–9 incisional hernia repair, laparoscopic 48
internal validity 13 oesophagogastric cancer, laparoscopic surgery 140
missing data 20–1 see also training
International Association of Pancreatology learning difficulties, patients with 7
(IAP) 198 left-sided malignant large bowel obstruction 86–8
interval appendicectomy (IA) 74–6 leg ulceration, superficial venous surgery in 294–6
intervention allocated versus intervention letrozole in breast cancer 349–51
received 18 leucovorin
intra-abdominal abscesses gastric cancer 144
laparoscopic appendicectomy 70 rectal cancer 252
non-operative management of complicated levamisole in colon cancer 255–6
appendicitis 71–3 Lichtenstein repair 39, 46
intraductal papillary mucinous neoplasm (IPMN) Lister, Joseph 26
management 198–200 liver metastases, surgical resection of
intravenous fluid therapy colorectal 205–7
hypotensive trauma patients 376–8 liver transplantation 316
after major abdominal surgery 30–1 local anaesthesia (LA)
carotid endarterectomy 270–1
Japan Pancreas Society 188, 190 groin hernia repair 42–3
long saphenous vein (LSV), incompetence 295, 296
kidney transplantation see renal transplantation long-term follow-up of clinical trials 11
King’s Fund for London, CEPOD study 102 longitudinal studies 5
Knee Arthroscopy Trial (KAT) 11 LOTUS study 120–2
lumpectomy 323
Ladies trial 94 lymphadenectomy
Langenbuch, Carl 156 gastric cancer 136–9
laparoscopy oesophageal cancer 132–5
achalasia 106–8 oesophagogastric cancer 140, 142
acute abdominal pain assessment 65–7
adrenalectomy 358–9 machine preservation in deceased donor
appendicectomy 68–70 transplantation 302–6
cholecystectomy (LC) MAGIC study 144, 145, 146, 151
biliary injury 163–7 magnetic resonance cholangiopancreatography
clinical trials 160–2 (MRCP), biliary duct injuries 165
early versus late in acute magnetic resonance imaging (MRI)
cholecystitis 168–70 acute abdomen assessment 64
evolution 156–9 circumferential resection margin in rectal
colorectal surgery 214–17 cancer 244, 245–6
donor nephrectomy 310–12 Mammoprint 336
gastro-oesophageal reflux disease masking randomized treatment (blinding) 14–15
medical therapy versus 120–2 mastectomy
technical aspects 123–5, 126 breast conservation versus 321–3
groin hernia repair 39–41 simple versus radical 318–20
incisional hernia repair 47–8 mechanical bowel preparation in colorectal
oesophagogastric cancer 140–3 surgery 210–11
SUBJECT INDEX 395
Medical Research Council necrosectomy

Gastric Cancer Surgical Trial 138 early versus late 178–80
neoadjuvant therapy in primary rectal minimally invasive versus open 181–3
cancer 252–4 necrotizing pancreatitis
STO3 study 147 antibiotic prophylaxis 175–7
Medicare patients enteral versus parenteral nutrition 171–4
and risk 54 necrosectomy
surgical volume and patient outcome 53, 54 early versus late 178–80
meta-analysis 22 minimally invasive versus open 181–3
methodological issues in surgical trials 1–2, 23 nephrectomy
analysis 18–21 laparoscopic versus mini-incision open
design issues 3–15 donor 310–12
protocol publication 22 surgical volume and patient outcome 53
randomized controlled trials 16–17 neuroendocrine tumour (NET) management 201–4
registration 22 Neurosign™ system 363
reporting 22 NHS Scotland 11
systematic reviews and meta-analysis 22 Nissen fundoplication in gastro-oesophageal reflux
methotrexate in breast cancer 322, 337–8, 339–40 disease
midazolam in open hernia repair 43 medical therapy versus 120–1
Miles, Sir Ernest 247 technical aspects 123–5, 126
Milligan–Morgan haemorrhoidectomy 226, 227 non-ignorable missingness 21
MIMIC study 286, 287, 288, 289 non-inferiority trials 8–9
mini-incision open donor nephrectomy 310–12 analysis sets 19
mini-laparotomy cholecystectomy 160–2 non-specific abdominal pain (NSAP) 65–7
missing data 13, 20–1 normothermic perfusion in deceased donor
mitomycin in anal cancer 258, 259, 260 transplantation 306
morphine administration prior to acute abdomen North American NET Society 204
assessment 60, 61 Nottingham Tenovus Primary Breast Cancer
Mouret, Phillipe 156 Study 334
mucosectomy 229 Nuffield Provincial Hospitals Trust 102
Mühe, Erich 156 null hypothesis in non-inferiority trials 9
multiple endocrine neoplasia 1 (MEN1) 201, 202
multiple imputation analysis 21 obesity, bariatric surgery for
mycophenolate mofetil (MMF) in renal health benefits 109–11
transplantation 298, 299, 301 SCOTT project 12
ABO-incompatible transplantation 313 observation, active
myotomy 106–8 acute abdomen 67
appendicitis, suspected 56–8
nasoenteric (NE) nutrition, acute pancreatitis 171–4 emergency surgery 56–8
National Cancer Institute, Milan 321, 323 observational studies 4, 10–12
National Confidential Enquiry into Perioperative validity 13
Deaths (NCEPOD) 31, 104 OEO2 trial 148–9, 150, 151
National Institute for Health and Clinical oesophageal adenocarcinoma (OAC) 127–8
Excellence (NICE) oesophageal cancer
groin hernia repair 41 in Barrett’s oesophagus 127–8
venous thromboembolism prophylaxis 29 dysphagia palliation 152–4
National Institute for Health Research (NIHR) minimally invasive surgery 140–3
Health Technology Assessment (HTA) multimodality treatment 146
programme 11, 12, 16, 20 pre-operative chemotherapy/
National Joint Registry 12 chemoradiotherapy 148–51
national registries 4 see also oesophagectomy
observational studies 11 oesophagectomy
National Research Council, trauma care system 370 laparoscopic 140–1, 142–3
National Surgical Adjuvant Breast and Bowel Project surgical volume and patient outcome 53, 54,
(NSABP) 129–31
B-15 study 339 trans-thoracic versus trans-hiatal 132–5
B-17 study 329–31 oesophagogastric cancer, minimally invasive surgery
B-18 study 355 for 140–3
B-24 study 329–31 oesophagogastric surgery
B-31 study 353, 354 achalasia 106–8
NCCTG-N9831 trial 353, 354 bariatric surgery 109–11
396 SUBJECT INDEX
oesophagogastric surgery (contd.) autoimmune, diagnosis 188–90

Barrett’s oesophagus, adenocarcinoma incidence endoscopic versus surgical therapy in
in 127–8 chronic 184–7
gastric cancer enteral versus parenteral nutrition
lymphadenectomy 136–9 in acute pancreatitis 171–4
minimally invasive surgery 140–3 effect on post-operative sepsis 97
multimodality treatment 144–7 necrosectomy
gastro-oesophageal reflux, laparoscopic surgery early versus late 178–80
medical therapy versus 120–2 minimally invasive versus open 181–3
technical aspects 123–6 pancreato-splenectomy for gastric
oesophagectomy cancer 137, 138, 139
hospital volume and outcome, relationship panel-type studies 5
between 129–31 PANTER study 183
trans-thoracic versus trans-hiatal 132–5 papaveretum prior to assessment of acute
oesophagogastric cancer abdomen 59, 60
dysphagia palliation 152–4 papillary thyroid carcinoma, management of the N0
minimally invasive surgery 140–3 neck 365–8
pre-operative chemotherapy or para-oesophageal hernia repair, laparoscopic 123–6
chemoradiotherapy 148–51 parathyroid gland hyperfunction, operative
para-oesophageal hernia repair 123–6 monitoring of 360–1
peptic ulcer bleeding parenteral nutrition
proton pump inhibitor treatment following in acute pancreatitis 171–4
endotherapy 117–19 effect on post-operative sepsis 95–8
single versus dual endotherapy 115–16 Parodi, Juan 278
upper GI haemorrhage, risk scoring for 112–14 partial gastrectomy for perforated duodenal
oestrogen receptors, role in breast cancer 332–3 ulcers 82
omeprazole following endotherapy for bleeding pattern mixture modelling 21
peptic ulcer 117–18 peptic ulcer
Oncotype DX 336 bleeding, endoscopic therapy for
opioid analgesia prior to assessment of acute proton pump inhibitor treatment
abdomen 59–61 following 117–19
Orange County, California trauma care single versus dual endotherapy 115–16
system 370–2 perforated duodenal ulcer
outcome assessment, blinded 15 eradication therapy after closure 80–2
oxaliplatin in colon cancer 257 non-operative treatment 77–9
Oxford study, intermittent claudication 286–7, per protocol (PP) analysis
288–9 analysis sets 19
non-inferiority trials 9, 19
paclitaxel in breast cancer 341 percutaneous drainage, necrotizing
paediatrics pancreatitis 181–3
ABO-incompatible renal transplantation 314 percutaneous trans-hepatic cholangiography (PTC),
ethical issues 7 biliary injuries 165, 166
liver transplantation 316 percutaneous transluminal angioplasty (PTA)
pain intermittent claudication 285–9
abdominal see acute abdomen; non-specific severe limb ischaemia 281–4
abdominal pain perforated appendix, and active observation 58
groin hernias perforated duodenal ulcer
asymptomatic 37 eradication therapy after closure 80–2
laparoscopic versus open repair 41 non-operative treatment 77–9
post-operative 44–5 perineal rectosigmoidecomy for rectal prolapse 229
incisional hernia repair 48 peri-operative mortality, CEPOD study 102–4
palliative care, dysphagia in oesophageal peripheral artery disease (PAD) 289
cancer 152–4 phaeochromocytoma 358
pancreatic cancer phenolization, ingrown toenails surgery 51–2
adjuvant therapy 195–7 pilonidal sinus disease 234–5
pre-operative biliary drainage 191–4 pilot studies 12
surgical volume and patient outcome in pancreatic placebo surgical trials, ethical issues 7
resection 53, 54 pneumatic dilation, achalasia 106–8
pancreatitis polydioxanone in abdominal wall closure 32–3
antibiotic prophylaxis in severe acute polypropylene in abdominal wall closure 32–3
necrotizing 175–7 post-hoc subgroup analyses 19
SUBJECT INDEX 397
post-operative infection rescue surgery 18

antibiotic prophylaxis in inguinal hernia validity 13, 14
repair 49–50 record linkage 4
ingrown toenails surgery 51, 52 observational studies 11
povidone–iodine solution for surgical skin rectal cancer
preparation 26–7 circumferential resection margin 244–6
pragmatic trials 4–5 extralevator abdomino-perineal excision of the
PRECIS tool 14 rectum 247–9
pre-endoscopy Rockall score 112–14 neoadjuvant therapy 250–4
pre-operative biliary drainage (PBD) for head of total mesorectal excision 240–3, 247, 249
pancreas cancer 191–4 circumferential resection margin 246
procedure for prolapse and haemorrhoids neoadjuvant thrapy 250, 251, 252, 253, 254
(PPH) 226–8 see also colorectal cancer
prophylaxis rectal enemas 211
antibiotic see antibiotic prophylaxis rectal prolapse, laparoscopic ventral mesh rectopexy
pulmonary embolism 28–9 for 229–30
venous thromboembolism 28–9 rectopexy, laparoscopic ventral mesh 229–30
prospective data 5 recurrent laryngeal nerve (RLN) monitoring in
prosthetic reinforcement in para-oesophageal hernia thyroid surgery 362–4
repair 123, 125, 126 REFLUX study 120–2
protocol publication 22 registration of trials 22
proton pump inhibitor (PPI) therapy registries
following endotherapy for bleeding peptic national 4, 11
ulcer 117–19 specialist surgical 11
for gastro-oesophageal reflux renal transplantation
disease 120–2 ABO-incompatible 313–15
publication bias calcineurin inhibitors, reduced exposure
registration of trials 22 to 298–301
UK Small Aneurysm Trial 273 donor nephrectomy, laparoscopic versus
underpowered studies 13 mini-incision open 310–12
pulmonary embolism (PE) prophylaxis 28–9 machine preservation versus static cold storage in
deceased donor transplantation 302–6
quadrantectomy 321–3 mortality benefit 307–9
quality of data 10 reporting 22
safety 16
radiofrequency ablation (RFA) rescue surgery 18
colorectal liver metastases 207 residual confounding 13
superficial venous incompetence 296 internal validity 13
varicose veins 291, 292, 293 observational studies 10–11
radiotherapy response rate, improving 20–1
anal cancer 258–60 retrospective case-control studies 4
breast cancer 322, 323 retrospective data 5
ductal carcinoma in situ 330, 331 rituximab in ABO-incompatible renal
external beam transplantation 313
anal cancer 258, 259 Rockall score 112–14
neuroendocrine tumours 203 Royal College of Anaesthetists 102
rectal cancer 250, 251–4 RTOG trial 260
randomization 4
analysis 18 sacral neuromodulation for faecal
causal modelling 20 incontinence 231–3
crossover studies 9 safety issues, RCTs 16
stepped wedge studies 10 analysis sets 19
see also randomized controlled trials sample size 12–13
randomized controlled trials (RCTs) 3–4 San Francisco County, California, trauma care
analysis sets 18–19 system 370–2
blinding 14 Scotland, national registry data 11
cluster 9 SCOTT project 12
delivering the intervention 16–17 Scottish Health Informatics Programme
future 23 (SHIP) 11
governance 16 segmental colectomy in left-sided malignant large
missing data 21 bowel obstruction 86–8
398 SUBJECT INDEX
self-expanding metal stents (SEMS), dysphagia Swedish Obese Subjects (SOS) study 109–11
palliation 152–4 systematic reviews 22
Sendai Guidelines, management of intraductal
papillary mucinous neoplasm 200 tacrolimus in renal transplantation 298, 299,
sentinel node biopsy (SNB) in breast cancer 324–6 300, 301
axillary dissection versus no axillary ABO-incompatible transplantation 313
dissection 327, 328 tamoxifen in breast cancer 349
sepsis as adjuvant therapy 343–6
goal-directed therapy in severe 99–101 aromatase inhibitors versus 347–8
post-operative complications, and enteral versus and chemotherapy 345–6
parenteral nutrition 95–7 ductal carcinoma in situ 330, 331
septic shock 101 taxanes in breast cancer 341–2, 356
severe limb ischaemia (SLI), BASIL trial 281–4 thyroid surgery
short course pre-operative radiotherapy (SCPRT) in papillary thyroid carcinoma, management of the
rectal cancer 250, 252–4 N0 neck 365–8
short gastric vessel (SGV) division, laparoscopic recurrent laryngeal nerve monitoring 362–4
anti-reflux surgery and para-oesophageal tibial nerve stimulation for faecal
repair 123, 124–5, 126 incontinence 233
Shouldice groin hernia repair 40 torso injuries, hypotensive trauma patients
sigmoidoscopy, flexible, in bowel cancer with 376–8
screening 236, 237–88, 239 total enteral nutrition (TEN), post-operative septic
sinus formation following abdominal wall complications 95–8
closure 33 total mesorectal excision (TME) for rectal
sirolimus in renal transplantation 298, 299, 300, 301 cancer 240–3, 247, 249
skin preparation, pre-surgical 26–7 circumferential resection margin 246
small bowel obstruction (SBO), water-soluble neoadjuvant thrapy 250, 251, 252, 253, 254
contrast in 83–5 total parenteral nutrition (TPN)
small-incision cholecystectomy 160–2 acute pancreatitis 171–4
small studies 12–13 post-operative septic complications 95–8
specialist surgical registries 11 totally extra-peritoneal (TEP) groin hernia
specialization repair 40, 41
and patient outcome 54 Toupet fundoplication in gastro-oesophageal reflux
rectal cancer surgery 242 disease 126
standardization of interventions 16–17 trachea-oesophageal fistula 154
stapled anastomosis in colorectal surgery 212–13 training
stapled haemorrhoidectomy 226–8 laparoscopic cholecystectomy 164
static cold storage in deceased donor laparoscopic versus open appendicectomy 70
transplantation 302–6 and patient outcome 54
statins see also learning curves
abdominal aortic aneurysms 280 tranexamic acid (TA) in the bleeding trauma
carotid endarterectomy 265, 271 patient 379–81
WOSCOPS 11 trans-abdominal pre-peritoneal (TAPP) groin hernia
status change forms 8 repair 40, 41
step-up approach, necrosectomy 181–3 trans-hiatal oesophagectomy (THO)
stepped wedge study design 9–10 laparoscopic 141
STO3 study 147 trans-thoracic oesophagectomy versus 132–5
strangulated groin hernias transient ischaemic attack (TIA) 265
asymptomatic 37, 38 transplant surgery
symptomatic 34–5 liver 316
subgroup analyses 19 renal
causal modelling 19–20 ABO-incompatible 313–15
subtotal colectomy in left-sided malignant large calcineurin inhibitors, reduced exposure
bowel obstruction 86–8 to 298–301
superficial femoral artery (SFA), intermittent donor nephrectomy, laparoscopic versus
claudication 286 mini-incision open 310–12
superficial venous surgery in leg ulceration 294–6 machine preservation versus static cold
superiority trials 8 storage in deceased donor
surgical volume and patient outcome 53–4 transplantation 302–6
oesophagectomy 53, 54, 129–31 mortality benefit 307–9
sutured versus biological mesh repair of para- trans-thoracic oesophagectomy (TTO) 132–5
oesophageal hernia 123, 125, 126 trastuzumab in breast cancer 352–4
SUBJECT INDEX 399
trauma surgery varicose veins, endovenous therapy 290–3

damage control surgery 373–5 vascular surgery
hypotensive patients with penetrating torso bypass or angioplasty for severe limb ischaemia
injuries 376–8 (BASIL) trial 281–4
systems of trauma care 370–2 carotid atherosclerosis, symptomatic 262–5
tranexamic acid in the bleeding trauma carotid endarterectomy, general versus local
patient 379–81 anaesthesia 270–1
treatment effects carotid stenosis, asymptomatic 266–9
pilot studies 12 EVAR 1 trial 275–7
underpowered studies 13 EVAR 2 trial 278–80
trial committees 16 intermittent claudication 285–9
Trial Steering Committees (TSCs) 16 leg ulceration, superficial venous surgery 294–6
tumour necrosis factor alpha (TNF-α) in Crohn’s UK Small Aneurysm Trial 272–4
disease 221–3 varicose veins, endovenous therapy 290–3
venous thromboembolism (VTE) prophylaxis 28–9
ulceration, leg 294–6 ventral mesh rectopexy for rectal prolapse 229–30
ulcerative colitis 223 Veterans Affairs 263–4
ultrasound video-assisted retroperitoneal debridement (VARD)
abdominal aortic aneurysms 272, 273 in necrotizing pancreatitis 182
acute abdomen 63 volume of surgical procedures and patient
biliary injuries 165 outcome 53–4
complicated appendicitis 73 oesophagectomy 53, 54, 129–31
laparoscopic cholecystectomy 167 vulnerable groups, ethical issues 7
necrotizing pancreatitis 183
ultrasound-guided foam sclerotherapy (UGFS) 292 water-soluble contrast (WSC) in small bowel
underpowered studies 12–13 obstruction 83–5
United Kingdom Clinical Research Collaboration West of Scotland Coronary Prevention Study
see National Institute for Health Research (WOSCOPS) 11
(NIHR) Health Technology Assessment (HTA) withdrawal in surgical trials 7–8
programme World Health Organization (WHO) 203
United States Food and Drug Administration wound infection
(FDA) 8 abdominal wall closure 33
upper gastrointestinal haemorrhage, risk scoring antibiotic prophylaxis in inguinal hernia
for 112–14 repair 49–50
ingrown toenails surgery 51, 52
vagotomy for perforated duodenal ulcers 82 surgical skin preparation 26–7
validity of studies 13–14
missing data 20 Z0011 trial 328

4 5800639631472986150

Uploaded by

Copyright:

Available Formats

4 5800639631472986150

Uploaded by

Document Information

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

4 5800639631472986150

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What information is provided on page 3?

What information is provided on page 3?

Landmark Papers in

1 Methodological considerations in surgical trials 1

Appendix: Levels of evidence 382

Gavin J. Bryce Julie C. Doughty

Iain R. Macpherson Patrick J. O’Dwyer

5-FU 5-fluorouracil CHD common hepatic duct

ECMO extra-corporeal membrane HR hazard ratio

MRA magnetic resonance angiography PV portal vein

UGFS ultrasound-guided foam VTE venous thromboembolism

Evidence-based medicine (EBM) is one of the cornerstones of effective clinical practice.

1.2 Designs for surgical trials

Logical Highly Highly Few

Fig. 1.1 Drug trials.

Observational versus randomized

Pragmatic versus explanatory

Prospective versus retrospective

Cross-sectional versus longitudinal

Fig. 1.2 Levels of complexity of surgical trials.

1.3 Specific design issues in surgical trials

Placebo surgery trials

Withdrawal in surgical trials

More complex designs

◆ The structure of the hypotheses is reversed in a non-inferiority trial—the null

Cluster randomized control trials

Sample size—the difficulty of underpowered studies

it affords an improvement in recurrence or even survival may need several thousand to

Internal and external validity

Masking randomized treatment (‘blinding’)

Maintaining the blind

Blinded assessment of outcomes

1.4 Issues in the conduct of randomized controlled trials

Delivering the intervention

1.5 Analysis of surgical trials

Intervention allocated versus intervention received

‘Crossover’ or rescue surgery

Avoiding missing data

Monitoring missing data

Missing completely at random, missing at random,

1.6 Other issues in surgical trials

Systematic reviews and meta-analysis

2.1 Surgical skin preparation

◆ A special applicator was used to clean using the 2% chlorhexidine gluconate–70%

◆ The relative risk of developing any post-operative wound infection in the

2.2 Venous thromboembolism (VTE) prophylaxis

◆ The treatment group received heparin 5000U subcutaneously 2 hours pre-operatively

Control (n = 2076) Low dose heparin p value

2.3 Intravenous fluid therapy after major abdominal surgery

Standard fluids Fluid restriction p value

2.4 Abdominal wall closure

2.5 Symptomatic groin hernias

2.6 Asymptomatic groin hernias

◆ Inclusion criteria were men with asymptomatic inguinal hernias

2.7 Laparoscopic versus open groin hernia repair

◆ Life -threatening complications occurred in 1.1% of laparoscopic cases and 0.1% of

2.8 Local or general anaesthesia for groin hernia repair