Prevalence, clinical profile, and cardiovascular

outcomes of atrial fibrillation patients

with atherothrombosis
Shinya Goto, MD, PhD, a Deepak L. Bhatt, MD, b Joachim Röther, MD, c Mark Alberts, MD, d
Michael D. Hill, MD, e Yasuo Ikeda, MD, f Shinichiro Uchiyama, MD, PhD, g Ralph D’Agostino, PhD, h
E. Magnus Ohman, MD, i Chiau-Suong Liau, MD, PhD, j Alan T. Hirsch, MD, k Jean-Louis Mas, MD, l
Peter W.F. Wilson, MD, m Ramón Corbalán, n Franz Aichner, MD, o and P. Gabriel Steg, MD, p
on behalf of the REACH Registry Investigators q Kanagawa and Tokyo, Japan; Boston, MA; Minden, Germany;
Chicago, IL; Calgary, Alberta, Canada; Durham, NC; Taipei, Taiwan; Minneapolis, MN;
Paris, France; Atlanta, GA; Linz, Austria; and Santiago, Chile

Background Atrial fibrillation (AF) is a major risk factor (RF) for ischemic stroke. Its prevalence and prognostic impact in
patients with atherothrombosis are unclear.
Methods Risk factors, drug usage, and 1-year cardiovascular (CV) outcomes (CV death, myocardial infarction [MI],
and stroke) were compared in AF and non-AF patients from the REduction of Atherothrombosis for Continued Health (REACH)
Registry, an international, prospective cohort of 68,236 stable outpatients with established atherothrombosis or ≥3
atherothrombotic RFs.
Results Atrial fibrillation and 1-year follow-up data are available for 63,589 patients. The prevalence of AF was, 12.5%,
13.7%, 11.5%, and 6.2% among coronary artery disease, CV disease, peripheral artery disease, and RF-only patients,
respectively. Of the 6,814 patients with AF, 6.7% experienced CV death, nonfatal MI, or nonfatal stroke within a year.
The annual incidence of nonfatal stroke (2.4% vs 1.6%, P b .0001) and unstable angina (6.0% vs 4.0%, P b .00001) was
higher, and CV death was more than double (3.2% vs 1.4%, P b .0001), in AF versus non-AF patients. In these patients with or
at high risk of atherothrombosis, most patients with AF received antiplatelet agents, but only 53.1% were treated with oral
anticoagulants. Even with high CHADS2 (congestive heart failure, hypertension, aging, diabetes mellitus, and stroke)
scores, anticoagulant use did not exceed (59%). The rate of bleeding requiring hospitalization was higher in AF versus non-AF
patients (1.5% vs 0.8%, P b .0001), possibly related to the more frequent use of anticoagulants (53.1% vs 7.1%).
Conclusions Atrial fibrillation is common in patients with atherothrombosis, associated with more frequent fatal and
nonfatal CV outcomes, and underuse of oral anticoagulants. (Am Heart J 2008;156:855-863.e2.)

Atrial fibrillation (AF) is a major risk factor for ischemic disease, and CV risk factors such as hypertension,
stroke, congestive heart failure, and mortality.1-3 The diabetes,7 obesity,8 and insulin resistance.9 The preva-
prevalence of AF is influenced by age, gender,4-6 lence of AF in the general population is estimated to
cardiovascular disease (CVD) such as valvular heart increase from 2.3 million in 2001 to 5.6 million in 2050 in

From the aDepartment of Medicine and Metabolic Disease Center, Tokai University School of Medicine, Atlanta, GA, nAcademic Teaching Hospital Wagner-Jauregg, Linz, Austria,
o
of Medicine, Department of Metabolic System Medicine, Research Institute of Medicine, Faculty of Medicine, Pontifica Universidad Catolica de Chile, Santiago, Chile, and
Tokai University, Kanagawa, Japan, bVA Boston Healthcare System and Brigham and p
INSERM U-698 et Université Paris VII-Denis Diderot, Hôpital Bichat-Claude Bernard,
Women Hospital, Boston, MA, cDepartment of Neurology, Klinikum Minden, Minden, Paris, France.
Germany, dDepartment of Neurology, Northwestern University Medical School, Chicago, q
For a list of REACH Registry Global Publication Committee members see Appendix A
IL, eDepartment of Clinical Neurosciences, Foothills Medical Centre, Calgary, Alberta, (available online).
Canada, fDepartment of Internal Medicine, Keio University School of Medicine, Tokyo, Author conflict of interest information is available in Appendix A available online.
Japan, gDepartment of Neurology, Tokyo Women's Medical College, Tokyo, Japan, Submitted March 26, 2008; accepted June 22, 2008.
h
Statistics and Consulting Unit, Boston University, Boston, MA, iDivision of Cardiology, Reprint requests: Shinya Goto, MD, Department of Medicine, Tokai University School of
Duke University, Durham, NC, jDepartment of Internal Medicine, National Taiwan Medicine, Isehara 2591143, Japan.
University Hospital and School of Medicine, Taipei, Taiwan, kMinneapolis Heart Institute E-mail: shinichi@is.icc.u-tokai.ac.jp
Foundation and Division of Epidemiology and Community Health, University of Minnesota 0002-8703/$ - see front matter
School of Public Health, Minneapolis, MN, lService de Neurologie, Centre Raymond © 2008, Mosby, Inc. All rights reserved.
Garcin, Hôpital Sainte-Anne, Paris, France, mCardiology Division, Emory University School doi:10.1016/j.ahj.2008.06.029
 American Heart Journal
856 Goto et al November 2008

the United States.10 However, the prevalence of AF Table I. Baseline demographics of patients with versus without
among patients with or at high risk of atherothrombosis is history of AF
still unknown.
AF + AF -
The REduction of Atherothrombosis for Continued
Variable (n = 6814) (n = 56 775) P
Health (REACH) Registry is a large, contemporary,
representative, and geographically diverse cohort of Patient characteristic
stable outpatients with or at high risk of atherothrombo- Mean age (SD), y 72.83 68.06 b.0001
sis.11,12 A total of 68,236 patients with either established (9.15) (10.08)
atherothrombotic disease (coronary artery disease [CAD], Gender (% male) 64.39 63.71 .2694
DM (%) 40.93 44.21 b.0001
peripheral artery disease [PAD], cerebrovascular disease Hypertension (%) 84.40 81.35 b.0001
[CVD]; n = 55,814) or at least 3 risk factors for Hypercholesterolemia (%) 66.84 72.74 b.0001
athrothrombosis (n = 12,422) were enrolled from 5,587 Waist circumference 99.00 97.76 b.0001
physician practices in 44 countries (in Europe, North and (SD, cm) (16.62) (15.98)
South America, and Asia) in 2003 to 2004. The primary Obesity (BMI N30, %) 29.83 29.89 .9187
BMI (SD) 28.05 28.11 .4313
aim of the REACH Registry was to describe demographic (5.59) (5.62)
characteristics and management and determine the risk Current smoker (%) 9.75 15.87 b.0001
of CV events in the global population and in each clinical Former smoker (%) 45.71 41.19 b.0001
subgroup. Secondary aims were to study specific Previous history of
atherosclerotic disease
subpopulations of the REACH population.11-13
History of CAD (%) 69.34 58.12 b.0001
The aim of the analysis in this article was to analyze Stable angina with 38.47 28.87 b.0001
the prevalence and prognostic impact of AF among documented CAD (%)
patients from the REACH Registry, focusing on risk factor Unstable angina with 16.14 12.16 b.0001
profiles and control, medication use, and 1-year rates of documented CAD (%)
History of PCI (%) 24.23 25.24 .0704
CV death, MI, and stroke in AF and non-AF patients. We
History of CABG (%) 27.08 19.51 b.0001
assessed the distribution of the CHADS2 (congestive History of MI (%) 37.54 30.83 b.0001
heart failure, hypertension, aging, diabetes mellitus, and History of CVD (%) 35.34 26.73 b.0001
stroke) score in that population and its relation to History of TIA (%) 18.90 12.21 b.0001
cerebrovascular outcomes.14 CHADS2 is a recognized History of stroke (%) 24.77 19.65 b.0001
History of PAD (%) 12.99 12.03 .0224
clinical prediction rule in estimating the risk of stroke in Claudication and history of 1.95 1.74 .2177
patient with AF,15 as well as for predicting various CV amputation (%)
outcomes (including CV death, MI, stroke, and com- History of peripheral 6.91 6.49 .1832
bined events) in patients with and at high risk angioplasty/stenting/surgery
of atherothrombosis. Claudication and ABI 14.17 13.65 .3562
b0.9 (%)
Three risk factors only (%) 10.46 19.11 b.0001
Methods Any history of symptomatic 89.54 80.89 b.0001
atherothrombosis (%)
Database of the REACH Registry CHADS2 score classification b.0001
This study was conducted using data from the REACH 0 2.82 7.09
Registry. The study design,12 baseline characteristics, and main 1 14.82 24.34
1-year outcomes have been published previously.11,13 Briefly, 2 26.34 33.20
the REACH Registry is a large-scale, prospective, international 3 26.80 20.12
4 17.41 11.60
cohort of stable outpatients aged ≥45 years with either
5 9.07 3.18
established atherothrombotic disease (CAD, CVD, and/or PAD)
6 2.74 0.47
or ≥3 risk factors for atherothrombosis (risk factors only [RFO]).
The enrollment criteria were predefined and have been BMI, Body mass index; PCI, percutaneous coronary intervention; CABG, coronary
artery bypass graft; ABI, ankle-brachial index.
published elsewhere.12 Exclusion criteria were limited to
participation in a clinical trial or anticipated difficulty in
returning for a follow-up visit to ensure a broad representation a history of AF, the patients were classified as “unknown” and
of the population. After approval of the study by the institutional excluded from this analysis.
review board in each country or hospital according to local Baseline demographic characteristics, previous history of
requirements, and obtaining written informed consent from atherothrombotic disease, and medication use at baseline were
each patient, data were collected centrally using a standardized documented. Patients with AF were also stratified using the
case report form. CHADS215 risk score in which one point is assigned to patients
with a history of congestive heart failure, hypertension, age
Definition of AF and patient follow-up ≥75 years, diabetes mellitus (DM), and 2 points for a history of
All patients were classified based on presence or absence of stroke or transient ischemic attack (TIA).15
AF, at the time of enrollment, as previously published. If At 12 ± 3 months from enrollment, data were collected
participating physicians could not confirm whether patients had regarding clinical outcomes, endovascular procedures, as well
American Heart Journal
Volume 156, Number 5
Goto et al 857

Table II. Baseline demographics and risk factor profile of patients with versus without history of AF in disease categories
RFO (n = 11 563) CAD (n = 37 724) CVD (n = 17 582) PAD (n = 7716)

AF+ AF− AF+ AF− AF+ AF− AF+ AF−

Sample size (n) 713 10 850 4725 32 999 2408 15 174 885 6831
Gender (male %) 53.7 49.2 68.2 70.2 59.9 59.5 72.6 70.6
P = .0205 P = .0056 P = .7217 P = .2132
Age (mean ± SD) 73.6 ± 8.3 68.8 ± 9.9 72.5 ± 9.3 67.7 ± 10.1 73.4- ± 9.2 68.9 ± 10.1 73.2 ± 9.3 68.8 ± 9.8
P b .0001 P b .0001 P b .0001 P b .0001
DM (%) 70.7 75.2 38.7 37.7 37.3 36.7 45.0 43.3
P = .0075 P = .1788 P = .5749 P = .3334
Hypertension (%) 92.7 90.0 83.9 79.5 84.5 82.9 86.6 80.6
P = .02 P b .0001 P = .0513 P b .0001
Hypercholesterolemia (%) 77.8 82.1 71.2 77.6 56.0 58.0 63.9 66.9
P = .0039 P b .0001 P = .0673 P = .0741
Waist circumference (cm) 103.30 (18.98) 100.2 (17.04) 99.5 (16.3) 98.3 (16.4) 96.8 (15.99) 95.4 (15.74) 98.3 (16.37) 97.6 (15.54)
P b .0001 P b .0001 P b .0001 P = .2753
Obesity (%) (BMI N30) 42.1 42.1 30.1 29.5 24.9 23.2 25.2 23.5
0.9771 0.4173 0.0620 0.2743
BMI (mean ± SD) 29.9 ± 6.2 29.9 ± 6.4 28.1 ± 5.5 28.1 ± 5.4 27.4 ± 5.6 27.1 ± 5.2 27.2 ± 5.1 27.1 ± 5.2
P = .6823 P = .9095 P = .0273 P = .7795
Current smoking (%) 11.3 19.8 9.4 13.5 9.5 15.1 15.8 25.4
P b .0001 P b .0001 P b .0001 P b .0001
Former smoking (%) 36.2 27.9 49.2 46.8 41.0 38.2 53.6 50.8
P b .0001 P = .0017 P = .0082 P = .1318

as chronic medications taken regularly since baseline. Clinical

Figure 1
outcomes of AF patients were compared with non-AF patients.
The current report is based on a database lock of July 30, 2006,
for analysis of the 1-year follow-up.

Statistical analysis
Continuous variables are expressed as mean (SD), and
categorical variables are expressed as frequencies and
percentages unless otherwise specified. Comparisons
between data in AF and non-AF patients were conducted
with Pearson χ2 for categorical variables and with Student
t test for continuous variables. Statistical analysis was
performed using SAS v.9 software (SAS Institute Inc, Cary,
NC). For analysis of 1-year CV events, all rates are reported
after adjustment for age, gender, and risk factors (smoking,
diabetes, hypertension, hypercholesterolemia) using a Cox
model. To take into account the multiplicity of tests
comparing the event rates between AF and non-AF patients, a
step-down Bonferroni adjustment was used within each Event rates for CV death/MI and stroke of patients with versus without
population (RFO, CAD, CVD, PAD).16 history of AF (adjusted for age, sex, smoking, diabetes, hypertension,
hypercholesterolemia). Combined event of CV death and/or nonfatal
MI and/or nonfatal stroke in patients with AF versus patients without
Results history of AF are shown after adjustment of age, gender and classic
Baseline characteristics of AF and non-AF patients risk factors. AF, Patients with a history of AF; non-AF, patients without
Among 68,236 patients initially recruited to the history of AF.
REACH Registry, 1-year follow-up was available for
64,977 patients of whom 63,589 patients had infor-
mation available regarding AF or non-AF at baseline
and represent the analysis sample.12 The prevalence of whereas it was 6.2% among patients with risk factors
AF in the patient groups with established athero- for atherothrombosis. The prevalence of AF in the total
thrombotic disease was 12.5%, 13.7%, and 11.5% for symptomatic atherothrombotic group (CAD, CVD, or
patients with CAD, CVD, and PAD respectively, PAD) was 11.7%.
 American Heart Journal
858 Goto et al November 2008

Table III. One-year CV outcomes of patients with AF versus those without AF (adjusted for age, sex, smoking, hypertension,
diabetes, hypercholesterolemia)
Total (n = 63 589) RFO (n = 11 563)

Variable AF+ AF− P⁎ AF+ AF− P⁎

n 6814 56 775 713 10 850

All-cause mortality 4.27 2.32 b.0001 2.80 1.34 .0051
Major CV events
CV death 3.16 1.42 b.0001 1.36 0.63 .0476
Nonfatal MI 1.36 1.11 .1205 1.18 0.74 1.0000
Nonfatal stroke 2.43 1.55 b.0001 0.39 0.85 1.0000
CV death/MI/stroke 6.66 3.88 b.0001 2.75 2.06 .7249
CV death/MI/stroke or hospitalization 17.88 12.09 b.0001 7.80 5.06 .0404
for atherothrombotic event(s)
Other events
Unstable angina 5.95 4.04 b.0001 1.49 1.17 1.0000
TIA 2.24 1.28 b.0001 0.42 0.67 1.0000
Other ischemic arterial event 1.99 1.25 b.0001 1.23 0.47 .1863
Congestive heart failure 8.32 2.74 b.0001 6.87 1.58 b.0001
Bleeding 1.51 0.77 b.0001 0.54 0.56 1.0000
Coronary angioplasty/stenting 2.80 2.56 .3259 1.51 0.81 .7249
CABG 1.21 1.00 .3259 0.40 0.52 1.0000
Carotid angioplasty/stenting 0.39 0.26 .3259 0.70 0.13 .0308
Peripheral bypass graft 1.03 0.66 .0129 0.26 0.19 1.0000
Peripheral angioplasty/stenting 1.27 0.99 .2069 0.34 0.40 1.0000
Amputation 0.44 0.30 .3259 0.29 0.22 1.0000
CAD (n = 37 724) CVD (n = 17 582) PAD (n = 7716)
⁎Step-down Bonferroni.

AF+ AF− P⁎ AF+ AF− P⁎ AF+ AF− P⁎

The baseline characteristics of patients with AF versus One-year outcome of AF and non-AF patients
those without AF were compared in the overall After adjustment for age, gender, and risk factors, the
population (Table I) and in each of the main subgroups presence of AF at baseline was associated with higher
(CAD, CVD, PAD, and RFO) (Table II). Overall, patients rates of adverse CV outcomes (combined CV death/MI/
with AF were older, had a higher prevalence of stroke) at 1-year follow-up (Figure 1). The all-cause
hypertension, but had lower prevalence of DM and (4.3%) and CV mortality (3.2%) rates of patients with AF
hypercholesterolemia (P b .0001). In addition, patients were substantially higher than in non-AF patients (2.3%
with AF had a larger waist circumference than non-AF and 1.4%, respectively; P b .0001) (Table III). This
patients (P b .0001), although there was no difference higher mortality with AF patients was consistently
in body mass index between the 2 groups (P = .4313). observed across all subgroups with established athero-
Despite the higher prevalence of former smoking in thrombosis (CAD, CVD, PAD) or at risk for athero-
patients with AF, current smoking was significantly thrombosis (Table III).
lower in AF versus non-AF patients (P b .0001) (Table Although numerically higher with AF, there was little
II). There were substantial differences in the risk factor difference in the rate of nonfatal MI between AF
profiles of AF versus non-AF patients in each established and non-AF patients (P = .1205). Conversely, and as
atherothrombotic lesion group (Table II). Indeed, lower expected, the rate of nonfatal stroke was higher
prevalence of DM in AF patients was true only for the (P b .0001) in AF (2.4%) versus non-AF patients (1.6%).
total study group and the subgroup of RFO. Lower The rate of nonfatal stroke was highest in AF patients
prevalence of hypercholesterolemia in AF patients was initially enrolled in REACH because of prior CVD
shown both in RFO and CAD subgroups, but not in the (4.9%). The rate of TIA was also higher in AF versus
CVD and PAD subgroups. non-AF patients (2.2% vs 1.3%, respectively; P b .0001)
A previous history of CVD was more common in AF (Table III). This result was consistent across all
patients (35.3% vs 26.7%, respectively; P b .0001) for both established atherothrombotic disease subgroups (CAD,
stroke (24.8% vs 19.7%, respectively; P b .0001) and TIA CVD, PAD) but not for the asymptomatic RFO
(18.9% vs 12.2%, respectively; P b .0001) (Table I). The patient group.
same was true for patients with a history of PAD (13.0% vs The combined end point of CV death/MI/stroke and/or
12.0%, P = .0224) and patients with a history of CAD hospitalization for atherothrombotic event was higher in
(69.3% vs 58.1%, P b .0001). AF patients across all categories (17.9% vs 12.1%,
American Heart Journal
Volume 156, Number 5
Goto et al 859

Table III (continued)

CAD (n = 37 724) CVD (n = 17 582) PAD (n = 7716)

AF+ AF− P⁎ AF+ AF− P⁎ AF+ AF− P⁎

4725 32 999 2408 15 174 885 6831

4.37 2.65 b.0001 4.90 2.94 b.0001 5.60 3.80 .1331
Major CV events
3.42 1.69 b.0001 3.54 1.88 b.0001 4.29 2.48 .0179
1.61 1.40 .6647 1.14 0.93 .9068 1.59 1.37 1.0000
2.24 1.24 b.0001 4.89 3.50 .0309 2.78 1.67 .3675
6.92 4.13 b.0001 9.16 6.07 b.0001 7.66 5.18 .0228
19.70 14.52 b.0001 20.76 13.41 b.0001 27.06 21.37 .0082

Other events
7.67 6.14 .0023 6.05 2.91 b.0001 6.61 4.05 .0149
2.22 1.06 b.0001 4.47 3.06 .0048 2.74 1.75 .4226
2.08 1.37 .0033 2.33 1.41 .0091 4.93 4.08 1.0000
9.38 3.76 b.0001 8.51 2.62 b.0001 10.44 3.62 b.0001
1.49 0.81 b.0001 1.91 0.82 b.0001 2.10 1.30 .4226
3.53 3.88 .9656 2.15 1.37 .0309 3.93 2.27 .0654
1.63 1.39 .9656 1.31 0.65 .0112 0.75 1.06 1.0000
0.40 0.29 .9656 0.33 0.35 .9068 0.45 0.57 1.0000
1.07 0.57 .0026 1.07 0.43 .0030 5.13 3.88 .8109
1.26 0.93 .2361 1.49 0.74 .0041 5.42 5.30 1.0000
0.29 0.23 .9656 0.59 0.17 .0030 2.33 1.59 0.9108

respectively; P b .0001), and highest in patients initially were treated with an antiplatelet agent (Table IV).
recruited with PAD (27.1%). Importantly, 16.9% of patients with AF received a
In comparison to non-AF patients, a sharp increase in combination of oral anticoagulants and antiplatelet
the incidence of heart failure (8.3% vs 2.7%, respectively; agents; despite this, almost half of patients did not receive
P b .0001), unstable angina (6.0% vs 4.0%, respectively; oral anticoagulants although these patients were at high
P b .0001), and bleeding requiring hospitalization and risk of stroke. There were modest differences between AF
transfusion (1.5% vs 0.8%, P b .0001) was observed in and non-AF patients in the use of antihypertensive agents
patients with AF in the year after the baseline visit. These among patients with hypertension, lipid-lowering agents
differences were consistent across subgroups (Table III). among patients with hypercholesterolemia, and antidia-
Revascularization procedures—including coronary betic agents among diabetic patients (Table IV).
angioplasty/stenting, coronary artery bypass grafting, and Lower serum cholesterol levels were noted in AF versus
carotid angioplasty/stenting—were generally more fre- non-AF patients with hypercholesterolemia (193.2 ±
quently performed during follow-up among AF patients 47.6 vs 197.3 ± 48.2 mg/dL, P b .0001), with lower use of
than non-AF patients. Likewise, the rates of peripheral lipid-lowering agents (87.7% vs 89.8%, P b .0001) and
artery bypass graft and peripheral angioplasty/stenting statins (81.5% vs 83.1%, P b .0001) (Table IV). Systolic and
were higher in patients with AF in the CAD and CVD diastolic blood pressures for patients with AF presenting
subgroups (Table III). with hypertension were lower than in non-AF patients
(Table IV). Fewer patients with AF presented with DM
Antithrombotic therapy and risk factor control in and were prescribed antidiabetic agents, and blood
patients with AF glucose levels were significantly lower in AF versus non-
Medication use and the achievement of risk factor AF patients (141.9 ± 52.0 vs 146.1 ± 52.5 mg/dL,
control in AF and non-AF patients in the whole study respectively; P = .0001) (Table IV).
population are summarized in Table IV. Use of antith-
rombotics was high among patients with AF, with Use of antithrombotic agents and CV outcome in AF
approximately 95% receiving an antithrombotic agent. patients classified with CHADS2 scoring
Approximately half (53.1%) were treated with oral Approximately 5% of patients with AF received no
anticoagulants (the use of oral anticoagulants in non-AF antithrombotic therapy at all and that proportion was
patients was 7.1%), and approximately 6 patients in 10 evenly distributed across CHADS2 score categories. The
 American Heart Journal
860 Goto et al November 2008

Table IV. Medication use and achievement of risk factor control

use of anticoagulants, whether alone or in combination
at 1 year in AF versus non-AF patients with antiplatelet agents, increased with higher CHADS2
scoring, from 44.7% in CHADS2 score 0 to 60.0% in
AF + AF −
CHADS2 score 4 patients. Approximately 15% of patients
Variable (n = 6814) (n = 56 775) P⁎
with AF were receiving a combination of antiplatelet and
Antithrombotic agents anticoagulant agents, and the proportion was similar
Any antithrombotic agent (%) 94.90 85.28 b.0001 across all CHADS2 categories (Table V).
Any antiplatelet agent (%) 58.39 81.10 b.0001 Analysis of the CHADS2 score in patients with AF
Antiplatelet agents only (%) 40.14 75.07 b.0001 demonstrated that CV events—especially nonfatal stroke,
Oral anticoagulants only (%) 36.17 3.83 b.0001
Oral anticoagulant and any 16.92 3.29 b.0001
and the combined end point of CV death/MI/stroke—
antiplatelet agent (%) were more frequent in the higher CHADS2 score
Aspirin alone (%) 49.51 69.44 b.0001 subgroup (Figure 2). The rate of CV death/MI/stroke
Other antiplatelet agent (%) 18.63 25.35 b.0001 varied from 3.9% in the CHADS2 score 0 group to 12.4% in
Any 2 antiplatelet agent (%) 9.66 13.58 b.0001 the CHADS2 score 6 group. There was a linear relation-
Lipid-lowering agents among
patients with
ship between the nonfatal stroke rate and CHADS2 scores
hypercholesterolemia (P b .0001) but no relation between the rates of nonfatal
At least one lipid-lowering 87.74 89.77 b.0001 MI and CHADS2 score (P = .7388).
agent (%)
Statins (%) 81.49 83.13 .0030
Other lipid-lowering 12.62 14.31 .0010
agents (%)
Discussion
Blood pressure–lowering agents The primary finding of this study is the high prevalence
among patients with HTN of AF among patients with atherothrombosis, ranging
At least one CV agent (%) 97.58 95.67 b.0001 from 11.5% in patients with PAD to 13.7% in patients with
β-Blockers (%) 56.97 47.99 b.0001
CVD. Indeed, the prevalence of AF in patients with
ACE inhibitors (%) 51.98 47.81 b.0001
Angiotensin receptor 23.17 25.56 b.0001 symptomatic atherothrombotic disease (CAD, CVD, or
blocker (%) PAD) was 11.7%. Even among RFO patients, the
Diuretics (%) 61.86 41.84 b.0001 prevalence of AF (6.2%) is substantially higher than the
Calcium channel 36.64 37.45 .2177 estimated prevalence in the general population aged
blockers (%)
Nitrates (%) 31.31 23.48 b.0001
40 years and older (2.3%) and also in the population aged
Other antihypertensives (%) 13.48 10.11 b.0001 65 years and older (5.9%).4 This higher than expected
Peripheral arterial 7.37 6.67 .0449 prevalence of AF among patients with atherothrombosis
claudication medications (%) has clinical implications because atherothrombotic
Antidiabetic agents among patients usually require chronic antiplatelet therapy and
patients with DM
At least one antidiabetic 81.76 86.85 b.0001
frequently protracted periods of dual antiplatelet therapy
agent (%) after acute coronary syndromes or percutaneous coron-
Sulfonylureas (%) 39.26 43.51 b.0001 ary intervention.17-20 As such, the fact that 12.5% of
Biguanides (%) 31.89 41.30 b.0001 patients with chronic CAD have AF highlights the
Insulin (%) 27.03 25.46 .0648
magnitude of this problem. Yet, prospective trials have
Thiazolidinediones (%) 13.07 17.46 b.0001
Other antidiabetic agents (%) 9.28 11.10 .0030 not addressed the optimal management of this subgroup
Blood pressure, cholesterol, of patients with respect to the combined use of oral
and sugar level anticoagulants and antiplatelet agents. The Atrial Fibrilla-
SBP (mm Hg) 138.65 140.69 b.0001 tion Clopidogrel Trial with Irbesartan for prevention of
(19.91) † (18.99) †
Vascular Events (ACTIVE-W) trial, however, showed that
DBP (mm Hg) 78.29 79.49 b.0001
(12.11) † (11.20) † antiplatelet therapy alone did not provide adequate
Serum cholesterol 193.15 197.33 b.0001 protection against the risk of stroke in AF patients,
(mg/dL) (%) (47.61) † (48.23) † compared with warfarin.21
Blood sugar (mg/dL) 141.94 146.11 .0001 Secondly, the study found that AF was associated with a
(51.98) † (52.52) †
major increase in CV mortality and morbidity, even after
HTN, hypertension; ACE, angiotensin-converting enzyme; SBP, systolic blood adjustment for age, gender, and risk factors. Although the
pressure; DBP, diastolic blood pressure.
⁎P value was calculated using Pearson χ2.
increase in mortality, risk of stroke, and TIA were
†SD. anticipated,1,22-24 it is also striking that unstable angina
was more frequent among patients with AF, particularly
for patients who were enrolled in REACH because of
use of antiplatelet therapy without oral anticoagulants prior CVD or PAD.
ranged from 50.5% in patients with CHADS2 scores of 0 to Finally, this study shows the higher incidence of
1 to 38.3% in patients with CHADS2 scores of 5 to 6. The chronic heart failure (requiring hospital admission)
American Heart Journal
Volume 156, Number 5
Goto et al 861

Table V. Use of antithrombotic agents in patients with AF classified based on CHADS2 scoring
CHADS2 score

0 (n = 192) 1 (n = 1010) 2 (n = 1795) 3 (n = 1826) 4 (n = 1186) 5 (n = 618) 6 (n = 187)

No antithrombotic agents 9 (4.74%) 51 (5.16%) 116 (6.57%) 87 (4.87%) 42 (3.60%) 33 (5.45%) 8 (4.47%)
Antiplatelet agents only (%) 96 (50.53%) 499 (50.46%) 745 (42.21%) 669 (37.48%) 425 (36.45%) 234 (38.68%) 67 (37.43%)
Oral anticoagulant only (%) 54 (28.42%) 278 (28.11%) 615 (34.84%) 717 (40.17%) 488 (41.85%) 236 (39.01%) 76 (42.46%)
Combination of antiplatelet and 31 (16.32%) 161 (16.25%) 289 (16.28%) 312 (17.45%) 211 (18.02%) 102 (16.80%) 28 (15.56%)
anticoagulant (%)
Antiplatelet agents or 183 (95.31%) 957 (94.94%) 1669 (93.50%) 1736 (95.23%) 1139 (96.44%) 582 (94.63%) 178 (95.70%)
anticoagulant (%)

Figure 2

Annual CV event risk in AF patients with various CHADS2 (congestive heart failure [C], history of hypertension [H], age N75 years [A], DM [D], or
history of stroke or TIA [S]) scoring (adjusted for age, sex, smoking, diabetes, hypertension, hypercholesterolemia). Annual event rate of CV death,
nonfatal stroke, and combined end point of CV death/nonfatal MI/nonfatal stroke are for patients with higher CHADS2 scoring, whereas the rate
of nonfatal MI was not influenced by CHADS2 scoring.

and severe bleeding in patients with AF at baseline. There is a consensus that the presence of AF is a major
Importantly, there were notable differences in the risk factor for ischemic stroke.25 However, marked
baseline characteristics and risk factor profile of heterogeneity was reported for the risk of ischemic
patients with AF compared with non-AF patients, with stroke in AF patients, ranging from b2% to N10%,
an older age, a higher prevalence of hypertension, depending on the associated risk factors.26 The CHADS2
and a larger waist circumference. These differences score is a simple and practical method for the risk
may contribute to worse CV outcomes for patients stratification of future stroke in patients with AF.15
with AF. However, the validity of CHADS2 score in predicting
 American Heart Journal
862 Goto et al November 2008

future CV events has not been tested in AF patients with of increased risk of subsequent serious CV events,
or at high risk of atherothrombosis. Our study demon- including CV mortality. We should also emphasize that
strates that the CHADS2 score classification was useful in our data and observations pertain to AF patients with
predicting not only stroke but also CV death in stable atherothrombosis and should not be generalized to the
outpatients with established or at high risk of athero- universal AF patient population.
thrombosis. However, it was not as useful in the In conclusion, in this large, global, and contemporary
prediction of nonfatal MI. registry, there is a high prevalence of AF among patients
Although there is a consensus that oral anticoagulants with, or at high risk of, atherothrombosis. These patients
offer better protection than antiplatelets against have a relatively low frequency of use of oral antic-
ischemic stroke in patients with AF,21,25-30 the use of oagulants, although they are at high risk of ischemic
oral anticoagulants in these patients in our study was stroke, probably due to the widespread use of anti-
low even with patients at high risk of stroke—an platelet agents for the treatment of atherothrombosis.
observation consistent with the findings of previous The presence of AF at baseline was associated with
studies, for example, the US Medicare Cohort Study.31 serious and multiple CV events including a higher rate of
The low use of oral anticoagulants in the REACH all-cause and CV mortality, nonfatal stroke, and a modest
population is coupled with a markedly high background increase in the risk of acute coronary events including
use of antiplatelet agents in patients with, or at high non-fatal MI and unstable angina. Clearly, atherothrom-
risk of, atherothrombosis. The increased risk of bleed- bosis patients with AF are a group at particularly high
ing associated with using combined therapy,32 in risk of major adverse cardiac events. Efforts should be
particular aspirin and clopidogrel,33 could account for devoted to improving the care of these patients,
the low use of oral anticoagulants observed. Increased particularly to ensure the appropriate use of oral
risk of bleeding was also associated with aspirin in anticoagulants. There is a need for the optimal
combination with oral anticoagulant in AF patients with antithrombotic therapy among AF patients to be clarified
CAD.34 Clinicians may favor long-term oral antiplatelet to balance the increased risk of thrombotic events and
therapy over oral anticoagulant or combined therapy: the increased risk of bleeding associated with combined
however, unlike previously published smaller studies,35 anticoagulant and antiplatelet therapy.
ACTIVE-W21 have established that oral anticoagulants
provide superior protection against stroke compared References
with antiplatelet therapy, even if the latter combines 1. Benjamin EJ, Wolf PA, D'Agostino RB, et al. Impact of atrial
aspirin and clopidogrel. fibrillation on the risk of death: the Framingham Heart Study.
There are several limitations to this analysis. Approxi- Circulation 1998;98:946-52.
mately 5% of patients were lost to follow-up and an 2. Wachtell K, Hornestam B, Lehto M, et al. Cardiovascular morbidity
and mortality in hypertensive patients with a history of atrial
additional 1,388 patients did not have their AF history
fibrillation: The Losartan Intervention For End Point Reduction in
documented at baseline; however, there were no
Hypertension (LIFE) study. J Am Coll Cardiol 2005;45:705-11.
significant differences in baseline characteristics 3. Miyasaka Y, Barnes ME, Gersh BJ, et al. Coronary ischemic events
between these patient subsets and the rest of the after first atrial fibrillation: risk and survival. Am J Med 2007;120:
cohort.13 Data on the classification of AF (paroxysmal, 357-63.
persistent, or chronic), duration of AF, or use of 4. Feinberg WM, Blackshear JL, Laupacis A, et al. Prevalence, age
antiarrhythmic agents were not available (although a distribution, and gender of patients with atrial fibrillation. Analysis
previous study36 and a meta-analysis demonstrate that and implications. Arch Intern Med 1995;155:469-73.
this does not influence the risk of thromboembolic 5. Heeringa J, van der Kuip DA, Hofman A, et al. Prevalence, incidence
and lifetime risk of atrial fibrillation: the Rotterdam study. Eur Heart J
events37). The presence of AF in the REACH Registry
2006;27:949-53.
population was determined by participating physicians 6. Lloyd-Jones DM, Wang TJ, Leip EP, et al. Lifetime risk for development
and not independently adjudicated. Given the large size of atrial fibrillation: the Framingham Heart Study. Circulation 2004;
of the population studied, some of the differences 110:1042-6.
observed in the baseline characteristics of AF and non-AF 7. Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for
patients, for example, blood pressure, may be statisti- atrial fibrillation in a population-based cohort. The Framingham
cally significant but of little clinical relevance. The Heart Study. JAMA 1994;271:840-4.
REACH Registry is an observational registry, which 8. Frost L, Hune LJ, Vestergaard P. Overweight and obesity as risk
makes it difficult to clarify the causality link between AF factors for atrial fibrillation or flutter: the Danish Diet, Cancer, and
Health Study. Am J Med 2005;118:489-95.
and the increased risk of CV events, as this may be in
9. Ostgren CJ, Merlo J, Rastam L, et al. Atrial fibrillation and its
part be confounded by differences in measured or association with type 2 diabetes and hypertension in a Swedish
unmeasured baseline characteristics. However, regard- community. Diabetes Obes Metab 2004;6:367-74.
less of whether AF is an independent predictor or simply 10. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed
a correlate of higher baseline risk, the presence of AF in atrial fibrillation in adults: national implications for rhythm
patients with atherothrombosis remains a major marker management and stroke prevention: the AnTicoagulation and Risk
American Heart Journal
Volume 156, Number 5
Goto et al 863

Factors in Atrial Fibrillation (ATRIA) Study. JAMA 2001;285: 25. Hart RG, Halperin JL, Pearce LA, et al. Lessons from the Stroke
2370-5. Prevention in Atrial Fibrillation trials. Ann Intern Med 2003;138:
11. Bhatt DL, Steg PG, Ohman EM, et al. International prevalence, 831-8.
recognition, and treatment of cardiovascular risk factors in out- 26. Hart RG, Palacio S, Pearce LA. Atrial fibrillation, stroke, and acute
patients with atherothrombosis. JAMA 2006;295:180-9. antithrombotic therapy: analysis of randomized clinical trials. Stroke
12. Ohman EM, Bhatt DL, Steg PG, et al. The REduction of 2002;33:2722-7.
Atherothrombosis for Continued Health (REACH) Registry: an 27. Mant J, Hobbs FD, Fletcher K, et al. Warfarin versus aspirin for
international, prospective, observational investigation in subjects at stroke prevention in an elderly community population with atrial
risk for atherothrombotic events—study design. Am Heart J 2006; fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged
151:786.e1-786.e10. Study, BAFTA): a randomised controlled trial. Lancet 2007;370:
13. Steg PG, Bhatt DL, Wilson PW, et al. One-year cardiovascular event 493-503.
rates in outpatients with atherothrombosis. JAMA 2007;297: 28. Fuster V, Ryden LE, Cannom DS, et al. ACC/AHA/ESC 2006
1197-206. guidelines for the management of patients with atrial fibrillation: a
14. Laguna P, Martin A, Del Arco C, et al. Differences among clinical report of the American College of Cardiology/American Heart
classification schemes for predicting stroke in atrial fibrillation: Association Task Force on Practice Guidelines and the European
implications for therapy in daily practice. Acad Emerg Med 2005;12: Society of Cardiology Committee for Practice Guidelines (Writing
828-34. Committee to Revise the 2001 Guidelines for the Management of
15. Gage BF, Waterman AD, Shannon W, et al. Validation of Patients With Atrial Fibrillation): developed in collaboration with the
clinical classification schemes for predicting stroke: results from European Heart Rhythm Association and the Heart Rhythm Society.
the National Registry of Atrial Fibrillation. JAMA 2001;285: Circulation 2006;114:e257-e354.
2864-70. 29. Lip GY, Edwards SJ. Stroke prevention with aspirin, warfarin and
16. Holland B, Copenhaver MD. An improved sequentially rejective ximelagatran in patients with non-valvular atrial fibrillation: a
Bonferroni test procedure. Biometrics 1987;43:417-23. systematic review and meta-analysis. Thromb Res 2006;118:
17. Yusuf S, Zhao F, Mehta SR, et al. Effects of clopidogrel in addition to 321-33.
aspirin in patients with acute coronary syndromes without ST-segment 30. Sato H, Ishikawa K, Kitabatake A, et al. Low-dose aspirin
elevation. N Engl J Med 2001;345:494-502. for prevention of stroke in low-risk patients with atrial
18. Mehta SR, Yusuf S, Peters RJ, et al. Effects of pretreatment with fibrillation: Japan Atrial Fibrillation Stroke Trial. Stroke 2006;37:
clopidogrel and aspirin followed by long-term therapy in patients 447-51.
undergoing percutaneous coronary intervention: the PCI-CURE study. 31. Lakshminarayan K, Solid CA, Collins AJ, et al. Atrial fibrillation and
Lancet 2001;358:527-33. stroke in the general Medicare population: a 10-year perspective
19. Beinart SC, Kolm P, Veledar E, et al. Long-term cost effectiveness of (1992 to 2002). Stroke 2006;37:1969-74.
early and sustained dual oral antiplatelet therapy with clopidogrel 32. Testa L, Zoccai GB, Porto I, et al. Adjusted indirect meta-analysis of
given for up to one year after percutaneous coronary intervention aspirin plus warfarin at international normalized ratios 2 to 3 versus
results: from the Clopidogrel for the Reduction of Events During aspirin plus clopidogrel after acute coronary syndromes. Am J
Observation (CREDO) trial. J Am Coll Cardiol 2005;46:761-9. Cardiol 2007;99:1637-42.
20. Laskey WK, Yancy CW, Maisel WH. Thrombosis in coronary drug- 33. DeEugenio D, Kolman L, DeCaro M, et al. Risk of major bleeding with
eluting stents: report from the meeting of the Circulatory System concomitant dual antiplatelet therapy after percutaneous coronary
Medical Devices Advisory Panel of the Food and Drug Administration intervention in patients receiving long-term warfarin therapy.
Center for Devices and Radiologic Health, December 7-8, 2006. Pharmacotherapy 2007;27:691-6.
Circulation 2007;115:2352-7. 34. Singer DE, Albers GW, Dalen JE, et al. Antithrombotic therapy
21. Connolly S, Pogue J, Hart R, et al. Clopidogrel plus aspirin versus oral in atrial fibrillation: the Seventh ACCP Conference on Antith-
anticoagulation for atrial fibrillation in the Atrial fibrillation rombotic and Thrombolytic Therapy. Chest 2004;126(3 Suppl):
Clopidogrel Trial with Irbesartan for prevention of Vascular Events 429S-56S.
(ACTIVE W): a randomised controlled trial. Lancet 2006;367: 35. Lorenzoni R, Lazzerini G, Cocci F, et al. Short-term prevention of
1903-12. thromboembolic complications in patients with atrial fibrillation with
22. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation: a major aspirin plus clopidogrel: the Clopidogrel-Aspirin Atrial Fibrillation
contributor to stroke in the elderly. The Framingham Study. Arch (CLAAF) pilot study. Am Heart J 2004;148:e6.
Intern Med 1987;147:1561-4. 36. Wyse DG, Waldo AL, DiMarco JP, et al. A comparison of rate control
23. Tanizaki Y, Kiyohara Y, Kato I, et al. Incidence and risk factors for and rhythm control in patients with atrial fibrillation. N Engl J Med
subtypes of cerebral infarction in a general population: the Hisayama 2002;347:1825-33.
study. Stroke 2000;31:2616-22. 37. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of
24. Kopecky SL, Gersh BJ, McGoon MD, et al. Lone atrial fibrillation in antithrombotic therapy in atrial fibrillation. Analysis of pooled data
elderly persons: a marker for cardiovascular risk. Arch Intern Med from five randomized controlled trials. Arch Intern Med 1994;154:
1999;159:1118-22. 1449-57.
American Heart Journal
Volume 156, Number 5
Goto et al 863.e1

Appendix A Ohman, Joachim Röther, Sidney C. Smith, Peter W.F.

Conflict of Interest Disclosures Wilson). All manuscripts in the REACH Registry are
Prof Goto has received honoraria and consulting fees prepared by independent authors who are not governed
from Bristol-Myers Squibb (Princeton, NJ) and sanofi- by the funding sponsors and are reviewed by an academic
aventis (Paris, France). Prof Goto also received a research publication committee before submission. The funding
grant from sanofi-aventis within the last 3 years. sponsors have the opportunity to review manuscript
Dr Bhatt discloses the following relationships with submissions but do not have authority to change any
Bristol-Myers Squibb and sanofi-aventis: honoraria aspect of a manuscript.
(donated to nonprofits for N2 years); speaker's bureau REACH Registry Global Publication Committee
(N 2 years ago); consultant/advisory board (any honoraria Mark Alberts, MD, Northwestern University Medical
donated to nonprofits); expert testimony regarding School, Chicago, IL; Deepak L. Bhatt, MD, VA Boston
clopidogrel (the compensation was donated to a non- Healthcare System and Brigham and Women's Hospital,
profit organization). Boston, MA, (chair); Ralph D'Agostino, MD, Boston
Professor Röther has received payment for speakers' University, Boston, MA; Kim Eagle, MD, University of
bureau and consultancy fees from sanofi-aventis and Michigan, Ann Arbor, MI; Shinya Goto, MD, PhD, Tokai
Bristol-Myers Squibb. University School of Medicine, Isehara, Kanagawa,
Professor Alberts has received research grants, honor- Japan; Alan T. Hirsch, MD, Minneapolis Heart Institute
aria, and consulting fees from Bristol-Myers Squibb and Foundation and Division of Epidemiology and Com-
sanofi-aventis. munity Health, University of Minnesota School of
Professor Hill has no conflicts of interest to declare Public Health, Minneapolis, MN; Chiau-Suong Liau, MD,
relevant to this publication. PhD, Taiwan University Hospital and College of
Dr Ikeda has no conflicts of interest to declare relevant Medicine, Taipei, Taiwan;
to this publication. Jean-Louis Mas, MD, Centre Raymond Garcin, Paris,
Professor Uchiyama has received honoraria and con- France; E. Magnus Ohman, MD, Duke University
sulting fees from sanofi-aventis within the last 3 years. Medical Center, Durham, NC; Joachim Röther, MD,
Professor D'Agostino has received consultancy fees Klinikum Minden, Minden, Germany; Sidney C. Smith,
from sanofi-aventis. MD, University of North Carolina at Chapel Hill,
Professor Ohman has received grant support from Chapel Hill, NC; P. Gabriel Steg, MD, Hôpital Bichat-
Bristol-Myers Squibb and sanofi-aventis. Claude Bernard, Paris, France (chair); Peter W.F.
Prof Liau has received honoraria from sanofi-aventis. Wilson, MD, Cardiology Division, Emory University
Dr Hirsch has received research grants from Bristol- School of Medicine, Atlanta GA.
Myers Squibb and sanofi-aventis, honoraria from sanofi- National Coordinators
aventis, and has served on the speakers' bureau Australia: Christopher Reid, Victoria. Austria: Franz
for sanofi-aventis. Aichner, Linz; Thomas Wascher, Graz. Belgium: Patrice
Prof Mas has no conflicts of interest to declare relevant Laloux, Mont-Godinne. Brazil: Denilson Campos de
to this publication. Albuquerque, Rio de Janeiro. Bulgaria: Julia Jorgova,
Dr Wilson has received research grants from sanofi- Sofia. Canada: Eric A. Cohen, Toronto, Ontario. Chile:
aventis within the last 3 years. Ramon Corbalan, Santiago. China: Chuanzhen L,
Prof Corbalán has no conflicts of interest to declare Shanghai; Runlin Gao, Beijing. Denmark: Per Hildeb-
relevant to this publication. randt, Frederiksberg. Finland: Ilkka Tierala, Helsinki.
Prof Aichner has no conflicts of interest to declare France: Jean-Louis Mas, Patrice Cacoub and Gilles
relevant to this publication. Montalescot, Paris. Germany: Jochum Senges and
Prof Steg has received honoraria for advisory board U Zeymer, Ludwigshafen; Michael Vogelpohl, Ulm.
attendance and consulting fees from Bristol Myers Squibb Greece: Moses Elisaf, Ioannina. Guatemala: Romulo
and sanofi-aventis; speakers bureau from Bristol-Myers Lopez, Guatemala City. Hong Kong: Juliana Chan,
Squibb and sanofi-aventis; and a research grant from Shatin. Hungary: György Pfliegler, Debrecen.
sanofi-aventis within the last 3 years. Indonesia: Bambang Sutrisna, Jakarta. Israel: Avi Porath,
Role of Funding Source Beer Sheva. Japan: Yasuo Ikeda, Tokyo. Lebanon:
The REACH Registry is sponsored by sanofi-aventis Ismail Khalil, Beirut. Lithuania: Ruta Babarskiene,
(Paris, France), Bristol-Myers Squibb (Princeton, NJ), and Kaunas. Malaysia: Robaayah Zambahari, Kuala Lumpur.
the Waksman Foundation (Tokyo, Japan). The sponsors Mexico: Efrain Gaxiola, Jalisco. The Netherlands: Don
provide logistical support. All the publication activity is Poldermans, Rotterdam. Philippines: Maria Teresa B.
controlled by the REACH Registry Global Publication Abola, Quezon City. Portugal: Victor Gil, Carnaxide.
Committee (P. Gabriel Steg, Deepak L. Bhatt, Mark Romania: Constantin Popa, Bucharest. Russia: Yuri
Alberts, Ralph D'Agostino, Kim Eagle, Shinya Goto, Alan Belenkov and Elizaveta Panchenko Pavlovna, Moscow.
T. Hirsch, Chiau-Suong Liau, Jean-Louis Mas, E. Magnus Kingdom of Saudi Arabia: Hassan Chamsi-Pasha,
 American Heart Journal
863.e2 Goto et al November 2008

Jeddah. Singapore: Yeo Tiong Cheng, Singapore. South Hastings. Ukraine: Vira Tseluyko, Kharkov. United
Korea: Oh Dong-Joo, Seoul. Spain: Carmen Suarez, States: Mark Alberts, Chicago, IL; Robert M. Califf,
Madrid. Switzerland: Iris Baumgartner, Bern. Taiwan: Durham, NC; Christopher P. Cannon, Boston, MA; Kim
Chiau-Suong Liau, Taipei. Thailand: Piyamitr Sritara, Eagle, Ann Arbor, MI; Alan T. Hirsch, Minneapolis, MN.
Bangkok. United Arab Emirates: Wael Al Mahmeed, The list of REACH Registry investigators is accessible
Abu Dhabi. United Kingdom: Jonathan Morrell, online at www.reachregistry.org.