NEJM CARDIOVASCULAR DISEASE ARTICLES - Part 8 pps

n engl j med

352;7

www.nejm.org february

17

,

2005

The

new england journal

of

medicine

670

verity of coronary artery disease as having no clini-
cally significant disease, one-, two-, or three-vessel
disease, or disease of the left main coronary artery.
Patients with three-vessel and left main coronary
artery disease had similar mortality rates and were
therefore combined into one group in the regres-
sion analysis. Analyses were performed with NT-
pro-BNP, in quartiles, as a categorical variable with

the lowest quartile serving as reference for the oth-
er three quartiles. A backward-elimination model
was applied. Tests were two-sided, and a P value of
less than 0.05 was considered to indicate statistical
significance. All calculations were generated by SAS
software, version 8.2.

demographic characteristics

After a median follow-up of 9.2 years, 288 of the
1034 patients (28 percent) had died. The median
NT-pro-BNP level for all subjects was elevated (169
pg per milliliter; interquartile range, 63 to 456) and
was significantly lower among patients who sur-
vived than among those who died (120 pg per milli-
liter [interquartile range, 50 to 318] vs. 386 pg per
milliliter [146 to 897], P<0.001). The patients were
divided into subgroups according to quartiles of
NT-pro-BNP. Patients with NT-pro-BNP in the up-
per quartile were older, had a higher left ventricular
results

Table 2. Baseline Clinical Characteristics According to the Left Ventricular Ejection Fraction (LVEF) and the Presence
or Absence of Clinically Significant Coronary Artery Disease (CAD).*
Characteristic LVEF ≥60% LVEF <60%

No CAD
(N=152)
CAD
(N=354) P Value

No CAD
(N=55)
CAD
(N=473) P Value
Age (yr) <0.001 0.02
Median 56 59 54 59
Interquartile range 47–63 52–65 46–67 54–66
Male sex (%) 33 76 <0.001 62 84 <0.001
Family history of ischemic
heart disease (%)
38 40 0.62 16 33 0.01
Medical history (%)
Suspected heart failure† 69 52 <0.001 67 63 0.61
NYHA functional class
I 13 22 17 17
II 64 60 35 58
III 17 15 48 21
IV 6 3 0 3
Hypertension 26 34 0.08 20 32 0.07
Diabetes 13 19 0.13 15 22 0.21
Previous myocardial infarction 17 43 <0.001 43 73 <0.001
Peripheral vascular disease 11 11 0.61 11 15 0.76
Cerebrovascular disease 8 6 0.26 2 5 0.33
Hypercholesterolemia 31 41 0.03 27 38 0.13
Angina pectoris 89 93 0.11 63 90 <0.001
CCS class
I 26 23 20 20
II 45 45 45 41
III 24 28 25 36
IV 5 4 10 6

Previous PTCA (%) 0.7 1.4 0.50 0 1.5 0.38
Previous CABG (%) 0.7 6 0.01 2 9 0.10
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .

n engl j med

352;7

www.nejm.org february

17, 2005

b-type natriuretic peptide and mortality in stable coronary disease

671

end-diastolic pressure, and were more likely to have
a history of myocardial infarction, clinically signifi-
cant coronary artery disease, and diabetes; the LVEF,
body-mass index (the weight in kilograms divided
by the square of the height in meters), and creati-
nine clearance rate were lower than among patients
with NT-pro-BNP levels in the lowest quartile. Ta-
ble 1 shows the baseline characteristics according
to quartiles of NT-pro-BNP.

relation of nt-

pro

-bnp levels
to angiographic status and lvef

The NT-pro-BNP level increased with the severity of
angiographic disease and of left ventricular systolic
dysfunction. In the subgroup of the population with
normal LVEF (i.e., ≥60 percent; 506 patients), the
level of NT-pro-BNP was significantly higher in pa-
tients with than in those without coronary artery
disease. Table 2 shows the baseline characteristics
of patients according to LVEF and the presence or
absence of coronary artery disease.

relation of nt-

pro

-bnp and mortality
from all causes

Kaplan–Meier estimates of survival for all subjects
according to quartiles of NT-pro-BNP are shown in
Figure 1. In a Cox regression analysis with the NT-
pro-BNP level as a categorical variable, the unad-
justed hazard ratios for death of patients with NT-
pro-BNP levels in the second, third, and fourth
quartiles, as compared with those in the first quar-
tile, were 2.1 (95 percent confidence interval, 1.3 to
3.3; P=0.002), 3.5 (95 percent confidence interval,

2.3 to 5.4; P<0.001), and 6.1 (95 percent confidence
interval, 4.0 to 9.2; P<0.001), respectively. In a mul-
tivariable Cox regression model, the hazard ratio for
death among patients in the second quartile of NT-
pro-BNP, as compared with those in the first quar-
tile, was 1.5 (95 percent confidence interval, 0.94 to
2.6; P=0.09), that for the third quartile was 1.9 (95
percent confidence interval, 1.2 to 3.0; P=0.007),
and that for the fourth quartile was 2.4 (95 percent
confidence interval, 1.5 to 4.0; P<0.001).
NT-pro-BNP added prognostic information
above and beyond that provided by age; sex; family
history with respect to ischemic heart disease; pres-
ence or absence of previous and recent myocardial
infarction; presence or absence of angina; Canadi-
an Cardiovascular Society (CCS) class; presence or
absence of hypertension, diabetes, suspected heart
failure, and prior revascularization; smoking status;
results of stress testing; body-mass index; creatinine

* NYHA denotes New York Heart Association, CCS Canadian Cardiovascular Society, PTCA percutaneous transluminal
coronary angioplasty, and CABG coronary-artery bypass grafting. Percentages may not total 100 because of rounding.
† Suspected heart failure includes patients with any symptoms of shortness of breath, leg edema, or both.

‡ The P value is for the comparison between the groups with normal LVEF and those with reduced LVEF.

Table 2. (Continued.)
Characteristic LVEF ≥60% LVEF<60%

No CAD

(N=152)
CAD
(N=354) P Value
No CAD
(N=55)
CAD
(N=473) P Value
Low-fat or low-cholesterol diet (%) 42 53 0.03 58 51 0.42
Regular exercise (%) 34 41 0.20 34 38 0.68
Current smoker (%) 35 40 0.32 49 41 0.33
Angiographic findings (%) <0.001‡
No CAD 100 0 100 0
1-Vessel disease 0 33 0 18
2-Vessel disease 0 30 0 24
3-Vessel disease 0 26 0 45
Left main coronary artery disease 0 11 0 13
NT-pro-BNP (pg/ml) 0.02 0.28
Median 65 83 343 305
Interquartile range 31–159 45–232 164–1085 147–738
Creatinine clearance (ml/min) 0.31 0.99
Median 75 79 74 74
Interquartile range 62–94 65–93 60–91 60–90
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .

n engl j med

352;7

www.nejm.org february

17

,

2005

The

new england journal

of

medicine

672

clearance rate; plasma lipid level; LVEF; left ven-
tricular end-diastolic pressure; and severity of cor-
onary artery disease at angiography. Excluding the
41 patients with CCS class 4 angina did not signifi-
cantly change the results (hazard ratio for the fourth
quartile as compared with the first, 1.8 [95 percent
confidence interval, 1.3 to 2.6]; P<0.001). Hazard
ratios after the backward elimination of the non-
significant variables are shown in Table 3. Adjusted
Kaplan–Meier estimates of survival according to
quartiles of NT-pro-BNP are shown in Figure 2.
Further analyses of subgroups of patients with
LVEF values of 60 percent or more and less than 60

percent yielded the same results. In patients with
LVEF values of less than 60 percent, the adjusted
hazard ratio for death in the fourth quartile of NT-
pro-BNP, as compared with the first quartile, was
3.1 (95 percent confidence interval, 1.2 to 8.1;
P<0.001). Among patients with an LVEF above 60
percent and clinically significant coronary artery
disease on angiography, NT-pro-BNP remained a
strong prognostic marker, with an adjusted hazard
ratio of 1.9 (95 percent confidence interval, 1.2 to
3.3; P=0.01) for the fourth versus the first quartile
of NT-pro-BNP. Among patients with an LVEF of
60 percent or more but without angiographic evi-
dence of coronary artery disease, the same trend
was observed, but was not significant, presumably
because of the lower number of patients.
Our study demonstrates that NT-pro-BNP measured
immediately before coronary angiography in pa-
tients with stable coronary heart disease provides
prognostic information on mortality from all caus-
es that is independent of invasive measurements of
left ventricular function and the severity of coronary
artery disease. These results extend currently avail-
able information about the value of NT-pro-BNP
and BNP as markers of risk in the general popula-
tion and among patients with acute coronary syn-
dromes to a new population of patients with stable
coronary disease who are at “intermediate” risk,
thereby widening the spectrum of clinical useful-
ness of NT-pro-BNP as a prognostic marker.

The findings of the present study support previ-
ous studies suggesting that the elevation of NT-pro-
BNP is associated with coronary heart disease.

9,10

However, unlike previous investigators, we used an-
giographically diagnosed coronary artery disease to
define existing coronary heart disease. Patients with
high NT-pro-BNP levels had a significantly higher
prevalence of coronary disease at angiography than
patients with low concentrations of NT-pro-BNP.
Interestingly, this association was also seen in the
group of patients with normal left ventricular sys-
tolic function.
In this study, NT-pro-BNP was elevated in pa-
tients with stable angina, a condition characterized
by transient ischemic episodes. It was also elevated
in patients with angiographically verified coronary
atherosclerosis, regardless of left ventricular systol-
ic function. Recent studies have suggested that is-
chemia itself, rather than changes in left ventricu-
lar wall stress secondary to ischemia, promotes the
release of BNP,

11

but the responsible mechanisms
still remain to be fully elucidated.
In support of this notion, it is well known that

acute myocardial infarction is associated with acti-
vation of the neurohormonal system that causes in-
creases in levels of natriuretic peptides,

12

in partic-
ular NT-pro-BNP,

13

which predicts poor short-term
and long-term outcome, independently of ventric-
ular function.

5,14-16

Nevertheless, only a few studies
have examined the association between BNP and
NT-pro-BNP and ischemia, and the results have
been conflicting.

17-19

The study by Bibbins-Domingo and colleagues,

19
discussion

Figure 1. Overall Survival among Patients with Stable Coronary Artery

Disease, According to Quartiles of NT-pro-BNP.

The NT-pro-BNP levels were as follows: first quartile, less than 64 pg per mil-
liliter; second quartile, 64 to 169 pg per milliliter; third quartile, 170 to 455 pg
per milliliter; and fourth quartile, more than 455 pg per milliliter. P<0.001 by
the log-rank test for the overall comparison among the groups.
Survival (%)
75
50
25
0
0 2 4 6 8 10
12
Years
1st quartile
2nd quartile
3rd quartile
4th quartile
100
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .

n engl j med

352;7

www.nejm.org february

17, 2005

b-type natriuretic peptide and mortality in stable coronary disease

673

which included 355 patients with stable coronary
disease, showed that elevated levels of BNP were
associated with inducible ischemia, suggesting an
explanation for the increased risk of subsequent
coronary events in patients with elevated BNP. Fur-
thermore, BNP is associated with the occurrence of
transient ischemic episodes during percutaneous
transluminal coronary angioplasty. In a study by
Tateishi and colleagues,

20

the transient increase in
BNP during balloon inflation was independent of
hemodynamic variables. However, there was no
correlation between BNP and the volume of ische-
mic myocardium or the duration of ischemia. The
mechanism involved remains unclear, but two stud-
ies have demonstrated increased cardiac BNP gene
expression in the ischemic left ventricle, suggesting
that elevated levels of BNP and NT-pro-BNP may
result from cardiac ischemia.

3,21

NT-pro-BNP has repeatedly been shown to be

elevated in patients with left ventricular dysfunc-
tion.

4,14,22

We measured the LVEF with single-plane
contrast ventriculography, previously regarded as
the gold standard for LVEF measurements. In our
study population, with more than half of patients
having had a previous myocardial infarction and
therefore likely to have abnormal left ventricular ge-
ometry, this technique may have some limitations,
as it relies on geometric assumptions and therefore
may overestimate LVEF. Despite these limitations,
previous studies have found a relatively good corre-
lation between LVEF as determined by echocardi-
ography and as determined by contrast ventriculog-
raphy.

23

However, the possibility remains that the
elevated NT-pro-BNP levels in patients with normal
LVEF, as determined by standard imaging methods
such as echocardiography, reflect unrecognized left
ventricular remodeling that is detectable only by
high-definition methods, such as magnetic reso-
nance imaging.

24

In our study of the prognostic importance of
NT-pro-BNP in intermediate-risk patients with sta-
ble coronary disease, we found NT-pro-BNP to be a
prognostic marker of long-term mortality from all
causes. BNP and NT-pro-BNP have previously been
shown to be predictors of cardiovascular morbidity
and mortality in the general population,

7

among pa-

* Relative risk and 95 percent confidence interval (CI) were
calculated as hazard ratios derived from the Cox propor-
tional-hazards regression model for overall mortality af-
ter backward elimination of the nonsignificant variables.
Covariates included in the initial model were age, sex,
NT-pro-BNP level, family history with respect to ische-
mic heart disease, presence or absence of previous and
recent myocardial infarction, presence or absence of an-
gina, Canadian Cardiovascular Society class, results of
exercise testing, presence or absence of previous revas-
cularization, hypertension, diabetes, suspected heart
failure, creatinine clearance rate, smoking status, lipid
levels, body-mass index, left ventricular ejection fraction
(LVEF), and severity of coronary disease at angiography.
† The hazard ratio is for three-vessel and left main coro-
nary artery disease (CAD) as compared with no clinically

significant coronary artery disease on angiography.

Table 3 Hazard Ratios for Death from Any Cause
in the Multivariable Model.*
Variable
Hazard Ratio
(95% CI) P Value

NT-pro-BNP (4th vs. 1st
quartile)
2.4 (1.5–4.0) <0.001
Age (per 10-yr increase) 1.6 (1.4–1.9) <0.001
Diabetes 1.7 (1.3–2.2) <0.001
Cigarette smoking 1.6 (1.2–2.0) <0.001
CAD (severe vs. none)† 1.8 (1.2–2.6) 0.002
LVEF (per 10% decrease) 1.2 (1.1–1.4) <0.001
Suspected heart failure 1.8 (1.4–2.4) <0.001

Figure 2. Adjusted Estimates of Overall Survival among Patients with Stable
Coronary Disease, According to Quartiles of NT-pro-BNP.

The survival estimates have been adjusted for age, presence or absence of di-
abetes, smoking status, left ventricular ejection fraction, presence or absence
of suspected heart failure, and severity of angiographic coronary disease. The
NT-pro-BNP levels were as follows: first quartile, less than 64 pg per milliliter;
second quartile, 64 to 169 pg per milliliter; third quartile, 170 to 455 pg per
milliliter; and fourth quartile, more than 455 pg per milliliter. P<0.001 by the
log-rank test for the overall comparison among the groups.
Survival (%)
75

50
25
0
0 2 4 6 8 10
12
Years
1st quartile
2nd quartile
3rd quartile
4th quartile
100
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .

n engl j med

352;7

www.nejm.org february

17

,

2005

The

new england journal

of

medicine

674

tients with acute coronary syndromes, and among
patients with heart failure.

4,5,25


The increased risk of death we observed among
patients with elevated NT-pro-BNP may be a conse-
quence of a higher frequency of coronary events.
The mortality rate among our patients was similar
to the rates in other recent studies of patients with
angiographic evidence of coronary artery disease.

26

Information on the causes of death was not avail-
able in our study, but other studies of patients with
stable coronary heart disease and similar mortality
rates have demonstrated that the chief causes of
death are myocardial infarction and sudden death.

27

Moreover, the association of NT-pro-BNP in our

study was independent of both LVEF and left ventric-
ular end-diastolic pressure, thus providing further
evidence in support of the hypothesis that ischemia
directly promotes the release of NT-pro-BNP, in a
manner that is independent of left ventricular wall
stress.
To explore this hypothesis further, we adjusted
our prognostic models for the severity of coronary
artery disease on angiography. Because angiograph-
ic measures may not fully account for eccentric ath-
erosclerotic lesions in the coronary vessel wall, the
possibility remains that the severity of coronary ar-
tery disease on angiography does not fully reflect the
functional ischemic burden. One could therefore
speculate that including the results of measure-
ments of cardiac ischemia that may be more sensi-
tive, such as radionuclide stress testing, in the mul-
tivariable analysis might attenuate the prognostic
value of NT-pro-BNP. Unfortunately, only 10 percent
of the patients in our study underwent radionuclide
stress testing; the data sample was therefore insuf-
ficient for any attempts at a meaningful analysis.
Other potential confounders of the association
between increasing levels of NT-pro-BNP and mor-
tality include atrial fibrillation and left ventricular
systolic dysfunction, which a convincing body of ev-
idence shows is strongly associated with BNP and
NT-pro-BNP. Unfortunately, information on atrial
fibrillation was not available. We adjusted for left
ventricular dysfunction, but the possibility of unde-

tected systolic and diastolic dysfunction remains.
For the purpose of optimal risk stratification
and for the targeting of treatment strategies, a mul-
timarker strategy has become increasingly com-
mon in the management of acute coronary syn-
dromes.

16,28,29

In stable coronary disease, early
identification of specific groups of patients at in-
creased risk or even at very low risk is equally justi-
fied. With limited resources, careful risk stratifica-
tion could potentially identify patients who would
benefit the most from specific treatment strategies
and make it possible to avoid overtreating patients
at low risk. In this study we have identified NT-pro-
BNP as a marker of increased risk, one independent
of invasive measures of left ventricular dysfunction
and the severity of coronary disease. Further studies
will show whether treatment strategies guided by
NT-pro-BNP levels will decrease morbidity and
mortality in patients with stable coronary disease
and whether NT-pro-BNP will find a place in the
routine clinical stratification of risk among such
patients.

Supported by the Danish Pharmacists Foundation. Dr. Kragelund
is the recipient of a junior research fellowship from the Danish Heart
Foundation (grant no. 00-2-9-10-22011). Roche Diagnostics sup-

plied the assay kits and measured NT-pro-BNP in patients’ samples.
Dr. Hildebrandt reports having received consulting and lecture
fees from Roche Diagnostics.

references

1.

Levin ER, Gardner DG, Samson WK.
Natriuretic peptides. N Engl J Med 1998;
339:321-8.

2.

Hall C. Essential biochemistry and
physiology of (NT-pro)BNP. Eur J Heart Fail
2004;6:257-60.

3.

Goetze JP, Christoffersen C, Perko M, et
al. Increased cardiac BNP expression associ-
ated with myocardial ischemia. FASEB J
2003;17:1105-7.

4.

Tsutamoto T, Wada A, Maeda K, et al.
Attenuation of compensation of endoge-
nous cardiac natriuretic peptide system in

chronic heart failure: prognostic role of
plasma brain natriuretic peptide concentra-
tion in patients with chronic symptomatic
left ventricular dysfunction. Circulation
1997;96:509-16.

5.

Omland T, Persson A, Ng L, et al. N-ter-
minal pro-B-type natriuretic peptide and
long-term mortality in acute coronary syn-
dromes. Circulation 2002;106:2913-8.

6.

de Lemos JA, Morrow DA, Bentley JH, et
al. The prognostic value of B-type natriuretic
peptide in patients with acute coronary syn-
dromes. N Engl J Med 2001;345:1014-21.

7.

Wang TJ, Larson MG, Levy D, et al. Plas-
ma natriuretic peptide levels and the risk of
cardiovascular events and death. N Engl J
Med 2004;350:655-63.

8.

Dodge HT, Sheehan FH. Quantitative

contrast angiography for assessment of ven-
tricular performance in heart disease. J Am
Coll Cardiol 1983;1:73-81.

9.

James SK, Lindahl B, Siegbahn A, et al.
N-terminal pro-brain natriuretic peptide
and other risk markers for the separate pre-
diction of mortality and subsequent myo-
cardial infarction in patients with unstable
coronary artery disease: a Global Utilization
of Strategies To Open occluded arteries
(GUSTO)-IV substudy. Circulation 2003;
108:275-81.

10.

Jernberg T, Stridsberg M, Venge P,
Lindahl B. N-terminal pro brain natriuretic
peptide on admission for early risk stratifi-
cation of patients with chest pain and no ST-
segment elevation. J Am Coll Cardiol 2002;
40:437-45.

11.

D’Souza SP, Baxter GF. B type natriuretic
peptide: a good omen in myocardial is-
chaemia? Heart 2003;89:707-9.

12.

Morita E, Yasue H, Yoshimura M, et al.
Increased plasma levels of brain natriuretic
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .

n engl j med

352;7

www.nejm.org february

17, 2005

b-type natriuretic peptide and mortality in stable coronary disease

675

peptide in patients with acute myocardial in-
farction. Circulation 1993;88:82-91.

13.

Gill D, Seidler T, Troughton RW, et al.
Vigorous response in plasma N-terminal
pro-brain natriuretic peptide (NT-BNP) to
acute myocardial infarction. Clin Sci (Lond)
2004;106:135-9.

14.

Omland T, Aakvaag A, Bonarjee VV, et
al. Plasma brain natriuretic peptide as an in-
dicator of left ventricular systolic function
and long-term survival after acute myocar-
dial infarction: comparison with plasma
atrial natriuretic peptide and N-terminal
proatrial natriuretic peptide. Circulation
1996;93:1963-9.

15.

Talwar S, Squire IB, Downie PF, et al.
Profile of plasma N-terminal proBNP fol-
lowing acute myocardial infarction: correla-
tion with left ventricular systolic dysfunc-
tion. Eur Heart J 2000;21:1514-21.

16.

Sabatine MS, Morrow DA, de Lemos JA,
et al. Multimarker approach to risk stratifi-
cation in non-ST elevation acute coronary
syndromes: simultaneous assessment of
troponin I, C-reactive protein, and B-type
natriuretic peptide. Circulation 2002;105:
1760-3.

17.

Talwar S, Squire IB, Downie PF, Davies
JE, Ng LL. Plasma N terminal pro-brain na-
triuretic peptide and cardiotrophin 1 are
raised in unstable angina. Heart 2000;84:
421-4.

18.

Marumoto K, Hamada M, Hiwada K. In-
creased secretion of atrial and brain natri-
uretic peptides during acute myocardial is-
chaemia induced by dynamic exercise in
patients with angina pectoris. Clin Sci
(Lond) 1995;88:551-6.

19.

Bibbins-Domingo K, Ansari M, Schiller
NB, Massie B, Whooley MA. B-type natri-
uretic peptide and ischemia in patients with
stable coronary disease: data from the Heart
and Soul study. Circulation 2003;108:2987-
92.

20.

Tateishi J, Masutani M, Ohyanagi M,
Iwasaki T. Transient increase in plasma

brain (B-type) natriuretic peptide after per-
cutaneous transluminal coronary angio-
plasty. Clin Cardiol 2000;23:776-80.

21.

D’Souza SP, Davis M, Baxter GF. Auto-
crine and paracrine actions of natriuretic
peptides in the heart. Pharmacol Ther 2004;
101:113-29.

22.

Richards AM, Nicholls MG, Yandle TG,
et al. Plasma N-terminal pro-brain natri-
uretic peptide and adrenomedullin: new
neurohormonal predictors of left ventricu-
lar function and prognosis after myocardial
infarction. Circulation 1998;97:1921-9.

23.

McGowan JH, Cleland JG. Reliability of
reporting left ventricular systolic function by
echocardiography: a systematic review of
3 methods. Am Heart J 2003;146:388-97.

24.

Nilsson JC, Groenning BA, Nielsen G,

et al. Left ventricular remodeling in the first
year after acute myocardial infarction and
the predictive value of N-terminal pro brain
natriuretic peptide. Am Heart J 2002;143:
696-702.

25.

Gardner RS, Ozalp F, Murday AJ, Robb
SD, McDonagh TA. N-terminal pro-brain
natriuretic peptide: a new gold standard in
predicting mortality in patients with ad-
vanced heart failure. Eur Heart J 2003;24:
1735-43.

26.

Cole JH, Miller JI III, Sperling LS, Wein-
traub WS. Long-term follow-up of coronary
artery disease presenting in young adults.
J Am Coll Cardiol 2003;41:521-8.

27.

Fox KM. Efficacy of perindopril in re-
duction of cardiovascular events among pa-
tients with stable coronary artery disease:
randomised, double-blind, placebo-con-
trolled, multicentre trial (the EUROPA
study). Lancet 2003;362:782-8.

28.

Wiviott SD, Cannon CP, Morrow DA, et
al. Differential expression of cardiac bio-
markers by gender in patients with unstable
angina/non-ST-elevation myocardial infarc-
tion: a TACTICS-TIMI 18 (Treat Angina with
Aggrastat and determine Cost of Therapy
with an Invasive or Conservative Strategy-
Thrombolysis In Myocardial Infarction 18)
substudy. Circulation 2004;109:580-6.

29.

Morrow DA, de Lemos JA, Sabatine MS,
et al. Evaluation of B-type natriuretic peptide
for risk assessment in unstable angina/non-
ST-elevation myocardial infarction: B-type
natriuretic peptide and prognosis in
TACTICS-TIMI 18. J Am Coll Cardiol 2003;
41:1264-72. [Erratum, J Am Coll Cardiol
2003;41:1852.]

Copyright © 2005 Massachusetts Medical Society.

clinical trial registration

The

Journal

encourages investigators to register their clinical trials in a public trials
registry. The members of the International Committee of Medical Journal Editors
plan to consider clinical trials for publication only if they have been registered
(see N Engl J Med 2004;351:1250-1). The National Library of Medicine’s
www.clinicaltrials.gov is a free registry, open to all investigators, that meets
the committee’s requirements.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
n engl j med 356;15 www.nejm.org april 12, 2007
1503
The new england
journal of medicine
established in 1812 april 12, 2007 vol. 356 no. 15
Optimal Medical Therapy with or without PCI
for Stable Coronary Disease
William E. Boden, M.D., Robert A. O’Rourke, M.D., Koon K. Teo, M.B., B.Ch., Ph.D., Pamela M. Hartigan, Ph.D.,
David J. Maron, M.D., William J. Kostuk, M.D., Merril Knudtson, M.D., Marcin Dada, M.D., Paul Casperson, Ph.D.,
Crystal L. Harris, Pharm.D., Bernard R. Chaitman, M.D., Leslee Shaw, Ph.D., Gilbert Gosselin, M.D.,
Shah Nawaz, M.D., Lawrence M. Title, M.D., Gerald Gau, M.D., Alvin S. Blaustein, M.D., David C. Booth, M.D.,
Eric R. Bates, M.D., John A. Spertus, M.D., M.P.H., Daniel S. Berman, M.D., G.B. John Mancini, M.D.,
and William S. Weintraub, M.D., for the COURAGE Trial Research Group*
A BS T R AC T
Affiliations for all authors are listed in the
Appendix. Address reprint requests to Dr.
Boden at the Division of Cardiology, Buf-
falo General Hospital, 100 High St., Buffalo,
NY 14203, or at wboden@kaleidahealth.
org.

*Members of the Clinical Outcomes Uti-
lizing Revascularization and Aggressive
Drug Evaluation (COURAGE) trial are list-
ed in the Appendix and in the Supplemen-
tary Appendix, available with the full text
of this article at www.nejm.org.
This article (10.1056/NEJMoa070829) was
published at www.nejm.org on March 26,
2007.
N Engl J Med 2007;356:1503-16.
Copyright © 2007 Massachusetts Medical Society.
Background
In patients with stable coronary artery disease, it remains unclear whether an initial
management strategy of percutaneous coronary intervention (PCI) with intensive
pharmacologic therapy and lifestyle intervention (optimal medical therapy) is superior
to optimal medical therapy alone in reducing the risk of cardiovascular events.
Methods
We conducted a randomized trial involving 2287 patients who had objective evidence
of myocardial ischemia and significant coronary artery disease at 50 U.S. and Cana-
dian centers. Between 1999 and 2004, we assigned 1149 patients to undergo PCI with
optimal medical therapy (PCI group) and 1138 to receive optimal medical therapy alone
(medical-therapy group). The primary outcome was death from any cause and non-
fatal myocardial infarction during a follow-up period of 2.5 to 7.0 years (median, 4.6).
Results
There were 211 primary events in the PCI group and 202 events in the medical-
therapy group. The 4.6-year cumulative primary-event rates were 19.0% in the PCI
group and 18.5% in the medical-therapy group (hazard ratio for the PCI group,
1.05; 95% confidence interval [CI], 0.87 to 1.27; P = 0.62). There were no significant
differences between the PCI group and the medical-therapy group in the composite
of death, myocardial infarction, and stroke (20.0% vs. 19.5%; hazard ratio, 1.05;

95% CI, 0.87 to 1.27; P = 0.62); hospitalization for acute coronary syndrome (12.4% vs.
11.8%; hazard ratio, 1.07; 95% CI, 0.84 to 1.37; P = 0.56); or myocardial infarction
(13.2% vs. 12.3%; hazard ratio, 1.13; 95% CI, 0.89 to 1.43; P = 0.33).
Conclusions
As an initial management strategy in patients with stable coronary artery disease,
PCI did not reduce the risk of death, myocardial infarction, or other major cardio-
vascular events when added to optimal medical therapy. (ClinicalTrials.gov number,
NCT00007657.)
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 356;15 www.nejm.org april 12, 2007
1504
D
uring the past 30 years, the use of
percutaneous coronary intervention (PCI)
has become common in the initial man-
agement strategy for patients with stable coronary
artery disease in North America, even though treat-
ment guidelines advocate an initial approach with
intensive medical therapy, a reduction of risk fac-
tors, and lifestyle intervention (known as optimal
medical therapy).
1,2
In 2004, more than 1 million
coronary stent procedures were performed in the
United States,
3
and recent registry data indicate
that approximately 85% of all PCI procedures are

undertaken electively in patients with stable cor-
onary artery disease.
4
PCI reduces the incidence of
death and myocardial infarction in patients who
present with acute coronary syndromes,
5-10
but
similar benefit has not been shown in patients with
stable coronary artery disease.
11-15
This issue has
been studied in fewer than 3000 patients,
16
many
of whom were treated before the widespread use
of intracoronary stents and current standards of
medical management.
17-28
Although successful PCI of flow-limiting ste-
noses might be expected to reduce the rate of
death, myocardial infarction, and hospitalization
for acute coronary syndromes, previous studies
have shown only that PCI decreases the frequency
of angina and improves short-term exercise per-
formance.
11,12,15
Thus, the long-term prognostic
effect of PCI on cardiovascular events in patients
with stable coronary artery disease remains un-

certain. Our study, the Clinical Outcomes Utiliz-
ing Revascularization and Aggressive Drug Evalu-
ation (COURAGE) trial, was designed to determine
whether PCI coupled with optimal medical ther-
apy reduces the risk of death and nonfatal myo-
cardial infarction in patients with stable coro-
nary artery disease, as compared with optimal
medical therapy alone.
Met hod s
Study Design
The methods we used in the trial have been de-
scribed previously.
29,30
Sponsorship and oversight
of the trial were provided by the Department of
Veterans Affairs Cooperative Studies Program.
Additional funding was provided by the Canadian
Institutes of Health Research. Supplemental cor-
porate support from several pharmaceutical com-
panies included funding and in-kind support. All
support from the pharmaceutical industry con-
sisted of unrestricted research grants payable to
the Department of Veterans Affairs.
The study protocol was approved by the hu-
man rights committee at the coordinating center
and by the local institutional review board at each
participating center. An independent data and
safety monitoring board oversaw the conduct, safe-
ty, and efficacy of the trial. Data management and
statistical analyses were performed solely by the

data coordinating center with oversight by the trial
executive committee, whose members, after un-
blinding, had full access to the data and vouch
for the accuracy and completeness of the data and
the analyses. The companies that provided finan-
cial support, products, or both had no role in the
design, analysis, or interpretation of the study.
Study Population
Patients with stable coronary artery disease and
those in whom initial Canadian Cardiovascular
Society (CCS) class IV angina subsequently stabi-
lized medically were included in the study. Entry
criteria included stenosis of at least 70% in at least
one proximal epicardial coronary artery and ob-
jective evidence of myocardial ischemia (substan-
tial changes in ST-segment depression or T-wave
inversion on the resting electrocardiogram or in-
ducible ischemia with either exercise or pharma-
cologic vasodilator stress) or at least one coronary
stenosis of at least 80% and classic angina with-
out provocative testing. Exclusion criteria included
persistent CCS class IV angina, a markedly posi-
tive stress test (substantial ST-segment depression
or hypotensive response during stage 1 of the
Bruce protocol), refractory heart failure or cardio-
genic shock, an ejection fraction of less than 30%,
revascularization within the previous 6 months,
and coronary anatomy not suitable for PCI. A de-
tailed description of the inclusion and exclusion
criteria is included in the Supplementary Appen-

dix (available with the full text of this article at
www.nejm.org). Patients who were eligible for the
study underwent randomization after providing
written informed consent.
Treatment
Patients were randomly assigned to undergo PCI
and optimal medical therapy (PCI group) or opti-
mal medical therapy alone (medical-therapy group).
A permuted-block design was used to generate
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se
n engl j med 356;15 www.nejm.org april 12, 2007
1505
random assignments within each study site along
with previous coronary-artery bypass grafting
(CABG) as a stratifying variable. All patients re-
ceived antiplatelet therapy with aspirin at a dose
of 81 to 325 mg per day or 75 mg of clopidogrel
per day, if aspirin intolerance was present. Patients
undergoing PCI received aspirin and clopidogrel,
in accordance with accepted treatment guidelines
and established practice standards. Medical anti-
ischemic therapy in both groups included long-
acting metoprolol, amlodipine, and isosorbide
mononitrate, alone or in combination, along with
either lisinopril or losartan as standard second-
ary prevention. All patients received aggressive
therapy to lower low-density lipoprotein (LDL)
cholesterol levels (simvastatin alone or in combi-

nation with ezetimibe) with a target level of 60 to
85 mg per deciliter (1.55 to 2.20 mmol per liter).
After the LDL cholesterol target was achieved, an
attempt was made to raise the level of high-den-
sity lipoprotein (HDL) cholesterol to a level above
40 mg per deciliter (1.03 mmol per liter) and lower
triglyceride to a level below 150 mg per deciliter
(1.69 mmol per liter) with exercise, extended-release
niacin, or fibrates, alone or in combination.
In patients undergoing PCI, target-lesion revas-
cularization was always attempted, and complete
revascularization was performed as clinically ap-
propriate. Success after PCI as seen on angiogra-
phy was defined as normal coronary-artery flow
and less than 50% stenosis in the luminal diam-
eter after balloon angioplasty and less than 20%
after coronary stent implantation, as assessed by
visual estimation of the angiograms before and
after the procedure. Clinical success was defined
as angiographic success plus the absence of in-
hospital myocardial infarction, emergency CABG,
or death. Drug-eluting stents were not approved
for clinical use until the final 6 months of the
study, so few patients received these intracoronary
devices.
Clinical Outcome
Clinical outcome was adjudicated by an indepen-
dent committee whose members were unaware of
treatment assignments. The primary outcome mea-
sure was a composite of death from any cause

and nonfatal myocardial infarction. Secondary out-
comes included a composite of death, myocardial
infarction, and stroke and hospitalization for un-
stable angina with negative biomarkers. The an-
gina status of patients was assessed according to
the CCS classification during each visit. Further
analyses of other secondary outcomes — includ-
ing quality of life, the use of resources, and cost-
effectiveness — are being conducted but have not
yet been completed.
The prespecified definition of myocardial in-
farction (whether periprocedural or spontaneous)
required a clinical presentation consistent with
an acute coronary syndrome and either new ab-
normal Q waves in two or more electrocardio-
graphic leads or positive results in cardiac bio-
markers. Silent myocardial infarction, as detected
by abnormal Q waves, was confirmed by a core
laboratory and was also included as an outcome
of myocardial infarction.
Statistical Analysis
We projected composite 3-year event rates of 21.0%
in the medical-therapy group and 16.4% in the PCI
group (relative difference, 22%) during a follow-
up period of 2.5 to 7.0 years. We also incorporated
assumptions about crossover between study groups
and loss to follow-up.
31
We estimated that the en-
rollment of 2270 patients would provide a power

of 85% to detect the anticipated difference in the
primary outcome at the 5% two-sided level of
significance. A detailed description of the sam-
ple-size calculation is included in the Supplemen-
tary Appendix.
Estimates of the cumulative event rate were
calculated by the Kaplan–Meier method,
32
and
the primary efficacy of PCI, as compared with
optimal medical therapy, was assessed by the
stratified log-rank statistic.
33
The treatment ef-
fect, as measured by the hazard ratio and its
associated 95% confidence interval (CI), was esti-
mated with the use of the Cox proportional-haz-
ards model.
34
Data for patients who were lost to
follow-up were censored at the time of the last
contact. Analyses were performed according to the
intention-to-treat principle. Categorical variables
were compared by use of the chi-square test or
the Wilcoxon rank-sum test, and continuous vari-
ables were compared by use of the Student t-test.
Adjusted analysis of the primary outcome was
performed with the use of a Cox proportional-
hazards regression model with eight preidentified
covariates of interest — age, sex, race, previous

myocardial infarction, extent or distribution of
angiographic coronary artery disease, ejection frac-
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 356;15 www.nejm.org april 12, 2007
1506
tion, presence or absence of diabetes, and health
care system (Veterans Affairs or non–Veterans
Affairs facility in the United States, or a Canadian
facility) — as well as the stratifying variable of
previous CABG. All other comparisons were un-
adjusted. A level of significance of less than 0.01
was used for all subgroup analyses and interac-
tions.
R e su lt s
Baseline Characteristics and Angiographic
Data
Between June 1999 and January 2004, a total of
2287 patients were enrolled in the trial at 50 U.S.
and Canadian centers (
Fig. 1
). Of these patients,
1149 were randomly assigned to the PCI group
and 1138 to the medical-therapy group. The base-
line characteristics of the patients were recently
published
35
and were similar in the two groups
(

Table 1
). The median time from the first episode
of angina before randomization was 5 months
(median, three episodes per week, with exertion
or at rest), and 58% of patients had CCS class II or
III angina. A total of 2168 patients (95%) had ob-
jective evidence of myocardial ischemia, whereas
the remaining 119 patients with classic angina
(CCS class III) and severe coronary stenoses did
not undergo ischemia testing (56 in the PCI group
and 63 in the medical-therapy group). Among pa-
tients who underwent myocardial perfusion im-
aging at baseline, 90% had either single (23%) or
multiple (67%) reversible defects for inducible is-
chemia. Two thirds of the patients had multivessel
coronary artery disease.
Of the 1149 patients in the PCI group, 46 never
underwent a procedure because the patient either
declined treatment or had coronary anatomy un-
suitable for PCI, as determined on clinical reas-
sessment. In 27 patients (2%), the operator was
unable to cross any lesions. PCI was attempted for
1688 lesions in 1077 patients, of whom 1006 (94%)
received at least one stent. In the stent group,
590 patients (59%) received one stent and 416
(41%) more than one stent. Drug-eluting stents
were used in 31 patients. On average, stenosis
in the luminal diameter, as evaluated on visual
assessment of angiograms, was reduced from a
mean (±SD) of 83±14% to 31±34% in the 244

lesions not treated with stents and from 82±12%
to 1.9±8% in the 1444 lesions treated with stents.
After PCI, successful treatment as seen on angi-
ography was achieved in 1576 of 1688 lesions
(93%), and clinical success (i.e., all lesions success-
fully dilated and no in-hospital complications)
was achieved in 958 of 1077 patients (89%).
Medication and Treatment Targets
Patients had a high rate of receiving multiple,
evidence-based therapies after randomization and
during follow-up, with similar rates in both study
groups (
Table 2
). At the 5-year follow-up visit,
70% of subjects had an LDL cholesterol level of
less than 85 mg per deciliter (2.20 mmol per liter)
(median, 71±1.3 mg per deciliter [1.84±0.03 mmol
per liter]); 65% and 94% had systolic and diastolic
blood pressure targets of less than 130 mm Hg
and 85 mm Hg, respectively; and 45% of patients
with diabetes had a glycated hemoglobin level of
no more than 7.0% (
Table 2
). Patients had high
rates of adherence to the regimen of diet, regular
exercise, and smoking cessation as recommended
by clinical practice guidelines,
1,2
although the
mean body-mass index did not decrease.

Follow-up Period
The median follow-up period was 4.6 years (inter-
quartile range, 3.3 to 5.7) and was similar in the
two study groups, with a total of 120,895 patient-
months at risk. Only 9% of patients were lost to
follow-up in the two groups (107 in the PCI group
and 97 in the medical-therapy group, P = 0.51) be-
fore the occurrence of a primary outcome or the
end of follow-up. Vital status was not ascertained
in 194 patients (99 in the PCI group and 95 in the
medical-therapy group, P = 0.81).
Primary Outcome
The primary outcome (a composite of death from
any cause and nonfatal myocardial infarction) oc-
curred in 211 patients in the PCI group and 202
patients in the medical-therapy group (Table 3).
The estimated 4.6-year cumulative primary event
rates were 19.0% in the PCI group and 18.5% in
the medical-therapy group (unadjusted hazard ra-
tio for the PCI group, 1.05; 95% CI, 0.87 to 1.27;
P = 0.62) (
Fig. 2
).
Secondary Outcomes
For the prespecified composite outcome of death,
nonfatal myocardial infarction, and stroke, the
event rate was 20.0% in the PCI group and 19.5%
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se

n engl j med 356;15 www.nejm.org april 12, 2007
1507
33p9
3071 Met eligibility criteria
2287 Consented to participate
(74% of patients with protocol eligibility)
35,539 Patients underwent assessment
32,468 Were excluded
8677 Did not meet inclusion criteria
5155 Had undocumented ischemia
3961 Did not meet protocol for vessels
6554 Were excluded for logistic reasons
18,360 Had one or more exclusions
4513 Had undergone recent (<6 mo) revascu-
larization
4939 Had an inadequate ejection fraction
2987 Had a contraindication to PCI
2542 Had a serious coexisting illness
1285 Had concomitant valvular disease
1203 Had class IV angina
1071 Had a failure of medical therapy
947 Had left main coronary artery stenosis
>50%
722 Had only PCI restenosis (no new lesions)
528 Had complications after myocardial
infarction
784 Did not provide consent
450 Did not receive physician’s
approval
237 Declined to give permission

97 Had an unknown reason
1149 Were assigned to PCI group
46 Did not undergo PCI
27 Had a lesion that could not be dilated
1006 Received at least one stent
107 Were lost to follow-up
1138 Were assigned to medical-therapy group
97 Were lost to follow-up
1149 Were included in the primary analysis 1138 Were included in the primary analysis
AUTHOR:
FIGURE:
JOB: ISSUE:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail
Line
H/T
Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd

3rd
Boden
1 of 3
04-12-07
ARTIST: ts
35615
Figure 1. Enrollment and Outcomes.
Of 35,539 patients who were assessed for eligibility in the trial, 32,468 were excluded for a variety of reasons (patients
could have more than one reason for exclusion). A total of 3071 patients met all inclusion criteria. Of these, 2287
(74%) consented to participate in the study (932 in Canada, 968 in U.S. Veterans Affairs facilities, and 387 in U.S.
facilities other than Veterans Affairs hospitals). Of these patients, 1149 were randomly assigned to the PCI group
and 1138 to the medical-therapy group. The median follow-up was 4.6 years for both study groups.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 356;15 www.nejm.org april 12, 2007
1508
Table 1. Baseline Clinical and Angiographic Characteristics.*
Characteristic
PCI Group
(N = 1149)
Medical-Therapy
Group (N = 1138) P Value
Demographic
Age — yr 61.5±10.1 61.8±9.7 0.54
Sex — no. (%) 0.95
Male 979 (85) 968 (85)
Female 169 (15) 169 (15)
Race or ethnic group — no. (%)† 0.64
White 988 (86) 975 (86)

Black 57 (5) 57 (5)
Hispanic 68 (6) 58 (5)
Other 35 (3) 47 (4)
Clinical
Angina (CCS class) — no. (%) 0.24
0 135 (12) 148 (13)
I 340 (30) 341 (30)
II 409 (36) 425 (37)
III 261 (23) 221 (19)
Missing data 3 (<1) 2 (<1)
Duration of angina — mo 0.53
Median 5 5
Interquartile range 1–15 1–15
Episodes/wk with exertion or at rest within last mo 0.83
Median 3 3
Interquartile range 1–6 1–6
History — no. (%)
Diabetes 367 (32) 399 (35) 0.12
Hypertension 757 (66) 764 (67) 0.53
Congestive heart failure 57 (5) 51 (4) 0.59
Cerebrovascular disease 100 (9) 102 (9) 0.83
Myocardial infarction 437 (38) 439 (39) 0.80
Previous PCI 174 (15) 185 (16) 0.49
CABG 124 (11) 124 (11) 0.94
Stress test‡
Total patients — no. (%) 972 (85) 977 (86) 0.84
Treadmill test — no. (%) 555 (57) 553 (57)
Duration of treadmill test — min 7.0±2.7 6.9±2.3 0.43
Pharmacologic stress — no. (%) 417 (43) 424 (43)
Echocardiography — no. (%) 63 (6) 54 (6)

Nuclear imaging — no. (%) 685 (70) 708 (72) 0.59
Single reversible defect§ 154 (22) 161 (23) 0.09
Multiple reversible defects§ 444 (65) 483 (68) 0.09

Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se
n engl j med 356;15 www.nejm.org april 12, 2007
1509
in the medical-therapy group (hazard ratio, 1.05;
95% CI, 0.87 to 1.27; P = 0.62) (Table 3 and
Fig. 2
).
The rates of hospitalization for acute coronary syn-
dromes were 12.4% in the PCI group and 11.8%
in the medical-therapy group (hazard ratio, 1.07;
95% CI, 0.84 to 1.37; P = 0.56), and adjudicated
rates of myocardial infarction were 13.2% and
12.3%, respectively (hazard ratio, 1.13; 95% CI,
0.89 to 1.43; P = 0.33). For death alone, the rates
were 7.6% and 8.3%, respectively (hazard ratio,
0.87; 95% CI, 0.65 to 1.16); the mortality curves
for the two groups were virtually identical during
the initial 4.6 years of the study. For stroke alone,
the rate was 2.1% in the PCI group and 1.8% in the
medical-therapy group (hazard ratio, 1.56; 95% CI,
0.80 to 3.04; P = 0.19). When the primary end point
was calculated with the exclusion of periproce-
dural myocardial infarction, the event rates were
16.2% and 17.9% (hazard ratio, 0.90; 95% CI, 0.73

to 1.10; P = 0.29).
At a median follow-up of 4.6 years, 21.1% of
patients in the PCI group had additional revascu-
larization, as compared with 32.6% of those in
the medical-therapy group (hazard ratio, 0.60;
95% CI, 0.51 to 0.71; P<0.001). In the PCI group,
77 patients subsequently underwent CABG, as com-
pared with 81 patients in the medical-therapy
group. Revascularization was performed for an-
gina that was unresponsive to maximal medical
therapy or when there was objective evidence of
worsening ischemia on noninvasive testing, at the
discretion of the patient’s physician. The median
time to subsequent revascularization was 10.0
months (interquartile range, 4.5 to 28.0) in the
PCI group and 10.8 months (interquartile range,
3.2 to 30.7) in the medical-therapy group.
There was a substantial reduction in the preva-
lence of angina in both groups during follow-up.
There was a statistically significant difference in
the rates of freedom from angina throughout
most of the follow-up period, in favor of the PCI
group (
Table 2
). At 5 years, 74% of patients in
the PCI group and 72% of those in the medical-
therapy group were free of angina (P = 0.35).
Subgroup Analyses
There was no significant interaction (P<0.01) be-
tween treatment effect and any predefined sub-

group variable (
Fig. 3
). Of note, among patients
with multivessel coronary artery disease, previous
myocardial infarction, and diabetes, the rate of
the primary end point was similar for both groups.
When subgroup variables were included in a multi-
variate analysis, the hazard ratio for treatment
was essentially unchanged (1.09; 95% CI, 0.90 to
1.33; P = 0.77).
Discu ssion
As an initial management strategy, PCI added to
optimal medical therapy did not reduce the pri-
mary composite end point of death and nonfatal
Table 1. (Continued.)
Characteristic
PCI Group
(N = 1149)
Medical-Therapy
Group (N = 1138) P Value
Angiographic
Vessels with disease — no. (%) 0.72
1 361 (31) 343 (30)
2 446 (39) 439 (39)
3 341 (30) 355 (31)
Disease in graft¶ 77 (62) 85 (69) 0.36
Proximal LAD disease 360 (31) 417 (37) 0.01
Ejection fraction 60.8±11.2 60.9±10.3 0.86
* Plus–minus values are means ±SD. Baseline data were missing for one patient in each study group. CCS denotes
Canadian Cardiovascular Society, CABG coronary-artery bypass grafting, and LAD left anterior descending artery.

† Race or ethnicity was reported by the patient at enrollment.
‡ Nuclear imaging could have been performed after either an exercise treadmill test or pharmacologic stress.
§ The percentage in this category is the proportion of patients who underwent imaging.
¶ The percentage in this category is the proportion of patients who had undergone previous CABG.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 356;15 www.nejm.org april 12, 2007
1510
Table 2. Clinical Status, Risk and Lifestyle Factors, and Use of Medication.*
Variable PCI Group (N = 1149) Medical-Therapy Group (N = 1138)
Baseline 1 Yr 3 Yr 5 Yr Baseline 1 Yr 3 Yr 5 Yr
median ±SE
Clinical status
No. evaluated 1148 1031 820 423 1137 1010 824 406
Blood pressure — mm Hg
Systolic 131±0.77 126±0.64 125±0.68 124±0.81 130±0.66 124±0.73 123±0.78 122±0.92
Diastolic 74±0.33 72±0.35 70±0.52 70±0.81 74±0.33 70±0.43 70±0.52 70±0.65
Cholesterol — mg/dl
Total 172±1.37 156±1.17 148±1.13 143±1.74 177±1.41 150±1.10 145±1.30 140±1.64
HDL 39±0.39 42±0.39 43±0.47 41±0.67 39±0.37 41±0.42 42±0.49 41±0.75
LDL 100±1.17 84±0.97 76±0.85 71±1.33 102±1.22 81±0.86 74±0.92 72±1.21
Triglycerides — mg/dl 143±2.96 129±2.74 124±2.79 123±4.13 149±3.03 133±2.90 126±2.84 131±4.70
Body-mass index
28.7±0.18 28.5±0.19 29.0±0.21 29.0±0.34 28.9±0.17 29.0±0.19 29.3±0.21 29.5±0.31
Angina-free — no. (%)†
135 (12) 680 (66) 602 (72) 316 (74) 148 (13) 595 (58) 558 (67) 296 (72)
Risk or lifestyle factor
Current smoker — no. (%) 260 (23) 206 (20) 156 (19) 74 (17) 259 (23) 206 (20) 160 (19) 80 (20)
AHA Step 2 diet — no. (%) 626 (55) 803 (78) 631 (77) 326 (77) 613 (54) 800 (79) 660 (80) 312 (77)

Moderate activity — no. (%)‡ 290 (25) 473 (46) 351 (42) 179 (42) 279 (25) 433 (43) 330 (40) 146 (36)
Glycated hemoglobin in patients
with diabetes
No. evaluated 319 239 197 97 336 286 233 123
Level — % 6.9±0.1 7.1±0.1 7.1±0.1 7.1±0.1 7.1±0.1 7.0±0.1 7.1±0.1 7.1±0.1
Medication
No. evaluated 1147 1044 837 428 1138 1028 838 417
ACE inhibitor — no. (%) 669 (58) 668 (64) 536 (64) 284 (66) 680 (60) 633 (62) 522 (62) 260 (62)
ARB — no. (%) 48 (4) 93 (9) 104 (12) 49 (11) 54 (5) 99 (10) 108 (13) 67 (16)
Statin — no. (%) 992 (86) 972 (93) 780 (93) 398 (93) 1014 (89) 972 (95) 769 (92) 386 (93)
Other antilipid — no. (%) 89 (8) 236 (23) 324 (39) 211 (49) 94 (8) 253 (25) 321 (38) 224 (54)
Aspirin — no. (%) 1097 (96) 995 (95) 792 (95) 408 (95) 1077 (95) 977 (95) 796 (95) 391 (94)
Beta-blocker — no. (%)
975 (85) 887 (85) 705 (84) 363 (85) 1008 (89) 916 (89) 724 (86) 357 (86)
Calcium-channel blocker — no. (%)§
459 (40) 415 (40) 360 (43) 180 (42) 488 (43) 501 (49) 418 (50) 217 (52)
Nitrates — no. (%)¶ 714 (62) 553 (53) 396 (47) 173 (40) 825 (72) 690 (67) 511 (61) 237 (57)
* Plus–minus values are medians ±SE, with the SE calculated with the use of the interquartile range. To convert cholesterol values to milli
-
moles per liter, multiply by 0.02586. To convert triglyceride values to millimoles per liter, multiply by 0.01129. ACE denotes angiotensin-
converting enzyme, and ARB angiotensin-receptor blocker.
† The comparison between the PCI group and the medical-therapy group was significant at 1 year (P<0.001) and 3 years (P = 0.02) but not at
baseline or at 5 years.
‡ This category includes at least 30 to 45 minutes of moderate activity five times per week or vigorous activity three times per week.
§ The comparison between the PCI group and the medical-therapy group was significant at 1 year (P<0.001), 3 years (P = 0.005), and 5 years
(P = 0.003).
¶ The comparison between the PCI group and the medical-therapy group was significant at all time points (P<0.001).
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se

n engl j med 356;15 www.nejm.org april 12, 2007
1511
myocardial infarction or reduce major cardiovas-
cular events, as compared with optimal medical
therapy alone, during follow-up of 2.5 to 7.0 years,
despite a high baseline prevalence of clinical co-
existing illnesses, objective evidence of ischemia,
and extensive coronary artery disease as seen on
angiography. Although the degree of angina re-
lief was significantly higher in the PCI group
than in the medical-therapy group, there was also
substantial improvement in the medical-therapy
group. All secondary outcomes and individual com-
ponents of the primary outcome showed no sig-
nificant differences between the study groups,
nor was there a significant interaction between
treatment effect and any prespecified subgroup
variable. For the primary outcome, the 95% CI
excludes a relative benefit of more than 13% in
the PCI group. Thus, it is highly unlikely that we
missed a prognostically important treatment ben-
efit in favor of the initial PCI strategy.
Table 3. Primary and Secondary Outcomes.*
Outcome Number of Events Hazard Ratio (95% CI)† P Value† Cumulative Rate at 4.6 Years
PCI Group
Medical-Therapy
Group PCI Group
Medical-Therapy
Group
%

Death and nonfatal myocardial
infarction‡
211 202 1.05 (0.87–1.27) 0.62 19.0 18.5
Death§ 68 74
Periprocedural myocardial
infarction
35 9
Spontaneous myocardial infarction 108 119
Death, myocardial infarction, and
stroke
222 213 1.05 (0.87–1.27) 0.62 20.0 19.5
Hospitalization for ACS 135 125 1.07 (0.84–1.37) 0.56 12.4 11.8
Death§ 85 95 0.87 (0.65–1.16) 0.38 7.6 8.3
Cardiac 23 25
Other 45 51
Unknown 17 19
Total nonfatal myocardial infarction 143 128 1.13 (0.89–1.43) 0.33 13.2 12.3
Periprocedural myocardial
infarction
35 9
Spontaneous myocardial infarction 108 119
Death, myocardial infarction, and ACS 294 288 1.05 (0.90–1.24) 0.52 27.6 27.0
Stroke 22 14 1.56 (0.80–3.04) 0.19 2.1 1.8
Revascularization (PCI or CABG)¶ 228 348 0.60 (0.51–0.71) <0.001 21.1 32.6
* ACS denotes acute coronary syndrome, PCI percutaneous coronary intervention, and CABG coronary-artery bypass grafting.
† The hazard ratio is for the PCI group as compared with the medical-therapy group, and P values were calculated by the log-rank test and are
unadjusted for multiple variables.
‡ The definition of myocardial infarction was the finding of new Q waves at any time; a spontaneous creatine kinase MB fraction of at least
1.5 times the upper limit of normal or a troponin T or I level of at least 2.0 times the upper limit of normal; during a PCI procedure, a cre-
atine kinase MB fraction of at least 3 times the upper limit of normal or a troponin T or I level of at least 5.0 times the upper limit of nor-

mal, associated with new ischemic symptoms; and after CABG, a creatine kinase MB fraction or a troponin T or I level of at least 10.0 times
the upper limit of normal. If periprocedural myocardial infarction is excluded from the primary outcome, the hazard ratio is 0.90 (95% CI,
0.73 to 1.10; P = 0.29).
§ Some patients had a nonfatal myocardial infarction before their subsequent death so that the number of deaths overall is greater than the
number of deaths in the primary outcome analysis, which includes the time until the first event.
¶ Values exclude the initial PCI procedure in patients who were originally assigned to the PCI group.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 356;15 www.nejm.org april 12, 2007
1512
Our findings may be explained, in part, by dif-
ferences in atherosclerotic plaque morphology and
vascular remodeling associated with acute coro-
nary syndromes, as compared with stable coronary
artery disease. Vulnerable plaques (precursors of
acute coronary syndromes) tend to have thin
fibrous caps, large lipid cores, fewer smooth-
muscle cells, more macrophages, and less colla-
gen, as compared with stable plaques, and are
associated with outward (expansive) remodeling
of the coronary-artery wall, causing less stenosis
of the coronary lumen.
36
As a result, vulnerable
plaques do not usually cause significant stenosis
before rupture and the precipitation of an acute
coronary syndrome.
36
By contrast, stable plaques

tend to have thick fibrous caps, small lipid cores,
more smooth-muscle cells, fewer macrophages,
and more collagen and are ultimately associated
with inward (constrictive) remodeling that nar-
rows the coronary lumen. These lesions produce
ischemia and anginal symptoms and are easily
detected by coronary angiography but are less like-
ly to result in an acute coronary syndrome.
37,38
Thus, unstable coronary lesions that lead to
myocardial infarction are not necessarily severely
stenotic, and severely stenotic lesions are not nec-
essarily unstable. Focal management of even
severely stenotic coronary lesions with PCI in our
study did not reduce the rate of death and myo-
cardial infarction, presumably because the treated
stenoses were not likely to trigger an acute coro-
nary event. Furthermore, our lower-than-projected
AUTHOR:
FIGURE:
JOB: ISSUE:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail
Line
H/T

Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd
3rd
Boden
2 of 3
04-12-07
ARTIST: ts
35615
39p6
No. at Risk
Medical therapy
PCI
30
35
192
200
408
417
638
637
834
833
959

952
1017
1013
1138
1149
Survival Free of Death from
Any Cause and Myocardial
Infarction
1.0
0.9
0.7
0.6
0.5
0.8
0
0 1 2 3 4 75 6
Years
Medical therapy
Medical therapy
Medical therapy
Medical therapy
PCI
PCI
Hazard ratio, 1.05; 95% CI (0.87–1.27); P=0.62
PCI
Hazard ratio, 0.87; 95% CI (0.65–1.16); P=0.38
Hazard ratio, 1.07; 95% CI (0.84–1.37); P=0.56
PCI
Hazard ratio, 1.13; 95% CI (0.89–1.43); P=0.33
A B

C
D
No. at Risk
Medical therapy
PCI
38
44
302
312
468
488
717
733
917
929
1029
1051
1073
1094
1138
1149
Overall Survival
1.0
0.9
0.7
0.6
0.5
0.8
0
0 1 2 3 4 75 6

Years
No. at Risk
Medical therapy
PCI
127
134
236
246
418
431
662
667
833
835
956
957
1025
1027
1138
1149
Survival Free of ACS
1.0
0.9
0.7
0.6
0.5
0.8
0
0 1 2 3 4 75 6
Years

No. at Risk
Medical therapy
PCI
120
134
192
200
409
418
638
637
834
833
962
954
1019
1015
1138
1149
Survival Free of Myocardial
Infarction
1.0
0.9
0.7
0.6
0.5
0.8
0
0 1 2 3 4 75 6
Years

Figure 2. Kaplan–Meier Survival Curves.
In Panel A, the estimated 4.6-year rate of the composite primary outcome of death from any cause and nonfatal myocardial infarction
was 19.0% in the PCI group and 18.5% in the medical-therapy group. In Panel B, the estimated 4.6-year rate of death from any cause
was 7.6% in the PCI group and 8.3% in the medical-therapy group. In Panel C, the estimated 4.6-year rate of hospitalization for acute
coronary syndrome (ACS) was 12.4% in the PCI group and 11.8% in the medical-therapy group. In Panel D, the estimated 4.6-year rate
of acute myocardial infarction was 13.2% in the PCI group and 12.3% in the medical-therapy group.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se
n engl j med 356;15 www.nejm.org april 12, 2007
1513
event rate in the medical-therapy group may be
explained by systemic therapy that reduced plaque
vulnerability through aggressive intervention for
multiple risk factors and evidence-based use of
medication.
Rates of angina were consistently lower in the
PCI group than in the medical-therapy group dur-
ing follow-up, and rates of subsequent revascu-
larization were likewise lower. However, there
was a substantial increase in freedom from an-
gina in patients in the medical-therapy group as
well, most of which had taken place at 1 year but
with a further improvement at 5 years. To what
extent this finding reflects a benefit of specific
39p6
0.25 0.50 1.00 2.001.751.50
Medical Therapy
Better
PCI Better

Overall
Sex
Male
Female
Myocardial infarction
Yes
No
Extent of CAD
Multivessel disease
Single-vessel disease
Smoking
Current
Not current
Diabetes
Yes
No
CCS angina class
0 or I
II or III
Ejection fraction
≤50%
>50%
Age
>65 yr
≤65 yr
Previous CABG
No
Yes
Race
White

Nonwhite
Health care system
Canadian
U.S. non-VA
U.S. VA
No. of
Patients Hazard Ratio (95% CI)
Medical Therapy
Event Rate for the Primary
Outcome
Baseline Characteristics
0.19
0.18
0.26
0.25
0.14
0.21
0.12
0.21
0.18
0.24
0.15
0.20
0.18
0.26
0.16
0.22
0.16
0.17
0.29

0.18
0.24
0.14
0.21
0.22
1.15 (0.93–1.42)
1.05 (0.87–1.27)
0.87 (0.54–1.42)
1.27 (0.90–1.78)
0.71 (0.44–1.14)
1.06 (0.80–1.38)
1.08 (0.87–1.34)
0.98 (0.52–1.82)
1.04 (0.84–1.29)
1.00 (0.77–1.32)
1.10 (0.83–1.46)
1.05 (0.84–1.32)
1.14 (0.77–1.70)
1.09 (0.85–1.40)
1.01 (0.75–1.38)
1.20 (0.92–1.56)
0.99 (0.73–1.32)
1.08 (0.86–1.36)
1.00 (0.71–1.41)
1.17 (0.76–1.80)
1.04 (0.84–1.30)
1.22 (0.93–1.60)
0.91 (0.69–1.21)
0.65 (0.40–1.06)
P Value

0.19
0.19
0.18
0.23
0.17
0.21
0.15
0.20
0.19
0.25
0.17
0.17
0.20
0.28
0.17
0.24
0.16
0.17
0.34
0.19
0.19
0.17
0.15
0.22
PCI
2287
1947
338
876
1371

1581
700
653
1631
766
1468
964
1371
406
1848
904
1381
2039
248
1963
322
932
387
968
0.03
0.15
0.65
0.71
0.33
0.73
0.72
0.62
0.81
0.43
0.17

AUTHOR:
FIGURE:
JOB: ISSUE:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail
Line
H/T
Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd
3rd
Boden
3 of 3
04-12-07
ARTIST: ts
35615
Figure 3. Subgroup Analyses.
The chart shows hazard ratios (black squares, sized in proportion to the number of subjects in a group), 95% CIs (horizontal lines), cumu-
lative 4.6-year event rates for the composite primary outcome (death from any cause and nonfatal myocardial infarction) for the PCI group

versus the medical-therapy group for the specified subgroups, and P values for the interaction between the treatment effects and sub-
group variables. P values were calculated with the use of the Wald statistic. There was no significant interaction between treatment and
subgroup variables as defined according to the prespecified value for interaction (P<0.01), although there was a trend for interaction with
respect to sex (P = 0.03). PCI denotes percutaneous coronary intervention, CAD coronary artery disease, CCS Canadian Cardiovascular
Society, CABG coronary-artery bypass grafting, and VA Veterans Affairs.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 356;15 www.nejm.org april 12, 2007
1514
antianginal medications (e.g., nitrates and beta-
blockers) or a favorable effect of therapies such
as statins on endothelial function and atheroscle-
rosis is unclear.
Our findings parallel those reported in recent
trials,
39,40
in which observed clinical-event rates
that were associated with optimal medical ther-
apy were lower than projected in the trial design.
These results are also concordant with a meta-
analysis of all previous trials involving PCI ver-
sus medical management.
16
In the aggregate,
these studies, including our own, include out-
come data on more than 5000 patients and show
that PCI has no effect in reducing major cardio-
vascular events.
The preponderance of male patients (85%) is

a limitation of our study, as is the lack of ethnic
diversity (14% of the patients were nonwhite). We
used bare-metal stents, since drug-eluting stents
were not available until late during accrual. Al-
though the latter factor may be perceived as a
limitation, published data indicate no benefit
(either short-term or long-term) with respect to
death and myocardial infarction in patients with
stable coronary artery disease who receive drug-
eluting stents, as compared with those who re-
ceive bare-metal stents.
41-46
Our findings reinforce existing clinical prac-
tice guidelines, which state that PCI can be safely
deferred in patients with stable coronary artery
disease, even in those with extensive, multivessel
involvement and inducible ischemia, provided that
intensive, multifaceted medical therapy is institut-
ed and maintained.
1,2
As an initial management
approach, optimal medical therapy without rou-
tine PCI can be implemented safely in the major-
ity of patients with stable coronary artery disease.
However, approximately one third of these pa-
tients may subsequently require revascularization
for symptom control or for subsequent develop-
ment of an acute coronary syndrome.
In summary, our trial compared optimal med-
ical therapy alone or in combination with PCI as

an initial management strategy in patients with
stable coronary artery disease. Although the ad-
dition of PCI to optimal medical therapy reduced
the prevalence of angina, it did not reduce long-
term rates of death, nonfatal myocardial infarc-
tion, and hospitalization for acute coronary syn-
dromes.
Supported by the Cooperative Studies Program of the U.S. De-
partment of Veterans Affairs Office of Research and Development,
in collaboration with the Canadian Institutes of Health Research;
and by unrestricted research grants from Merck, Pfizer, Bristol-
Myers Squibb, Fujisawa, Kos Pharmaceuticals, Datascope, Astra-
Zeneca, Key Pharmaceutical, Sanofi-Aventis, First Horizon, and
GE Healthcare. All industrial funding in support of the trial was
directed through the U.S. Department of Veterans Affairs.
Dr. Boden reports receiving consulting fees and lecture fees
from Kos Pharmaceuticals, PDL BioPharma, Pfizer, CV Thera-
peutics, and Sanofi-Aventis, and grant support from Merck and
Abbott Laboratories; Dr. O’Rourke, consulting fees from King
Pharmaceuticals, Lilly, and CV Therapeutics; Dr. Teo, grant sup-
port from Boehringer Ingelheim; Dr. Knudtson, lecture fees from
Medtronic and Lilly; Dr. Harris, having equity ownership in
Amgen; Dr. Chaitman, receiving consulting fees from CV Thera-
peutics, Merck, and Bayer, lecture fees from Pfizer, AstraZeneca,
and CV Therapeutics, and grant support from Pfizer, CV Thera-
peutics, and Sanofi-Aventis; Dr. Shaw, grant support from Bristol-
Myers Squibb and Astellas Healthcare; Dr. Booth, grant support
from Actelion; Dr. Bates, consulting fees from Sanofi-Aventis
and AstraZeneca and lecture fees from Sanofi-Aventis; Dr. Sper-
tus, consulting fees from Amgen and United Healthcare and

grant support from Amgen, Roche Diagnostics, and Lilly (and in
the past, consulting fees and grant support from CV Therapeu-
tics and owning the copyright for the Seattle Angina Question-
naire, the Peripheral Artery Questionnaire, and the Kansas City
Cardiomyopathy Questionnaire); Dr. Berman, consulting fees
and lecture fees from Bristol-Myers Squibb, Astellas, Tyco, and
Siemens and grant support from Bristol-Myers Squibb and As-
tellas; Dr. Mancini, consulting and lecture fees from Pfizer,
Abbott, and GlaxoSmithKline, lecture fees from Merck and
Sanofi-Aventis, and grant support from Cordis and GlaxoSmith-
Kline; and Dr. Weintraub, consulting fees from Sanofi-Aventis
and Bristol-Myers Squibb and grant support from Sanofi-Aventis.
No other potential conflict of interest relevant to this article was
reported.
Appendix
The authors’ affiliations are as follows: Western New York Veterans Affairs (VA) Healthcare Network and Buffalo General Hospital–
SUNY, Buffalo, NY (W.E.B.); South Texas Veterans Health Care System–Audie Murphy Campus, San Antonio, TX (R.A.O., P.C.); Mc-
Master University Medical Center, Hamilton, ON, Canada (K.K.T.); VA Cooperative Studies Program Coordinating Center, VA Con-
necticut Healthcare System, West Haven, CT (P.M.H); Vanderbilt University Medical Center, Nashville (D.J.M); London Health Sciences
Centre, London, ON, Canada (W.J.K.); Foothills Hospital, Calgary, AB, Canada (M.K.); Hartford Hospital, Hartford, CT (M.D.); VA
Cooperative Studies Program Clinical Research Pharmacy Coordinating Center, Albuquerque, NM (C.L.H.); Saint Louis University, St.
Louis (B.R.C.); Cedars–Sinai Medical Center, Los Angeles (L.S., D.S.B.); Montreal Heart Institute, Montreal (G. Gosselin); Sudbury
Regional Hospital, Sudbury, ON, Canada (S.N.); Queen Elizabeth Health Sciences Centre, Halifax, NS, Canada (L.M.T.); Mayo Clinic,
Rochester, MN (G. Gau); Houston VA Medical Center, Houston (A.S.B.); Lexington VA Medical Center, Lexington, KY (D.C.B.); Uni-
versity of Michigan Medical Center, Ann Arbor (E.R.B.); Mid America Heart Institute, Kansas City, MO (J.A.S.); Vancouver Hospital and
Health Sciences Centre, Vancouver, BC, Canada (G.B.J.M.); and Christiana Care Health System, Newark, DE (W.S.W.).
The members of the COURAGE trial were as follows: Writing Committee: W. Boden (study cochair), R. O’Rourke (study cochair) K. Teo
(study cochair), P. Hartigan. W. Weintraub, D. Maron, J. Mancini; Executive Committee: W. Weintraub (chair), W. Boden, R. O’Rourke,
K. Teo, P. Hartigan, M. Knudtson, D. Maron, E. Bates, A. Blaustein, D. Booth, R. Carere, S. Ellis, G. Gosselin, G. Gau, A. Jacobs,
Copyright © 2007 Massachusetts Medical Society. All rights reserved.

Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se
n engl j med 356;15 www.nejm.org april 12, 2007
1515
S. King, III, W. Kostuk, C. Harris, J. Spertus; P. Peduzzi (ex officio); Data and Safety Monitoring Board: T. Ryan (chair), B. Turnbull, T.
Feldman, R. Bonow, W. Haskell, P. Diehr, P. Lachenbruch, D. Waters, D. Johnstone; Adjudication Committee: L. Cohen (chair), B.
Cantin, W. Hager, F. Samaha, J. Januzzi, J. Arrighi, B. Chaitman; Economics Committee: W. Weintraub (chair), P. Hartigan, R.
O’Rourke, W. Boden, P. Barnett, J. Spertus, R. Goeree; Optimal Medical Therapy Committee: D. Maron (chair), W. Boden, R. O’Rourke,
K. Teo, W. Weintraub.
The following members assisted in coordination of the study: VA Cooperative Studies Program Coordinating Center, VA Connecticut
Healthcare System, West Haven, CT — P. Peduzzi (director); M. Antonelli, (associate director of operations); J. Smith (project manager);
R. Kilstrom, B. Hunter (coordinators); L. Durant (quality assurance officer); S. O’Neil (end points coordinator); T. Economou, J. Nabors
(programmers); A. Kossack (data clerk); VA Cooperative Studies Program Clinical Research Pharmacy Coordinating Center, Albuquer-
que, NM — M. Sather (director), C. Harris (assistant director), W. Gagne (project manager), C. Fye (pharmacist); VA Cooperative Stud-
ies Human Rights Committee, West Haven, CT — R. Marottoli (chair), H. Allore, D. Beckwith, W. Farrell, R. Feldman, R. Mehta, J.
Neiderman, E. Perry, S. Kasl, M. Zeman; VA Office of Research and Development, Clinical Science Research and Development, Wash-
ington, DC — T. O’Leary (acting director), G. Huang (deputy director, Cooperative Studies Program); Study Chairs Offices — Western
New York VA Healthcare Network and Buffalo General Hospital–SUNY, Buffalo, NY — W. Boden (study cochair), M. Dada, K. Potter
(national coordinators), T. Rivera (program assistant); South Texas Veterans Health Care System, San Antonio, TX — R. O’Rourke
(study cochair), P. Casperson (national coordinator), A. O’Shea (program assistant); McMaster University Medical Center, Hamilton,
ON, Canada — K. Teo (study cochair), G. Woodcock (coordinator); Laboratories: Christiana Care Center for Outcomes Research,
Newark, DE, and Emory University, Atlanta — W. Weintraub; Health Economics Research Center, Menlo Park, CA — P. Barnett; Pro-
gram for Assessment of Technology in Health, Hamilton, ON, Canada — R. Goeree, B. O’Brien; Vancouver Hospital, Cardiovascular
Imaging Research Core Laboratory, Vancouver, BC, Canada — G.B.J. Mancini, E. Yeoh; Washington University Central Lipid Core
Laboratory, St. Louis — J. Ladenson, V. Thompson; Saint Louis University ECG Core Laboratory, St. Louis — B. Chaitman, T. Bertran;
Cedars–Sinai Medical Center Nuclear Core Laboratory, Los Angeles — D. Berman, J. Gerlach, R. Littman, L. Shaw; San Diego State
University PACE Program, San Diego, CA — K. Calfas, J. Sallis.
The following investigators are listed according to their clinical study sites: VA: South Texas Veterans Health Care System, San Antonio, TX
— R. O’Rourke, P. Baker, J. Bolton; VA Medical Center, Houston — A. Blaustein, C. Rowe; VA Medical Center, Durham, NC — K. Morris, S.
Hoffman; VA Health Care System, New York — S. Sedlis, M. Keary; VA Health Care System, Ann Arbor, MI — C. Duvernoy, C. Majors; VA Medical

Center, Lexington, KY — Booth, M. Shockey; James A. Haley Veterans Hospital, Tampa, FL — R. Zoble, I. Fernandez; VA Health Care System, Puget
Sound, WA — K. Lehmann, A. Sorley, M. Abel; VA Health Care System, Albuquerque, NM — M. Sheldon, K. Wagoner; Portland VA Medical
Center, Portland, OR — E. Murphy, K. Avalos; Iowa City VA Medical Center, Iowa City — J. Rossen, K. Schneider; Central Arkansas Veterans Health
Care System, Little Rock, AR — B. Molavi, L. Garza, P. Barton; VA Medical Center, Atlanta — K. Mavromatis, Z. Forghani; Tennessee Valley
Health Care System, Nashville — R. Smith, C. Mitchell; VA Medical Center, Memphis, TN — K. Ramanathan, T. Touchstone. Canada: London
Health Sciences Centre, London, ON — W. Kostuk, K. Sridhar, S. Carr, D. Wiseman; Sudbury Regional Hospital, Sudbury, ON — S. Nawaz, C.
Dion; Montreal Heart Institute, Montreal — G. Gosselin, J. Theberge, M. Cuso; Queen Elizabeth II Health Care Center, Halifax, NS — L. Title, P.
Simon, L. Carroll, K. Courtney-Cox; Sunnybrook Health Care Centre, Toronto — E. Cohen, E. Hsu; University Health Network–Toronto Hospital,
Toronto — V. Dzavik, J. Lan; Foothills Hospital, Calgary, AB — M. Knudtson, D. Lundberg; Hamilton General Hospital–McMaster Clinic, Hamilton,
ON — M. Natarajan, G. Cappelli; St. Michael’s Hospital, Toronto — M. Kutryk, A. DiMarco, B. Strauss; Vancouver Hospital, Vancouver, BC — A.
Fung, J. Chow; Saint John Regional Hospital, Saint John, NB — D. Marr, F. Fitzgerald; St. Paul’s Hospital, Vancouver, BC — R. Carere, T. Nacario;
University of Alberta Hospital, Edmonton — W. Tymchak, L. Harris; Trillium Health Care, Newmarket, ON — C. Lazzam, A. Carter; Hôpital du Sacre
Coeur de Montreal, Montreal — D. Palisaitis, C. Mercure. U.S. Non-VA: Mayo Clinic, Rochester, MN — M. Bell, M. Peterson; MIMA Century
Research Associates, Melbourne, FL — R. Vicari, M. Carroll; University of Michigan Medical Center, Ann Arbor — E. Bates, A. Luciano; Southern
California Kaiser Permanente Medical Group, CA — P. Mahrer; S. Reyes; University of Oklahoma, Oklahoma City — J. Saucedo, D. vanWieren; Mid
America Heart Institute, St. Louis — J. O’Keefe, P. Kennedy; Boston Medical Center, Boston — A. Jacobs. C. Berger, S. Mayo; Emory University
Hospital, Atlanta — J. Miller, T. Arnold; Hartford Hospital, Hartford, CT — F. Kiernan, D. Murphy; Henry Ford Health System, Detroit — A.
Kugelmass, R. Pangilinan; University of Rochester Medical Center, Rochester, NY — R. Schwartz, L. Caufield; Vanderbilt University Hospital, Nash-
ville — D. Hansen, C. Mitchell; SUNY University Hospital, Syracuse, NY — R. Carhart, A. Pennella; Cleveland Clinic, Cleveland — S. Ellis, C.
Stevenson; Barnes–Jewish Hospital, St. Louis — R. Krone, J. Humphrey; Mayo Clinic, Scottsdale, AZ — C. Appleton, J. Wisbey; Christiana
Care Health Systems, Wilmington, DE — M. Stillabower, A. DiSabatino; Rush–Presbyterian–St. Luke’s Medical Center, Chicago — M. Davidson,
J. Mathien.
References
Gibbons RJ, Abrams J, Chatterjee K,
et al. ACC/AHA 2002 guideline update for
the management of patients with chronic
stable angina — summary article: a report
of the American College of Cardiology/
American Heart Association Task Force
on practice guidelines (Committee on the

Management of Patients with Chronic Sta-
ble Angina). J Am Coll Cardiol 2003;41:
159-68.
Smith SC Jr, Feldman TE, Hirshfeld
JW Jr, et al. ACC/AHA/SCAI 2005 guide-
line update for percutaneous coronary in-
tervention — summary article: a report of
the American College of Cardiology/Amer-
ican Heart Association Task Force on Prac-
tice Guidelines (ACC/AHA/SCAI Writing
Committee to Update the 2001 Guidelines
1.
2.
for Percutaneous Coronary Intervention).
Circulation 2006;113:156-75.
Rosamond W, Flegal K, Friday G, et al.
Heart disease and stroke statistics —
2007 update: a report from the American
Heart Association Statistics Committee
and Stroke Statistics Subcommittee. Cir-
culation 2007;115(5):e69-e171.
Feldman DN, Gade CL, Slotwiner AJ,
et al. Comparison of outcomes of percuta-
neous coronary interventions in patients
of three age groups (<60, 60 to 80, and
>80 years) (from the New York State Angio-
plasty Registry). Am J Cardiol 2006;98:
1334-9.
Antman EM, Anbe DT, Armstrong PW,
et al. ACC/AHA guidelines for the man-

agement of patients with ST-elevation myo-
3.
4.
5.
cardial infarction — executive summary:
a report of the American College of Cardi-
ology/American Heart Association Task
Force on Practice Guidelines (Writing Com-
mittee to Revise the 1999 Guidelines for
the Management of Patients with Acute
Myocardial Infarction). Circulation 2004;
110:588-636. [Erratum, Circulation 2005;
111:2013.]
Keeley EC, Boura JA, Grines CL. Prima-
ry angioplasty versus intravenous throm-
bolytic therapy for acute myocardial infarc-
tion: a quantitative review of 23 randomised
trials. Lancet 2003;361:13-20.
Fragmin and Fast Revascularisation
during InStability in Coronary artery dis-
ease Investigators. Invasive compared with
non-invasive treatment in unstable coro-
6.
7.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
n engl j med 356;15 www.nejm.org april 12, 2007
1516
Optimal Medical Ther apy with or withou t PCI for Stable Coronary Dise a se
nary-artery disease: FRISC II prospective

randomised multicentre study. Lancet 1999;
354:708-15.
Cannon CP, Weintraub WS, Demopou-
los LA, et al. Comparison of early invasive
and conservative strategies in patients with
unstable coronary syndromes treated with
the glycoprotein IIb/IIIa inhibitor tirofi-
ban. N Engl J Med 2001;344:1879-87.
Fox KA, Poole-Wilson PA, Henderson
RA, et al. Interventional versus conserva-
tive treatment for patients with unstable
angina or non-ST-elevation myocardial
infarction: the British Heart Foundation
RITA 3 randomised trial: Randomized In-
tervention Trial of unstable Angina. Lancet
2002;360:743-51.
Mehta SR, Cannon CP, Fox KA, et al.
Routine vs selective invasive strategies in
patients with acute coronary syndromes:
a collaborative meta-analysis of random-
ized trials. JAMA 2005;293:2908-17.
Parisi AF, Folland ED, Hartigan P.
A comparison of angioplasty with medi-
cal therapy in the treatment of single-ves-
sel coronary artery disease. N Engl J Med
1992;326:10-6.
Coronary angioplasty versus medical
therapy for angina: the second Randomised
Intervention Treatment of Angina (RITA-2)
trial. Lancet 1997;350:461-8.

Pitt B, Waters D, Brown WV, et al. Ag-
gressive lipid-lowering therapy compared
with angioplasty in stable coronary artery
disease. N Engl J Med 1999;341:70-6.
Hueb W, Soares PR, Gersh BJ, et al.
The Medicine, Angioplasty, or Surgery
Study (MASS-II): a randomized, controlled
clinical trial of three therapeutic strate-
gies for multivessel coronary artery dis-
ease: one-year results. J Am Coll Cardiol
2004;43:1743-51.
Henderson RA, Pocock SJ, Clayton TC,
et al. Seven-year outcome in the RITA-2
trial: coronary angioplasty versus medical
therapy. J Am Coll Cardiol 2003;42:1161-
70.
Katritsis DG, Ioannidis JP. Percutane-
ous coronary intervention versus conser-
vative therapy in nonacute coronary artery
disease: a meta-analysis. Circulation 2005;
111:2906-12.
Yusuf S, Wittes J, Friedman L. Overview
of results of randomized clinical trials in
heart disease. I. Treatments following myo-
cardial infarction. JAMA 1988;260:2088-
93.
A randomized trial of propranolol in
patients with acute myocardial infarction.
I. Mortality results. JAMA 1982;247:1707-
14.

Pfeffer MA, Braunwald E, Moyé LA,
et al. Effect of captopril on mortality and
morbidity in patients with left ventricular
dysfunction after myocardial infarction:
results of the Survival and Ventricular En-
largement trial. N Engl J Med 1992;327:
669-77.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
The SOLVD Investigators. Effect of
enalapril on survival in patients with re-
duced left ventricular ejection fractions
and congestive heart failure. N Engl J Med
1991;325:293-302.
The Heart Outcomes Prevention Eval-
uation Study Investigators. Effects of an
angiotensin-converting–enzyme inhibitor,
ramipril, on cardiovascular events in high-
risk patients. N Engl J Med 2000;342:145-
53. [Errata, N Engl J Med 2000;342:748,

1376.]
Fox KM. Efficacy of perindopril in re-
duction of cardiovascular events among
patients with stable coronary artery dis-
ease: randomised, double-blind, placebo-
controlled, multicentre trial (the EUROPA
study). Lancet 2003;362:782-8.
Randomised trial of cholesterol low-
ering in 4444 patients with coronary heart
disease: the Scandinavian Simvastatin Sur-
vival Study (4S). Lancet 1994;344:1383-
9.
Heart Protection Study Collaborative
Group. MRC/BHF Heart Protection Study
of cholesterol lowering with simvastatin
in 20,536 high-risk individuals: a ran-
domised placebo-controlled trial. Lancet
2002;360:7-22.
LaRosa JC, Grundy SM, Waters DD,
et al. Intensive lipid lowering with atorvas-
tatin in patients with stable coronary dis-
ease. N Engl J Med 2005;352:1425-35.
The Clopidogrel in Unstable Angina
to Prevent Events Trial Investigators. Ef-
fects of clopidogrel in addition to aspirin
in patients with acute coronary syndromes
without ST-segment elevation. N Engl J
Med 2001;345:494-502. [Errata, N Engl J
Med 2001;345:1506, 1716.]
de Lorgeril M, Salen P, Martin JL,

Monjaud I, Delaye J, Mamelle N. Mediter-
ranean diet, traditional risk factors, and
the rate of cardiovascular complications
after myocardial infarction: final report
of the Lyon Diet Heart Study. Circulation
1999;99:779-85.
O’Connor GT, Buring JE, Yusuf S, et al.
An overview of randomized trials of reha-
bilitation with exercise after myocardial
infarction. Circulation 1989;80:234-44.
Boden WE, O’Rourke RA, Teo KK, et al.
Design and rationale of the Clinical Out-
comes Utilizing Revascularization and Ag-
gressive DruG Evaluation (COURAGE) trial
Veterans Affairs Cooperative Studies Pro-
gram no. 424. Am Heart J 2006;151:1173-9.
Weintraub WS, Barnett P, Chen S, et al.
Economics methods in the Clinical Out-
comes Utilizing percutaneous coronary
Revascularization and Aggressive Guide-
line-driven drug Evaluation (COURAGE)
trial. Am Heart J 2006;151:1180-5.
Shih JH. Sample size calculation for
complex clinical trials with survival end-
points. Control Clin Trials 1995;16:395-
407.
Kaplan EL, Meier P. Nonparametric es-
20.
21.
22.

23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
timation from incomplete observations.
J Am Stat Assoc 1958;53:457-81.
Schoenfeld DA, Tsiatis AA. A modified
log rank test for highly stratified data.
Biometrika 1987;74:167-75.
Cox DR. Regression models and life-
tables. J R Stat Soc [B] 1972;34:187-220.
Boden WE, O’Rourke RA, Teo KK, et
al. The evolving pattern of symptomatic
coronary artery disease in the United States
and Canada: baseline characteristics of the
Clinical Outcomes Utilizing Revascular-
ization and Aggressive DruG Evaluation
(COURAGE) trial. Am J Cardiol 2007;99:
208-12.
Naghavi M, Libby P, Falk E, et al. From
vulnerable plaque to vulnerable patient:
a call for new definitions and risk assess-
ment strategies. Circulation 2003;108:1664-
72.

Ambrose JA, Tannenbaum MA, Alexo-
poulos D, et al. Angiographic progression
of coronary artery disease and the devel-
opment of myocardial infarction. J Am
Coll Cardiol 1988;12:56-62.
Little WC, Constantinescu M, Apple-
gate RJ, et al. Can coronary angiography
predict the site of a subsequent myocar-
dial infarction in patients with mild-to-
moderate coronary artery disease? Circu-
lation 1988;78:1157-66.
Hochman JS, Lamas GA, Buller CE,
et al. Coronary intervention for persistent
occlusion after myocardial infarction.
N Engl J Med 2006;355:2395-407.
de Winter RJ, Windhausen F, Cornel
JH, et al. Early invasive versus selectively
invasive management for acute coronary
syndromes. N Engl J Med 2005;353:1095-
104.
Curfman GD, Morrissey S, Jarcho JA,
Drazen JM. Drug-eluting coronary stents:
promise and uncertainty. N Engl J Med
2007;356:1059-60.
Spaulding C, Daemen J, Boersma E,
Cutlip DE, Serruys PW. A pooled analysis
of data comparing sirolimus-eluting stents
with bare-metal stents. N Engl J Med 2007;
356:989-97.
Stone GW, Moses JW, Ellis SG, et al.

Safety and efficacy of sirolimus- and pacli-
taxel-eluting coronary stents. N Engl J Med
2007;356:998-1008.
Kastrati A, Mehilli J, Pache J, et al.
Analysis of 14 trials comparing sirolimus-
eluting stents with bare-metal stents.
N Engl J Med 2007;356:1030-9.
Lagerqvist B, James SK, Stenestrand U,
Lindbäck J, Nilsson T, Wallentin L. Long-
term outcomes with drug-eluting stents
versus bare-metal stents in Sweden. N Engl
J Med 2007;356:1009-19.
Mauri L, Hsieh W, Massaro JM, Ho
KKL, D’Agostino R, Cutlip DE. Stent throm-
bosis in randomized clinical trials of drug-
eluting stents. N Engl J Med 2007;356:1020-
9.
Copyright © 2007 Massachusetts Medical Society.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.

45.
46.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at RIKSHOSPITALET HF on February 18, 2008 .
original article
T he
ne w e ng l an d j ou r na l of m edi ci ne
n engl j med 357;22 www.nejm.org november 29, 2007
2262
Patent Foramen Ovale and Cryptogenic
Stroke in Older Patients
Michael Handke, M.D., Andreas Harloff, M.D., Manfred Olschewski, M.Sc.,
Andreas Hetzel, M.D., and Annette Geibel, M.D.
From the Departments of Cardiology and
Angiology (M.H., A.G.), Neurology and
Neurophysiology (A. Harloff, A. Hetzel),
and Medical Biometry and Statistics (M.O.),
University Hospital Freiburg, Freiburg, Ger-
many. Address reprint requests to Dr.
Handke at the Department of Cardiology,
University Hospital Basel, Petersgraben 4,
4031 Basel, Switzerland, or at handkem@
uhbs.ch.
Drs. Handke and Harloff contributed equal-
ly to this article.
N Engl J Med 2007;357:2262-8.
Copyright © 2007 Massachusetts Medical Society.
A bs t r ac t
BACKGROUND
Studies to date have shown an association between the presence of patent foramen

ovale and cryptogenic stroke in patients younger than 55 years of age. This associa-
tion has not been established in patients 55 years of age or older.
METHODS
We prospectively examined 503 consecutive patients who had had a stroke, and we
compared the 227 patients with cryptogenic stroke and the 276 control patients with
stroke of known cause. We examined the prevalences of patent foramen ovale and
of patent foramen ovale with concomitant atrial septal aneurysm in all patients, using
transesophageal echocardiography. We also compared data for the 131 younger
patients (<55 years of age) and those for the 372 older patients (≥55 years of age).
RESULTS
The prevalence of patent foramen ovale was significantly greater among patients
with cryptogenic stroke than among those with stroke of known cause, for both
younger patients (43.9% vs. 14.3%; odds ratio, 4.70; 95% confidence interval [CI],
1.89 to 11.68; P<0.001) and older patients (28.3% vs. 11.9%; odds ratio, 2.92; 95% CI,
1.70 to 5.01; P<0.001). Even stronger was the association between the presence of
patent foramen ovale with concomitant atrial septal aneurysm and cryptogenic
stroke, as compared with stroke of known cause, among both younger patients
(13.4% vs. 2.0%; odds ratio, 7.36; 95% CI, 1.01 to 326.60; P = 0.049) and older pa-
tients (15.2% vs. 4.4%; odds ratio, 3.88; 95% CI, 1.78 to 8.46; P<0.001). Multivariate
analysis adjusted for age, plaque thickness, and presence or absence of coronary
artery disease and hypertension showed that the presence of patent foramen ovale
was independently associated with cryptogenic stroke in both the younger group
(odds ratio, 3.70; 95% CI, 1.42 to 9.65; P = 0.008) and the older group (odds ratio, 3.00;
95% CI, 1.73 to 5.23; P<0.001).
CONCLUSIONS
There is an association between the presence of patent foramen ovale and crypto-
genic stroke in both older patients and younger patients. These data suggest that
paradoxical embolism is a cause of stroke in both age groups.
Downloaded from www.nejm.org on February 18, 2008 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Patent For amen Ovale and Cr y p togenic Strok e in Older Patien ts

n engl j med 357;22 www.nejm.org november 29, 2007
2263
T
he cause of stroke remains uniden-
tified by routine diagnostic testing in about
40% of patients.
1
One potential cause of
embolic stroke is a patent foramen ovale, which
enables right-to-left intracardiac shunting. The
foramen ovale remains open in about one fourth
of the general population. The prevalence decreas-
es gradually with increasing age, from 34% dur-
ing the first three decades to 20% during the
ninth decade.
2
An association between the presence of pat-
ent foramen ovale and cryptogenic stroke has
been reported; however, the study populations
consisted primarily of younger patients.
3-6
Only
a few studies have included older patients.
7-9
A
meta-analysis showed that the presence of patent
foramen ovale in patients younger than 55 years
of age is significantly associated with cryptogenic
stroke, but the relationship has remained uncon-
firmed in patients 55 years of age or older.

10

Proof of a significant relationship would have
implications for diagnostic and therapeutic man-
agement.
The objectives of our study were to clarify
whether there is a significant association between
the presence of patent foramen ovale and crypto-
genic stroke in patients 55 years of age or older
and to compare the findings with those for
younger patients. Our study population included
all consecutive patients 18 to 85 years of age who
were admitted to our stroke unit or our neuro-
logic intensive care unit. We examined all patients
for the presence or absence of patent foramen
ovale, using transesophageal echocardiography.
Met hod s
Study Population
A total of 596 consecutive patients admitted to
our stroke unit or neurologic intensive care unit
fulfilled the inclusion criteria of an age of 18 to
85 years and the presence of suspected acute brain
ischemia. These patients were enrolled in the
study during the 16-month period between Janu-
ary 2001 and April 2002. Written informed con-
sent was obtained from each patient or, if the
patient was incapable of providing consent, from
the patient’s relatives. The study was approved by
the local ethics committee.
Thirty-seven patients were excluded from analy-

ses because transesophageal echocardiography
could not be performed: 10 were too ill to un-
dergo the procedure, 10 were transferred to an-
other clinic or died before the procedure could be
performed, 14 had an uncontrollable gag reflex
or gastrointestinal tract obstruction that preclud-
ed our performing the procedure, and 3 had un-
dergone the procedure in another clinic but the
method of data acquisition was inconsistent with
our protocol. An additional 24 patients declined
transesophageal echocardiography. Another 32 pa-
tients were excluded from analyses because their
diagnosis on discharge was other than brain ische-
mia. Our analyses focused on the remaining 503
patients.
The following routine diagnostic tests were
performed in all patients: cranial computed to-
mography, magnetic resonance imaging (MRI) of
the brain, or both; and duplex sonography of the
extracranial and intracranial arteries with a 4- to
7-MHz linear-array scanner (model ATL HDI
5000/3500, Advanced Technology Laboratories).
The degree of stenosis of the internal carotid ar-
tery was defined according to the European Carot-
id Surgery Trial protocol.
11
Angiographic imag-
ing of the intracranial and vertebrobasilar arteries
was performed in 231 patients (45.9%), by means
of time-of-flight angiography in patients who

underwent MRI or by means of digital subtrac-
tion angiography before intraarterial thromboly-
sis therapy in the remaining patients. Angiogra-
phy was not performed in 272 patients (54.1%).
All patients underwent transthoracic echocardiog-
raphy and electrocardiography. In patients with
suspected paroxysmal atrial fibrillation (those
with palpitations or a history of atrial fibrillation)
or ambiguous results of routine diagnostic tests,
we also performed 24-hour Holter monitoring.
Before transesophageal echocardiography was
performed, we classified the cause of infarction
according to the modified Trial of Org 10172 in
Acute Stroke Treatment (TOAST) criteria.
12
The
classification was done in advance to make pos-
sible a comparison of our data and those from
earlier studies of the role of patent foramen ovale
in stroke.
10
The TOAST classification subdivides
the cause of stroke into five subtypes, on the ba-
sis of clinical features and the results of diagnos-
tic tests: large-artery atherosclerosis, cardioembo-
lism, small-vessel occlusion, stroke of other known
cause, and stroke of unknown cause. The pa-
tients with stroke whose cause was classified as
one of the four known-cause subtypes were as-
signed to the known-cause group in our study.

We deviated from the original TOAST criteria by
assigning patients with two or more causes of
Downloaded from www.nejm.org on February 18, 2008 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 357;22 www.nejm.org november 29, 2007
2264
stroke to the known-cause group. Patients with
stroke of causes that were unknown despite ex-
tensive routine diagnostic testing before trans-
esophageal echocardiography were assigned to
our cryptogenic-stroke group.
Echocardiographic Examinations
An ultrasonography system (ATL HDI 3500) was
used for transthoracic examinations (with a 2-MHz
transducer) and transesophageal examinations
(with a 5-MHz transducer). Routine transthoracic
echocardiography of the heart was performed in
each patient. Within a median of 2 days after
stroke onset (mean ±SD, 3±2), all patients under-
went transesophageal echocardiography. A con-
trast agent (oxypolygelatin [Gelifundol, Biotest])
was injected while the patient was at rest and
while a Valsalva maneuver was performed. Patent
foramen ovale was diagnosed when microbub-
bles were detected in the left atrium within four
cycles after right-atrial opacification. Atrial sep-
tal aneurysm was diagnosed when the excur-
sion of an abnormally redundant and mobile atrial
septum was over 10 mm.
13

The ascending aorta
and the aortic arch, including the outlet of the
left subclavian artery, were examined for aortic
plaques. The thickest plaque was considered for
classification.
14
Statistical Analysis
Data are presented as means ±SD for continuous
variables and as absolute numbers and relative
percentages for categorical variables. Group com-
parisons were performed by means of the Wil-
coxon rank-sum test for continuous variables and
Fisher’s exact test for categorical variables. Uni-
variate and multivariate logistic-regression analy-
ses were used to estimate the unadjusted and ad-
justed odds ratios and the corresponding 95%
confidence intervals. The characteristics that af-
fected the univariate analysis were included in the
multivariable models: age, plaque thickness, pres-
ence or absence of coronary artery disease, and
presence or absence of hypertension. All statisti-
cal tests were two-sided, and P values of less than
0.05 were considered to indicate statistical signifi-
cance. All analyses were performed with the SAS
statistical package (version 8.2).
R e su lt s
Baseline Characteristics
The overall age range of the patients was 20 to 84
years (mean, 62.2±13.1). Of the 503 patients, 131
(26.0%) were younger than 55 years of age (mean,

45.3±8.3), and 372 (74.0%) were 55 years of age
or older (mean, 68.0±7.0). The cause of stroke
could be identified by means of routine diagnos-
tic testing in 276 patients (54.9%). The stroke was
classified as cryptogenic in the remaining 227
patients (45.1%), including 82 of the 131 patients
younger than 55 years of age (62.6%) and 145 of
the 372 patients 55 years of age or older (39.0%).
As compared with the patients with stroke of
known cause, the patients with cryptogenic stroke
were on average 6 years younger and more of them
had patent foramen ovale with or without a con-
comitant atrial septal aneurysm (
Table 1
). On the
Table 1. Baseline Characteristics of Patients with Cryptogenic Stroke or with Stroke of Known Cause.*
Characteristic
Cryptogenic Stroke
(N = 227)
Stroke of Known Cause
(N = 276) P Value
Age — yr 58.2±13.9 64.5±10.4 <0.001
Female sex — no. (%) 94 (41.4) 97 (35.1) 0.17
PFO — no. (%) 77 (33.9) 34 (12.3) <0.001
PFO–ASA — no. (%) 33 (14.5) 11 (4.0) <0.001
Hypertension — no. (%) 143 (63.0) 222 (80.4) <0.001
Diabetes — no. (%) 48 (21.1) 74 (26.8) 0.15
Hyperlipidemia — no. (%) 81 (35.7) 111 (40.2) 0.31
History of smoking — no. (%) 68 (30.0) 76 (27.5) 0.55
Coronary artery disease — no. (%) 41 (18.1) 82 (29.7) 0.003

Peripheral artery disease — no. (%) 12 (5.3) 20 (7.2) 0.46
Aortic plaque — mm 2.72±1.83 3.06±1.55 <0.001
* Plus–minus values are means ±SD. PFO denotes patent foramen ovale, and ASA atrial septum aneurysm.
Downloaded from www.nejm.org on February 18, 2008 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Patent For amen Ovale and Cr y p togenic Strok e in Older Patien ts
n engl j med 357;22 www.nejm.org november 29, 2007
2265
other hand, patients with cryptogenic stroke had
a lower prevalence of coronary artery disease or
hypertension and slightly thinner aortic plaques
than patients with stroke of known cause. There
were no significant differences between the two
groups with regard to sex or the presence or
absence of diabetes, hyperlipidemia, history of
smoking, or peripheral artery disease.
Association between Patent Foramen Ovale
and Cryptogenic Stroke
The prevalence of patent foramen ovale was sig-
nificantly greater among patients with cryptogen-
ic stroke than among patients with stroke of
known cause. This held true both for patients
younger than 55 years of age (43.9% vs. 14.3%,
P<0.001) and for patients 55 years of age or older
(28.3% vs. 11.9%, P<0.001) (Fig. 1A). The preva-
lence of patent foramen ovale with concomitant
atrial septal aneurysm was also greater among
patients with cryptogenic stroke than among
those with stroke of known cause, both in the
younger group (13.4% vs. 2.0%, P = 0.03) and in
the older group (15.2% vs. 4.4%, P<0.001) (Fig. 1B).

In the unadjusted univariate analysis, the odds
ratios for patients younger than 55 years of age
with cryptogenic stroke, as compared with stroke
of known cause, were 4.70 (95% confidence in-
terval [CI], 1.89 to 11.68; P<0.001) for the pres-
ence of patent foramen ovale and 7.36 (95% CI,
1.01 to 326.60; P = 0.049) for the presence of pat-
ent foramen ovale with concomitant atrial septal
aneurysm. Among patients 55 years of age or
older, the odds ratios were smaller: 2.92 (95% CI,
1.70 to 5.01; P<0.001) for the presence of patent
foramen ovale and 3.88 (95% CI, 1.78 to 8.46;
P<0.001) for the presence of patent foramen ovale
with concomitant atrial septal aneurysm. In the
multivariate analysis, the presence of patent fora-
men ovale was independently associated with
cryptogenic stroke (Fig. 2), both in the overall
study population and in the younger and older
groups. In contrast to the unadjusted odds ratios
from the univariate analysis, the adjusted odds
ratios from the multivariate analysis showed only
a slight difference in the prevalence of patent fora-
men ovale among patients with cryptogenic stroke,
as compared with stroke of known cause, in the
younger group (odds ratio, 3.70; P = 0.008) and the
older group (odds ratio, 3.00; P<0.001).
Among patients 55 years of age or older with
cryptogenic stroke, atherosclerotic plaque thick-
ness was significantly less in patients with patent
foramen ovale (2.78±1.56 mm) and in those with

patent foramen ovale with concomitant atrial
septal aneurysm (2.64±1.36 mm) than in patients
without patent foramen ovale (3.65±1.95 mm)
(Fig. 3). These three subgroups did not differ
significantly with respect to mean age (67.8±8.0
years, 68.6±8.0 years, and 67.9±7.0 years, respec-
tively; P = 0.81). Among patients 55 years of age or
older, the multivariate analysis also showed a
nearly significant relationship between the thick-
ness of the aortic plaque and the risk of crypto-
genic stroke (odds ratio for patients with crypto-
genic stroke vs. stroke of known cause, 1.15 per
1-mm increase in plaque thickness; 95% CI, 1.00
to 1.31; P = 0.05).
Discu ssion
We found an association between the presence of
patent foramen ovale and cryptogenic stroke, not
only in patients younger than 55 years of age but
also in those 55 years of age or older. The rela-
tionship between the presence of patent foramen
ovale and cryptogenic stroke was even more pro-
Patent Foramen Ovale
Patent Foramen Ovale with Atrial Septal Aneurysm
16p6
Percent of Patients
50
40
30
10
20

0
≥55 Yr<55 Yr
B
A
P<0.001 P<0.001
AUTHOR:
FIGURE:
JOB: ISSUE:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail
Line
H/T
Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd
3rd
Handke
1 of 3
10-18-07

ARTIST: ts
35716
Cryptogenic Known cause
Percent of Patients
25
20
15
5
10
0
≥55 Yr<55 Yr
P=0.03 P<0.001
Cryptogenic Known cause
Figure 1. Prevalences of Patent Foramen Ovale (PFO)
and PFO with Concomitant Atrial Septal Aneurysm
among Patients with Cryptogenic Stroke and Those
with Stroke of Known Cause, According to Age Group.
Downloaded from www.nejm.org on February 18, 2008 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
T he ne w eng l a nd jo ur na l of m edic in e
n engl j med 357;22 www.nejm.org november 29, 2007
2266
nounced among patients who had concomitant
atrial septal aneurysm. Moreover, multivariate
analysis showed that the presence of a patent fo-
ramen ovale was independently associated with
cryptogenic stroke in both age groups.
Previous studies have been limited by selection
bias, with older patients undergoing transesoph-
ageal echocardiography less often than younger
patients.

10
In our study, transesophageal echo-
cardiography was performed for all patients. The
prevalence of cryptogenic stroke in our study was
relatively high, at 45%. However, we classified
the cause of stroke before transesophageal echo-
cardiography was performed, as was done in a
recently published study.
15
Two studies that re-
ported prevalences of cryptogenic stroke lower
than those in our study used available data from
transesophageal echocardiography to classify the
cause of stroke.
16,17
We may have slightly overesti-
mated the prevalence of cryptogenic stroke, be-
cause magnetic resonance angiography or digital
subtraction angiography was performed in only
about half the patients. As a result, some patients
with intracranial large-artery atherosclerosis may
have been erroneously assigned to the crypto-
genic-stroke group.
It has long been debated whether the presence
of patent foramen ovale actually does play a caus-
al role in stroke or whether there is only a non-
causal statistical relationship. However, there is
considerable evidence
18
that a patent foramen

ovale can cause ischemic stroke, by means of
paradoxical embolism. A positive relationship has
been shown between the size of the shunt and
the risk of stroke,
19
patients with a residual shunt
after occlusion of the patent foramen ovale have an
increased rate of recurrence of stroke,
20
the rate
of stroke is increased among patients with pul-
monary embolism and a patent foramen ovale,
21

and the migration of a thrombus through the pat-
ent foramen ovale can be directly visualized re-
peatedly,
22
even in very elderly patients.
23,24
The
relationship is still controversial with respect to
older patients,
25
because the available data are
contradictory and are based on studies that used
different diagnostic tests.
7-9
In a study involving
transthoracic echocardiography, Di Tullio et al.

reported an increased prevalence of patent fora-
men ovale in all age groups with cryptogenic
stroke.
7
In one study using transesophageal echo-
cardiography, and another using transcranial Dop-
pler ultrasonography, for detection of right-to-left
shunting after administration of contrast agent,
there were no significant associations between the
presence of patent foramen ovale and cryptogenic
stroke.
8,9
Data from these studies were insuffi-
cient to draw a conclusion for older patients with
cryptogenic stroke.
10
In addition to the significant association found
in our study, additional considerations provide
support for the hypothesis that patent foramen
ovale plays a role in stroke in older patients. With
increasing age, there is an increasing potential
for paradoxical embolism,
26
since the incidence
of venous thromboembolism increases exponen-
tially with increasing age.
27
The combination of
more frequent formation of thromboembolic ma-
terial and hemodynamic changes promoting right-

to-left shunting could contribute to an increased
probability of paradoxical embolism in older pa-
tients.
28
A recent retrospective analysis showed
that the presence of patent foramen ovale increas-
es the risk of adverse events in older patients with
cryptogenic stroke who are receiving aspirin or
warfarin, but not in younger patients.
28
In addi-
tion, the diameter of the patent foramen ovale in-
creases with age,
2
and this could make older pa-
tients more susceptible to paradoxical embolism.
7
33p9
−1.0 1.0 3.0 9.0 11.0
Positive AssociationNegative Association
All patients
Patients <55 yr
Patients ≥55 yr
Stroke of Known Cause
(N=276) Adjusted Odds Ratio (95% CI)Group
3.12 (1.98–5.10)
3.00 (1.73–5.23)
3.70 (1.42–9.65)
34/276
7/49

27/227
Cryptogenic Stroke
(N=227)
77/227
36/82
41/145
5.0 7.0
AUTHOR:
FIGURE:
JOB:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail
Line
H/T
Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd
3rd
Handke

2 of 3
10-18-07
ARTIST: ts
35716 ISSUE:
Figure 2. Odds Ratios for the Presence of Patent Foramen Ovale among Patients with Cryptogenic Stroke, as Compared
with Those with Stroke of Known Cause.
Odds ratios were adjusted for age, plaque thickness, presence or absence of coronary artery disease, and presence
or absence of hypertension.
Downloaded from www.nejm.org on February 18, 2008 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.