Values in the candesartan column may not sum to total values due to
missing data for some patients. Error bars indicate 95% confidence intervals.
ACE indicates angiotensin-converting enzyme; IHD, ischemic heart disease;
LVEF, left ventricular ejection fraction; MI, myocardial infarction; SBP,
systolic blood pressure.
Demers C, McMurray JJV, Swedberg K, Pfeffer MA, Granger CB, Olofsson B, McKelvie RS, Östergren J, Michelson EL, Johansson PA, Wang D, Yusuf S, CHARM Investigators FT. Impact of Candesartan on Nonfatal Myocardial Infarction and Cardiovascular
Death in Patients With Heart Failure. JAMA. 2005;294(14):1794-1798. doi:10.1001/jama.294.14.1794
Author Affiliations: McMaster University, Hamilton,
Ontario (Drs Demers, McKelvie, and Yusuf); University of Glasgow, Glasgow,
Scotland (Dr McMurray); Sahlgrenska University Hospital/Östra, Göteborg,
Sweden (Dr Swedberg); Cardiovascular Division, Brigham and Women’s Hospital,
Boston, Mass (Dr Pfeffer); Duke University Medical Center, Durham, NC (Dr
Granger); AstraZeneca Research and Development, Mölndal, Sweden (Dr Olofsson
and Mr Johansson); Karolinska Hospital, Stockholm, Sweden (Dr Östergren);
AstraZeneca LP, Wilmington, Del (Dr Michelson); and London School of Hygiene
and Tropical Medicine, London, England (Dr Wang).
Context Angiotensin-converting enzyme (ACE) inhibitors reduce the risk of myocardial
infarction (MI), but it is not known whether angiotensin receptor blockers
have the same effect.
Objective To assess the impact of the angiotensin receptor blocker candesartan
on MI and other coronary events in patients with heart failure.
Design, Setting, and Participants The Candesartan in Heart Failure: Assessment of Reduction in Mortality
and Morbidity (CHARM) program, a randomized, placebo-controlled study enrolling
patients (mean age, 66 [SD, 11] years) with New York Heart Association class
II to IV symptoms who were randomly allocated to receive candesartan (target
dose, 32 mg once daily) or matching placebo given in addition to optimal therapy
for heart failure. Patients were enrolled from March 1999 through March 2001.
Of 7599 patients allocated, 4004 (53%) had experienced a previous MI, and
1808 (24%) currently had angina. At baseline, 3125 (41%) were receiving an
ACE inhibitor; 4203 (55%), a β-blocker; 3153 (42%), a lipid-lowering
drug; 4246 (56%), aspirin; and 6286 (83%), a diuretic.
Main Outcome Measure The primary outcome of the present analysis was the composite of cardiovascular
death or nonfatal MI in patients with heart failure receiving candesartan
Results During the median follow-up of 37.7 months, the primary outcome of cardiovascular
death or nonfatal MI was significantly reduced in the candesartan group (775
patients [20.4%]) vs the placebo group (868 [22.9%]) (hazard ratio [HR], 0.87;
95% confidence interval [CI], 0.79-0.96; P = .004;
number needed to treat [NNT], 40). Nonfatal MI alone was also significantly
reduced in the candesartan group (116 [3.1%]) vs the placebo group (148 [3.9%])
(HR, 0.77; 95% CI, 0.60-0.98; P = .03;
NNT, 118). The secondary outcome of fatal MI, sudden death, or nonfatal MI
was significantly reduced with candesartan (459 [12.1%]) vs placebo (522 [13.8%])
(HR, 0.86; 95% CI, 0.75-0.97; P = .02;
NNT, 59). Risk reductions in cardiovascular death or nonfatal MI were similar
across predetermined subgroups and the component CHARM trials. There was no
impact on hospitalizations for unstable angina or coronary revascularization
procedures with candesartan.
Conclusion In patients with heart failure, candesartan significantly reduces the
risk of the composite outcome of cardiovascular death or nonfatal MI.
Angiotensin-converting enzyme (ACE) inhibitors reduce cardiovascular
death, hospitalization for heart failure, and myocardial infarction (MI) in
patients with heart failure or left ventricular systolic dysfunction and in
high-risk patients with coronary artery disease or diabetes.1- 9 This
effect is assumed to be due to the action of these drugs to reduce angiotensin
II production, although ACE inhibitors also prevent bradykinin breakdown,
which may have additional beneficial effects.10 This
raises the question of whether angiotensin II receptor blockers (ARBs) are
as protective as ACE inhibitors in preventing MI. Conversely, blockade of
the renin-angiotensin-aldosterone system by ACE inhibitors may be incomplete,
especially during long-term treatment in patients with an activated system;
in these patients, there is evidence of continued production of angiotensin
II by non–ACE-dependent pathways.11 This
raises the possibility that an ARB in combination with an ACE inhibitor may
be effective in further reducing MI.
In the placebo-controlled Candesartan in Heart Failure: Assessment of
Reduction in Mortality and Morbidity (CHARM) program, the effect of candesartan
on total mortality, cardiovascular death, and hospitalization for heart failure
was examined in patients with heart failure receiving recommended therapy.12,13 This article describes the effects
of candesartan on MI and on hospitalization for unstable angina and coronary
revascularization procedures in the overall CHARM program.
The design and primary results of CHARM have been published and are
summarized here.12,13 In brief,
the CHARM program consisted of 3 component trials that compared the effects
of adding candesartan or placebo to optimal background therapy in consenting
patients with heart failure and either preserved left ventricular ejection
fraction (LVEF) (CHARM-Preserved) or reduced LVEF (CHARM-Added, enrolling
patients treated with an ACE inhibitor, and CHARM-Alternative, enrolling those
not receiving an ACE inhibitor because of documented intolerance). Patients
were enrolled from March 1999 through March 2001.
Patients with New York Heart Association functional class II to IV were
eligible and were randomly allocated according to LVEF (≤40%, >40%) and
treatment with an ACE inhibitor. All patients received candesartan or matching
placebo, starting at a dosage of 4 or 8 mg once daily, which was increased
as tolerated to the target of 32 mg once daily. All sites received approval
from local ethics committees for the conduct of this trial, and all patients
provided written informed consent.
After the initial dose-titration period, follow-up visits occurred every
4 months, with a minimum planned duration of 2 years. At each visit, the occurrence
of study outcomes was ascertained according to the intention-to-treat principle.
Causes of death and reasons for hospital admissions were classified on standard
forms by the investigator, without knowledge of treatment assignment, and
confirmed or refuted by a blinded central adjudication process. Death was
classified as cardiovascular unless an unequivocal noncardiovascular cause
of death was confirmed by the central adjudication committee. Cardiovascular
death included sudden death; death due to MI, heart failure, or stroke; procedure-related
death (cardiovascular investigation/procedure/operation); death due to other
specified cardiovascular causes; and presumed cardiovascular deaths (ie, those
for which a noncardiovascular cause had not been clearly established).
A diagnosis of MI was made if (1) levels of creatine kinase or creatine
kinase-MB (or troponin I or T if these were not available) were more than
twice the upper limit of normal or if levels of these same markers were 3
times the upper limit of normal within 24 hours of percutaneous coronary intervention
or 5 times the upper limit of normal within 24 hours of coronary artery bypass
graft surgery and if, in addition, the patient had (2) electrocardiographic
changes in 2 or more contiguous leads showing new Q waves (or R waves in V1 and V2), left bundle-branch block, or ischemic ST–T-wave
changes, or (3) typical clinical presentation with cardiac ischemic-type pain
lasting more than 20 minutes, pulmonary edema, or cardiogenic shock not otherwise
explained. All reported nonfatal MI events underwent blinded central adjudication.
Information on hospitalization for unstable angina and coronary revascularization
procedures (percutaneous coronary intervention or coronary artery bypass graft
surgery) were based on the events reported by the investigator and were not
The primary composite outcome of this analysis was cardiovascular death
or nonfatal MI. All randomized patients were included in the analyses except
for 2 individuals for whom no data were available. Hazard ratios were estimated
by finding the values of the regression coefficients in a Cox regression model
(stratified for study) that maximized the partial likelihood. The Cox proportional
hazards assumption was confirmed by plotting the hazards against follow-up
time. The Wald statistic was used to test each coefficient separately, and
95% confidence intervals were calculated. Tests of heterogeneity of hazard
ratios across studies were performed. Survival curves were estimated by the
Kaplan-Meier procedure. Cox regression analyses were used to determine the
uniformity of treatment effects across prespecified subgroups for the CHARM-Overall
study. Analyses were performed using SAS version 8.2 (SAS Institute Inc, Cary,
NC); P<.05 was used to determine statistical significance.
Baseline patient characteristics for the CHARM-Overall study were previously
published.13 Mean age was 66 (SD, 11) years.
Of 7599 patients randomly allocated, 5199 (68%) were men, 4004 (53%) had experienced
a previous MI, 1808 (24%) had current angina, 4681 (62%) had heart failure
of ischemic etiology, 2160 (28%) had diabetes mellitus, 1228 (16%) had undergone
percutaneous coronary intervention, and 1791 (24%) had undergone coronary
artery bypass graft surgery. At baseline, 3125 patients (41%) were receiving
an ACE inhibitor (100% in CHARM-Added, 0% in CHARM-Alternative, 19% in CHARM-Preserved);
4203 (55%), a β-blocker; 3153 (42%), a lipid-lowering drug; 4246 (56%),
aspirin; and 6286 (83%), a diuretic.
There was a significant reduction in the primary composite outcome of
cardiovascular death or nonfatal MI and in the secondary outcome of nonfatal
MI alone in patients receiving candesartan compared with placebo (Table and Figure
1). There was a nonsignificant trend of fewer patients who experienced
sudden death or fatal MI, but the composite outcome of fatal MI, sudden death,
or nonfatal MI was significantly reduced with candesartan.
The effect of candesartan on the composite outcome of cardiovascular
death or nonfatal MI was consistent across the component CHARM trials (Added,
Alternative, and Preserved) (Figure 2).
The impact of candesartan on cardiovascular death or nonfatal MI compared
with placebo was also consistent across relevant subgroups (Figure 2). Candesartan had no effect on hospitalization for unstable
angina or coronary revascularization procedures.
In the CHARM program, the addition of the ARB candesartan to conventional
therapies for heart failure resulted in a significant reduction in the combined
outcome of cardiovascular death or nonfatal MI in patients with symptomatic
heart failure. These findings were consistent across all subgroups examined,
including patients treated with other therapies proven to be effective in
reducing the risk of MI or reinfarction. The prevention of MI broadens the
potential benefit of candesartan in this patient population.
It is of interest to compare this effect of candesartan with that of
ACE inhibitors. In the Studies of Left Ventricular Dysfunction (SOLVD) treatment
and prevention trials,1 the ACE inhibitor enalapril
decreased the risk of nonfatal MI by 23% (95% confidence interval, 11%-34%; P<.001). Although similar reductions in MI were described
in the Heart Outcomes Prevention Evaluation (HOPE)7 and
the European Trial on Reduction of Cardiac Events With Perindopril in Stable
Coronary Artery Disease (EUROPA)8 studies with
ACE inhibitor treatment in high- and intermediate-risk patient populations
without documented heart failure or left ventricular systolic dysfunction,
the Prevention of Events With Angiotensin Converting Enzyme Inhibitors (PEACE)
study9 showed no effect on nonfatal MI of adding
trandolapril in low-risk patients. However, a meta-analysis of the HOPE, EUROPA,
and PEACE trials indicates a reduction in total mortality.14 Prior
to CHARM, it was unknown whether ARBs would also reduce MI in patients with
heart failure or other cardiovascular conditions. It is notable, therefore,
that the magnitude of the reduction in cardiovascular death and nonfatal MI
in CHARM was similar to that observed in SOLVD and other trials. Furthermore,
the beneficial impact of candesartan was observed in patients taking β-blockers,
lipid-lowering agents, or aspirin, indicating an additive and independent
effect to standard therapy, as seen with ACE inhibitors. Importantly, however,
candesartan also further reduced risk in patients receiving an ACE inhibitor,
suggesting additional protection against cardiovascular events through increased
blockade of the renin-angiotensin-aldosterone system, at least in patients
with heart failure.
Although nonfatal MI alone and the composite outcome of cardiovascular
death or nonfatal MI were significantly reduced by candesartan, there was
only a nonsignificant reduction of the composite of sudden death or fatal
MI. In CHARM, the number of fatal MIs was small, leading to wide confidence
intervals and statistical uncertainty about the effect of treatment. Second,
although central adjudication of potential end points was used in CHARM, it
is difficult to precisely classify the cause of death in patients with heart
failure. While broad categories such as cardiovascular deaths vs noncardiovascular
deaths are likely reliable, further subcategories may not be.15 For
example, an autopsy substudy of the Assessment of Treatment with Lisinopril
And Survival (ATLAS) trial found that a high proportion of “sudden deaths”
had evidence of coronary occlusion, as did many patients thought to have died
from “pump failure.”16 This underscores
why the combined outcome of cardiovascular death or nonfatal MI is more reliable
and better reflects the impact of candesartan on fatal or nonfatal MI in the
CHARM program. In this context, our approach is consistent with that used
in previous large trials.1,7
Candesartan did not reduce hospitalizations for unstable angina and
coronary revascularization procedures. Although this differs from SOLVD, these
results are similar to HOPE and EUROPA.7,8 In
the SOLVD treatment and prevention trials, hospitalizations for unstable angina
were documented in 499 patients (14.7%) treated with enalapril and in 595
(17.5%) in the placebo group (risk reduction, 20%; 95% confidence interval,
9%-29%; P = .001).1 The
SOLVD trials were conducted from 1985-1990, when the use of β-blockers
and aspirin was lower. Use of aspirin and β-blockers was 46% and 18%,
respectively, in the SOLVD trials, compared with 56% and 55% in CHARM. This
difference in the use of proven anti-ischemic therapy may in part explain
the smaller number of events leading to hospitalization for unstable angina
with wide confidence intervals in the CHARM program. In a similar fashion,
the number of coronary revascularization procedures was small, with no significant
effect of candesartan. Furthermore, hospital admission for unstable angina
and coronary revascularization may not necessarily reflect disease progression
but rather may be due to variations in physician practice styles.
The Valsartan in Acute Myocardial Infarction (VALIANT) trial and the
Optimal Trial in Myocardial Infarction With the Angiotensin II Antagonist
Losartan (OPTIMAAL) study evaluated the effects of an ARB compared with the
ACE inhibitor captopril in patients with acute MI.17,18 The
VALIANT trial also compared the combination of the ARB valsartan plus captopril
with captopril alone. Neither trial showed superiority of the ARB over captopril,
with respect to the primary end point of all-cause mortality. In OPTIMAAL,
the risk of fatal or nonfatal reinfarction was comparable in patients treated
with losartan and captopril (relative risk, 1.03; 95% confidence interval,
0.89-1.18; P = .72).18 In
VALIANT, the number of patients who experienced an MI was similar in the groups
treated with valsartan (3-year Kaplan-Meier rate, 14.9%), captopril (14.2%),
and the combination of valsartan plus captopril (14.1%).19 The
results of CHARM provide additional information on the effect of the ARB candesartan
alone or in combination with ACE inhibitors compared with ACE inhibitors alone.
The conclusion drawn from an overview of the complete data available refutes
that of a recent but more selective review, which suggested that ARBs, unlike
ACE inhibitors, may not reduce MI.20
Our observations in the CHARM-Alternative component especially suggest
that the possible anti-MI effect of candesartan (and, by inference, ACE inhibitors)
is angiotensin II–dependent. Furthermore, that candesartan seemed to
have a beneficial effect independent of ACE inhibition suggests that non-ACE
angiotensin II generation might be contributing to the continuing risk of
MI in patients treated with an ACE inhibitor. Large prospective trials are
needed to test these hypotheses, and at least 2 are under way (Ongoing Telmisartan
Alone and in Combination With Ramipril Global Endpoint Trial [ONTARGET] and
Telmisartan Randomized Assessment Study in ACE Intolerant Subjects With Cardiovascular
In conclusion, these results from the CHARM program suggest that the
use of candesartan in patients optimally treated for heart failure reduces
the risk of cardiovascular death or nonfatal MI. This apparent benefit is
in addition to that of other agents known to decrease MI. Further studies
are required to confirm this benefit and elucidate the mechanisms responsible
for the actions of candesartan on ischemic cardiovascular events in this patient
Corresponding Author: Salim Yusuf, DPhil,
Population Health Research Institute, McMaster University, Room 252, McMaster
Clinic, Hamilton General Hospital, 237 Barton St E, Hamilton, Ontario, Canada
L8L 2X2 (firstname.lastname@example.org).
Author Contributions: Dr Demers had full access
to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: McMurray, Swedberg,
Pfeffer, Granger, McKelvie, Michelson, Yusuf.
Acquisition of data: McMurray, Swedberg, Pfeffer,
McKelvie, Östergren, Wang.
Analysis and interpretation of data: Demers,
McMurray, Swedberg, Pfeffer, Olofsson, Östergren, Michelson, Johansson,
Drafting of the manuscript: Demers, McMurray,
Swedberg, Pfeffer, Yusuf.
Critical revision of the manuscript for important
intellectual content: Demers, McMurray, Swedberg, Pfeffer, Granger,
Olofsson, McKelvie, Östergren, Michelson, Johansson, Wang, Yusuf.
Statistical analysis: Olofsson, Johansson,
Obtained funding: McMurray, Swedberg, Michelson.
Administrative, technical, or material support:
McKelvie, Östergren, Michelson, Yusuf.
Study supervision: Demers, McMurray, Swedberg,
Pfeffer, Östergren, Yusuf.
Financial Disclosures: Drs Demers, McMurray,
Swedberg, Pfeffer, Granger, McKelvie, Östergren, and Yusuf have received
research grants, honoraria for lectures, and/or consulting fees from a number
of pharmaceutical companies manufacturing and marketing inhibitors of the
renin-angiotensin-aldosterone system, including AstraZeneca, Boehringer-Ingelheim,
Bristol-Myers Squibb/Sanofi-Aventis, Merck, Novartis, Pfizer, and Takeda.
Dr Wang reported no disclosures.
Funding/Support: AstraZeneca funded the CHARM
Role of the Sponsor: Representatives from AstraZeneca
were involved in protocol design; in collection, management, analysis, and
interpretation of data; and in manuscript preparation.
Independent Statistical Review: Independent
statistical review of the data included in this analysis was perfomed by Stuart
Pocock, PhD (also served on the CHARM Data Safety and Monitoring Committee)
and Duolao Wang, both of the London School of Hygiene and Tropical Medicine.