Discussion
Our results show that treatment of
a broad spectrum of patients with
symptomatic heart failure with
candesartan resulted in a reduction in
deaths, albeit of borderline significance,
notably because of a significant 12%
reduction in cardiovascular deaths. In
the overall CHARM programme, the
risk of death and, particularly, death
attributed to cardiovascular causes was
strongly affected by left-ventricular systolic function. The
annual cardiovascular death rate among the placebo group
who had reduced LVEF was around 9%22,23 and was only
4% in the placebo group of CHARM-Preserved.24
However, the annual non-cardiovascular death rate of
about 2% per year in the placebo group was similar in all
component trials. In the patients with LVEF higher than
40%, who had a substantially lower risk of dying of a
cardiovascular cause than did patients with lower LVEF,
candesartan did not seem to alter survival. However, the
absence of heterogeneity in treatment effect across trials
and by LVEF, and the reduction in hospital admissions for
heart failure in this lower-risk group, does provide an
indication that worthwhile clinical benefits were derived
from candesartan.
Our prespecified analysis of the mortality results
from the two low LVEF trials shows a clear prolongation
in survival with candesartan, a significant reduction in
all deaths, and a 16% lower rate of cardiovascular
death. This reduction in cardiovascular deaths was
complemented by significant reductions in hospital
admissions for the management of heart failure in all
component trials.
The concept that inhibition of the renin-angiotensin
system by blockade at the angiotensin II type 1 receptor
can result in more complete inhibition of the adverse
cardiovascular effects of angiotensin II while leaving
unopposed other potentially desirable actions modulated
by different angiotensin II receptors has stimulated much
interest and clinical investigation. Angiotensin-receptor
blockers effectively reduce important non-fatal clinical
events in hypertensive patients who have diabetes and
nephropathy,25,26 hypertensive patients with electrocardiographic
evidence of left-ventricular hypertrophy,27
elderly patients with hypertension,28 and those with
symptomatic heart failure and depressed ejection
fraction.29 However, a survival advantage produced by
angiotensin-receptor blockers in patients with heart failure
and reduced LVEF has not been clearly shown,29,30 nor in
any other high-risk population studied. We show that
candesartan offers survival benefits in patients with CHF
and reduced LVEF. However, we cannot tell to what
extent this improvement in survival was related to the
32 mg daily target dose of candesartan or other potentially
distinctive pharmacological properties.31
In patients with heart failure and reduced LVEF, it has
been suggested that additional survival benefits could not
be achieved with angiotensin-receptor blockers among
those already taking proven effective treatments,
including angiotensin-converting-enzyme inhibitors and
blockers.32 Our two cohorts of patients with
symptomatic heart failure (LVEF 40%) prespecified by
use of angiotensin-converting-enzyme inhibitor and with
substantial -blocker use, is particularly well suited to
address this question. The reductions in cardiovascular
death with candesartan were similar in patients taking
angiotensin-converting-enzyme inhibitors and not taking
them because of intolerance. Similarly, the similarity of
clinical-outcome benefits irrespective of blocker use
suggests that candesartan offers additive benefits and
complementary mechanisms to these other proven
treatments.
The primary outcome of all the component CHARM
trials, time to cardiovascular death or adjudicated
hospital admission for CHF, was consistently reduced in
symptomatic heart-failure patients. This finding, based on
more than 20 000 patient-years and more than 2450
events, is robust, and suggests that patients who have
symptomatic heart failure will derive important clinical
benefits from candesartan. The absence of heterogeneity
in results underscores that this benefit was achieved across
a broad spectrum of patients. Subgroup analyses must
be interpreted cautiously since the most rigorous test of
the study hypothesis is derived from the entire population,
in which consistency of this benefit was seen. Similar
reductions in mortality and morbidity outcomes with the
use of candesartan were obtained in women and men,
those with and without diabetes, and importantly, across
age-groups, a substantial number of patients being older
than 75 years. The beneficial effects of candesartan in the
CHARM programme were not altered by baseline use
of blockers, spironolactone, digoxin, aspirin, and
lipid-lowering treatments. This added efficacy on top
of blockers is particularly noteworthy, since 55% of
CHARM patients at baseline were receiving these drugs,
and a previous subgroup analysis from the Valsartan
Heart Failure Trial29 suggested less benefit of an
angiotensin-receptor blocker in patients already receiving
blockers.
The use of candesartan in patients with symptomatic
heart failure did not significantly reduce the risk of
myocardial infarction, stroke, or use of coronary
revascularisation procedures. However, the significant
benefits of candesartan treatment were maintained when
these non-fatal cardiovascular events were incorporated
with admission to hospital for heart failure and
cardiovascular death in a prespecified analysis of time to
first event. Lowering of blood pressure was more
pronounced in CHARM-Preserved than in the other
component trials and did not seem to be related to
improved clinical outcome. The frequency of new
diabetes was lower in the candesartan group than in the
placebo group, which is an effect that has been seen in
other large populations treated with inhibitors of
angiotensin-converting enzyme33 and angiotensin-receptor
blockers.27,28
Candesartan was generally well tolerated but was
associated with a greater occurrence of discontinuation
of study medication than was placebo because of
hypotension, hyperkalaemia, and an increase in serum
creatinine, underscoring the need to monitor patients.
Although more cancer deaths occurred in the candesartan
group, we attributed this imbalance to the play of
chance, since the investigator-reported rate of non-fatal
neoplasms did not differ between treatment groups.
Moreover, including CHARM together with the entire
previous candesartan placebo-controlled trial experience
(AstraZeneca, data on file), there were 523 (144 fatal)
investigator-reported neoplasms, including cancers, during
20 692 patient-years of exposure to candesartan compared
with 491 (125 fatal) in 20 135 patient-years with placebo
(p=0·6 and p=0·4, respectively). No consistent differences
in fatal or non-fatal neoplasms at different sites have been
noted between candesartan and placebo.
Our findings show that candesartan, given in titrated
doses as tolerated, can prolong survival, particularly
in patients with LVEF of 40% or less, and provide
incremental clinical benefits across the broad spectrum
of patients with symptomatic heart failure, including
reductions in hospital admissions for heart failure and
prevention of diabetes. This effect is consistent for the
combined cardiovascular mortality and morbidity
outcome, irrespective of other effective concomitant
treatments, ejection fraction, age, and sex. The clinical
effectiveness we report for candesartan in the treatment of
chronic heart failure offers the opportunity to further
reduce cardiovascular mortality and morbidity in this
expanding segment of our ageing population.