Customize your JAMA Network experience by selecting one or more topics from the list below.
Hjalmarson Å, Goldstein S, Fagerberg B, et al. Effects of Controlled-Release Metoprolol on Total Mortality, Hospitalizations, and Well-being in Patients With Heart FailureThe Metoprolol CR/XL Randomized Intervention Trial in Congestive Heart Failure (MERIT-HF). JAMA. 2000;283(10):1295–1302. doi:10.1001/jama.283.10.1295
Context Results from recent studies on the effects of β1-blockade
in patients with heart failure demonstrated a 34% reduction in total mortality.
However, the effect of β1-blockade on the frequency of hospitalizations,
symptoms, and quality of life in patients with heart failure has not been
Objective To examine the effects of the β1-blocker controlled-release/extended-release
metoprolol succinate (metoprolol CR/XL) on mortality, hospitalization, symptoms,
and quality of life in patients with heart failure.
Design Randomized, double-blind controlled trial, preceded by a 2-week single-blind
placebo run-in period, conducted from February 14, 1997, to October 31, 1998,
with a mean follow-up of 1 year.
Setting Three hundred thirteen sites in 14 countries.
Participants Patients (n = 3991) with chronic heart failure, New York Heart Association
(NYHA) functional class II to IV, and ejection fraction of 0.40 or less who
were stabilized with optimum standard therapy.
Interventions Patients were randomized to metoprolol CR/XL, 25 mg once per day (NYHA
class II), or 12.5 mg once per day (NYHA class III or IV), titrated for 6
to 8 weeks up to a target dosage of 200 mg once per day (n = 1990); or matching
placebo (n = 2001).
Main Outcome Measures Total mortality or any hospitalization (time to first event), number
of hospitalizations for worsening heart failure, and change in NYHA class,
by intervention group; quality of life was assessed in a substudy of 741 patients.
Results The incidence of all predefined end points was lower in the metoprolol
CR/XL group than in the placebo group, including total mortality or all-cause
hospitalizations (the prespecified second primary end point; 641 vs 767 events;
risk reduction, 19%; 95% confidence interval [CI], 10%-27%; P<.001); total mortality or hospitalizations due to worsening heart
failure (311 vs 439 events; risk reduction, 31%; 95% CI, 20%-40%; P<.001), number of hospitalizations due to worsening heart failure
(317 vs 451; P<.001); and number of days in hospital
due to worsening heart failure (3401 vs 5303 days; P<.001).
NYHA functional class, assessed by physicians, and McMaster Overall Treatment
Evaluation score, assessed by patients, both improved in the metoprolol CR/XL
group compared with the placebo group (P = .003 and P = .009, respectively).
Conclusions In this study of patients with symptomatic heart failure, metoprolol
CR/XL improved survival, reduced the need for hospitalizations due to worsening
heart failure, improved NYHA functional class, and had beneficial effects
on patient well-being.
Chronic heart failure is a common disease that has a poor prognosis
and periods of incapacitating symptoms necessitating recurrent hospital admissions.1,2 The most common modes of death are
sudden death or death from worsening heart failure.3
The discovery of the pathophysiological importance of neuroendocrine activation
in heart failure and the possibility of modifying such mechanisms of the disease
process have greatly improved treatment in clinical practice.4
Thus, angiotensin-converting enzyme (ACE) inhibitors have been established
as standard therapy for patients with chronic heart failure due to left ventricular
systolic dysfunction, with proven effects on mortality and symptoms related
to worsening heart failure.4,5
Despite the benefits of this mode of therapy, mortality and morbidity remain
high for patients with heart failure.
The role of β-blocker treatment in the management of chronic heart
failure has taken time to clarify. The results from meta-analyses of previous
smaller studies of various β-blockers in heart failure, including the
carvedilol studies, have indicated beneficial effects.6-8
Two studies on the survival effects of β1-blockade published
in 1999, the Cardiac Insufficiency Bisoprolol Study (CIBIS) II9
and the present Metoprolol CR/XL Randomized Intervention Trial in Congestive
Heart Failure (MERIT-HF),10 demonstrated that
total mortality was reduced by 34%.
Although the survival benefit of β1-blockade in chronic
heart failure due to systolic dysfunction has been established, the need for
hospital care, safety aspects, symptom alleviation, and improved quality of
life are additional important aspects of treatment, for both the patient and
the clinician. However, the impact of β-blockers on these outcomes has
not been fully explored. Accordingly, the MERIT-HF was designed to study the
effects of controlled-release/extended-release metoprolol succinate (metoprolol
CR/XL) on mortality, as previously reported,10
as well as hospitalizations, symptoms, and quality of life.
The MERIT-HF was a randomized, double-blind placebo-controlled trial
with a single-blind, 2-week placebo run-in period. Randomization was performed
according to an optimal allocation procedure, which balanced the metoprolol
CR/XL and placebo groups for investigational site, age, sex, race/ethnicity,
cause of heart failure, previous acute myocardial infarction (AMI), and, within
the previous AMI group, time since last AMI, diabetes mellitus, ejection fraction,
and New York Heart Association (NYHA) functional class. An interactive voice
recording system (Covance, Princeton, NJ) was used to provide investigators
with the computer-generated study drug number based on the optimal allocation
A total of 3991 patients with symptomatic chronic heart failure and
decreased ejection fraction who were stabilized with standard treatment were
randomized (Figure 1) at 313 investigational
sites in the United States and 13 European countries (Belgium, Czech Republic,
Denmark, Finland, Germany, Hungary, Iceland, the Netherlands, Norway, Poland,
Sweden, Switzerland, and the United Kingdom). The study was approved by the
institutional review board of each hospital and all patients provided written
The Independent Endpoint Committee, whose 5 members were unaware of
treatment status, classified all events from copies of medical charts and
other documents according to prespecified definitions. An Independent Safety
Committee monitored safety issues during the study. The stopping rule for
efficacy was based on the total number of expected deaths, analyzed based
on the intent-to-treat principle. The study used an asymmetrical group sequential
procedure to monitor total mortality. A Peto-type boundary11
was used for monitoring a positive trend. This approach favors a large critical Z-value for all interim tests before the end of the trial.
The cumulative α level was planned to be .0012, .0024, and .0036 at
the first, second, and third interim analyses to take place when 25%, 50%,
and 75%, respectively, of the total number of expected deaths had occurred.
The cumulative probability of early stopping for harm was planned to be .005,
.010, and .015 at the first, second, and third interim analyses, respectively.
There were 2 primary outcome measures: total mortality and the combined
end point of total mortality or all-cause hospitalization (time to first event).
The MERIT-HF was stopped early, on October 31, 1998, because the second preplanned
interim analysis showed a significant 34% reduction in total mortality in
the metoprolol CR/XL group.10 As previously
reported, 145 patients died in the metoprolol CR/XL group compared with 217
in the placebo group.10
The following combined end points (time to first event) were also predefined:
total mortality or hospitalization due to worsening heart failure; death or
heart transplantation; cardiac death or nonfatal AMI; and total mortality
or hospitalization due to worsening heart failure or emergency department
visit due to worsening heart failure. Other end points were number of hospitalizations
due to heart failure and other cardiovascular causes, withdrawal of study
drug due to worsening heart failure, and change in NYHA functional class.
Effect on quality of life was assessed in a substudy that was conducted in
the United States, United Kingdom, Sweden, Norway, and the Netherlands.
The major inclusion criteria were symptomatic heart failure for at least
3 months, corresponding to NYHA class II to IV, and a left ventricular ejection
fraction of 0.40 or less in men and women aged 40 to 80 years. For patients
with an ejection fraction between 0.36 and 0.40, it was mandatory that a 6-minute
walk test result did not exceed 500 yd (450 m). Resting heart rate had to
be 68/min or more. Patients had to be receiving optimal treatment (defined
as any combination of diuretics and an ACE inhibitor) for at least 2 weeks
prior to randomization. If an ACE inhibitor was not tolerated, hydralazine,
long-acting nitrate, or an angiotensin II blocker could be used. Digitalis
also could be prescribed. In addition, the inclusion criteria included a stable
clinical condition during the 2-week placebo run-in phase before randomization
The main exclusion criteria included AMI or unstable angina pectoris
within 28 days before randomization, indication or contraindication for treatment
with β1-blockade, severe decompensated heart failure (eg,
pulmonary edema, hypoperfusion), or supine systolic blood pressure of less
than 100 mm Hg. A more detailed description of the study protocol has been
At the randomization visit, patients were allocated to treatment with
metoprolol CR/XL or placebo administered once daily. The starting dosage was
one 25-mg tablet once per day (half of a 25-mg tablet was recommended for
patients with NYHA functional class III or IV). It was recommended to double
the dosage after each 2-week period to reach the target dosage level of 200
mg/d of metoprolol CR/XL or placebo. This regimen could be modified according
to the judgment of the investigator. If a patient did not tolerate increased
titration of study drug, temporary reduction in dosage or increase in diuretic
dosage was advocated. During follow-up, patient visits were scheduled every
At each visit, the investigators judged and documented the patient's
NYHA functional class. The Minnesota Living with Heart Failure questionnaire
was completed by patients at randomization, after each 6-month treatment period,
and at study closure.13 This questionnaire
consists of 21 items; the total score ranges from 0 to 105, with lower scores
indicating better quality of life. The McMaster Overall Treatment Evaluation
questionnaire (OTE) was completed by the patients after each 6-month treatment
period and at study closure.14 This questionnaire
has 3 items that assess the overall effect according to whether a patient
experienced any change in activity limitation, symptoms, or feelings since
the treatment started, using 7-point scales. Any improvement or deterioration
was subsequently scored by the patient in terms of magnitude and importance
to the patient's ability to carry out daily activities.
Hospitalizations were defined as care at an acute-care hospital lasting
for 24 hours or more and had to be separated from other hospitalizations by
separate dates for discharge and admission. Transfer from one ward to a different
type of hospital ward was counted as 1 hospitalization. Hospitalization due
to heart failure was defined as documentation in the medical charts indicating
worsening heart failure as the reason for hospitalization. If competing reasons
were judged to be of equal importance, the heart failure diagnosis took preference.
Emergency department visit was defined as care in an urgent fashion with urgent-care
treatment such as intravenous medication.
The power calculation showed that the mean follow-up time had to be
2.4 years if 1600 patients were randomized to each treatment group over 14
months. This was based on a significance level of α = .04 (2-sided)
for the first primary end point of total mortality and α = .01 for the
second primary end point of total mortality or any hospitalization (time to
first event), a power of at least 80% (β≤.2), and the following assumptions:
a 9.4% mean annual mortality in the placebo group, a mean risk-reducing effect
of metoprolol CR/XL of 30% (with treatment), and a withdrawal rate from study
drug of 20% the first year and 5% annually thereafter.12
Because patient recruitment proceeded faster than planned, 3991 patients were
randomized during the recruitment period, thereby increasing the power of
The analysis was by intent to treat. The main analyses used the log-rank
test for the comparison of the 2 randomized groups and the Cox proportional
hazards model to calculate relative risk and 95% confidence intervals (CIs).
Additional Cox proportional hazards regression analyses of the combined end
points of total mortality or all-cause hospitalizations (time to first event)
and total mortality or hospitalizations due to heart failure (time to first
event) were performed to explore any unfavorable outcome in prespecified risk
groups, defined by entry characteristics as previously described.10,12 For ejection fraction, systolic and
diastolic blood pressure, and heart rate, patients in the lowest tertile were
compared with those in the middle and upper tertiles. Regarding age, the upper
tertile was compared with the middle and lower tertiles. New York Heart Association
class, etiology of heart failure, smoking status, sex, previous AMI, diabetes
mellitus, and hypertension were also prespecified as risk groups. Ischemic
and nonischemic heart disease have been defined as the 2 major causes of heart
failure. Hypertension was defined as pharmacologically treated high blood
pressure, and diabetes mellitus was defined as a clinical diagnosis made by
the investigator. More than 180 events in any such subgroup would yield a
power of at least 70% to detect a 30% increase in risk. Data on complementary
subgroups having less than 180 events have also been depicted.
The sample size calculation for the quality of life substudy showed
that with 419 patients in each group, it would be possible to detect a difference
of 3 units on total Living with Heart Failure score between the treatment
groups based on the following assumptions: SD for change = 16, α = .05,
and β = .20. A net difference of 3 units was judged to be a clinically
meaningful change. The changes in NYHA class and OTE score were tested by
means of a permutation test using raw data scores. Changes in Living with
Heart Failure score were analyzed using an analysis of covariance model with
adjustment for the baseline Living with Heart Failure score. A 2-sided P<.05 was regarded as statistically significant.
Randomization began on February 14, 1997, and the last patient was randomized
on April 14, 1998. The International Steering Committee stopped the study
on October 31, 1998, on recommendation from the Independent Safety Committee.
The second preplanned interim analysis (at the halfway point) had shown that
the predefined criterion for termination of the study was met and exceeded.
In total, 2004 patient-years were accumulated in the metoprolol CR/XL group
and 1977 in the placebo group (total mortality). The corresponding patient-years
for the combined end point of total mortality or all-cause hospitalization
were 1650 vs 1600 patient-years, and for total mortality or hospitalization
for worsening heart failure were 1880 vs 1840 patient-years, respectively.
The mean follow-up time was 1 year.
The 2 groups were similar at entry (Table 1). Furosemide daily dosage at baseline and during follow-up
was 66 mg/d and 70 mg/d in the metoprolol CR/XL group and 65 mg/d and 73 mg/d
in the placebo group, respectively. The ACE inhibitor daily dosage was also
similar at baseline and during follow-up in both randomization groups. For
enalapril, it was 14 mg/d at baseline and 15 mg/d at follow-up in both groups;
corresponding dosages at baseline and follow-up, respectively, for captopril
were 68 mg/d vs 70 mg/d in the metoprolol CR/XL group and 60 mg/d vs 64 mg/d
in the placebo group, and for lisinopril were 17 mg/d vs 17 mg/d in the metoprolol
CR/XL group and 16 mg/d vs 16 mg/d in the placebo group.
The patients who participated in the quality of life substudy (n = 741)
had characteristics similar to those of the entire group (mean age, 64.4 years;
female sex, 28%; NYHA class II, 39%; NYHA class III, 56%; NYHA class IV, 5%;
mean ejection fraction, 0.27; previous AMI, 51%; and treatment with ACE inhibitor
or angiotensin II blocker, 96%).
Metoprolol CR/XL significantly reduced all combined end points (time
to first event) compared with placebo (Table 2, Figure 2, and Figure 3). Total mortality or all-cause hospitalizations
(the prespecified second primary end point) was reduced by 19% (Figure 2), total mortality or hospitalization for worsening heart
failure by 31% (Figure 3), death
or heart transplantation by 32%, cardiac death or nonfatal AMI by 39% (Figure 3), and total mortality or hospitalization
due to worsening heart failure or emergency department visit due to worsening
heart failure by 32%. No significant increase in total mortality or all-cause
hospitalizations (time to first event) or in total mortality or hospitalizations
due to worsening heart failure (time to first event) were observed in any
of the predefined subgroups analyzed for safety reasons (Figure 4).
Compared with placebo, metoprolol CR/XL reduced the number of patients
with any hospitalization, the total number of hospitalizations, and the total
number of days in the hospital due to all causes (Figure 5, Table 3).
This was mainly explained by a reduction in the number of patients who were
hospitalized for worsening heart failure, accompanied by decreases in the
total number of hospitalizations and total number of days in the hospital
due to heart failure (Table 3, Figure 5). To account for the improved survival
with metoprolol CR/XL, the proportion of days spent alive outside the hospital
was also calculated as 95% in the metoprolol CR/XL group and 93% in the placebo
Physicians classified NYHA functional class at baseline and at the last
visit in the study. Improvement was recorded in 28.6% vs 25.8% of the metoprolol
CR/XL and placebo groups, respectively (26.0% vs 24.3% improved 1 class; 2.6%
vs 1.5% improved 2 classes); 65.4% vs 66.7% were unchanged; 6.0% vs 7.5% deteriorated
(5.7% vs 6.8% deteriorated 1 class, 0.3% vs 0.7% deteriorated 2 classes).
These data show a more favorable change in NYHA class in the metoprolol CR/XL
group compared with the placebo group (P = .003).
There was a statistically significant improvement in the OTE score in
the metoprolol CR/XL group compared with placebo (P
= .009; Figure 6). In the metoprolol
CR/XL group, 185 patients (50%) reported improvement, and patients' evaluations
of the importance of this change were available for 184 patients, showing
that 132 patients (72%) judged this improvement as important, very important,
or extremely important to carry out daily activities. In the placebo group,
148 patients (40%) reported improvement that was judged to be important, very
important, or extremely important by 72% of these patients.
Living with Heart Failure forms completed at randomization and at the
last visit were available for 670 patients. Scores were similar at randomization
in the 2 study groups. The total Living with Heart Failure score, adjusted
for the score at baseline, decreased (improved) by 0.7 in the metoprolol CR/XL
group (n = 331)and increased (deteriorated) by 0.2 in the placebo group (n
= 339) (mean difference, –0.9; 95% CI, –3.4 to 1.6; P = .20).
The most frequent adverse events necessitating withdrawal of study drug
were worsening heart failure, atrial fibrillation, and angina pectoris, and
these events were less frequent in the metoprolol CR/XL group than in the
placebo group (Table 4). Dizziness,
bradycardia, and hypotension occurred slightly more frequently in the metoprolol
Permanent withdrawal of study drug due to any cause during the study
is shown in Figure 2 and occurred
in 279 patients in the metoprolol CR/XL group and 310 patients in the placebo
group (risk reduction for withdrawal decreased by 10% in the metoprolol CR/XL
group; 95% CI, −5% to 24%; P = .18). Permanent
withdrawal of study drug due to any adverse event occurred in 196 patients
in the metoprolol CR/XL group and 234 patients in the placebo group (risk
reduction, 17%; 95% CI, −1% to 31%; P = .06).
Worsening heart failure was the main reason for withdrawal in 64 patients
(3.2%) in the metoprolol CR/XL group and 85 (4.2%) in the placebo group (risk
reduction, 25%; 95% CI, −4% to 46%; P = .08).
This study demonstrated that metoprolol CR/XL, a β1-blocker
given once per day in addition to conventional therapy to patients with chronic
heart failure, improved survival as previously reported,10
reduced the need for hospital admissions due to worsening heart failure, and
improved symptoms and well-being.
In the MERIT-HF, there were no differences between the study groups
in underlying pharmacological treatment for heart failure at baseline or during
follow-up. All-cause mortality or hospitalizations (the prespecified second
primary end point, time to first event) was significantly reduced by 19% in
the metoprolol CR/XL group. Correspondingly, total deaths or hospitalizations
due to worsening heart failure were reduced by 31%. The annual mortality in
the placebo group was 11.2% and total mortality was reduced by 34% in the
metoprolol CR/XL group.10 Despite this improved
survival and more patients at risk for hospital admissions, fewer patients
were hospitalized due to any cause in the metoprolol CR/XL group, mainly due
to a 35% reduction in the number of patients hospitalized for worsening heart
failure. The total number of days in the hospital due to heart failure was
reduced to a similar degree in the metoprolol CR/XL group, with no increase
in hospitalizations for other reasons. Conversely, the number of days alive
without need for hospital care was higher in the metoprolol CR/XL group than
in the placebo group. Given the comparatively low cost of β-blocker therapy
and the high cost of hospitalizations, the 36% reduction in days spent in
the hospital for worsening heart failure suggests a positive effect on health
care costs with metoprolol CR/XL treatment in patients with chronic heart
failure. It is noteworthy that there was considerable comorbidity among these
patients because hospitalizations due to noncardiovascular causes accounted
for 30% of all days spent in the hospital and one quarter of all days was
due to cardiovascular causes other than worsening heart failure (Table 3, placebo group).
Our results are consistent with data from the CIBIS-II study, in which
treatment with the β1-blocker bisoprolol also reduced hospital
admissions due to any cause and due to worsening heart failure.9
The smaller carvedilol studies also have shown reductions in hospitalizations.7,8
The mechanisms involved in the beneficial effects of β1-blockade
in patients with chronic heart failure are not completely known. However,
it is well established that in patients with chronic heart failure due to
systolic dysfunction of various etiologies, metoprolol has favorable effects
on left ventricular geometry and function, myocardial energy balance, and
In patients with dilated cardiomyopathy, metoprolol has been found to reduce
heart transplantations and to improve NYHA functional class and quality of
life.17,18 In line with these
previous observations, the MERIT-HF results showed that symptoms of heart
failure improved according to judgments made by the responsible physicians,
as did quality of life as assessed by patients on the OTE. Improvement in
quality of life was of considerable importance for a large proportion of patients.
The observation that the OTE but not the Living with Heart Failure questionnaire
showed significant improvement in quality of life may reflect the different
constructions of these 2 instruments. The former is a global assessment of
treatment effects on activity limitation, symptoms, and feelings, whereas
the latter is a 21-item questionnaire used at baseline and during follow-up
(see "Methods" section). Given the decrease in hospitalizations and the improvement
in both NYHA class and in OTE score, the sensitivity of the Living with Heart
Failure questionnaire to assess beneficial changes in quality of life might
be questionable. A recent study showed that retrospective measures, such as
the 7-point scale used in the OTE, may be more sensitive to change than serial
measures and also correlate more strongly with patient's' satisfaction with
change.19 The previously published documentation
of the effect of β-blockade on the Living with Heart Failure score in
patients with heart failure relates mainly to nonselective β-blockade
with carvedilol therapy, showing no statistically significant effect.20-22
Metoprolol CR/XL was well tolerated. Withdrawal of study drug from all
causes was 10% lower10 and withdrawal due to
worsening heart failure was 25% lower in the metoprolol CR/XL group compared
with the placebo group. These findings are of interest, especially against
the background that no metoprolol CR/XL test dose had been given prior to
initiating double-blind treatment. For the most frequent adverse reactions
leading to withdrawal of study drug, including worsening heart failure, atrial
fibrillation, and angina pectoris, withdrawal was more common in the placebo
group. Fewer than 1 of 100 patients treated for 1 year withdrew from metoprolol
CR/XL treatment because of bradycardia, dizziness, or hypotension. There were
no specific safety concerns observed in any of the preidentified risk groups.
In this study, controlled-release/extended-release metoprolol succinate,
once per day, was used. This formulation leads to a more pronounced and even β-blockade
over 24 hours compared with conventional immediate-release metoprolol tartrate
tablets, 50 mg 3 times per day.23 In patients
with chronic heart failure, the dosing schedule can be simplified with metoprolol
CR/XL and the target dosage also can be increased to 200 mg once per day compared
with 50 mg 3 times per day with the conventional formulation, without increasing
the peak plasma concentration of the drug.24
The titration schedule started with a low once-daily dosage, 25 mg/d for those
in NYHA class II and 12.5 mg/d for those in NYHA class III or IV, with increased
titration every 2 weeks. The target dosage, 200 mg/d, was reached by 64% of
the patients, and 87% received 100 mg/d or more.10
The mean dosage was 159 mg/d.10 The combined
results from MERIT-HF and CIBIS II demonstrate that it is safe to treat patients
with heart failure with β1-blockers by using a low starting
dosage and gradual increased titration.
This study has several limitations. Several categories of patients were
not included (eg, patients with severe heart failure who were confined to
bed, patients with heart failure and an ejection fraction of more than 0.40,
and patients with heart failure early after AMI). The group of patients in
NYHA functional class IV was small, resulting in wide 95% CIs overlapping
those in NYHA functional classes II and III.10
However, results from a recent meta-analysis of several studies indicate that
treatment with β-blockers confers significant beneficial effects on the
clinical outcome in patients in NYHA class IV.25
In conclusion, the MERIT-HF study demonstrates that treatment with metoprolol
CR/XL once daily added to standard therapy for patients with mild to severe
heart failure due to left ventricular systolic dysfunction improves survival,
reduces the need for hospital admissions due to worsening heart failure, improves
symptoms of heart failure, and increases well-being.