Review of the COMET trial

Comparison of carvedilol and metoprolol on clinical outcomes in patients with chronic heart failure in the Carvedilol Or Metoprolol European Trial (COMET): randomised controlled trial

Lancet 2003;362:7-13

Background: Beta blockers have different pharmacologic properties. Selective beta-blockers like Metoprolol and Bisoprolol have a high specificity for beta-1 adrenergic receptors, which are found primarily in the heart. Blocking these receptors lowers the heart rate and reduces contractility. Non-selective beta-blockers like Carvedilol and Bucindolol block beta-1 receptors as well as beta-2, and alpha-1 adrenergic receptors and also have blood pressure lowering effects. Over a period of months (not days or weeks) both types of beta-blockers increase ejection fraction by restoring the density of beta-adrenergic receptors in the heart, which are down regulated due to sympathetic overactivity in chronic heart failure.

Thus far, we have reviewed clinical trials involving the selective beta-blockers Metoprolol Succinate (MERIT-HF) and Bisoprolol (CIBIS-II), which were both positive for reducing all-cause mortality. We also reviewed 2 trials involving the non-selective beta blocker Carvedilol (here and here), which significantly reduced mortality as well as 1 trial involving Bucindolol (BEST), which did not find a significant mortality reduction. At the time the COMET trial was undertaken, only 1 of the 5 trials mentioned above had been published - it was a small trial involving Carvedilol with several significant limitations.

The Carvedilol Or Metoprolol European Trial (COMET) was performed to directly compare the effects of Carvedilol and Metoprolol Tartrate on mortality and morbidity in patients with mild to severe chronic systolic heart failure. Of note, Metoprolol Tartrate was used in this trial because Metoprolol Succinate was not available at the start of the trial, which began enrolling patients in December of 1996.

Patients: The COMET trial included men and women with symptomatic, chronic systolic heart failure (NYHA class II-IV) who had at least 1 cardiovascular admission during the previous 2 years. Patients had to be on stable heart failure therapy with an ACE inhibitor for at least 4 weeks, and on treatment with diuretics (>/=40 mg of furosemide or equivalent) for at least 2 weeks. Other drug classes could be used at the discretion of investigators. LVEF had to be </=35% within the previous 3 months.

Patients were excluded if there had a recent change of heart failure treatment within the 2 weeks prior to randomization, if they required intravenous inotropic therapy, if they were currently being treated with calcium channel blockers, amiodarone or class-I antiarrhythmic drugs. They were also excluded if they had unstable angina, myocardial infarction, or coronary revascularization with the previous 2 months. Other exclusion criteria included uncontrolled hypertension, resting heart rate <60 bpm, sick sinus syndrome, 2nd or 3rd degree AV block unless treated with a pacemaker, SBP <85 mmHg, history of asthma or COPD, peripheral arterial disease with rest symptoms, unstable insulin-dependent diabetes, hemodynamically significant valvular disease, significant ventricular arrhythmias, pregnancy, history of drug or alcohol abuse, poor adherence with medical treatment or any other serious systemic illness that could complicate management or reduce life expectancy.

Baseline Characteristics: The average age of patients was 62 years, 80% were male, 48%, 48%, and 3% were NYHA class II, III, and IV, respectively. Almost all patients were white (99%). The mean LVEF was 26%. Slightly over half of all patients had ischemic cardiomyopathy and 18% had hypertensive heart disease. The mean SBP was 126 mmHg and the resting heart rate was 81 bpm. The mean duration of heart failure was 3.5 years. Atrial fibrillation was present in 20% of patients. Nearly all patients were on an ACE inhibitor or ARB, 11% were on an aldosterone antagonist and 60% were on digoxin.

Trial Procedures: At randomization, patients were assigned to receive either 3.125 mg of carvedilol twice daily or 5 mg of metoprolol tartrate twice daily. Investigators and patients were blinded to the treatment. Each drug was increased to the target dose by doubling the dose every 2 weeks (i.e., carvedilol went from 3.125 to 6.25 to 12.5 to 25 mg twice daily and metoprolol went from 5 mg to 12.5 mg to 25 mg to 50 mg twice daily). Thus, for each drug, it took 6 weeks to reach the target dose, which was 25 mg twice daily of carvedilol and 50 mg twice daily of metoprolol tartrate. Once the target dose was achieved, patients were followed every 4 months until the end of the study.

Endpoints: The primary endpoint of the trial was all-cause mortality. A second co-primary endpoint was added after the start of the study that included all-cause mortality or all-cause hospitalization. This was an event-driven trial and it was estimated that 1020 fatal events would yield 80% power to detect a 20% relative risk reduction for the primary endpoint of all-cause mortality. This was based on a yearly mortality rate of 8.1%. It was estimated that 2400 events would yield 80% power to detect a 15% relative risk reduction in the composite endpoint. The type 1 error was split - 0.04 for all-cause mortality and 0.01 for the composite endpoint.

Results: A total of 3029 patients were enrolled with 1511 randomized to carvedilol and 1518 to metoprolol tartrate. The average follow-up time was 58 months or 4.8 years and 1112 deaths occurred. Compared to metoprolol tartrate, carvedilol significantly reduced the primary endpoint of all-cause death (34% vs 40%; HR 0.83; 95% CI 0.74-0.93) as well as the secondary endpoint of cardiovascular death (29% vs 35%; HR 0.80; 95% CI 0.70-0.90). Carvedilol did not significantly reduce the co-primary endpoint of death or hospitalization (74% vs 76%; HR 0.94; 95% CI 0.86-1.02).

The mean daily dose of carvedilol was 42 mg with 75% receiving the target dose of 25 mg twice daily. The mean daily dose of metoprolol tartrate was 85 mg with 78% receiving the target dose of 50 mg twice daily. After 4 months of treatment, the mean heart rate decreased from baseline by 13 bpm in the carvedilol group and by 12 bpm in the metoprolol tartrate group and the mean SBP decreased from baseline by 4 mmHg in the carvedilol group and 2 mmHg in metoprolol group, respectively. The drug was permanently discontinued for reasons other than death in 32% of patients in both groups.

Conclusions: When carvedilol is titrated to a target dose of 25 mg twice daily, all-cause death is significantly reduced compared to when metoprolol tartrate is titrated to a target dose of 50 mg twice daily. The number of patients of who need to be treated with this carvedilol regimen compared to this metoprolol tartrate regimen to prevent 1 death over 4.8 years is approximately 17 patients. While on the surface, this may seem like a resounding victory for the non-selective beta-blocker carvedilol over the selective beta-blocker metoprolol, we have reservations.

When the COMET trial was designed and began enrolling patients, metoprolol succinate (the long-acting form of the drug) was not available. This is important because metoprolol succinate is the form of the drug that was used in the MERIT-HF trial, which we have reviewed previously and was published prior to completion of COMET. MERIT-HF found a significant reduction in all-cause death that was similar in scale to the reductions reported in trials involving carvedilol (here and here). However, in MERIT-HF, the target dose of metoprolol succinate was 200 mg daily and the average daily dose achieved was 159 mg with 64% of patients receiving the full target dose. In COMET, the target dose of metoprolol tartrate was only 100 mg daily and the average daily dose achieved was 85 mg. Thus, patients in COMET who were randomized to metoprolol tartrate did not receive an equivalent dose of metoprolol compared to the MERIT-HF trial and arguably, they did not receive an equivalent dose relative to the dose of carvedilol that was used, which likely biased the results in favor of carvedilol.

In the discussion of COMET, the investigators provided their rationale for why the the dose of metoprolol tartrate was chosen and also provide a defense of why they feel the dose of metoprolol did not bias the trial in favor of carvedilol. It is clear from the discussion that the investigators believe carvedilol is the superior agent in this patient population; however, we do not concur with their conclusions and feel that their is insufficient evidence to declare carvedilol superior to metoprolol.

It should be noted that metoprolol tartrate is not generally considered an acceptable beta-blocker to be used in patients with chronic systolic heart failure due to the results of this trial. In our opinion, that is an over-extrapolation; however, metoprolol succinate is recommended and does not pose a hardship to patients in terms of cost or tolerability relative to tartrate and so we have no reservations using it in its place.

In our opinion, both agents, metoprolol and carvedilol, are effective for reducing morbidity and mortality in well-selected patients with chronic systolic heart failure. Despite the results of COMET, we feel there is insufficient evidence to declare 1 agent is better than the other or to declare that non-selective beta blockers are better than selective ones for patients with chronic systolic heart failure.