BACKGROUND: Guidelines recommend that patients with heart failure receive beta-blockers in doses used in the trials that have proven their efficacy. Although the adverse effects of beta-blockade are dose-related, it is unclear whether the benefits are. PURPOSE: To determine whether the survival benefits of beta-blockade in heart failure are associated with the magnitude of heart rate reduction or the beta-blocker dose. DATA SOURCES: MEDLINE, EMBASE, CINAHL, SIGLE, Web of Science, and the Cochrane Central Register of Controlled Trials, supplemented by hand-searches of bibliographies. STUDY SELECTION: Randomized, placebo-controlled heart failure trials that reported all-cause mortality. DATA EXTRACTION: Two reviewers independently extracted data on study characteristics, beta-blocker dosing and heart rate reduction, and death. DATA SYNTHESIS: The mean left ventricular ejection fraction in the 23 beta-blocker trials ranged from 0.17 to 0.36, and more than 95% of the 19 209 patients had systolic dysfunction. The overall risk ratio for death was 0.76 (95% CI, 0.68 to 0.84); however, heterogeneity testing revealed moderate heterogeneity among trials (I (2) = 30%), which was associated with the magnitude of heart rate reduction achieved within each trial (P for meta-regression = 0.006). For every heart rate reduction of 5 beats/min with beta-blocker treatment, a commensurate 18% reduction (CI, 6% to 29%) in the risk for death occurred. No significant relationship between all-cause mortality and beta-blocker dosing was observed (risk ratio for death, 0.74 CI, 0.64 to 0.86) in high-dose beta-blocker trials vs. 0.78 CI, 0.63 to 0.96 in low-dose beta-blocker trials; P for meta-regression = 0.69). LIMITATIONS: The analysis is based on aggregate data and resting heart rates. Few patients in these trials had bradycardia or diastolic dysfunction at baseline. CONCLUSION: The magnitude of heart rate reduction is statistically significantly associated with the survival benefit of beta-blockers in heart failure, whereas the dose of beta-blocker is not.
Finlay A. McAlister (2009) studied this question.