Michael J Domanski, Heidi Krause-Steinrauf, Barry M Massie, Prakash Deedwania, Dean Follmann, David Kovar, David Murray, Ron Oren, Yves Rosenberg, James Young, Michael Zile, Eric Eichhorn, and null for the Best Investigators
Background Recent large randomized, controlled trials (BEST [β-blocker Evaluation of Survival Trial], CIBIS-II [Cardiac Insufficiency Bisoprolol Trial II], COPERNICUS [Carvedilol Prospective Randomized Cumulative Survival Study], and MERIT-HF [Metoprolol Randomized Intervention Trial in Congestive Heart Failure]) have addressed the usefulness of β-blockade in the treatment of advanced heart failure. CIBIS-II, COPERNICUS, and MERIT-HF have shown that β-blocker treatment with bisoprolol, carvedilol, and metoprolol XL, respectively, reduce mortality in advanced heart failure patients, whereas BEST found a statistically nonsignificant trend toward reduced mortality with bucindolol. We conducted a post hoc analysis to determine whether the response to β-blockade in BEST could be related to differences in the clinical and demographic characteristics of the study populations. We generated a sample from BEST to resemble the patient cohorts studied in CIBIS-II and MERIT-HF to find out whether the response to β-blocker therapy was similar to that reported in the other trials. These findings are further compared with COPERNICUS, which entered patients with more severe heart failure. Methods To achieve conformity with the entry criteria for CIBIS-II and MERIT-HF, the BEST study population was adjusted to exclude patients with systolic blood pressure 80 years (exclusion criteria employed in those trials). The BEST comparison subgroup (BCG) was further modified to more closely reflect the racial demographics reported for patients enrolled in CIBIS-II and MERIT-HF. The association of β-blocker therapy with overall survival and survival free of cardiac death, sudden cardiac death, and progressive pump failure in the BCG was assessed. Results In the BCG subgroup, bucindolol treatment was associated with significantly lower risk of death from all causes (hazard ratio (HR) = 0.77 [95% CI = 0.65, 0.92]), cardiovascular death (HR = 0.71 [0.58, 0.86]), sudden death (HR = 0.77 [0.59, 0.999]), and pump failure death (HR = 0.64 [0.45, 0.91]). Conclusions Although not excluding the possibility of differences resulting from chance alone or to different properties among β-blockers, this study suggests the possibility that different heart failure population subgroups may have different responses to β-blocker therapy.