Correspondence
Effects of Beta-Blocker Therapy in Severe Chronic Heart Failure
N Engl J Med 2001; 345:998-999September 27, 2001
- Article
To the Editor:
Packer et al. (May 31 issue)1 report that the Carvedilol Prospective Randomized Cumulative Survival Study was stopped early on the recommendation of the data and safety monitoring board because of a significant beneficial effect of carvedilol on survival. The patients in the study had a mean ejection fraction of 19.8 percent. The mean duration of follow-up was 10.4 months. At the time that the study was terminated, the mortality rate was 16.8 percent in the placebo group and 11.2 percent in the carvedilol group.
In the Beta-Blocker Evaluation of Survival Trial (BEST), reported in the same issue of the Journal, 2 the patients who received bucindolol, although they had a mean ejection fraction of 23 percent, actually had a slightly worse outcome than the patients who received carvedilol in the study by Packer et al. In the BEST study, the mortality rate was 33 percent in the placebo group and 30 percent in the bucindolol group at 24 months, a difference that was not statistically significant. This study was also terminated early by the data and safety monitoring board. At 10 months, there was apparently no difference in survival between the groups, and at this time, the mortality rate was slightly higher overall in the bucindolol study than in the carvedilol study.
One major difference in the base-line therapy received by the patients in the two studies might explain the apparently contradictory results. Twenty percent of the patients enrolled in the carvedilol study received spironolactone, but only 4 percent of the patients in the bucindolol study received this medication.
In the Randomized Aldactone Evaluation Study (RALES),3 a mean dose of 26 mg of spironolactone given daily reduced mortality among patients with severe heart failure (New York Heart Association class III or IV; mean ejection fraction, 25 percent), with similar effects in patients receiving concomitant beta-blocker therapy (11 percent of the patients in the spironolactone group). Reductions in the risk of death and hospitalization occurred after just 2 to 3 months of treatment and persisted for the duration of the study (mean follow-up, 24 months). Could the authors of the carvedilol and bucindolol studies provide insight into the effects of combined therapy with a beta-blocker and spironolactone?
3 ReferencesChristian Negut, M.D.
University of Padua, 35100 Padua, Italy
bonvicini_negut@tin.it 1
Packer M ,Coats AJS ,Fowler MB , et al. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med 2001;344:1651-1658
Full Text | Web of Science | Medline2
The Beta-Blocker Evaluation of Survival Trial Investigators
Full Text | Web of Science | Medline3
Pitt B ,Zannad F ,Remme WJ , et al. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med 1999;341:709-717
Full Text | Web of Science | Medline
To the Editor:
Packer et al. conclude that carvedilol, added to the conventional therapy given to patients with severe chronic heart failure, reduced the risk of death by 35 percent and reduced the combined risk of death or hospitalization by 24 percent. The mean dose of angiotensin-converting–enzyme (ACE) inhibitors received by patients in the placebo and carvedilol groups was not specified. Although this was a randomized, double-blind trial, we do not know whether the ACE-inhibitor dose was similar in the two groups; a difference in the mean dose might have led to bias with respect to the risk of hospitalization.
José Mario Sabio, M.D.
Juan Jiménez-Alonso, M.D.
Virgen de las Nieves University Hospital, 18012 Granada, Spain
jualso@hvn.sas. cica. es Author/Editor Response
The authors reply:
To the Editor: Both Dr. Negut and Drs. Sabio and Jiménez-Alonso inquire about the background therapy received by patients with severe heart failure in our study — particularly, the use of drugs that are known to prolong life in such patients.
Drs. Sabio and Jiménez-Alonso ask whether the treatment groups were similar with respect to the dose of ACE inhibitors. There were 11 different ACE inhibitors used by patients in the study at the time of randomization. To compare doses in the carvedilol and placebo groups, we classified patients according to whether they received no ACE inhibitor (12 percent of patients in the carvedilol group and 11 percent of those in the placebo group) or a low dose (29 percent and 28 percent, respectively), a medium dose (33 percent and 31 percent, respectively), or a high dose (27 percent and 30 percent, respectively), on the basis of the range of doses for each ACE inhibitor. Thus, the distribution of doses of ACE inhibitors was very similar in the two groups.
Dr. Negut asks whether differences in the frequency of use of spironolactone might explain the different results of the carvedilol and bucindolol trials. To answer this question, we subdivided the patients in our study according to the presence or absence of spironolactone therapy at the time of randomization. The estimates of the treatment effect (a 35 percent reduction in the risk of death and a 24 percent reduction in the risk of death or hospitalization for any reason) were identical for the subgroup of patients who received spironolactone and the subgroup of patients who did not receive spironolactone. Hence, the use of spironolactone does not explain the favorable effects of carvedilol in our study. Our findings complement those of RALES, which showed a benefit of spironolactone whether or not patients were taking a beta-blocker.1
Dr. Negut evaluates the mortality rates in the carvedilol and bucindolol trials by comparing the crude proportions of patients who died, but this comparison is inappropriate, since enrollment was staggered in the trials, and they had substantially different follow-up periods. A more appropriate comparison of the mortality rates in the placebo groups would be either the Kaplan–Meier rates at one year or the annualized mortality (hazard) rates. In the bucindolol trial, the annualized mortality rate in the placebo group was 17 percent. In the carvedilol trial, the annualized mortality rate in the placebo group (as noted in the discussion) was 19.7 percent. Hence, the mortality rate in the placebo group in the bucindolol trial was lower — not higher — than that in the carvedilol study.
1 ReferencesMilton Packer, M.D.
Columbia University College of Physicians and Surgeons, New York, NY 10032
mp65@columbia.edu Ellen B. Roecker, Ph.D.
University of Wisconsin, Madison, WI 537061
Pitt B ,Zannad F ,Remme WJ , et al. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med 1999;341:709-717
Full Text | Web of Science | Medline
To the Editor:
We would like to thank Dr. Negut for his comments concerning the possible effect of combined treatment with aldosterone and beta-blockade in patients with advanced heart failure. In RALES,1 patients with advanced heart failure (New York Heart Association class III or IV) were randomly assigned to receive spironolactone or placebo. The study demonstrated a 30 percent reduction in mortality with spironolactone as compared with placebo, which was statistically significant. In that study, subgroup analysis revealed approximately a 60 percent reduction in mortality among the patients taking a beta-blocker (in the spironolactone group as compared with the placebo group), whereas there was less than a 30 percent reduction in mortality among patients randomly assigned to receive spironolactone (as compared with placebo) but not taking a beta-blocker. We must keep in mind that this was a subgroup analysis in a study that was not designed to examine a treatment-effect interaction of beta-blockade with spironolactone. Since the trial was not stratified according to beta-blocker use, the base-line comparability of the spironolactone and placebo groups cannot be ensured for subgroups based on beta-blocker use. In addition, since only 10 percent of patients in the study were taking a beta-blocker, the confidence intervals for the beta-blocker group were wide.
In the BEST study, we found no statistical interaction between spironolactone and bucindolol with respect to the outcome, but there were so few patients taking spironolactone (4 percent of the whole sample) that the test for an interaction had very low power. Spironolactone was used by a larger proportion of patients in the carvedilol study, which has more information on this interaction. It should be kept in mind that since treatment with spironolactone was not randomly assigned in either beta-blocker study, the results may not be interpretable in terms of treatment effects.
1 ReferencesEric J. Eichhorn, M.D.
University of Texas Southwestern Medical Center, Dallas, TX 75216
eric.eichhorn@utsouthwestern. edu Heidi Krause-Steinrauf, M.S.
National Heart, Lung, and Blood Institute, Bethesda, MD 20817Philip W. Lavori, Ph.D.
Veterans Affairs Medical Center, Menlo Park, CA 940251
Pitt B ,Zannad F ,Remme WJ , et al. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med 1999;341:709-717
Full Text | Web of Science | Medline
