Article Text
Statistics from Altmetric.com
QUESTION: In patients who have had myocardial infarction, heart failure, or hypertension, do β blockers increase depressive symptoms, fatigue, and sexual dysfunction?
Data sources
Studies were identified by searching Medline (1966–2001) with the keywords myocardial infarction, heart failure, or hypertension in combination with the keywords adrenergic β antagonists and trial; and by scanning reference lists.
Study selection
English language studies were selected if they were randomised controlled trials with a placebo comparison, were not crossover trials, enrolled ≥100 patients, and had ≥6 months of follow up.
Data extraction
Data were extracted on number of patients; presence of heart failure, hypertension, or myocardial infarction; length of follow up; type of β blocker; and patient reported adverse events (ie, depressive symptoms, fatigue, and sexual dysfunction).
Main results
15 studies (42 409 patients) were included. Follow up ranged from 6–59 months. β blockers led to an increase in fatigue (10 studies, 17 682 patients); no difference existed between β blockers and placebo for depressive symptoms (7 studies, 10 662 patients) or sexual dysfunction (6 studies, 14 897 patients) (table 1). Withdrawals because of fatigue (10 studies, 29 454 patients) and sexual dysfunction (4 studies, 11 260 patients) were higher in the β blocker group than in the placebo group; withdrawals for depressive symptoms did not differ between groups (4 studies, 5803 patients) (table 1).
β blockers v placebo in myocardial infarction at ≥6 months*
Conclusions
In patients who have had myocardial infarction, hypertension, or heart failure, β blockers increase fatigue and withdrawals because of fatigue or sexual dysfunction. β blockers do not increase depressive symptoms.
Commentary
β noradrenergic antagonists have broad utility, but they are considered to have troublesome side effects. Ko et al found that β blockers were associated with increased fatigue and increased withdrawal because of fatigue or sexual dysfunction. Yet, closer examination sheds doubt on whether the evidence supports differential effects across the 3 problems and raises questions about the interpretation of side effects in placebo controlled trials.
Both placebo and active drug effects ranged widely across studies and were strongly correlated (table 2). The range across studies dwarfed the relatively small apparent differences between placebo and active drug. The significant difference between placebo and active drug for withdrawal because of fatigue or sexual dysfunction may reflect a true adverse effect of β blockers, although with sexual dysfunction the difference may not be meaningful because almost all withdrawals were in the same study.
Adverse effects of β blockers
The heterogeneity across trials, with close tracking of placebo and active drug effects, is consistent with a nocebo effect, whereby negative expectations can result in unfavorable outcomes.1 Side effects of placebo have been documented to resemble those of the reference drug.2 Patients in randomised clinical trials receive detailed information about potential side effects of the reference drug. This contributes to similar “side effect” rates for active drug and placebo but would not account for the wide range of event rates. Such a range could result from varying sources, including patient characteristics, actual β blockers used, or study design.
In summary, a prominent apparent nocebo effect probably biases toward underdetection of side effects in placebo controlled trials. Furthermore, the studies were so heterogeneous in event rates that it is difficult to interpret their results in combination. Even with these biasing factors, differences in the incidence of fatigue and sexual dysfunction emerge.
Supplementary materials
. Additional Comments
Reply to Dr Swann (commentator) by the original authors, Ko et al.
Response from Dr Swann to the reply
Files in this Data Supplement:
Footnotes
-
Source of funding: not stated.
-
For correspondence: Dr H M Krumholz, Yale University School of Medicine, New Haven, CT, USA. harlan.krumholz{at}yale.edu
-
Abstract and commentary also published in ACP Journal Club