Do Randomized Controlled Trials Always Trump Case Reports? A Second Look at Propranolol and Depression.

Study Objective. To explore reasons for discrepancies between findings from case reports and those from a meta-analysis of randomized controlled trials regarding the association between β-adrenergic blockers and depression. Design. Systematic review. Data Source. PubMed/MEDLINE database. Measurements and Main Results. We reviewed 24 published case reports showing an association between β-blockers and depression and eight randomized controlled trials included in a recta-analysis of the adverse effects of these drugs. We abstracted the β-blocker taken, patients' age and sex, diagnoses, history of depression, type of depressive symptoms reported, and method and timing of the assessment of depression. Naranjo criteria were used to evaluate the strength of evidence from each case report for a possible association between β-blockers and depression. Twelve case reports had a Naranjo score of 5 or more (suggesting a likely causal relationship), nine of which involved propranolol. In all nine, depression began soon after treatment, and in four, the patient had a history of depression. Three randomized controlled trials assessed propranolol. Depression rates in the control groups of these studies differed substantially from each other (0-40%, p<0.0001). In only one randomized controlled trial did investigators assess depression systematically; they evaluated depression after 1 year of treatment and eliminated patients who had previously been prescribed an antidepressant. Conclusion. A criterion standard to assess the true relationship between β-blockers and depression is lacking. Factors such as the lack of systematic assessment of depression, the timing of assessments, and the selection of patients may have reduced the ability of researchers in the randomized controlled trials to detect depression as an adverse effect. Evidence from case reports should be carefully considered when relevant randomized controlled trials have not been adequately designed to detect adverse effects. [ABSTRACT FROM AUTHOR]