Design makes a difference: a meta-analysis of antidepressant response rates in placebo-controlled versus comparator trials in late-life depression

Abstract

Objective: Qualitative reviews of late-life antidepressant clinical trials suggest that antidepressant response rates in comparator trials are higher than antidepressant response rates in placebo-controlled trials. No quantitative review has been conducted to test this hypothesis.

Methods: A meta-analysis was conducted of all published articles in peer-reviewed journals from 1985 to the present to identify randomized clinical trials contrasting antidepressant pharmacotherapy to placebo or an active comparator in late-life depressed outpatients. Sixteen studies (nine comparator trials and seven placebo-controlled trials) were identified. Antidepressant response rates in both placebo-controlled and comparator trials were extracted and submitted for analysis using multilevel meta-analysis procedures.

Results: The authors found significant variability in antidepressant response rates beyond chance. This variability decreased by 27% when the authors included study type in the model. As expected, antidepressant response rates in comparator trials were significantly higher (60%) than antidepressant response rates in placebo-controlled trials (46%).

Conclusion: Antidepressant response rates are higher in comparator trials as compared to placebo-controlled trials. These findings have important implications for combined medication and psychotherapy trials that use placebo-controlled medication conditions because the response rates from these conditions are likely to be lower than those from unblinded conditions.