A systematic review and meta-analysis of placebo-controlled antidepressant studies in people with depression and dementia

Abstract

Objectives: To determine the efficacy of antidepressants in people with depression and dementia.

Design: A systematic review and meta-analysis based on a literature search of Medline and Cochrane Trials Registry for acute-phase, double-blind, placebo-controlled, parallel-design, random-assignment trials of antidepressants marketed in the United States.

Setting: Outpatient clinics, inpatient units, residential settings.

Participants: People with criterion-based diagnoses of dementia and depression.

Measurements: Numbers of participants randomized; baseline and end point depression scale scores; and response, remission, and discontinuation rates were extracted. Random-effects meta-analyses were performed for response and remission rates, change scores using standardized mean differences, and discontinuation rates. Sensitivity analyses were planned to examine effects of depression diagnosis, severity, and trial duration.

Results: Seven trials with 330 participants met selection criteria. The odds ratio (OR) for six trials reporting response rates with antidepressant and placebo was 2.12 (95% confidence interval (CI)=0.95-4.70; Z=1.84, P=.07). The OR for five trials reporting remission rates was 1.97 (95% CI=0.85-4.55; Z=1.59, P=.11). Both analyses demonstrated heterogeneity. The standardized mean difference in trials was 0.29 (95% CI=0.02-0.60, Z=1.86, P=.06). This analysis did not demonstrate significant heterogeneity. Adverse event discontinuation rates (9.0%) were not significantly higher with drug than placebo (6.0%), and were low.

Conclusion: The evidence for antidepressant treatment of people with depression and dementia, although suggestive, does not confirm efficacy. All of the trials were significantly underpowered to detect differences, resulting in inconclusive findings. Variable trial methods, comorbid conditions, and differences in antidepressants employed further confounded findings.