作者
Darragh O'Sullivan,Jennifer M. Rice,Jacob D. Meyer,Seana Smith,Brett R. Gordon,Cillian McDowell,Ian Sherwin,Mark Lyons,Matthew Payton Herring
摘要
Importance: Resistance exercise training (RET) has established antidepressant effects compared with inactive control conditions; however, its effects vs active comparisons remain unclear. Objective: To quantify the association of RET efficacy with active and inactive comparisons for depressive symptoms and explore variability across moderators. Data Sources: Published articles were identified using Google Scholar, MEDLINE, PubMed, Web of Science, and PsycInfo from inception to July 7, 2025. Study Selection: Randomized clinical trials (RCTs) with at least 3 weeks of RET or any non-RET comparison and a validated depressive symptom measure at baseline and at least 3 weeks were included. Active comparisons were aerobic exercise training, relaxation, flexibility, mind-body exercise, nutrition, education, advice or attention, mixed activities, and balance exercises. Inactive comparisons were no treatment, waiting lists, and continued usual care. Data Extraction and Synthesis: Three reviewers independently extracted data. Meta-regression was used to test moderators, including primary moderators (age, sex, mental health status, RET duration, intensity, concealed allocation). Study quality was assessed using a modified Detsky scale. Main Outcomes and Measures: All available associations of RET vs comparison conditions with depressive symptoms were quantified from baseline to at least 3 weeks. Random-effects models were used. Results: Across 80 RCTs including 6123 participants (2880 participants [47.0%] in RET groups and 3243 [53.0%] in non-RET comparison groups; mean [SD] age, 57 [16] years; 3412 female [67.1%]), the number of effect sizes examined was 176; 57 RCTs enrolled participants with an illness. Resistance exercise training was associated with a significant, heterogenous, small-magnitude antidepressant effect size (Δ = 0.32 [95% CI, 0.25-0.39]; z = 9.31; P < .001) across all comparisons. The effect size was larger vs inactive comparisons (Δ = 0.54 [95% CI, 0.39-0.68]; z = 7.22; P < .001) than active comparisons (Δ = 0.21 [95% CI, 0.14-0.27]; z = 6.20; P < .001). Versus active comparisons, RET was associated with a significant, heterogenous, small-magnitude antidepressant effect size, with larger effect sizes among participants with a mental illness (Δ = 0.40 [95% CI, 0.29-0.50]) vs without a mental illness (Δ = 0.11 [95% CI, 0.04-0.18]). Versus inactive comparisons, RET was associated with a significant, heterogenous, moderate-magnitude antidepressant effect size, with larger effect sizes for all-female cohorts (Δ = 0.87 [95% CI, 0.60-1.14]), RET frequencies of at least 4 days per week (Δ = 0.46 [95% CI, 0.31-0.60]), and session durations of 30 minutes or less (Δ = 1.75 [95% CI, 0.97-2.54]), and smaller effect sizes with concealed allocation (Δ = 0.32 [95% CI, 0.20-0.44]). Conclusions and Relevance: This meta-analysis and meta-regression analysis found that RET was associated with moderate-magnitude, clinically meaningful reductions in depressive symptoms that were greater than other activities with antidepressant benefits, supporting the recommendation of RET in the treatment of depression.