LAUSR

Objective To assess the comparative effectiveness of exercise, antidepressants and their combination for alleviating depressive symptoms in adults with non-severe depression. Design Systematic review and network meta-analysis. Data sources Embase, MEDLINE, PsycINFO, Cochrane Library, Web of Science, Scopus and SportDiscus. Eligibility criteria Randomised controlled trials (1990–present) that examined the effectiveness of an exercise, antidepressant or combination intervention against either treatment alone or a control/placebo condition in adults with non-severe depression. Study selection and analysis Risk of bias, indirectness and the overall confidence in the network were assessed by two independent investigators. A frequentist network meta-analysis was performed to examine postintervention differences in depressive symptom severity between groups. Intervention drop-out was assessed as a measure of treatment acceptability. Results Twenty-one randomised controlled trials (n=2551) with 25 comparisons were included in the network. There were no differences in treatment effectiveness among the three main interventions (exercise vs antidepressants: standardised mean differences, SMD, −0.12; 95% CI −0.33 to 0.10, combination versus exercise: SMD, 0.00; 95% CI −0.33 to 0.33, combination vs antidepressants: SMD, −0.12; 95% CI −0.40 to 0.16), although all treatments were more beneficial than controls. Exercise interventions had higher drop-out rates than antidepressant interventions (risk ratio 1.31; 95% CI 1.09 to 1.57). Heterogeneity in the network was moderate (τ2=0.03; I2=46%). Conclusions The results suggest no difference between exercise and pharmacological interventions in reducing depressive symptoms in adults with non-severe depression. These findings support the adoption of exercise as an alternative or adjuvant treatment for non-severe depression in adults. Systematic review registration PROSPERO CRD4202122656.