Across 41 randomised trials in 2,264 adults, exercise produced large reductions in depressive symptoms compared with non-active controls, with a number needed to treat of 2. The effect remained moderate when analysis was restricted to the 12 low risk-of-bias trials.
What it shows: Meta-analysis of 41 RCTs (2,264 participants post-intervention); many trials were small with high heterogeneity in methods, and the effect shrank to moderate in the 12 low risk-of-bias studies.Meta-analysis · Br J Sports Med