LAUSR

OBJECTIVE To perform a systematic review and meta-analysis to clarify the effect of programmed exercise on depressive symptoms (DSs) in midlife and older women. METHODS We carried out a structured search of PubMed-Medline, Web of Science, Scopus, Embase, Cochrane Library and Scielo, from database inception through June 29, 2017, without language restriction. The search included the following terms: "depression", "depressive symptoms", "exercise", "physical activity", "menopause", and "randomized controlled trial" (RCTs) in midlife and older women. The US, UK and Australian Clinical Trials databases were also searched. We assessed randomized controlled trials (RCTs) that compared the effect of exercise for at least 6 weeks versus no intervention on DSs as the outcome (as defined by trial authors). Exercise was classified according to duration as "mid-term exercise intervention" (MTEI; lasting for 12 weeks to 4 months), and "long-term exercise intervention" (LTEI; lasting for 6-12 months). Mean changes (±standard deviations) in DSs, as assessed with different questionnaires, were extracted to calculate Hedges' g and then used as the effect size for meta-analysis. Standardized mean differences (SMDs) of DSs after intervention were pooled using a random-effects model. RESULTS Eleven publications were included for analysis related to 1943 midlife and older women (age range 44-55 years minimum to 65.5±4.0 maximum), none of whom was using a hormone therapy. Seven MTEIs were associated with a significant reduction in DSs (SMD=-0.44; 95% CI -0.69, -0.18; p=0.0008) compared with controls. The reduction in DSs was also significant in six LTEIs (SMD=- 0.29; 95% CI -0.49; -0.09; p=0.005). Heterogeneity of effects among studies was moderate to high. Less perceived stress and insomnia (after exercise) were also found as secondary outcomes. CONCLUSION Exercise of low to moderate intensity reduces depressive symptoms in midlife and older women.