Structured effects of exercise on circulating brain-derived neurotrophic factor (BDNF) in postmenopausal women: a meta-analysis

Abstract

To systematically evaluate the effect of structured exercise interventions on circulating brain-derived neurotrophic factor (BDNF) levels among postmenopausal women, and to clarify the moderating effects of intervention modality, population characteristics, training duration, exercise intensity and energy-supply pattern on intervention outcomes, so as to provide evidence-based references for precise exercise interventions targeting brain health in postmenopausal women. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, electronic databases including PubMed, Web of Science, EBSCO and CNKI were searched for original studies regarding the effects of various exercise interventions on BDNF in postmenopausal women, from database inception to August 1, 2026. The Cochrane Risk-of-Bias Tool was applied for randomized studies. For quasi-randomized and non-randomized intervention studies, we explicitly acknowledge limitations of this tool and interpret methodological assessments cautiously; studies with insufficient randomization reporting were rated as unclear risk of bias. Meta-analysis was performed using Review Manager 5.4 software. Effect sizes were presented as standardized mean difference (SMD) with 95% confidence interval (95% CI). Subgroup analysis, sensitivity analysis and publication-bias evaluation were also conducted. Twelve studies involving 351 participants were finally included, with 173 subjects in the intervention group and 178 subjects in the control group. All included trials adopted chronic long-term structured exercise interventions without acute single-bout exercise protocols. Pooled meta-analysis demonstrated that structured exercise interventions significantly elevated BDNF levels in postmenopausal women compared with control conditions [SMD = 0.64, 95% CI (0.11, 1.17), P = 0.02], with substantial between-study heterogeneity (I² = 81%, P < 0.0001). Subgroup analyses (only subgroups with ≥3 studies included for quantitative synthesis) indicated that intervention modality, population characteristics, training duration and energy-supply pattern were potential main sources of heterogeneity. Statistically-significant maximum effect sizes were observed in the resistance training subgroup [SMD = 1.07, 95% CI (0.12, 2.01), P = 0.03], healthy older/postmenopausal-women subgroup [SMD = 0.30, 95% CI (-0.05, 0.65), P = 0.009], anaerobic-energy-supply-predominant subgroup [SMD = 1.07, 95% CI (0.12, 2.01), P = 0.03]. Some potential moderators could not be quantitatively evaluated because corresponding subgroups contained fewer than three primary studies, and those comparisons are only discussed as exploratory clues rather than robust conclusions. Sensitivity analysis revealed no substantial shift in pooled effect direction after sequential exclusion of individual studies. Funnel-plot inspection and Egger’s test detected no statistical evidence of publication asymmetry; nevertheless, these assessments were limited by the small number of included studies and cannot rule out publication bias definitively. Structured exercise interventions can significantly increase peripheral BDNF levels in postmenopausal women, which suggests beneficial impacts on neuroplasticity and brain health for this population. Chronic resistance-based exercise regimens dominated by anaerobic energy-supply exert superior regulatory effects among healthy postmenopausal women. https://www.crd.york.ac.uk/PROSPERO/view/CRD420261468069 , identifier CRD420261468069.