Optimizing exercise modalities to enhance brain-derived neurotrophic factor levels in older adults: a Bayesian network and dose-response meta-analysis
Background Global aging leads to increased age-related cognitive decline and neurodegenerative diseases like dementia. Brain-derived neurotrophic factor (BDNF) is crucial for neuronal survival, synaptic plasticity, and cognitive function, but its levels often decline with age. Physical exercise can potentially elevate BDNF levels in older adults. Previous meta-analyses had limitations in comparing diverse exercise types simultaneously and exploring dose-response relationships. Objective To comprehensively compare the relative effects of exercise types on BDNF levels in older adults and explore dose-response relationships using a multilevel Bayesian network meta-analysis. Methods A systematic search of PubMed, Embase, Web of Science, and the Cochrane Library for randomized controlled trials (RCTs) involving exercise interventions in older adults was conducted until April 2025. Data on BDNF levels, exercise type, duration, frequency, and intensity were collected. A multilevel Bayesian network meta-analysis was performed to compare exercise interventions and explore dose-response relationships. Results Forty-seven RCTs with 1815 participants were included. Aquatic exercise showed the largest point estimate (SMD = 1.00), followed by mind-body exercise (SMD = 0.61) and mixed exercise (SMD = 0.30). However, the AQE estimate was based on sparse evidence and should be interpreted cautiously. Several exercise modalities exceeded a 0.5-SD distribution-based interpretability threshold. Aerobic, combined aerobic and resistance, and resistance exercises also improved BDNF levels, whereas HIIT remained uncertain in the primary analysis, although the active-comparator-exclusion sensitivity analysis yielded a positive estimate that requires confirmation. Dose-response analyses showed non-linear relationships, with effects observed across dosages and after intervention cessation. Advanced age was linked to diminished effects. Conclusion Several exercise modalities may increase BDNF levels in older adults, but AQE and HIIT should be interpreted as preliminary or hypothesis-generating. Complex dose-response relationships exist, and age and sex may moderate effectiveness. Treatment rankings should be interpreted cautiously in light of evidence certainty.