The Optimal Type and Dose of Exercise for Elevating Brain-Derived Neurotrophic Factor Levels in Patients With Depression: A Systematic Review With Pairwise, Network, and Dose-Response Meta-Analyses

Abstract

Background: Reduced brain-derived neurotrophic factor (BDNF) levels have been linked to increased depression risk. While physical exercise is known to alleviate depressive symptoms and elevate BDNF levels, the effects of different exercise modalities and doses, along with their dose-response relationships, remain unclear. Objective: This study aims to systematically evaluate the effects of various exercise types and doses on BDNF levels in patients with depression through pairwise meta-analysis, network meta-analysis (NMA), and dose-response NMA and to provide personalized exercise prescription recommendations. Methods: A comprehensive search identified randomized controlled trials (RCTs) examining exercise's impact on BDNF levels in depression. Pairwise and NMA compared six exercise modalities: continuous aerobic exercise (CAE), resistance exercise (RE), combined aerobic and resistance exercise (AERE), yoga, Qigong, and mindfulness. Dose-response NMA was used to assess the relationships between exercise dose and BDNF levels. Results: Thirty-six RCTs with 2515 participants were included. The pairwise meta-analysis indicated that all exercise interventions significantly elevated BDNF levels in patients with depression, with AERE, RE, and yoga demonstrating the most substantial effects. NMA rankings suggested that AERE was the most effective intervention, followed by RE, yoga, Qigong, mindfulness, and CAE. Dose-response NMA revealed a positive nonlinear dose-response relationship between total exercise volume and BDNF levels, with an optimal effective dose identified at ~610 METs-min/week. Beyond 1000 metabolic equivalent of tasks (METs)-min/week, increases in BDNF levels appeared to plateau. Moreover, each exercise type had distinct dose-response patterns, with RE and AERE having relatively higher optimal effective dose ranges, while CAE, yoga, Qigong, and mindfulness exhibited lower optimal ranges. Conclusions: AERE, RE, and yoga are effective interventions for enhancing BDNF levels in patients with depression, with Qigong, mindfulness, and CAE being comparatively less effective. A positive nonlinear dose-response relationship between exercise volume and BDNF levels was observed. Further research is needed to refine dose-response relationships in this population.

Keywords: BDNF; depression; dose–response network meta-analysis; exercise.

Figure 1

Flow diagram of the study selection.

Figure 2

Risk assessment plots: (A) risk of bias and (B) funnel plot for publication bias. Note: Green indicates low risk, red indicates high risk, and blank indicates unclear risk.

Figure 3

Forest plot and network ranking table of the effects of different exercise interventions on BDNF levels: (A) forest plot and (B) network ranking table. Note: The results of the network meta-analysis are shown in the lower left of (3B), while the results of the pairwise meta-analyses are presented in the upper right. AERE, combined aerobic and resistance exercise; BDNF, brain-derived neurotrophic factor; CAE, continuous aerobic exercise; CG, control group; RE, resistance exercise.

Figure 4

Network meta-analysis: league table and SUCRA ranking panel. (A) League table presenting the direct and indirect comparisons among different exercise interventions. (B) Bayesian ranking panel illustrating SUCRA values for each exercise intervention. (4A) displays the network of comparisons, where node sizes represent sample sizes and the thickness of the edges denotes the number of studies comparing each pair of interventions. (4B) shows the SUCRA rankings, with higher values indicating superior therapeutic effects. AERE, combined aerobic and resistance exercise; CAE, continuous aerobic exercise; CG, control group; RE, resistance exercise; SUCRA, surface under the cumulative ranking curve.

Figure 5

Dose–response relationship between total exercise volume and BDNF levels. The x-axis represents the total exercise dose, ranging from 0 to 1500 METs-min/week, while the y-axis shows the predicted BDNF response value at each dose level. Higher response values indicate a more pronounced effect. The solid black line illustrates the predicted response curve across different doses, depicting the trend of BDNF response as exercise volume increases. The gray shaded area represents the 95% confidence interval for the predicted response. BDNF, brain-derived neurotrophic factor; MD, mean difference; METs, metabolic equivalent of tasks.

Figure 6

Depicts the dose–response NMA for different exercise interventions on BDNF levels. Note: The horizontal axis represents the dose levels of exercise interventions, while the vertical axis illustrates the predicted response of BDNF at these doses. Higher response values indicate stronger therapeutic effects. The solid black line represents the predicted response curve across various exercise doses. Dashed curves denote the 95% confidence intervals, reflecting the uncertainty in the predicted responses. Vertical curve lines show the predicted changes in the BDNF levels at specific exercise doses. The gray shaded area highlights the variability in dose effects within the key dose range. AERE, aerobic and resistance exercise; BDNF, brain-derived neurotrophic factor; CAE, continuous aerobic exercise; MBNMA, Bayesian model-based network meta-analysis; MD, mean difference; METs, metabolic equivalent of tasks; Mind, mindfulness; RE, resistance exercise.