The Optimal Exercise Modality and Dose for Blood Pressure Management in Middle-Aged and Older Adults: A Systematic Review with Bayesian Model-Based, and Dose–Response Network Meta-Analysis of RCTs

Background Exercise is an established non-pharmacological strategy for lowering blood pressure (BP), but the comparative effectiveness and optimal dose of different exercise modalities in middle-aged and older adults remain unclear. Objective This study aimed to systematically compare the effects of different exercise modalities and doses on resting systolic and diastolic BP (SBP/DBP) in middle-aged and older adults (≥ 45 years). Methods Randomized controlled trials (RCTs) were identified from six electronic databases from inception to May 2025. Bayesian random-effects network meta-analysis was used to compare BP-lowering effects across exercise modalities. Model-based dose–response analyses were conducted to examine associations between exercise dose, expressed as metabolic equivalent of task minutes per week (METs-min/week), and SBP reduction. Network meta-regression was used to explore potential effect modifiers. Results A total of 159 RCTs ( n = 10,821) were included. All exercise modalities were associated with significant reductions in SBP and DBP compared with control. Circuit-based training (CBT) showed the largest estimated SBP reduction (mean difference [MD] − 13.52 mmHg; 95% credible interval [CrI] − 18.59 to − 8.44), followed by combined training (CT), high-intensity interval training, moderate-intensity continuous training, dynamic resistance training, and isometric exercise. CBT and CT generally remained among the highest-ranked modalities across sensitivity analyses, supporting the potential value of integrated aerobic–resistance training (IART). Higher baseline SBP was associated with greater SBP reductions ( β = − 2.49 mmHg per 10-mmHg increase; 95% CrI − 3.73 to − 1.24). Dose–response analyses suggested a nonlinear association between exercise dose and SBP reduction, with the largest model-estimated reduction occurring around 1200 METs-min/week, corresponding to the upper range of World Health Organization (WHO) physical activity recommendations. CBT showed a model-estimated peak effect at approximately 580 METs-min/week, close to the lower range of WHO recommendations. Conclusions All exercise modalities were associated with significant BP reductions in middle-aged and older adults, although effect magnitude, certainty, and robustness varied. IART, including CBT and CT, appeared to be a promising approach for BP management. CBT showed the largest estimated effect and may offer a time-efficient option, but its apparent superiority over CT should be interpreted cautiously given the limited CBT-specific evidence, low-to-very-low certainty of evidence, and sensitivity analysis findings. Given that no network comparison was rated as high confidence and many included studies had some concerns or high risk of bias, comparative rankings should be interpreted cautiously and should not be considered definitive evidence of superiority.