Effects of exercise modalities and doses on brachial artery flow-mediated dilation in adults with overweight or obesity: a Bayesian dose-response network meta-analysis

Background: Endothelial dysfunction is an early vascular abnormality in adults with overweight or obesity. Although exercise can improve vascular function, the most favorable exercise modality and weekly dose for improving brachial artery flow-mediated dilation (brachial FMD) remain unclear. This study compared exercise modalities and weekly dose-response relationships for brachial FMD. Methods: We conducted a systematic review and Bayesian dose-response network meta-analysis of parallel-group randomized controlled trials. Eligible trials enrolled adults with overweight or obesity, compared exercise lasting at least 4 weeks with non-exercise control or another modality, and reported pre- and post-intervention brachial FMD. Exercise dose was expressed as MET-min/week. Effects were expressed as mean differences (MDs) in change in brachial FMD. Eight dose-response functions were compared using MBNMAdose in R, and certainty of evidence was assessed with CINeMA. Results: Forty-seven trials (2,211 participants) were included. The quadratic random-effects model provided the best fit (DIC = 322.6). Overall, the largest model-estimated improvement occurred at 1,500 MET-min/week (MD: 4.72 percentage points, 95% CrI: 3.43, 5.91). Among modalities, combined aerobic and resistance training (AE-RT) showed the largest estimated improvement at 1,500 MET-min/week (MD: 5.77, 95% CrI: 4.04, 7.52). AE-RT at 1,200-1,600 MET-min/week was most consistently ranked among the highest predicted responses, with 1,400 MET-min/week ranked first. Subgroup patterns were broadly consistent across baseline brachial FMD, BMI, health status, and duration. Sensitivity analysis excluding high-risk-of-bias trials remained consistent (1,500 MET-min/week; MD: 4.84, 95% CrI: 3.53, 6.05). Meta-regression did not identify statistically clear effect modification by age, baseline brachial FMD, intervention length, or baseline BMI. Across all 13 network comparisons, the certainty of evidence was low or very low. The mind-body exercise optimum lay below the observed range, an unreliable extrapolation. Conclusion: In adults with overweight or obesity, exercise showed dose-dependent improvement in brachial FMD, with AE-RT around 1,500 MET-min/week showing the most favorable model-estimated profile. These estimates should be interpreted as provisional, model-based prescription anchors rather than ready-to-apply clinical thresholds, and should be individualized according to clinical status, exercise tolerance, and safety considerations. The low-to-very-low certainty of evidence warrants cautious interpretation and confirmation in higher-quality head-to-head randomized trials using standardized brachial FMD protocols. Systematic Review Registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD420261402756, PROSPERO CRD420261402756.