Abstract
This study assessed the optimal exercise modality and dose for HbA1c reduction in individuals with prediabetes, taking baseline BMI into account. We searched six databases and conducted pairwise, network, and dose-response meta-analyses. Thirty-nine RCTs (3421 participants) showed that all exercise modalities significantly reduced HbA1c compared with controls, with HIIT showing the greatest effect, followed by combined, aerobic, resistance, and mind-body training. Dose-response analysis identified an optimal dose of 850 METs·min/week, whereas the minimum effective dose varied according to BMI and exercise type. These findings indicate that exercise improves HbA1c in prediabetes through a nonlinear dose-response and support BMI-stratified, individualized exercise prescriptions to optimize glycemic control.