Efficacy of exercise-based interventions for prediabetes: an umbrella review of meta-analyses of randomised controlled trials.

Background: Exercise is a key lifestyle strategy for managing prediabetes, yet varied exercise modalities and outcomes across studies have led to inconsistent findings, hindering clinical guidance. This umbrella review synthesises up-to-date evidence on the efficacy of exercise-based interventions in improving glycaemic control and cardiometabolic health in individuals with prediabetes. Methods: For this umbrella review of meta-analyses of randomised controlled trials, we searched seven databases (PubMed, EMBASE, Cochrane, CINAHL, Scopus, SPORTDiscus, and Web of Science) between database inception and June 01, 2025, for systematic reviews with meta-analyses comparing various exercise modalities (aerobic, resistance, high-intensity interval training [HIIT], mind-body, combined) to non-exercise. Outcomes included glycaemic control (glycosylated haemoglobin [HbA1c], fasting blood glucose [FBG], oral glucose tolerance test [OGTT], fasting blood insulin [FBI], and Homeostatic Model Assessment for Insulin Resistance [HOMA-IR] and other clinically relevant cardiometabolic parameters (lipid profile, blood pressure, and weight status). Methodological quality was assessed using AMSTAR-2, and risk of bias in primary studies was evaluated with RoB2. Publication bias was examined using Egger's test and funnel plots.; certainty of evidence via the GRADE framework; and heterogeneity via the I2 statistic. This umbrella review was registered with PROSPERO, CRD420251070016. Findings: From 3064 records, 12 systematic reviews with meta-analyses, representing 162 randomised controlled trials and 18,583 participants were included. Participants were predominantly middle-aged adults (59.0% female) with diverse geographical backgrounds. Overall, exercise-based interventions significantly improved all glycaemic control parameters (all p < 0.001): HbA1c (weighted mean difference: -0.25%, 95% confidence interval -0.33 to -0.17; I2 = 92%), FBG (-0.33 mmol/L, -0.42 to -0.24; I2 = 92%), OGTT (-0.94 mmol/L, -1.38 to -0.50; I2 = 94%), FBI (-0.84 μU/mL, -1.20 to -0.47; I2 = 89%), and HOMA-IR (-0.47 arbitrary unit, -0.68 to -0.26; I2 = 95%). Secondary outcomes of cardiometabolic health also showed consistent improvements (all p < 0.001). Subgroup analyses indicated that aerobic exercise improved all glycaemic control and cardiometabolic health parameters except FBI, while resistance exercise, combined training, HIIT, and mind-body exercise enhanced specific glycaemic control and cardiometabolic risk parameters. Interpretation: This novel umbrella review with large-scale meta-analyses synthesises robust evidence that exercise-based interventions, across various modalities, significantly improve glycaemic control and cardiometabolic health. The findings underscore the versatility of exercise, supporting tailored, patient-centred prescriptions for managing prediabetes. Future research should prioritise high-quality RCTs focusing on patient-relevant outcomes, such as adherence and safety, while exploring the interactions among various moderators. Funding: The Health and Medical Research Fund (HMRF) Research Fellowship Scheme, funded by the Health Bureau of the Government of the Hong Kong Special Administrative Region.