Exercise and Insulin‐Resistance Biomarkers in Adults With Overweight/Obesity: A Network Meta‐Analysis of Randomised Controlled Trials

OBJECTIVE: To compare associations between different exercise training modalities-aerobic training (AT), resistance training (RT), combined training (CT) and high-intensity interval training (HIIT)-and changes in insulin-resistance biomarkers in adults with overweight/obesity and to evaluate the dose-response relationship for optimal exercise dose. METHODS: We systematically searched PubMed, Embase, Cochrane Central Register of Controlled Trials and Web of Science from inception to May 2026. Two reviewers independently selected randomised controlled trials (RCTs) comparing exercise interventions in overweight/obese adults. Risk of bias was assessed using the Cochrane Risk of Bias tool version 2.0 (RoB 2.0) and confidence in the network meta-analysis evidence was evaluated using the CINeMA framework. Data were synthesised using frequentist network meta-analysis, with consistency assessed via the node-splitting method and dose-response meta-regression using restricted cubic spline models was performed to examine the relationship between exercise volume and biomarker changes. RESULTS: A total of 86 RCTs involving 4358 adults (68% women) were included. The RoB 2.0 assessment revealed that the most common concerns were blinding of participants and personnel (performance bias) and missing outcome data. CINeMA ratings indicated that the certainty of evidence across comparisons ranged from low to moderate. Node-splitting analyses showed no significant inconsistency (all p > 0.05), supporting the consistency assumption of the network. Compared with usual care, all exercise modalities significantly reduced fasting insulin (FINS) and Homeostasis Model Assessment of Insulin Resistance (HOMA-IR), with HIIT demonstrating the greatest effects (FINS: MD = -2.24 μU/mL; HOMA-IR: MD = -0.67). HIIT, AT and RT also significantly lowered Fasting Blood Glucose (FBG). Notably, exercise plus caloric restriction showed no significant additional benefit over exercise alone for insulin-resistance markers, but this low-confidence indirect evidence (CINeMA) warrants caution. Dose-response analyses revealed nonlinear associations. The optimal exercise dose associated with the greatest improvement in fasting insulin was approximately 1265 MET-min/week, while for HOMA-IR, it was approximately 1091 MET-min/week. These findings translate to a practical recommendation of approximately 1000-1200 MET-minutes per week. No significant dose-response relationship was observed for FBG. In the probability ranking, HIIT ranked first for reducing FINS (SUCRA: 85.1%), FBG (SUCRA: 97.8%) and HOMA-IR (SUCRA: 80.0%). CONCLUSIONS: This systematic review and network meta-analysis indicate that HIIT, AT, RT and CT are all associated with improvements in insulin-resistance markers among adults with overweight/obesity, with an optimal weekly dose of approximately 1000-1200 MET-minutes. HIIT demonstrated the highest probability of being among the more effective modalities evaluated, while CT and RT represent reasonable alternatives. Given the low-to-moderate certainty of evidence (CINeMA), these findings should be interpreted cautiously and warrant confirmation in future head-to-head trials, particularly those comparing different exercise modalities or combined interventions and those assessing long-term adherence.