Network meta-analysis compares exercise types for glycemic control in type 2 diabetes

IntroductionType 2 diabetes mellitus (T2DM) is highly prevalent among middle-aged and older adults and is associated with adverse metabolic, cardiovascular, and functional outcomes. Although aerobic exercise, resistance training, and combined training are commonly recommended, comparative evidence regarding the efficacy and safety of different exercise intensities remains limited. Therefore, we conducted a systematic review and network meta-analysis to compare the effects of different aerobic and resistance exercise modalities and intensities in this population.MethodsPubMed, Embase, Cochrane Library, and Web of Science were systematically searched from inception to May 5, 2026, without language restrictions. A combined strategy of controlled vocabulary and free-text terms was applied. The primary outcomes were glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), peak oxygen uptake (VO₂peak), systolic blood pressure (SBP), and resting heart rate (HR). Risk of bias was assessed using the Cochrane Risk of Bias 2 tool, and the certainty of evidence was evaluated with the CINeMA framework. The protocol was registered in PROSPERO (CRD420251270452).ResultsTwenty-nine randomized controlled trials comprising 1,301 participants and 10 exercise interventions were included. For HbA1c, moderate certainty evidence indicated that high-intensity resistance training (HIRT; MD − 0.62, 95% CI − 0.93 to −0.30), low-certainty evidence indicated that moderate-intensity aerobic training (MIAT; MD − 0.58, 95% CI − 1.10 to −0.05), high-intensity aerobic training combined with moderate-intensity resistance training (HIAT–MIRT; MD − 0.54, 95% CI − 1.02 to −0.06), and low-intensity resistance training (LIRT; MD − 0.54, 95% CI − 1.00 to −0.09) significantly reduced HbA1c compared with usual care (UC), with HIRT ranking highest (SUCRA 78.6%). For FPG, moderate certainty evidence indicated that only moderate-intensity resistance training (MIRT; MD − 29.13 mg/dL, 95% CI − 57.68 to −0.58) showed a statistically significant reduction versus UC and ranked first (SUCRA 80.5%). For VO₂peak, high certainty evidence indicated that HIAT–HIRT (MD 3.75 mL/kg/min, 95% CI 1.11 to 6.38), HIAT (MD 3.14 mL/kg/min, 95% CI 1.33 to 4.95), and moderate certainty evidence indicated that HIAT–MIRT (MD 1.80 mL/kg/min, 95% CI 0.11 to 3.48) were associated with significant improvements compared with UC. For SBP, low certainty evidence indicated that HIRT demonstrated a significant reduction (MD − 3.68 mmHg, 95% CI − 7.20 to −0.16). No intervention significantly reduced resting HR relative to usual care. Global inconsistency testing did not indicate major network-wide incoherence; however, loop-specific analysis suggested potential inconsistency in the FPG network. Sensitivity analyses generally supported the stability of the main findings when network connectivity was preserved, although sparse networks, particularly for VO₂peak and resting HR, warrant cautious interpretation.ConclusionAmong middle-aged and older adults with T2DM, exercise interventions demonstrate outcome-specific effects across metabolic control, cardiorespiratory fitness, and cardiovascular parameters. Moderate-certainty evidence indicated that HIRT was associated with clinically meaningful reductions in HbA1c, whereas low-certainty evidence suggested similar associations for MIAT, HIAT–MIRT, and LIRT. Moderate certainty evidence indicated that MIRT showed may be associated with a potential advantage for improving FPG. For VO₂peak, HIAT–HIRT and HIAT were associated with the largest improvements compared with usual care, but these findings were based on a sparse evidence network and should be considered exploratory. However, because only a limited number of studies were included for this outcome and the evidence network was sparse, these results should be interpreted with caution. For blood pressure control, low-certainty evidence suggested that HIRT may be the only modality associated with a significant reduction in SBP. Although HIAT showed potential benefits for resting HR, the current evidence remains inconclusive. However, the certainty of evidence for most comparisons involving FPG and SBP was rated as low or very low; therefore, these findings should be interpreted cautiously. To bolster the credibility of the evidence and guide tailored exercise recommendations for this group, we require extensive, meticulously planned randomized studies featuring uniform outcome measures, extended follow-up periods, and direct comparisons among primary exercise types.Systematic review registrationhttps://www.crd.york.ac.uk/PROSPERO/view/CRD420261306289.