Heavy lifting topped the table for HbA1c

Type 2 diabetes is common in middle-aged and older adults and carries metabolic, cardiovascular and functional consequences. Aerobic exercise, resistance training and combined programmes are all recommended, but guidelines rarely specify intensity with confidence, because head-to-head trials are scarce. That leaves an obvious practical gap: if someone can only sustain a limited amount of training, which mode and which intensity buys the most glycaemic control, and does the answer change if the goal is fitness or blood pressure instead? A network meta-analysis can combine direct and indirect comparisons to attempt that ranking.

The authors searched PubMed, Embase, the Cochrane Library and Web of Science from inception to 5 May 2026 without language restrictions, using controlled vocabulary and free-text terms. They included randomised controlled trials of aerobic and resistance exercise at differing intensities in middle-aged and older adults with type 2 diabetes. Primary outcomes were HbA1c, fasting plasma glucose, peak oxygen uptake, systolic blood pressure and resting heart rate. Risk of bias was assessed with the Cochrane Risk of Bias 2 tool and certainty with the CINeMA framework. Twenty-nine trials comprising 1,301 participants and 10 exercise interventions were analysed, with the protocol registered in advance.

For HbA1c, moderate-certainty evidence favoured high-intensity resistance training (MD -0.62, 95% CI -0.93 to -0.30), which ranked highest with a SUCRA of 78.6%; low-certainty evidence supported moderate-intensity aerobic training (-0.58), high-intensity aerobic combined with moderate-intensity resistance training (-0.54) and low-intensity resistance training (-0.54). For fasting plasma glucose, only moderate-intensity resistance training reached significance (-29.13 mg/dL) and ranked first (SUCRA 80.5%). For peak oxygen uptake, high-certainty evidence favoured combined high-intensity aerobic and resistance training (3.75 mL/kg/min) and high-intensity aerobic training alone (3.14 mL/kg/min). High-intensity resistance training was the only modality significantly reducing systolic pressure (-3.68 mmHg), and no intervention lowered resting heart rate.

The results push back on the assumption that aerobic training is automatically the glycaemic intervention of choice, since resistance work at higher intensity ranked first for HbA1c here. They also show that no single modality wins across outcomes: fitness, fasting glucose and blood pressure each had a different leader. Confidence varies sharply by outcome, with the fasting glucose and blood pressure comparisons rated low or very low certainty and loop-specific inconsistency flagged in the fasting glucose network. The peak oxygen uptake network was sparse, so those rankings are exploratory, and the authors call for larger trials with uniform outcomes and direct mode comparisons.

This is a reason to treat resistance training as a serious glycaemic tool in later life rather than an add-on to cardio, particularly when it is performed at higher intensity. It also suggests that the best programme depends on which outcome you care about, since fitness gains clustered around high-intensity aerobic work while blood glucose responded to resistance work. The certainty behind most comparisons is limited, so the rankings should not be read as a definitive league table. Anyone managing type 2 diabetes is looking at decisions to make with a clinician, not from pooled averages.