Summary: In a small 12-week trial in prediabetes using detailed muscle MRI, aerobic exercise (alone or with metformin) improved skeletal muscle composition and achieved normoglycaemia in 80 to 90% of participants, whereas metformin alone had little muscle effect; the combination uniquely reduced visceral fat and increased muscle size.
PICO Summary
| Element | Detail |
|---|---|
| Population | 42 prediabetic adults (mean age 48) in four small groups (control 10, metformin 10, exercise 11, combined 11); 12-week randomised controlled trial with multiparameter MRI, China. |
| Intervention | Aerobic exercise alone, or aerobic exercise combined with metformin. |
| Comparison | Metformin alone, or no-intervention control. |
| Outcome | Normoglycaemia was reached by 90% of the combined group, 80% of exercise, 20% of metformin, and 10% of controls. Exercise reduced intermuscular adipose tissue, T2, and apparent diffusion coefficient and increased fractional anisotropy, while metformin alone had no significant MRI effect. The combination did not further improve blood biomarkers versus exercise but uniquely reduced visceral adipose tissue and increased muscle cross-sectional areas. MRI parameters correlated with glucose and insulin measures. |
Exercise vs Metformin in Prediabetes
RCT · prediabetes · 12 weeks
Aerobic exercise restored normoglycaemia in 80% of prediabetic adults versus 20% with metformin alone over 12 weeks. The combination added visceral-fat reduction and muscle-size gains.
Expert Commentary
This is a mechanistically rich small trial whose imaging adds texture to a familiar message, that exercise drives metabolic improvement in prediabetes. The striking gap in normoglycaemia, 80 to 90% with exercise-containing arms versus 20% with metformin and 10% with control, echoes the Diabetes Prevention Program and is a useful counter to any assumption that the drug substitutes for activity, and the MRI findings give a structural correlate, with exercise reducing muscle fat and altering diffusion measures consistent with healthier tissue. The nuance worth keeping is that metformin was not useless but worked differently, adding visceral-fat reduction and muscle-size gains in combination rather than changing muscle quality itself. The limitations are real and temper enthusiasm about the dramatic percentages: only 10 to 11 participants per arm, a brief 12 weeks that cannot speak to durable diabetes prevention, and undetailed exercise and metformin dosing. Can I use this with my patients? Yes, to reinforce priorities. It strengthens the case for making aerobic exercise the cornerstone of prediabetes care, with metformin as a complement rather than a replacement, while I would present the high remission rates as encouraging short-term findings from a small study rather than guaranteed outcomes.
References
Yu F, Dong Y, Ye N, et al. Multiparameter MRI assessment of metformin and exercise effects on skeletal muscle in prediabetes: a randomized controlled trial. Eur Radiol Exp. 2025;9(1):123. doi:10.1186/s41747-025-00658-y