Design: Systematic review and meta-analysis of randomised trials
Population: 11 RCTs, about 740 older adults with type 2 diabetes
Primary outcome: Lean mass, muscle strength and physical performance
Effect: Strength g 0.67 (0.46–0.88); performance g 0.74; HbA1c −0.32%

This meta-analysis pooled 11 randomised trials (about 740 participants, mean age 65 or over) comparing structured exercise of at least four weeks with no exercise in type 2 diabetes. Outcomes were lean mass, strength, physical performance and glycaemic control.

Exercise produced moderate-to-large gains in muscle strength (g 0.67, 95% CI 0.46–0.88) and physical performance (g 0.74, 0.44–1.03), and lowered HbA1c by 0.32 percentage points. HOMA-IR improved, fasting glucose barely changed, and lean mass showed only a small, non-significant gain (g 0.14). Adverse events were rare.

Older people with diabetes lose strength faster than their peers, and the usual focus on glucose leaves that loss untreated. This analysis supports prescribing resistance or multicomponent exercise as part of diabetes care: it improves function and glycaemia together, even when a scan would show little new muscle.

Prescribe resistance or multicomponent exercise as part of diabetes care in older adults
Measure grip strength or chair-stand time at baseline and follow-up
Review hypoglycaemia risk when starting exercise in those on insulin or sulfonylureas
Ensure adequate protein intake alongside training
Judge benefit by function, not by muscle mass alone

Why it matters

Moves exercise from general advice to a treatment with measurable effects on both function and glucose.

The statistics, in plain English

Effect sizes of 0.67 and 0.74 are moderate to large, meaning clear improvements in strength and performance. An HbA1c fall of 0.32 percentage points is modest but similar to adding a mild glucose-lowering drug. The lean mass confidence interval includes zero, so exercise may improve function through nerve and muscle quality rather than size.