Three light lifting sessions a week lowered blood sugar in type 2 diabetes

Most people with type 2 diabetes are overweight or obese, a combination strongly associated with worse quality of life and a higher risk of long-term complications, and keeping blood glucose controlled remains the central clinical challenge. Resistance training is already recognised as an effective way to improve HbA1c, the standard marker of long-term blood sugar, but the practical question has never been settled: how often, how heavy, how long, and how much total work? Without established dose parameters, prescriptions vary widely, so this analysis set out to map the dose-response relationship between lifting variables and HbA1c reduction.

The authors ran a systematic review with a Bayesian network meta-analysis, a method that lets different training doses be compared against each other across trials. They searched PubMed, Embase and the Cochrane Library from database inception through 19 January 2026 for randomised controlled trials of resistance training in overweight or obese adults with type 2 diabetes. Ten studies with 633 participants qualified, with a mean age of 60.03 years, mean BMI of 30.69, and 32.7% men. Extracted data covered training frequency, intensity, session time, exercises, repetitions, sets, weekly volume and programme length.

Resistance training variables were effective at improving HbA1c across the board. The effective ranges were 3 sessions per week, intensities of 30 to 62% of one-rep max, programmes of 17 to 23 weeks, 45 minutes per session, 10 exercises, 13 to 18 repetitions, 2 to 4 sets, and weekly volumes of 1000 to 2100 repetitions. The suggested optimum sat at 3 sessions per week, 44% of one-rep max, 23 weeks, 45-minute sessions, 10 exercises, 18 repetitions, 3 sets, and around 1,500 repetitions per week, with the weekly volume figure carrying the largest estimated reduction at 0.51. The individual dose estimates came with wide uncertainty ranges.

The finding that resistance training improves HbA1c confirms what was already accepted; the contribution here is the dose detail, which had not been systematically established. The most striking element is the intensity: the effective and optimal loads sit at the light-to-moderate end, around 44% of one-rep max, well below what a typical strength programme prescribes. With only 10 trials and 633 participants behind the numbers, the optima are best read as estimates rather than precise targets, and whether heavier loading, longer programmes or different populations would shift them remains open.

For a reader managing type 2 diabetes or its risk, the useful information is that lifting appears to work for long-term blood sugar without requiring heavy weights: the doses tied to benefit here were light loads, higher repetitions and steady weekly volume over several months. That reframes resistance training as something accessible rather than intimidating for this purpose. The figures are pooled estimates with wide uncertainty, so they describe what worked on average in trials rather than a formula, and anyone with diabetes making training changes is in normal shared territory with their own clinician.