Every exercise type improves insulin resistance, but HIIT ranks top

Insulin resistance sits upstream of type 2 diabetes and much of the metabolic trouble that accompanies excess weight, so anything that reliably improves it matters for a very large group of people. Exercise is the standard prescription, but guidance rarely says which kind: steady aerobic work, resistance training, a combination, or high-intensity intervals. It is also unclear how much weekly exercise gives the best return, and whether adding calorie restriction buys anything extra. This analysis set out to rank the modalities against each other and to find the dose where the benefit peaks.

The researchers searched four major databases from inception to May 2026 for randomised controlled trials of exercise in adults with overweight or obesity, ending with 86 trials and 4358 participants, 68% of them women. They used network meta-analysis, a technique that lets treatments be compared even when they were never tested head to head in the same trial. Outcomes were fasting insulin, fasting blood glucose and HOMA-IR, a standard index of insulin resistance. They also ran dose-response models relating weekly exercise volume, measured in MET-minutes, to the size of improvement.

Compared with usual care, every modality significantly reduced fasting insulin and HOMA-IR, and HIIT, aerobic and resistance training also lowered fasting blood glucose. HIIT showed the largest effects, cutting fasting insulin by 2.24 μU/mL and HOMA-IR by 0.67, and it ranked first in the probability analysis for all three markers. The dose-response curves were nonlinear: improvement peaked at roughly 1265 MET-minutes per week for fasting insulin and 1091 for HOMA-IR, which the authors translate to a practical 1000-1200 MET-minutes per week. Notably, adding calorie restriction to exercise showed no significant extra benefit for these markers, though that rested on low-confidence indirect evidence, and no dose-response pattern emerged for fasting glucose.

That exercise improves insulin sensitivity is not news; the value here is the ranking across modalities and the attempt to put a number on the optimal weekly dose. The HIIT advantage echoes earlier work suggesting intensity matters for metabolic outcomes, but the authors are careful: evidence certainty ranged from low to moderate, and many trials had problems with blinding and missing data. The null result for adding calorie restriction complicates common advice and needs direct head-to-head trials before anyone leans on it. Long-term adherence, arguably the variable that decides real-world results, was not assessed.

If you carry extra weight and want to improve how your body handles glucose, this is evidence that the type of exercise matters less than doing some, since every modality moved the markers. It is also a reason to see intervals as a strong option if you tolerate them, with resistance or combined training as reasonable alternatives rather than inferior ones. The sweet-spot finding suggests more is not endlessly better for these particular markers. None of this replaces individual medical advice, and the authors themselves flag the certainty as low to moderate.