HbA1c in type 2 diabetes: a small fall of uncertain sizeMixed evidence
Several reviews agree that yoga lowers HbA1c, but not on how much. In the newest (28 trials, 2,241 adults, 22 trials measuring HbA1c), yoga lowered it by 0.64 percentage points against usual care or a waiting list. Earlier reviews found 0.47 points in 12 trials and a small to moderate effect in 23 studies.
No trial was at low risk of bias, 20 of the 28 were from India, and the reviewers rated the certainty low. A check for missing small studies shrank the effect to 0.32 points, with a range that included no effect, so the true benefit may be smaller than the headline figure.
Sources 1, 2, 3
Fasting blood sugarMixed evidence
Fasting blood sugar was 1.36 mmol/L lower with yoga in 24 trials, again at low certainty, with signs that small studies with positive results were more likely to be published.
Source 1
More benefit when blood sugar starts higherMixed evidence
Across trials, the fall was bigger where HbA1c started higher: the predicted effect went from no benefit at an HbA1c of 7% to a fall of about half a point at 10%. This compares trials rather than people, so it's a pointer, not proof.
Source 1
How long and how oftenMixed evidence
In 22 Asian trials, programmes of 12 weeks or more lowered HbA1c and shorter ones didn't reach significance. In the newest review, how often people practised, for how long, and whether the classes were supervised made no clear difference, but there were too few trials to tell.
Sources 1, 18
Yoga compared with other exerciseEmerging evidence
In a comparison of 85 trials (5,535 people), HbA1c fell by 0.78 points with interval training, 0.54 with aerobic plus strength training, 0.45 with yoga and 0.42 with steady aerobic exercise, with no significant difference between them, on low-quality evidence. Smaller reviews favoured yoga a little over other exercise (by 0.39 points in 8 trials) and over walking (by 0.20 points in 16 trials), and a small network found yoga and interval training alike.
Sources 4, 19, 20, 21
Weight, cholesterol and blood pressureMixed evidence
Pooled trials found lower total and LDL ("bad") cholesterol, higher HDL ("good") cholesterol, and better blood pressure, BMI (body mass index) and waist-to-hip ratio with yoga, but the changes were small and the trials varied a lot. In Asian trials, weight change was inconsistent. Our yoga for weight loss and yoga for high blood pressure guides cover these.
Sources 2, 3, 18, 22
Strength and balanceMixed evidence
In a small network of trials in type 2 diabetes, yoga improved muscle strength, at very low certainty. The ADA says activities including yoga, tai chi and resistance training can improve flexibility, muscle strength, balance and walking in older adults with diabetes.
Sources 23, 10
Numb feet from diabetes (peripheral neuropathy)Mixed evidence
In 7 small studies (227 people, 4 of them randomised trials), 8 to 12 weeks of yoga went with less pain and better balance, leg strength, walking speed and quality of life, with no serious side effects. The studies were at high risk of bias.
Source 24
Type 1 diabetes: little researchMixed evidence
Of 92 teenagers randomised, 46 finished; in them, 3 months of added yoga lowered HbA1c and fear of hypos. In 15 adults, a 50-minute yoga session and cycling both lowered blood sugar, with 1 hypo after yoga against 7 after cycling over the next 12 hours. In 62 teenage girls, 8 weekly online sessions eased depression scores.
A 2026 review called these approaches promising and asked for bigger trials. Breakthrough T1D says flexibility exercise such as yoga, tai chi and Pilates "will have a minimal effect on your blood glucose levels".
Sources 25, 26, 27, 28, 29
Prediabetes: one large trialEmerging evidence
In 974 Indian adults with prediabetes, adding a structured 40-minute routine of poses and yoga breathing to standard lifestyle advice lowered the number who developed diabetes over 3 years: 11.5% in the yoga group against 18.9% with advice alone. Most kept it up (77% did the yoga more than three-quarters of the time).
It's one open-label trial (everyone knew which group they were in) in one country, and we read its abstract only, which doesn't say how many times a week the routine was meant to be done.
Source 14
Prediabetes: smaller and shorter studiesMixed evidence
A trial that randomised 80 groups of people in India reported far fewer new cases of diabetes at 3 months, but a quarter of the people weren't followed up. Smaller trials found lower blood sugar after meals, LDL cholesterol and waist size (184 people), or more weight loss than walking but no difference in blood sugar (41 people), and a feasibility trial (65 people) found yoga acceptable and safe. A 2019 review concluded that yoga helped, but most of its pooled results could have been chance.
Sources 30, 31, 32, 33, 34
Diabetes in pregnancyMixed evidence
Exercise in pregnancy roughly halved the odds of gestational diabetes in a 2025 analysis, and yoga was among the types that helped, though from very few trials. One small Thai trial combined mindful eating and yoga in women who already had gestational diabetes. Our exercise with gestational diabetes and yoga in pregnancy guides cover what to do.
Sources 35, 36
Where the main evidence liesStrong evidence
Medicines, eating well and regular activity carry the strongest evidence. The NHS asks for at least 2.5 hours a week of activity that leaves you out of breath, the ADA recommends at least 150 minutes of aerobic activity a week plus strength training 2 to 3 times a week, and Diabetes UK says exercise alone can't lower blood sugar enough for remission. Yoga is one way to be active within that plan.
Sources 16, 10, 13