Yoga does not defeat diabetes, but it helps. Pooled trials show that 10 to 24 weeks of yoga, added to usual medicines, lowers HbA1c by about 0.5 percentage points and fasting glucose by a modest amount. The ten asanas and pranayama below come from the protocol used in India's largest yoga-diabetes trial; some of them are unsafe for people with eye disease, nerve damage or a tendency to low blood sugar.
Does yoga lower blood sugar in type 2 diabetes?
Yes, by a modest amount, in trials that lasted a few months. Chen and colleagues pooled 13 randomised controlled trials with 1,335 adults with type 2 diabetes in a 2022 meta-analysis in Frontiers in Sports and Active Living. Yoga lowered HbA1c by 0.47 percentage points more than control (95% CI −0.77 to −0.16), lowered fasting blood glucose (standardised mean difference −0.92) and lowered post-meal glucose (SMD −0.53). Programmes ran 10 to 24 weeks, most often 12 weeks.
Thind and colleagues reached a similar answer in a 2017 meta-analysis in Preventive Medicine covering 23 studies and 2,473 adults, 18 of them from India. Yoga improved HbA1c (effect size d+ = 0.36), fasting glucose (d+ = 0.58) and post-meal glucose (d+ = 0.40). The typical programme was 12 weeks of 60-minute sessions.
The caveats matter. Chen's review found heterogeneity of 82% for HbA1c and 96% for fasting glucose, judged risk of bias as mostly unclear or high, and noted that every trial gave yoga on top of drug treatment. Thind's review found studies met only 41% of quality criteria on average, and 87% of them followed people for 3 months or less. No included trial reported adverse events, so the reviews could not say whether harms occurred. The American Diabetes Association's 2016 position statement on physical activity calls the benefits of yoga "less established" than those of aerobic and resistance exercise.
What do Indian trials of yoga for diabetes show?
The largest Indian trial shows that a yoga-based lifestyle programme reduced short-term progression from prediabetes to diabetes. The Niyantrita Madhumeha Bharat (NMB) trial, funded by the Ministry of Health and Family Welfare and the Ministry of AYUSH and published in Frontiers in Endocrinology in 2021, randomised 80 village and census-block clusters with 4,450 adults aged 20 to 70 who had prediabetes (HbA1c 5.7 to 6.4%) and a high Indian Diabetes Risk Score. The yoga arm attended nine consecutive days of two-hour camps, then practised daily at home with a weekly supervised class for three months. At three months, 11.21% of the yoga group had developed diabetes against 32.01% of controls, a relative risk reduction of 63.81%.
The authors list the limits themselves: follow-up was only three months, 26% of the yoga group dropped out, participants could not be blinded, and each result rested on a single HbA1c reading.
A Chandigarh sub-study of the same programme, published in Diabetology and Metabolic Syndrome in 2021, randomised 184 people at high risk of diabetes to the Diabetic Yoga Protocol or a waiting list. After three months HbA1c did not differ between groups (−0.02 points, p = 0.077), while post-meal glucose, LDL cholesterol and waist circumference improved relative to controls.
The scale of the problem explains the government interest. The ICMR-INDIAB survey of 113,043 adults, published in The Lancet Diabetes and Endocrinology in 2023, put diabetes prevalence at 11.4% and prediabetes at 15.3% of Indian adults.
Which 10 yoga poses and breathing practices are used in diabetes trials?
The ten practices below are taken from the 60-minute Diabetic Yoga Protocol published with the Chandigarh NMB sub-study, the same protocol family used across the national trial. The protocol holds each asana for about one minute, runs Surya Namaskara for nine minutes, and ends with breathing practices and 15 minutes of cyclic meditation. The caution column applies the American Diabetes Association precautions explained in the next section.
| # | Practice | What it involves | Main caution |
|---|---|---|---|
| 1 | Surya Namaskara (sun salutation) | 12-step flowing sequence; 6 fast and 1 slow round | Vigorous and jarring; avoid with unstable retinopathy; the protocol has a chair version |
| 2 | Trikonasana (triangle) | Standing side-bend, legs wide | Balance; use a wall with neuropathy or vision loss |
| 3 | Bhujangasana (cobra) | Prone back-bend lifting the chest | No breath-holding |
| 4 | Dhanurasana (bow) | Prone, holding the ankles, chest and thighs lifted | Strain and breath-holding; avoid with unstable retinopathy |
| 5 | Pawanmuktasana (wind-relieving) | Supine, knees hugged to the chest | Gentle; used between poses |
| 6 | Ardha Matsyendrasana or Vakrasana | Seated spinal twist | Gentle; stop if dizzy |
| 7 | Paschimottanasana (forward bend) | Seated, folding over straight legs | Head-down; avoid with unstable retinopathy |
| 8 | Mandukasana (frog) | Kneeling, fists on the abdomen, folding forward | Abdominal pressure; not studied on its own |
| 9 | Viparita Karani (legs up) | Supine inversion, legs raised | Head-down; avoid with retinopathy or postural dizziness |
| 10 | Kapalabhati, Nadi Shuddhi, Bhramari | 1 min rapid exhalations, 6 min alternate-nostril breathing with a 2-second hold, 3 min humming breath | Forced breathing and breath-holds; ADA lists both as risks with unstable retinopathy |
No cited trial tested these poses one at a time; the effect sizes belong to the whole hour, meditation and diet advice included.
Who should avoid which poses?
People with unstable diabetic eye disease, nerve damage in the feet, or autonomic neuropathy need a modified programme, and some should not start without a doctor's clearance. The American Diabetes Association position statement sets out the rules.
- Retinopathy. With severe non-proliferative or unstable proliferative retinopathy, "vigorous aerobic or resistance exercise; jumping, jarring, head-down activities; and breath holding should be avoided". No exercise at all during a vitreous haemorrhage. That rules out inversions such as Viparita Karani, deep forward bends, strong back-bends and forced breathing until an ophthalmologist says the retina is stable.
- Peripheral neuropathy. Check the feet daily, keep them dry, and use well-fitting footwear or a mat; the ADA says moderate activity "is not likely to increase risk of foot ulcers", but weight-bearing activity "should be avoided with unhealed ulcers". Standing balance poses need a wall or chair if gait is altered. The NIDDK adds that "some activities that need balance may be unsafe for people with nerve damage or vision problems caused by diabetes".
- Autonomic neuropathy. Avoid rapid changes of posture, which can cause fainting, avoid practising in heat, and drink water. With cardiac autonomic neuropathy, get physician approval first.
- Everyone. The NIDDK advises talking to the care team before starting any new activity and asking whether insulin or carbohydrate needs adjusting.
See a doctor the same day if a session brings sudden floaters, flashes or a curtain over part of the vision, chest pain, or a foot wound that was not there before. Call 112 for chest pain with sweating or breathlessness.
How do you prevent low blood sugar during a yoga session?
Check glucose before the session, carry fast sugar, and know the thresholds. The NIDDK defines low blood glucose for most people with diabetes as a reading below 70 mg/dL, and says the risk is highest in people taking insulin or sulfonylureas. Activity beyond the usual routine can lower glucose for up to 24 hours afterwards, so a low can come at night after a morning class.
The ADA statement gives pre-exercise targets of 90 to 250 mg/dL (5.0 to 13.9 mmol/L):
- Below 90 mg/dL: take 15 to 30 g of fast-acting carbohydrate before starting.
- 90 to 150 mg/dL: start, and keep carbohydrate at hand for longer sessions.
- 150 to 250 mg/dL: start, and hold off on carbohydrate until glucose falls below 150.
- 250 to 350 mg/dL: test for ketones and do not exercise if they are moderate or large.
If symptoms of a low appear during practice, shakiness, sweating, hunger, a racing heart, blurred vision or confusion, stop. The NIDDK's treatment is 15 to 20 g of glucose or carbohydrate (four glucose tablets, half a cup of fruit juice, or a tablespoon of sugar), recheck after 15 minutes, and repeat until the reading is back in range. The NIDDK notes that some diabetes medicines slow carbohydrate digestion, and people taking them need glucose tablets or gel instead.
A quiet, slow yoga session rarely drops glucose as sharply as a run, but the trials that showed benefit included Surya Namaskara and kapalabhati, which are not quiet. Treat a 60-minute class as exercise.
How much yoga, and how do you start in India?
The trials that found benefit used about an hour a day, most days, for at least 12 weeks. Thind's meta-analysis reports a median of 50 sessions of 60 minutes over 12 weeks, a total of about 60 hours. The NMB trial began with a nine-day supervised camp and continued with daily home practice and one supervised class a week. Nothing in these studies supports the idea that ten poses done occasionally change HbA1c.
A practical sequence:
- Get a baseline HbA1c test and a dilated eye examination, so the programme can be matched to the state of the retina.
- Tell the doctor who prescribes the diabetes medicines; the NIDDK advises asking whether insulin or sulfonylurea doses need adjusting around exercise.
- Learn the protocol from a trained instructor for the first two weeks, as the trials did, before practising alone.
- Keep the aerobic target. The NIDDK and the NHS both recommend 150 minutes a week of activity that leaves you slightly out of breath; yoga adds to this, it does not replace it.
- Repeat HbA1c after three months, the point at which the trials measured change.
The NMB trial was funded through the Ministry of AYUSH's Central Council for Research in Yoga and Naturopathy (CCRYN), and its protocol is published in full in the Chandigarh sub-study paper, so an instructor can follow the same sequence.
Does yoga replace diabetes medicines?
No. In every trial in the 2022 meta-analysis, yoga was added to drug treatment, not substituted for it, and the reviewers wrote that yoga's independent effect is therefore hard to isolate. The NHS advice for type 2 diabetes is plain: "Do not stop taking your medicine unless your healthcare professional or doctor tells you to."
The size of the effect also sets expectations. A 0.47-point fall in HbA1c is useful, but it is not a cure, and the trials did not test whether it lasts beyond the 10 to 24 weeks they ran. The ADA's view is that yoga and tai chi "can be included based on individual preferences to increase flexibility, strength, and balance", with aerobic and resistance training as the core.
Yoga sits best inside a plan an endocrinologist or general physician reviews, with a dietician handling the meal pattern, which the NMB protocol also included. Someone whose HbA1c keeps rising despite three months of regular practice needs a medicine review, not more asanas.
This guide is general information from the sources listed below, not medical advice; a registered doctor makes decisions with you.
