Can type 2 diabetes be reversed or put into remission naturally?

What remission means. An expert group convened by the American Diabetes Association defined remission as an HbA1c (a blood test of average sugar over about three months) below 6.5% (48 mmol/mol), measured at least 3 months after stopping glucose-lowering medicine. It is not a cure, and they advise continuing to observe people in remission⁴.

Weight is the main lever. In DiRECT (306 adults diagnosed within the past 6 years, not on insulin), remission rose with weight lost: none of the 76 people who gained weight, 7% of those who lost up to 5 kg, 34% at 5 to 10 kg, 57% at 10 to 15 kg and 86% at 15 kg or more. The programme withdrew diabetes and blood pressure medicines, used a formula diet of 825 to 853 kcal a day for 3 to 5 months, then reintroduced food and offered long-term support¹. A trial in Qatar in adults aged 18 to 50 with diabetes for 3 years or less found remission in 61% versus 12% with usual care at 12 months⁵.

Keeping it off is harder. At two years, 36% of the DiRECT group were in remission; at five years, 11 of 85 people (13%) who received continued support were². A standard behavioural programme with counselling, calorie reduction and activity (4,503 US adults) produced partial or complete remission in 11.5% in year 1 and 7.3% by year 4, versus 2.0% with usual education³.

Food patterns. Low-carbohydrate diets (under 130 g of carbohydrate a day) raised remission at 6 months (57% versus 31%, with remission defined by HbA1c alone), but the benefit was smaller if remission required being off medicines and was unclear by 12 months⁶. A Mediterranean-style diet delayed the need for medicine in newly diagnosed people⁷. See our Mediterranean diet article.

Exercise and fasting. Structured exercise lowers HbA1c on its own⁸. Two non-consecutive days a week of 500 to 600 kcal gave similar HbA1c results to daily calorie restriction over a year⁹. See our insulin resistance guide.

Supplements are minor. In 10 trials, cinnamon lowered fasting glucose by about 25 mg/dL on average but did not significantly change HbA1c¹⁰; see our cinnamon article.

What remission does not prove. A large trial of intensive weight-loss lifestyle changes did not reduce the rate of cardiovascular events¹¹. Keep your regular checks.

DiRECT and DIADEM-I

DiRECT randomised 49 UK primary care practices (306 adults aged 20 to 65, diabetes under 6 years, BMI 27 to 45, no insulin). At 12 months, remission (HbA1c under 6.5% at least 2 months off all diabetes drugs) occurred in 68 of 149 people (46%) versus 6 of 149 (4%) (odds ratio 19.7, 95% CI 7.8 to 49.8). Mean weight fell 10.0 kg versus 1.0 kg. Nine serious adverse events occurred in the intervention group and two in controls; two (biliary colic and abdominal pain) were possibly related¹. At 24 months, 36% of the intervention group were in remission². The five-year extension followed 85 people who continued support: they lost 6.1 kg on average and 13% were in remission; among all original intervention participants, 27% of visits were in remission versus 4% for controls, and 26% of those in remission at year 2 were still in remission at year 5². DIADEM-I (158 randomised, 147 analysed) used a similar formula-diet approach with physical activity support: weight fell 11.98 kg versus 3.98 kg, and remission occurred in 61% versus 12% (OR 12.03, 5.17 to 28.03)⁵. Both were open-label, in early diabetes, and used intensive supervised programmes.

Look AHEAD

The Look AHEAD lifestyle programme (2,241 intensive versus 2,262 education) achieved weight loss of 7.9% more than controls at year 1 and 3.9% more at year 4. Any remission was 11.5% in year 1 and 7.3% at year 4, versus 2.0%; sustained remission for 4 years was 3.5% versus 0.5%. The authors called absolute rates modest³. In the main trial (5,145 adults, stopped early for futility at a median 9.6 years), the primary cardiovascular outcome occurred at 1.83 versus 1.92 events per 100 person-years (HR 0.95, 95% CI 0.83 to 1.09)¹¹.

Diet and exercise trials

A meta-analysis of 23 trials (1,357 people) found low-carbohydrate diets gave 57% (76 of 133) remission at 6 months versus 31% (41 of 131) (risk difference 0.32, 95% CI 0.17 to 0.47; eight studies, 264 people), with smaller non-significant effects when remission meant off medicine, and a marked drop in studies including insulin users. At 12 months data were sparse, and benefits for weight, triglycerides and insulin sensitivity diminished. Certainty was moderate to low⁶. In 215 newly diagnosed overweight adults, 44% on a Mediterranean-style diet versus 70% on a low-fat diet needed drugs by year 4 (HR 0.63, 0.51 to 0.86); investigators starting drugs were not blinded and intake was self-reported⁷. A meta-analysis of 47 trials (8,538 patients) found structured exercise lowered HbA1c by 0.67 points (95% CI 0.49 to 0.84), aerobic by 0.73, resistance by 0.57 and combined by 0.51; more than 150 minutes a week lowered it by 0.89 versus 0.36 for 150 minutes or less; advice alone changed nothing unless paired with diet advice⁸.

Fasting and supplements

In 137 adults with type 2 diabetes, two days a week of 500 to 600 kcal and daily restriction of 1,200 to 1,500 kcal produced equivalent HbA1c changes at 12 months (-0.3% versus -0.5%; 90% CI of difference -0.2 to 0.5); 97 completed the trial, and weight loss was not shown to be equivalent. Glucose events affected 35% of those taking sulfonylureas or insulin in the first two weeks⁹. Cinnamon at 120 mg to 6 g a day for 4 to 18 weeks (10 trials, 543 patients) lowered fasting glucose by 24.6 mg/dL (95% CI 8.7 to 40.5) but not HbA1c (-0.16%, -0.39 to 0.02), with high heterogeneity¹⁰.

Where the evidence disagrees

Remission rates depend on the definition and programme: 46% to 61% with formula diets, about 7% to 12% with standard behavioural programmes, and 57% at six months for low-carbohydrate diets defined by HbA1c alone^1,3,5,6. Durability is the weak point, with most gains lost by five years without continued support². Remission has not been shown to prevent heart attacks in a randomised trial¹¹, and the DiRECT-type programmes are not "natural" in the sense of unsupervised: they were medically managed^1,5. SGLT2-inhibitor ketoacidosis can occur with low-carbohydrate diets or fasting¹².

Bottom line (confidence 4 of 5). Early type 2 diabetes can go into remission with large, sustained weight loss, and exercise and food-pattern changes help glucose control. It is a supervised, effortful process that often needs ongoing support, and medicines remain central for many people.