Can a plant-based diet prevent or reverse type 2 diabetes?

The objection

“Type 2 diabetes runs in my family, and I don’t see how going vegan would help. A diet of bread, rice, potatoes, beans and fruit is a diet of carbohydrate, the very thing that pushes blood sugar up, and doctors now recommend cutting carbs. The studies behind the vegan claim only show that health-conscious people get less diabetes, and the trials are small, short and run by people who already believe in the diet.”

The answer

Prevention has the largest body of evidence, though all of it is observational. A 2023 meta-analysis of 76 publications on plant-based eating and chronic disease, covering 2.2 million adults in all and 60,718 new cases of type 2 diabetes, found an 18% lower risk of type 2 diabetes among people whose diets were most plant-based (relative risk 0.82, 95% CI 0.77 to 0.86), and 21% when the score favoured healthy plant foods (0.79, 0.72 to 0.87) [1]. Results varied a good deal from study to study, and plant-based diets built on unhealthy plant foods went with a higher risk [1].

An earlier 2019 meta-analysis of nine cohorts, which the chart on this page shows, followed 307,099 adults, 23,544 of whom developed type 2 diabetes, and found a 23% lower risk among those whose diets were most plant-based (relative risk 0.77, 95% CI 0.71 to 0.84) [2]. Using a version of the score that favours healthful plant foods such as fruit, vegetables, whole grains, legumes and nuts strengthened the link to 30% (0.70, 0.62 to 0.79), while in three of the studies an unhealthful plant-based pattern, heavy in refined grains, starches and sugars, went with a higher risk [2]. A vegan diet of chips and white bread can’t borrow these results.

Part of the link runs through body weight. In the six studies that could be compared both ways, the pooled relative risk was 0.53 before allowing for body mass index and 0.79 after (2019) [2], so weight accounts for much of the gap, though not all of it. These are observational studies of people in high-income countries who reported their own diets on food questionnaires, and the authors say that confounding by something unmeasured can’t be ruled out [2]. The review gives relative risks only.

Vegans themselves have been studied most closely in the Adventist Health Study-2, which recruited Seventh-day Adventist church members across North America between 2002 and 2006. When they joined, 2.9% of the vegans had type 2 diabetes against 7.6% of the non-vegetarians, and after allowing for age, sex, ethnicity, education, income, exercise, television watching, sleep, alcohol and body mass index, the vegans’ odds were about half (odds ratio 0.51, 0.40 to 0.66) (2009) [3]. Over the next two years, 0.54% of the 3,545 vegans who began without diabetes reported developing diabetes, against 2.12% of the 17,695 non-vegetarians, an adjusted odds ratio of 0.381 (0.236 to 0.617) (2013) [4]. The question did not ask which type of diabetes, the diagnoses were self-reported, and the authors warn that so few vegans developed diabetes that their result should be read with care [4]. Pooling the studies that looked at vegans specifically, a 2025 meta-analysis put their risk 35% lower, though the estimate was imprecise (0.65, 0.42 to 1.00) [5].

Does a vegan diet lower blood sugar if you already have type 2 diabetes?

In randomised trials a vegan or vegetarian diet does lower blood sugar, by a modest margin that depends heavily on weight loss and on whether medicines change along the way. A 2019 meta-analysis of nine trials with 664 participants, carried out to update the European Association for the Study of Diabetes guidelines, found that vegetarian diets lowered HbA1c, a measure of average blood sugar, more than the comparison diets did (mean difference −0.29 percentage points, 95% CI −0.45 to −0.12), and cut body weight by 2.15 kg more, with moderate certainty of evidence [6]. A 2024 review carried out with the Academy of Nutrition and Dietetics’ Evidence Analysis Center, covering seven trials and 770 adults, found a slightly larger fall (−0.40 percentage points, 95% CI −0.59 to −0.21, moderate certainty) and concluded that vegetarian or vegan patterns belong among the options in nutrition care for adults with type 2 diabetes [7].

If you take insulin or tablets that lower blood sugar, such as gliclazide, talk to your GP or diabetes nurse before changing how you eat, because blood sugar can fall quickly and doses may need reducing. The NHS warns that gliclazide can cause low blood sugar, known as a hypo, and says to follow your doctor’s advice about diet while you take it [11]. In the Barnard trial described below, a study endocrinologist lowered medicine doses whenever blood sugar ran low [9]. That decision belongs to your own doctor, so talk to your GP before changing or stopping any medicine; the NHS says not to stop taking diabetes medicine unless your doctor tells you to [12]. Metformin, one of the medicines usually offered first, can cause vitamin B12 deficiency [12][13], so on a vegan diet a reliable B12 supplement or fortified foods matter even more.

The largest reviews here, of prevention and of treatment, included only adults [1][7]. The NHS says anyone with gestational diabetes should be referred to a dietitian [14], and Diabetes UK says a vegan diet needs particular care when it is planned for children [23], so if you are pregnant, or the diet is for a child or teenager with type 2 diabetes, plan any change with the diabetes team.

The best-known single trial, published in 2006, randomly assigned 99 people with type 2 diabetes to a low-fat vegan diet or to a diet following American Diabetes Association guidelines [8]. After 22 weeks, 43% of the vegan group and 26% of the comparison group had reduced their diabetes medicines. Counting everyone, HbA1c fell 0.96 points against 0.56, a difference that was not statistically significant (P = 0.089), while among those whose medicines stayed the same it fell 1.23 points against 0.38 (P = 0.01), and the vegan group lost 6.5 kg against 3.1 kg [8].

The same participants were followed to 74 weeks (2009) [9]. By then HbA1c had changed by −0.34 against −0.14 points, no longer a significant difference (P = 0.43), and weight loss of 4.4 kg against 3.0 kg didn’t differ significantly either. Counting each person’s last reading before any change of medicine, HbA1c fell 0.40 points on the vegan diet and rose 0.01 on the conventional one (P = 0.03), and across the trial weight change tracked HbA1c change (r = 0.50) [9]. The vegan group had no limits on portions or calories, while comparison participants who were overweight were set a daily deficit of 500 to 1,000 kcal, and by the end about half of each group (51% and 48%) met all their dietary targets [9].

The lead author of both papers, Neal Barnard, is president of the Physicians Committee for Responsible Medicine, and the trial was funded by the US National Institute of Diabetes and Digestive and Kidney Diseases and the Diabetes Action Research and Education Foundation [9].

A newer and larger trial in the Marshall Islands, which its authors describe as the largest and longest so far to compare a whole-food, plant-based programme with standard medical care, randomly assigned 169 adults with type 2 diabetes to standard medical care or to a whole-food, plant-based diet with moderate exercise (2025) [10]. The diet allowed small amounts of animal food after the first two weeks, and for the first 12 weeks it came with prepared meals, exercise sessions and classes. At 24 weeks HbA1c had fallen 1.4 percentage points against 0.7 (p = 0.01), 63% of the plant-based group on diabetes medicines had reduced them against 24%, and the plant-based group had lost 2.7 kg more [10]. The gap was larger at 12 weeks, while meals were provided (HbA1c down 2.0 points against 0.7), and gains in fasting glucose, insulin, cholesterol and triglycerides seen at earlier visits had faded by week 24 [10]. Of the 79 people assigned to the plant-based programme, 66 were analysed, against 70 of the 90 on standard care. Only the laboratory analysts were blinded, no diet records were kept, and with meals and exercise bundled in, the trial can’t isolate the food [10]. The trial was funded by the US Army and by Canvasback Missions, the organisation that developed and ran the programme at its own diabetes centre, and one author has written books on diet for diabetes [10].

HbA1c changes are in percentage points, and the Barnard figures here include people whose medicines changed.

Trial	HbA1c change, plant-based against comparison	Weight change
Barnard (2006), 99 people, 22 weeks	−0.96 against −0.56 (P = 0.089)	−6.5 kg against −3.1 kg
Barnard (2009), same 99, 74 weeks	−0.34 against −0.14 (P = 0.43)	−4.4 kg against −3.0 kg
Hanick (2025), 169 people, 24 weeks	−1.4 against −0.7 (p = 0.01)	2.7 kg more lost
Viguiliouk (2019), 9 trials, 664 people	Difference −0.29 (95% CI −0.45 to −0.12)	2.15 kg more lost
Guest (2024), 7 trials, 770 people	Difference −0.40 (95% CI −0.59 to −0.21)	BMI 0.96 lower

Can a plant-based diet put type 2 diabetes into remission?

A plant-based diet may help some people into remission, but the evidence is thin, and the main remission trials have tested other diets, with sustained weight loss as the dominant driver [16][19]. The NHS says lifestyle changes can sometimes lower blood glucose to a normal level and stop diabetes, which is called remission [12]. In the UK’s DiRECT trial, adults aged 20 to 65 with a body mass index of 27 to 45, diagnosed within the previous six years and not on insulin, were put on a low-calorie formula diet of 825 to 853 kcal a day for 12 to 20 weeks, followed by support to keep the weight off (2019) [15]. Of the 149 on the programme, 46% were in remission at one year and 36% at two years, against 3% of those on usual care at two years, and across the whole trial 64% of the participants who kept off at least 10 kg were in remission at two years [15]. Remission tends to fade, though: by five years, with lighter support, 13% of the 85 people still in the extended programme were in remission, against 5% of the 93 original usual-care participants with results, and only about a quarter (26%) of those in remission at two years still were (2024) [16]. Diabetes UK says a low-calorie diet works for most people but is not the only route, and that people have reached remission through different diets [17]. Low-carbohydrate diets have their own remission evidence: a 2021 meta-analysis of 23 trials found that at six months, in the eight trials that reported it, 57% of people on them had an HbA1c below 6.5%, against 31% on control diets, but the difference was smaller and not significant when remission had to be reached without diabetes medicine, and data at 12 months were sparse [18].

In the Marshall Islands trial, 5 people in the plant-based group (8%) went into remission within 24 weeks, meaning an HbA1c below 6.5% after at least three months off diabetes medicines, against none on standard care. All five had started with an HbA1c below 9%, and a post hoc analysis put remission at 23% of that subgroup, against none of the standard-care participants in it, a difference that was not statistically significant. The five lost no more than about 6 kg, and one gained weight. Remission was not one of the trial’s main outcomes [10]. Researchers in Toronto wrote in 2026 that the evidence for remission rests on weight-loss surgery and low-calorie meal-replacement diets, and their REPAIR trial of 160 adults is testing a plant-based version that aims for a 15% weight loss [19].

The NHS says not to switch to a very strict, low-calorie diet without talking to a health professional first, because it may not be suitable for people taking insulin or other diabetes medicines [12], and Diabetes UK advises speaking to your healthcare team before you change how you eat to try for remission [17].

If cutting back on food or chasing weight loss starts to take over, or you have had an eating disorder, talk to your GP before starting; help is also available from Beat in the UK (beateatingdisorders.org.uk), Bodywhys in Ireland (bodywhys.ie), ANAD in the US (anad.org), NEDIC in Canada (nedic.ca), the Butterfly Foundation in Australia (butterfly.org.au) and, for families and carers, EDANZ in New Zealand (ed.org.nz).

Is a high-carbohydrate diet bad for type 2 diabetes?

A diet high in carbohydrate from whole plant foods can still bring blood sugar down, as the vegan trials show, though cutting carbohydrate has the stronger evidence in the short term. The low-fat vegan diet in the 74-week trial aimed for about 75% of calories from carbohydrate, and the vegan group’s carbohydrate rose from 47.7% to 66.3% of calories against 46.3% to 46.5% in the comparison group. HbA1c still fell further on the vegan diet in an analysis that set aside readings taken after medicine changes, though the main comparison at 74 weeks was not significant (2009) [9].

The American Diabetes Association’s 2019 consensus report said that reducing overall carbohydrate intake had shown the most evidence for improving blood sugar and could be applied within a variety of eating patterns, which it judged acceptable for managing diabetes, and it put the average effect of vegetarian and vegan eating plans at 0.3 to 0.4 percentage points of HbA1c [20]. A 2021 report by the UK’s Scientific Advisory Committee on Nutrition, whose working group was chaired jointly with Diabetes UK, found that lower-carbohydrate diets lowered HbA1c more than higher-carbohydrate ones over three to six months, but that the evidence from 12 to 24 months was inconsistent, with no difference in HbA1c at 24 months and none in weight loss in the longer term [21]. It judged them an effective short-term option, said weight management remains the main goal, and warned that people on diabetes medicines may be at risk of hypos [21]. The nutrition study group of the European Association for the Study of Diabetes, updating its recommendations in 2023, found that a range of foods and eating patterns suit diabetes and advised eating minimally processed plant foods such as whole grains, vegetables, whole fruit, legumes, nuts and seeds while minimising red and processed meat, sugar-sweetened drinks and refined grains [22].

On a vegan diet the type of carbohydrate matters: Diabetes UK points out that plant sources of protein generally contain more carbohydrate, so intake may rise when you switch, and advises choosing low-glycaemic-index, high-fibre foods; it says vegan diets fit well with current dietary guidelines for people with diabetes and that there is no reason not to follow one if you wish, though you should discuss any queries or concerns with your diabetes team [23]. The NHS lists wholegrain foods and pulses such as chickpeas and lentils as part of a healthy diet for type 2 diabetes [12]. If you want both approaches at once, a low-carb vegan diet is possible, but on diabetes medicines plan it with your diabetes team first, because cutting carbohydrate sharply can bring on hypos [21].

Can a vegan diet help type 1 diabetes?

A vegan diet can’t prevent or cure type 1 diabetes, because it is an autoimmune condition in which the body can’t make insulin, and everyone who has it needs insulin every day [24]. The NHS says it is managed by checking blood glucose regularly, tracking what you eat and drink, and adjusting how much insulin you take [24]. A 2025 meta-analysis of 35 trials in type 1 diabetes found no effect on glucose management in the single trial of a vegan diet, with low certainty of evidence, while carbohydrate-restricted diets improved time in the target range and reduced insulin needs but not HbA1c [25]. A vegan diet can fit around type 1, but a big change in how much carbohydrate you eat means changing insulin doses, so plan it with your diabetes team [24].