Keto diet lowers blood sugar

In adults with type 2 diabetes, keto lowers blood sugar more than comparison diets for the first several months, and the advantage fades by about a year. A 2020 meta-analysis of 37 randomized trials in 3,301 people found that very-low-carbohydrate diets lowered HbA1c, the blood test used to track average blood sugar over recent months, by about 0.4 percentage points more than comparison diets at three and six months, with no difference at 12 or 24 months. A 2022 review of strict keto trials lasting at least six months found a difference of 0.01 points at 12 months across the four trials it could pool, while two further trials analysed separately found 0.65 points lower on keto. The largest pooled count of keto-level diets so far, a 2024 review of 29 studies in 2,359 people that included some non-randomized studies, put HbA1c about 0.3 points lower on keto overall and rated that evidence very low certainty.

In people without diabetes, and for anyone taking medicines that lower blood sugar, the picture differs. A 2020 meta-analysis of 14 trials found HbA1c fell about half a point more on keto than on low-fat diets in people with type 2 diabetes, but found no significant difference in glycemic measures in the 290 participants without diabetes. In a 2024 trial that randomized 60 healthy adults, fasting glucose was lower on keto than on the comparison diet at four weeks but no longer at 12 weeks, and the rise in blood sugar after a carbohydrate-containing test meal was larger on keto at both points. The UK guideline for type 2 diabetes makes reducing the risk of hypoglycaemia (blood sugar falling too low) a particular aim when carbohydrate intake is set for people using insulin or drugs that stimulate insulin release, and the prescribing label for empagliflozin (Jardiance), an SGLT2 inhibitor, lists a ketogenic diet among conditions that can precipitate ketoacidosis, a medical emergency.

Part of what the trials measure is medication change. In one strict keto trial pooled in the 2022 review, 95% of the keto group reduced diabetes medication against 62% of the comparison group, so HbA1c comparisons in these trials reflect diet and medication changes together. A 2021 review of 23 trials in 1,357 adults found that 57% of people on low-carbohydrate diets reached an HbA1c under 6.5% at six months, against 31% on comparison diets, when diabetes medication was allowed. Counting only people who reached that level without medication, the difference was smaller and not statistically significant, it shrank in trials that included insulin users, and by 12 months the data were sparse. A 2022 trial enrolled 150 adults with HbA1c between 6.0% and 6.9% who were not on diabetes medication and set a keto-level carbohydrate target for the first three months. At six months HbA1c was 0.23 points lower than on a usual diet and weight was 5.9 kg lower, and the trial could not separate the effect of the diet from the weight loss; participants reported eating 66 to 75 grams of net carbohydrate a day against targets of 40 and then 60.

The advantage does not hold over longer periods in most analyses. In the 2020 meta-analysis, the broader group of low-carbohydrate diets, averaging 36% of calories from carbohydrate, lowered HbA1c by 0.17 points at three months, showed no difference at six and 12 months, and was 0.23 points higher than comparison diets at 24 months. A 2023 pooled analysis of 11 randomized keto trials in type 2 diabetes found no significant difference in blood sugar control over periods of up to two years. The 2024 review found HbA1c 0.23 points lower on keto in studies running longer than six months; two reports of one non-randomized study make up about a quarter of all the participants that review counted. Carbohydrate intake on keto also drifted upward: in one two-year trial pooled in the 2022 review, the keto group ate about 57 grams a day at 24 weeks, 74 grams at one year and 83 grams at two years, against a target of under 50.

Against a Mediterranean-style diet that also cut added sugar and refined grains, HbA1c results did not differ. In a 12-week randomized crossover trial of 33 adults with prediabetes or type 2 diabetes, HbA1c improved from baseline on both diets and did not differ between them, while LDL cholesterol rose 10% on keto and fell 5% on the Mediterranean diet. A 2026 pooled analysis of randomized trials comparing keto with Mediterranean diets likewise found no significant difference in fasting blood sugar or HbA1c.

Nearly all of this comes from adults with overweight or obesity, most with type 2 diabetes. The no-difference finding in people without diabetes rests on 290 participants and includes one trial in children and adolescents, and the trial in healthy adults lasted 12 weeks. None of these sources tests keto in type 1 diabetes or pregnancy, and none follows blood sugar beyond two years.