Diets
ketogenic diet induces T2D remission
In plain terms: Can a keto diet lastingly reverse type 2 diabetes?
Leans against Diets 💰 Industry COI noted🔬 Includes disconfirming
Part of: 🥗 ketogenic diet
RefutedContestedStrong support
consensus score -0.27
Probably not — early gains mostly fade by a year as the diet gets hard to keep.
📅 Last reviewed: 2026-07-15 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: All trials, pooled (Meta-analysis)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
3 support 3 contradict 1 tested null 6 mixed · 13 sources, 6 independent groups
What the evidence shows
The 6-month T2D remission/weight/HbA1c advantage of low-carb diets largely DISAPPEARS by 12 months - adherence collapse, not a permanent metabolic reset. Keto is a short-term lever, not a proven durable cure.
The evidence (15)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Churuangsuk C, Hall J, Reynolds A, Griffin SJ, Combet E, Lean MEJ 2022 · Diabetologia | meta-analysis | supports | high | Two-part review for EASD guidelines: an umbrella review of 19 weight-loss meta-analyses plus a systematic review (no pooling) of 14 remission studies. Neither part contains its own new data. For ketogenic diets it found no RCT of remission; the only study was one non-randomised controlled trial (Hallberg 2018, the Virta cohort, already counted on this claim through its 2-year and 5-year reports) reporting 20 percent remission at 1 year (52 of 262 by ITT) against none in usual care, rated GRADE very low for serious risk of bias. The earlier extract's 'not superior for remission, benefits converge by 12 months' came from the weight-loss umbrella half (low-carb no better than low-fat for weight, GRADE high), not from any remission result. |
| Parry-Strong A, Wright-McNaughton M, Weatherall M, Hall RM, Coppell KJ, Barthow C, Krebs JD 2022 · Diabetes Obes Metab | meta-analysis | contradicts | moderate | Meta-analysis of VLCKD RCTs: pooled 12-month HbA1c difference vs control ~0.01% (null); no durable glycemic advantage over other diets. |
| Goldenberg JZ, Day A, Brinkworth GD, Sato J, Yamada S, Jönsson T, Beardsley J, Johnson JA, Thabane L, Johnston BC 2021 · BMJ | meta-analysis | mixed | moderate | FT 6mo remission RD0.32 BUT med-free RD0.05 NS; 12mo remission NS, weight benefit vanishes |
| Das R, Mohammad N, Islam MS, Das S, Abdullah F, Kader MA, Mamun MAA, Dutta S, Unwin D, Mazak I 2025 · Cureus | observational | mixed | low | DiRECT 62% yr1->13% yr5; weight regain; HbA1c back to baseline; narrative SR no meta |
| Guess ND 2022 · Diabetes Ther | observational | mixed | moderate | Review: durable T2D remission predicated on sustained >=10kg weight loss, not keto per se; adherence wanes |
| Iyengar S, Kaur J, Singh A, Sahay I, Kumar A, Kolke S, Raut M, Singh RK 2026 · Frontiers in nutrition | n-of-1 | supports | low | N-of-1: HbA1c 7.2%->4.7-5.3% medication-free sustained for 10 years on low-carb diet (~100g/day) |
| Jayedi A, Zeraattalab-Motlagh S, Jabbarzadeh B, Hosseini Y, Jibril AT, Shahinfar H, Mirrafiei A, Hosseini F, Bidar SS 2022 · Am J Clin Nutr | meta-analysis | tested-null | moderate | Dose-response MA: HbA1c/weight benefit weakening by 12 months |
| Lan J, Chen M, Zhang X, Yang J 2025 · Asia Pacific journal of clinical nutrition | meta-analysis | contradicts | moderate | Dose-response MA, 38 RCTs/2831pts: HbA1c/FG/HOMA-IR improved per 10% carb cut, but 'effects diminished beyond' 6 months |
| Saslow LR, Daubenmier JJ, Moskowitz JT, Kim S, Murphy EJ, Phinney SD, Ploutz-Snyder R, Goldman V, Cox RM, Mason AE, Moran P, Hecht FM 2017 · Nutrition & Diabetes | RCT | supports | moderate | 12-month RCT: very-low-carb ketogenic diet produced greater HbA1c reduction than moderate-carb low-fat diet in T2D/prediabetes. |
| Athinarayanan SJ, Adams RN, Hallberg SJ, McKenzie AL, Bhanpuri NH, Campbell WW, Volek JS, Phinney SD, McCarter JP 2019 · Front Endocrinol | observational | supports | moderate | Virta 2-year non-randomized CCI with nutritional ketosis: sustained HbA1c, weight, medication reductions and diabetes remission/reversal vs usual care. |
| Zinn C, Campbell JL, Fraser L, Davies G, Hawkins M, Currie O, Cannons J, Unwin D, Crofts C, Stewart T, Schofield G 2025 · Nutrients | observational | mixed | low | Retrospective NZ audit, median follow-up 19mo (IQR 6-32): HbA1c -3 mmol/mol (p=0.004), 32% T2D reversal at final follow-up |
| Mongkolsucharitkul P, Surawit A, Pimsen A, Winitchayothin S, Pumeiam S, Pinsawas B, Ophakas S, Suta S, Pasookhush P, Mayurasakorn K 2025 · Diabetes Res Clin Pract | meta-analysis | mixed | moderate | SR of 27 RCTs: low-carb diets improved HbA1c short-term (-0.29%) but benefits diminished over time; no arm sustained very-low-carb intake. |
| Choy KYC, Louie JCY 2023 · Diabetes & metabolic syndrome | meta-analysis | contradicts | moderate | Meta-analysis (11 RCTs): over 2 years ketogenic diet showed NO additional glycemic-control or weight benefit vs control, only lipid changes. |
| McKenzie AL, Athinarayanan SJ, Van Tieghem MR, Volk BM, Roberts CGP, Adams RN, Volek JS, Phinney SD, Hallberg SJ 2024 · Diabetes Res Clin Pract | observational | mixed | moderate | Virta 5-year extension: only 20% of completers in remission and heavy attrition (262 to 122)—durability real but partial and dropout-prone; industry-funded, non-randomized. |
| Li S, Ding L, Xiao X 2021 · International journal of endocrinology | meta-analysis | mixed | moderate | Meta-analysis (12 RCTs): low-carb diets reduced HbA1c (-0.35%) and weight more than low-fat, with longer-term cardiovascular benefit. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.