What are the most effective diet patterns (low-carb vs low-fat vs Mediterranean) for reversing prediabetes in randomized trials?
Executive summary
Randomized trials show that carbohydrate-restricted diets (low-carbohydrate and ketogenic) and Mediterranean-style diets can all improve glycemia and induce diabetes remission in some people with prediabetes or early type 2 diabetes, but the size and durability of those effects depend heavily on weight loss, intensity of support, and adherence; pooled trial data suggest low‑carbohydrate approaches increase short‑term remission rates versus mostly low‑fat comparators, while Mediterranean patterns show strong prevention signals in at‑risk populations [1] [2] [3].
1. Randomized‑trial landscape: what has been tested and in whom
High‑quality randomized trials relevant to prediabetes include multi‑arm weight‑loss trials (DIRECT) comparing low‑fat, Mediterranean, and low‑carbohydrate diets over two years (DIRECT, n≈322) [4] [3], crossover trials comparing ketogenic versus Mediterranean patterns in mixed prediabetes/T2D populations (Keto‑Med, n≈33 completers) [5] [6], and multiple smaller randomized trials and systematic reviews/meta‑analyses that pool LCD versus LFD trials in overweight and diabetic populations [7] [8] [9].
2. Glycemic control and remission outcomes in randomized trials
Meta‑analyses and pooled randomized data report meaningful short‑term benefits for carbohydrate restriction: a BMJ meta‑analysis found low‑carbohydrate diets were associated with a large (≈32%) increase in diabetes remission at six months versus mostly low‑fat control diets in randomized trials [1]. Individual trials report mixed specific outcomes: Keto‑Med found no difference in HbA1c between ketogenic and Mediterranean phases after 12 weeks, though both improved from baseline [6], while DIRECT reported more favorable glycemic effects with the Mediterranean diet subgroup in some analyses [3] [2].
3. Weight loss versus macronutrient composition: the proximate driver
A consistent theme across randomized studies is that weight loss itself is a major mediator of improved glucose regulation and remission; trials that achieved greater or faster weight loss—whether via carb restriction, caloric restriction, or structured programs—tended to produce larger glycemic benefits [4] [10]. Some trials show greater early weight loss with low‑carbohydrate approaches and attendant short‑term glycemic gains [7] [11], but long‑term advantage often narrows as adherence wanes [4] [10].
4. Mediterranean diet evidence: prevention and clinically meaningful glycemic change
The Mediterranean pattern carries strong trial evidence for diabetes prevention and modest glycemic improvements: the PREDIMED randomized trial reported a ~52% reduction in incident diabetes in high‑cardiovascular‑risk older adults assigned to Mediterranean diets enriched with extra‑virgin olive oil or nuts versus a low‑fat control [2], and randomized comparisons have noted HbA1c reductions and favorable cardiometabolic profiles with Mediterranean patterns versus low‑fat diets [2] [3].
5. Adherence, sustainability, and safety signals that matter clinically
Adherence distinguishes efficacy from effectiveness: crossover and follow‑up data show ketogenic or very low‑carbohydrate diets can produce pronounced short‑term benefits but are often harder to sustain, with many participants reverting to less restrictive patterns post‑trial [6] [12]. Safety and lipid signals are mixed—some low‑carbohydrate trials report triglyceride improvements but potential LDL increases—so cardiovascular risk considerations matter when choosing an approach [6] [10] [11].
6. Practical synthesis: which pattern best “reverses” prediabetes in RCTs
Randomized evidence supports three conclusions: carbohydrate‑restricted diets can increase short‑term remission rates versus mainly low‑fat comparators [1]; Mediterranean diets show strong prevention effects and clinically meaningful glycemic improvements with likely better long‑term sustainability [2] [3]; and across trials, weight loss and program intensity—not a single macronutrient—are primary determinants of reversal or remission [4] [7].
7. Limitations, open questions, and where the evidence is thin
Existing trials vary widely in definitions (prediabetes vs T2D), carbohydrate thresholds, caloric targets, duration, and behavioral support, and many randomized studies are small or short‑term—limiting conclusions about durability, long‑term cardiovascular outcomes, and which subgroups benefit most; these gaps are acknowledged across the trials and meta‑analyses [7] [8] [10].