For type 2 diabetes, does a ketogenic diet actually beat a Mediterranean one on blood sugar?

Eric Charles WestmanMD, MHS · internist and Associate Professor of Medicine, Duke University

The trial that found no difference was biased against the ketogenic arm, and its own sensitivity analysis favours keto once the medication confound is removed.

What it rests onA published letter making four methodological objections: medication reduction differed between the two periods in a way that affects the primary outcome of HbA1c, adjusting glycaemic change for weight loss removes an advantage of the ketogenic arm, the crossover had no washout so the second period is contaminated, and no power calculation was reported. He quotes the trial's own text that excluding the four participants who discontinued medications produced a significant difference for both HbA1c and HDL cholesterol favouring keto. He opens by applauding the study and declares his equity in Adapt Your Life, Inc. in the same document. Behind the objection sits his own 2008 randomised trial, in which diabetes medications were reduced or eliminated in 95.2 percent of the ketogenic group against 62 percent of the comparison group.

Christopher D. GardnerPhD · nutrition scientist, Stanford Prevention Research Center

HbA1c did not differ between the two diets, and the medication protocol was matched to each dietary pattern on clinical safety grounds rather than to favour either arm.

What it rests onA randomised crossover trial in 40 adults with prediabetes or type 2 diabetes, each following a well-formulated ketogenic diet and a Mediterranean-plus diet for 12 weeks in random order, with both arms excluding added sugar and refined grain so the comparison isolates legumes, fruit and whole intact grains. HbA1c did not differ at 12 weeks; triglycerides fell more on keto, LDL rose about 10 percent on keto, and nutrient intakes were lower on keto. In his published reply he concedes two points in print: 'The effect of following a practical guideline improves the generalizability but reduces the fairness of the comparison. This could have been highlighted better in the article, we agree', and 'The primary analysis is not adjusted for weight change. It was only in the sensitivity analyses that we report the findings.' He notes 69 percent of participants with type 2 diabetes kept identical regimens across both phases, defends the absent washout as impractical and unethical since it would require regaining weight, and declines a post hoc power calculation on methodological grounds. His conflicts line reads 'The authors report no conflicts of interest.'