Same 10% down on three diets. Muscle insulin sensitivity rose about 50% on keto, Mediterranean, and very-low-fat. Liver fat and hepatic insulin sensitivity did not.

Petersen, Klein, and colleagues at Washington University randomized adults with obesity, prediabetes, and hepatic steatosis. Cell Metabolism, 2026. Forty-two completed, 14 per group. All meals provided. Adherence over 95%. Weight loss about 10% in each arm over roughly 5 months.

Clamp glucose disposal, the muscle number, rose about 50% in all three groups. No diet difference. Fat-free mass loss was similar.

Hepatic insulin sensitivity increased two- to three-fold more on keto. MRI liver fat fell from 19.4% to 6.2% on keto, about 67%, versus about 45% on Mediterranean and very-low-fat. De novo lipogenesis fell more on keto. HbA1c 5.6% to 5.0% on keto. Fasting triglycerides down 82 mg/dL versus 41 and 24. Prediabetes remitted in 50% on keto, 29% Mediterranean, 7% very-low-fat. Small numbers.

LDL-C, ApoB, and 24-hour triglycerides did not differ among groups, even with 23% of keto energy from saturated fat. I still measure those labs. They did not rise on keto relative to the other arms in this trial.

Meals were packed out. The n is small. The authors cannot say a milder carb cut without ketosis would do the same, or that these liver effects hold if keto starts only after the weight is already off.

Weight loss still improved muscle insulin sensitivity on every diet. After weight loss was matched, keto still moved liver fat and hepatic insulin action more.

Petersen MC, Smith GI, Farabi SS, et al. Effect of diet macronutrient content on the cardiometabolic response to weight loss: A randomized clinical trial. Cell Metabolism. 2026;38:1-12. https://doi.org/10.1016/j.cmet.2026.07.020

This is educational, not medical advice for your specific case.