In adults with metabolically unhealthy obesity (prediabetes plus hepatic steatosis), ~10% weight loss across three diets all improved muscle insulin sensitivity by roughly 50% — but the very-low-carbohydrate ketogenic diet produced two-to-three-fold greater gains in hepatic insulin sensitivity compared to Mediterranean and very-low-fat plant-forward diets (p < 0.001). Intrahepatic triglyceride content, hepatic de novo lipogenesis, HbA1c, and 24-hour glucose and insulin profiles all fell most sharply in the ketogenic group. Critically, LDL-cholesterol, ApoB, and 24-hour triglycerides did not differ significantly between diets — a finding that defuses a common clinical objection to ketogenic eating.
This randomized trial matters because it isolates macronutrient composition from weight loss magnitude — the usual confound that makes diet comparisons murky. The liver is the engine of metabolic disease: hepatic insulin resistance drives fasting hyperglycemia and fatty liver, both of which accelerate cardiovascular and diabetes risk. Showing that diet quality, not just calorie deficit, reshapes hepatic metabolism is clinically meaningful.
Limitations are real: the trial likely ran over a defined, controlled period with supervised diets, which rarely reflects real-world adherence; the cohort is specifically people with prediabetes and steatosis, so findings may not generalize to metabolically healthy individuals. Still, for the large population of adults with fatty liver and prediabetes, this is more than incremental — it provides mechanistic justification for prescribing very-low-carbohydrate diets as a therapeutic, not merely a weight-loss, tool.
