Keto Diet Cut Liver Fat by 67 Percent in New Study

A clinical trial published August 27 in the journal Cell Metabolism (DOI 10.1016/j.cmet.2026.07.020) found that weight loss does not affect the liver and blood sugar the same way, depending on which diet is used to achieve it. The trial was run by a team from the Washington University School of Medicine in St. Louis and compared a ketogenic, a Mediterranean and a plant based diet in adults with obesity, prediabetes and fatty liver disease.

Obesity affects roughly four in ten American adults, and it is usually accompanied by higher risk of type 2 diabetes, fatty liver disease and cardiovascular complications. Weight loss has already been shown to reduce those risks, but it has not been clear whether the composition of the diet, meaning the ratio of carbohydrates, fat and protein, changes the outcome on its own, independent of how many pounds are lost. Our outlet has not previously covered this topic.

The number of participants differs across the available statements. Washington University’s own release says 55 people with metabolically unhealthy obesity took part, while a release distributed through EurekAlert cites 42 people with obesity, prediabetes and fatty liver disease. The trial ran four to five months. Participants received all of their food from the research team for the entire period and met weekly with a dietitian to ensure they stuck to their assigned diet. Caloric intake was adjusted for each participant so that everyone, regardless of group, lost about 10 percent of their starting body weight.

The three diets differed in macronutrient makeup: the ketogenic diet drew 4 percent of energy from carbohydrates and 73 percent from fat, the Mediterranean diet drew 50 percent from carbohydrates and 35 percent from fat, and the plant based diet, built around plant foods, drew 70 percent from carbohydrates and 15 percent from fat. All three groups ended up losing about the same amount of weight, roughly 10 percent of body mass, and muscle sensitivity to insulin improved by about 50 percent across all three groups equally, suggesting that weight loss itself, rather than diet choice, drives that part of metabolic health.

The difference showed up in the liver. Liver fat dropped 67 percent in the ketogenic group, compared with 45 percent in the Mediterranean and plant based groups. Liver sensitivity to insulin, which governs how well the liver keeps its own glucose production in check, improved two to three times more on the ketogenic diet than on the other two. Prediabetes resolved in 50 percent of participants on the ketogenic diet after the trial, compared with 29 percent on the Mediterranean diet and just 7 percent on the plant based diet. Blood sugar, measured continuously over 24 hours, fell 20 percent in the ketogenic group versus 8 percent in the other two, and blood insulin fell 74 percent versus 44 percent on the Mediterranean diet and 27 percent on the plant based diet.

“We know that weight loss can improve metabolic health in people who are obese,” Samuel Klein, the study’s senior author from the Washington University School of Medicine in St. Louis, said in the university’s release. “Our data show that if you have high liver fat content and prediabetes, cutting carbohydrate intake can have important therapeutic effects on metabolism, beyond weight loss itself,” Klein added.

The paper’s first author, assistant professor Mahesh (Max) Petersen, pointed to an unexpected safety finding in the university’s release: “What surprised us most was that participants in the ketogenic diet group did not see a rise in blood fat and cholesterol levels, even though they consumed the most fat.”

The research was funded through the National Institutes of Health, under grant numbers P30 DK56341, P30 DK20579, UL1 TR002345 and S10 OD027006, and through the Barnes Jewish Hospital Foundation. Alongside Klein and Petersen, associate professor Gordon Smith is listed as a co-author on the paper.

The trial has clear limits. The sample is small, ranging from 42 to 55 people depending on which release is cited, and limited to people who already have metabolically unhealthy obesity, prediabetes and fatty liver disease, so the findings cannot be extended to the wider population without further study. All food was supplied and controlled for participants, a condition that is hard to replicate in everyday life outside a research setting. The follow up period was limited to four to five months, so the study does not show whether the ketogenic diet’s advantages hold up over the longer term. The authors themselves say their next step is to clarify the mechanisms behind these differences, including how they relate to the effects of GLP-1 weight loss drugs.