A closely controlled clinical trial has found that a ketogenic diet produced bigger improvements in liver fat and blood sugar control than either a Mediterranean diet or a low-fat, plant-forward diet, even though all three approaches led to similar weight loss. The results, from a team at Washington University School of Medicine in St. Louis, add to evidence that what people eat — not just how much weight they lose — can shape metabolic health.
The trial, published in Cell Metabolism, enrolled 42 adults with obesity, prediabetes and excess liver fat, randomly assigning them to a ketogenic diet (73 percent fat, 4 percent carbohydrate), a Mediterranean diet or a high-carbohydrate, low-fat plant-forward diet for four to five months. Researchers supplied all of participants' meals and met with them weekly, a level of control meant to isolate the effect of diet composition from the usual challenges of self-reported eating. All three groups lost about 10 percent of their body weight. But the ketogenic group's liver fat fell by 67 percent on average, versus 45 percent in the other two groups, and prediabetes was reversed in half of the keto participants, compared with 29 percent on the Mediterranean diet and just 7 percent on the plant-forward diet.
A surprise on cholesterol
One finding ran counter to a common worry about very-high-fat diets: despite eating the most fat by far, the ketogenic group did not see a rise in blood cholesterol or other blood fats. According to the research team, hepatic insulin sensitivity — a measure of how well the liver responds to insulin — improved two to three times more in the keto group than in the other two.
"Our data suggest that if you have high liver fat content and prediabetes, reducing carbohydrate intake can have important therapeutic effects on metabolism beyond just weight loss alone."
— Samuel Klein, Washington University School of Medicine, senior author
Fatty liver disease, now often called metabolic dysfunction-associated steatotic liver disease, affects an estimated one in three American adults and is closely tied to obesity and prediabetes; left unmanaged it can progress to cirrhosis. Diet trials in this population have often struggled to separate the effects of weight loss itself from the effects of what, specifically, people ate to lose it — a gap this trial was designed to address by matching weight loss across all three groups.
The study is notably small — 42 participants split across three diet arms, meaning roughly 14 people per group — and lasted only four to five months, with all meals provided by researchers rather than prepared at home. That combination makes the results informative about biological mechanism but not necessarily representative of how people would fare, or whether they could sustain such a restrictive diet, outside a tightly monitored trial. The National Institutes of Health-funded team says longer follow-up is needed to see whether the metabolic gains persist and whether they translate into fewer cases of advanced liver disease or diabetes over time.
