Keto vs Mediterranean: What a New 2026 Study Found

Three people lose the same 10% of their body weight. One eats almost no carbs, one eats a Mediterranean plate, one goes low-fat and plant-heavy. They end up lighter, but not equally healthy inside. That is the surprising takeaway from a clinical trial published August 27, 2026 in Cell Metabolism.

Researchers at Washington University in St. Louis wanted to know something specific: once you strip out weight loss itself, does the type of diet still matter for metabolic health? The short answer is yes, at least for liver fat and blood sugar.

Key Takeaway: In a 2026 trial, three very different diets all produced about 10% weight loss, but the ketogenic diet led to the largest improvements in liver fat and prediabetes markers.

What the study actually did

The trial followed 42 adults who completed the study, all living with obesity, prediabetes, and excess liver fat. Each person was randomly assigned to one of three eating plans and given every meal for roughly five months, with weekly check-ins from a dietitian.

The three diets sat far apart on the macronutrient map. The keto plan drew just 4% of calories from carbs and 73% from fat. The Mediterranean plan landed at 50% carbs and 35% fat. The low-fat, plant-forward plan pushed carbs up to 70% and fat down to 15%.

Stat: The ketogenic arm ate 4% of calories as carbohydrate; the low-fat arm ate 70%. Despite that gap, everyone hit the same 10% weight-loss target.

Calorie intake was tuned for each person so that all three groups lost a similar amount of weight. That design is the clever part. If everyone loses the same weight, any leftover difference in health markers points to the food, not the scale.

Where keto pulled ahead

All three diets improved metabolic health. Body fat dropped, insulin sensitivity improved, and participants across every group got healthier on paper. Weight loss did most of the heavy lifting, and that holds no matter which plan you pick.

The ketogenic group stood out on two measures. About half of the keto participants no longer met the criteria for prediabetes by the end, and the group saw the biggest reduction in liver fat. First author Max Petersen noted that several separate markers of liver health moved together, which is what made the result convincing rather than a fluke.

Key Takeaway: Roughly 50% of people on the keto diet no longer met the definition of prediabetes after four to five months, a larger shift than the other two diets produced.

One detail worth holding onto: this was not junk-food keto. The plan leaned on eggs, fish, meat, cheese, and plenty of vegetables. That is a well-formulated low-carb diet, not a bacon free-for-all, and it may explain why the results looked as clean as they did. If you are curious about the flip side of very low-carb eating, we covered a separate 2026 study on keto and cancer risk.

Why this does not crown a winner

Here is where I want to pump the brakes a little. The trial was small, only 42 people finished, and it ran for months, not years. It tells us what happened to specific metabolic markers over a short window. It does not tell us who keeps the weight off, who sticks with the diet, or what happens to heart health over a decade.

The people in the study also had a specific profile: obesity plus prediabetes plus fatty liver. If that describes you, the liver and blood-sugar findings are directly relevant. If it does not, the results may not transfer, and the Mediterranean diet still has the deepest long-term evidence base for heart health and diabetes prevention, as we discussed in our look at the Mediterranean diet and diabetes risk.

Stat: Only 42 participants completed the trial, and it lasted about five months, so the findings describe short-term metabolic change, not long-term outcomes.

What to take from it

The practical message is not "go keto." It is that weight loss helps almost everyone, and the composition of your plate can add a second layer of benefit for certain people. Reversing prediabetes does not always require dropping pounds either, something we explored in how to reverse prediabetes without losing weight.

If you do experiment with a lower-carb approach, the version that mirrors this study is built on whole foods and real vegetables, and it tracks more than just carbs. Keeping an eye on protein, fat, and fiber, not only calories, is what separates a structured plan from guesswork. The best diet is still the one you can actually live with for longer than five months.

Sources

FAQ

Did the keto diet cause more weight loss than the others? No. All three diets were calorie-tuned so participants lost about the same 10% of body weight. The keto group's advantage showed up in liver fat and prediabetes markers, not on the scale. Weight loss itself improved metabolic health across every group.

Does this mean keto is the healthiest diet for everyone? Not necessarily. The trial only included people with obesity, prediabetes, and fatty liver, and it ran about five months with 42 finishers. The findings may not apply to healthy adults, and long-term effects were not measured.

What kind of keto diet did the study use? A well-formulated one. Meals featured eggs, fish, meat, cheese, and vegetables rather than processed food. Carbohydrates made up just 4% of calories while fat provided 73%. This is different from an unstructured high-saturated-fat version.

Is the Mediterranean diet still worth following? Yes. It improved metabolic health in the trial and has the strongest long-term evidence for heart health and diabetes prevention. This study does not replace that record; it adds nuance for a specific high-risk group.

-- Selena