If you saw the headline that keto “slashes liver fat by 67%,” you’re asking the right question: does keto beat a Mediterranean diet for fatty liver? In a Washington University trial published online in Cell Metabolism on Aug. 27, 2026, liver fat fell 67% on a ketogenic diet and 45% on a Mediterranean diet or a low-fat plant-forward diet. All three groups lost about the same 10% of body weight, so keto’s extra effect was about 22 percentage points, and both other diets worked too.
Because weight loss was matched, the extra effect is not explained by weight loss. A ScienceDaily report re-ran the news on Nov. 20, 2026; the trial itself was published in August. Here’s what it found, what it means for fatty liver and prediabetes, who it applies to, and what to ask your clinician.
Disclosure: Before you read on, this page contains affiliate links. Buying through them may earn us a commission.
What did the trial find?
Each group got all of its food and met with a dietitian every week for about five months, with the aim of losing about 10% of body weight. The Washington University release reports the design and results below. The main figures here come from the university's release and ScienceDaily's report of the paper.
Quick facts, as of 2026-11-20
| Item | Detail |
|---|---|
| Published | Cell Metabolism, online Aug. 27, 2026 |
| Study type | Randomized clinical trial with three diet groups |
| Who took part | Adults with obesity, prediabetes and fatty liver |
| Participants | 55 adults per the Washington University release; 42 per ScienceDaily |
| Length | About five months per the Washington University release; 4 months in the ScienceDaily headline |
| How food was handled | All food provided, with weekly dietitian meetings |
| Weight-loss target | About 10%; all three groups lost about the same amount |
| Keto diet shares | 4% carbs, 73% fat (ScienceDaily) |
| Mediterranean diet shares | 50% carbs, 35% fat (ScienceDaily) |
| Low-fat plant-forward diet shares | 70% carbs, 15% fat (ScienceDaily) |
| Funding | NIH grants; Foundation for Barnes-Jewish Hospital |
How much did liver fat and blood sugar change?
Liver fat fell 67% on keto and 45% on each of the other two diets. Here’s how the main measures lined up.
| Measure | Keto | Mediterranean | Low-fat plant-forward |
|---|---|---|---|
| Liver fat change | -67% | -45% | -45% |
| Blood insulin change | -74% | -44% | -27% |
| No longer meets prediabetes criteria at end of diet | 50% | 29% | 7% |
Liver insulin sensitivity improved two to three times more on keto. Over 24 hours, glucose fell 20% on keto and 8% on the other two diets. Muscle insulin sensitivity rose about 50% in all three groups.
Prediabetes was checked at the end of the diet. Half the keto group no longer met the criteria, compared with 29% of the Mediterranean group and 7% of the plant-forward group. In a trial this small, those shares are small numbers of people.
Does this mean the Mediterranean diet failed?
No. Both comparison diets cut liver fat by 45%, with the same weight loss as keto. Keto’s edge was the extra 22 points on top of that.
Can the Mediterranean diet reverse fatty liver disease?
This trial measured fat in the liver. In the Mediterranean group, that fell 45% over about five months, alongside a 10% weight loss. Whether that counts as reversing the disease is a question for your liver specialist, who can read your own tests.
For context, the NIDDK page on a NASH treatment study describes NASH as a more severe form of fatty liver disease, with fat buildup, inflammation and liver damage that can progress to cirrhosis, liver cancer or liver failure.
What about heart health and cholesterol on keto?
ScienceDaily quotes first author Max C. Petersen on blood fat and cholesterol. He says participants “didn’t experience an increase in blood fat or cholesterol levels.” A five-month trial cannot show long-term safety.
A different trial, Keto-Med, was published in 2022 in The American Journal of Clinical Nutrition. It enrolled 40 adults with prediabetes or type 2 diabetes and compared a keto-style diet with a Mediterranean-plus diet, 12 weeks on each. Its main analysis included 33 people with complete data. The keto-style diet also led to lower intakes of fiber and three other nutrients. The Keto-Med trial abstract names those as potential risks, including LDL cholesterol.
| Measure (12 weeks) | Keto-style (WFKD) | Mediterranean-plus |
|---|---|---|
| LDL cholesterol change | +10% | -5% |
| Triglyceride change | -16% | -5% |
| Body weight change | -8% | -7% |
The Keto-Med abstract also flags a diet-by-order interaction for body weight and HDL, and HbA1c (a blood sugar measure) did not differ between the diets, so read the weight row with care.
If you want a heart specialist’s view on keto, ask your cardiologist, and bring those LDL numbers with you.
Could I just eat this way on my own?
The Washington University trial provided every meal and held weekly dietitian meetings. Keto-Med also had a stretch where people prepared their own food. Adherence ranged widely on both diets, according to the Keto-Med adherence analysis. Its 12-week follow-up data point to the Mediterranean-plus diet as the easier of the two to keep up, per the Keto-Med trial abstract.
Doing it yourself means sourcing, measuring and remembering every meal. That’s the work the trial’s provided food and dietitians took on, and it’s the part to plan for.
What about bread, pasta, burgers, sandwiches, condiments and cheese?
Both trials compared whole eating patterns, not single foods. So neither one gives a yes or no on a particular bread, pasta, burger, sandwich, condiment or cheese. Here’s what they did set out:
- Washington University trial (per ScienceDaily): keto at 4% carbs and 73% fat; Mediterranean at 50% carbs and 35% fat; low-fat plant-forward at 70% carbs and 15% fat.
- Keto-Med: both diets were built around nonstarchy vegetables and left out added sugars and refined grains.
- Keto-Med: the Mediterranean-plus diet also included legumes, fruits and whole, intact grains, while the keto-style diet avoided them.
What did the Keto-Med Mediterranean-plus diet leave out?
This describes the separate Keto-Med trial’s diet, not the Mediterranean diet in the Washington University trial (50% carbs, 35% fat). In the Keto-Med trial, the Mediterranean-plus diet avoided added sugars and refined grains, the same as its keto-style diet. It leaned on nonstarchy vegetables, legumes, fruits and whole, intact grains.
Who is this for, and who should be careful?
The Washington University trial enrolled adults with obesity, prediabetes and fatty liver. Keto-Med enrolled adults with prediabetes or type 2 diabetes. If those describe you, these are the trials closest to your situation.
If you have diabetes and take a medicine that lowers blood sugar, or you have kidney disease, are pregnant, or have a history of an eating disorder, talk with your clinician before you change how you eat. Your prescriber is the one who decides whether a medicine should change alongside your diet.
How big was the trial, and what is still open?
The Washington University release gives 55 adults, and ScienceDaily gives 42. This was a small trial that lasted about five months and provided every meal, and it’s still an open question whether the liver gains last after the diet ends. No longer meeting prediabetes criteria at the end of a five-month diet is not a cure or a lasting reversal.
Who paid for the trial?
The Washington University release lists NIH grants and support from the Foundation for Barnes-Jewish Hospital. Samuel Klein, the corresponding author, reports advisory fees from AbbVie, 89Bio and Boehringer Ingelheim, plus an investigator-initiated grant from Merck and support for an industry-initiated trial from Viking Therapeutics.
What are the common mix-ups?
- “Slashes 67%” is keto alone. The extra over the other diets is about 22 percentage points, and all three groups lost about 10% of body weight.
- Keto beat Mediterranean does not mean Mediterranean failed. Both comparison diets cut liver fat 45%.
- “Reversed prediabetes.” Some write-ups use that phrase. The trial reports only who no longer met prediabetes criteria at the end of the diet, not a lasting reversal.
- “4 months” vs “five months.” ScienceDaily’s headline says 4 months and the Washington University release says about five.
- A diet trial, not a product trial. No supplement, product or brand was part of it.
- A republished story. The trial was published online Aug. 27, 2026; ScienceDaily’s report is dated Nov. 20, 2026.
Questions to bring to your clinician
- What stage is my fatty liver, and which tests would you use to track it?
- Do my blood sugar and prediabetes results point to a diet change, and which pattern fits me?
- If I take a medicine for blood sugar, how should a diet change be handled alongside it?
- Given the LDL rise in the Keto-Med trial, is my cholesterol a concern on a higher-fat diet?
- Can a registered dietitian help me build a food list I can keep up for months?
Your next step
Take the list above to your next appointment. If cravings are what make a plan hard to keep, read weight loss and appetite control. Build your daily routine around meals, movement and tracking, and how long to walk for weight loss covers the movement side. For the wider picture, see weight-loss plan evidence questions, and Ketogenic Diet Research: What Studies Actually Tested (our research library) for more on keto studies.
By the New Life Wellness Guide Editorial Team
This article is for general information purposes only and does not constitute medical advice. Consult your doctor or qualified healthcare provider before making changes to your health routine.