Same Weight Loss, Different Liver

Weight loss is supposed to be the whole story for a fatty liver. Drop enough weight, the thinking goes, and the liver will follow the scale. That idea is not empty. It is just incomplete.

In August 2026 a feeding trial from Washington University School of Medicine in St. Louis put that idea under a tighter light. Three diets. Same target weight loss. Food provided. The scale moved about the same. The liver did not.

This is what we here at Ascent Metabolic Coaching call trail intel. It is a map of what a specific study showed, including the parts that are weaker than a caption. You stay in control. If you later sign up as a client, you will be the climber. Nothing in this piece is a reason to start, stop, or change medication on your own.

What the internet did with it

Two physicians posted the paper on X in late August. One stayed close to the design: matched weight loss, carbohydrate reduction, liver fat. The other used the word "crushes." That is the usual split. A careful trial arrives. The caption chooses a winner.

The paper is not a personality contest between diet tribes. It is a small, tightly controlled look at whether the mix of carbohydrate, fat, and protein changes what weight loss does inside the liver.

Who ran it, and on whom

The first author is Max C. Petersen, an assistant professor of medicine at WashU Medicine. The senior author is Samuel Klein, a professor of medicine and nutritional science at the same school. The trial is registered as NCT02706262. The journal is Cell Metabolism. The paper is titled "Effect of diet macronutrient content on the cardiometabolic response to weight loss: A randomized clinical trial." It appeared online on 27 August 2026. DOI: 10.1016/j.cmet.2026.07.020.

They randomly assigned 55 adults with metabolically unhealthy obesity, meaning obesity plus prediabetes plus fat stored inside the liver, to one of three eating patterns. Forty-two people completed the study and went into the main analysis, 14 on each diet.

The three patterns:

• A very-low-carbohydrate ketogenic diet: about 4 percent of calories from carbohydrate, 73 percent from fat, 23 percent from protein.

• A Mediterranean-style diet: about 50 percent carbohydrate, 35 percent fat, 15 percent protein.

• A very-low-fat, plant-forward diet: about 70 percent carbohydrate, 15 percent fat, 15 percent protein.

Every meal and snack was supplied. Calories were adjusted so each group would lose about 10 percent of starting weight over roughly five months. Participants met a study dietitian every week. That last point matters. This is not how most people eat at home on a Tuesday in Ontario.

Before the weight-loss phase, everyone ate a provided standard Western pattern (50 percent carbohydrate, 35 percent fat, 15 percent protein) so the starting line was shared.

What the trial showed

Muscle insulin sensitivity (how well muscle responds to insulin) rose by about 50 percent on all three diets. Weight loss helped the muscle. The diet label did not decide that part.

The liver was different.

Intrahepatic triglyceride (fat stored inside the liver) fell by about 67 percent on the ketogenic diet and by about 45 percent on each of the other two. Hepatic insulin sensitivity (how well the liver responds to insulin, including how well it turns down extra glucose production) improved two to three times more on the ketogenic diet than on the other patterns.

WashU's newsroom summary of the paper also reported that 24-hour blood glucose fell about 20 percent from baseline on the ketogenic diet, versus about 8 percent on the other two, and that circulating insulin fell further on the ketogenic arm (about 74 percent versus 44 percent on Mediterranean and 27 percent on the high-carbohydrate diet). The paper itself says glycated hemoglobin (HbA1c, a roughly three-month average of blood glucose) and 24-hour glucose and insulin fell most on the very-low-carbohydrate arm.

Lab-defined prediabetes clearance, meaning people no longer met the lab criteria the team used for prediabetes, was more common on the ketogenic diet in this small sample: 50 percent versus 29 percent on Mediterranean and 7 percent on the plant-forward diet. That is a lab threshold in 14-person groups. It is not a promise that a reader will "reverse" anything.

LDL cholesterol (the cholesterol particle often used as a heart-risk marker) and apolipoprotein B (a count of the particles that carry that cholesterol) did not differ by diet. Twenty-four-hour plasma triglycerides also did not differ among groups. In a trial where the ketogenic diet was high in fat, including saturated fat, those blood lipids did not rise relative to the other two arms. That is a finding. It is not a lifetime cardiac verdict.

The team also looked at de novo lipogenesis, the liver making new fat from carbohydrate. Press summaries of the paper say that new-fat production fell on the ketogenic and Mediterranean diets, and fell most on ketogenic, but not on the plant-forward diet. Treat that line as supporting mechanism, not as the headline.

Why a matched scale can still hide a liver difference

The usual public story is simple: lose weight and metabolic health follows. Petersen put the question another way. If the kilograms match, does the mix of the food still change the liver?

Carbohydrate is the nutrient that most strongly drives insulin in most mixed meals. Insulin is one of the signals that tells the liver to store fat and to keep making glucose. Cut the carbohydrate load far enough, and two things can move together: less substrate for the liver to turn into new fat, and a lower insulin signal telling it to do that work. That is a mechanism already in view on our public Evidence page. This trial is a cleaner test of the same idea because the kitchen was controlled and the weight loss was matched.

It does not follow that Mediterranean or plant-forward eating "failed." Both still lowered liver fat by about 45 percent. Both improved muscle insulin sensitivity. Klein was explicit that all three diets improved metabolic health through weight loss. The extra claim the data support is narrower: in this group, with food provided, a very-low-carbohydrate pattern added liver and glucose effects beyond the matched 10 percent weight loss.

To be fair

Hold the limits in the same hand as the percentages.

Forty-two completers. Fourteen people per diet. That is large enough to see a liver-fat gap in a feeding study. It is not large enough to settle what "works" for a country.

The study ran about five months, not five years. We do not know who could keep this pattern in a free-living kitchen, or what the liver looks like after the provided meals stop.

Every calorie was handed to the participant. Weekly dietitian contact. Empty containers counted. That is a strength of the science and a limit of the life lesson. Most people are not living inside a metabolic ward with a packed cooler.

The people studied were adults with obesity, prediabetes, and fatty liver, average age in the low forties and a high body-mass index in the published summaries. If that is not your picture, the size of the liver gap may not be yours either.

"Remission" here means lab criteria for prediabetes. It does not mean a disease was cured. It does not mean medication can be stopped from a blog. It does not mean the next scan will match this sample.

Ketogenic eating is hard for some people to live, socially and practically. A feeding trial cannot measure birthday cake, shift work, or a teenager's pasta night. A pattern you cannot repeat is not a method. It is a temporary experiment.

Funding came from the U.S. National Institutes of Health (including P30 DK56341, P30 DK20579, UL1 TR002345 / KL2 TR002346, and S10 OD027006) and the Foundation for Barnes-Jewish Hospital. Klein reported advisory fees from AbbVie, 89Bio, and Boehringer Ingelheim, an investigator-initiated grant from Merck, and support for an industry-initiated trial from Viking Therapeutics. Those facts belong beside the findings. They do not automatically cancel them. They are why we keep this in the promising-with-limits column, the same column we already use on the Evidence page (https://www.ascentmetaboliccoaching.ca/the-evidence) for supervised carbohydrate restriction.

This paper does not license a supplement stack, a 30-day transformation, or a fight with your family about bread. It also does not say the Mediterranean pattern is useless. It says the liver still noticed carbohydrate load after the scale had been matched.

One thing to do this week

Do not start a ketogenic protocol from this article. That would be a medical-adjacent project, and it is not the assignment.

Do this instead.

Pick one meal you already eat most days. Leave the protein and the vegetables as they are. Drop or shrink the refined starch or sugary drink that usually sits beside them. Keep the rest of the week ordinary.

For seven days, notice two things only: how you feel in the three hours after that meal, and how hungry you are when the next meal arrives. Write it down in a sentence if that helps. No score. No app required.

If fatty liver or prediabetes is already on your chart, take the paper itself to the clinician who owns that chart. The useful question is not "Should I go keto." The useful question is "If I lose some weight, does the pattern of the food change the liver outcome for me?" That decision stays with you and that clinician.

What this will not do

It will not make the scale the only number that matters. It will not make the scale irrelevant either.

It will not tell you which tribe you belong to. It will not replace Canada's food guidance, your pharmacist, or the person who actually reads your labs.

It will not turn five months of provided meals into a personality.

Weight loss helped every group in this trial. Carbohydrate quantity still shaped what happened in the liver. That is the finding, hedged, as the source showed it. Your next step can be as small as one meal you already eat.

If you want a guide who has made this climb and will be there every step of the way, book my free Trailhead Chat. Fifteen minutes. No cost. You leave knowing where you are standing on the trail.

Graham Kingma is a health and wellness coach, not a physician. Nothing here is medical advice. If you have a medical condition or take medication, especially for blood sugar, blood pressure, or liver disease, talk with your doctor before changing how you eat. Results vary. There is no guaranteed outcome.

Sources

Petersen MC, Smith GI, Farabi SS, Palacios HH, Shankaran M, Hellerstein MK, Patterson BW, Klein S. Effect of diet macronutrient content on the cardiometabolic response to weight loss: A randomized clinical trial. Cell Metabolism. Published online 27 August 2026. DOI: 10.1016/j.cmet.2026.07.020. ClinicalTrials.gov: NCT02706262.

WashU Medicine newsroom summary, 27 August 2026: "Low-carb diet delivers added liver benefits beyond weight loss." https://medicine.washu.edu/news/low-carb-diet-delivers-added-liver-benefits-beyond-weight-loss/

Ascent public method page: https://www.ascentmetaboliccoaching.ca/the-evidence