A ketogenic diet did 2 to 3 times more for liver insulin sensitivity at matched weight loss

Every argument about diets runs aground on the same problem. If people on one diet lose more weight than people on another, and they end up healthier, nobody can say whether the food did it or the weight did.

A trial published in Cell Metabolism removed that problem by removing the difference. Three diets with markedly different macronutrient composition were each run until participants had lost about a tenth of their body weight, and only then were the results compared.

At equal weight loss, the diets were not equal. The very-low-carbohydrate one caused a two-to-three-fold greater increase in hepatic insulin sensitivity than the others.

What the liver has to do with blood sugar

The organ in question is not the one most people associate with diabetes.

Between meals the liver releases glucose to keep the blood supplied. Insulin is the signal telling it to stop, and in people heading toward type 2 diabetes that signal stops landing. The liver keeps producing sugar into a bloodstream that already has too much.

Fat inside the organ is part of how that happens. Steatotic liver disease is a condition in which fat builds up in your liver, and it commonly travels with prediabetes, which means your blood glucose levels are higher than normal, but not high enough to be called diabetes.

The people in this trial had both. That combination, described here as metabolically unhealthy obesity, is the group for whom the question actually matters.

Three diets, one target

The comparison was between a very-low-carbohydrate ketogenic diet, a Mediterranean diet, and a very-low-fat plant-forward diet. Very different foods, very different macronutrient splits, one shared instruction: keep going until about 10% of body weight is gone.

Previously, trials of this kind have compared diets that produced unequal weight loss and then tried to adjust for the difference statistically. Building the match into the design instead is what makes this result interpretable.

What was the same and what was not

The muscles did not care which diet delivered the weight loss.

Weight loss increased muscle insulin sensitivity by about 50% in all groups. Muscle is where most glucose ends up after a meal, and losing a tenth of your body weight improved how well it responds to insulin by half, regardless of what was on the plate.

The liver told a different story. Alongside the insulin finding, intrahepatic triglyceride content, hepatic de novo lipogenesis, glycated hemoglobin, and 24-hour serial plasma glucose and insulin decreased most in the very-low-carbohydrate group. In plainer terms: fat stored in the liver, fresh fat being manufactured there, the three-month average of blood sugar, and the glucose and insulin swings across an ordinary day all fell furthest when the carbohydrates went.

That last group of measures is worth pausing on, because it is not a single lab value moving. Four different windows onto the same organ shifted together, in the same direction, in the same group.

The cholesterol question, answered in the negative

The standing objection to eating this way is that replacing carbohydrate with fat should push cholesterol up.

It did not. There were no differences among groups in LDL cholesterol, apoB, or the blood fats measured across a full day. ApoB counts the particles that carry cholesterol around the body, and it predicts risk better than the cholesterol figure most people are given.

A null result is easy to skip past, and this one deserves the opposite. The most common reason a clinician hesitates over a very-low-carbohydrate diet is the fear of what it will do to a lipid panel, and across three diets at the same weight loss, the panels came out level.

What a trial like this cannot tell you

It measured biology, not events. Nobody in this study had a heart attack or a diabetes diagnosis counted as an outcome, and improvements in insulin sensitivity are a strong bet on those outcomes rather than proof of them.

It also ran to a target, not to a horizon. Reaching 10% weight loss is not the same as holding it, and the hardest part of every diet is the part after the trial ends. Nothing here says which of the three is easiest to live with, which is the question that decides most real outcomes.

And the population was chosen for a reason. These were people with fat in the liver and blood sugar already drifting. Someone metabolically healthy has far less to gain and is not represented in this comparison.

What it changes

The advice that has held for years is that the best diet is the one you can stick to, because they all come out the same once the weight comes off. This trial is a careful, specific dent in that.

For a liver already carrying fat, the kind of calories you remove appears to matter as well as the number. The muscles responded to the weight loss alone; the liver responded to the carbohydrates.

That is a narrower claim than the headlines this study will generate, and it is more useful. It does not make ketogenic eating the right answer for everyone, and it does not make the Mediterranean diet a failure, since every group improved. It says that for one organ, in one group of people who are already in trouble, the composition of the diet is doing work of its own.