Five hundred calories nobody chose

Nutrition research is mostly questionnaires and regret. This one is not.

In 2019 Kevin Hall and colleagues published a trial in Cell Metabolism that did something the field almost never manages: it admitted people to a hospital ward, controlled everything they ate, and randomised the order.

Twenty weight-stable adults lived at the NIH Clinical Center. Each spent two weeks on an ultra-processed diet and two weeks on an unprocessed one, in randomised order. They could eat as much as they wanted at every meal.

The part that makes it matter

The two diets were matched — for presented calories, energy density, sugar, fat, sodium, fibre and the whole macronutrient profile. On paper, on a nutrition label, they were the same food.

On the ultra-processed diet, people ate about 500 kcal a day more. They gained weight. On the unprocessed one they lost it. Nobody was told to eat more or less. Nobody was on a diet.

Same calories offered. Same sugar. Same fibre. Five hundred calories a day of difference in what people chose to eat.

This is the first randomised controlled trial to show a causal link between ultra-processed food and weight gain in humans. Before it, everything we had was people remembering what they ate.

What it does not prove

Twenty people is twenty people. Four weeks is four weeks. And it cannot tell us why — whether it is the softness, the speed of eating, the texture, the absence of anything that requires chewing, or something about the formulation nobody has isolated yet.

It also does not draw a clean line around the word "ultra-processed". The classification is contested, and it lumps a supermarket sandwich in with a fizzy drink. That is a real weakness in the field, not a detail.

What we do in the kitchen

What we take from it is narrow and, we think, safe:

• The label is not the food. Two things can match on every declared number and behave completely differently in a body. That is the finding.
• Effort is a feature. Food that takes chewing and time is food you eat less of, without deciding to.
• Not a moral category. There is no virtue in this and no failure in a packet of biscuits. It is a system that makes eating more the easy path.
• Start with what you eat most. Breakfast, five days a week, is worth more than an immaculate Sunday.

It pairs with the other piece in this kitchen: Stanford's fermented-food trial found something measurable in ten weeks on ordinary groceries. Between them, they say the same quiet thing — the strongest food findings we have are about what is on the plate, not what is in a bottle.