Yes, in the controlled trials so far, people offered ultra-processed diets ate more than when offered less-processed ones, even with the menus matched on paper. In the best-known trial, 20 adults ate about 508 kcal a day more on the ultra-processed diet1.
Why they ate more is less settled. Below: what was tested, what the extra calories were, and what newer trials added.
What "ultra-processed" means
Most studies use the NOVA system, which sorts foods into four groups by how they are made1. The fourth group, ultra-processed foods, are industrial formulations built mostly from refined ingredients and additives, with little whole food left. Examples named in the 2025 trial paper include breakfast cereals, sweets and mass-produced bread5.
"Ultra-processed" describes how a food is made, not its calories or nutrients. That is why the trials matched the diets for nutrients: to see whether processing matters beyond them.
The 2019 ward trial
Twenty weight-stable adults, 10 men and 10 women with an average age of 31, lived on a research ward at the US National Institutes of Health for 28 days1. A metabolic ward is a hospital research unit where every bite is weighed.
Each person ate two diets for two weeks each, in random order: ultra-processed and unprocessed. This is a crossover design, so everyone is compared with themselves. The menus were matched for the calories, calories per gram, carbohydrate, fat, protein, sugar, salt and fibre presented1. About twice what people needed was served, and they could eat as much or as little as they wanted.
| What | Detail |
|---|---|
| People | 20 adults, 10 men and 10 women, average age 31 |
| Setting | research ward, 28 days |
| Design | randomised crossover, 2 weeks per diet |
| Calories presented | about 5,435 kcal/day on both diets |
| Matched on paper | calories, calories per gram, macronutrients, sugar, salt, fibre |
In plain terms: two menus that look the same on a nutrition label, and a free choice of how much to eat.
What the studies found
About 500 kcal a day more. On the ultra-processed diet, people ate 508 kcal a day more on average (standard error 106)1. The extra was carbohydrate (+280 kcal) and fat (+230 kcal). Protein intake did not change (−2 kcal).
Extra daily intake on the ultra-processed diet, by nutrient · kcal/day
Body weight followed intake. Group-average weight rose 0.9 kg during the ultra-processed fortnight and fell 0.9 kg during the unprocessed one1.
Not tastier, but faster. People rated the meals on both diets as similarly pleasant and familiar1. Hunger and fullness ratings were alike. But they ate faster on the ultra-processed diet: about 17 kcal more per minute.
Denser where it counts. The menus matched on calories per gram only when drinks were included. The ultra-processed diet's solid foods were about 85% denser, and the match came from fibre-added drinks1. The authors say this "likely contributed" to the extra intake.
The result replicated in a small trial. In Tokyo, nine men with overweight ate an ultra-processed and a non-ultra-processed diet for one week each, matched for energy and macronutrients2. They ate about 814 kcal a day more on the ultra-processed week (95% CI 342 to 1,285), and chewed less per calorie.
Extra daily intake on ultra-processed vs less-processed diets · kcal/day
Not an order effect. The 2019 trial had no break between diets. A 2024 reanalysis checked whether diet order skewed the results3. For the ultra-processed trial, order made no significant difference to intake or body weight.
In plain terms: with the menus matched on paper, people still ate more of the ultra-processed diet, mostly from carbohydrate and fat, and they ate it faster.
Why it might happen
The 2019 trial was not designed to find the cause, and its authors say so1. They proposed a few candidates:
- Eating speed. Softer food that is easier to chew and swallow may be eaten faster, before fullness catches up. This is a hypothesis in the 2019 paper.
- Calories per gram. Denser solid foods pack more energy into each bite.
- Protein. A lower share of protein can push intake up. The authors calculated that this could explain at most about half the gap1.
Later trials tested pieces of this. A 2026 trial kept two ultra-processed diets matched for calories per gram and changed only their texture, which set how fast they were eaten6. Intake was about 369 kcal a day lower on the slow-to-eat diet, though body weight did not differ between the two diets over two weeks. So speed alone can move intake, even within ultra-processed food.
A 2025 trial raised protein inside an ultra-processed diet: 30% of energy vs 13%, for 54 hours in a sealed metabolic room4. People ate about 196 kcal a day less on the high-protein version, but still overate on both: intake ran 18% and 32% above the energy they used.
In plain terms: speed and density look like the strongest leads; the label "ultra-processed" bundles several features that each can matter.
What this means in practice
These nutrition trials describe diets, not a verdict on single foods. They show that, on average, a diet mostly of ultra-processed foods led people to eat more when food was unlimited. They do not rank single foods, and flexible approaches such as IIFYM count the same calories and macros either way.
The mechanisms point to features you can see on the plate: how fast a food is eaten and how many calories each gram carries. For the energy-balance side, see calories in, calories out.
The studies at a glance
| Study | Who was studied | Key result | Year |
|---|---|---|---|
| Hall et al., *Cell Metab*1 | 20 adults on a research ward; 2 weeks per diet, crossover | About 508 kcal/day more eaten on ultra-processed; faster eating | 2019 |
| Hamano et al., *Diabetes Obes Metab*2 | 9 men with overweight, Japan; 1 week per diet (abstract) | About 814 kcal/day more on ultra-processed; fewer chews per calorie | 2024 |
| Sciarrillo et al., *Am J Clin Nutr*3 | Reanalysis of the 2019 trial by diet order (abstract) | Diet order did not significantly change the ultra-processed result | 2024 |
| Hägele et al., *Nat Metab*4 | 21 healthy young adults; 54 hours, two ultra-processed diets | High-protein version: about 196 kcal/day less, but overeating on both | 2025 |
| Dicken et al., *Nat Med*5 | 55 adults in England, BMI 25–40; 8 weeks per diet at home | Group-average weight change −2.06% minimally processed vs −1.05% ultra-processed | 2025 |
| Forde et al., *Am J Clin Nutr*6 | 41 healthy adults; two 14-day ultra-processed diets | Slow-to-eat texture: about 369 kcal/day less | 2026 |
Limits worth knowing
- Small and short. The 2019 trial had 20 people for two weeks per diet; the Tokyo trial nine men for one week1.
- Artificial settings. A research ward, with free food chosen by the team, is not everyday life. The authors say generalising is "difficult"1.
- Cause not isolated. The 2019 trial showed people ate more, not which feature of the food caused it.
- Intake not measured at home. The 2025 home trial could not measure energy intake or eating speed directly5.
- Mostly women in the home trial. About 91% of its 55 participants were female5.
- Abstract only for two papers. The 2024 Tokyo trial and the 2024 reanalysis were read as abstracts; the rest in full.
Latest research
The biggest newer trial is from 20255. Fifty-five adults in England, with a body-mass index of 25 to 40, received all their food at home for two eight-week periods: one diet minimally processed, one ultra-processed. Both followed the UK's national healthy-eating guide, and people ate as much as they liked.
Group-average weight fell on both diets: by 2.06% on the minimally processed diet and 1.05% on the ultra-processed one. The gap between diets was about 1 percentage point (95% CI 0.14 to 1.87 points), larger on minimally processed food5. Energy intake itself could not be measured at home, so the trial shows the outcome but not the eating behind it.
Group-average weight change over 8 weeks, by diet · %
The short version
- In a ward trial with menus matched on paper, people ate about 508 kcal a day more on ultra-processed food.
- The extra was carbohydrate and fat, and people ate faster; the meals were not rated tastier.
- Eating speed and calories per gram are the strongest leads for why; each has been shown to move intake.
- A 2025 home trial found a smaller group-average weight change on ultra-processed food when both diets followed guidelines.
Educational overview only — not medical advice.
