Ultra-processed foods and overeating through the trial evidence

Weight loss also occurred on an ultra-processed diet

Dicken and colleagues compared diets arranged around UK dietary guidelines but differing in food processing. The trial enrolled 55 adults in England with BMI of at least 25 and below 40. They normally obtained at least half of their dietary energy from ultra-processed foods. Participants were randomly assigned the order in which they received an ultra-processed diet and a minimally processed diet, each provided for 8 weeks.[5]

The analysis included 50 participants. The primary outcome was the difference between the diets in each participant’s percentage weight change. Average weight loss was 2.06% on the minimally processed diet, with a 95% confidence interval corresponding to a loss of 1.13–2.99%. On the ultra-processed diet, average weight loss was 1.05%, with a 95% confidence interval corresponding to a loss of 0.13–1.98%.[5] Participants lost weight under both dietary conditions.

The minimally processed diet produced 1.01 percentage points more weight loss, with a 95% confidence interval of 0.14–1.87 percentage points.[5] This compares percentages of weight change. It is not a difference of 1.01 kg, and it does not mean participants gained weight on the ultra-processed diet. Keeping the unit attached to the number changes how the result reads: there was weight loss in both conditions, with a larger reduction in the minimally processed condition.

This may appear to conflict with the inpatient study if the name of the processing category is treated as a complete description of the diet. Hall’s trial compared particular meal combinations over short periods. Dicken’s trial provided diets following dietary guidelines over longer periods.[1][5] The shared label does not mean that the meal designs were identical. Comparing the studies makes overall dietary composition important to the interpretation, alongside the extent of processing.

It would also be premature to conclude that processing becomes irrelevant once a diet follows nutritional guidance. There was still a difference in the amount of weight lost between the two guideline-based diets.[5] The findings support considering both the overall arrangement of meals and opportunities to choose less-processed foods. They do not support the claim that every ultra-processed meal prevents weight loss, or the claim that there was no difference between the trial conditions.

Mild gastrointestinal symptoms were common on both diets.[5] This is another reason to keep the full study context rather than ranking diets using only the main weight result. What people were given, who took part and what happened during the intervention all belong in the interpretation. The trial’s central comparison was weight change, but that does not make the other reported experiences disappear. A smaller weight number alone cannot describe every aspect of a dietary condition.

A separate study by Hägele and colleagues compared two ultra-processed diets with different nutrient proportions. Twenty-one healthy young adults received a higher-protein, lower-carbohydrate diet and a normal-protein diet for 54 hours each. The meals were matched for calories, fat, fiber and palatability, and participants could eat as much as they wanted.[10] This comparison asks about differences within ultra-processed diets, rather than replacing the category with unprocessed foods.

Intake was 196 ± 396 kcal per day lower on the higher-protein diet, while energy expenditure was 128 ± 98 kcal per day higher. Nevertheless, overeating remained.[10] The higher-protein condition supplied 30% of energy from protein and 29% from carbohydrate; the normal condition supplied 13% and 46%, respectively. The diet changed both protein and carbohydrate proportions, so describing the intervention solely as an added protein label would leave out part of the comparison.

Energy balance, the difference between intake and expenditure, was positive under both conditions. Intake exceeded expenditure by 18% in the higher-protein condition and 32% in the normal condition.[10] A reduction in intake relative to another diet is therefore different from eliminating excess intake relative to expenditure. The higher-protein diet reduced the excess without eliminating it.

Long-term associations and causal trials answer different questions

Observational research helps examine whether dietary patterns are associated with weight beyond the short periods used in feeding trials. Askari and colleagues conducted a systematic review and meta-analysis, collecting studies by a defined method and combining their results. Of the 14 included studies, 13 were cross-sectional studies measuring diet and weight at a point in time. Only 1 was a cohort study following people over time.[8]

The pooled effect size for obesity was 1.26, with a 95% confidence interval of 1.13–1.41, indicating a positive association with ultra-processed food consumption.[8] This is a combined measure of association across studies, rather than a predicted weight change for an individual. The authors also called for studies that follow people over time and adequately address other factors to clarify causality. The type of evidence remains relevant even when the combined result is statistically significant.

A larger source of longer-term data is Córdova and colleagues’ cohort of 348,748 European men and women aged 25–70. The median interval between weight measurements was 5 years. Researchers assessed dietary habits using questionnaires, classified food by processing and statistically adjusted for other factors before examining the association with subsequent weight.[16] Unlike a feeding trial, this study observed what people normally ate rather than randomly assigning their diets.

Each increase of 1 standard deviation in ultra-processed food consumption was associated with 0.12 kg greater weight gain over 5 years, with a 95% confidence interval of 0.09–0.15 kg.[16] A standard deviation is a unit describing variation within a group. It does not mean one packet, serving or meal. Without that distinction, the number can easily be attached to a familiar food quantity that the study did not actually use as its comparison unit.

Comparing the highest-consuming group with the lowest-consuming group, the relative risk of becoming overweight or obese was 15% higher among participants initially at normal weight. Among those already overweight, the risk of becoming obese was 16% higher.[16] These are relative differences between groups, and the outcomes depended on starting body size. They are not statements that every person eating more ultra-processed foods will gain a fixed amount of weight.

Large numbers and longer follow-up make the study valuable, but they do not turn an observed diet into a randomly assigned intervention. Statistical adjustment can address measured differences while leaving other factors incompletely accounted for.[8][16] Cohorts describe longer-term associations; controlled trials examine responses when conditions are changed. Understanding the method, in addition to the sample size, helps decide what a result can contribute to a claim about cause.

The limitations also appear in reviews of randomized evidence. Aramburu and colleagues assessed educational interventions and a controlled feeding trial involving a total of 455 participants, with a median follow-up of 12 weeks. No significant effects were found for 30 of the 42 outcomes evaluated.[12] Some outcomes, including weight and intake, showed benefits, but the limited number of studies and methodological weaknesses prevented definitive conclusions. Counting all outcomes matters when evaluating the overall evidence.

Louie’s perspective examines interpretation rather than reporting a new experiment. It argues that even direct comparisons of ultra-processed and minimally processed diets can change several dietary features at once, making processing itself difficult to isolate as the cause. It also highlights diversity within the category and the tendency to emphasize findings that fit expectations.[11] This is a prompt to examine how results are used, rather than a reason to discard every measured difference.

Taken together, these studies support examining food processing without treating it as a complete explanation for every result. Intake differences in controlled comparisons, weight loss on guideline-based diets, variation in food texture and uncertainty about longer-term causality all belong in the same account.[1][5][9][12] For everyday decisions, the evidence is most useful when it guides a closer look at meals. It is less useful when a single result becomes a verdict on every food or a forecast of personal weight.

Common misconceptions

The idea that all processed foods undermine weight management is too broad. Food measurements showed variation in energy intake rate within each processing category, and participants lost weight on an ultra-processed diet designed around dietary guidelines.[5][9] A group average cannot provide a definitive assessment of every product or every person’s dinner. It is more useful to consider the combination of foods, their nutritional arrangement and how they are eaten.

The claim that matching nutrition labels prevents overeating also conflicts with the inpatient trial. Presented calories and nutritional characteristics were matched, but freely chosen intake differed.[1] Labels describe the foods; they cannot specify how much additional food someone will eat afterward. Looking at a label remains a way to understand what is being offered, while observing portions eaten and additional food addresses a different part of the meal.

A daily intake difference cannot simply be multiplied into a forecast of future weight gain. The trials measured responses over their particular periods and under their particular conditions.[1][4] Another meal design produced weight loss in both processing conditions.[5] Short-term differences can justify revisiting food choices, but these studies do not establish a long-term personal prediction. Keeping the duration and comparator alongside each number avoids asking the evidence to answer a different question.

What you can do today

• Use the 2-week unprocessed-diet comparison as a starting point for replacing your usual main dish with a less-processed option, while planning the staple and sides together. Buy the replacement ingredients when shopping and choose a combination you can prepare again the following day.[1] The 2 weeks describe the study period, rather than a deadline for achieving weight loss at home.
• At your next shop, choose a meal combination with something that requires chewing rather than relying entirely on foods that are easy to swallow quickly. The texture comparison ran for 14 days and found lower intake with slower-eating textures.[15] That duration describes the experiment; it does not guarantee the same intake reduction in your own meals.
• Try a meal combination with more protein in the main dish, and review whether you have been leaving that dish out and filling its place with other foods. In the ultra-processed diet comparison, changing protein’s share of energy from 13% to 30%, while also lowering carbohydrate, reduced intake but did not prevent overeating.[10] Use the finding to reconsider the meal’s composition rather than adopting those percentages as personal targets.
• If extra food after meals concerns you, record what the meal contained and the foods and quantities you added. Separate what was prepared from what was actually eaten. Apply the same observation to meals selected for a high-protein label, rather than assuming the label answers the question of how much you consumed.[1][10]

Beyond food labels, keep moving at On the Shore Tachikawa

A processing label leaves out details of the meal; an activity name also leaves room to explore the lesson itself. On the Shore Tachikawa offers a place to choose a lesson and keep an exercise habit going alongside attention to everyday meals.

The studio is a 1-minute walk from the North Exit of JR Tachikawa Station. Its address is 3F Etoile Bldg, 2-14-10 Akebonocho, Tachikawa, Tokyo, and it is open every day 8:00–23:30. Many of our yoga instructors speak English, and the studio’s owner often helps guests from the US bases in English herself.

Pregnant guests cannot join lava-stone hot yoga; room-temperature maternity yoga is available. If a doctor has told you not to exercise, you cannot participate at the studio.

The lesson options include lava-stone hot yoga and room-temperature yoga, with more than 25 kinds of yoga lessons. Pilates, personal training, HIIT, boxercise and women-only kickboxercise are also available.

For a first visit, the yoga trial is a 60-minute lesson costing ¥1,980 including tax. Rental includes a yoga mat, 1 face towel and 2 bath towels. The prices page provides pricing information, and the trial booking page is where you can book the trial.

References

1. Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake — K. Hall et al., 2019, Cell metabolism. DOI: 10.1016/j.cmet.2019.05.020
2. Ultra‐processed foods cause weight gain and increased energy intake associated with reduced chewing frequency: A randomized, open‐label, crossover study — Shoko Hamano et al., 2024, Diabetes. DOI: 10.1111/dom.15922
3. Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial — S. Dicken et al., 2025, Nature Medicine. DOI: 10.1038/s41591-025-03842-0
4. Ultra-processed food and the risk of overweight and obesity: a systematic review and meta-analysis of observational studies — Mohammadreza Askari et al., 2020, International Journal of Obesity. DOI: 10.1038/s41366-020-00650-z
5. Ultra-Processing or Oral Processing? A Role for Energy Density and Eating Rate in Moderating Energy Intake from Processed Foods — C. Forde et al., 2020, Current Developments in Nutrition. DOI: 10.1093/cdn/nzaa019
6. Short-term effects of high-protein, lower-carbohydrate ultra-processed foods on human energy balance — Franziska A. Hägele et al., 2025, Nature Metabolism. DOI: 10.1038/s42255-025-01247-4
7. Interpreting ultraprocessed food trial evidence: evidentiary limits, reporting practices, and policy relevance — J. C. Y. Louie, 2026, The American Journal of Clinical Nutrition. DOI: 10.1016/j.ajcnut.2026.101395
8. Ultra-processed foods consumption and health-related outcomes: a systematic review of randomized controlled trials — Adolfo Aramburu et al., 2024, Frontiers in Nutrition. DOI: 10.3389/fnut.2024.1421728
9. Eating rate has sustained effects on energy intake from ultraprocessed diets: a 2-week ad libitum dietary randomized controlled crossover trial — C. Forde et al., 2025, The American Journal of Clinical Nutrition. DOI: 10.1016/j.ajcnut.2025.11.012
10. Consumption of ultra-processed foods associated with weight gain and obesity in adults: A multi-national cohort study. — R. Córdova et al., 2021, Clinical nutrition. DOI: 10.1016/j.clnu.2021.08.009

Cover photo: Eggplants, radishes, carrots and other fresh vegetables and fruit arranged at a market stall. (Photo: OnlyJanz / CC BY-SA 4.0 / Wikimedia Commons)

Information as of October 2026. For your own health, consult a doctor.

More from Diet Science (English)

• Metabolic adaptation and why your weight loss can slow down
• Body composition and weight loss beyond the bathroom scale
• Food logging and weight loss: what matters beyond perfect records
• Protein and weight loss beyond fullness and the bathroom scale
• Dietary fiber and weight loss: what gut research really shows
• Low-carb vs low-fat diets: what long-term weight loss shows
• Time-restricted eating and weight loss: what trials reveal
• Breakfast and weight loss: what eating or skipping really changes
• Sleep deprivation and appetite: why nighttime snacks add up
• Stress and emotional eating beyond the cortisol explanation
• Aerobic vs resistance training for fat loss and lean mass
• HIIT and body fat loss: what shorter workouts can really offer
• Strength training, metabolism and the science of weight regain
• Yoga and weight loss: what body composition research shows
• Hot yoga calories and sweat: what a lighter scale really means
• Pilates and body composition: what posture and fat studies show
• NEAT and weight loss: what everyday movement can really do
• Daily steps and weight loss: what 8,000 steps really tells us
• Keeping weight off: what long-term maintainers do and how they recover
• Daily self-weighing and weight loss: how to use the numbers
• Water before meals and weight loss: what 500 ml trials show
• Sugary drinks and sweeteners: what replacement means for weight
• Alcohol and weight loss through calories, appetite and fat use
• Eating rate and weight loss: what slower chewing can change
• Menopause weight gain and why your waist can change first
• Postpartum weight loss through diet, exercise and support
• Sarcopenia and protein: protecting muscle as you grow older
• Visceral fat in men and what exercise changes beyond weight
• Supervised exercise and weight loss: building lasting habits

All Diet Science articles (English)