Randomised controlled trialcited by the Tayebat archivefull text checked★ key study A-UPF-01 in-depth review
Ultra-processed vs unprocessed diet in hospital
Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake.
Overview
Hall KD et al.. Cell Metab 2019 · PMID 31105044 · DOI 10.1016/j.cmet.2019.05.008
People ate about 500 calories more per day and gained weight on ultra-processed food.
Does the study support the specific claim? Yes · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Do ultra-processed foods affect energy intake?
- How was the study done?
Inpatient randomized crossover trial, 2 weeks per diet, ad libitum.
- How many people, animals or samples were studied?
20; Weight-stable adults, mean age 31, BMI 27
- What was actually found?
Energy intake +508 ± 106 kcal/day on UPF (p=0.0001); carbohydrate +280, fat +230 kcal/day; protein unchanged; weight +0.9 ± 0.3 kg (UPF) vs -0.9 ± 0.3 kg (unprocessed).
- How large was the observed effect?
+508 kcal/day; +0.9 kg vs -0.9 kg
- What do the results mean in everyday life?
Ultra-processed food can make people eat more.
- What are the study's limitations?
Only 20 people, 2 weeks per diet, in hospital conditions; no washout between diets; participants knew which diet they ate; the cause of the extra eating (e.g. energy-dense food, faster eating) is not clear.
- What does it say about Elawady's specific claim?
Core evidence for C-UPF-02.
- Do the results apply to Germany and Europe?
High.
- What can explicitly not be concluded from this study?
Long-term health outcomes.
Scientific in-depth review
1. Full reference
- Authors
- Hall KD, Ayuketah A, Brychta R, Cai H, Cassimatis T, Chen KY, Chung ST, Costa E, Courville A, Darcey V, Fletcher LA, Forde CG, Gharib AM, Guo J, Howard R, Joseph PV, McGehee S, Ouwerkerk R, Raisinger K, Rozga I, Stagliano M, Walter M, Walter PJ, Yang S, Zhou M
- Journal
- Cell Metab
- Year
- 2019
- Identifier
- PMID 31105044 DOI 10.1016/j.cmet.2019.05.008 · PMC7946062
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
PubMed lists: ErratumIn (Cell Metab. 2019 Jul 2;30(1):226. doi: 10.1016/j.cmet.2019.05.020.), ErratumIn (Cell Metab. 2020 Oct 6;32(4):690. doi: 10.1016/j.cmet.2020.08.014.)
- Extent checked
- full text
- Conflicts of interest and funding
NIH Intramural Research Program (NIDDK). C.G. Forde: conference reimbursements from companies selling nutritional products, scientific advisory council of Kerry Taste and Nutrition, academic consortium funded by Abbott Nutrition, Nestec and Danone; other authors none.
2. Research question
Do ultra-processed foods affect energy intake?
3. Study design and population
- Study design
- Randomised controlled trial – Randomised crossover inpatient feeding trial (NCT03407053): 28 continuous days on a metabolic ward, 2 weeks ultra-processed then 2 weeks unprocessed diet or the reverse, without run-in or washout; ad libitum meals and snacks.
- Population
10 men and 10 women, weight-stable, age 31.2±1.6 y (mean±SE), BMI 27±1.5 kg/m2
- Country
USA (NIH Clinical Center)
- Period
Not stated in full text (registered 2018; published 2019)
- Sample size
20 randomised; 20 completed
- Inclusion and exclusion criteria
Age 18-50, BMI >18.5, weight-stable (<±5% over 6 months); excluded anaemia, diabetes, cancer, thyroid disease, eating disorders or other psychiatric conditions, food allergies or special diets (e.g. vegetarian, vegan, kosher).
4. Methods and endpoints
- Exposure / intervention
Ultra-processed diet (83.5% of energy from NOVA group 4 in meals) offered at twice estimated energy requirements
- Comparator
Unprocessed diet (0% ultra-processed) matched for presented energy, energy density (including beverages), macronutrients, sugars, sodium and fibre; not matched for non-beverage energy density (1.957 vs 1.057 kcal/g in meals), added-to-total sugar (~54% vs 1%), saturated fat, omega-6:omega-3 ratio; fibre in the UPF diet partly added as dissolved supplements in beverages
- Endpoints
Primary: mean ad libitum energy intake over each 14-day period; secondary: body weight and composition (DXA), energy expenditure (respiratory chamber, doubly labelled water), appetite ratings, eating rate, glucose tolerance, CGM, liver fat, fasting hormones and lipids
- Statistical methods
ANOVA (SAS PROC GLM), least-squares means ±SE, two-sided t-tests; powered (0.8) to detect 125-150 kcal/d difference with n=20.
5. Actual results
- Results
Energy intake +508±106 kcal/d on UPF (p=0.0001); final-week difference +459±105 kcal/d (p=0.0003); carbohydrate +280±54, fat +230±53, protein -2±12 kcal/d. No order or sex effect. Weight +0.9±0.3 kg (UPF, p=0.009) vs -0.9±0.3 kg (unprocessed, p=0.007); weight change correlated with energy intake difference (r=0.8). Body fat +0.4±0.1 kg vs -0.3±0.1 kg. Eating rate higher on UPF (+17±1 kcal/min). Pleasantness, familiarity, hunger and fullness ratings did not differ. Sodium intake 5.8 vs 4.6 g/d; total sugars 93.3 vs 96.6 g/d and fibre 48.5 vs 45.8 g/d (not different). No differences in glucose tolerance, Matsuda index (3.9 vs 4.5, p=0.1) or liver fat. PYY higher and ghrelin lower on the unprocessed diet.
- Absolute risks
not applicable (short-term physiological trial; absolute differences reported: +508 kcal/d, +0.9 vs -0.9 kg over 2 weeks)
- Authors' conclusion
Eliminating ultra-processed foods decreases energy intake and results in weight loss, whereas a diet high in ultra-processed food increases energy intake and leads to weight gain; limiting ultra-processed foods may be an effective strategy for obesity prevention and treatment. The study was not designed to identify the cause of the intake difference; results may not generalise to free-living conditions.
6. Strengths
- Randomised crossover with each participant as own control
- Metabolic-ward control with weighed intake; objective energy expenditure (DLW, chambers)
- Pre-registered; no dropouts
- Diets matched on several presented nutrients
7. Weaknesses
- Small (n=20) and short (2 weeks per diet)
- No run-in or washout period; possible carry-over (authors)
- Participants and staff not blinded to diet
- Non-beverage energy density ~85% higher and eating rate faster on UPF diet; protein lower (14% vs 15.6%); fibre added via beverages - mechanism unresolved
- Inpatient setting, free food, no cost/convenience factors: limited generalisability (authors)
- Body composition and energy balance estimates imprecise (authors)
8. Risk of bias
- Orienting tool
- RoB 2 (crossover trial)
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Structured single-reviewer judgement from the full text; a complete formal RoB 2 assessment requires two independent reviewers and the trial protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| D1 Randomisation process | some concerns | Randomised diet order; method of sequence generation and concealment not described in the full text; order had no effect on the outcome (p=0.64). |
| D2 Deviations from intended interventions (incl. carry-over, period effects) | some concerns | No washout; carry-over possible, partly addressed by final-week analysis; open-label feeding but intervention delivered entirely by ward staff; participants told it was not a weight-loss study. |
| D3 Missing outcome data | low | All 20 completed; one DLW measurement failed QC. |
| D4 Measurement of the outcome | low | Energy intake by weighing leftovers; objective, though nutrient databases estimate metabolizable energy. |
| D5 Selection of the reported result | low | Registered primary endpoint (ad libitum energy intake) reported as planned. |
Full-text review (25 October 2026): Numbers confirmed. Added eligibility criteria, registration (NCT03407053), no run-in/washout, unmatched non-beverage energy density (1.957 vs 1.057 kcal/g), eating rate, sodium, no glucose-tolerance or liver-fat differences, correlation of weight change with energy intake (r=0.8), power calculation, full COI. Domain-level RoB 2 (overall some_concerns, unchanged).
9. Assessment of the statistical interpretation
Large, precise within-person effect on energy intake (well above the 125-150 kcal/d the study was powered for). Weight results are short-term and partly fluid shifts (sodium intake correlated with fat-free-mass change). The trial shows UPF diets as delivered increase intake; it does not show which property (energy density, eating rate, protein, sodium) is responsible.
10. Relevance to the specific claim
Core evidence for C-UPF-02.
11. Independent contrary findings
None contradicting.
12. Transferability to Germany and Europe
High.
13. Overall methodological judgement
Landmark, well-controlled proof-of-concept trial showing that, under ward conditions, an ultra-processed diet increased energy intake by ~500 kcal/d and caused short-term weight gain, with weight change closely tracking energy intake. Small, short and unblinded; mechanism unresolved.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | essentially correct but incomplete |
14. Open questions
Long-term effects; mechanism.
15. Linked fact checks
- Ultra-processed foods drive overeating and weight gain: in an NIH randomized trial, people ate about 500 kcal more per day and gained 0.9 kg in two weeks on an ultra-processed diet. – Studies supporting the claim
- Calories are not a relevant concept ('there is no such thing as calories'); his system is based on high calorie intake, and people lose weight while eating sugar, juices, fats, potatoes and rice, because with 'zero insulin' no fat is formed (the passage continues: 'even if you drink oil, even if you swallow bags of sugar'). Elsewhere he says he drinks 6 litres of sugar-cane juice and eats baladi ghee, rice and potatoes every day. – Studies contradicting or limiting the claim
Use by the Tayebat archive
Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.
- What the archive says about the study
NIH inpatient RCT — an ultra-processed diet caused ~500 extra kcal/day and weight gain versus a matched unprocessed diet.
- A. What was actually studied?
Ad libitum energy intake on UPF vs unprocessed diets.
- B. What did the authors find?
+508 kcal/day; +0.9 vs -0.9 kg.
- C. What conclusions do the authors themselves draw?
Limiting UPF may help prevent obesity.
- D. What does the archive derive from it?
Called 'the single most important piece of evidence' and 'decisive'.
- E. Is this derivation scientifically justified?
Results reported correctly; 'decisive' overstates a 20-person, 4-week trial. The trial also shows that weight change tracked energy intake (r=0.8) and that total sugar intake was similar on both diets, which is relevant to claims that calories do not matter.
- Adequacy of the interpretation (by the archive)
- essentially correct but incomplete