Cohort studycited by the Tayebat archivefull text checked★ key study A-UPF-12 in-depth review
Ultra-processed food and inflammatory bowel disease in 21 countries
Association of ultra-processed food intake with risk of inflammatory bowel disease: prospective cohort study.
Overview
Narula N et al.. BMJ 2021 · PMID 34261638 · DOI 10.1136/bmj.n1554
Five or more servings a day: 82% higher IBD risk; legumes, dairy and vegetables showed no link.
Does the study support the specific claim? Yes · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Is UPF intake associated with IBD?
- How was the study done?
Prospective multinational cohort (PURE), country-specific FFQs.
- How many people, animals or samples were studied?
116,087; 467 IBD cases; Adults 35-70
- What was actually found?
>=5 servings/day HR 1.82 (1.22-2.72); 1-4/day 1.67 (1.18-2.37); soft drinks, refined sweetened foods, salty snacks, processed meat each associated; meat, dairy, starch, fruit, vegetables, legumes not associated.
- How large was the observed effect?
HR 1.82 (1.22-2.72)
- What do the results mean in everyday life?
Association; IBD remains rare.
- What are the study's limitations?
Only 467 cases; most diagnoses were self-reported and only about one in five checked against medical records; diet measured once; the higher risk was similar at 1-4 and at 5 or more servings per day.
- What does it say about Elawady's specific claim?
Supports C-UPF-05.
- Do the results apply to Germany and Europe?
Moderate.
- What can explicitly not be concluded from this study?
It cannot prove that ultra-processed food causes IBD or which ingredient might be responsible.
Scientific in-depth review
1. Full reference
- Authors
- Narula N, Wong ECL, Dehghan M, Mente A, Rangarajan S, Lanas F, Lopez-Jaramillo P, Rohatgi P, Lakshmi PVM, Varma RP, Orlandini A, Avezum A, Wielgosz A, Poirier P, Almadi MA, Altuntas Y, Ng KK, Chifamba J, Yeates K, Puoane T, Khatib R, Yusuf R, Boström KB, Zatonska K, Iqbal R, Weida L, Yibing Z, Sidong L, Dans A, Yusufali A, Mohammadifard N, Marshall JK, Moayyedi P, Reinisch W, Yusuf S
- Journal
- BMJ
- Year
- 2021
- Identifier
- PMID 34261638 DOI 10.1136/bmj.n1554 · PMC8279036
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
PURE funded by Population Health Research Institute, CIHR, Heart and Stroke Foundation, Ontario Ministry of Health and unrestricted grants from pharmaceutical companies (AstraZeneca, Sanofi-Aventis, Boehringer Ingelheim, Servier, GSK, Novartis, King Pharma) and national bodies (including Unilever Health Institute Brazil and the South Africa Sugar Association); funders had no input. ICMJE: no competing interests declared.
2. Research question
Is UPF intake associated with IBD?
3. Study design and population
- Study design
- Cohort study – Prospective multinational cohort (PURE) with a single baseline country-specific validated FFQ; ultra-processed food defined as a list of packaged/formulated food groups (processed meat, breakfast cereal, sauces, soft drinks, refined sweetened foods, chips, ice cream, pastries, biscuits, fruit drinks), not a full item-level NOVA classification.
- Population
Adults aged 35-70 from 21 low-, middle- and high-income countries (incl. Iran, Palestine, Saudi Arabia, UAE, Turkey; no Western European country except Sweden and Poland)
- Country
21 countries, 7 regions incl. Middle East
- Period
Enrolment 2003-2016; outcomes to 5 July 2019; median follow-up 9.7 years (IQR 8.9-11.2); IBD diagnoses recorded systematically only after a 2014 protocol amendment
- Sample size
116,087 in abstract (116,037 in results text); 467 incident IBD (90 Crohn's disease, 377 ulcerative colitis); the main table shows 428 events among 113,610 participants
- Inclusion and exclusion criteria
Plausible energy intake (500-5000 kcal/day), >=1 follow-up cycle, complete FFQ (10,625 excluded for incomplete FFQ); prevalent IBD and diagnoses within 1 year of the FFQ excluded.
4. Methods and endpoints
- Exposure / intervention
Ultra-processed food servings/day (<1, 1-4, >=5) from baseline FFQ; also g/day and subgroups
- Comparator
<1 serving/day
- Endpoints
Self-reported incident IBD (Crohn's disease, ulcerative colitis); about 20% of diagnoses validated against medical records
- Statistical methods
Cox models: unadjusted, minimally adjusted (age, sex, region) and fully adjusted (plus education, alcohol, smoking, BMI, energy, urban/rural; covariates selected partly by univariable P<0.15); additional AHEI adjustment; multiple imputation; heterogeneity between CD and UC and across regions.
5. Actual results
- Results
IBD, fully adjusted vs <1 serving/day: 1-4 servings HR 1.67 (1.18-2.37); >=5 servings HR 1.82 (1.22-2.72); P trend 0.006 (minimally adjusted 1.41 and 1.42). With AHEI: 1.75 and 1.92. Crohn's disease >=5/day HR 4.50 (1.67-12.13, 30 events); ulcerative colitis >=5/day HR 1.46 (0.93-2.28), not significant. Subgroups: processed meat >=1/day 2.07 (1.14-3.76); soft drinks >=3/week 1.94 (1.42-2.66); refined sweetened foods >=100 g/day 2.58 (1.44-4.62); salty snacks >=100 g/day 2.06 (1.41-3.00); fried food >=1/day 3.02 (1.51-6.03). Urinary sodium not associated. White meat, unprocessed red meat, dairy, starch, fruit/vegetables/legumes not associated.
- Absolute risks
Cumulative incidence over ~9.7 years: 0.26% (<1 serving/day; 199/76,415), 0.53% (1-4/day; 134/25,453), 0.81% (>=5/day; 95/11,742) - crude, unadjusted.
- Authors' conclusion
Higher ultra-processed food intake was positively associated with IBD risk; it might be the processing rather than the food itself; residual confounding possible; further studies needed to identify contributory factors.
6. Strengths
- Large multinational prospective cohort including Middle Eastern countries
- Diet assessed before diagnosis; exclusion of diagnoses within 1 year of FFQ
- Consistency across regions (I2=0%)
7. Weaknesses
- Few cases (467), especially Crohn's disease (90): imprecise estimates
- Self-reported IBD; only ~20% validated; systematic recording started only in 2014 (authors)
- Single baseline FFQ; not designed for NOVA; UPF defined by selected food groups
- No clear dose-response between 1-4 and >=5 servings/day; ulcerative colitis (most cases) not significant at >=5/day
- Full adjustment increased HRs versus minimal adjustment; data-driven covariate selection
- Multiple comparisons across many food groups (authors)
- Residual confounding (e.g. early-life antibiotics) acknowledged
- Small inconsistencies in reported numbers (116,087 vs 116,037 participants; 467 vs 428 events in the main table)
8. Risk of bias
- Orienting tool
- ROBINS-E
- 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
Outcome measurement judged high risk; overall kept at some_concerns because misclassification is likely partly non-differential and results were consistent across regions, but the overall could reasonably be rated high. Structured single-reviewer judgement from the full text; a complete formal ROBINS-E assessment requires two independent reviewers and a protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | some concerns | Adjusted for main socio-demographic and lifestyle factors and AHEI; covariates partly selected by univariable P values; residual confounding acknowledged. |
| Exposure measurement | some concerns | Single baseline FFQ; UPF defined by a list of food groups; FFQ suited to relative ranking only (authors). |
| Selection of participants | some concerns | 8.3% excluded for incomplete FFQ (multiple imputation similar); outcome recording only after 2014 may miss early events. |
| Post-exposure interventions | some concerns | Dietary westernisation during follow-up not captured (authors). |
| Missing data | low | Multiple imputation gave similar results. |
| Outcome measurement | high | Self-reported IBD with validation of only ~20%; misclassification plausible and could be related to health-care access, which differs by region and UPF intake. |
| Selection of reported result | some concerns | Many exposures tested; number inconsistencies between abstract, text and table. |
Full-text review (25 October 2026): Headline HRs confirmed. Added: UPF defined by food-group list (not item-level NOVA), self-reported IBD with ~20% validation and recording only from 2014, crude incidence 0.26%/0.53%/0.81%, CD vs UC estimates (UC >=5/day non-significant), subgroup HRs, minimally vs fully adjusted HRs, pharma/industry funding of PURE, numeric inconsistencies. Outcome-measurement domain rated high; overall some_concerns (unchanged); precision lowered to low.
9. Assessment of the statistical interpretation
Moderate relative associations with wide CIs and small absolute risks (<1% over ~10 years). The plateau between 1-4 and >=5 servings/day and the null fully adjusted ulcerative colitis estimate weaken the dose-response argument; the Crohn's estimate rests on 30 events in the top group.
10. Relevance to the specific claim
Supports C-UPF-05.
11. Independent contrary findings
None.
12. Transferability to Germany and Europe
Moderate (few European sites; includes Middle Eastern countries relevant to Egypt).
13. Overall methodological judgement
Relevant multinational signal linking UPF intake to IBD, but based on few, mostly unvalidated self-reported cases and a crude UPF definition; hypothesis-supporting rather than confirmatory.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | low |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | unknown |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | correctly reported |
14. Open questions
Validation with clinically confirmed cases; mechanisms (e.g. emulsifiers); replication with item-level NOVA assessment.
15. Linked fact checks
- Higher ultra-processed food intake raises the risk of inflammatory bowel disease (PURE study: HR 1.82 for >=5 servings/day). – Studies supporting 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
PURE across 21 countries (116,087 adults) — higher UPF intake associated with greater IBD risk (HR 1.82 for >=5 servings/day).
- A. What was actually studied?
UPF and IBD.
- B. What did the authors find?
As above.
- C. What conclusions do the authors themselves draw?
As above.
- D. What does the archive derive from it?
Association.
- E. Is this derivation scientifically justified?
Correctly reported as an association. Not mentioned: few cases, self-reported diagnoses with ~20% validation, and that the ulcerative colitis association was not significant at >=5 servings/day.
- Adequacy of the interpretation (by the archive)
- correctly reported