Cohort studycited by the Tayebat archivefull text checked★ key study A-WHT-14 in-depth review
Cereal fibre and diabetes in Europe
Dietary fibre and incidence of type 2 diabetes in eight European countries: the EPIC-InterAct Study and a meta-analysis of prospective studies.
Overview
InterAct Consortium. Diabetologia 2015 · PMID 26021487 · DOI 10.1007/s00125-015-3585-9
Each extra 10 g/day of cereal fibre linked to 25% lower diabetes risk.
Does the study support the specific claim? Partly · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Fibre and diabetes
- How was the study done?
Case-cohort + meta-analysis
- How many people, animals or samples were studied?
11,559 cases
- What was actually found?
Cereal fibre protective
- How large was the observed effect?
RR 0.75/10 g
- What do the results mean in everyday life?
Cereal fibre helpful
- What are the study's limitations?
Observational; European data not significant after accounting for body weight; 25% estimate from heterogeneous meta-analysis
- What does it say about Elawady's specific claim?
Supports fibre, not glue mechanism
- Do the results apply to Germany and Europe?
High
- What can explicitly not be concluded from this study?
Cause
Scientific in-depth review
1. Full reference
- Authors
- InterAct Consortium
- Journal
- Diabetologia
- Year
- 2015
- Identifier
- PMID 26021487 DOI 10.1007/s00125-015-3585-9 · PMC4472947
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
InterAct funded by EU FP6 (LSHM-CT-2006-037197); individual authors supported by national cancer charities, research councils, ministries, WCRF, MRC; one author (PWF) also supported by Novo Nordisk among other sources. Authors declare no duality of interest.
2. Research question
Fibre types and T2D
3. Study design and population
- Study design
- Cohort study – Nested case-cohort study within EPIC (EPIC-InterAct, 8 countries) plus dose-response meta-analysis of 19 prospective cohorts (18 published + EPIC-InterAct)
- Population
EPIC-InterAct: 340,234 eligible EPIC participants; analysed 26,088 (11,559 incident T2D cases, subcohort 15,258 incl. 729 cases), mean subcohort age 52.4 years. Meta-analysis: 617,968 participants, 41,066 cases (8 US, 5 European, 3 Australian, 3 Asian cohorts)
- Country
Denmark, France, Germany, Italy, Netherlands, Spain, Sweden, UK (EPIC-InterAct); meta-analysis adds USA, Australia, Finland, Japan, China, Taiwan
- Period
EPIC recruitment 1990s, follow-up to 31 Dec 2007 (median 10.8 years); meta-analysis PubMed search to 24 January 2014
- Sample size
EPIC-InterAct 26,088 (11,559 cases; subcohort 15,258); meta-analysis 19 cohorts, 617,968 participants, 41,066 cases
- Inclusion and exclusion criteria
EPIC-InterAct exclusions: prevalent diabetes (548), missing diabetes status (129), post-censoring diabetes (4), extreme energy reporting top/bottom 1% (736), missing education/physical activity/smoking/BMI (955). Meta-analysis: prospective studies of fibre and T2D found in PubMed (single database), details in ESM.
4. Methods and endpoints
- Exposure / intervention
Energy-adjusted total, cereal, fruit and vegetable fibre (AOAC gravimetric method; Englyst method in UK) from country-specific FFQs/diet histories at baseline only; cereal fibre in EPIC came 81% from bread, 8% pasta/rice, 7% breakfast cereals
- Comparator
Lowest quartile (EPIC-InterAct: total fibre <18.9 g/day; cereal fibre <5.7 g/day); per 10 g/day in dose-response meta-analysis
- Endpoints
Verified incident type 2 diabetes (self-report, registers, medication, hospital records, verified by medical records)
- Statistical methods
Prentice-weighted Cox models per country, age as time scale, pooled by random-effects meta-analysis (I2); 4 nested adjustment models (age/sex; lifestyle; diet incl. carbohydrate, Mg, vitamin B1, SFA; BMI); fibre types mutually adjusted. Meta-analysis: random effects, Greenland-Longnecker dose-response, fractional polynomials for non-linearity, Egger test and funnel plots.
5. Actual results
- Results
EPIC-InterAct Q4 vs Q1 (lifestyle+diet model): total fibre HR 0.82 (0.69, 0.97), cereal fibre 0.81 (0.70, 0.93), vegetable fibre 0.84 (0.74, 0.96), fruit fibre 0.98 (0.89, 1.08). After BMI adjustment: total 0.91 (0.81, 1.03), cereal 0.95 (0.83, 1.08), fruit 0.96 (0.83, 1.10), vegetable 0.93 (0.84, 1.03); BMI explained 50% of the total-fibre association. Meta-analysis per 10 g/day: total fibre RR 0.91 (0.87, 0.96; I2 29.4%, n=15); cereal fibre 0.75 (0.65, 0.86; I2 75.1%, n=12; non-linear, P=0.004); fruit fibre 0.95 (0.87, 1.03); vegetable fibre 0.93 (0.82, 1.05); insoluble fibre 0.73 (0.62, 0.86; 3 studies); soluble fibre 0.70 (0.47, 1.04). Highest vs lowest: total 0.85 (0.77, 0.94), cereal 0.77 (0.68, 0.87; I2 77.7%). Cereal fibre Egger P=0.08 (possible small-study bias); restricted to studies with >=1,000 cases RR 0.76 (0.63, 0.92).
- Absolute risks
not reported in full text
- Authors' conclusion
In EPIC-InterAct, total, cereal and vegetable fibre were inversely associated with T2D but not significantly after BMI adjustment; the meta-analysis found inverse associations for total and cereal fibre independent of BMI; a diet rich in fibre, especially cereal fibre, may lower T2D risk.
6. Strengths
- Large European case-cohort with 11,559 verified cases and measured height/weight in most participants
- Country-specific analyses pooled with low between-country heterogeneity (I2 0-34%)
- Stepwise adjustment shows the role of BMI transparently; fibre types mutually adjusted
- Updated dose-response meta-analysis with non-linearity testing and small-study bias assessment
7. Weaknesses
- EPIC-InterAct: associations null after BMI adjustment (BMI may be confounder and mediator)
- Diet measured only at baseline; FFQ measurement error
- Residual confounding by healthy dietary pattern acknowledged by authors
- Meta-analysis searched PubMed only; selection details only in ESM; no registered protocol; no formal risk-of-bias tool for included cohorts
- High heterogeneity for cereal fibre (I2 75.1%) and possible small-study bias (Egger P=0.08)
- Authors note cereal fibre inverse associations were consistently found in studies with self-reported BMI but less often with measured BMI (possible residual confounding)
- No absolute risks; per-10 g/day increment is large relative to EPIC cereal fibre intakes (Q4 vs Q1 medians 13.7 vs 4.3 g/day)
8. Risk of bias
- Orienting tool
- ROBINS-E (EPIC-InterAct component); AMSTAR 2-oriented note for the meta-analysis component
- 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
Overall unchanged. The European primary data are null after BMI adjustment; the 25% per 10 g/day figure comes from the heterogeneous meta-analysis. Simplified single-reviewer judgement from the full text; a formal ROBINS-E/AMSTAR 2 assessment would need two independent reviewers and the protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | some concerns | Adjusted for lifestyle, education, diet components and BMI; residual confounding by overall healthy diet acknowledged; BMI acts as confounder and mediator, and the association disappears after BMI adjustment. |
| Measurement of exposure | some concerns | Validated but country-specific FFQs/diet histories at baseline only; different fibre analytical methods (Englyst in UK); non-differential error likely attenuates. |
| Selection of participants | low | Random subcohort and all incident cases; exclusions modest and authors report no baseline differences between included and excluded participants. |
| Post-exposure interventions | some concerns | Dietary change during 10.8 years not captured. |
| Missing data | low | Complete-case exclusions of about 3% for covariates; outcome ascertainment via registers and records. |
| Measurement of outcome | low | Incident T2D verified by multiple sources and medical record review. |
| Selection of reported result | low | All four models and all fibre sources reported, including null BMI-adjusted results. |
| Meta-analysis component (AMSTAR 2-oriented) | some concerns | Single database (PubMed), no protocol, no RoB tool for included cohorts; heterogeneity and small-study bias for cereal fibre; sensitivity analysis restricted to large studies gives a similar estimate. |
Full-text review (25 October 2026): Added EPIC-InterAct fibre-source HRs before/after BMI (cereal 0.81 -> 0.95), BMI explaining 50%, meta-analysis heterogeneity (cereal I2 75.1%), non-linearity, small-study bias (Egger P=0.08), single-database search, funding details incl. Novo Nordisk support for one author. Bias overall unchanged (some_concerns) with full domains; consistency set to mixed; archive derivation refined.
9. Assessment of the statistical interpretation
Relative risks/hazard ratios from observational data. Cereal fibre RR 0.75 per 10 g/day is heterogeneous (I2 75.1%), non-linear and possibly inflated by small-study bias, though large studies give a similar estimate (0.76). In European EPIC-InterAct data alone, BMI-adjusted cereal fibre HR 0.95 (0.83-1.08) is not significant.
10. Relevance to the specific claim
Supports C-WHT-05.
11. Independent contrary findings
12. Transferability to Germany and Europe
High.
13. Overall methodological judgement
Good-quality European case-cohort plus a useful but methodologically limited meta-analysis. Supports an inverse association of total and cereal fibre with T2D; the magnitude is uncertain and partly explained by body weight; causality not established.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | indirect |
| Reproducibility | replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | essentially correct but incomplete |
14. Open questions
Whether cereal fibre itself or whole-grain foods/healthier diets explain the association; dose-response at Egyptian intake levels; role of BMI as mediator.
15. Linked fact checks
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
Each 10 g/day cereal fibre dropped T2D risk 25% - the very fibre stripped out when wheat is milled; fruit and vegetable fibre showed no comparable effect.
- A. What was actually studied?
Fibre and T2D.
- B. What did the authors find?
EPIC-InterAct: inverse associations for total, cereal and vegetable fibre that became non-significant after BMI adjustment; meta-analysis: cereal fibre RR 0.75 per 10 g/day (I2 75%), total fibre 0.91, fruit and vegetable fibre not significant.
- C. What conclusions do the authors themselves draw?
Total and cereal fibre inversely related; partly explained by body weight.
- D. What does the archive derive from it?
Links to milling of wheat.
- E. Is this derivation scientifically justified?
Essentially correct for the meta-analysis figure, but incomplete: it omits that the study's own European (EPIC-InterAct) data showed no significant association after BMI adjustment, that the 25% estimate is highly heterogeneous with possible small-study bias, and that vegetable fibre was inversely associated in EPIC before BMI adjustment. 'Dropped' implies causation. 81% of EPIC cereal fibre came from bread, so the finding supports higher-fibre/whole-grain cereal products rather than avoiding wheat.
- Adequacy of the interpretation (by the archive)
- essentially correct but incomplete