Meta-analysisindependent cross-checkfull text checked★ key study I-LEG-112 in-depth review
Legumes and type 2 diabetes in 27 cohorts
Associations of Total Legume, Pulse, and Soy Consumption with Incident Type 2 Diabetes: Federated Meta-Analysis of 27 Studies from Diverse World Regions.
Overview
Pearce M et al.. J Nutr 2021 · PMID 33693815 · DOI 10.1093/jn/nxaa447
Overall, legume intake showed a very weak positive link with diabetes, seen only in some European cohorts; pulses were not linked.
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?
Pool individual data on legumes and diabetes.
- How was the study done?
Federated meta-analysis.
- How many people, animals or samples were studied?
807,785 adults, 27 cohorts.
- What was actually found?
IRR 1.02 per 20 g/d; Europe 1.05.
- How large was the observed effect?
Small.
- What do the results mean in everyday life?
Legumes do not clearly protect against diabetes, but the evidence does not show meaningful harm either.
- What are the study's limitations?
Dietary measurement error; confounding.
- What does it say about Elawady's specific claim?
Tempers claims of benefit; does not support a ban.
- Do the results apply to Germany and Europe?
Relevant to Europe.
- What can explicitly not be concluded from this study?
It cannot show cause and effect.
Scientific in-depth review
1. Full reference
- Authors
- Pearce M et al.
- Journal
- J Nutr
- Year
- 2021
- Identifier
- PMID 33693815 DOI 10.1093/jn/nxaa447 · PMC8112771
- Verification
- verified – PubMed efetch 2026-10-26 + Crossref works API
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
EU FP7 (InterConnect) and national research councils (full text).
2. Research question
Are total legume, pulse and soy intakes associated with incident T2D across world regions?
3. Study design and population
- Study design
- Meta-analysis – Federated individual-participant meta-analysis of 27 prospective cohorts (InterConnect)
- Population
807,785 adults without diabetes in the Americas, Eastern Mediterranean (Golestan, Iran), Europe and Western Pacific
- Country
Multinational
- Period
Cohort follow-up periods varied
- Sample size
807,785 participants; up to 42,473 incident cases
- Inclusion and exclusion criteria
Cohorts with legume intake and T2D data whose investigators agreed to participate.
4. Methods and endpoints
- Exposure / intervention
Legume, pulse, soy intake (per 20 g/d)
- Comparator
Lower intake
- Endpoints
Incident T2D
- Statistical methods
Harmonised Poisson/Cox models; random-effects meta-analysis
5. Actual results
- Results
Total legumes IRR 1.02 (95% CI 1.01-1.04) per 20 g/d, I2 74%; Americas 1.01 (0.99-1.02); Eastern Mediterranean 1.07 (0.98-1.18; single Iranian cohort); Western Pacific 1.00; Europe 1.05 (1.01-1.10), driven by Germany, UK and Sweden. No association for pulses or soy.
- Absolute risks
Not reported
- Authors' conclusion
No evidence of association in several regions; the European positive association may reflect the dietary context in which legumes are eaten.
6. Strengths
- Individual-participant data
- 26 of 27 cohorts unpublished on this question (less publication bias)
- Full text read
7. Weaknesses
- Heterogeneous FFQs and portion assumptions
- Residual confounding
- Only one Eastern Mediterranean cohort
8. Risk of bias
- Orienting tool
- ROBINS-E
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Observational; measurement error; region-specific confounding.
9. Assessment of the statistical interpretation
Very small effect; extrapolated to 100 g/d the overall IRR would be about 1.10 (stats/C-LEG-13_legume_meta_ci.py), driven by Europe; pulses (the Egyptian type of legume) showed no association.
10. Relevance to the specific claim
Honest counterweight: does not support a protective effect on diabetes and shows a weak positive association in some European cohorts; but pulses showed no association and there is no signal of substantial harm.
11. Independent contrary findings
I-LEG-113 (RCTs: pulses improve glycaemic control), I-LEG-111.
12. Transferability to Germany and Europe
European subgroup directly relevant; mechanism unclear.
13. Overall methodological judgement
Best evidence against a diabetes benefit; does not establish harm.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | high |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Why European cohorts differ (foods eaten with legumes, e.g. processed meat, baked beans).
15. Linked fact checks
- Legumes as a whole (fava beans, lentils, chickpeas, beans, cowpeas, peas, lupin, peanuts) are harmful and should be fully forbidden. – Studies with mixed or unclear results
- Fava beans and lentils (together with flour products) are toxins that cause inflammation and deteriorate health; the fava bean hull is indigestible and the bean contains toxins that ferment 'dirtily'. – Studies supporting the claim