Meta-analysisindependent cross-checkfull text checked★ key study I-MEAT-03 in-depth review
More red and processed meat, more bowel cancer
Red and processed meat and colorectal cancer incidence: meta-analysis of prospective studies.
Overview
Chan DS et al.. PLoS One 2011 · PMID 21674008 · DOI 10.1371/journal.pone.0020456
Fresh red meat RR 1.17 per 100 g/day; processed meat RR 1.18 per 50 g/day.
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?
Meat and colorectal cancer.
- How was the study done?
Meta-analysis.
- How many people, animals or samples were studied?
Prospective cohorts.
- What was actually found?
Higher risk with more meat.
- How large was the observed effect?
RR 1.17 per 100 g/d red meat.
- What do the results mean in everyday life?
Limiting meat is sensible.
- What are the study's limitations?
Observational; individual studies not quality-rated; only PubMed searched; no association in Asian studies and for rectal cancer when red and processed meat are analysed separately.
- What does it say about Elawady's specific claim?
Supports limiting red meat.
- Do the results apply to Germany and Europe?
Moderate-high.
- What can explicitly not be concluded from this study?
A specific safe frequency (e.g. beef once a week); risk at low intakes; absolute risk.
Scientific in-depth review
1. Full reference
- Authors
- Chan DS et al.
- Journal
- PLoS One
- Year
- 2011
- Identifier
- PMID 21674008 DOI 10.1371/journal.pone.0020456
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Funded by World Cancer Research Fund International (grant 2007/SP01); funder had no role in design, analysis or publication. Authors declare no competing interests.
2. Research question
Red and processed meat and colorectal cancer incidence.
3. Study design and population
- Study design
- Meta-analysis – Systematic review and meta-analysis (highest vs lowest, linear and non-linear fractional-polynomial dose-response) of prospective studies (19 cohort, 2 case-cohort, 3 nested case-control), WCRF/AICR Continuous Update Project.
- Population
Adults in 24 prospective studies: 12 North American, EPIC (10 European countries), Finland, Netherlands, Japan (2 each), Australia, Canada, Sweden, China, Singapore; 13 mixed-sex, 3 male, 8 female cohorts.
- Country
Multinational
- Period
PubMed searched 1966 to 31 March 2011.
- Sample size
42 articles from 28 prospective studies identified; 24 studies in highest-vs-lowest and 21 in dose-response analyses. Red+processed meat dose-response: 11 studies, 11,358 cases; red meat: 8 studies, 4,314 cases; processed meat: 9 studies, 10,863 cases. Total number of participants not reported.
- Inclusion and exclusion criteria
Prospective studies reporting estimates for red meat, processed meat or both with colorectal, colon or rectal cancer; dose-response required case numbers and denominators. Duplicate cohort publications resolved by longer follow-up/more cases. UK Dietary Cohort Consortium (null pooled result, 579 cases) and the Pooling Project (null preliminary abstract) could not be included.
4. Methods and endpoints
- Exposure / intervention
Red meat (mostly beef, veal, pork, mutton, lamb), processed meat (ham, bacon, sausages, cured/preserved meats), and the combined item; servings converted with 120 g (red/red+processed) and 50 g (processed) standard portions.
- Comparator
Lowest intake; per dose
- Endpoints
Colorectal, colon, rectal cancer.
- Statistical methods
Random-effects (DerSimonian-Laird) and fixed-effects pooling; generalized least-squares trend estimation; fractional polynomial non-linear models; Cochran Q and I2; meta-regression; Egger's test and funnel plots; influence analysis. No risk-of-bias/quality assessment of individual studies.
5. Actual results
- Results
Red+processed meat: highest vs lowest colorectal RR 1.22 (95% CI 1.11-1.34), colon 1.19 (1.06-1.34), rectal 1.51 (1.31-1.75); per 100 g/day colorectal 1.14 (1.04-1.24; I2=56%), colon 1.25 (1.10-1.43), rectal 1.31 (1.13-1.52); non-linear (P=0.03), roughly linear up to ~140 g/day then flatter. Red meat: highest vs lowest colorectal 1.10 (1.00-1.21); per 100 g/day colorectal 1.17 (1.05-1.31; I2=0%), colon 1.17 (1.02-1.33), rectal 1.18 (0.98-1.42). Processed meat: highest vs lowest colorectal 1.17 (1.09-1.25); per 50 g/day colorectal 1.18 (1.10-1.28), colon 1.24 (1.13-1.35), rectal 1.12 (0.99-1.28). Red meat and colon cancer by region per 100 g/day: Europe 1.29 (1.08-1.54), North America 1.11 (0.86-1.44), Asia-Pacific 0.94 (0.69-1.27). Sex-specific estimates mostly non-significant (few studies).
- Absolute risks
Not reported as absolute risks; the discussion cites external population preventable fractions for red meat of 5% (US, UK) and 7% (Brazil, China).
- Authors' conclusion
High intake of red and processed meat is associated with increased risk of colorectal, colon and rectal cancers; evidence supports limiting red and processed meat in high consumers; prevention should address multiple diet and lifestyle factors.
6. Strengths
- Large number of prospective studies and cases
- Linear and non-linear dose-response analyses
- Heterogeneity explored by meta-regression and influence analysis
- Independent charity funding; no COI declared
7. Weaknesses
- Observational studies; residual confounding cannot be ruled out (authors); association attenuated in some better-adjusted analyses (e.g. NHS/HPFS)
- No formal risk-of-bias/quality assessment of included studies
- Only PubMed searched (plus hand-searching); two null pooled analyses (UK Dietary Cohort Consortium, Pooling Project) not included
- Portion-size conversions and midpoint assignment introduce exposure misclassification; moderate heterogeneity for combined meat (I2 56-60%)
- Rectal-cancer associations for red and processed meat separately not significant; no association in Asia-Pacific studies; sex-specific results imprecise
- Small effect sizes (RR 1.1-1.2) in the range where confounding can matter
8. Risk of bias
- Orienting tool
- AMSTAR 2
- 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
- high
- Rationale
Single-reviewer AMSTAR 2-oriented appraisal from the full text; formal assessment needs two independent reviewers. One critical flaw (no risk-of-bias assessment of primary studies) plus a limited single-database search gives low confidence in the review under AMSTAR 2 rules, hence overall 'high' (was 'some_concerns' from the abstract). The direction of the association is consistent with later WCRF/IARC evaluations, but this review alone cannot establish causality.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 2 (critical) a-priori protocol | low | WCRF/AICR CUP protocol and search strategy published online. |
| Item 4 (critical) comprehensive search | some concerns | PubMed only plus reference lists; no language restriction; two null pooled analyses not includable. |
| Items 5-6 duplicate selection and extraction | some concerns | Several reviewers / two reviewers described, but independent duplicate process not explicitly stated. |
| Item 7 (critical) excluded studies justified | low | Excluded articles and studies not in dose-response analysis listed with reasons (Table S2). |
| Item 9 (critical) risk of bias of included studies | high | No quality or risk-of-bias assessment of individual studies; only adjustment level explored in meta-regression. |
| Item 10 funding of included studies | high | Not reported. |
| Item 11 (critical) meta-analytic methods | low | Appropriate random-effects and dose-response methods. |
| Item 13 (critical) RoB considered in interpretation | some concerns | Residual confounding and attenuation with adjustment discussed, but no formal RoB to inform conclusions. |
| Item 14 heterogeneity | low | Explored by meta-regression (unit conversion, adjustment level) and stratification. |
| Item 15 (critical) publication bias | low | Funnel plots and Egger's tests; possible missing small inverse studies for processed meat and colon cancer noted. |
| Item 16 COI | low | Funding and COI declared. |
Full-text review (25 October 2026): Study counts and case numbers added; colon/rectal, regional (Asia-Pacific null) and sex-specific results added; WCRF funding and no COI documented; full text shows no risk-of-bias assessment of primary studies and PubMed-only search -> AMSTAR 2 overall some_concerns -> high, methodological_quality high -> moderate, consistency consistent -> mixed; absolute risk now cites preventable fractions mentioned.
9. Assessment of the statistical interpretation
Modest relative risks (about 1.1-1.2 per portion increment) with a dose-response; consistent for colon cancer in Europe/North America but null in Asia-Pacific studies and imprecise by sex and for rectal cancer when meats are analysed separately. Absolute risks not reported; small relative increases translate into small absolute differences.
10. Relevance to the specific claim
Supports C-MEAT-01.
11. Independent contrary findings
12. Transferability to Germany and Europe
Moderate-high: European studies (EPIC etc.) show the clearest red meat-colon cancer association.
13. Overall methodological judgement
Dose-response association of red and processed meat with colorectal cancer in prospective studies, strongest for processed meat and colon cancer; review lacks risk-of-bias assessment and searched only PubMed. Supports limiting intake in high consumers but does not define safe frequencies such as 'once a week'.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | high |
| Precision of results | high |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Causality for unprocessed red meat; why no association in Asia-Pacific cohorts; effect at low intakes (1-3 portions/week); role of cooking methods and residual confounding.
15. Linked fact checks
- Red meat should be limited: beef/buffalo at most once a week and lamb twice a week (protein not every day). – Studies supporting the claim