Meta-analysisindependent cross-checkfull text checked★ key study I-VEG-106 in-depth review
About five portions of fruit and vegetables a day linked to lower death rates
Fruit and Vegetable Intake and Mortality: Results From 2 Prospective Cohort Studies of US Men and Women and a Meta-Analysis of 26 Cohort Studies.
Overview
Wang DD et al.. Circulation 2021 · PMID 33641343 · DOI 10.1161/CIRCULATIONAHA.120.048996
Compared with 2 servings a day, 5 servings were linked to 13% lower mortality in 2 US cohorts and in a meta-analysis of 26 cohorts with nearly 1.9 million people.
Does the study support the specific claim? No · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Find the fruit and vegetable intake linked to the lowest mortality.
- How was the study done?
Two long cohorts plus a dose-response meta-analysis.
- How many people, animals or samples were studied?
108,735 US adults; 1.9 million in meta-analysis.
- What was actually found?
Lowest mortality at ~5 servings/day; green leafy vegetables beneficial; potatoes and juice not.
- How large was the observed effect?
HR 0.87 (0.85-0.90).
- What do the results mean in everyday life?
Eating around 3 servings of vegetables and 2 of fruit a day is associated with longer life.
- What are the study's limitations?
Observational; confounding and measurement error possible.
- What does it say about Elawady's specific claim?
Contradicts the claim that avoiding vegetables is compatible with good health.
- Do the results apply to Germany and Europe?
Applies to European populations (included cohorts).
- What can explicitly not be concluded from this study?
Cannot prove that vegetables themselves cause the lower mortality.
Scientific in-depth review
1. Full reference
- Authors
- Wang DD et al.
- Journal
- Circulation
- Year
- 2021
- Identifier
- PMID 33641343 DOI 10.1161/CIRCULATIONAHA.120.048996
- Verification
- verified – PubMed efetch 2026-10-26; Crossref 2026-10-26
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
COI statement: Conflict of Interest Disclosures. Yanping Li has received research support from the California Walnut Commission and SwissRe Management Ltd. The remaining authors disclose no conflicts. | Funding (PubMed grant list): Medical Research Council; NCI NIH HHS; NHLBI NIH HHS; NIDDK NIH HHS
2. Research question
What intake of fruit and vegetables is associated with the lowest total and cause-specific mortality?
3. Study design and population
- Study design
- Meta-analysis – Two prospective cohorts (NHS, HPFS) with repeated FFQs plus dose-response meta-analysis of 26 cohorts
- Population
66,719 women (NHS 1984-2014) and 42,016 men (HPFS 1986-2014) free of CVD, cancer and diabetes; meta-analysis 1,892,885 participants
- Country
USA (cohorts); multiple (meta-analysis)
- Period
1984-2014
- Sample size
108,735 in own cohorts (33,898 deaths); 1,892,885 in meta-analysis (145,015 deaths)
- Inclusion and exclusion criteria
Excluded baseline CVD, cancer, diabetes; diet updating stopped at diagnosis of intermediate conditions
4. Methods and endpoints
- Exposure / intervention
Fruit and vegetable servings/day and subgroups (green leafy, cruciferous, starchy etc.)
- Comparator
2 servings/day (reference)
- Endpoints
Total, CVD, cancer and respiratory mortality
- Statistical methods
Cox models with cumulative averages; restricted cubic splines; random-effects dose-response meta-analysis; I2; meta-regression; small-study tests
5. Actual results
- Results
5 vs 2 servings/day: HR 0.87 (0.85-0.90) total, 0.88 (0.83-0.94) CVD, 0.90 (0.86-0.95) cancer, 0.65 (0.59-0.72) respiratory mortality; plateau above ~5 servings/day (2 fruit + 3 vegetables). Meta-analysis RR 0.87 (0.85-0.88) for 5 servings/day; considerable heterogeneity and evidence of small-study bias. Green leafy, non-starchy and cruciferous vegetables inversely associated; starchy vegetables, potatoes and fruit juice not associated.
- Absolute risks
not reported
- Authors' conclusion
Higher intakes of fruit and vegetables were associated with lower mortality, plateauing at about 5 servings/day; supports recommendations, but not for fruit juices and potatoes.
6. Strengths
- Large cohorts with repeated diet measures
- Excluded baseline chronic disease; stopped updating diet at diagnosis
- Meta-analysis of 26 cohorts with heterogeneity exploration
7. Weaknesses
- Observational: residual confounding by health consciousness possible
- FFQ measurement error
- Considerable heterogeneity and small-study bias in the meta-analysis
- Mainly white US health professionals in own cohorts
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
Confounding is the main concern; authors discuss reverse causation and measurement error (full text read).
9. Assessment of the statistical interpretation
Ratio CIs are log-symmetric within rounding; the narrow meta-analytic CI (0.85-0.88) shows rounding asymmetry only (stats/C-VEG-16_mortality.py). Illustrative ARR 1.3% over 10 years at 10% baseline risk if causal.
10. Relevance to the specific claim
Direct for C-VEG-16 (very low vegetable intake is the reference category with the highest mortality) and indirect for C-VEG-01.
11. Independent contrary findings
Feng 2022 (UK Biobank) suggests much of the vegetable-CVD association may be residual confounding; Feng 2023 MR null for CHD.
12. Transferability to Germany and Europe
Meta-analysis includes European and Asian cohorts; direction consistent across regions.
13. Overall methodological judgement
Strongest recent observational evidence that very low fruit and vegetable intake is associated with higher mortality; causality probable but not proven.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | high |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Causal magnitude for vegetables alone (vs fruit) and in populations with different confounding structures.
15. Linked fact checks
- Avoiding (nearly) all vegetables, salads and legumes is compatible with good health. – Studies contradicting or limiting the claim
- Humans, having a single stomach, cannot digest the cellulose in vegetables (unlike ruminants), so greens and legumes cause gas and bloating and are food made for animals, not for humans. – Studies contradicting or limiting the claim