Meta-analysisindependent cross-checkfull text checked★ key study I-VEG-02 in-depth review
Fruit, vegetables and death rates
Fruit and vegetable intake and the risk of cardiovascular disease, total cancer and all-cause mortality-a systematic review and dose-response meta-analysis of prospective studies.
Overview
Aune D et al.. Int J Epidemiol 2017 · PMID 28338764 · DOI 10.1093/ije/dyw319
Each extra 200 g of fruit and vegetables a day was linked to 10% lower death risk; green leafy vegetables and salads were also protective.
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?
F&V and outcomes
- How was the study done?
Meta-analysis
- How many people, animals or samples were studied?
95 studies
- What was actually found?
Lower CVD, cancer, mortality
- How large was the observed effect?
RR 0.90 per 200 g
- What do the results mean in everyday life?
Eat plenty of vegetables.
- What are the study's limitations?
Observational
- What does it say about Elawady's specific claim?
Contradicts avoidance
- Do the results apply to Germany and Europe?
High
- What can explicitly not be concluded from this study?
Cannot prove causation.
Scientific in-depth review
1. Full reference
- Authors
- Aune D et al.
- Journal
- Int J Epidemiol
- Year
- 2017
- Identifier
- PMID 28338764 DOI 10.1093/ije/dyw319
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Olav og Gerd Meidel Raagholt's Stiftelse, Central Norway RHA-NTNU Liaison Committee, Imperial College NIHR BRC; funders had no role; authors declare no conflicts of interest.
2. Research question
Fruit/vegetable intake and CVD, cancer, mortality.
3. Study design and population
- Study design
- Meta-analysis – Systematic review and dose-response meta-analysis of prospective studies
- Population
Adult general-population prospective cohorts: 44 from Europe, 26 USA, 20 Asia, 5 Australia; no cohorts from Africa, West Asia or Latin America
- Country
Multiple (Europe, USA, Asia, Australia)
- Period
PubMed and Embase from inception to 19 July 2016, updated to 29 September 2016
- Sample size
95 cohort studies (142 publications); up to 2,123,415 participants per analysis; 71,160-94,235 deaths in all-cause analyses
- Inclusion and exclusion criteria
Prospective studies reporting adjusted RRs with 95% CIs for CHD, stroke, CVD, total cancer or all-cause mortality; >=3 intake categories for dose-response; excluded studies listed in supplement
4. Methods and endpoints
- Exposure / intervention
Fruit and vegetable intake (per 200 g/day; subtypes per 100 g/day; 80 g serving)
- Comparator
Lower intake
- Endpoints
CHD, stroke, CVD, cancer, all-cause mortality
- Statistical methods
DerSimonian-Laird random effects; Greenland-Longnecker linear and restricted-cubic-spline non-linear dose-response; Egger tests; leave-one-out; subgroup analyses; Newcastle-Ottawa quality; attributable fractions (Miettinen) using World Health Survey and GBD 2013 data
5. Actual results
- Results
Fruit and vegetables per 200 g/day: CHD RR 0.92 (95% CI 0.90-0.94; I2 0%); stroke 0.84 (0.76-0.92; I2 73%); CVD 0.92 (0.90-0.95); total cancer 0.97 (0.95-0.99); all-cause mortality 0.90 (0.87-0.93; I2 83%). Non-linear: 800 g/day vs 0-40 g/day associated with 31% lower all-cause mortality; lowest cancer risk at ~600 g/day. All-cause mortality per 100 g/day: green leafy vegetables/salads 0.78 (0.71-0.86); cruciferous vegetables 0.90 (0.85-0.95); cooked vegetables 0.89 (0.80-0.99); tinned fruit associated with higher mortality (high vs low 1.13, 1.04-1.23). Findings persisted across most subgroups, including geographical region and study quality (most studies 7-9 NOS stars). EPIC data indicate measurement error attenuates associations.
- Absolute risks
Attributable deaths, assuming causality: 5.4 million (text; abstract states 5.6 million) premature deaths worldwide in 2013 attributable to intake below 500 g/day and 7.8 million below 800 g/day; no absolute risk differences per individual reported
- Authors' conclusion
Fruit and vegetable intake was inversely associated with CVD, cancer and all-cause mortality, with the lowest risks at about 800 g/day (600 g/day for cancer), supporting recommendations to increase intake.
6. Strengths
- Very large evidence base with detailed linear and non-linear dose-response
- Analyses of specific fruit and vegetable subtypes
- Consistent results across regions, sex, study quality and adjustment subgroups
- High average study quality (NOS ~8)
7. Weaknesses
- Observational: residual confounding by healthy lifestyle and healthcare access possible (acknowledged)
- High heterogeneity for all-cause mortality and stroke
- No registered protocol reported; extraction by one author checked by two others; only two databases
- Single baseline diet measurement; measurement error
- No cohorts from Africa/West Asia/Latin America
- Attributable-death estimates rest on causality assumption; abstract/text discrepancy (5.6 vs 5.4 million)
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
- some concerns
- Rationale
Single-reviewer structured AMSTAR 2 appraisal from full text; not formal (needs two independent reviewers). Strict AMSTAR 2 rating would be 'low' confidence because of no registered protocol, but the analysis is otherwise thorough; the main limitation is the observational evidence base.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 1 PICO components | low | Clearly defined. |
| Item 2 Protocol registered a priori (critical) | high | No protocol registration reported. |
| Item 3 Explanation of included designs | low | Prospective studies justified. |
| Item 4 Comprehensive search (critical) | some concerns | PubMed and Embase only, plus reference checks; broad terms. |
| Item 5 Duplicate selection | some concerns | Not clearly described as duplicate. |
| Item 6 Duplicate extraction | some concerns | Single extractor with accuracy check by two co-authors. |
| Item 7 Excluded studies justified (critical) | low | Supplementary Table 2 lists excluded studies. |
| Item 8 Description of included studies | low | Supplementary tables of characteristics. |
| Item 9 RoB assessment (critical) | some concerns | Newcastle-Ottawa scale, a limited tool for confounding. |
| Item 10 Funding of included studies | high | Not reported. |
| Item 11 Appropriate meta-analytic methods (critical) | low | Standard random-effects and dose-response methods. |
| Item 12 Impact of RoB on results | low | Subgroups by study quality and adjustment. |
| Item 13 RoB considered in interpretation (critical) | low | Confounding, measurement error and causality discussed. |
| Item 14 Heterogeneity explained | low | Extensive subgroup analyses. |
| Item 15 Publication bias (critical) | low | Egger tests and funnel plots with sensitivity analyses. |
| Item 16 Conflicts of interest | low | Declared none. |
Full-text review (25 October 2026): Added cohort geography, participants/deaths, subtype estimates (green leafy/salads 0.78 per 100 g/d, cruciferous 0.90), non-linear 31% reduction at 800 g/d, measurement-error data, regional gaps, abstract/text discrepancy in attributable deaths, funding/COI; AMSTAR 2 item-level appraisal (overall unchanged).
9. Assessment of the statistical interpretation
Consistent, precise inverse associations of modest size (e.g. 10% lower all-cause mortality per 200 g/day); green leafy vegetables/salads and cruciferous vegetables, the foods specifically targeted by the claims, show inverse rather than harmful associations. Causality is plausible but not proven.
10. Relevance to the specific claim
Directly addresses vegetables including leafy greens and salads.
11. Independent contrary findings
None.
12. Transferability to Germany and Europe
High (44 of 95 cohorts European; no regional heterogeneity). No cohorts from Africa or West Asia, so direct data for Egypt are lacking.
13. Overall methodological judgement
Key evidence.
| 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 | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Causality; generalisability to Middle East/North Africa; residue-specific effects.
15. Linked fact checks
- 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
- Nitrate from fertilisers (together with pesticides) in lettuce, rocket and other leaves harms the people who eat them. – Studies contradicting or limiting the claim
- Raw cruciferous greens such as bok choy suppress the thyroid (glucosinolates) and can cause myxedema coma. – Studies contradicting or limiting the claim
- Vegetables themselves are often harmful because of heavy pesticide use; organic diets show how much exposure conventional vegetables cause. – Studies contradicting or limiting the claim
- A graded exception list: zucchini (2nd degree), okra (3rd), eggplant (4th) and molokhia (5th degree) may be eaten only rarely, ranked from least to most harmful; eating them weekly causes inflammation lasting 4-5 days. – Studies contradicting or limiting the claim
- Avoiding (nearly) all vegetables, salads and legumes is compatible with good health. – Studies contradicting or limiting the claim