Meta-analysisindependent cross-checkabstract checked★ key study I-SUP-108 in-depth review
Antioxidant pills and death: Cochrane review
Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases.
Overview
Bjelakovic G et al.. Cochrane Database Syst Rev 2012 · PMID 22419320 · DOI 10.1002/14651858.CD007176.pub2
In 78 trials, antioxidant pills did not lower death rates; beta-carotene and vitamin E slightly increased them in the best trials.
Does the study support the specific claim? Yes · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
To see whether antioxidant supplements prevent death.
- How was the study done?
Cochrane meta-analysis of randomised trials.
- How many people, animals or samples were studied?
78 trials; 296,707 people.
- What was actually found?
No benefit; small increase in deaths with beta-carotene and vitamin E in low-bias trials.
- How large was the observed effect?
Low-bias RR 1.04 (1.01-1.07); beta-carotene 1.05; vitamin E 1.03.
- What do the results mean in everyday life?
Antioxidant pills do not help you live longer and some may slightly shorten life.
- What are the study's limitations?
Mixed populations; older search.
- What does it say about Elawady's specific claim?
- Do the results apply to Germany and Europe?
Applicable.
- What can explicitly not be concluded from this study?
Does not apply to antioxidants in food.
Scientific in-depth review
1. Full reference
- Authors
- Bjelakovic G et al.
- Journal
- Cochrane Database Syst Rev
- Year
- 2012
- Identifier
- PMID 22419320 DOI 10.1002/14651858.CD007176.pub2 · PMC8407395
- Verification
- verified – PubMed efetch 2026-10-26; Crossref works API 2026-10-26
- Retracted
- no
- Corrections
none Crossref update-to: new version of an earlier Cochrane review (not a retraction).
- Extent checked
- abstract only
- Conflicts of interest and funding
COI statement in record: None known. The funding sources had no role in the conduct of the study, collection of data, management, analysis, interpretation of the data, or preparation of the manuscript.
2. Research question
Do antioxidant supplements (beta-carotene, vitamins A, C, E, selenium) affect mortality in primary or secondary prevention?
3. Study design and population
- Study design
- Meta-analysis – Cochrane systematic review and meta-analysis of randomised trials with trial sequential analysis
- Population
Adults, healthy or with various diseases in a stable phase
- Country
Multinational
- Period
Search to February 2011
- Sample size
78 trials; 296,707 participants (56 low-bias trials, 244,056)
- Inclusion and exclusion criteria
Randomised trials of antioxidant supplements vs placebo or no intervention
4. Methods and endpoints
- Exposure / intervention
Beta-carotene, vitamin A, vitamin C, vitamin E, selenium (alone or combined)
- Comparator
Placebo/no intervention
- Endpoints
All-cause mortality
- Statistical methods
Random- and fixed-effect meta-analysis; meta-regression; trial sequential analysis
5. Actual results
- Results
All trials: RR 1.02 (0.98-1.05) random effects (fixed effect 1.03, 1.01-1.05); I2 12%. Low-bias trials: RR 1.04 (1.01-1.07). Low-bias trials by agent: beta-carotene RR 1.05 (1.01-1.09); vitamin E RR 1.03 (1.00-1.05); vitamin A RR 1.07 (0.97-1.18); vitamin C RR 1.02 (0.98-1.07); selenium RR 0.97 (0.91-1.03). Vitamin A dose associated with mortality in meta-regression.
- Absolute risks
Pooled event proportions in low-bias trials 12.9% vs 10.6% (crude; not comparable because of factorial designs; stats/C-SUP-01_bjelakovic2012.py)
- Authors' conclusion
No evidence supports antioxidant supplements for primary or secondary prevention; beta-carotene and vitamin E seem to increase mortality, and so may higher doses of vitamin A.
6. Strengths
- Cochrane methods
- Large number of trials
- Risk-of-bias stratification and trial sequential analysis
7. Weaknesses
- Mixed populations and doses
- Overlaps with A-SUP-24 (same group, 2007)
- Search ends 2011
8. Risk of bias
- Orienting tool
- AMSTAR 2
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- low
- Rationale
Cochrane review; some debate about inclusion and fixed vs random effects.
9. Assessment of the statistical interpretation
Crude proportions (11.7% vs 10.2%) overstate the effect because factorial trials put more people in active arms; the valid pooled excess is small (RR 1.03-1.05).
10. Relevance to the specific claim
Updates A-SUP-24 (not independent); supports lack of benefit and a small mortality excess for beta-carotene and vitamin E.
11. Independent contrary findings
Random-effects analysis of all trials non-significant; USPSTF 2022 found vitamin E mortality OR 1.02 (n.s.).
12. Transferability to Germany and Europe
Applicable.
13. Overall methodological judgement
Best overall summary of antioxidant-supplement mortality; small harms for beta-carotene and high-dose vitamin E.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | low |
| 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
Effects in deficient populations.
15. Linked fact checks
- When popular oral supplements (calcium, beta-carotene, vitamin E, omega-3, multivitamins, glucosamine, selenium, vitamin A) were tested in large long-term placebo-controlled trials, most failed to deliver the promised prevention of disease. – Studies supporting the claim
- Beta-carotene supplements (with or without vitamin A) increased lung cancer and mortality in smokers and asbestos-exposed workers (CARET, ATBC), and the USPSTF recommends against them. – Studies supporting the claim
- High-dose vitamin E (400 IU/day) increases prostate cancer (SELECT, +17%) and may increase all-cause mortality (Miller 2005: +39 per 10,000). – Studies supporting the claim