Cohort studycited by the Tayebat archivefull text checked★ key study A-SUP-06 in-depth review
Supplements and mortality in older women
Dietary supplements and mortality rate in older women: the Iowa Women's Health Study.
Overview
Mursu J et al.. Arch Intern Med 2011 · PMID 21987192 · DOI 10.1001/archinternmed.2011.445
Several supplements were linked to slightly more deaths; calcium to fewer.
Does the study support the specific claim? Partly · Methodological quality: low
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Mortality with supplements.
- How was the study done?
Cohort.
- How many people, animals or samples were studied?
38,772 women.
- What was actually found?
Small associations both ways.
- How large was the observed effect?
HR 1.06 (MV), 0.91 (Ca).
- What do the results mean in everyday life?
Not proof of harm.
- What are the study's limitations?
Confounding.
- What does it say about Elawady's specific claim?
Partly supports; contradicts calcium claim.
- Do the results apply to Germany and Europe?
Moderate.
- What can explicitly not be concluded from this study?
Causation.
Scientific in-depth review
1. Full reference
- Authors
- Mursu J, Robien K, Harnack LJ, Park K, Jacobs DR
- Journal
- Arch Intern Med
- Year
- 2011
- Identifier
- PMID 21987192 DOI 10.1001/archinternmed.2011.445 · PMC4114071
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
NCI grant R01 CA39742; Academy of Finland, Finnish Cultural Foundation, Fulbright; sponsors had no role. One author unpaid member of the California Walnut Commission scientific advisory board; others none.
2. Research question
Are vitamin/mineral supplements associated with total mortality in older women?
3. Study design and population
- Study design
- Cohort study – Prospective cohort (Iowa Women's Health Study)
- Population
Postmenopausal women aged 55-69 in 1986 (mean 61.6 y), 99% white, Iowa
- Country
USA
- Period
1986-2008 (mean follow-up 19.0 years); supplement use reported in 1986, 1997, 2004
- Sample size
38,772 women (15,594 deaths, 40.2%); interval analyses with 29,230 (from 1997) and 19,124 (from 2004)
- Inclusion and exclusion criteria
IWHS participants who adequately completed the 1986 questionnaire including FFQ and supplement use
4. Methods and endpoints
- Exposure / intervention
Self-reported use of 15 supplements (multivitamins, vitamins A, beta-carotene, B6, folic acid, B-complex, C, D, E, iron, calcium, copper, magnesium, selenium, zinc); dose for some
- Comparator
Non-users of the respective supplement
- Endpoints
Total mortality (State Health Registry of Iowa, National Death Index); cause-specific mortality (CVD, cancer, other)
- Statistical methods
Cox models: model 1 age+energy, model 2 + lifestyle/clinical factors, model 3 + dietary factors; shorter-interval analyses; dose-response; exclusion of baseline CVD/diabetes/cancer; 15 primary tests (authors note Bonferroni threshold P=0.0033 not applied)
5. Actual results
- Results
Model 3 HRs vs non-users: multivitamins 1.06 (1.02-1.10), B6 1.10 (1.01-1.21), folic acid 1.15 (1.00-1.32), iron 1.10 (1.03-1.17), magnesium 1.08 (1.01-1.15), zinc 1.08 (1.01-1.15), copper 1.45 (1.20-1.75); calcium 0.91 (0.88-0.94). In age/energy-adjusted models vitamins B-complex, C, D, E and calcium were associated with lower mortality; adjustment for lifestyle reversed or attenuated several associations. Interval analyses: iron HRs 1.20, 1.43, 1.56; calcium 0.89, 0.90, 0.88; folic acid significant only in last interval. Iron: dose-response and HR 1.60 (1.04-2.46) when reported at all three surveys; inverse calcium association lost at highest dose.
- Absolute risks
Absolute risk increases over ~19 years (authors' calculation): multivitamins 2.4%, B6 4.1%, folic acid 5.9%, iron 3.9%, magnesium 3.6%, zinc 3.0%, copper 18.0%; calcium absolute risk reduction 3.8%
- Authors' conclusion
Several commonly used supplements may be associated with increased total mortality in older women, most strongly iron, while calcium was associated with decreased risk; authors state a causal statement is not advisable from these observational data and see little justification for widespread supplement use without medical cause.
6. Strengths
- Large cohort with long follow-up and high mortality ascertainment
- Repeated exposure assessment (three surveys) and interval analyses
- Sensitivity analyses excluding baseline chronic disease; dose-response for iron
7. Weaknesses
- Observational; confounding by indication and reverse causation (e.g. iron for anaemia due to illness) possible
- Self-reported supplement use; supplement questionnaire not separately validated
- 15 supplements tested without multiplicity correction; several HRs (B6, folic acid, magnesium, zinc) would not meet the authors' own Bonferroni threshold of P=0.0033
- Associations sensitive to model specification (direction changed with lifestyle adjustment)
- Only white older women; no nutritional status data
- No data on folic acid with B12 and cancer in this study
8. Risk of bias
- Orienting tool
- ROBINS-E
- 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 structured ROBINS-E judgement from full text; not formal (needs two independent reviewers and protocol).
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | high | Supplement users differ systematically (healthier lifestyle); direction of associations changed with adjustment; indication for use (e.g. iron, B vitamins) not captured. |
| Exposure measurement | some concerns | Self-report without dose for several supplements; repeated at three time points. |
| Selection of participants | low | Population-based cohort; small exclusions. |
| Post-exposure interventions | some concerns | Supplement use changed markedly over time (63% to 85%). |
| Missing data | low | Mortality linkage complete; questionnaires declined over time but deaths captured. |
| Outcome measurement | low | All-cause death from registries. |
| Selection of reported result | high | Many exposures and models; no multiplicity correction; emphasis on model 3 where several associations reach significance only after adjustment. |
Full-text review (25 October 2026): Added model-dependence (inverse associations in minimal models reversed after lifestyle adjustment), authors' own Bonferroni threshold (P=0.0033) that several HRs would not meet, interval and dose-response results for iron and calcium, absolute risk increases over ~19 years, funding/COI; ROBINS-E domains (overall unchanged: high).
9. Assessment of the statistical interpretation
Small hazard ratios (1.06-1.15 for most supplements) with lower CI bounds close to 1, no multiplicity correction and model-dependent direction; only iron (consistent, dose-related) and calcium (inverse) are relatively robust within this dataset. Hypothesis-generating only.
10. Relevance to the specific claim
C-SUP-06; C-SUP-04 (calcium inverse).
11. Independent contrary findings
Macpherson 2013 (A-SUP-26) RCT meta-analysis: RR 0.98.
12. Transferability to Germany and Europe
High: biology and trial populations comparable to European adults.
13. Overall methodological judgement
Hypothesis-generating; contradicted by RCT evidence for multivitamins.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | low |
| Risk of bias | high |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | not replicated |
| Consistency with the overall evidence | inconsistent |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | partly correct but over-interpreted |
14. Open questions
15. Linked fact checks
- Calcium supplements without vitamin D raise the risk of myocardial infarction by about 31%, and adding vitamin D does not remove this harm. – Studies with mixed or unclear results
- Multivitamins, iron, folic acid, B6, magnesium, zinc and copper supplements increase mortality, and folic acid (with B12) increases cancer. – Studies supporting the claim
Use by the Tayebat archive
Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.
- What the archive says about the study
Across 22 years, multivitamins, iron, folic acid, B6, magnesium, zinc and copper were associated with higher mortality.
- A. What was actually studied?
Supplement use vs mortality.
- B. What did the authors find?
Small HRs above 1 for multivitamins, B6, folic acid, iron, magnesium, zinc, copper (several not robust to multiplicity correction); iron most consistent; calcium HR 0.91 (inverse).
- C. What conclusions do the authors themselves draw?
May be associated with increased mortality.
- D. What does the archive derive from it?
Summarised as supplements 'linked to higher mortality' and part of 'clear-cut harm'.
- E. Is this derivation scientifically justified?
Partly: associations correctly listed, but observational, uncorrected for 15 tests, and explicitly not causal according to the authors; the archive omits the inverse association for calcium that conflicts with its calcium claim; the study does not address folic acid/B12 and cancer.
- Adequacy of the interpretation (by the archive)
- partly correct but over-interpreted