Miller ER III, Pastor-Burriuso R, Dalal D, Riemersma RA, Appel LJ, Guallar E. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann Intern Med 2005;142(1):37-46.
Research question
Does supplementation with vitamin E increase risk of death?
Type of article and design
Meta-analysis.
Relevance to family physicians
Vitamin E is a commonly used supplement. Patients using vitamin E supplements hope to obtain benefit for many common and serious conditions, such as cardiovascular disease, prostate cancer, and Alzheimer disease. Vitamin E is one of the few complementary agents with evidence of efficacy for certain conditions.1 The proposed action of vitamin E is through its antioxidant effects, which protect the brain and heart from oxidative stress. Evidence from both laboratory studies and epidemiologic data has promoted use of this supplement.
Laboratory research has suggested that oxidative chemical events are the cause of such conditions as cardiovascular disease, dementia, cancer, and chronic lung disease.2-4 This research is supported by epidemiologic and cross-sectional population studies that compare those using high doses of vitamin E with those using low doses and reveal lower rates of specific diseases in high-dose users.2,5
Approximately 22% of adults use vitamin E, usually in the form of alpha-tocopherol, at 400 IU daily.6 Despite its widespread use, no strong evidence supports its beneficial effect on any condition. This disconnection between evidence and the widespread use of vitamin E might be because many patients and physicians strongly believe that vitamin E has few, if any, known side effects. People think, “It won’t hurt and might help, so why not take it?”2
Publication of this meta-analysis and similar studies implicating complementary products in an increased risk of death will likely result in patients visiting their health care providers for assistance in interpreting this new information.
Overview of study and outcomes
This study used standard meta-analysis methods to search for articles meeting the following criteria: the study was randomized; vitamin E was used either alone or in combination with other supplements; duration of use was at least 1 year; and the trial had at least 10 deaths. Of 36 studies abstracted by three independent researchers, 19 qualified for review. This meta-analysis assessed 135 967 patients aged 47 to 84 years. One trial included only men and another trial included only women; these two trials were assessed separately during analysis. Follow up ranged from 1.4 to 8.2 years; median dose of vitamin E was 400 IU/d (16.5 to 2000 IU/d).
Results
There was no difference in all-cause mortality among patients taking vitamin E supplements in general, but when 11 trials with “high doses” were analyzed, the risk ratio for supplementation was 1.04. High-dose administration was defined as > 400 IU/d. The all-cause mortality risk difference was 39/10 000 people (95% confidence interval 3 to 74, P = .035). In dose-response analysis, mortality increased with doses >150 IU/d. A non-significant decrease in mortality was associated with doses <150 IU/d. Factors that could have confounded results, such as sex, age, and follow up, did not affect the conclusions. No one study dominated the findings when analysis was done removing results from each of the individual trials in turn.
Analysis of methodology
Various authors list criteria for appraising meta-analyses. Five criteria are commonly cited. First, is the design appropriate? This study used only randomized controlled trials and was fortunate to find several large, well performed multicentre trials.
Second, was the search strategy comprehensive? The authors explain that they did an in-depth search using clearly stated methods. This included contacting investigators of the primary studies for information not available in published reports and including all relevant trials.
Third, were important outcomes specified? All-cause mortality was the specified outcome and is the most reliably measured health outcome. Specific determinants of outcome, such as dose, were also collected. Positive results, such as a non-statistical trend to benefit for doses below 150 IU/d, could be determined. In addition, secondary guides to validity, such as temporal relationships and a dose gradient, were also demonstrated. Both the linear and quadratic components of the dose-response gradients were statistically significant.
Fourth, is data collection appropriate? The search strategy was well stated, but specific quality criteria were not applied to the primary studies. Assessing their individual validity is thus difficult, yet a review of primary studies meeting the inclusion criteria shows they are high-quality studies. In addition, a sensitivity analysis was done that showed that no one study affected the final results if it was removed from the analysis.
Finally, was it reasonable to combine the results? A potential weakness of this meta-analysis could be that results of trials on heterogeneous populations (patients with Alzheimer disease, Parkinson disease, and cancer) were combined. The effect of vitamin E, however, was similar from study to study (test of heterogeneity was negative). Therefore, pooling of results is reasonable and including heterogeneous studies strengthens the generalizability of the results.
Application to clinical practice
How sure are we that vitamin E increases the chance of death? Should we be telling tens of millions of our patients to stop taking their medications? The issue is difficult for physicians who would like to see patients involved in their own care and to be perceived as open to alternative comprehensive care. Telling a patient who fears Alzheimer disease to discontinue a drug that might prevent it based on a relatively complex methodology, such as a meta-analysis, can be daunting.
To make this decision, we have to ask how great the risk to patients is and what the adverse consequences of reducing or eliminating the harmful medications are. The authors of this trial demonstrated a statistically significant excess in all-cause mortality of 39 per 10 000 people when patients took high doses (>400 IU/d) of vitamin E (confidence interval did not cross unity). Though this risk seems small, given the outcome and the number of patients potentially taking the product, a reasonable health policy would be to recommend against use of high-dose vitamin E.
This recommendation must be balanced against the likely benefits to individual patients who could be at high risk of specific diseases, such as prostate cancer, where the risk of death from the disease might be much higher than risks from other causes. Findings of this study are consistent with those of other studies7-9; they suggest that antioxidants are more harmful than beneficial. In fact, vitamin E might displace other fat-soluble antioxidants resulting in an increase in oxidative damage. Our message to our patients should be based on the evidence, which points to harm as opposed to benefit.
Bottom line
Vitamin E supplementation at doses above 400 IU/d is associated with a small increase in all-cause mortality.
Patients taking high doses of vitamin E should be counseled to discontinue or reduce their medication.
Critical Appraisal reviews important articles in the literature relevant to family physicians. Reviews are by family physicians, not experts on the topics. They assess not only the strength of the studies but the “bottom line” clinical importance for family practice.
Biographies
Dr Lock teaches at the University of Western Ontario and is Interim Head of the Department of Radiation Oncology at the London Regional Cancer Program.
Dr Loblaw is a clinician-scientist working at the University of Toronto’s Toronto-Sunnybrook Regional Cancer Centre.
References
- 1.Zandi PP, Anthony JC, Khachaturian AS, Stone SV, Gustafson D, Tschanz JT, et al. Reduced risk of Alzheimer disease in users of antioxidant vitamin supplements: the Cache County Study. Arch Neurol. 2004;61:82–88. doi: 10.1001/archneur.61.1.82. [DOI] [PubMed] [Google Scholar]
- 2.Greenberg ER. Vitamin E supplements: good in theory, but is the theory good? Ann Intern Med. 2005;142:75–76. doi: 10.7326/0003-4819-142-1-200501040-00112. [DOI] [PubMed] [Google Scholar]
- 3.Finkel T, Holbrook NJ. Oxidants, oxidative stress and the biology of ageing. Nature. 2000;408:239–247. doi: 10.1038/35041687. [DOI] [PubMed] [Google Scholar]
- 4.Klaunig JE, Jamendulis LM. The role of oxidative stress in carcinogenesis. Ann Rev Pharmacol Toxicol. 2004;44:239–267. doi: 10.1146/annurev.pharmtox.44.101802.121851. [DOI] [PubMed] [Google Scholar]
- 5.Stanner SA, Hughes J, Kelly CN, Buttriss J. A review of the epidemiological evidence for the ‘antioxidant hypothesis.’. Public Health Nutr. 2004;7:407–422. doi: 10.1079/phn2003543. [DOI] [PubMed] [Google Scholar]
- 6.Miller AE, Dodd KW, Subar AF. Use of vitamin, mineral, non-vitamin, and non-mineral supplements in the United States. The 1987, 1992, and 2000 National Health Interview Survey Results. J Am Diet Assoc. 2004;104:942–950. doi: 10.1016/j.jada.2004.03.022. [DOI] [PubMed] [Google Scholar]
- 7.Vivekananthan DP, Penn MS, Sapp SK, Hsu A, Topol EJ. Use of antioxidant vitamins for the prevention of cardiovascular disease: meta-analysis of randomized trials. Lancet. 2003;361:2017–2023. doi: 10.1016/S0140-6736(03)13637-9. [DOI] [PubMed] [Google Scholar]
- 8.Shekelle PG, Morton SC, Jungvig LK, Udani J, Spar M, Tu W, et al. Effect of supplemental vitamin E for the prevention and treatment of cardiovascular disease. J Gen Intern Med. 2004;19:380–389. doi: 10.1111/j.1525-1497.2004.30090.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Eidelman RS, Hollar D, Hebert PR, Lamas GA, Hennekens CH. Randomized trials of vitamin E in the treatment and prevention of cardiovascular disease. Arch Intern Med. 2004;164:1552–1556. doi: 10.1001/archinte.164.14.1552. [DOI] [PubMed] [Google Scholar]