• DOI:10.7326/0003-4819-142-1-200501040-00110
• Corpus ID: 35030072

Meta-Analysis: High-Dosage Vitamin E Supplementation May Increase All-Cause Mortality

@article{Miller2005MetaAnalysisHV,
  title={Meta-Analysis: High-Dosage Vitamin E Supplementation May Increase All-Cause Mortality},
  author={Edgar R. Miller and Roberto Pastor-Barriuso and Darshan Dalal and Rudolph A. Riemersma and Lawrence J. Appel and Eliseo Guallar},
  journal={Annals of Internal Medicine},
  year={2005},
  volume={142},
  pages={37-46}
}

• E. Miller, R. Pastor-Barriuso, +3 authors E. Guallar
• Published 4 January 2005
• Medicine
• Annals of Internal Medicine

Context Does vitamin E supplementation increase mortality in adults? Contribution This meta-analysis of 19 randomized, controlled trials involving more than 135000 participants found that high-dosage vitamin E supplementation (400 IU/d for at least 1 year) increased all-cause mortality. Benefits or harms of lower-dosage supplementation were unclear. Cautions Trials that tested high dosages involved adults with chronic diseases, and these findings may not be generalizable to healthy adults. Some trials evaluated multivitamin combinations. The findings don't clearly establish that the lowest dosage of supplementation is associated with increased mortality risk. Implications Avoid high-dosage vitamin E supplementation. The Editors On the basis of the premise that vitamin E reduces oxidative stress, many clinical trials have tested vitamin E supplementation as a therapy to prevent various chronic diseases. The results of these trials have been largely disappointing (1-3). Three recent meta-analyses, which did not consider doseresponse relationships, reported no overall effect of vitamin E on survival (4-6). However, several trials of high-dosage vitamin E supplementation have reported nonstatistically significant increases in total mortality. Because individual trials typically tested only 1 dosage of vitamin E and large-scale trials with several dosages are not feasible, we performed this doseresponse meta-analysis to evaluate a potential dose-dependent effect of vitamin E supplementation. We focused on all-cause mortality because this end point has unambiguous clinical relevance and, in contrast to cause-specific events such as cardiovascular morbidity or death, is resistant to miscoding. Methods Search Strategy and Inclusion Criteria We searched for all reports of clinical trials (with no language restrictions) that tested the effect of vitamin E supplementation in humans. We performed a MEDLINE search by using the Medical Subject Heading (MeSH) terms vitamin E, antioxidant vitamins, alpha tocopherol, tocopherol, and clinical trials. The search period was 1966 through August 2004. We complemented the MEDLINE search by searching the Cochrane database of randomized, controlled trials; reviewing the reference lists from original research, review articles, and previous meta-analyses; and reviewing the files of the investigators. Our prespecified inclusion criteria were 1) random allocation of participants, 2) use of vitamin E supplementation alone or combined with other vitamins or minerals, 3) presence of a control or placebo group, 4) study sample limited to men or nonpregnant women, 5) duration of vitamin E supplementation and follow-up longer than 1 year, and 6) occurrence of at least 10 deaths in the trial. The restriction to include studies with follow-up longer than 1 year and at least 10 deaths was determined a priori because we anticipated that many small trials did not collect mortality data. We contacted the investigators of the original studies if information required to establish inclusion criteria or information on trial design or mortality results was not available in the published reports. Three investigators independently abstracted the articles. They resolved disagreements by consensus. We converted the dosage of vitamin E in each trial to international units per day (IU/d) (7). Only 1 trial (8) used 2 vitamin E dosage groups (400 and 800 IU/d); however, all-cause mortality data were available for both vitamin E groups combined (9), not for the separate dosages. For this trial, we assigned all deaths in the 2 vitamin E groups to the average dosage of 600 IU/d. Statistical Analysis We conducted all analyses according to the intention-to-treat principle. For trials with a factorial design, we based main results on 2-way analyses, that is, all trial participants receiving vitamin E were compared with all participants not receiving it. We used a 2-level hierarchical logistic regression model (10, 11) to evaluate the association between vitamin E supplementation and all-cause mortality. In the first within-study level, we specified a logistic model for the probability of death as a function of randomized assignment to vitamin E treatment in each trial. In the second between-study level, we first introduced an indicator variable for high-dosage vitamin E trials (400 IU/d) to evaluate the differences in effect of high- and low-dosage vitamin E supplementation. For doseresponse analyses, we replaced the above indicator variable with a quadratic-linear spline model (12) for log-dose of vitamin E. The quadratic-linear spline assumed a quadratic trend below the median vitamin E dosage across studies (400 IU/d) and a linear trend more than this dosage. We selected the lower quadratic term to allow for nonlinear responses, while ensuring a null effect for 0 dose level. We restricted the upper term to be linear to avoid implausible shapes at high dosages of vitamin E. Thus, this quadratic-linear spline model is more flexible, less influenced by extreme values, and more plausible than the usual linear or quadratic polynomial models (12-14). We performed several sensitivity analyses to evaluate the robustness of the results. First, we replicated the analyses assuming different knot values (100, 200, 300, 400, or 500 IU of vitamin E per day) for the quadratic-linear, doseresponse model. The results were virtually insensitive to the selected knot for the quadratic-linear model, and the shape of the doseresponse curve remained essentially unchanged (not shown). Then, we repeated the analyses on the basis of 4-way analysis of factorial trials. In this analysis, we restricted data from factorial trials to participants who were not exposed to the second factorial intervention. In addition, we included other study-specific explanatory variables as second-level covariates in the categorical and doseresponse hierarchical models. Because of the limited number of trials, we separately evaluated the effect of adding the following variables: sex distribution, mean age, use of other vitamins or minerals combined with vitamin E, and average duration of follow-up. Finally, we evaluated the influence of each trial on the results by removing each individual study from the analysis. In all analyses, we estimated and inferred vitamin E effects by using population-average models with robust standard errors (11, 15). Population-average models estimate the effects at different vitamin E dosages averaged across all trials, whereas robust standard errors are relatively insensitive to model misspecification. We transformed the logistic model results into risk differences and risk ratios of the effect of vitamin E supplementation compared with control for easier clinical interpretability. We obtained CIs for risk differences and risk ratios by using the delta method. We performed analyses by using S-PLUS 2000 (Mathsoft, Inc., Seattle, Washington) (16) and HLM 5 (Scientific Software International, Lincolnwood, Illinois) (17). Role of the Funding Sources The funding sources had no role in the collection, analysis, and interpretation of the data or in the decision to submit the manuscript for publication. Results Study Description Figure 1 summarizes the trial selection process. We identified 36 randomized, controlled trials with follow-up longer than 1 year. We excluded 12 trials with fewer than 10 deaths (18-29), 3 trials with unavailable mortality data (30-32), and 2 trials in which mortality data were not separated from a composite end point (33, 34). Nineteen trials (8, 35-53) met our inclusion criteria (Table 1). The selected trials were reported between 1993 and 2004, and their sample sizes varied between 196 and 29584 participants. The total number of participants was 135967, with 12504 all-cause deaths. The reported mean age ranged from 47 to 84 years. Seventeen trials included men and women, 1 trial included only men, and 1 trial included only women. Most trials targeted populations at high risk for a chronic disease, most often coronary heart disease. Nine of 19 trials used vitamin E alone, while the other 10 trials combined vitamin E with other vitamins or minerals. All but 3 trials (42, 43, 51) were placebo-controlled and double-blind. Average follow-up ranged from 1.4 to 8.2 years. Vitamin E dosage varied between 16.5 and 2000 IU/d, with a median dosage of 400 IU/d. High-dosage (400 IU/d) trials, however, were smaller on average, and only 40thinsp;950 patients (30.1% of all randomly assigned patients) participated in high-dosage trials. Figure 1. Flow diagram of the trial selection process. Table 1. Clinical Trials of Vitamin E Supplementation and Risk for All-Cause Mortality, Ordered by Dosage of Vitamin E Combined Effect of Vitamin E on All-Cause Mortality The average death risk across trials in the control groups was 1022 per 10000 persons. Overall, vitamin E supplementation did not affect all-cause mortality. The pooled risk difference comparing vitamin E with control was 10 per 10000 persons (95% CI, 18 to 38 per 10000 persons) and the risk ratio was 1.01 (CI, 0.98 to 1.04; P> 0.2) (Figure 2). However, there was significant heterogeneity of study results (P= 0.020) that was explained by differences in results between low-dosage (<400 IU/d) and high-dosage (400 IU/d) vitamin E trials (P= 0.18 for residual heterogeneity). In 8 trials testing low-dosage vitamin E supplementation, the pooled risk difference comparing vitamin E with control was 16 per 10000 persons (CI, 41 to 10 per 10000 persons) and the risk ratio was 0.98 (CI, 0.96 to 1.01; P> 0.2). In 11 trials testing high-dosage vitamin E, the pooled risk difference was 39 per 10000 persons (CI, 3 to 74 per 10000 persons) and the risk ratio was 1.04 (CI, 1.01 to 1.07; P= 0.035) (Figure 2). Figure 2. Risk difference in all-cause mortality for randomized, controlled trials of vitamin E supplementation and pooled results for low-dosage (<400 IU/d) and high-dosage (400 IU/d)