Decision analysis supports the paradigm that indiscriminate supplementation of vitamin E does more harm than good.

OBJECTIVE For many years, the prevailing concept was that LDL oxidation plays a central role in atherogenesis. As a consequence, supplementation of antioxidants, particularly vitamin E, became very popular. Unfortunately, however, the major randomized clinical trials have yielded disappointing results on the effects of vitamin E on both mortality and morbidity. Moreover, recent meta-analyses have concluded that vitamin E supplementation increases mortality. This conclusion has raised much criticism, most of it relating to three issues: (1) the choice of clinical trials to be included in the meta-analyses; (2) the end point of these meta-analyses (only mortality); and (3) the heterogeneity of the analyzed clinical trials with respect to both population and treatment. Our goal was to bring this controversy to an end by using a Markov-model approach, which is free of most of the limitations involved in using meta-analyses. METHODS AND RESULTS We used a Markov model to compare the vitamin E supplemented virtual cohorts with nonsupplemented cohorts derived from published randomized clinical trials that were included in at least one of the major meta-analyses. The difference between the virtual supplemented and nonsupplemented cohorts is given in terms of a composite end point denoted quality-adjusted life year (QALY). The vitamin E supplemented virtual cohort had 0.30 QALY (95%CI 0.21 to 0.39) less than the nontreated virtual cohort. CONCLUSIONS Our study demonstrates that in terms of QALY, indiscriminate supplementation of high doses of vitamin E is not beneficial in preventing CVD. Selective supplementation of vitamin E to individuals under oxidative stress requires further investigation.