Response to Letter Regarding Article, “Effects of Random Allocation to Vitamin E Supplementation on the Occurrence of Venous Thromboembolism: Report From the Women’s Health Study”
We thank Drs Violi and Pignatelli for discussing additional mechanisms whereby supplementation with vitamin E might influence the risk of venous thromboembolism (VTE). The primary end points of the Women’s Health Study were cardiovascular disease and cancer, and the a priori mechanisms that motivated the trial focused on these end points. Although it was prospectively evaluated and its occurrence required confirmation through record review by the trial’s end points committee, VTE was a secondary end point, and potential mechanisms identified after trial initiation should be interpreted cautiously.
The different relationships of vitamin E with venous versus arterial events in the Women’s Health Study, along with the alternative mechanisms described by Drs Violi and Pignatelli, suggest that vitamin E might differentially affect risk of venous versus arterial events. In further support of this possibility, prospective observational studies have found quite different relationships of cardiovascular risk factors with risk of VTE.1,2 Our trial3 provides no information on potential interference with oral anticoagulants by vitamin E because women taking anticoagulants at baseline were excluded from participation.
Drs Violi and Pignatelli also comment on subgroup findings in our trial, notably our age-specific results. Although it is intriguing that subgroups at the highest absolute risk of VTE (ie, those with a history of VTE, the factor V Leiden or prothrombin mutation, and older women) had risk reductions somewhat larger than the overall treatment effect, the most striking aspect of subgroup analyses was that confidence intervals for all subgroups broadly overlapped the overall trial estimate of a 21% reduction in the rate of VTE associated with vitamin E. Overall, the Women’s Health Study found a substantial benefit of vitamin E treatment in its entire population, with no statistically significant interactions indicative of heterogeneity across subgroups. Nonetheless, we agree that absolute levels of risk should play a role in treatment decisions and that use of vitamin E is unwarranted for those at low absolute risk. In addition to age, prior history and genetic predisposition are important factors that might indicate use of vitamin E for prophylaxis against VTE.
Glynn RJ, Rosner B. Comparison of risk factors for the competing risks of coronary heart disease, stroke, and venous thromboembolism. Am J Epidemiol. 2005; 162: 975–982.
Glynn RJ, Ridker PM, Goldhaber SZ, Zee RY, Buring JE. Effects of random allocation to vitamin E supplementation on the occurrence of venous thromboembolism: report from the Women’s Health Study. Circulation. 2007; 116: 1497–1503.