Vitamin E and the Age-Related Eye Disease Study Supplementation for Age-Related Macular Degeneration

An oral presentation at the November 2004 meeting of the American College of Physicians in New Orleans, La, which summarized a meta-analysis of controlled clinical trials of vitamin E, received much attention from the press, with headlines such as “Vitamin E Supplements Might Kill You” and “Vitamin E Can Be Deadly.” The article,¹ published in the January 2005 edition of the Annals of Internal Medicine, reports that taking high-dose vitamin E supplements does not lower the risk of death and might be associated with a small increase in the risk of death. These study conclusions have resulted in confusion among the treating physicians and persons who are at moderate or high risk for developing age-related macular degeneration (AMD) and who are taking the Age-Related Eye Disease Study (AREDS) supplements that include 400 international units (IU) of vitamin E daily.²

In this new study, the authors conducted a meta-analysis to determine the risk of death by combining data from 19 clinical trials, which tested vitamin E as a possible treatment for a variety of diseases. One of those 19 studies was the AREDS. The authors conclude in their abstract that “high-dosage (≥400 IU/d) vitamin E supplements may increase all-cause mortality.”^{1 (p37)} Their data do seem to show that persons taking particularly high dosages of vitamin E (500-2000 IU/d) over varying periods may have some increased risk of death, but even at these high dosages the risks observed for patients ranged from a benefit of reduced mortality of 2% to an increased risk of mortality of 5%. In addition, it is not clear that this possible increased risk with high dosages would apply to persons taking 400 IU/d, the dosage in the AREDS formulation.

In their analysis, the authors divided their population into 2 groups of vitamin E consumption, low dose, less than 400 IU/d, and high dose, 400 IU/d or more. Inclusion of 400 IU, a common dose in many vitamin E preparations, in the high-dose group rather than the low-dose group seems somewhat arbitrary. The dosage of 400 IU/d was the median dosage for the 19 studies included in their meta-analysis. (The median dosage for the 135 976 participants included in the meta-analysis was about 60 IU/d.) Of all the studies included in this new analysis of vitamin E, 3 studies, including the AREDS, evaluated dosages of about 400 IU/d of vitamin E (2 of 400 IU/d and 1 of 440 IU/d).^{3 - 4} There were more than 15 000 patients followed in these 3 studies and more than 1600 total deaths in this generally elderly population. In total, the group taking vitamin E was actually just slightly more likely to be living after 5 years (801 deaths of 7564 persons in the vitamin E group and 806 deaths of 7598 in the placebo group) (Table). The pooled risk ratio for these 3 studies is 0.998 with a risk difference of –1.8 per 10 000 persons (in the direction of benefit of reduced mortality).

The authors used a statistical model to estimate risk over the range of doses studied. This statistical model is dependent on assumptions regarding the dose-response behavior that may or may not be true. For example, the slope of the beneficial risk ratio at low doses of vitamin E is driven in large part by the results seen in populations likely to be vitamin deficient. This model produces estimates of risks that may overestimate or underestimate the actual risk at specific vitamin E levels in the United States. A look at the data available at dosage levels near 400 IU/d, a dosage commonly used by the community, does not suggest the type of risk estimated by the statistical model.

If the only reason for taking vitamin E is to reduce one’s mortality risk, we agree with the authors that there seems to be little evidence of benefit and some possible risk of harm at very high doses. However, the decision of whether to take the AREDS formulation must balance possible risks with possible benefits of the complete formulation. The AREDS formulation contains beta-carotene (15 mg), vitamin C (500 mg), zinc (80 mg as zinc oxide), and copper (2 mg as cupric oxide), in addition to vitamin E. A review of the mortality experience in the AREDS shows that those taking the AREDS formulation had a 14% reduction in mortality risk after an average of 6.5 years of supplementation compared with placebo (relative risk, 0.86; 95% confidence interval, 0.65-1.12) (Table).⁵ For these reasons, combined with little evidence from the meta-analysis of harm at the 400 IU/d dosage, we do not think there is increased risk of mortality associated with the AREDS supplements that include 400 IU/d of vitamin E. The benefits of the AREDS supplementation for those at risk for advanced AMD include a 25% reduction in the risk of developing advanced AMD and a 19% reduction in the risk of moderate vision loss.² Persons who are at risk for advanced AMD (ie, those who have extensive intermediate drusen or large drusen, or advanced AMD—either neovascular or central geographic atrophy—in 1 eye) should still consider taking the AREDS formulation to reduce the risk of development of advanced AMD. Patients taking multiple drugs secondary to other chronic diseases should consult their physicians, and smokers should probably not take the AREDS formulation because of the high-dose beta carotene. A letter prepared for the AREDS participants explaining the results of the meta-analyses of vitamin E and the use of the AREDS formulation can be found at http://spitfire.emmes.com/study/areds/.

The National Eye Institute is planning a new study to increase our understanding of the role of the vitamins and minerals used in the AREDS formulation as well as potential contributions of lutein/zeaxanthin and ω-3 long-chain polyunsaturated fatty acids in delaying the progression of vision loss. Such a study will provide evidence-based treatment strategies for patients at risk for developing advanced AMD.

Correspondence: Dr Chew, Division of Epidemiology and Clinical Research, National Institute of Health, Bldg 10 CRC, Room 3-2531, 10 Center Dr, MSC-1204, Bethesda, MD 20892-1204 (echew@nei.nih.gov).