The Antioxidant Vitamins and Cardiovascular Disease: A Critical Review of Epidemiologic and Clinical Trial Data

Cardiovascular disease is responsible for about 40% of the deaths in industrialized countries and many deaths in developing countries [1, 2]. Simple, accessible, and cost-effective preventive therapies that decrease the incidence of cardiovascular disease could greatly affect public health. Substantial interest has recently focused on the hypothesis that the naturally occurring antioxidant vitamins such as vitamin E, vitamin C, and -carotene may prevent myocardial infarction, progression of coronary heart disease, or strokes [3, 4]. We critically review the human prospective observational studies and randomized clinical trials of antioxidant vitamins (vitamins E and C and -carotene) in cardiovascular disease, highlighting the differences in the methods and results of the two study designs. Methods Data Sources We searched the scientific literature for all epidemiologic studies (prospective cohort, prospective nested casecontrol, retrospective casecontrol, or geographic correlations) and randomized clinical trials of antioxidants (using the terms vitamin E, vitamin C, -carotene, vitamins, and antioxidants) and cardiovascular disease (using the terms coronary heart disease, cerebrovascular disease, peripheral vascular disease, atherosclerosis, and mortality). We also examined all reports in the cancer literature to identify any studies of the effects of antioxidants on total mortality or other possible cardiovascular outcomes. We searched the MEDLINE database and the Science Citation Index for articles published from 1965 to 1994. We identified additional studies by examining bibliographies of original articles, review articles, and textbooks and by personally contacting the researchers. We also provide in the Appendix a list of the most relevant reviews as a supplement to the references. Study Selection We included only studies that specifically defined the type of intake of antioxidants, that is, studies that quantified the components of dietary intake or supplement use. We excluded studies that did not quantify intake (such as intake of fruits and vegetables or randomized trials of different diets). We used standard definitions to define epidemiologic studies [5] and did not consider case series. Randomized clinical trials were those in which previously established criteria [6] were used to randomly assign patients and to blind the investigators to treatment assignment. Study Synthesis We only selected prospective cohort studies; the rationale for this decision was consistent with discussions of epidemiologic causality [5]. The epidemiologic cohort studies varied greatly in terms of study design, sample size, and assessment of antioxidant intake. To consistently summarize the results of these studies, we emphasized the consistency of the reductions in risk estimates across studies and antioxidant groups; when possible, we reviewed dose-response relations. We excluded retrospective studies because they cannot determine whether reported antioxidant intake or biological levels were altered by the development of cardiovascular disease. We also excluded broad geographic correlations of population-based intake of antioxidants and cardiovascular disease rates because such correlations are retrospective in nature; in addition, extrapolation of these results to individual clinical decisions is difficult. These two types of studies have been reviewed previously [7]. We included all randomized trials with more than 100 participants; only five randomized trials had fewer than 100 participants. These five trials consisted of a total of 268 patients, and all trials had a maximum treatment duration of 6 months. Thus, their exclusion did not alter our results. To consistently present results for randomized clinical trials, we used the one-step modified Mantel-Haenszel-Peto method [6] to estimate a relative risk reduction (RRR) for each trial. Mechanisms of Antioxidants Substantial laboratory, animal, and human data suggest that oxidation of low-density lipoprotein (LDL) cholesterol is an important step in the pathogenesis of atherosclerotic lesions. More extensive reviews have been published [4, 8, 9]. Because oxidized LDL cholesterol is believed to have different properties than nonoxidized LDL cholesterol, LDL cholesterol accumulates in the cells that line the blood vessels. Various chemotactic and proliferative mechanisms lead to fatty streaks and later to atherosclerotic lesions. It is unknown whether oxidation of LDL cholesterol is important in both the initiation and progression of plaque or increases the risk for plaque rupture [10]. The major lipid-soluble antioxidant vitamins are vitamin E (-tocopherol) and -carotene, a precursor of vitamin A. The major water-soluble antioxidant vitamin is vitamin C (ascorbic acid). Vitamin E is important in preventing oxidation of LDL cholesterol. In vitro studies have shown that this process does not begin until oxidative stress depletes the host vitamin E content [11]. -Carotene prevents oxidation of LDL cholesterol [12], although this finding is inconsistent [13]. Vitamin C prevents oxidation of LDL cholesterol and preserves vitamin E and -carotene levels during oxidative stress [14]. A combination of antioxidants is not believed to be clearly more efficacious in preventing oxidation of LDL cholesterol than any one antioxidant alone [13]. Supple-mentation with daily doses greater than 200 IU of vitamin E, 1000 mg of vitamin C, and 25 mg of -carotene increase blood levels of the corresponding vitamin to the same extent as do higher doses of these vitamins [15-17] (1.49 IU of vitamin E equals 1 mg; 10 IU of -carotene equals 1 g of retinol, which equals 1 retinol equivalent. Vitamin C is expressed in mg rather than IU). Antioxidant vitamins presumably exert their effects through protection of oxidation; however, some studies have shown that the vitamins may also preserve endothelial function [18], affect hemostasis [19], and lower both LDL cholesterol levels [20, 21] and blood pressure [22]. Epidemiologic Studies of Antioxidant Vitamins and Cardiovascular Disease In the prospective observational studies, a large group of persons was followed over time and the incidence of cardiovascular disease in persons with high intakes (through diet or vitamin supplementation) or high blood levels of antioxidant vitamins was compared with the incidence in patients with lower intakes or levels. In the nested casecontrol studies, vitamin levels in blood samples collected at baseline were compared among persons who subsequently did and did not develop cardiovascular disease. Vitamin E Prospective Cohort Studies In a large prospective cohort study, the U.S. Nurses' Health Study [23] (Table 1; Figure 1), 87000 female nurses were followed for an average of 8 years. About 13% of women regularly used vitamin E supplements. These women had a statistically significant RRR of 31% (95% CI, 3% to 51%) for nonfatal myocardial infarction and death from cardiovascular disease compared with women who did not use the supplements, after adjustment for age, smoking, alcohol use, menopausal status, hormone use, exercise, aspirin use, hypertension, cholesterol intake, diabetes, caloric intake, and vitamin C and -carotene intake. The absolute risk reduction was 3.4 per 10000 years of follow-up (8.5 compared with 5.2 per 10000 years of follow-up). Table 1. Prospective Observational Studies of Vitamin E and Cardiovascular Disease* Figure 1. Prospective observational studies and randomized trials of vitamin E: effects on cardiovascular end points. Reduced risk was only seen with vitamin E supplementation (at least 100 IU/d) and not with multivitamin use (about 30 IU of vitamin E per day). No dose-response relation was seen with increasing duration of use, presumably because of the small numbers of women who used vitamin E for a prolonged period. However, only use of vitamin E beyond 2 years was associated with significant reductions in the risk for cardiovascular disease. A significantly reduced risk for myocardial infarction or death from cardiovascular disease was seen only with the fourth or highest quintile of vitamin E intake (mean intake for the fourth quintile, 17 IU/d) compared with the lowest quintile (mean intake, 3 IU/d); this corresponded to at least a fivefold difference in intake levels. Although persons using vitamin E more commonly took vitamin C or -carotene, the effect of vitamin E was independent of these antioxidants. A nonsignificant reduction in ischemic stroke was seen with the use of vitamin E supplements (RRR, 29% [CI, 31% to 61%]). Use of vitamin E supplements was found to offer similar protection in another cohort study in which 39000 male health professionals were followed for 4 years [24]. About 17% of the men took vitamin E supplements. Compared with men in the lower quintile of vitamin E intake (median intake, 6 IU/d), men in the upper quintile (median intake, 419 IU/d) had a 40% RRR (CI, 19% to 56%) for nonfatal myocardial infarction, death from coronary heart disease, or coronary revascularization after adjustment for age, smoking, body mass, caloric intake, fiber intake, alcohol use, hypertension, aspirin use, exercise, family history, profession, and vitamin C and -carotene use. The absolute risk reduction was 6.0 per 10000 years of follow-up (50.9 compared with 44.9 per 10000 years of follow-up) among persons using vitamin E supplements. Use of vitamin E supplements for less than 2 years was not associated with a reduced risk for cardiovascular events (RRR, 5%; P > 0.05). As in the Nurses' Health Study, only the two upper quintiles of vitamin E intake were associated with significant risk reductions; the intakes in these quintiles were at least four times the intake in the lowest quintile. In a prospective cohort study in Finland [25], 2748 men and 2385 women were followed for a mean of 14 years. During follow-up, 186 men and 58 women died of coronary heart disease. Only 3% of the st