| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Supplements |
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONHILL CRITERIAUS PREVENTIVE SERVICES TASK...SUMMARYREFERENCES |
|---|
Key Words: 
-Tocopherol • vitamin E • coronary heart disease • epidemiologic studies • prospective studies • causal inference
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONHILL CRITERIAUS PREVENTIVE SERVICES TASK...SUMMARYREFERENCES |
|---|
One set of criteria by which epidemiologic studies can be evaluated to determine whether there is sufficient evidence to infer a causal relation between dietary factors and risk of chronic diseases was outlined in the National Research Council's report
Diet and Health: Implications for Reducing Chronic Disease Risk
(
1
). In this report, the Committee on Diet and Health presented 6 of the criteria suggested by Hill (
6
) as a principal basis for evaluating the evidence. These are familiar to most students of epidemiology and are listed in
Table 1
. If causality can be inferred, then there is a basis for developing public health recommendations.
|
). For example, if an association is established between a dietary exposure and disease outcome in a randomized, controlled trial, that is taken as the best evidence of causality to support development of recommendations. In the absence of randomized trials, evidence from analytic epidemiologic studies (ie, cohort or case-control studies) is considered to be the best source of information. In studies of dietary exposures, cohort studies—in which food intake is measured before onset of disease—are considered superior to case-control studies, because the latter may have biased conclusions because of differential recall of past dietary habits (
8
), although whether this occurs and has practical implications is controversial (
9
). The least impressive evidence to support recommendations would be opinions of experts or expert committees in the absence of adequate human studies. In recent years there has been considerable interest in the possible role of dietary antioxidants in the etiology and prevention of coronary heart disease. Although many antioxidant compounds are found in the diet, vitamin E has been the focus of much of this research. This article is focused on the application of the Hill and US Preventive Services Task Force criteria to the association of dietary vitamin E (whether from food or supplements) and coronary heart disease. Note that this article is not meant to be a comprehensive review of this literature; for example, no studies related to plasma or tissue amounts of vitamin E and risk of coronary disease are included. Rather, it is intended as an illustration of how these criteria may be applied to determine causality while also showing the nature of the evidence related to a recommendation to consume vitamin E to prevent coronary heart disease.
|
HILL CRITERIA |
|---|
|
TOPABSTRACTINTRODUCTIONHILL CRITERIAUS PREVENTIVE SERVICES TASK...SUMMARYREFERENCES |
|---|
. Other characteristics of these 6 studies are also shown in this table.
|
10-fold relative risk of lung cancer mortality in those who smoke 1 pack cigarettes/d compared with nonsmokers in most prospective studies that reported on this association (
16
). For vitamin E intake from supplements, the relative risk for those who had the highest reported dose of daily supplement intake compared with those who did not take vitamin E supplements ranged from 1.09 to 0.53. Another study reported a relative risk of 0.21 for supplement users compared with nonusers, but this estimate was not adjusted for potentially confounding variables ( 15 ). Thus, the strongest reported association corresponded to an approximate halving of the risk of coronary disease in those who took vitamin E supplements compared with those who did not.
Dose-response relation
Relative risk estimates of coronary heart disease with increasing intake of vitamin E, again from foods, supplements, or both combined are shown in
Table 3
. Four prospective studies presented data on foods that allowed investigation of the effect of dosage of vitamin E intake from foods, 3 presented information regarding dosage of supplemental vitamin E intake, and 2 compared those who took supplements with those who did not and presented no information on dosage.
|
Of the other 2 studies, one somewhat supported an inverse association of dietary vitamin E and risk of coronary disease, with the relative risk of 0.79 for a comparison of the highest fifth of intake with the lowest; however, there was no clear dose-response relation, and relative risks from lowest to highest fifths were 1.0, 1.10, 1.17, 0.97, and 0.79 ( 11 ). The other study did not support an association of dietary vitamin E intake with coronary disease ( 10 ). However, this latter study was based on a food-frequency questionnaire that listed about half as many food items as did the questionnaires in the other 3 studies, and it may not have accurately assessed individual vitamin E intake from foods.
None of the studies strongly supported a dose-response relation for supplemental intake with coronary disease. In 2 of the studies it appeared that although intake of any supplements (except for low doses) was inversely associated with risk of coronary disease, there was little evidence of a dose response (
10
,
11
). For example, in a study of female nurses, relative risks of coronary disease were 0.56, 0.56, and 0.58 in those taking daily doses of 100–250, 300–500, and 600 IU (1 IU = 1 mg -TE), respectively, compared with those who did not take supplements (
10
). In the other study, the relative risks of coronary disease in male health professionals in the highest 3 categories of supplemental intake compared with those who took no supplements were 0.78, 0.54, and 0.70 (
11
). The third study with dose information did not observe an association of vitamin E supplement intake with mortality from coronary heart disease (
13
).
Temporally correct association
The studies presented in
Tables
2
and
3
are all prospective cohort studies; participants with coronary disease at baseline were excluded from follow-up. Thus, in each of these studies, dietary habits were assessed before the occurrence of disease or death, and the results of these studies are directly relevant to the question of whether vitamin E intake influences the risk of subsequent coronary heart disease. As noted below, results from these studies are consistent with the theory that vitamin E intake from foods and supplements is associated with decreased risk of coronary heart disease.
Not all studies of diet and coronary disease satisfy this criterion, however. One example is the Scottish Heart Health Study, a cross-sectional survey of diet and coronary disease in 10359 men and women ( 17 ). In this study, dietary habits were assessed in participants by using a 50-item food-frequency questionnaire. Participants were classified as having diagnosed coronary heart disease if they reported that they had medically diagnosed angina or myocardial infarction. They were classified as having undiagnosed coronary heart disease if they answered appropriately on the angina questionnaire of Rose et al ( 18 ) or if they had abnormal electrocardiograph results. All remaining participants were classified as control subjects. Dietary habits, including vitamin E intake, were then compared among the 3 groups.
The investigators in this study recognized that study participants with diagnosed coronary heart disease had probably modified their diets as a result of their diagnosis; however, they suggested that comparisons between the undiagnosed coronary disease group and the control subjects may reflect etiologically meaningful differences ( 17 ). For vitamin E intake, for example, the odds ratio for diagnosed coronary heart disease increased significantly with increasing vitamin E intake, whereas the odds ratio for undiagnosed coronary heart disease decreased modestly with increasing vitamin E intake ( 17 ). However, participants, including those with undiagnosed coronary heart disease, may have altered their diets as a result of their symptoms, and thus this study is inadequate for providing support for causality.
Consistency of association
Generally, the more studies that observe a similar relation between an exposure and disease, preferably in different study populations and perhaps with different study designs, the more likely it is that the association is causal. The information in
Tables
2
and
3
can be examined to determine whether there is consistency of the association of vitamin E with coronary heart disease among the prospective cohort studies that have examined this relation. Among the 4 studies and 5 cohorts to report associations of vitamin E intake from foods and coronary heart disease, only one, the Nurses' Health Study (
10
), did not suggest an inverse association of vitamin E intake with coronary heart disease incidence (
Table
2
). Among the studies reporting on the association of supplemental vitamin E intake, only one, the Iowa Women's Health Study (
13
), did not suggest an inverse association of supplemental vitamin E with coronary disease. In the Nurses' Health Study, vitamin E from foods was not inversely associated with coronary disease, but supplemental vitamin E was; conversely, in the Iowa study, vitamin E from foods but not supplements was inversely associated with coronary heart disease. Thus, overall, among the 6 prospective studies to examine the association of vitamin E intake with coronary disease, all reported some evidence of an inverse association of some measure of vitamin E intake with risk of coronary disease.
Specificity of association
In all but one of the studies listed in
Table
2
, the reported associations of vitamin E with coronary heart disease were adjusted for several other coronary heart disease risk factors. Thus, the association was generally independent of other known and measured risk factors for coronary heart disease. However, in only 2 of the studies was information available on blood cholesterol concentrations, a known risk factor for coronary heart disease (
12
,
15
); in one of these, only crude associations of vitamin E intake with coronary mortality were presented (
15
).
Coronary heart disease is recognized as a multifactorial disease entity. Among dietary factors, dietary cholesterol ( 10 , 19 , 20 ), various fatty acids ( 21 , 22 ), and dietary fiber ( 20 , 23 , 24 ) have been recognized as factors that influence the risk of coronary heart disease. Similarly, vitamin E and possibly other vitamins and minerals may influence the risk of coronary heart disease. Vitamin E may also influence the risk of other diseases, including cancer ( 25 ).
In studies of dietary factors and chronic diseases such as coronary heart disease, the criterion of specificity of association is rarely met. A specific factor such as inadequate vitamin E intake is not the only cause of the disease, and coronary heart disease is not necessarily the only consequence of low vitamin E intakes. However, this also holds for well-established risk factors for coronary heart disease, such as cigarette smoking. Heart disease risk may still be elevated for other reasons in nonsmokers, and cigarette smoking is recognized as a cause of several other illnesses. Thus, in the context of chronic diseases with multifactorial etiologies, failure to meet the Hill criterion of specificity of association does not necessarily mitigate inference that an association is causal.
For dietary variables, the ability to ascribe specificity of an association to a given dietary factor is complicated by the multicollinearity inherent in dietary exposures. Specifically, nutrients and other dietary factors are usually consumed as foods rather than as discrete items. For example, foods that are high in vitamin E also tend to be high in polyunsaturated fatty acids. Thus, associations that are observed for vitamin E may be results of other dietary factors. Although this multicollinearity is avoided to a large extent in studies of supplements, confounding by other factors that may by correlated with supplement-taking behavior may remain.
Biological plausibility
It is increasingly being recognized that oxidized LDL has greatly enhanced ability to infiltrate the subendothelial layer of arteries and that this oxidative modification also enhances the uptake of LDL by macrophages, thereby creating foam cells (
26
). Factors that may prevent the oxidative modification of LDL may therefore inhibit the development of foam cells, fatty streaks, atherosclerosis, and coronary heart disease. Vitamin E is a potent lipid-soluble antioxidant that is carried in lipoproteins; this provides a plausible biological mechanism by which increased vitamin E intake may decrease risk of coronary heart disease. High levels of vitamin E intakes may result in high concentrations of vitamin E in lipoprotein particles, thereby inhibiting oxidative modification of these lipoproteins.
Although there are plausible biological mechanisms by which vitamin E may decrease risk of coronary heart disease, it should be stressed that recognition of biological mechanisms does not adequately determine that a causal association exists. The ß-carotene drama provides a cautionary tale in this regard. There were biologically plausible mechanisms for a protective effect of ß-carotene on carcinogenesis related to its antioxidant capabilities and its role as a precursor of vitamin A ( 25 ). There was also supporting epidemiologic evidence from case-control studies of vegetable and fruit intake showing that high ß-carotene intake was associated with decreased risk of lung cancer ( 27 ) and from nested case-control studies comparing blood concentrations of ß-carotene in people who subsequently developed lung cancer with those in the same cohort who did not ( 28 ). However, in 2 of 3 large, randomized trials of ß-carotene supplementation and lung cancer risk, subjects randomly assigned to receive ß-carotene supplementation had a higher rate of lung cancer than did control subjects ( 29 , 30 ). In the other study, there was no significant difference in lung cancer rates between treatment groups ( 31 ). This suggests that evidence from studies of different design, not just case-control or prospective studies, may be required to establish causality.
|
US PREVENTIVE SERVICES TASK FORCE CRITERIA |
|---|
|
TOPABSTRACTINTRODUCTIONHILL CRITERIAUS PREVENTIVE SERVICES TASK...SUMMARYREFERENCES |
|---|
Criterion I: evidence obtained from at least one properly designed randomized controlled trial
Only 1 randomized, controlled trial of vitamin E intake and primary prevention of coronary heart disease has been published. This study is the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study, a randomized, placebo-controlled trial of -tocopherol and ß-carotene in 29133 male cigarette smokers in Finland (
29
). Approximately equal numbers of participants were randomly assigned to one of 4 groups: those who received vitamin E supplements (50 mg/d) alone, those who received ß-carotene supplements (20 mg/d) alone, those who received vitamin E and ß-carotene, and those who received placebo. Of relevance to the vitamin E–related analyses, 14564 men received vitamin E supplements and 14569 did not. Results for different endpoints in the ATBC Study and from a secondary prevention trial (
32
) are shown in
Table 4
.
|
In addition to effects on mortality, the effects of vitamin E and ß-carotene supplements on occurrence of angina pectoris were also examined in the ATBC Study ( 33 ). In these analyses, 6864 men with evidence of coronary heart disease at baseline who were excluded from analysis were evenly distributed among the study's 4 treatment groups. As with coronary mortality, men who received vitamin E supplements were at slightly reduced risk of angina compared with those who did not, with a relative risk of 0.91. This relative risk was attenuated to 0.97 when differences in rates of angina were compared between those who received only vitamin E and those who received placebo because men who received ß-carotene supplementation experienced a slightly higher rate of angina pectoris than those who did not receive ß-carotene. Overall, this study provides only modest evidence in support of a protective effect of vitamin E supplements against coronary heart disease.
There have been at least 2 secondary prevention trials of vitamin E supplementation and coronary heart disease. One of these was an analysis in the ATBC Study of the men with baseline evidence of myocardial infarction; 1862 such men were randomly assigned in roughly equivalent numbers to the 4 study treatments ( 34 ). Overall, the men who received only vitamin E supplements experienced a significantly reduced rate of recurrence of myocardial infarction compared with those who received placebo alone, with a relative risk of 0.62. However, this was offset by a nonsignificant increase in fatal coronary heart disease (relative risk: 1.33), which resulted in a modest, nonsignificant decrease in total coronary events in the vitamin E–supplemented group (relative risk: 0.90). Men who received either ß-carotene alone or a combination of vitamin E and ß-carotene had nonsignificantly increased risks of total coronary events compared with those who received placebo.
The other secondary prevention trial was the Cambridge Heart Antioxidant Study (CHAOS), a randomized, double-blind, placebo-controlled study of 2002 patients with angiographically proven coronary atherosclerosis ( 32 ). In this study, 1035 patents were randomly assigned to receive intervention of either 800 or 400 IU vitamin E/d; results were apparently similar for these 2 doses. As in the ATBC Study, those who received vitamin E experienced a significantly reduced risk of developing nonfatal myocardial infarction compared with those who received placebo (relative risk: 0.23), whereas the risk of fatal myocardial infarction and total cardiovascular deaths (relative risk: 1.18) was elevated in this group. Overall, the vitamin E–supplemented group had a reduced risk of cardiovascular deaths or nonfatal myocardial infarction of 0.53 ( P < 0.005).
The general impression from these randomized trials is that there may be some influence of vitamin E supplementation on reducing risk of coronary events, particularly nonfatal events, but that the evidence is modest. The strongest evidence comes from CHAOS, but in this study of secondary prevention there was an elevated risk of fatal endpoints ( 32 ). Although this increased coronary mortality rate may have been due to chance, the fact that the ATBC Study also reported an elevated risk of coronary mortality with vitamin E supplementation suggests caution in dismissing this finding. In the only primary prevention study to date, vitamin E supplements appear to have had only a modest effect on decreasing occurrence of either angina or fatal coronary heart disease ( 29 ). Other studies of vitamin E supplementation and coronary heart disease are currently underway ( 35 ), but none are aimed at increasing vitamin E intake from foods. The Women's Health Initiative, perhaps the only major ongoing dietary intervention trial that has the ability to examine the primary prevention of coronary heart disease, has as its dietary intervention focus a low-fat diet ( 36 ). Because the richest dietary sources of vitamin E are oil-rich foods from vegetable sources, the intervention group in this trial may inadvertently be consuming lower amounts of vitamin E than the comparison group if the dietary intervention focuses on reduction of all fats in the diet without attention to the type or source of fat. In any case, the trial is not designed to examine the effect of alterations in vitamin E intake on risk of coronary heart disease.
Criterion II
Evidence obtained from well-designed controlled trials without randomization.
No controlled trials without randomization regarding the association of vitamin E and coronary heart disease have been published.
Evidence obtained from well-designed cohort or case-control analytic studies, preferably from more than one center or research group.
Several prospective cohort studies have published results directly pertaining to the question of whether vitamin E intake, from foods or supplements, is associated with risk of coronary heart disease. As shown in
Table
2
, most of the studies noted an inverse association. Although 2 of the 6 studies were from the same research group (
10
,
11
) and a third (
13
) used a dietary assessment instrument similar to that used in the former 2, other studies have been conducted in other populations. The largest studies were the 3 that assessed dietary and supplemental vitamin E intake using similar food-frequency questionnaires (
10
,
11
,
13
).
Case-control studies of diet and disease may be influenced by differential measurement error ( 8 ) and therefore are generally considered to be a weaker basis on which to make causal inferences regarding dietary factors. Few such studies of diet and heart disease have been conducted.
Evidence obtained from multiple time series with or without the intervention.
At least one international correlation study examined the association of food disappearance data from the Food and Agricultural Organization with premature mortality from coronary heart disease (
37
). This study included 24 developed countries, 19 of which were European. Of the various commodities and nutrients examined, vitamin E showed among the strongest inverse associations, with a correlation of -0.8. Within each country, changes in vitamin E availability and changes in coronary heart disease mortality from 1970 to 1987 supported an inverse association between these 2 variables.
There has also been at least one ecologic study of the association of plasma vitamin E concentrations and mortality from coronary heart disease in several European populations that are participants in the World Health Organization Monitoring Trends and Determinants in Cardiovascular Disease (WHO/MONICA) Study ( 38 ). In this study, populations with the highest mortality rates from coronary heart disease had the lowest plasma concentrations of vitamin E, which supports an inverse association of vitamin E with coronary mortality.
Criterion III: opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees
No expert committee has as yet pronounced definitively that vitamin E is inversely associated with coronary heart disease and that increased vitamin E intake or ingestion of vitamin E supplements is a recommended course of action for the prevention of coronary heart disease. However, listed in the American Heart Association's top 10 heart and stroke research advances for 1996 is "Vitamin E may prevent heart disease" (American Heart Association Office of Communications, December 18, 1996). Studies that were specifically mentioned in this designation were the secondary prevention trial CHAOS (
32
) and the Iowa Women's Health Study (
13
), one of the prospective cohort studies.
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONHILL CRITERIAUS PREVENTIVE SERVICES TASK...SUMMARYREFERENCES |
|---|
To date, only 2 randomized trials, both using vitamin E supplements, have addressed the question of whether vitamin E intake may be related to coronary heart disease ( 29 , 32 ). One of these, the ATBC Study, a primary prevention trial in male smokers, provided little evidence that vitamin E intake may decrease the incidence of angina pectoris or mortality rates from coronary heart disease ( 29 , 33 ). The other trial, a secondary prevention trial, provided evidence that vitamin E supplements may decrease risk of recurrence of nonfatal myocardial infarction, but not fatal cardiovascular events ( 32 ). In a follow-up of subjects with definite myocardial infarction at entry into the ATBC Study, similar findings were also observed ( 34 ). Overall, these studies, although perhaps broadly consistent with an inverse association of vitamin E intake with coronary disease, are neither definitive nor convincing in this matter.
The use of the Hill criteria to examine whether there is evidence of a causal relation between vitamin E and coronary heart disease shows that there is some consistency among the results of the prospective cohort studies, with all of them suggesting that vitamin E intake from foods or from supplements is inversely associated with risk of coronary disease. However, the evidence for a dose-response relation is less consistent among the studies, whereas the strength of association is modest although relatively strong for associations of dietary factors with chronic disease endpoints. Plausible biological mechanisms related to prevention of oxidative modification of LDL are based on numerous laboratory studies examining the role of vitamin E and other antioxidants in atherogenesis.
Following the US Preventive Services Task Force criteria, there is little evidence from randomized trials that vitamin E supplementation may reduce risk of coronary heart disease, although it may play a role in prevention of nonfatal myocardial infarction in those with coronary heart disease. Evidence from prospective cohort studies is reasonably strong and consistent, and the one ecologic study to examine associations of food availability and coronary heart disease rates is consistent with an inverse association of vitamin E and coronary heart disease. Few case-control studies and no pronouncements from expert committees exist on this topic. Overall, the evidence can be deemed to support an inverse association of vitamin E intake and coronary heart disease. Whether there is sufficient evidence on which to base public health nutrition recommendations is a matter of debate.
|
FOOTNOTES |
|---|
2 Reprints not available. Address correspondence to LH Kushi, Division of Epidemiology, University of Minnesota School of Public Health, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015. E-mail: kushi{at}epi.umn.edu.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONHILL CRITERIAUS PREVENTIVE SERVICES TASK...SUMMARYREFERENCES |
|---|
-tocopherol and ß-carotene supplements on incidence of major coronary events in men with previous myocardial infarction. Lancet
1997
;349
:1715
–20.[Medline]
This article has been cited by other articles:
|
|
 |
|
  J. M. Genkinger, E. A. Platz, S. C. Hoffman, G. W. Comstock, and K. J. Helzlsouer Fruit, Vegetable, and Antioxidant Intake and All-Cause, Cancer, and Cardiovascular Disease Mortality in a Community-dwelling Population in Washington County, Maryland Am. J. Epidemiol., December 15, 2004; 160(12): 1223 - 1233. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Morris and S. Carson Routine Vitamin Supplementation To Prevent Cardiovascular Disease: A Summary of the Evidence for the U.S. Preventive Services Task Force Ann Intern Med, July 1, 2003; 139(1): 56 - 70. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Potischman Biologic and Methodologic Issues for Nutritional Biomarkers J. Nutr., March 1, 2003; 133(3): 875S - 880. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. El-Sohemy, A. Baylin, A. Ascherio, E. Kabagambe, D. Spiegelman, and H. Campos Population-based study of {alpha}- and {gamma}-tocopherol in plasma and adipose tissue as biomarkers of intake in Costa Rican adults Am. J. Clinical Nutrition, September 1, 2001; 74(3): 356 - 363. [Abstract] [Full Text]
|
|
|
|
 |
|
 I. A. Hininger, A. Meyer-Wenger, U. Moser, A. Wright, S. Southon, D. Thurnham, M. Chopra, H. Van Den Berg, B. Olmedilla, A. E. Favier, et al. No Significant Effects of Lutein, Lycopene or {beta}-Carotene Supplementation on Biological Markers of Oxidative Stress and LDL Oxidizability in Healthy Adult Subjects J. Am. Coll. Nutr., June 1, 2001; 20(3): 232 - 238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Byers, B. Lyle, and W. Participants Summary statement Am. J. Clinical Nutrition, June 1, 1999; 69 (6): 1365S - 1367S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
