Vitamin E, vitamin C, and β-carotene were not associated with a decreased risk for stroke in men
ACP J Club. 1999 Nov-Dec;131:77. doi:10.7326/ACPJC-1999-131-3-077
Ascherio A, Rimm EB, Hernán MA, et al. Relation of consumption of vitamin E, vitamin C, and carotenoids to risk for stroke among men in the United States. Ann Intern Med. 1999 Jun 15;130:963-70.
Is intake of vitamin E, vitamin C, or β-carotene associated with a decreased risk for stroke in men?
8-year cohort study of men participating in the Health Professionals Follow-up Study.
Population-based study in the United States.
43 738 male health professionals (age range 40 to 75 y in 1986). Exclusion criteria were daily caloric intake < 800 kcal or > 4200 kcal, missing data on food questionnaires, cardiovascular disease, or diabetes mellitus.
Assessment of risk factors
Questionnaires completed at baseline and every 2 years thereafter were used to gather data on medical history, lifestyle, frequency of intake of 131 foods during the previous year, use of vitamin and mineral supplements, and potential risk factors for stroke (smoking history, body mass index, history of hypercholesterolemia, physical activity, parental history of myocardial infarction, and alcohol consumption). For analysis, men were divided into quintiles on the basis of intake of vitamin E, vitamin C, β-carotene, and other nutrients (energy-adjusted to 2000 kcal/d); the lowest quintile was assigned a relative risk (RR) of 1.0. Multivariate RRs were adjusted for total energy intake, smoking, alcohol use, hypertension, parental history of myocardial infarction, profession, body mass index, physical activity, and age.
Main outcome measures
Fatal or nonfatal (ischemic, hemorrhagic, or unclassified) stroke was confirmed by blinded review of medical records, autopsy reports, and death certificates.
During follow-up, 328 strokes (210 ischemic, 70 hemorrhagic, and 48 unclassified) occurred and 50 were fatal. Compared with men in the bottom quintile of vitamin E intake (mean 5.4 IU/d), men in the top quintile (mean 411 IU/d) had a multivariate RR for total stroke of 1.25 (95% CI 0.88 to 1.78, P for trend > 0.2). Compared with men in the bottom quintile (mean 95 mg/d), men in the top quintile of vitamin C intake (mean 1167 mg/d) had a multivariate RR for total stroke of 0.95 (CI 0.66 to 1.35, P for trend > 0.2). Compared with men in the bottom quintile (mean 2042 IU/d), men in the top quintile of β-carotene intake (mean 9194 IU/d) had a multivariate RR for total stroke of 0.77 (CI 0.54 to 1.08, P for trend > 0.2). No associations were seen between dose or duration of use of vitamin E or C and stroke of any type (P ≥ 0.10).
Vitamin E, vitamin C, and β-carotene were not associated with a decreased risk for total, ischemic, or hemorrhagic stroke in men.
Source of funding: National Institutes of Health.
For correspondence: Dr. A. Ascherio, Department of Nutrition, Harvard School of Public Health, 665 Huntington Avenue, Boston, MA 02115, USA. FAX 617-432-2435.
The failure to identify an association between dietary intake of vitamin E, vitamin C, and β-carotene and risk for stroke in this study does not exclude such an association because potential biases were inherent in the study design.
First, the participants were not representative of the general population; they were a self-selected group of male health professionals, had no history of symptomatic vascular, disease, and managed to complete at least half of the dietary questionnaire.
Second, the validity of the estimate of dietary intake of vitamin E, vitamin C, and β-carotene is questionable. The only attempt to validate the dietary information was by asking 0.3% of the participants (127 men) to complete two 1-week diet records and provide blood samples. Furthermore, the validity of the "gold standards" (diet records and plasma vitamin levels) is uncertain, and the correlations between food frequency and plasma vitamin levels are not impressive (0.40 to 0.51).
Third, the completeness of outcome measurement and the validity of the diagnosis of stroke are unclear. The absolute rate of stroke was very low, possibly because notification depended on self-report of an incident stroke (i.e., men needed to be able to recognize that they had had a stroke), and the diagnosis of stroke was based on retrospective review of medical records. Furthermore, the rate of ischemic stroke (64%) was lower than usual (expected rate 80% to 85%) and the rate of hemorrhagic stroke (21%) was higher than usual (expected rate 10% to 15%). Because more hemorrhagic strokes than ischemic strokes are fatal, the proportional inconsistency in pathologic stroke subtype in this study suggests that the investigators missed a considerable number of nonfatal ischemic strokes.
Fourth, the lack of association may be confounded by a higher prevalence of other risk factors or direct causes of stroke among men with a high intake of vitamins.
Finally, the 95% CI of the RRs are wide and the lower CI estimates are consistent with vitamin E, vitamin C, and β-carotene intakes associated with a 12%, 34%, and 46% reduction in stroke, respectively. Nevertheless, the most obvious interpretation of the evidence is that these substances do not provide protection from stroke.
Graeme J. Hankey, MBBS, MD
Royal Perth Hospital andUniversity of Western AustraliaPerth, Western Australia, Australia