Dietary Intakes of Antioxidant Vitamins and Mortality From Cardiovascular Disease The Japan Collaborative Cohort Study (JACC) Study

Transcription

1 Dietary Intakes of Antioxidant Vitamins and Mortality From Cardiovascular Disease The Japan Collaborative Cohort Study (JACC) Study Yoshimi Kubota, PhD; Hiroyasu Iso, MD, PhD, MPH; Chigusa Date, PhD; Shogo Kikuchi, MD, PhD; Yoshiyuki Watanabe, MD, PhD; Yasuhiko Wada, MD, PhD; Yutaka Inaba, MD, PhD; Akiko Tamakoshi, MD, PhD; the JACC Study Group Background and Purpose Only a few reports have dealt with the association of antioxidant vitamin intakes with mortality or morbidity from cardiovascular disease in Asia. We investigated the relation of dietary vitamins A, E, and C intake with mortality from cardiovascular disease for Japanese men and women. Methods The subjects were men and women aged 40 to 79 years without a history of cardiovascular disease or cancer who responded to the food frequency questionnaire as part of the Japan Collaborative Cohort Study for Cancer Risk (JACC) Study. They were followed up for a median period of 16.5 years. Hazard ratios were calculated per quintile of dietary vitamins A, E, and C intake by using Cox proportional hazard model. Results During the person-year follow-up, there were 2690 deaths (1343 men and 1347 women) from cardiovascular disease, comprising 1227 (607 men and 620 women) from stroke and 557 (311 men and 246 women) from coronary heart disease. The multivariable hazard ratios (95% CI) associated with the highest versus lowest quintiles of vitamin C intake were 0.70 (0.54 to 0.92) for total stroke, 0.63 (0.41 to 0.97) for coronary heart disease, and 0.79 (0.66 to 0.94) for total cardiovascular disease for women, but the inverse associations observed were weak and did not reach statistical significance for men. No significant association was observed between vitamins A or E intake and risk of mortality for either men or women. Conclusions Vitamin C intake is inversely associated with mortality from cardiovascular disease for Japanese women. (Stroke. 2011;42: ) Key Words: antioxidants cardiovascular disease vitamin A vitamin C vitamin E Antioxidant vitamins such as vitamins A, E, and C may attenuate the development of atherosclerosis through a reduction in oxidation of low-density lipoprotein cholesterol, inhibition of smooth muscle proliferation and platelet adhesion, and regeneration of unoxidized vitamin E by vitamin C. 1 3 Previous studies have showed inverse associations of vegetable and fruit intakes and risk of cardiovascular disease. 4 7 Vegetables are a primary dietary source of vitamins A and E, and fruit is a primary dietary source of vitamin C for the Japanese. 8 However, whether this protective effect is in fact caused by antioxidant vitamins A, E, or C remains unclear. Observational studies and clinical trials in Europe and the United States have reported on associations of antioxidant vitamin with reduced mortality or morbidity from cardiovascular disease. 9,10 However, the evidence for Asia is limited. 11 We therefore investigated the relationship between dietary antioxidant vitamin intake and mortality from cardiovascular disease in a large prospective study of Japanese men and women. Materials and Methods Populations The Japan Collaborative Cohort (JACC) Study for Evaluation of Cancer Risks, sponsored by the Ministry of Education, Sport, and Science, conducted a baseline survey from 1988 to 1990 in 45 areas throughout Japan. Participants completed self-administered questionnaires concerning their lifestyles and medical histories of previous cardiovascular disease and cancer. The details of this survey have been described elsewhere. 12 The participants comprised subjects ( men and women) aged 40 to 79 years old and the overall participation rate was 83%. In 34 of the 45 areas, men and women were eligible for the nutrition Received September 23, 2010; final revision received December 19, 2010; accepted December 28, From Public Health (Y.K., H.I.), Department of Social and Medicine, Graduate School of Medicine, Osaka University, Yamadaoka, Osaka, Japan; the Department of Food Science and Nutrition (C.D.), School of Human Science and Environment, University of Hyogo, Hyogo, Japan; the Department of Public Health (S.K., A.T.), Aichi Medical University School of Medicine, Aichi, Japan; the Department of Epidemiology for Community Health and Medicine (Y. Watanabe), Kyoto Prefectural University of Medicine Graduate School of Medical Science, Kyoto, Japan; the Department of Nutrition, Faculty of Health Science (Y. Wada), Kochi Women s University, Kochi, Japan; and the Department of Food and Health Sciences (Y.I.), Faculty of Human Life Sciences, Jissen Women s University, Tokyo, Japan. Correspondence to Hiroyasu Iso, MD, PhD, MPH, Professor, Public Health, Department of Social and Environmental Medicine, Osaka University, Graduate School of Medicine, 2-2 Yamadaoka, Suita, Osaka-fu , Japan. iso@pbhel.med.osaka-u.ac.jp 2011 American Heart Association, Inc. Stroke is available at DOI: /STROKEAHA

2 1666 Stroke June 2011 survey using a self-administered 40-item food frequency questionnaire (FFQ). A total of men and women, 72% of the eligible subjects, completed the FFQ satisfactorily to calculate nutrients. After exclusion of those persons who had a history of cardiovascular disease or cancer, we used the data for subjects ( men and women). Mortality Surveillance Investigators conducted a systematic review of death certificates in each of the communities for mortality surveillance, all of which were forwarded to their respective public health centers. Mortality data were then centralized at the Ministry of Health and Welfare, and the underlying cause of death was coded for the National Vital Statistics according to the International Classification of Diseases. Registration of death is required by the Family Registration Law in Japan and it is believed to be adhered to throughout Japan. All deaths within the cohort were ascertained by death certificates from public health centers, except for subjects who died after they had moved from their original community, in which case they were treated as censored cases. The participants were followed up to determine mortality from cardiovascular disease having occurred by the end of 2006, except for 4 areas in 1999 and 4 areas in 2003 where the follow-up had been terminated. The cause-specific mortality from cardiovascular disease was determined based on the International Classification of Diseases, 10th Revision: I60 to 69 for stroke, I20 to 25 for coronary heart disease, and I00 to I99 for total cardiovascular disease. Assessment of Vitamins A, E, and C Intake Dietary intakes of vitamins A, E, and C for each of the participant were estimated on the basis of a self-administered FFQ that has been newly developed for the JACC Study. The FFQ covers the frequency of consumption of rice, miso soup, 33 other food items, alcoholic beverages, coffee, and 3 kinds of tea for a total of 40 items. Five responses were possible for the 33 food items, ranging from rarely, once or twice per month, once or twice per week, 3 to 4 times per week, and almost everyday. Details of the validation study for the FFQ were reported previously. 13 For this validation study, we established a 12-day weighted dietary record conducted for 3 consecutive days every 3 months and administered the FFQ twice 1 year apart for 85 subjects. Because the questionnaire provided no option for the size of food portions, we determined the size for each food item according to the 12-day weighted dietary record and determined the daily amount of food items consumed for each subject by weighing the frequency weight as 0, 0.05, 0.214, 0.5, and 1.0 corresponding to the 5 responses. For the calculation of nutrients, we excluded participants with incomplete responses for any of the following: rice, miso soup, 5 of the other 33 food items, and alcohol beverages for men. We calculated nutrient intakes based on the Standard Tables of Food Composition in Japan, fifth revised and enlarged edition. 14 The Spearman rank correlation coefficients between the 12-day weighted dietary records and the second FFQ were 0.30 for vitamin A, 0.29 for vitamin E, and 0.18 for vitamin C of energy-adjusted intakes. The Spearman rank correlation coefficients between the 2 FFQs administered 1 year apart were 0.73, 0.60, and 0.64, respectively. Although we did not take vitamin supplement intakes into consideration for nutrition calculations, the questionnaire included questions about habitual vitamin supplement use over the previous year with 3 responses for frequency, that is, almost everyday, sometimes, and do not use, and 7 choices for type of supplement, namely, multivitamin, vitamin B1, vitamin C, vitamin E, liver oil, and other supplement. We reanalyzed the data after excluding participants who used vitamin supplements almost everyday to examine the potential influence of vitamin supplements. Statistical Analysis We calculated the person-years of follow-up for each participant from the baseline in 1988 to 1990 to the first end point: death, moving from the community and the end of follow-up. Sex-specific age-adjusted means and proportions of selected cardiovascular risk factors were calculated by quintile of energy-adjusted intakes of vitamins A, E, and C. Dietary intakes of vitamins A, E, and C were energy-adjusted with the residual method. 15 Sex-specific hazard ratios (HRs) with 95% CIs were calculated after adjustment for age and other potential confounding factors by using Cox proportional hazard models. The confounding factors included body mass index (sex-specific quintiles), history of hypertension and diabetes (yes or no), smoking status (never, exsmoker, and current smoker of 1 to 19 or 20 cigarettes/day), alcohol intake (never, exdrinker, and current drinker with an ethanol intake of 1 to 22, 23 to 45, 46 to 48, or 69 g/day), perceived mental stress (low, medium, or high), walking (rarely, 30, 30 to 60, or 60 minutes/day), sports (rarely, 1 to 2, 3 to 4, 5 hours/week), education level (age of education completion: 13, 13 to 15, 16 to 18, or 19), total energy intake (kcal/day), dietary cholesterol, saturated fatty acid, n-3 fatty acids, and sodium intake (sex-specific energy-adjusted quintiles). These factors had been known as cardiovascular risk factors. We tested the assumption of proportional hazards according to vitamins A, E, and C intake and found no violation for proportionality. The test for linear trends of confounding variables or HRs across vitamin intake quintiles was conducted by means of linear regression or Cox proportional hazard model using the median variable for each vitamin quintile. The analysis was conducted by adjusting further for sex-specific quintiles of energy-adjusted fiber or potassium intakes to examine potential confounding by other dietary factors rich in fruits and vegetables like antioxidant vitamins. We repeated the analysis by excluding deaths within the first 2 years of follow-up or stratified by hypertension (yes or no), diabetes (yes or no), or smoking status (ever or never smokers) to examine their potential effect modifications. We used SAS Version software (SAS Institute Inc, Cary, NC) for statistical analysis. All statistical tests were 2-tailed and values of P 0.05 were regarded as significant. Results Table 1 shows the baseline sex-specific age-adjusted mean values and proportions of cardiovascular risk according to quintiles of vitamins A, E, and C intake. The male participants with higher vitamins A, E, or C intake tended to be older, more educated, less likely to be smokers, to have a history of hypertension and higher ethanol intake, to show higher proportions of sports participation for 5 hours/week, of walking for 1 hour/day, daily use of vitamin supplements, and higher energy-adjusted intakes of cholesterol, saturated fatty acids, n-3 polyunsaturated fatty acids, and sodium. Total energy intake was not associated with vitamins E or C intake, a history of diabetes mellitus was associated with vitamin A intake, and body mass index was positively associated with vitamin C intake. The female participants showed similar trends as the men. During person-years of follow-up of persons ( men and women), we documented 1227 deaths from total stroke (607 men and 620 women), including 459 deaths from hemorrhagic stroke (196 men and 263 women), 413 deaths from ischemic stroke (222 men and 191 women), 355 deaths from undetermined type of stroke (189 men and 166 women), and 557 deaths from coronary heart disease (311 men and 246 women) including 428 deaths from myocardial infarction (243 men and 185 women), 906 deaths from other cardiovascular disease (425 men and 481 women) and 2690 deaths from total cardiovascular disease (1343 men and 1347 women). Table 2 shows sex-specific age-adjusted and multivariable HRs (95% CI) for mortality from total stroke, coronary heart disease, and total cardiovascular disease according to quintiles of vitamins A, E, and C intake. Dietary intake of vitamin A was

3 Kubota et al Dietary Intakes of Antioxidant Vitamins, CVD 1667 Table 1. Sex-Specific Age-Adjusted Mean Values and Proportions of Cardiovascular Risk Factors According to Quintiles of Dietary Vitamins A, E, and C Intake Vitamin A, g/d P for Vitamin E, mg/d P for Vitamin C, mg/d P for Q1 Q3 Q5 Trend Q1 Q3 Q5 Trend Q1 Q3 Q5 Trend Men Median intake Age, y Body mass index, kg/m History of hypertension, % History of diabetes, % Current smoker, % Ethanol intake, g/d Sports 5 h/wk, % Walking 1 h/d, % Hours of sleep, h/d College or higher education, % High perceived mental stress, % Use of vitamin supplement daily, % Nutrient intake* Energy, kcal/d Dietary cholesterol, mg/d Saturated fatty acids, g/d N-3 fatty acids, g/d Sodium, mg/d Women Median intake Age, y Body mass index, kg/m History of hypertension, % History of diabetes, % Current smoker, % Ethanol intake, g/d Sports 5 h/wk, % Walking 1 h/d, % Hours of sleep, h/d College or higher education, % High perceived mental stress, % Use of vitamin supplement daily, % Nutrient intake* Energy, kcal/d Dietary cholesterol, mg/d Saturated fatty acids, g/d N-3 fatty acids, g/d Sodium, mg/d *Nutrient intake: energy-adjusted. Q indicates quintile. inversely associated with age-adjusted risk of mortality from total stroke for women, and vitamin E intake was associated with age-adjusted risk of mortality from coronary heart disease and total cardiovascular disease for women. These associations with vitamins A and E intake, however, were no longer statistically significant after adjustment for known cardiovascular risk factors. Dietary vitamin C intake was inversely associated with mortality from total stroke for men and mortality from total stroke, coronary heart disease, and total cardiovascular disease for women. These associations for women but not for men remained statistically significant after adjustment for confounding variables. The multivariable HRs (95% CI) for the highest

4 1668 Stroke June 2011 Table 2. Sex-Specific Age-Adjusted and Multivariable* Hazard Ratios (HR) and 95% CIs for Cardiovascular Diseases According to Quintiles of Dietary Vitamins A, E, and C Intake Q1 Q2 Q3 Q4 Q5 P for Trend Men Vitamin A intake Range, g/d Median, g/d No. at risk Person-years Total stroke, no Age-adjusted HR (95% CI) ( ) 1.00 ( ) 0.91 ( ) 1.08 ( ) 0.08 Multivariable HR (95% CI) ( ) 1.09 ( ) 1.02 ( ) 1.15 ( ) 0.06 Coronary heart disease, no Age-adjusted HR (95% CI) ( ) 0.95 ( ) 0.93 ( ) 0.97 ( ) 0.63 Multivariable HR (95% CI) ( ) 1.01 ( ) 0.94 ( ) 0.99 ( ) 0.64 Total cardiovascular disease, no Age-adjusted HR (95% CI) ( ) 0.94 ( ) 0.95 ( ) 1.05 ( ) 0.08 Multivariable HR (95% CI) ( ) 1.02 ( ) 1.04 ( ) 1.09 ( ) 0.10 Vitamin E intake Range, mg/d Median, mg/d No. at risk Person-years Total stroke, no Age-adjusted HR (95% CI) ( ) 0.92 ( ) 0.75 ( ) 0.96 ( ) 0.74 Multivariable HR (95% CI) ( ) 1.15 ( ) 0.93 ( ) 1.20 ( ) 0.48 Coronary heart disease, no Age-adjusted HR (95% CI) ( ) 0.77 ( ) 0.99 ( ) 0.80 ( ) 0.27 Multivariable HR (95% CI) ( ) 0.87 ( ) 1.12 ( ) 0.98 ( ) 0.99 Total cardiovascular disease, no Age-adjusted HR (95% CI) ( ) 0.82 ( ) 0.81 ( ) 0.87 ( ) 0.14 Multivariable HR (95% CI) ( ) 0.93 ( ) 0.88 ( ) 0.91 ( ) 0.48 Vitamin C intake Range, mg/d Median, mg/d No. at risk Person-years Total stroke, no Age-adjusted HR, 95% CI) ( ) 0.85 ( ) 0.84 ( ) 0.78 ( ) 0.10 Multivariable HR (95% CI) ( ) 0.90 ( ) 0.89 ( ) 0.84 ( ) 0.36 Coronary heart disease, no Age-adjusted HR (95% CI) ( ) 1.05 ( ) 0.78 ( ) 0.74 ( ) 0.14 Multivariable HR (95% CI) ( ) 1.16 ( ) 0.90 ( ) 0.86 ( ) 0.62 Total cardiovascular disease, no Age-adjusted HR (95% CI) ( ) 0.86 ( ) 0.80 ( ) 0.82 ( ) 0.04 Multivariable HR (95% CI) ( ) 0.92 ( ) 0.85 ( ) 0.88 ( ) 0.31 (Continued)

5 Kubota et al Dietary Intakes of Antioxidant Vitamins, CVD 1669 Table 2. Continued Q1 Q2 Q3 Q4 Q5 P for Trend Women Vitamin A intake Range, g/d Median, g/d No. at risk Person-years Total stroke, no Age-adjusted HR (95% CI) ( ) 0.93 ( ) 0.89 ( ) 0.70 ( ) Multivariable HR (95% CI) ( ) 1.06 ( ) 1.10 ( ) 0.89 ( ) 0.23 Coronary heart disease, no Age-adjusted HR (95% CI) ( ) 0.63 ( ) 0.61 ( ) 0.81 ( ) 0.96 Multivariable HR (95% CI) ( ) 0.83 ( ) 0.84 ( ) 1.06 ( ) 0.28 Total cardiovascular disease, no Age-adjusted HR (95% CI) ( ) 0.79 ( ) 0.77 ( ) 0.83 ( ) 0.10 Multivariable HR (95% CI) ( ) 0.94 ( ) 0.95 ( ) 1.04 ( ) 0.58 Vitamin E intake Range, mg/d Median, mg/d No. at risk Person-years Total stroke, no Age-adjusted HR (95% CI) ( ) 0.92 ( ) 0.83 ( ) 0.75 ( ) 0.12 Multivariable HR (95% CI) ( ) 0.88 ( ) 0.78 ( ) 0.71 ( ) 0.23 Coronary heart disease, no Age-adjusted HR (95% CI) ( ) 0.83 ( ) 0.57 ( ) 0.64 ( ) Multivariable HR (95% CI) ( ) 1.12 ( ) 0.76 ( ) 0.83 ( ) 0.46 Total cardiovascular disease, no Age-adjusted HR (95% CI) ( ) 0.87 ( ) 0.77 ( ) 0.74 ( ) Multivariable HR (95% CI) ( ) 0.97 ( ) 0.84 ( ) 0.82 ( ) 0.23 Vitamin C intake Range, mg/d Median, mg/d No. at risk Person-years Total stroke, no Age-adjusted HR (95% CI) ( ) 0.70 ( ) 0.75 ( ) 0.67 ( ) Multivariable HR (95% CI) ( ) 0.73 ( ) 0.79 ( ) 0.70 ( ) Coronary heart disease, no Age-adjusted HR (95% CI) ( ) 0.66 ( ) 0.62 ( ) 0.54 ( ) Multivariable HR (95% CI) ( ) 0.81 ( ) 0.77 ( ) 0.63 ( ) Total cardiovascular disease, no Age-adjusted HR (95% CI) ( ) 0.76 ( ) 0.74 ( ) 0.72 ( ) Multivariable HR (95% CI) ( ) 0.82 ( ) 0.82 ( ) 0.79 ( ) Q indicates quintile; HR, hazard ratio. *Adjusted for age, history of hypertension and diabetes, smoking status, alcohol consumption, body mass index, mental stress, walking, sports, education levels, dietary intakes of total energy, cholesterol, saturated fatty acids, n-3 fatty acids, and sodium. versus lowest quintiles of vitamin C intake for women were 0.70 (0.54 to 0.92; P for trend 0.006) for total stroke, 0.63 (0.41 to 0.97; P for trend 0.046) for coronary heart disease, and 0.79 (0.66 to 0.94; P for trend 0.004) for total cardiovascular disease. Among stroke subtypes, the inverse association was confined to ischemic stroke. We repeated the analysis after the exclusion of participants who used supplements daily (n 6012), but the results did not change materially. The respective multivariate HRs (95% CI) for women were 0.73 (0.55 to 0.97; P for trend 0.019), 0.59 (0.37 to 0.93; P for trend 0.028), and 0.80 (0.66 to 0.97; P for trend 0.010).

6 1670 Stroke June 2011 Table 3. Multivariable* Hazard Ratios (HR) and 95% CIs for Cardiovascular Diseases According to Quintiles of Dietary Vitamin C Intake Stratified by Hypertension, Diabetes or Smoking Status for Women Q1 Q2 Q3 Q4 Q5 P for Trend Hypertension Yes No. at risk Person-years Total stroke, no Multivariable HR (95% CI) ( ) 0.62 ( ) 0.69 ( ) 0.72 ( ) Coronary heart disease, no Multivariable HR (95% CI) ( ) 1.24 ( ) 0.82 ( ) 0.67 ( ) 0.24 Total cardiovascular disease, no Multivariable HR (95% CI) ( ) 0.80 ( ) 0.67 ( ) 0.79 ( ) No No. at risk Person-years Total stroke, no Multivariable HR (95% CI) ( ) 0.79 ( ) 0.85 ( ) 0.64 ( ) Coronary heart disease, no Multivariable HR (95% CI) ( ) 0.61 ( ) 0.78 ( ) 0.66 ( ) 0.17 Total cardiovascular disease, no Multivariable HR (95% CI) ( ) 0.84 ( ) 0.93 ( ) 0.77 ( ) 0.07 Diabetes Yes No. at risk Person-years Total stroke, no Multivariable HR (95% CI) ( ) 0.34 ( ) 0.60 ( ) 0.61 ( ) 0.60 Coronary heart disease, no Multivariable HR (95% CI) ( ) 0.54 ( ) 1.30 ( ) 1.54 ( ) 0.57 Total cardiovascular disease, no Multivariable HR (95% CI) ( ) 0.47 ( ) 0.81 ( ) 0.69 ( ) 0.63 No No. at risk Person-years Total stroke, no Multivariable HR (95% CI) ( ) 0.76 ( ) 0.81 ( ) 0.70 ( ) Coronary heart disease, no Multivariable HR (95% CI) ( ) 0.78 ( ) 0.73 ( ) 0.56 ( ) Total cardiovascular disease, no Multivariable HR (95% CI) ( ) 0.85 ( ) 0.83 ( ) 0.79 ( ) Smoking status Ever-smokers No. at risk Person-years Total stroke, no Multivariable HR (95% CI) ( ) 1.19 ( ) 0.81 ( ) 0.85 ( ) 0.20 Coronary heart disease, no Multivariable HR (95% CI) ( ) 0.74 ( ) 0.26 ( ) 1.66 ( ) 0.60 Total cardiovascular disease, no Multivariable HR (95% CI) ( ) 1.07 ( ) 0.52 ( ) 1.10 ( ) 0.39 (Continued)

7 Kubota et al Dietary Intakes of Antioxidant Vitamins, CVD 1671 Table 3. Continued Q1 Q2 Q3 Q4 Q5 P for Trend Never smokers No. at risk Person-y Total stroke, no Multivariable HR (95% CI) ( ) 0.69 ( ) 0.76 ( ) 0.62 ( ) Coronary heart disease, no Multivariable HR (95% CI) ( ) 0.79 ( ) 0.91 ( ) 0.58 ( ) 0.06 Total cardiovascular disease, no Multivariable HR (95% CI) ( ) 0.76 ( ) 0.82 ( ) 0.73 ( ) Q indicates quintile; HR, hazard ratio. *Adjusted for age, history of hypertension and diabetes, smoking status, alcohol consumption, body mass index, mental stress, walking, sports, education levels, dietary intakes of total energy, cholesterol, saturated fatty acids, n-3 fatty acids, and sodium. After further adjustment for fiber intake, the inverse associations with vitamin C intake for women did not change materially; the respective multivariable HRs (95% CI) were 0.58 (0.42 to 0.80; P for trend 0.001), 0.64 (0.38 to 1.08; P for trend 0.12), and 0.71 (0.57 to 0.89; P for trend 0.002). However, when potassium intake was adjusted further, the inverse association was attenuated for total cardiovascular disease but did not change materially for total stroke and coronary heart disease with enlarged CIs. The respective multivariable HRs (95% CI) were 0.69 (0.45 to 1.06; P for trend 0.09), 0.68 (0.35 to 1.33; P for trend 0.30), and 0.91 (0.68 to 1.22; P for trend 0.51). When deaths within 2 years were excluded, the result did not change. The respective multivariable HRs (95% CI) were 0.69 (0.52 to 0.91; P for trend 0.004), 0.59 (0.38 to 0.91; P for trend 0.025), and 0.79 (0.65 to 0.95; P for trend 0.004). Table 3 shows multivariable HRs (95% CI) for mortality from cardiovascular disease mortality according to quintiles of vitamin C intake stratified by hypertension, diabetes, or smoking status for women. The inverse association vitamin C intake and cardiovascular disease mortality for women did not vary by hypertension, diabetes, or smoking status. For men, no association between vitamin C intake and cardiovascular disease mortality was observed when stratified by hypertension, diabetes, and smoking status (not shown in the table). Discussion We found inverse associations between dietary vitamin C intake and mortality from stroke, coronary heart disease, and total cardiovascular disease for women even after known cardiovascular risk factors were taken into account. However, the association between vitamin C intake and stroke mortality was weak and no longer statistically significant for men after adjustment for confounding factors. No association was found between vitamins A or E intake and mortality from cardiovascular disease for either sex. The inverse association between vitamin C intake and stroke for women but not for men was consistent with the finding of a Finnish cohort. 16 The authors of this study explained this finding as possibly the result of the effect of dietary vitamin C being overwhelmed by a less efficient antioxidant capacity or more potent coronary risk factors for men. The multivariate HR (95% CI) for the highest versus lowest quintiles of vitamin C intake was 0.49 (0.24 to 0.98) for women and 1.00 (0.68 to 1.45) for men. Cohort studies in Europe or the United States 9 have indicated that high dietary intakes of vitamin E as well as vitamin C are associated with reduced mortality or incidence of cardiovascular disease, including acute myocardial infarction, stroke, and coronary heart disease. The summary OR (95% CI) determined in these studies for high versus low antioxidant intakes was 0.89 (0.79 to 0.99) for vitamin C and 0.74 (0.66 to 0.83) for vitamin E. In the current study, we did not find any significant associations between dietary vitamin E intake and mortality from cardiovascular disease for either sex. This lack of association may constitute support for the notion that food constituents other than vitamin E protect against coronary heart disease in Western countries but not in Japan because of major differences in food sources. In Japan, 15% to 31% of vitamin E is obtained from green and yellow vegetables, 13% to 27% from fish and shellfish, and 11% to 21% from vegetable oil, whereas in Europe or the United States, 33% comes from oil and fats and 17% from cereals and doughnuts. 8,17 Our study has several limitations. First, vitamin intakes were estimated on the basis of a limited number of foods (n 40) in the self-administered FFQ without options for food portion size. In the present study, the validity of estimated vitamin C intake was lower than that of other FFQs. 18,19 Because the misclassification in the estimate of vitamin C intake seems to be nondifferential for mortality outcome, the real association between vitamin C intake and cardiovascular disease mortality would be stronger. Second, vitamin intake from supplements was not taken into account in the present analysis. However, the apparent percentage of daily supplement users was 10% and the relationship between dietary vitamin intakes and mortality from cardiovascular disease did not change after the exclusion of supplement users. Third, it is possible that the observed association with vitamin C intake for women was confounded from other nutrients. When we adjusted further for fiber or potassium intake, the inverse associations did not change materially except for total cardiovascular disease after adjustment for potassium intake. However, the enlarged CIs observed after adjustment for potassium intake were probably due to the high correlation between vitamin C and potassium intakes (Spearman correlation coeffi-

8 1672 Stroke June 2011 cients, 0.85 for men and 0.83 for women). Fourth, because the outcomes were expressed in terms of mortality but not incidence, the contribution of these vitamin intakes to cardiovascular outcomes may be influenced by medical care and lifestyle change after an incidence of cardiovascular disease. The strength of the present study was our use of large samples from 34 areas in Japan so that our findings can probably be generalized for the Japanese population as a whole. In conclusion, we found that vitamin C intake is inversely associated with mortality from total stroke, coronary heart disease, and total cardiovascular disease for Japanese women. A similar but nonsignificant inverse association was observed between vitamin C intake and mortality from total stroke for Japanese men. Appendix Study Investigators Dr Akiko Tamakoshi (present chairperson of the study group), Aichi Medical University School of Medicine; Drs Mitsuru Mori and Fumio Sakauchi, Sapporo Medical University School of Medicine; Dr Yutaka Motohashi, Akita University School of Medicine; Dr Ichiro Tsuji, Tohoku University Graduate School of Medicine; Dr Yosikazu Nakamura, Jichi Medical School; Dr Hiroyasu Iso, Osaka University School of Medicine; Dr Haruo Mikami, Chiba Cancer Center; Dr Michiko Kurosawa, Juntendo University School of Medicine; Dr Yoshiharu Hoshiyama, University of Human Arts and Sciences; Dr Naohito Tanabe, Niigata University School of Medicine; Dr Koji Tamakoshi, Nagoya University Graduate School of Health Science; Dr Kenji Wakai, Nagoya University Graduate School of Medicine; Dr Shinkan Tokudome, National Institute of Health and Nutrition; Dr Koji Suzuki, Fujita Health University School of Health Sciences; Dr Shuji Hashimoto, Fujita Health University School of Medicine; Dr Shogo Kikuchi, Aichi Medical University School of Medicine; Dr Yasuhiko Wada, Faculty of Health Science, Kochi Women s University; Dr Takashi Kawamura, Kyoto University Center for Student Health; Dr Yoshiyuki Watanabe, Kyoto Prefectural University of Medicine Graduate School of Medical Science; Dr Kotaro Ozasa, Radiation Effects Research Foundation; Dr Tsuneharu Miki, Kyoto Prefectural University of Medicine Graduate School of Medical Science; Dr Chigusa Date, School of Human Science and Environment, University of Hyogo; Dr Kiyomi Sakata, Iwate Medical University; Dr Yoichi Kurozawa, Tottori University Faculty of Medicine; Drs Takesumi Yoshimura and Yoshihisa Fujino, University of Occupational and Environmental Health; Dr Akira Shibata, Kyushu University Institute of Health Science; Dr Naoyuki Okamoto, Kanagawa Cancer Center; and Dr Hideo Shio, Moriyama Municipal Hospital. Acknowledgments We thank all staff members involved in this study for their valuable help in conducting the baseline survey and follow-up. Sources of Funding This study was supported by the ministry of Education, Science, Sports and Culture of Japan (Monbusho) and the Japanese Ministry of Education, Culture, Sports, Science, and Technology (Monbukagaku-sho): grant numbers , , , , , , , , , , , , , and None. Disclosures References 1. Cherubini A, Vigna GB, Zuliani G, Ruggiero C, Senin U, Fellin R. Role of antioxidants in atherosclerosis: epidemiological and clinical update. Current Pharmaceutical Design. 2005;11: Fairfield KM, Fletcher RH. Vitamins for chronic disease prevention in adults. JAMA. 2002;287: Padayatty SJ, Katz A, Wang Y, Eck P, Kwon O, Lee J, Chen S, Corpe C, Dutta A, Dutta S, Levine M. Vitamin C as an antioxidant: evaluation of its role in disease prevention. J Am Coll Nutr. 2003;22: Dauchet L, Amouyel P, Hercberg S, Dallongeville J. Fruit and vegetable consumption and coronary heart disease: a meta analysis of cohort studies. J Nutr. 2006;136: He FJ, Nowson CA, MacGregor GA. Fruit and vegetable consumption and stroke: meta-analysis of cohort studies. Lancet. 2006;367: Takachi R, Inoue M, Ishihara J, Kurahashi N, Iwasaki M, Sasazuki S, Iso H, Tsubono Y, Tsugane S; JPHC Study Group. Fruit and vegetable intake and risk of total cancer and cardiovascular disease: Japan Public Health Center-Based Prospective Study. Am J Epidemiol. 2008;167: Nagura J, Iso H, Watanabe Y, Maruyama K, Date C, Toyoshima H, Yamamoto A, Kikuchi S, Koizumi A, Kondo T, Wada Y, Inaba Y, Tamakoshi A; the JACC Study Group. Fruit, vegetable and bean intake and mortality from cardiovascular disease among Japanese men and women: the JACC Study. Br J Nutr. 2009;102: Kato Y, Ikehara S, Maruyama K, Inagawa M, Oshima M, Yokota K, Yamazaki T, Kishi M, Murai S, Umesawa M, Ma E, Yamagishi K, Tanigawa T, Kurokawa M, Sato S, Shimamoto T, Iso H. Trends in dietary intakes of vitamins A, C and E among Japanese men and women from 1974 to Public Health Nutr. 2009;12: Asplund K. Antioxidant vitamins in the prevention of cardiovascular disease: a systematic review. J Intern Med. 2002;251: Willcox BJ, Curb JD, Rodriguez BL. Antioxidants in cardiovascular health and disease: key lessons from epidemiologic studies. Am J Cardiol. 2008;101:75D 86D. 11. Yokoyama T, Date C, Kokubo Y, Yoshiike N, Matsumura Y, Tanaka H. Serum vitamin C concentration was inversely associated with subsequent 20 year incidence of stroke in a Japanese rural community: the Shibata study. Stroke. 2000;31: Tamakoshi A, Yoshimura T, Inaba Y, Ito Y, Watanabe Y, Fukuda K, Iso H; JACC Study Group. Profile of the JACC Study. J Epidemiol. 2005; 15:S4 S Date C, Fukui M, Yamamoto A, Wakai K, Ozeki A, Motohashi Y, Adachi C, Okamoto N, Kurosawa M, Tokudome Y, Kurisu Y, Watanabe Y, Ozasa K, Nakagawa S, Tokui N, Yoshimura T, Tamakoshi A; JACC Study Group. Reproducibility and validity of a self-administered food frequency questionnaire used in the JACC study. J Epidemiol. 2005;15: S9 S Council for Science and Technology; Ministry of Education, Culture, Sports, Science and Technology; the Government of Japan. Standard Tables of Food Composition in Japan, fifth revised and enlarged edition. Tokyo: National Printing Bureau; Willet WC, Stampfer MJ. Total energy intake: implications for epidemiologic analysis. Am J Epidemiol. 1986;124: Knekt P, Reunanen A, Järvinen R, Seppänen R, Heliövaara M, Aromaa A. Antioxidant vitamin intake and coronary mortality in a longitudinal population study. Am J Epidemiol. 1994;139: Subar AF, Krebs-Smith SM, Cook A, Kahle LL. Dietary sources of nutrients among US adults, 1989 to J Am Diet Assoc. 1998;98: Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, Hennekens CH, Speizer FE. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 1985;122: Ishihara J, Inoue M, Kobayashi M, Tanaka S, Yamamoto S, Iso H, Tsugane S; JPHC FFQ Validation Study Group. Impact of the revision of a nutrient database on the validity of a self-administered food frequency questionnaire (FFQ). J Epidemiol. 2006;16: