点击显示 收起
1 From the Division of Epidemiology, School of Public Health, University of Minnesota (D-HL, ARF, LH, and DRJ); the Department of Preventive Medicine, School of Medicine, Kyungpook National University, Daegu, Korea (D-HL); the Department of Biochemistry, National University of Singapore (BH); and the Department of Nutrition, University of Oslo (DRJ)
2 The contents of this manuscript are solely the responsibilities of the authors and do not necessarily represent the official view of the National Cancer Institute. 3 Supported by grant R01-HL053560-08 (YALTA) from the National Heart, Lung, and Blood Institute and by grant 03/1/21/18/213 from the Singapore Biomedical Research Council for support (to BH).The Iowa Women's Health Study was supported by a grant (RO1 CA39742) from the National Cancer Institute. 4 Address reprint requests to DR Jacobs, Jr, University of Minnesota, Division of Epidemiology, School of Public Health, 1300 South 2nd Street, Suite 300, Minneapolis, MN 55454. E-mail: jacobs{at}epi.umn.edu.
ABSTRACT
Background: Vitamin C acts as a potent antioxidant; however, it can also be a prooxidant and glycate protein under certain circumstances in vitro. These observations led us to hypothesize that a high intake of vitamin C in diabetic persons might promote atherosclerosis.
Objective: The objective was to examine the relation between vitamin C intake and mortality from cardiovascular disease.
Design: We studied the relation between vitamin C intake and mortality from total cardiovascular disease (n = 281), coronary artery disease (n = 175), and stroke (n = 57) in 1923 postmenopausal women who reported being diabetic at baseline. Diet was assessed with a food-frequency questionnaire at baseline, and subjects initially free of coronary artery disease were prospectively followed for 15 y.
Results: After adjustment for cardiovascular disease risk factors, type of diabetes medication used, duration of diabetes, and intakes of folate, vitamin E, and ß-carotene, the adjusted relative risks of total cardiovascular disease mortality were 1.0, 0.97, 1.11, 1.47, and 1.84 (P for trend < 0.01) across quintiles of total vitamin C intake from food and supplements. Adjusted relative risks of coronary artery disease were 1.0, 0.81, 0.99, 1.26, and 1.91 (P for trend = 0.01) and of stroke were 1.0, 0.52, 1.23, 2.22, and 2.57 (P for trend < 0.01). When dietary and supplemental vitamin C were analyzed separately, only supplemental vitamin C showed a positive association with mortality endpoints. Vitamin C intake was unrelated to mortality from cardiovascular disease in the nondiabetic subjects at baseline.
Conclusion: A high vitamin C intake from supplements is associated with an increased risk of cardiovascular disease mortality in postmenopausal women with diabetes.
Key Words: Vitamin C • cardiovascular diseases • diabetes mellitus • iron
INTRODUCTION
Vitamin C acts as a potent antioxidant by scavenging physiologically relevant reactive oxygen, chlorine, and nitrogen species (1, 2). Vitamin C can also be a prooxidant in vitro under certain circumstances (2-4) and can glycate protein (5). Plasma vitamin C is often subnormal in diabetic persons (6, 7). The implication is that the correction of these subnormal concentrations would be beneficial, but few data exist to support this hypothesis. Therefore, we examined the relation of vitamin C intake to mortality from coronary artery disease (CAD), stroke, or total cardiovascular disease (CVD) in women with diabetes by analyzing data from the Iowa Women's Health Study.
SUBJECTS AND METHODS
Iowa Women's Health Study Cohort
Methods for the Iowa Women's Health Study recruitment and data collection were published previously (8). Briefly, this study was designed to examine associations between several host, dietary, and lifestyle factors and the incidence of cancer and mortality in 41 836 postmenopausal women aged 55–69 y who completed a 16-page self-administered questionnaire at baseline in January 1986. Of the women who returned the 1986 baseline questionnaire, 99% were white, 34% resided in cities with a population of 10 000, and 27% lived in a rural area. Follow-up questionnaires were mailed in 1987, 1989, 1992, and 1997 (response rates of 91%, 89%, 83%, and 79%, respectively) to identify emigrants from Iowa and deaths.
The baseline questionnaire included self-report questions concerning several risk factors pertinent to CVD, including history of diabetes or high blood pressure, smoking, physical activity, alcohol consumption, and hormone replacement therapy. Weight, height, and waist and hip circumferences were measured by each participant using a fixed protocol. Blood lipids and blood pressure were not measured. A 127-item food-frequency questionnaire similar to that used in the 1984 survey of the Nurses' Health Study (9) was also included. Food-composition values were obtained from the Harvard University Food Composition Database derived from US Department of Agriculture sources (10) and were supplemented with manufacturer information and other published values. To evaluate the validity of the questionnaire, we compared the vitamin C intake in a subgroup of 44 women with their mean intake estimated from five 24-h dietary recall interviews (11). The correlation coefficients between estimates from the food-frequency questionnaire and the average of the dietary recalls were 0.76 for the total intake of vitamin C and 0.53 for vitamin C intake excluding supplements, similar to those from the Nurses' Health Study (12).
At baseline, women were considered to have diabetes if they responded "yes" to one of the following questions at baseline: 1) Have you ever been told by a doctor that you have sugar diabetes? or 2) Have you ever taken insulin or pills for sugar diabetes (or to lower blood glucose)? Although no specific information was obtained about whether the diabetes was type 1 or type 2, it is likely that the great majority of cases were type 2 because only 5.8% of diabetic patients were diagnosed before age 30 y. Data on the duration of diabetes and type of treatment were collected.
Deaths in Iowa were identified annually through the State Health Registry of Iowa. Deaths for subjects who did not respond to the follow-up questionnaires or who had emigrated from Iowa were found through the National Death Index. Underlying cause of death was assigned by state vital registries via the International Classification of Disease (ICD). We defined 1) CAD by codes 410 through 414 or 429.2 in the 9th revision of the ICD or codes I20 through I25 or I51.6 in the 10th revision, 2) stroke by codes 430 through 438 in the 9th revision or I60 through I69 in the 10th revision, and 3) all CVD by codes 390 through 459 in the 9th revision or I00 through I99 in the 10th revision.
Data analysis
Of 2881 women reporting diabetes at baseline, we excluded from analysis those who reported implausibly high (>5000 kcal) or low (<600 kcal) energy intakes (n = 59), left 30 items blank on the food-frequency questionnaire (n = 320), were pre- or perimenopausal (n = 26), or reported angina, heart disease, or heart attack at baseline (n = 686). After the exclusions were accounted for, 1923 women remained eligible for the study.
The length of follow-up for each woman was calculated from the time of completion of the baseline questionnaire to the date of death or 31 December 2000, whichever came first. Of the women eligible for follow-up, 281 died of CVD; 175 from CAD, and 57 from stroke.
Because vitamins can be ingested from both foods and supplements, 3 exposure indexes were studied: total vitamin C intake from food and supplements, vitamin C intake from food alone, and vitamin C intake from supplements only.
The participants were categorized according to quintiles of total vitamin C intake or dietary vitamin C intake. Supplemental intake of vitamin C was classified as 0, 1–99, 100–299, and 300 mg/d. We used proportional hazards regression to account for potential effects of other risk factors for CVD. Multivariate models adjusted simultaneously for baseline values of age (continuous), total energy intake (continuous), history of hypertension, BMI (continuous), waist-hip ratio (WHR, continuous), physical activity score (low, medium, or high), smoking status (current smoker, exsmoker, or nonsmoker), amount smoked (none or 1–19, 20–39, or 40 pack-years), alcohol consumption (g/d), hormone replacement therapy (current, former, or never), main type of diabetes medication used (insulin, oral medication, or none), and duration of diabetes (<5, 5–9, 10–15, 20–24, 25–30, and 30 y). An additional model adjusted for intake of saturated fat, trans fat, polyunsaturated fat, folate, ß-carotene, and vitamin E (all as continuous variables). When dietary vitamin C and supplemental vitamin C were analyzed separately, the amount of vitamin C intake from one source was adjusted for that from the other source.
In tests for trend, median values within categories of total, dietary, or supplemental vitamin C intake were used. For all relative risks (RRs), we calculated 95% CIs by using SAS software (version 8.2; SAS Institute, Cary, NC). All P values were two-tailed.
RESULTS
The distribution of CVD risk factors according to intake of vitamin C from food and supplements is shown in Table 1. After adjustment for age and total energy intake, the women who consumed more vitamin C tended to have healthier behaviors and dietary habits. The women in the highest versus the lowest quintile tended to have lower WHR and fewer of them were current smokers, more were physically active, and more took postmenopausal hormones. In addition, women who consumed more vitamin C consumed less saturated fat, less trans fat, more vitamin E, more ß-carotene, and more folate. There was no association between total vitamin C intake and type of diabetic medication used (32% of which was insulin; the remainder of which were oral agents) or duration of diabetes. Supplemental vitamin C use (especially intakes 300 mg/d) increased greatly with quintile of total vitamin C intake. Relations among risk factors and supplemental vitamin C were very similar to those with total vitamin C intake (data not shown). The women with diabetes had a higher CVD mortality if they had a history of hypertension, had a higher WHR, were cigarette smokers, had a higher saturated fat intake, reported less physical activity, did not use postmenopausal hormones, or had a lower folate intake. Compared with women who did not take any diabetic medication, users of insulin or oral agents had a higher CVD mortality.
View this table:
TABLE 1. Age- and total energy intake–adjusted baseline characteristics according to quintile of vitamin C intake from food and supplements in postmenopausal women with diabetes in the Iowa Women's Health Study, 19861
After adjustment for age and total energy intake, total vitamin C intake from food and supplements tended to be positively associated with mortality from total CVD, CAD, and stroke. Given the strong associations of vitamin C with many factors that are inversely related to CVD, additional adjustment for CVD risk factors, type of diabetes medication, duration of diabetes, and other nutrients made the association stronger (Table 2). After additional adjustment for dietary factors, the RRs for total CVD mortality across quintiles of total vitamin C intake were 1.00, 0.97, 1.11, 1.47, and 1.84 (P for trend < 0.01). Adjusted RRs showed a similar pattern for death from CAD or stroke.
View this table:
TABLE 2. Relative risks (RRs) and 95% CIs for mortality from cardiovascular disease, coronary artery disease, and stroke according to quintile of vitamin C intake from food and supplements in postmenopausal women with diabetes in the Iowa Women's Health Study, 1986–2000
In an attempt to disentangle the associations of CVD deaths with vitamin C from food or supplements, models were run that simultaneously included terms for food and for supplements. Vitamin C intake from food only showed no clear trend with mortality outcomes (Table 3). In contrast, vitamin C intake from supplements of 300 mg/d showed a positive association with mortality outcomes (Table 4). Total vitamin E intake and total ß-carotene intake were not associated with CVD, whereas total folate intake showed a trend toward an inverse association (Table 5).
View this table:
TABLE 3. Relative risks (RRs) and 95% CIs for mortality from cardiovascular disease, coronary artery disease, and stroke according to quintile of vitamin C intake from food in postmenopausal women with diabetes in the Iowa Women's Health Study, 1986–2000
View this table:
TABLE 4. Relative risks (RRs) and 95% CIs for mortality from cardiovascular disease, coronary artery disease, and stroke according to quintile of vitamin C intake from supplements in postmenopausal women with diabetes in the Iowa Women's Health Study, 1986–2000
View this table:
TABLE 5. Adjusted relative risks and 95% CIs for mortality from cardiovascular disease, coronary artery disease, and stroke per quintile of vitamin C, folate, vitamin E, and ß-carotene intakes from food and supplements in postmenopausal women with diabetes in the Iowa Women's Health Study, 1986–20001
A supplemental analysis was run to determine whether the observed associations were specific to women with prevalent diabetes or generalized to nondiabetic persons. Of 32 569 women who did not have diabetes at baseline, total vitamin C intake was not associated with the disease outcomes; the adjusted RRs for total CVD mortality (n = 1486) across quintiles of total vitamin C intake were 1.00, 0.90, 0.98, 0.88, and 1.03 (P for trend = 0.29, P for interaction with prevalent diabetes = 0.08). Similarly, for supplemental vitamin C, the P value for the interaction was 0.03 and the adjusted RRs among nondiabetic persons were 1.0, 0.92, 0.89, and 1.02 (P for trend = 0.64).
DISCUSSION
Vitamin C intake from supplements was positively related to mortality from all CVD, CAD, and stroke among postmenopausal women who reported diabetes at baseline. No such association was observed among women who did not have diabetes at baseline.
Although the reasons for accelerated atherosclerosis in diabetic subjects are still not completely understood, CVD risk in diabetes largely depends on both diabetes-related variables such as glycemic control, diabetes duration, proteinuria, and overall artery disease as well as the classic risk factors blood pressure, smoking, and dyslipidemia (13). We adjusted for most classic risk factors. Although we did not measure blood pressure or blood lipids, typically these have not been correlated strongly with vitamin C intake (14, 15). Furthermore, because vitamin C supplement users tend to have healthy behaviors and diet patterns, consumers of high amounts of vitamin C likely had lower blood pressures and lower atherogenic blood lipid concentrations than did consumers of low amounts. Adjustment for these factors would likely strengthen the association. Moreover, vitamin C intake was not associated with medication type or duration of diabetes; therefore, the severity of diabetes is unlikely to be a strong confounder in the association between vitamin C consumption and CVD mortality.
Nevertheless, there is still a possibility that patients with more severe diabetes tend to consume more vitamin C, including that from supplements. Vitamin C is a known antioxidant, and patients with type 2 diabetes might interpret general advice to consume more vegetables and fruit to include supplemental vitamin C. However, if this were true, other nutrients with characteristics similar to vitamin C for which similar recommendations exist should also show a positive association with CVD mortality. However, vitamin E and ß-carotene intakes were not associated with CVD mortality, and folate intake showed an inverse association. Therefore, we think that confounding was an unlikely explanation of the positive association between supplementary vitamin C and CVD among these postmenopausal women with diabetes.
If the observed association of supplemental vitamin C and risk for CVD is not a chance occurrence, what might be the mechanism of this apparent deleterious effect of extra vitamin C in diabetes? We present 3 possibilities. It is known that vitamin C may act as a prooxidant in the presence of free iron in vitro (2, 3). In some persons with diabetes, there might be iron overload and a disturbance of iron metabolism (16-18), possibly inducing free iron. Alternatively, vitamin C can cause damage by glycating proteins (5) or by stimulating lipid peroxidation (4).
Even though vitamin C is a well-known strong antioxidant and a higher intake of fruit and vegetables has been known to be beneficial in preventing CVD (19, 20), the findings concerning a protective effect of vitamin C intake on CVD have been unclear. Observational evidence that vitamin C intake decreases CVD risk is inconsistent (21-27), especially after adjustment for vitamin E intake. A recent double-blind, randomized clinical trial of the combination of supplements of 400 IU vitamin E and 500 mg vitamin C daily compared with placebo showed an increased rate of progression of coronary atherosclerosis in postmenopausal women (28). Clinical trials focusing solely on vitamin C supplements have not yet been conducted. Recently, the Nurses' Health Study reported a beneficial effect of vitamin C supplements among diabetic persons and among all study subjects in the only other study that explored the relation among diabetic persons (27). Although a reduced risk of fatal and nonfatal CAD was observed in the entire group of women who took high doses of vitamin C supplements, the results for diabetic women who took high doses of vitamin C supplements (ie, 300 mg) were not presented separately. In partial agreement with the findings in the women in the Nurses' Health Study, the diabetic women in the Iowa Women's Health Study who took 1–99 mg vitamin C/d showed a weak inverse trend (Table 4).
In previous reports of the Iowa Women's Health Study (25, 26), a nonsignificant positive association between vitamin C intake and the risk of death from CAD (25) or a U-shaped association with stroke (26) was seen among the whole cohort. Vitamin C can act as an antioxidant or a prooxidant (3, 4); therefore, the lack of association between vitamin C and CVD mortality in the whole cohort might reflect the opposing actions of vitamin C. On the other hand, some studies found that vitamin C supplementation reduces oxidative stress in type 2 diabetic persons (29, 30). However, in these studies, only the short-term effects on serum markers of oxidative stress or endothelial function were observed. Short-term beneficial effects of vitamin C might not lead to long-term beneficial effects, a finding that is similar to that observed with supplemental ß-carotene (31).
In the current study, only vitamin C from supplements—and not from foods—increased the risk of CVD mortality. Although a higher absolute intake from supplements might explain this finding, a more likely explanation is that antioxidants naturally present in food are balanced biochemically, ie, they are part of a mixture of redox agents in oxidized form and in reduced form, whereas every supplement pill lacks this balance (32). In addition, the body's defense against oxidative stress is accomplished by interconnecting systems of antioxidant micronutrients with a range of physical properties; therefore, the use of high doses of a single antioxidant could perturb the antioxidant-prooxidant balance (33).
Our study had several limitations. First, our findings were based on only one dietary and health behavior assessment at baseline. Although this limitation may have resulted in misclassification, generally the nondifferential misclassification of exposure variables leads to a null association rather than to a spurious association. Second, the subjects with diabetes were ascertained by self reports. According to our validation study of self-reported diabetes (34), our subjects tended to overreport diabetes. Nonvalidated positive reports may nonetheless reflect some level of glucose intolerance (35). Furthermore, the fasting glucose criterion for diabetes was 140 mg/dL in 1988, but it was subsequently lowered to 126 mg/dL. Thus, some women who falsely reported a diagnosis of diabetes may still have had some level of underlying glucose abnormality. If vitamin C supplementation is unrelated to CVD death in nondiabetic women, the mislabeling of some nondiabetic women as diabetic would tend to attenuate the true association between vitamin C supplementation and CVD. Third, we cannot totally rule out the possibility that our findings occurred by chance because of the small number of cases in some subgroups. Fourth, because our study was limited to postmenopausal women, the results may not be generalizable to other groups. Fifth, we did not collect CVD incidence data; however, associations of risk factors with CVD mortality endpoints generally parallel those for incidence (13, 36).
In summary, our results suggest that a high vitamin C intake from supplements was associated with increased mortality from CVD among postmenopausal women who reported having diabetes. Vitamin C supplements are commonly consumed. It is sometimes assumed that the common condition of subnormal plasma vitamin C concentrations in diabetic persons is unfavorable and should be corrected. Our data challenge this assumption. Indeed, it is possible that low vitamin C concentrations may be a beneficial response to diabetes at sites of tissue damage (37).
ACKNOWLEDGMENTS
D-HL was responsible for the data analysis, conception, and first draft of the manuscript. ARF, LH, and BH provided intellectual input and edited the manuscript. DRJ carried out the data analysis, provided intellectual input, and edited the manuscript. None of the authors had a financial or personal interest involved in the sponsorship of this research study.
REFERENCES