Vitamin C and hyperglycemia in the European Prospective Investigation into Cancer--Norfolk (EPIC-Norfolk) study: a population-based study.

Abstract

To examine the cross-sectional association between plasma vitamin C, self-reported diabetes, and HbA1c. Data from a population-based study of diet, cancer, and chronic disease were analyzed. A total of 2,898 men and 3,560 women 45-74 years of age who were registered with general practices in Norfolk, U.K., were recruited to the European Prospective Investigation Into Cancer-Norfolk study between 1995 and 1998. Mean plasma vitamin C levels were significantly higher in individuals with HbA1c levels < 7% than in those with self-reported diabetes or prevalent undiagnosed hyperglycemia (HbA1c > or = 7%). An inverse gradient of mean plasma vitamin C was found in both sexes across quintiles of HbA1c distribution < 7%. The odds ratio (95% CI) of having prevalent undiagnosed hyperglycemia per 20 micromol/l (or 1 SD) increase in plasma vitamin C was 0.70 (0.52-0.95) (adjusted for sex, age, BMI, waist-to-hip ratio, tertiary education, any use of dietary supplements, vegetarian diet, alcohol consumption, physical activity, dietary vitamin E, dietary fiber, dietary saturated fat, and smoking history). The unadjusted change in HbA1c per 20 micromol/l increase in vitamin C estimated by linear regression was -0.12% (-0.14 to -0.09) in men and -0.09% (-0.11 to -0.07) in women. After adjusting for the possible confounders, these values were -0.08% (-0.11 to -0.04) in men and -0.05% (-0.07 to -0.03) in women. An inverse association was found between plasma vitamin C and HbA1c. Dietary measures to increase plasma vitamin C may be an important public health strategy for reducing the prevalence of diabetes.