Relations of Serum Ascorbic Acid and α-Tocopherol to Diabetic Retinopathy in the Third National Health and Nutrition Examination Survey

Amy E. Millen 1 Michael Gruber 1 Ronald Klein 1 Barbara E. K. Klein 1 Mari Palta 0 Julie A. Mares ) 1 0 Department of Preventive Medicine, University of Wisconsin Medical School , Madison, WI 1 Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School , Madison, WI The protective relation of ascorbic acid and -tocopherol to the development of diabetic retinopathy has not been thoroughly evaluated in epidemiologic studies. The association of prevalent diabetic retinopathy with serum ascorbic acid and -tocopherol was studied among participants with type 2 diabetes (40 years) (n = 998) in the Third National Health and Nutrition Examination Survey (1988-1994); 20% of the sample (n = 199) had prevalent retinopathy. The overall odds ratio for retinopathy among participants in quartile 4 compared with quartile 1 for serum ascorbic acid was 1.3 (95% confidence interval: 0.8, 2.3), with a p for trend = 0.60 after adjustment for the confounders of smoking, race, waist/hip ratio, hypertension, and duration of diabetes. The overall odds ratio for retinopathy among participants in quartile 4 compared with quartile 1 for serum -tocopherol was 2.7 (95% confidence interval: 1.6, 4.6), with a p for trend = 0.14 after adjustment for confounders. After removal of supplement users of vitamin C (n = 307) or vitamin E (n = 298), the odds ratio changed direction or was attenuated: adjusted odds ratios for retinopathy among participants in quartile 4 compared with quartile 1 for serum ascorbic acid and -tocopherol = 0.7 (95% confidence interval: 0.3, 1.4) and 1.6 (95% confidence interval: 0.9, 2.9), respectively. In summary, no significant associations were observed between serum levels of major dietary antioxidants and retinopathy. Recent use of supplements for treatment of complications of diabetes may explain the direct associations. ascorbic acid; diabetic retinopathy; nutrition surveys; vitamin E Abbreviation: NHANES III, Third National Health and Nutrition Examination Survey. - Persons with diabetes suffer from micro- and macrovascular disease such as diabetic retinopathy, a disease of the retinal microvasculature. Retinopathy occurs in just over 2.5 percent of the population aged 18 years or more (1). Ascorbic acid and -tocopherol are nutrients hypothesized to prevent retinopathy by affecting presumed pathogenic factors: protein glycosylation, insulin sensitivity, retinal blood flow, and oxidative stress (224). Not all studies, however, support these hypotheses (2530). The majority of research investigating the potential protective effect of ascorbic acid and -tocopherol has been conducted in diabetic animal models and short-term supplementation trials. Only one population-based study has investigated the relation between the risk of severity of diabetic retinopathy and these antioxidant nutrients, using dietary intake from one 24-hour recall (31). This previous study showed an increased risk of diabetic retinopathy with increasing intake of vitamins C and E (31). Additional research is needed to assess the association between these micronutrients and retinopathy in the general population of persons with diabetes. The Third National Health and Nutrition Examination Survey (NHANES III) provides the opportunity to investigate the association between biomarkers of ascorbic acid and -tocopherol intakes and the odds of prevalent diabetic retinopathy in a large sample of free-living persons (32). The purpose of this research was to investigate the association between serum ascorbic acid and -tocopherol concentrations and prevalent retinopathy using data from NHANES III. We hypothesized that participants in the highest, compared with the lowest, quartile of serum ascorbic acid and -tocopherol would have less retinopathy. MATERIALS AND METHODS Study design NHANES III was conducted by the Centers for Disease Control and Prevention between 1988 and 1994 using a nationally representative probability sample of the US noninstitutionalized civilian population. A multistage probability sample design was used with oversampling of certain subgroups to allow more precise estimates, including nonHispanic Blacks, Mexican Americans, and adults over 60 years of age. All eligible participants were invited to complete a household interview to obtain information on demographic and socioeconomic characteristics, health history, and dietary behavior. All interviewed persons were invited to participate in a medical examination that was conducted at a mobile examination center. The examination included body measurements, fundus photography, urine collection, a 24-hour dietary recall interview, and phlebotomy to measure a number of hematologic factors. Sample selection Persons with type 2 diabetes were identified from a total of 14,464 persons 40 years of age or more who were invited to participate in NHANES III. Of the 14,464 persons aged 40 or more years who were eligible for the survey, 11,448 persons were interviewed and 9,737 were examined in the mobile examination center. These analyses were restricted to nonHispanic Whites, non-Hispanic Blacks, and Mexican Americans. Persons of other races or ethnicities were excluded (n = 371), because the sample sizes were too small to allow statistically reliable estimates. A total of 1,434 persons were classified as having type 2 diabetes. Participants who reported being diagnosed with diabetes by their physician and who recalled being diagnosed at age 30 or more years were considered to have type 2 diabetes (n = 1,137). If a participant fasted for 10 or more hours, completed an oral glucose tolerance test, and had a plasma glucose level equal to or greater than 200 mg/dl, he/ she was also considered to have type 2 diabetes (n = 297) (33). Persons with missing or ungradable fundus photographs for retinopathy were excluded (n = 256). Persons were also excluded if they had missing serum ascorbic acid or -tocopherol values (n = 75), missing food frequency data (n = 1), or missing data for potential confounders or other significant covariates (n = 104). Thus, a total of 998 participants with type 2 diabetes were eligible for inclusion in these analyses, and 20 percent (n = 199) had retinopathy. Overall, the sample used in these analyses was 70 percent of all nonHispanic White, non-Hispanic Black, or Mexican-American persons 40 years of age or more classified with type 2 diabetes in NHANES III (n = 1,434). To understand the differences between the sample we selected for our study (n = 998) and those persons who were excluded (n = 436) from the original sample of non-Hispanic White, non-Hispanic Black, and Mexican-American persons with type 2 diabetes for missing data, we compared the demographic, physical, lifestyle, dietary, and medical characteristics of these two samples. The persons with type 2 diabetes who were excluded from this study were older (65 vs. 62 years, p < 0.0001), more likely to be Mexican American and non-Hispanic (...truncated)