Methods for analyzing multivariate binary data, with association between outcomes of interest.

A likelihood based method is proposed for multivariate categorical data. It is assumed that, together with the marginal outcomes, the set of pairwise associations between outcomes is of scientific interest. The focus is on binary outcomes and it is indicated how the proposed method generalizes to categorical outcomes. A connection with second-order generalized estimating equations (GEE2) is established. The method is applied to analyze data from a developmental toxicity study.

Association of zinc and antioxidant nutrients with age-related maculopathy.

OBJECTIVE: To quantify relationships between dietary intake of zinc and antioxidant nutrients and early and late age-related maculopathy (ARM). DESIGN: A retrospective longitudinal cohort design using data pertaining to diets in the past (1978-1980), which were assessed retrospectively using a food frequency questionnaire. SETTING: Beaver Dam, Wis. PATIENTS: A 50% random sample of free-living Beaver Dam Eye Study participants, 43 to 86 years of age (N = 1968). MAIN OUTCOME MEASURE: The presence of early and late ARM determined from fundus photography. RESULTS: People in the highest vs lowest quintiles for intake of zinc from foods had lower risk for early ARM (odds ratio = 0.6, 95% confidence interval, 0.4-1.0, P for trend < .05). This relationship appeared to be stronger for some types of early ARM (increased retinal pigment) than for others. Zinc intake was unrelated to late ARM. However, small numbers (n = 30) of people with this condition limit the ability to draw conclusions about this later stage. Levels of carotenoids were unrelated to early or late ARM. Odds for early ARM were lower in people in the highest vs lowest quintiles for the intake of vitamins C or E. However, these associations were not statistically significant. CONCLUSIONS: The data are weakly supportive of a protective effect of zinc on the development of some forms of early ARM. Prospective studies are needed to further evaluate the potential influence of these and other nutritional factors on different types and stages of age-related macular degeneration.

Alcohol use and age-related maculopathy in the Beaver Dam Eye Study.

PURPOSE: Because there are few epidemiologic data examining the relationship between alcohol use and age-related maculopathy, we investigated this relationship via the population-based Beaver Dam Eye Study. METHODS: Alcohol consumption in the year before examination was determined by a standardized questionnaire. Age-related maculopathy status was determined by grading stereoscopic color fundus photographs. RESULTS: While controlling for other factors, consumption of beer in the past year was related to greater odds of increased retinal pigment degeneration (odds ratio, 1.13; 95% confidence interval, 1.02, 1.25) and exudative macular degeneration (odds ratio, 1.41; 95% confidence interval, 1.05, 1.88). Current consumption of neither wine nor liquor was related to early or late age-related maculopathy. CONCLUSIONS: These data suggest a relationship between beer consumption and greater odds of having exudative macular degeneration. Whether this is from a toxic effect specific to beer or from other unknown confounders cannot be determined.

Diet and nuclear lens opacities.

Relations between diet and nuclear opacities in the lens of the eye were investigated in a population-based cohort of middle-aged and older adults who lived in Beaver Dam, Wisconsin. Nuclear sclerosis was assessed from photographs of the lens taken during 1988-1990 in 1,919 persons in the Beaver Dam Eye Study. Diets in the past (1978-1980) were assessed retrospectively with the use of a food frequency questionnaire in home interviews. Relations with intake of foods and nutrients were evaluated using logistic regression analyses. In men, after controlling for age, smoking, and heavy drinking, intakes of numerous nutrients in the highest versus lowest quintile were associated with 40-50 percent reduced odds of more severe nuclear sclerosis. Relations with some nutrients (vitamins A, C, and E, riboflavin, thiamin, niacin) were at least partly explained by previously identified inverse associations with multivitamin use. Relations with other nutrients (folate, alpha-carotene, and dietary fiber) appeared to reflect associations with intake of foods, particularly vegetables. Inverse associations with individual nutrients and foods were often weaker or nonexistent in women. These data indicate that the intake of vitamin supplements (in men and women) and certain foods (particularly in men) may explain associations of several nutrients with risk for nuclear sclerosis.

Relation between lens opacities and vitamin and mineral supplement use.

PURPOSE: The authors evaluated the cross-sectional and longitudinal relations between lens opacities and intake of vitamins and minerals in supplements. METHODS: Type and quantity of vitamins and minerals taken currently and 10 years before interview were obtained in a population-based sample of persons in the Nutritional Factors in Eye Disease Study (n = 2152). RESULTS: In persons without diabetes, regular use of multivitamin preparations 10 years in the past was associated with decreased risk for nuclear sclerosis (odds ratio [OR], 0.6; 95% confidence interval [CI], 0.5, 0.8) and increased the risk for cortical opacities (OR, 1.6; CI, 1.1, 2.2). In persons with diabetes, past multivitamin use was not associated with nuclear sclerosis (OR, 1.1; CI, 0.4,2.7) but with decreased risk for cortical opacities (OR, 0.1; CI, 0.0, 0.9). Cross-sectional associations with current intake of supplements were weaker, but in the same direction. Multivitamin use was not cross-sectionally or longitudinally related to posterior subcapsular cataract in persons with or without diabetes. CONCLUSION: Use of vitamin supplements is longitudinally associated with nuclear and cortical opacities. However, the direction of the association is influenced by the type of opacity and diabetes status.

Is there evidence of an estrogen effect on age-related lens opacities? The Beaver Dam Eye Study.

OBJECTIVE: To evaluate a possible association between estrogen and lens opacities. DESIGN: A prevalence survey of age-related eye disease. SETTING: Beaver Dam, Wis. PARTICIPANTS: A total community sample of persons 43 through 84 years of age at the time of a private census were recruited for examination, medical history, and photographs of the lenses of the eyes. Data from women are included in this report. MAIN OUTCOME MEASURES: Photographs were graded by protocol for the presence and severity of age-related lens opacities. RESULTS: Current use of postmenopausal estrogens was associated with a decreased risk of more severe nuclear sclerosis. Younger age at menarche was also associated with a protective effect regarding nuclear sclerosis. Older age at menopause was associated with a decreased risk of cortical opacities. CONCLUSION: These data are suggestive of a modest protective effect of estrogen exposure on the lenses of women. Further evaluation of this may be possible in studies currently under way in which postmenopausal estrogens are being tested for other (protective) effects.

Relationship of drinking alcohol and smoking to prevalence of open-angle glaucoma. The Beaver Dam Eye Study.

PURPOSE: To evaluate a possible association of glaucoma with alcohol ingestion and cigarette-smoking behavior. METHODS: A population-based survey was performed of persons 43 to 84 years of age in Beaver Dam, Wisconsin (n = 4926). The diagnosis of glaucoma was based on visual field testing, measurement of intraocular pressure, and cup-to-disc ratios based on detailed protocols for examination, measurement, and gradings of photographs. Alcohol intake and smoking history were based on questionnaire responses. The prevalence of glaucoma in "heavy drinkers," whether current or past, was contrasted with the prevalence in those not reporting heavy drinking. Prevalences in never, past, and current smokers were compared. RESULTS: There was no difference in frequency of glaucoma by drinking status. Similarly, there was no difference in the frequency of glaucoma by cigarette-smoking status. CONCLUSION: In these prevalence data, neither heavy drinking nor cigarette-smoking behavior was related to the prevalence of open-angle glaucoma. This finding may reflect a real absence of a pathologic effect.

Glycosylated hemoglobin concentrations and vitamin E, vitamin C, and beta-carotene intake in diabetic and nondiabetic older adults.

Studies indicate that large doses of all-rac-alpha-tocopherol in people with diabetes or ascorbic acid in nondiabetic subjects reduces protein glycosylation. The mechanisms by which these nutrients influence glycosylation are poorly understood but may be related to their ability to function as antioxidants. We examined the relationship between glycosylated hemoglobin (GHb) and intake of vitamins E and C and beta-carotene in a population-based sample of middle-aged and older adults participating in the Beaver Dam Eye Study. In people with diabetes, no significant associations were observed between GHb and intake of vitamins E and C and beta-carotene. In people without diabetes, energy-adjusted vitamin C intake was negatively associated with GHb after age and sex were controlled for (dietary, P = 0.02; total, P = 0.04). No significant relationships between GHb and intake of vitamin E and beta-carotene were observed.

The association of microalbuminuria with diabetic retinopathy. The Wisconsin Epidemiologic Study of Diabetic Retinopathy.

PURPOSE: To investigate the relationship between microalbuminuria and the presence and severity of diabetic retinopathy in a large population-based cohort of individuals with diabetes. METHODS: Microalbuminuria was measured by an agglutination inhibition assay in random urine samples obtained from participants in the Wisconsin Epidemiologic Study of Diabetic Retinopathy (n = 1139) who did not have hematuria, gross proteinuria, or a history of renal disease. Retinopathy was determined from stereoscopic color fundus photographs graded according to a modification of the Airlie House Classification System. RESULTS: Younger-onset (diagnosed with diabetes before 30 years of age) and older-onset (diagnosed with diabetes when 30 years of age or older) individuals with microalbuminuria were more likely to have retinopathy than those without microalbuminuria. Younger-onset individuals who had microalbuminuria at the time of examination were more likely to have proliferative retinopathy than younger-onset subjects with normoalbuminuria. These relationships remained after controlling for glycemia, hypertension, duration of diabetes, and other potential confounders. CONCLUSIONS: Microalbuminuria is associated cross-sectionally with the presence of retinopathy in persons with diabetes and with the presence of proliferative disease in younger-onset individuals. These data suggest that microalbuminuria may be a marker for the risk of proliferative retinopathy developing. If longitudinal studies confirm these findings, individuals with insulin-dependent diabetes mellitus (IDDM) who have microalbuminuria may benefit from ophthalmologic follow-up.

A diet history questionnaire ranks nutrient intakes in middle-aged and older men and women similarly to multiple food records.

We evaluated the reproducibility of a modified version of the National Cancer Institute diet history questionnaire and also studied comparability of nutrient estimates from this questionnaire to those from four 2-d food records. Subjects (n = 211) were from a population-based sample of middle-aged and older adults participating in the Beaver Dam Eye Study in south-central Wisconsin. Median age-specific correlation coefficients between the questionnaire, administered twice at a 3-mo interval, were 0.8 in men and 0.7 in women (range = 0.5 to 0.9). Questionnaire estimates of protein, cholesterol, thiamin, niacin, iron and zinc were consistently lower than food record estimates whereas estimates of fat were higher. Correlation coefficients between estimates from the questionnaire and records were generally > 0.5, indicating overall good agreement in ranking. However, the range in correlation coefficients was wide [ranging from 0.06 for iron (without supplements) in middle-aged women to 0.8 for alcohol in middle-aged men and women]. Questionnaire estimates similarly classified persons into the lowest two food record quintiles of intake 50 to 93% of the time depending on nutrient. These results suggest that the questionnaire produces nutrient estimates that rank individuals on the basis of intake of most nutrients similarly to estimates from multiple food records.

Nutrient supplements contribute to the dietary intake of middle- and older-aged adult residents of Beaver Dam, Wisconsin.

Intake of macronutrients and micronutrients was estimated from a diet history questionnaire in a population-based sample of persons (n = 2152), 43-86 y of age. Men and women 65 y of age or older consumed less protein, fat, cholesterol, niacin, riboflavin, calcium, iron, zinc and lycopene than their middle-aged counterparts. After including nutrient intake from supplements, age differences in intake were no longer significant for zinc in women or for niacin and iron in men. After further adjusting for energy intake, only protein, cholesterol, and lycopene mean intakes were lower in older compared with middle-aged men and women. Regular use of supplements was most prevalent among older men and women (40 and 48%, respectively). Supplement use was greater in the current time period compared with retrospective reports of supplement use 10 y before interview. The impact of supplement intake on median values of intake was greatest in women, particularly for vitamins A, C and E, riboflavin and calcium. The high prevalence of supplement use emphasizes the importance of considering this source of nutrients in future investigations in which dietary intake is measured.

Alcohol use and lens opacities in the Beaver Dam Eye Study.

The relationship between alcohol use and lens opacities was examined in a large (N = 4926) population-based study of adults aged 43 to 86 years in Beaver Dam, Wis. These data were collected from 1988 to 1990. Alcohol history was determined by a standardized questionnaire. Prevalence and severity of cataract were determined by masked grading of photographs obtained using a slit-lamp camera and retroillumination. A history of heavy drinking was related to more severe nuclear sclerotic, cortical, and posterior subcapsular opacities (odds ratios, 1.34, 1.38, and 1.57, respectively). These relationships remained after adjusting for other risk factors such as smoking. Moderate liquor consumption was associated with less severe nuclear sclerosis (odds ratio, 0.81). Participants who drank wine had less severe nuclear sclerosis (odds ratio, 0.84) and cortical opacities (odds ratio, 0.84) than those who did not. Increased consumption of beer was related to increased risk of cortical opacities.