OBJECTIVE To determine the prevalence of diabetes and examine its association with food intake, anthropometric and metabolic variables, and other coronary risk factors in urban and rural older Mexican populations. DESIGN A cross-sectional study. SETTING Three Mexican communities (urban areas of medium and low income and a rural area). PARTICIPANTS A total of 121 men and 223 women aged 60 years and older and 93 men and 180 women aged 35 to 59 years were selected randomly for inclusion in the survey, which was derived from the CRONOS study (Cross-Cultural Research on Nutrition in the Older Adult Study Group) promoted by the European Economic Community. MEASUREMENTS A personal interview assessed demographic information, personal medical history, and functional status, and a 24-hour diet recall was obtained. A physical examination included anthropometric and blood pressure measurements. A fasting blood sample was obtained for measurements of lipids, insulin, and glucose. RESULTS Diabetes prevalence was higher in men than in women for all age groups: 16.7% versus 9.5% in younger adults and 30.8% versus 22.8% in older adults. For all age groups, diabetes was more highly prevalent in urban communities. Using a multivariate stepwise logistic regression, variables associated independently with diabetes in older individuals were: gender (male sex: OR = 2.1; P < .009); diminished carbohydrate intake in the diet (OR = 0.77; P < .03); central distribution of adiposity (OR = 1.9; P < .03); and functional disability (OR = 2.3; P < .01). This relationship was not observed with living area, income, education, fiber and alcohol intake, body mass index, or age. Individuals 80 years and older had a diminished atherogenic risk profile. Diabetes in older people was associated significantly with hypertriglyceridemia, impaired functional status, and an increased prevalence of ischemic heart disease; in younger adults diabetes was associated with low density lipoprotein (LDL) hypercholesterolemia, hypertriglyceridemia, and a proportionally higher fat intake. CONCLUSION This survey confirms the high prevalence of diabetes in the older Mexican population - particularly in men and in individuals living in urban areas - associated with an increased prevalence of other coronary risk factors. Diabetes was associated with higher fat, low carbohydrate, low fiber diets and increased prevalence of central distribution of adiposity. In the older subjects, diabetes was associated significantly with hypertriglyceridemia, impaired functional status, and increased prevalence of ischemic heart disease. A bias produced by early mortality and a survivorship effect must be considered in studies of older individuals. The health situation in the older Mexican population presents a complex problem that needs correct diagnosis and better strategies to benefit those segments of the population at increased risk.