Associations of cardiovascular risk factors, including several measures of adiposity, with hyperinsulinemia were assessed in 3562 elderly (71 to 93 years of age) Japanese American men from the Honolulu Heart Program who were examined between 1991 and 1993. In addition, cardiovascular risk factors measured 25 years earlier were also examined in relation to hyperinsulinemia. Hyperinsulinemia was defined as fasting insulin > or = 95th percentile (20 microU/mL) among the subset of subjects (n = 504) who were nonobese and free of clinical diabetes and glucose intolerance. When this definition was applied to the entire population, the prevalence of hyperinsulinemia declined cross-sectionally with age (P < 0.001) from 24.2% in men aged 71 to 74 years to 16.4% in men aged 85 to 93 years. Factors having a positive and independent association with hyperinsulinemia included body mass index (BMI), triglycerides, glucose, hematocrit, use of diabetic medication, heart rate, and hypertension. The association with physical activity was negative. Triglycerides, BMI, diabetic medication, hypertension, and smoking levels measured 25 years earlier were also associated independently with hyperinsulinemia. Associations were similar in nondiabetic subjects. Three measures of adiposity (BMI, waist circumference, and subscapular skinfold thickness) were independently related to hyperinsulinemia cross-sectionally. However, associations involving a difference between the 80th and 20th percentiles in each adiposity measure appeared strongest for BMI (odds ratio (OR) = 4.5, 95% confidence interval (CI) = 3.7 to 5.6) and waist circumference (OR = 4.1, 95% CI = 3.3-5.1) and slightly weaker for subscapular skinfold thickness (OR = 2.1, 95% CI = 1.8-2.5). These findings suggest that features of an insulin resistance syndrome including dyslipidemia, glucose intolerance, hypertension, and obesity, assessed both cross-sectionally and 25 years previously, are associated independently with hyperinsulinemia in elderly Japanese American men.