The role of insulin in cardiovascular disease is uncertain, and studies in elderly or minority populations are infrequent. Fasting and 2-hour insulin concentrations and their cross-sectional associations with cardiovascular risk factors were examined in 3562 elderly (aged 71 to 93 years) Japanese American men from the Honolulu Heart Program who were reexamined between 1991 and 1993. Insulin distributions were skewed (mean and median: 16.8 and 12 microU/mL for fasting; 117.2 and 93 microU/mL for 2-hour); fasting but not 2-hour insulin levels declined significantly with age (P < .0001 and P = .54, respectively). Factors most strongly correlated with insulin included measures of obesity, fat distribution, and levels of triglyceride, glucose (r = .38 to r = .50 fasting, r = .21 to r = .27 2-hour), and HDL cholesterol (r = -.41 and r = -.22, respectively). Other correlates included fibrinogen, hematocrit, heart rate, blood pressure, cigarettes per day (all positive), alcohol, physical activity, and forced vital capacity (negative). Associations were also evident across risk factor quintiles. Insulin levels were significantly elevated in men with hypertension and diabetes. In multiple linear regression analyses, log10 fasting insulin was positively and independently associated with body mass index, triglycerides, glucose, fibrinogen, hematocrit, heart rate, diabetes, and hypertension and negatively associated with HDL cholesterol, physical activity, and forced vital capacity. In general, results were similar for log10 2-hour insulin and when subjects who fasted < 12 hours or had diabetes were excluded. Substitution of medication use and blood pressure for hypertension indicated independent associations of medication use but not blood pressure with insulin.(ABSTRACT TRUNCATED AT 250 WORDS)