The relationships among blood pressure, obesity, glucose tolerance, and serum insulin concentration were studied in 2873 Pima Indians aged 18-92 yr (mean 37 yr). Age- and sex-adjusted to the Pima population, the prevalence of hypertension (systolic blood pressure greater than or equal to 160 mmHg, diastolic blood pressure greater than or equal to 95 mmHg, or receiving drug treatment) was 7.1% for subjects with normal glucose tolerance compared with 13.0% for subjects with impaired glucose tolerance (IGT) and 19.8% for those with non-insulin-dependent diabetes mellitus (NIDDM) (P less than 0.001). The prevalence ratio of hypertension was 1.8 (95% confidence interval [CI] 1.2-2.5) for IGT and 2.6 (95% CI 2.0-3.2) for NIDDM compared with normal glucose tolerance, controlled for age, sex, and body mass index (BMI). In logistic regression analysis, hypertension was positively related to age, male sex, BMI, glucose tolerance, and fasting but not 2-h postload serum insulin concentration. Among subjects not taking antihypertensive drugs, however, neither fasting nor 2-h postload serum insulin was significantly related to hypertension. Furthermore, in 2033 subjects receiving neither antihypertensive nor antidiabetic drugs, blood pressure was not significantly correlated to fasting insulin concentration, and 2-h postload serum insulin was negatively correlated with diastolic blood pressure. In conclusion, insulin is not significantly related to blood pressure in Pima Indians not receiving antihypertensive drugs. Higher insulin concentrations in drug-treated hypertensive patients might result from the treatment rather than contribute to the pathogenesis of hypertension. Thus, these data do not support a major role for insulin in determining the occurrence of hypertension or regulation of blood pressure in Pima Indians.