The relationships between body fatness, adipose tissue distribution, plasma glucose, insulin levels, lipoprotein levels, and resting blood pressure were studied in 81 men aged 36.0 +/- 3.3 years (mean +/- s.d.) (body mass index (BMI): 27.4 +/- 3.8 kg/m2, percentage body fat: 26.4 +/- 6.6%). Systolic and diastolic blood pressures (BP) were significantly associated with the BMI (r = 0.31, r = 0.33, P < 0.01), the waist circumference (r = 0.33, r = 0.27; P < 0.01) as well as with adipose tissue areas measured by computerized tomography (CT) (0.27 < or = r < or = 0.36, P < 0.01). Furthermore, the relative accumulation of subcutaneous abdominal fat, as estimated by the ratio of abdominal to femoral adipose tissue areas measured by CT, was positively correlated with systolic and diastolic BP (P < 0.01). Fasting plasma insulin level (r = 0.30, P < 0.01) as well as the insulin area measured during an oral glucose tolerance test (0.34 < or = r < or = 0.37, P < 0.01) were significantly correlated with blood pressure. Systolic and diastolic BP were significantly associated with HDL2-cholesterol (C) as well as with the HDL2-C/HDL3-C ratio (-0.24 < or = r < or = -0.34), whereas triglycerides (r = 0.23) and the HDL-C/C ratio (r = -0.23) were significantly correlated with diastolic BP only (P < 0.05). Multivariate analysis indicated that the insulin area was the most important variable associated with blood pressure and that this association was independent of total body fatness and regional adipose tissue distribution. Plasma insulin levels explained 14% and 11% of the variance observed in the systolic and diastolic blood pressures respectively. These results suggest that most of the association between abdominal obesity and high blood pressure is mediated by the hyperinsulinemia and/or the related insulin resistant state.