The relations of regional adipose tissue (AT) distribution measured by computed tomography (CT) to plasma insulin-glucose homeostasis and lipoprotein-lipid levels were studied in 58 obese and 29 lean control men. In the group of obese men, the visceral AT area measured by CT was positively correlated with fasting plasma triglyceride and insulin levels and with glucose and insulin areas under the curves measured during a 75-g oral glucose tolerance test. Visceral AT area was also negatively associated with plasma high-density lipoprotein (HDL) and HDL2 cholesterol levels. The relative accumulation of abdominal fat, estimated by the ratio of abdominal to femoral AT areas obtained by CT, was also a significant correlate of indices of carbohydrate metabolism and was the best univariate correlate of plasma lipoprotein levels. No significant associations were observed between the visceral AT area, the ratio of abdominal to femoral AT areas, and indices of carbohydrate and lipoprotein metabolism in the group of lean men. On the other hand, the subcutaneous abdominal AT area was a significant correlate of the glucose area under the curve in both groups of men, but this association was not independent from the percentage of total body fat. No relationship was observed between the femoral AT area and indices of carbohydrate metabolism in either lean or obese groups. In obese men, however, the femoral AT area was negatively correlated with plasma triglyceride concentration and positively correlated with plasma HDL and HDL2 cholesterol levels. Comparison between two subgroups of equally obese men showing either low or high levels of visceral AT and a group of lean men generally revealed that only obese men with high levels of visceral AT showed significant metabolic alterations compared with lean men. These results demonstrate that the amount of visceral AT and the ratio of abdominal to femoral AT measured by CT are important correlates of the alterations in carbohydrate and lipoprotein metabolism observed in obese men. In addition, our results suggest that obesity is required to observe significant associations between body fat distribution measured by CT and metabolic variables. Finally, we suggest that, in men, a high accumulation of femoral fat may be protective against the adverse effects of obesity, particularly abdominal obesity, on plasma lipoprotein levels.