In 197 patients with coronary artery disease (CAD) who underwent coronary angiography, 75 g oral glucose tolerance test (OGTT) was performed, measuring plasma glucose(PG) and plasma insulin (IRI) at 4 time points (0, 30, 60 and 120 min). Subjects were separated into two groups by their insulinogenic index (I.I. = delta IRI/delta PG from 0 up to 30 min), 99 cases with good insulin response (I.I. > or = 0.4) and 98 cases with poor insulin response (I.I. < 0.4). Only two subjects were diabetic in good insulin response, while 37 were diabetic in poor insulin response. The severity of coronary atherosclerosis was expressed as a coronary index (CI), calculated according to Balcon's method. Fasting PG and the sum of PG were significantly higher in the latter group, while the sum of IRI was significantly lower. CI was not significantly different, however. In the group with good insulin response, the severity of CAD was significantly correlated to fasting IRI (n = 99, r = -0.387, P < 0.02), but, there was no such relationship in the other group. We conclude that hyperinsulinemia might be a risk factor for ischemic heart disease, but in diabetics it is difficult to link the relationship between fasting IRI and CI.