The diabetogenic effects of glucocorticoids appear to be dose dependent. To determine the effects of different doses of dexamethasone on glucose metabolism, we performed frequently sampled intravenous glucose tolerance tests in 20 healthy young men. Glucose kinetics were analysed by the minimal model. Ten subjects received low-dose dexamethasone (2 mg/day) for 3 days, and the other 10 received high-dose dexamethasone (6 mg/day) for 3 days. The rate of glucose disappearance (KG) did not decrease in the low-dose group (2.46 +/- 0.20 to 2.19 +/- 0.11% min-1, P = 0.35). In contrast, KG in the high-dose group did decrease significantly (2.43 +/- 0.29 to 1.81 +/- 0.11% min-1, P < 0.05). The factor responsible for the decline in KG in the high-dose group was not glucose effectiveness because these values did not change in either group. The insulin sensitivity decreased significantly, by 46% in the low-dose group and 69% in the high-dose group [17.1 +/- 2.7 to 9.2 +/- 1.5 and 18.5 +/- 3.7 to 5.8 +/- 0.9 x 10(-5) min-1 (pmol/L)-1, P < 0.001 and P < 0.01, respectively]. The insulin area (0-20 min) increased significantly, by 104% in the low-dose group and 114% in the high-dose group [3412.6 +/- 609.7 to 6972.7 +/- 1450.1 and 4086.7 +/- 864.5 to 8750.0 +/- 1451.6 (pmol/L) min, P < 0.01 and P < 0.01, respectively]. Insulin sensitivity x insulin area as an estimate of insulin-dependent glucose uptake and insulin's action to suppress hepatic glucose production decreased significantly in the high-dose group (0.588 +/- 0.112 to 0.441 +/- 0.073, P < 0.05), but did not change in the low-dose group (0.436 +/- 0.050 to 0.484 +/- 0.032, P = 0.77). Therefore, the decline in KG in the high-dose group may be associated with the compensatory failure of pancreatic beta-cells against for the insulin resistance.