Despite recent interest in the therapeutic potential of recombinant human insulin-like growth factor-I (rhIGF-I) in the treatment of diabetes mellitus, its mechanism of action is still not defined. We have studied the effects of low-dose bolus subcutaneous rhIGF-I (40 microg/kg and 20 microg/kg) on insulin sensitivity, growth hormone (GH) and glucagon levels in seven young adults with insulin-dependent diabetes mellitus (IDDM) using a randomized double-blind placebo-controlled crossover study design. Each was subjected to a euglycemic clamp (5 mmol/L) protocol consisting of a variable-rate insulin infusion clamp (6:00 PM to 8:00 AM) followed by a two-dose hyperinsulinemic clamp (insulin infusion of 0.75 mU x kg(-1) x min(-1) from 8 to 10 AM and 1.5 mU x kg(-1) x min(-1) from 10 AM to 12 noon) incorporating [6,6 2H2]glucose tracer for determination of glucose production/utilization rates. Following rhIGF-I administration, the serum IGF-I level (mean +/- SEM) increased (40 microg/kg, 655 +/- 90 ng/mL, P < .001; 20 microg/kg, 472 +/- 67 ng/mL, P < .001; placebo, 258 +/- 51 ng/mL). Dose-related reductions in insulin were observed during the period of steady-state euglycemia (1 AM to 8 AM) (40 microg/kg, 48 +/- 5 pmol/L, P = .01; 20 microg/kg, 58 +/- 8 pmol/L, P = .03; placebo, 72 +/- 8 pmol/L). The mean overnight GH level (40 microg/kg, 9.1 +/- 1.4 mU/L, P = .04; 20 microg/kg, 9.6 +/- 2.0 mU/L, P = .12; placebo, 11.3 +/- 1.7 mU/L) and GH pulse amplitude (40 microg/kg, 18.8 +/- 2.9 mU/L, P = .04; 20 microg/kg, 17.0 +/- 3.4 mU/L, P > .05; placebo, 23.0 +/- 3.7 mU/L) were also reduced. No differences in glucagon, IGF binding protein-1 (IGFBP-1), acetoacetate, or beta-hydroxybutyrate levels were found. During the hyperinsulinemic clamp conditions, no differences in glucose utilization were noted, whereas hepatic glucose production was reduced by rhIGF-I 40 microg/kg (P = .05). Our data demonstrate that in subjects with IDDM, low-dose subcutaneous rhIGF-I leads to a dose-dependent reduction in the insulin level for euglycemia overnight that parallels the decrease in overnight GH levels, but glucagon and IGFBP-1 levels remain unchanged. The decreases in hepatic glucose production during the hyperinsulinemic clamp study observed the following day are likely related to GH suppression, although a direct effect by rhIGF-I cannot be entirely discounted.