To investigate whether recombinant human insulin-like growth factor-I (rhIGF-I) has direct effects on the insulin requirement to maintain euglycemia independent of the growth hormone (GH) level, nine subjects with insulin-dependent diabetes mellitus ([IDDM] seven females; median (range) age, duration of diabetes, and hemoglobin A(1C) [HbA(1C)], 16.9 (12.5 to 21.9) years, 11.8 (4.8 to 16.8) years, and 9.8% (7.9% to 14.1%), respectively) underwent two euglycemic studies (6:00 PM to 8:00 AM) after double-blind subcutaneous administration of rhIGF-I/placebo (40 μg/kg). Octreotide infusion (300 ng/kg/h) suppressed endogenous GH, and three identical discrete GH pulses were infused on both nights. Variable-rate insulin infusion maintained euglycemia. Samples were taken every 15 minutes (glucose and GH), 30 minutes (insulin and intermediate metabolites), and 60 minutes (IGF-I end nonesterified fatty acids [NEFA]). Variables were analyzed during the steady-state period of euglycemia (4:00 to 8:00 AM). Data are expressed as the mean ± SEM. The insulin infusion rate and free-insulin level were both significantly reduced after rhIGF-I administration (0.13 ± 0.03 v placebo 0.23 ± 0.05 mU/kg/min, P = .04, and 8.4 ± 1.3 v placebo 12.1 ± 1.4 mU/L, P = .03, respectively). GH pulse-related changes in the insulin requirement observed after placebo were not present after rhIGF-I. Glucagon levels were equally suppressed on both nights. Insulin clearance was not altered after rhIGF-I administration. NEFA end ketone levels also were not different on the 2 nights. In conclusion, in adolescents and young adults with diabetes, rhiGF-I administration directly affected insulin requirements independent of GH levels, but had no effect on fatty acid or ketone levels. This difference is related to the abolition of changes in the insulin requirement after GH pulses, and would suggest a complex interaction between GH and IGF-I on insulin action.