At supraphysiological levels, IGF-I bypasses some forms of insulin resistance and has been proposed as a therapeutic agent in the treatment of diabetes. Unfortunately, side effects of high-dose IGF-I (100-250 µg/kg) have precluded its clinical use. Low-dose IGF-I (40-80 µg/kg), however, shows minimal side effects but has not been systematically evaluated. In our previous study under conditions of declining glucose, low-dose IGF-I infusion was more effective in stimulating glucose utilization, but less effective in suppressing glucose production and lipolysis than low-dose insulin. However, under conditions of hyperglycemia, we could not observe any differential effects between high-dose infusions of IGF-I and insulin. To determine whether the differential effects of IGF-I and insulin are dose-related or related to the prevailing glucose level, 3 h glucose clamps were performed in the same animal model as in the previous studies, i.e. the moderately hyperglycemic (175 mg/dl) insulin-infused depancreatized dog, with additional infusions of low-dose IGF-I (67·8 µg/kg, i.e. 29·1 µg/kg bolus plus 0·215 µg/kg per min infusion; n=5) or insulin 49·5 mU/kg (9 mU/kg bolus plus 0·45 mU/kg per min; n=7). As in the previous study under conditions of declining glucose, low-dose IGF-I had significant metabolic effects in vivo, in our model of complete absence of endogenous insulin secretion. Glucose production was similarly suppressed with both IGF-I and insulin, by 54 3 and 56 2% .. (P=NS) respectively. Glucose utilization was stimulated to the same extent (IGF-I 5·2 0·2, insulin 5·5 0·3 mg/kg per min, P=NS). Glucagon, free fatty acid, glycerol, alanine and betahydroxybutyrate, were suppressed, while lactate and pyruvate levels were raised, similarly with IGF-I and insulin. We conclude that: (i) differential effects of IGF-I and insulin may be masked under hyperglycemic conditions, independent of the hormone dose; (ii) low-dose IGF-I has no selective advantage over additional insulin in suppressing glucose production and lipolysis, nor in stimulating glucose utilization during hyperglycemia and subbasal insulin infusion when insulin secretion is absent, as in type 1 diabetes mellitus.