To establish potential effects of glucagon-like peptide I (GLP-I) on blood glucose control in insulin-deficient states, GLP-I [GLP-I(7-36) amide; 10 pmol x kg(-1) x min(-1)] was infused intravenously in six fasting, canine C-peptide-negative, chronically diabetic dogs for 8 h. Blood samples were saved for the analysis of hormones, metabolites, and turnover rates of glucose (6-(3)H-glucose), alanine (U-(14)C-alanine), and urea ((15)N(2)-urea) starting 22 h after the last subcutaneous dose of exogenous insulin. Circulating plasma GLP-I levels rose under infusion from 2.9 +/- 0.8 to 41.4 +/- 10.1 pmol/l. This was efficient to significantly reduce the preexisting diabetic hyperglucagonemia. Since in the utilized model functioning pancreatic beta-cells are lacking, GLP-I had no insulinogenic effect. Compared with control experiments in the same animals receiving saline infusion, glycemia dropped from 20.8 +/- 1.9 to 16.2 +/- 1.0 mmol/l (P < 0.05). This was in parallel to the infusion of GLP-I and was most likely caused by a decrease of elevated glucose production since overall glucose turnover decreased with no alteration in glucose metabolic clearance. Alanine turnover was significantly reduced, obviously reflecting a decline in alanine production in relation to changed muscle glucose uptake under conditions of lower glycemia and overall glucose turnover. There was, however, neither an effect of GLP-I on alanine conversion into circulating glucose nor an effect on urea production rate, indicating unchanged gluconeogenesis from amino acid precursors. We conclude that the blood glucose-lowering effect of GLP-I in an animal model of insulinopenia was shown to be due to a reduction in hepatic glucose output, possibly secondary to reduction in glucagon concentrations leading to decreased glycogenolysis. Whether GLP-I might be therapeutically useful in clinical insulin-deficient diabetes needs to be verified.