The metabolic consequences of the prolonged systemic insulin delivery associated with human pancreas transplantation have not been precisely defined. To determine if systemic insulin delivery in the absence of immunosuppressive agents results in alterations in hepatic or extrahepatic insulin action, three groups of dogs were studied 2 months after either a sham operation or after their pancreatic venous drainage was severed and anastomosed to the inferior vena cava or portal vein (sham, peripheral and portal groups, respectively). The pattern of venous drainage was documented by measuring vena cava and portal insulin concentrations before and after glucose injection. Systemic insulin concentrationswerehigher(p < 0.05)intheperipheralgroupthaninthe portal group both following a 14-h fast and after intravenous glucose. During a hyperinsulinaemic euglycaemic clamp (lmU.kg-l.min-1), glucose utilization (measured using [63H] glucose) was slightly lower (p = 0.07) in the peripheral than in the portal group. Hepatic glucose release was equal in all groups. Carbon dioxide incorporation into glucose (an estimate of gluconeogenesis) was higher in the portal than peripheral group in the fasted state but not during insulin infusion. Plasma concentrations and flux rates of fatty acids and amino acids did not differ between groups. We conclude that chronic systemic insulin delivery results in a) systemicbut not portal hyperinsulinaemia, b)a minimal impairment in insulin-stimulated glucose uptake, without altering insulin-induced suppression of hepatic glucose release, and c) no effect on fatty acid or amino acid turnover. Although chronic systemic insulin delivery appears to have a minimal effect on insulin action, it remains to be determined whether it has other deleterious effects such as enhancing atherogenesis.