To evaluate the metabolic consequences of pancreas transplantation with systemic venous drainage on beta-cell function, we examined insulin and C-peptide responses to glucose and arginine in type I (insulin-dependent) diabetic pancreas recipients (n = 30), nondiabetic kidney recipients (n = 8), and nondiabetic control subjects (n = 28). Basal insulin levels were 66 +/- 5 pM in control subjects, 204 +/- 18 pM in pancreas recipients (P less than 0.0001 vs. control), and 77 +/- 17 pM in kidney recipients. Acute insulin responses to glucose were 416 +/- 44 pM in control subjects, 763 +/- 91 pM in pancreas recipients (P less than 0.01 vs. control), and 589 +/- 113 pM in kidney recipients (NS vs. control). Basal and stimulated insulin levels in two pancreas recipients with portal venous drainage were normal. Integrated acute C-peptide responses were not statistically different (25.3 +/- 4.3 nM/min in pancreas recipients, 34.2 +/- 5.5 nM/min in kidney recipients, and 23.7 +/- 2.1 nM/min in control subjects). Similar insulin and C-peptide results were obtained with arginine stimulation, and both basal and glucose-stimulated insulin-C-peptide ratios in pancreas recipients were significantly greater than in control subjects. We conclude that recipients of pancreas allografts with systemic venous drainage have elevated basal and stimulated insulin levels and that these alterations are primarily due to alterations of first-pass hepatic insulin clearance, although insulin resistance secondary to immunosuppressive therapy (including prednisone) probably plays a contributing role. To avoid hyperinsulinemia and its possible long-term adverse consequences, transplantation of pancreas allografts into sites with portal rather than systemic venous drainage should be considered.