To test whether pancreatic hormonal changes that occur during exercise are necessary for the postexercise enhancement of insulin-stimulated net hepatic glucose uptake, chronically catheterized dogs were exercised on a treadmill or rested for 150 min. At the onset of exercise, somatostatin was infused into a peripheral vein, and insulin and glucagon were infused in the portal vein to maintain basal levels (EX-Basal) or simulate the response to exercise (EX-Sim). Glucose was infused as needed to maintain euglycemia during exercise. After exercise or rest, somatostatin infusion was continued in exercised dogs and initiated in dogs that had remained sedentary. In addition, basal glucagon, glucose, and insulin were infused in the portal vein for 150 min to create a hyperinsulinemic-hyperglycemic clamp in EX-Basal, EX-Sim, and sedentary dogs. Steady-state measurements were made during the final 50 min of the clamp. During exercise, net hepatic glucose output (mg x kg(-1) x min(-1)) rose in EX-Sim (7.6 +/- 2.8) but not EX-Basal (1.9 +/- 0.3) dogs. During the hyperinsulinemic-hyperglycemic clamp that followed either exercise or rest, net hepatic glucose uptake (mg x kg(-1) x min(-1)) was elevated in both EX-Basal (4.0 +/- 0.7) and EX-Sim (4.6 +/- 0.5) dogs compared with sedentary dogs (2.0 +/- 0.3). Despite this elevation in net hepatic glucose uptake after exercise, glucose incorporation into hepatic glycogen, determined using [3-3H]glucose, was not different in EX-Basal and sedentary dogs, but was 50 +/- 30% greater in EX-Sim dogs. Exercise-induced changes in insulin and glucagon, and consequent glycogen depletion, are not required for the increase in net hepatic glucose uptake after exercise but result in a greater fraction of the glucose consumed by the liver being directed to glycogen.