The present study was designed to characterize further the role of insulin in the regulation of glucose utilization during a 2-h exercise at 40% VO2max in 14 h fasted, healthy subjects. Endogenous insulin and glucagon were suppressed by somatostatin infusion and replaced singly or in combination to match the hormonal concentrations observed during similar exercise in saline-treated control subjects. Glucose kinetics were determined by a tracer method using D-[2,3,4,6,6-2H]glucose. In the exercising controls, during the last hour of the exercise, plasma glucose remained stable (4.26 +/- 0.06 mmol/L) and glucose utilization (Rd) increased significantly (p < 0.05) from 12.2 +/- 0.2 to 28.6 +/- 1.3 mumol.kg-1.min-1. During insulin deficiency without glucagon replacement, plasma glucose was maintained at 3.74 +/- 0.10 mmol/L by dextrose infusion, but with glucagon replacement plasma glucose increased to 6.69 +/- 0.24 mmol/L (p < 0.05). These hormonal changes were associated with an increase in Rd to 18.6 +/- 1.1 mumol.kg-1.min-1 (p = ns versus resting controls) and to 37.9 +/- 1.9 mumol.kg-1.min-1 (p < 0.05 versus resting controls), respectively. When insulin was replaced without glucagon replacement, plasma glucose was maintained at 3.85 +/- 0.06 mmol/L by dextrose infusion and Rd increased significantly (p < 0.05) from the resting value to 25.9 +/- 0.7 mumol.kg-1.min-1. When insulin was replaced together with glucagon, the plasma glucose (4.29 +/- 0.15 mmol/L) and the Rd (32.1 +/- 0.9 mumol.kg-1.min-1, p < 0.05 versus the resting value) obtained were similar to the values from the saline exercising control. Glucose metabolic clearance rate (MCR) significantly increased (p < 0.05) during exercise in all protocols. When insulin was made deficient, MCR increased 2-fold (p < 0.05) during exercise (2.7 to 4.8 and 5.4 mL.kg-1.min-1, respectively, with and without glucagon deficiency). However, when insulin was present, with and without glucagon deficiency, it increased further to 6.7 and 7.5 mL.kg-1.min-1, respectively, and values were different (p < 0.05) from glucose MCRs during insulin deficiencies. It is concluded that in 14 h fasted, healthy subjects, exercise per se can stimulate whole body glucose uptake even when insulin is made deficient. Insulin is necessary, however, for optimal glucose utilization during prolonged mild intensity exercise.