It should be possible to avoid variations in plasma glucose concentration during anesthesia by adjusting glucose infusion rate to whole-body glucose uptake. To study this hypothesis, we measured glucose utilization and production, before and during halothane anesthesia. After an overnight fast, six adolescents between 12 and 17 yr of age were infused with tracer doses of [6,6-2H2]glucose for 2 h before undergoing anesthesia, and the infusion was continued after induction, until the beginning of surgery. Plasma glucose concentration was monitored throughout, and free fatty acids, lactate, insulin, and glucagon concentrations were measured before and during anesthesia. Despite the use of a glucose-free maintenance solution, plasma glucose concentration increased slightly but significantly 5 min after induction (5.3 +/- 0.4 vs. 4.5 +/- 0.4 mmol.l-1, P < 0.05). This early increase corresponded to a significant increase in endogenous glucose production over basal conditions (4.1 +/- 0.4 vs. 3.6 +/- 0.2 mg.kg-1.min-1, P < 0.05), with no concomitant change in peripheral glucose utilization. Fifteen minutes after induction, both glucose utilization and production rates decreased steadily and were 20% less than basal values by 35 min after induction (2.9 +/- 0.3 vs. 3.6 +/- 0.2 mg.kg-1.min-1, P < 0.05). Similarly, glucose metabolic clearance rate decreased by 25% after 35 min. Despite the increase in blood glucose concentration, anesthesia resulted in a significant decrease in plasma insulin concentration. These data suggest that halothane anesthesia per se affects glucose metabolism. The decrease in peripheral glucose utilization and metabolic clearance rates and the blunted insulin release question the relevance of glucose infusion in these clinical settings.