The perioperative blood glucose regulatory response was compared in 20 healthy children (aged 1-5 yr) presenting for minor surgery and allocated randomly to either a fasted or a glucose group. All children received a milk feed at midnight. The fasted group received no oral intake thereafter, whereas the glucose group received 5% dextrose water 10 ml kg-1 orally about 4 h before operation. The mean plasma glucose concentrations in the two groups were similar before operation and were within normal limits. The pattern of change in the concentrations of plasma glucose, insulin, cortisol, growth hormone and glucagon were also similar between the two groups. Ten percent of patients in the fasted group and 33% in the glucose group had gastric aspirates in excess of 0.4 ml kg-1. The pH of all gastric samples was less than 2.5. The results suggest that healthy preschool children were able to maintain glucose homeostasis after 8 h of fasting. Feeding within 4-6 h before surgery may increase the risk of pulmonary aspiration.