We have measured blood concentrations of glucose, 3-hydroxybutyrate (3-OHBA), non-esterified fatty acid (NEFA) and lactate before and after glucose administration in 105 patients to determine the optimal glucose dose during the period of preoperative fasting state. The patients, scheduled for laparotomy in the afternoon, received fluid infusion containing an arbitrary glucose dose for 4 hours prior to surgery. The glucose dose showed a negative correlation with the ratio of 3-OHBA or NEFA prior to and after glucose administration (r = 0.40, r = 0.41 respectively, P < 0.05). There was no correlation between lactate ratio and glucose dose (r = 0.06). Glucose administration more than 0.4 g.kg-1 significantly suppressed 3-OHBA production (P < 0.05). Hyperglycemia (> 200 mg.dl-1) developed more often in the groups given 0.5 g.kg-1 or more glucose (P < 0.05), while significant hypoglycemia (< 60 mg.dl-1) occurred in the groups given less than 0.2 g.kg-1 of glucose (P < 0.05). This study demonstrates that preoperative infusion of glucose 0.4-0.5 g.kg-1 is useful to maintain the energy metabolism during the fasting state in patients scheduled for operation in the afternoon.