Oxygen consumption in the immediate postoperative period was assessed in a series of twelve patients undergoing maxillofacial surgery, randomly assigned to two groups (n = 6 for each). The anaesthetic protocol was the same in all patients: premedication with flunitrazepam 30-mu-g.kg-1, induction with flunitrazepam 20-mu-g.kg-1, fentanyl 5-mu-g.kg-1, thiopentone 5 mg.kg-1, pancuronium bromide 80-mu-g.kg-1, and maintenance during the first three hours with flunitrazepam 10-mu-g.kg-1.h-1 and pancuronium 0.03-mu-g.kg-1.h-1. All the patients were given a continuous infusion of fentanyl during the whole length of the surgery: in Group I (GI), 3-mu-g.kg-1.h-1 and in Group II (GII), 1-mu-g.kg-1.h-1. The patients in Group II were also given enflurane 1.2 +/- 0.3 vol %. During the immediate postoperative period, the parameters studied were measured by means of an Engstrom Metabolism Calculator at the 15th, 30th, 45th, 60th, 120th and 240th minutes, and at the time of extubation. In GII, postoperative heart rate and lactic acid levels were higher than in GI (p < 0.05). Both groups had oxygen consumption values greater than theoretical resting levels. However, postoperative oxygen consumption was greatly increased in GII with respect to GI during the first 4 hours (p < 0.01). The respiratory quotient was greater in GI than in GII at the end of the study (p < 0.05). These data indicate that the intraoperative analgesic technique influences postoperative oxygen consumption. A dose of fentanyl of 3-mu-g.kg-1.h-1 is more efficient than one of 1-mu-g.kg-1.h-1 associated with enflurane.