The effect of flumazenil reversal of midazolam-induced anesthesia on whole body oxygen uptake (VO2) was investigated in a double-blind trial in 48 patients (ASA, 1 or 2) undergoing elective surgery under general anesthesia. VO2 was measured in spontaneously breathing patients during recovery from anaesthesia induced with midazolam 0.25 mg.kg-1 and maintained with nitrous oxide 60% in oxygen and halothane. The level of sedation was evaluated by a subjective score. To reverse midazolam-induced anesthesia, patients were randomly allocated to receive placebo or flumazenil (6 micrograms.kg-1). No significant changes in VO2 (160 +/- 53 vs 150 +/- 39 ml.min-1.m-2 or sedation score (2.5 +/- 1.0 vs 2.1 +/- 0.9) were observed in the placebo group. After flumazenil administration, the sedation score significantly (P less than 0.05) improved (2.9 +/- 1.0 vs 1.3 +/- 0.8) whereas no significant change in VO2 was observed (158 +/- 67 vs 157 +/- 61 ml O2.min-1.m-2). These data show that reversal of benzodiazepine effects with flumazenil resulted in no significant change in oxygen uptake.