Forty adult patients undergoing elective surgery, anaesthetized with diazepam, alfentanil and nitrous oxide in oxygen, and paralysed with atracurium were given flumazenil or placebo i.v. in a double-blind randomized study to assess the efficacy of flumazenil. Awake state, heart rate, arterial pressure, rate of ventilation and arterial blood-gas values were measured at 0, 5, 30, 120 and 240 min after administration of flumazenil or placebo. Flumazenil was found to antagonize the sedative effects of diazepam; there was no resedation within the time of observation. There were no intergroup differences in any of the other measured variables. A median dose of diazepam 0.33 mg kg-1 during surgery lasting a median of 98 min was antagonized by a median dose of flumazenil 0.35 mg. No side effects related to flumazenil were observed.