Fifty patients undergoing endoscopy (laryngoscopy, bronchoscopy, mediastinoscopy) were anaesthetised in a double-blind prospective trial using total intravenous anaesthesia. Half of the patients were anaesthetised with an infusion of a solution of 250 mg ketamine and 12.5 mg midazolam. The other patients received an infusion of a solution of 250 mg ketamine and 20 mg diazepam. In addition, both groups were given increments of 50-100 micrograms of fentanyl. The immediate awakening time (t1) was not significantly different between groups, but the patients who had received midazolam-ketamine, had a significantly shorter time to more complete recovery (t2), a significantly lower frequency of emergence reactions and were more satisfied with the anaesthetic than the patients who had received diazepam-ketamine. There was no difference between groups with respect to intraoperative heart rate and blood pressure. No awareness during anaesthesia was reported.