In a randomised clinical study, two total intravenous anaesthesia techniques for microlaryngoscopic laser surgery were compared. After an induction dose of 100 mg methohexital, Group I received a maintenance infusion of 10 mg.min-1. In Group II anaesthesia was obtained by 15 mg midazolam followed by 0.1 mg.min-1 continuously and terminated by the injection of flumazenil. For analgesia 5 mg alfentanil were administered. Opiate-induced respiratory depression was antagonised by 0.08 to 0.12 mg naloxone. Prior to, during, and after surgery, adrenergic response was assessed by HPLC-analysis of blood taken from a peripheral vein. Haemodynamic responses to the operation and during the post-operative period were almost identical in both groups. In Group I, the mean recovery period of 14 min was significantly longer than in Group II (9 min), where patients received a mean dose of 0.53 mg (+/- 0.15) flumazenil. Resedation could be observed in all patients receiving flumazenil within 60 min after antagonisation, which was associated with a mean decrease in O2-saturation from 95% to 88%. There was no difference in epinephrine and norepinephrine blood levels between the two groups prior to and during anaesthesia. In all patients, arousal was associated with a significant increase in the epinephrine plasma concentration. While blood levels in Group I decreased during the post-operative period to levels prior to surgery, the concentrations in Group II remained elevated. In one patient who received no naloxone, the reversal of midazolam action induced a 16-fold increase in catecholamine levels (from 50 to 800 ng.l-1) associated with a tachycardia of 170 b.min-1 and hypertension of 160 mmHg.(ABSTRACT TRUNCATED AT 250 WORDS)