The goal of this study was to assess the degree of postoperative anxiety flumazenil might provide in patients when used to reverse the sedation induced by midazolam. Twenty-four patients, aged 18 to 60 yrs, and scheduled for elective orthopaedic surgery of the upper limb, were included in the study Regional anaesthesia (brachial plexus block) was carried out 20 min after the patient had been premedicated with 0.15 mg.kg-1 intramuscular midazolam. When the block was clinically complete, 0.12 mg.kg-1 midazolam was given intravenously. Sedation was maintained throughout surgery using 0.03 mg.kg-1 midazolam every quarter of an hour (mean total dose 0.206 mg.kg-1). At the end of the procedure, the patients were randomly allocated to one of two groups: A+, the patients were given 0.1 mg intravenous flumazenil repeated every min until the patient was fully awake (mean 4.5 +/- 2.6 micrograms.kg-1); and A-, the patients were allowed to recover spontaneously. A wakefulness, and a determination just before premedication (E1, P1, H1 respectively). The degree of recovery was assessed in both groups, and, when the value of the wakefulness test had returned to that of E1 (E2) (E2 = E1), P2 and H2 were carried out. There were no statistically significant differences between P1 and P2, and between H1 and H2, and this for either group of patients. So, the doses of flumazenil used, which awoke all the patients within 3 min, did not create any anxiety. In accordance with most of the recently published studies, flumazenil can be safely used to reverse benzodiazepine induced sedation.