One of the major problems with total intravenous anaesthesia (TIVA) is postoperative sedation, possibly with respiratory depression. The aim of the present study was to evaluate the recovery characteristics after TIVA using a continuous infusion of a mixture of midazolam and alfentanil with flumazenil reversal before extubation. This method was compared to balanced anaesthesia using midazolam, alfentanil and nitrous oxide without flumazenil reversal. The degree of sedation was measured by reaction time test, Glasgow Coma Scale, cipher copying test and subtraction test. We found significantly faster reaction times postoperatively in the TIVA group (n = 15) compared to the balanced group (n = 13), despite larger doses of both midazolam (median 21 mg versus 9 mg) and alfentanil (median 5.9 mg versus 4.5 mg). The other tests revealed no difference between the groups. One patient became resedated after flumazenil. We conclude that the TIVA technique described here resulted in slightly better recovery characteristics, offering a usable alternative to balanced anaesthesia.