In spite of several advantages, the need for postoperative ventilatory support limits the use of high-dose opioid anesthesia. We prospectively evaluated the effectiveness of naloxone infusion for the reversal of high-dose fentanyl anesthesia. Anesthesia was maintained with fentanyl in patients undergoing major abdominal surgery. After anesthesia, the trachea was extubated when intravenous naloxone, which was titrated in separate 50- micro g doses, established an acceptable level of consciousness and arterial blood gas (ABG) status under spontaneous respiration; this was followed by continuous infusion started at the rate of the sum of the bolus doses per hour. The naloxone infusion was terminated based on evaluation of the level of consciousness, ABG, and acute abstinence symptoms. Postoperative pain was evaluated using self-reported four-step categorical terms (none, mild, moderate, and severe). Plasma concentrations of fentanyl and naloxone were analyzed in 12 patients, using high-performance liquid chromatography. Fifty-seven out of 59 eligible patients were successfully extubated at 34 +/- 14 min after termination of fentanyl (total dose, 127 +/- 64 micro g.kg(-1); mean +/- SD) with naloxone (total bolus, 3.4 +/- 2.6 micro g.kg(-1)). All these patients recovered fully without ventilatory support under the naloxone infusion, which was terminated at 11 +/- 7 h. The reduction of the naloxone infusion rate effectively relieved the increased pain, and no supplemental analgesic was used in any patients during the naloxone infusion. Pharmacokinetic analysis did not indicate any correlations between plasma fentanyl and naloxone concentrations. The results suggest that naloxone infusion with individual dose titration facilitates the use of high-dose opioid anesthesia, maintaining the advantager of this anesthesia.