The authors anesthetized 18 patients with good pulmonary and ventricular function for coronary artery bypass grafting with high doses of fentanyl. When the patients were arousable and their vital signs stable in the intensive care unit, the authors administered nalbuphine or placebo (randomly and double-blinded) until extubation criteria were met, and subsequently gave nalbuphine for analgesia. In one of ten placebo patients, tracheal extubation was accomplished without nalbuphine. This patient then retained CO2 and required nalbuphine; the other nine placebo patients could not be extubated after placebo trials and were given nalbuphine. In all other patients in both groups, tracheal extubation was successful following nalbuphine (median dose 60 micrograms/kg, range 30-180 micrograms/kg). One patient became renarcotized 4 h after tracheal extubation without an increase in plasma fentanyl concentration; he received an additional dose of nalbuphine and recovered without further incident. Nine patients required treatment with vasoactive agents or beta-blockers for hypertension or tachycardia associated with the administration of nalbuphine. Eight of 18 patients were not satisfied with nalbuphine analgesia, and required morphine for relief of their pain. Recurrent elevations of fentanyl concentrations in plasma were observed and appeared to be related to increasing motor activity. Nalbuphine is an effective opioid antagonist after fentanyl anesthesia, but its use is associated with side effects, and analgesia for the post-sternotomy patient may be unsatisfactory unless the dose is carefully titrated to the minimum required to antagonize respiratory depression.