Fentanyl in doses of 50-60 microgram/kg has been reported to produce anesthesia with remarkable hemodynamic stability in patients with coronary artery disease (CAD). Because the authors had observed hypertension and tachycardia in response to noxious stimulation during aortocoronary bypass (ACB) operations in patients so anesthetized, they studied the hemodynamic changes and anesthetic conditions produced by fentanyl/O2/relaxant anesthesia in patients undergoing elective ACB. Twelve patients with left ventricular (LV) ejection fractions greater than 0.4 were maintained on propranolol until 10 hours before operation and were premedicated with fentanyl, diazepam, and scopolamine. Cannulae were inserted before the study commenced for measurement of intravascular pressures, arterial blood gases, and thermodilution cardiac output. The patients breathed 100 per cent oxygen throughout the study. Controlled ventilation aided by succinylcholine to reduce truncal rigidity maintained PaCO2 at 30-45 torr. Measurements were made after each of the following: breathing oxygen (control), 10 microgram/kg fentanyl, 50 microgram/kg fentanyl, and 0.1 mg/kg pancuronium, tracheal intubation, skin incision, and sternotomy. Fentanyl alone produced no significant hemodynamic changes. Fentanyl and pancuronium in combination produced increased heart rate and reduced stroke volume. Significant and progressively greater increases in mean arterial pressure and systemic vascular resistance followed intubation, skin incision, and sternotomy. Chest rigidity occurred in every patient at a lower fentanyl dose than did unresponsiveness. While fentanyl, 62.4 +/- 2.9 microgram/kg (SE), produced minor hemodynamic changes, it failed to block hemodynamic responses to noxious stimulation. Such changes resulted in increased cardiac work, and could have affected myocardial oxygen balance unfavorably. In eight of the 12 patients, following the last set of measurements, supplementary anesthetic agents were required to maintain hemodynamic stability during the surgical procedure. The authors suggest that this fentanyl/O2/relaxant technique should be modified for patients with severe CAD and reasonably good LV function.