The authors conducted a prospective, randomized, double-blind comparison of an epidural fentanyl infusion versus patient-controlled analgesia (PCA) with morphine in the management of postthoracotomy pain. Thirty-six patients were randomized into one of two groups. The epidural group received an epidural fentanyl infusion, 10 micrograms/mL, and saline through their PCA machine. The PCA group received an epidural saline infusion and morphine, 1.0 mg/mL, through their PCA device. The infusions were escalated according to a study protocol when pain relief was deemed inadequate by the patients. Pain relief was evaluated by a visual analog pain scale (VAS), both at rest and during coughing, and by verbal rating scores (VRS) of pain relief. Degree of sedation and the frequency of nausea, vomiting, and pruritus were also noted. The VAS, VRS, degree of sedation, and side effects were evaluated every 2 h from 7 AM to 7 PM, for 72 h after surgery. Forced vital capacities were determined before surgery and at 24, 48, and 72 h after surgery. The VAS were significantly lower (P = 0.001), and the Total Pain Relief scores higher (P < 0.02) in the epidural group, signifying better analgesia. There were no differences in postoperative forced vital capacity between the two groups. More patients in the PCA group had greater degrees of sedation on postoperative day 1 (P = 0.005), whereas pruritus was more frequent (P < 0.02) in the epidural group. We conclude that an epidural fentanyl infusion is superior to that of PCA with morphine in the management of pain after thoracotomy.