Epidural infusions of fentanyl, in a 10 micrograms.ml-1 concentration, combined with bupivacaine 0.1% were compared with epidural infusions of fentanyl alone for postoperative analgesia following abdominal or thoracic surgery. There were no detectable differences between the two groups in analgesia (mean visual analogue scale pain scores ranging between 15-35 mm), average infusion rates of 7-9 ml.hr-1, and serum fentanyl concentrations which reached 1-2 ng.ml-1. There was no difference in postoperative pulmonary function (pH, PaCO2, SaO2), or bowel function (time to flatus or po fluids). The incidence of side-effects including somnolence, nausea and vomiting, pruritus and postural hypotension was also similar. Of the patients receiving fentanyl and bupivacaine 0.1%, three developed a transient unilateral sensory loss to pinprick and ice, and two of these patients had unilateral leg weakness equal to a Bromage 1 score. The addition of bupivacaine 0.1% does not improve epidural infusions of fentanyl using a 10 micrograms.ml-1 concentration following abdominal or thoracic surgery.