Few anesthesia studies have explored perioperative continuous epidural infusion of neostigmine. We examined such a regimen in thoracotomy patients. Ninety patients were randomized to one of three groups in this double-blind trial. Before anesthesia induction, an epidural catheter was inserted in all patients at T5-8 levels under local anesthesia. Pre-neo patients received bolus 500-microg epidural neostigmine before anesthesia induction followed by infusion of 125 microg/h until the end of surgery. Post-neo patients received epidural saline during the same time periods plus bolus 500-microg epidural neostigmine at end of surgery. Patients in the control group received saline placebo during all three periods. Patients in the neostigmine groups postoperatively received patient-controlled epidural analgesia with morphine 0.02 mg/mL, bupivacaine 0.08 mg/mL, and neostigmine 7 microg/mL. Control patient-controlled epidural analgesia excluded neostigmine. Data were recorded for 6 postoperative days. Daily patient-controlled epidural analgesia consumption (mL) for Pre-neo patients was significantly less than that of post-neo and control group patients for postoperative days 1-6 (at least 10% and 16% less, respectively; P < 0.05). There was a modest decrease in pain intensity on postoperative days 3-6 for pre-neo patients versus other groups (P < 0.05). These results suggest that continuous thoracic epidural neostigmine started before anesthesia provided preemptive, preventive analgesia and an analgesic-sparing effect that improved postoperative analgesia for these patients without increasing the incidence of adverse effects.