We determined the early postoperative analgesia using intraoperative continuous epidural infusion of lidocaine during general anesthesia in patients undergoing upper abdominal surgery in a prospective double-blind manner. After insertion of an epidural catheter at the T10-T11 interspace, general anesthesia was induced. Thirty patients were randomly allocated to receive continuous epidural infusion of either 0.5% (n = 15) or 1% (n = 15) plain lidocaine at 10 ml/hr. The infusion was continued from 10 to 15 minutes before surgery until the end of surgery. Visual analog pain scale (0-10) within 30 minutes after the end of surgery was significantly lower in the 1% lidocaine group (5.6 +/- 0.9, mean +/-SE) than in the 0.5% lidocaine group (8.2 +/- 0.8), however, it was unsatisfactory in both groups. Plasma concentrations of lidocaine and its principal metabolite, monoethylglycinexylidide, gradually increased through epidural infusion, but remained below the toxic range in both groups. We conclude that continuous epidural lidocaine during general anesthesia offered limited analgesia in the early postoperative period.