We measured venous blood concentrations of lidocaine in 15 patients undergoing lower abdominal or extremity surgery after epidural injection of lidocaine. Patients were divided into either an infusion group (Group I, n = 6) or a bolus injection group (Group II, n = 9). We administered 15 ml of 2% lidocaine into the epidural space at 1 mL/min in Group I and at 1 mL/3 in Group II. Venous blood was drawn at 3, 6, 9, 12, 15, 18, 21, 30, 60, and 90 min after injection for measurement of plasma lidocaine concentrations. The peak plasma concentration (Cmax) of lidocaine in Group I was significantly less than that in Group II (P < 0.0005). We conclude that slow epidural infusion can produce a lower Cmax of lidocaine compared with that of a bolus administration and thereby decrease the potential for systemic toxicity of the local anesthetic.