Twenty ASA 1 pregnant women at term, undergoing elective Caesarean section were included in this study. They were randomly assigned to one of two groups, receiving either a spinal or an epidural anaesthesia. Before induction, in order to prevent hypotension, all patients were given an i.v. infusion of 1000 ml of Ringer-lactate and a subcutaneous injection of ephedrine 30 mg. They were positioned on the operating table with a 15 degrees left lateral tilt. Spinal anaesthesia was performed with hyperbaric bupivacaine 0.5 p. cent (0.08 mg.cm-1 of height). Epidural anaesthesia was obtained with a bolus dose of 0.5 p. cent plain bupivacaine, followed by a continuous infusion through the epidural catheter until the level of surgical block reached T6 bilaterally. Bupivacaine was assayed in plasma by high performance liquid chromatography (HPLC). Following pharmacokinetic parameters of bupivacaine were determined: Cmax (maximal concentration), Tmax (time to reach maximum), AUC (area under curve), Cl (total plasma clearance), Vz (volume of distribution during the elimination phase), T1/2 (elimination half-life). Bupivacaine concentration was also measured in samples obtained at birth from umbilical vein and umbilical artery. The mean dose of bupivacaine used was 12.8 +/- 0.6 mg in the spinal group and 118.6 +/- 17.8 mg in the epidural group. The time of onset of surgical anaesthesia was significantly shorter with spinal anaesthesia (7.6 +/- 4.4 vs 31 +/- 11.1 min; p < 0.01). The sensory block had a longer duration in epidural group (223.2 +/- 15 vs 291 +/- 13.8; p < 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)