In this study, we compared postoperative liver function in patients with liver cirrhosis between isoflurane and sevoflurane anesthesia with nitrous oxide (N(2)O) and epidural block. Forty cirrhotic patients with Child-Pugh Grade A, aged 40 to 70 yr, scheduled for liver segmentectomy, had anesthesia induced with midazolam 0.1 mg/kg and fentanyl 4 micro g/kg. For maintenance, intermittent epidural administration of 1.5% lidocaine 4 to 6 mL and sevoflurane (sevoflurane group) or isoflurane (isoflurane group) with N(2)O 3 L/min in oxygen 3 L/min was used. Aspartate aminotransferase, alanine aminotransferase, total bilirubin, alkaline phosphatase, choline esterase, albumin, prothrombin time, and platelet count were measured before and 1, 3, and 7 days after surgery. Aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase increased significantly, with the peaks at 3 days after surgery in both groups. The increases in these variables were significantly larger in the isoflurane group than those in the sevoflurane group. No patient developed hepatic failure. All increases in liver enzymes were small and of questionable clinical relevance. Whether sevoflurane might be a better anesthetic when combined with N(2)O and epidural block for cirrhotic patients than isoflurane with respect to liver damage remains to be determined. CONCLUSIONS In cirrhotic patients with Child-Pugh Grade A, isoflurane induced more of an increase in serum concentrations of liver enzymes after surgery than sevoflurane when combined with nitrous oxide and epidural block. However, the increases were small, and there was no clinical liver damage.