Background. Stable and reproducible large animal models of hepatic failure, which allow the assessment of liver-assist devices, are not available. Our objective was to develop a physiologically stable animal model of hepatic failure on which the safety and efficacy of an extracorporeal liver-assist device can be tested. We hypothesized that a surgical model which consists of an end-to-side portocaval shunt combined with common bile duct ligation and transection would create hepatic failure with: (1) elevations in amino transferases, total bilirubin, and ammonia; (2) a decrease in the ratio of branched chain to aromatic amino acids; and (3) histologic evidence of hepatic injury. Methods. Eleven mongrel dogs underwent common bile duct transection and an end-to-side portocaval shunt. Aminotransferases (AST, ALT), total bilirubin, ammonia, and branched chain and aromatic amino acids were measured prior to operation (baseline) and after 9 days. A necropsy was performed on Postoperative Day 9 and liver biopsies were obtained for histology. Results. By Postoperative Day 9, AST, ALT, total bilirubin, and ammonia values were significantly elevated compared to baseline (P < 0.02). The ratio of branched chain to aromatic amino acids was significantly reduced compared to baseline (P < 0.003). There was histologic evidence of cholestasis and inflammation. Conclusion. Portocaval shunt with common bile duct transection produces liver failure with elevations in aminotransferases, total bilirubin, and ammonia, a decreased branched chain to aromatic amino acid ratio, and histologic inflammation. Unlike ischemic or chemically induced models of liver failure, the dogs were hemodynamically and neurologically stable. This model can be used to test the safety and efficacy of liver-assist devices aimed at temporizing the detoxification functions of the failing liver. (C) 2000 Academic Press.
