The objective of this study was to assess the prognostic value of spontaneous portosystemic shunting and liver function for survival and spontaneous hepatic encephalopathy after end-to-side portacaval shunt in cirrhotic patients. One hundred ninety-eight patients with variceal hemorrhage as shown by endoscopy were evaluated. Forty-five were excluded because of uncontrollable hemorrhage; 84 were rejected because they were poor operative risk, had portal vein thrombosis or had been previously treated with β-blockers, sclerotherapy or surgery. The remaining 69 patients were enrolled in this prospective study. There were 43 patients with alcoholic cirrhosis, 23 with cryptogenic cirrhosis and three with primary biliary cirrhosis. The severity of liver disease was assessed according to the Pugh classification: 37 patients (54%) had Pugh's score 5 to 7, 26 (38%) had 8 to 10 and six (8%) had 11 to 12. Indocyanine green intrinsic clearance was used as a probe of preoperative liver function and lidocaine systemic availability as an index of spontaneous preoperative shunting. All the patients underwent an elective end-to-side portacaval shunt. The length of minimal follow-up was 40 months. One-year survival was 76% and 5-year survival was 46%. During follow-up, 25 patients died from their liver disease and 11 patients died from various causes unrelated to their liver disease. Spontaneous chronic encephalopathy occurred in 16 patients (23%). Age, Pugh's score, active alcoholism, indocyanine green intrinsic clearance and lidocaine systemic availability were tested as prognostic factors in a multivariate analysis. A stepwise analysis using the Cox proportional hazard demonstrated that, in this population of cirrhotic patients, the Pugh score was the only variable significantly related to survival (β = 0.3; S.E. = 0.12, p < 0.05). A stepwise multiple logistic regression analysis showed that none of the studied parameters was significantly related to the onset of postoperative spontaneous encephalopathy. These findings suggest that in this population of cirrhotic patients the Pugh score remains the best predictive variable for survival and that pharmacokinetic parameters have no role in the prediction of postoperative outcome after end-to-side portacaval shunt.