In the 1990s, liver transplantations and transjugular intrahepatic portosystemic shunts (TIPS) have become the most common methods to decompress portal hypertension. This center has continued to use surgical shunts for variceal bleeding in good-risk patients who continue to bleed through endoscopic and pharmacologic treatment. This article reports this center's experience with surgical shunts and TIPS shunts from 1992 through 1999. Sixty-three patients (Child A, 43 patients; Child B, 20 patients) received surgical shunts: distal splenorenal, 54 patients; splenocaval, 4 patients; coronary caval, 1 patient; and mesocaval, 4 patients. Sixty-two patients had refractory variceal bleeding, and 1 patient had ascites with Budd-Chiari syndrome. Two hundred patients (Child A, 24 patients; Child B, 62 patients; Child C, 114 patients) received TIPS shunts. One hundred forty-nine patients had refractory variceal bleeding, and 51 patients had ascites, hydrothorax, or hepatorenal syndrome. Data were collected by prospective databases, protocol follow-up, and phone contact. The 30-day mortality rate was 0% for surgical shunts and 26% for TIPS shunts; the overall survival rate was 86% (median follow-up, 36 months) for surgical shunts and 53% (median follow-up, 40 months) for TIPS shunts. For surgical shunts, the portal hypertensive rebleeding rate was 6.3%; the overall rebleeding rate was 14.3%. For TIPS shunts, the overall rebleeding rate was 25.5% (30-day, 9.4%; late, 22.4%). There were 4 reinterventions for surgical shunts (6.3%); the reintervention rate for TIPS shunts in the bleeding group was 33%, and the reintervention rate in the ascites group was 9.5%. Encephalopathy was severe in 3.1% of the shunt group and mild in 17.5%; this was not systematically evaluated in the TIPS shunts patients. Surgical shunts still have a role for patients whose condition was classified as Child A and B with refractory bleeding, who achieve excellent outcomes with low morbidity and mortality rates. TIPS shunts have been used in high-risk patients with significant early and late mortality rates and have been useful in the control of refractory bleeding and as a bridge to transplantation. The comparative role of TIPS shunts versus surgical shunt in patients whose condition was classified as Child A and B is under study in a randomized controlled trial.