HYPOTHESIS In good-risk patients with variceal bleeding undergoing portal decompression, surgical shunt is more effective, more durable, and less costly than angiographic shunt (transjugular intrahepatic portasystemic shunt [TIPS]). DESIGN Retrospective case-control study. SETTING Academic referral center for liver disease. PATIENTS Patients with Child-Pugh class A or B cirrhosis with at least 1 prior episode of bleeding from portal hypertension (gastroesophageal varices, portal hypertensive gastropathy). INTERVENTION Portal decompression by angiographic (TIPS) or surgical (portacaval, distal splenorenal) shunt. MAIN OUTCOME MEASURES Thirty-day and long-term mortality, postintervention diagnostic procedures (endoscopic, ultrasonographic, and angiographic studies), hospital readmissions, variceal rebleeding episodes, blood transfusions, shunt revisions, and hospital and professional charges. RESULTS Patients with Child-Pugh class A or B cirrhosis undergoing TIPS (n = 20) or surgical shunt (n = 20) were followed up for 385 and 456 patient-months, respectively. Thirty-day mortality was greater following TIPS compared with surgical shunt (20% vs 0%; P =.20); long-term mortality did not differ. Significantly more rebleeding episodes (P<.001); rehospitalizations (P<.05); diagnostic studies of all types (P<.001); shunt revisions (P<.001); and hospital (P<.005), professional (P<.05), and total (P<. 005) charges occurred following TIPS compared with surgical shunt. CONCLUSIONS Operative portal decompression is more effective, more durable, and less costly than TIPS in Child-Pugh class A and B cirrhotic patients with variceal bleeding. Good-risk patients with portal hypertensive bleeding should be referred for surgical shunt.