Between 1975 and 1983, 303 cirrhotic patients with endoscopically proven major variceal hemorrhage were admitted to the participating hospitals of the Boston-New Haven Collaborative Liver Group. Of these, 274 were evaluated for admission to a prospective, randomized controlled trial comparing portal-systemic shunts with distal splenorenal shunts. The criteria for inclusion were as follows: (i) a portohepatic pressure gradient ≥ 12 mmHg; (ii) angiographic evidence of antegrade portal venous flow; (iii) angiographic demonstration that the inferior vena cava and portal, splenic and left renal veins were anatomically suitable for either a portal-systemic or distal splenorenal shunt, and (iv) the patient was a reasonable operative risk. Eighty-one patients from the six participating hospitals fulfilled the criteria and consented to participate. Thirty-eight patients were randomly assigned to have portal-systemic shunt and 43 to have distal splenorenal shunt.

After a follow-up period of 11 years (mean = 3.5 years for all patients), survival was found to be similar in the two groups of patients. The 30-day operative mortality was 13% for the portal-systemic shunt group and 9% for the distal splenorenal shunt patients. Late mortality was 55% for the portal-systemic shunt and 37% for the distal splenorenal shunt group. Total mortality was 68% for the portal-systemic shunt and 46% for the distal splenorenal shunt group. None of these differences is statistically significant. In those patients who survived >30 days after surgery, recurrent variceal hemorrhage occurred in four (12%) in the portal-systemic shunt group compared to seven in the distal splenorenal shunt group (18%) (NS). Portal-systemic encephalopathy occurred in 15 (45%) of the operative survivors in the portal-systemic shunt group compared to 20 (51%) in the distal splenorenal shunt group. Severe portal-systemic encephalopathy was observed with equal frequency in the two groups. There was no difference in the time of onset nor the severity of encephalopathy in the two groups.

We conclude that the operative mortality and cumulative survival are similar for the two operations and that the frequency and severity of hepatic encephalopathy following distal splenorenal shunt are similar to that after portal-systemic shunt. Based on our observations and previously published data, recurrent hemorrhage from varices appears to be more frequently prevented by portacaval anastomosis than by distal splenorenal shunt.