THE ROLE of a portacaval shunt in the prevention of hemorrhage from esophagogastric varices and in the prolongation of life of patients with cirrhosis is unknown. Although two prospective studies1-4 have reported early survival analysis of this operation as a prophylaxis against hemorrhage and death from bleeding varices, there has been no controlled attempt to determine its effect upon the full spectrum of hepatic disease and the management of gastrointestinal hemorrhage. Since Jan 1, 1962, investigators at 12 Veterans Administration hospitals have screened, classified, and randomized for therapy, patients admitted to their institutions and qualifying for the study under a uniform protocol. No patient was randomized until he was considered a satisfactory surgical risk and had agreed to accept an operation as a part of his treatment program. All patients were followed at regular intervals until death or lost to follow-up.