Controversy still surrounds the place of portalsystemic shunting in the therapy of bleeding esophageal varices. Recently, a selective shunt, the distal splenorenal shunt, has achieved some degree of popularity and, apparently, is associated with less chronic encephalopathy. Because of this, a trial was initiated at the Massachusetts General Hospital and continued at the University of Cincinnati Medical Center, prospectively randomizing central and distal splenorenal shunts in consecutive elective cases of patients with established variceal bleeding. Preoperative evaluation included endoscopic examination at the time of hemorrhage, angiography and upper gastrointestinal series, emphasis on mental function including EEG, amino acids, neurologic examination, as well as standard liver chemistries. Nineteen patients underwent central splenorenal shunts and 23 distal splenorenal shunt. There was one operative death from hemorrhagic pancreatitis in a Child's Class A patient with distal splenorenal shunt. Four late deaths, from gunshot wound, auto accident, overwhelming pneumonitis similar to postsplenectomy syndrome, and metastatic carcinoma (2.5 years after operation), have been recorded in the distal splenorenal shunt group, and none in the central splenorenal shunt group. On follow-up angiographic examination, six shunts have clotted, with three patients requiring reoperation, generally mesocaval shunt. There has been no chronic encephalopathy, three individual episodes of encephalopathy, two in the central splenorenal shunt group and one in the distal splenorenal shunt group, two associated with gastrointestinal bleeding and one with intercurrent infection and overdiuresis. Follow-up liver chemistries and amino acids which may be useful as an indicator of hepatic function suggest that although the distal shunt group had a better amino acid pattern before operation, branched-chain amino acids tend to become lower in the distal group while remaining the same in the central group. Aromatic amino acids increase post shunt, equally in the two groups. The results do not support the contention that distal splenorenal shunt is associated either with greater survival or freedom from encephalopathy than central splenorenal shunt, a small side-to-side shunt. Ascites seems better controlled by the central splenorenal shunt.