The results of a prospective randomized controlled trial of elective endoscopic intravariceal sclerotherapy carried out over a 36-mo period in comparison with elective percutaneous transhepatic obliteration of varices (PTO) are presented. Sixty-six patients with nonalcoholic cirrhosis were randomized after they had stabilized, usually between 7 and 14 days after variceal bleeding had stopped following medical treatment (balloon tamponade and vasopressin infusion). Thirty-three patients were assigned to the sclerotherapy group and the other 33 patients were assigned to the PTO group. The mean follow-up period was similar in both groups. There was no significant difference in demographic, clinical, and laboratory data between the two groups. Six patients (18%) in the sclerotherapy group and 21 (64%) in the PTO group had at least one episode of gastrointestinal bleeding during the follow-up period (p less than 0.005). Three patients in the sclerotherapy group and 1 patient in the PTO group bled from lesions other than varices; therefore the incidence of variceal bleeding was 9% in the former and 61% in the latter (p less than 0.005). The cumulative variceal bleeding rate was significantly lower in the sclerotherapy group than the PTO group (p less than 0.05). Five patients in the sclerotherapy group died during the follow-up period but none died of recurrent variceal bleeding. Nineteen patients in the PTO group died and 10 of them died of bleeding from varices. The cumulative survival rate was significantly better in the sclerotherapy group (p less than 0.05). These results indicate that elective endoscopic intravariceal sclerotherapy is superior to elective PTO in the prevention of recurrent variceal hemorrhage and mortality in nonalcoholic cirrhosis.