Between July 1984 and March 1986, we conducted a prospective randomized trial comparing propranolol and endoscopic sclerotherapy in the prevention of recurrent variceal hemorrhage in a group of non selected alcoholic cirrhotics. Seventy-six patients with variceal hemorrhage were randomized to receive propranolol (P) (34 patients), or sclerotherapy (S) (42 patients) approximately 12 days after initial bleeding. The 2 groups were similar as concern age, sex, etiology of cirrhosis, severity of liver failure, the number of previous hemorrhages, and the severity of initial hemorrhage. No side effects were observed in the P group; 20 patients (48 percent) in the group S had at least one side effect although minor. After an average follow-up of 36 months, 18 patients in group P (53 percent) and 23 in group S (55 percent) had hemorrhagic recurrence. Rebleeding occurred from other sources than esophageal varices in 5 patients, in the group S only. Five patients in group P and 8 patients in group S died of rebleeding. During the follow-up period, 8 patients in group P (23 percent) and 13 patients in group S (31 percent) died. No significant difference could be demonstrated between the 2 groups as regards the percentages of patients without variceal rebleeding or survival, calculated according to the Kaplan Meier method. In conclusion, in this trial, no significant difference could be demonstrated between propranolol and endoscopic sclerosis in the prevention of recurrent variceal hemorrhage in alcoholic cirrhotic patients.