We conducted a prospective randomized trial of propranolol for the prevention of recurrent variceal bleeding in 48 patients with cirrhosis of the liver. During a follow-up period of up to 21 months, 12 of 26 patients in the propranolol group and 11 of 22 in the control group had rebleeding from esophageal varices. There was no significant difference in rebleeding between the two groups. This contrasts with a previous report of the efficacy of propranolol in preventing recurrent gastrointestinal bleeding in alcoholic cirrhosis. The difference in results may be due to the inclusion in our study of patients with other causes of cirrhosis and more severe liver disease. Propranolol may not be indicated for the prophylaxis of variceal rebleeding in such patients, and we advocate that its use be limited at present to controlled clinical trials.