Background: Hemodynamic studies have shown the efficacy of nitrates in reducing portal pressure in cirrhosis. We therefore studied the efficacy of isosorbide-5-mononitrate vs. propranolol in the prevention of first bleeding within a prospective controlled trial. Methods: One hundred eighteen cirrhotics with esophageal varices were blindly randomized to receive 20 mg of isosorbide-5-mononitrate three times a day (n = 57) or propranolol (n = 61) up to the maximum tolerated dose. Both groups also received ranitidine (150 mg/ day). Results: The median follow-up was 29 months. Twenty-six patients dropped out (13 in the isosorbide group) because of poor compliance or complications unrelated to treatment. Eighteen patients died (9 in the isosorbide-treated group), 6 due to bleeding. The 1-and 2-year actuarial percentages of patients free of bleeding were 90.8% and 82.2% in the isosorbide-5-mononitrate-and 93.9% and 85.8% in the propranolol-treated groups, respectively (P = NS). These values are higher than those expected from the North Italian Endoscopic Club predicting scores. There were few major side effects in either group. The 2-year survival rate did not differ between the two groups (82.2% vs. 85.4%). Conclusions: lsosorbide-5-mononitrate administered orally is a safe and effective alternative to propranolol in the prophylaxis of bleeding in cirrhosis.