The two main causes of gastrointestinal bleeding in cirrhosis are oesophageal varices and portal hypertensive gastropathy (PHG). Rebleeding from varices can be prevented by beta-blockers, but it is not clear whether these drugs effectively reduce rebleeding from PHG. 54 cirrhotic patients with acute or chronic bleeding from severe PHG took part in a randomised, controlled trial to investigate the efficacy of propranolol in prevention of rebleeding from PHG. 26 patients were randomised to receive propranolol daily at a dose that reduced the resting heart rate by 25% or to 55 bpm (20-160 mg twice daily), throughout mean follow-up of 21 (SD 11) months. 28 untreated controls were followed-up, with the same examinations, for 18 (13) months. The actuarial percentages of patients free of rebleeding from PHG were significantly higher in the propranolol-treated patients than in the untreated controls at 12 months (65% vs 38%; p less than 0.05) and at 30 months of follow-up (52% vs 7%; p less than 0.05). Propranolol-treated patients had fewer episodes of acute bleeding than controls (0.010 [0.004] vs 0.120 [0.040] per patient per month). Multivariate analysis showed that absence of propranolol treatment was the only predictive variable for rebleeding. Actuarial survival was slightly higher in the propranolol group than in the controls, but the difference was not significant. Thus, long-term propranolol treatment significantly reduces the frequency of rebleeding from severe PHG, and may improve the prognosis of cirrhotic patients with this disorder.