Strong inhibition of acid secretion could be able to decrease gastric and duodenal ulcer early rebleeding. OBJECTIVE--The aim of this double blind randomized trial was to compare early rebleeding rates of 2 groups of patients treated with ranitidine (600 mg/day) or lansoprazole (60 mg/day) per os for 6 consecutive days. METHODS--Seventy five patients with a high risk of rebleeding (clinical and endoscopical criteria) were included in this trial. These ulcers were Ia (n = 10), Ib (n = 20), IIa (n = 13), IIb (n = 32) in Forrest classification. RESULTS--Nineteen out of 75 patients rebled (25.3%): 11 out of 37 (30%) and 8 out of 38 (21%) in the ranitidine and lansoprazole groups respectively. Rates of rebleeding were 10%, 12.5%, 36% and 29% respectively in the ulcers grade Ia (previously treated with endoscopic sclerosis), Ib, IIa and IIb in the Forest classification. CONCLUSION--The rates of rebleeding were not statistically different in the 2 groups of treatment. The high rebleeding rates observed with Forrest IIa and IIb and duodenal ulcers support the need of haemostatic endoscopic therapy associated to antisecretory treatment in such patients.