Ninety-eight patients (26 females), who presented with erosive and/or ulcerative oesophagitis, despite at least a 3-month period of treatment with standard doses of cimetidine (greater than or equal to 1200 mg daily) or ranitidine (greater than or equal to 300 mg daily), were included in a double-blind, randomized trial to compare omeprazole (40 mg o.m.) with a high dose of ranitidine (300 mg b.d.). The treatment was given for 4-12 weeks; endoscopy assessment and laboratory screening were performed on entry to the trial and thereafter every fourth week. Endoscopic healing was defined as complete epithelialization of all macroscopic erosions or ulcers in the squamous epithelium. An 'intention-to-treat' analysis of the clinical data revealed omeprazole to be superior to ranitidine: 63% of those patients who were given omeprazole were healed endoscopically after a 4-week period of treatment, compared with only 17% of those given ranitidine. This difference in healing rate persisted during the 12-week study period (90% vs 47% after 12 weeks; P less than 0.0001). Reflux symptoms were more rapidly and completely relieved with omeprazole: heartburn resolved completely in 86% of patients treated with omeprazole for 4 weeks compared with 32% in the ranitidine group (P less than 0.0001). The mean basal gastrin concentrations increased only in those given omeprazole from 18.9 pmol/L at pre-entry to a mean value of 31.7 pmol/L on the last day of omeprazole administration. In ranitidine-treated patients no significant increase in basal gastrin concentration was observed. Both drugs were well tolerated with few adverse events, which were mainly mild and transient. These results demonstrate the superiority of omeprazole over a high dose of ranitidine in the treatment of resistant reflux oesophagitis.