Secretion of gastric acid and volume, serum gastrin concentration, and ambulatory 24-hr esophageal pH monitoring were evaluated prospectively in 12 patients with idiopathic gastric acid hypersecretion (basal acid output greater than 10.0 meq/hr) undergoing treatment for refractory chronic long-standing pyrosis. Treatment lasted six months and consisted of three months of ranitidine (mean 2150 mg/day, range 1200-3000 mg/day), followed by three months of omeprazole (mean 33 mg/day, range 20-60 mg/day). Both ranitidine and omeprazole significantly reduced gastric acid output (P < 0.001) and gastric volume output (P < 0.001) compared to a basal evaluation and resulted in complete disappearance of pyrosis. Total reflux time (percent 24 hr intraesophageal pH less than 4) was significantly reduced by ranitidine (P < 0.02) and omeprazole (P < 0.001) compared to basal evaluation; however, the effects of omeprazole were significantly greater than ranitidine (P < 0.05). Omeprazole caused a significant increase in serum gastrin concentration compared to both basal and ranitidine (P < 0.05). Endoscopically documented erosive esophagitis was present in nine of the 12 patients, and seven of the 12 patients had Barrett's epithelium. All 12 patients had complete resolution of pyrosis and healed esophagitis by six months, but no significant endoscopic regression was observed in the extent of Barrett's epithelium. No side effects occurred with these high doses of ranitidine or omeprazole. These results indicate that high-dose ranitidine and omeprazole are effective therapy for refractory gastroesophageal reflux disease. However, with omeprazole, total reflux times are reduced more than with ranitidine, often into the normal range. That marked reduction of gastric acid secretion with omeprazole, which greatly reduces total reflux times, accounts for the significant elevation of serum gastrin concentration seen during omeprazole therapy.