We prospectively evaluated gastric acid output (mEq/h), gastric volume output (ml/h), ambulatory 24-h esophageal pH monitoring, and the endoscopic appearance of the esophagus in 23 patients undergoing treatment of chronic long-standing pyrosis. Twelve of these 23 individuals (52%) remained symptomatic after 3 mo of standard antisecretory treatment with ranitidine, 150 mg twice daily. When compared with initial responders, those patients who did not experience complete symptomatic relief on therapy had significantly higher basal acid output (p less than 0.001), basal volume output (p less than 0.02), and basal upright (but not supine) reflux time (p less than 0.05). Nine of the 12 patients who did not respond to initial treatment had gastric acid hypersecretion (basal acid output greater than 10 mEq/h), and 10 of the 12 had Barrett's epithelium compared with only 1 patient in the initial-responder group (p less than 0.001). All 12 nonresponders were treated for an additional 3 mo with increased doses of ranitidine (mean, 1280 mg/day; range, 600-1800 mg/day), and complete disappearance of pyrosis occurred in 10 of the 12, although no significant endoscopic regression was observed in the extent of the underlying columnar mucosa in those with Barrett's esophagus over the 6-mo duration of the study. A significant correlation was shown between the daily ranitidine dose required to eliminate symptoms and the pretreatment basal acid output (r = 0.81, p less than 0.001); gastric acid output had to be almost totally suppressed (i.e., less than 1 mEq/h) for pyrosis to disappear completely. No side effects occurred with any of these high doses of ranitidine. We conclude that a subgroup of patients with long-standing symptomatic gastroesophageal reflux disease who do not respond to standard ulcer-healing doses of histamine2-receptor antagonists are hypersecretors of basal gastric acid and require increased acid-suppressive therapy. Many of these individuals also have underlying Barrett's epithelium.