In 108 patients the healing and relapse of reflux esophagitis, defined endoscopically by the presence of epithelial defects (erosions and ulcerations) of the esophageal mucosa, were studied. In the first study, with open treatment of ranitidine, the healing rate after 6 wk was 50%. The most important factor that negatively influenced healing was the extent of esophageal erosions. Patients with isolated erosions had a 6-wk healing rate of 78%; the healing rate was 38% in patients with longitudinally confluent lesions and 23% in those with circumferential erosions of the distal esophagus. Smoking also had an unfavorable effect. Age, sex, duration of history, body weight, and alcohol consumption were not related to outcome. Symptoms improved during treatment with ranitidine, but the correlation between symptoms and endoscopic findings at 6 wk was weak. In the second study, relapse was investigated in 61 patients with healed esophagitis in a randomized, double-blind trial comparing placebo and ranitidine (150 mg at bedtime for 6 mo). In both groups, relapse occurred in more than one-third of the patients, with no significant difference between ranitidine and placebo treatment. Patients with worse daytime symptoms at the time of previous healing had a higher relapse rate. The initial severity of esophagitis and smoking did not influence recurrence. Thus, the initial endoscopic findings are of prognostic value in reflux esophagitis. Smoking retards healing. Low-dose maintenance treatment with ranitidine does not prevent relapse.