In a double-blind, randomized study, the clinical effects of 5 mg and 10 mg of cisapride three times daily were compared with those of 10 mg of metoclopramide three times daily in 114 patients with symptoms of gastroesophageal reflux, mainly diurnal and nocturnal heartburn and regurgitation. The symptoms significantly (P less than 0.001) improved in the three groups; the mean severity score decreased by at least 78% after four weeks of treatment. Initial symptoms were more severe in the cisapride-treated patients, especially in those receiving 10 mg three times daily; however, the patients' condition after four weeks was similar in the three groups. Central nervous system side effects were reported by one patient from each of the cisapride-treated groups and by nine of the 43 metoclopramide-treated patients (P less than 0.02). Six metoclopramide-treated patients and one cisapride-treated patient dropped out of the study because of side effects. These findings favor the use of cisapride when prokinetic treatment of gastroesophageal reflux is considered.