A double-blind crossover study was conducted of two gastric prokinetic drugs in 23 patients with gastroesophageal reflux. Patients were divided into two groups on the basis of a dual-isotope mixed-meal study of their gastric emptying (GE). Group I had normal GE and group II delayed GE. Nine gastrointestinal symptoms were assessed for frequency and severity before treatment. The trial had three 1-month treatment periods using metoclopramide 10 mg q.i.d., domperidone 20 mg q.i.d., or placebo on a random basis. Symptoms were reassessed at the end of each month. Taken as a whole, the group showed a significant symptomatic response in all three treatment periods (p less than 0.0001), but patients with delayed or normal GE did not differ significantly in their symptomatic response. Eleven patients complained of side effects with metoclopramide and three stopped therapy before the 1-month course was completed. Two patients described side effects with domperidone, including one woman with galactorrhea after 36 h of treatment. Three patients on placebo also complained of important side effects. We conclude that a significant placebo effect is present in the treatment of gastroesophageal reflux. No significant difference was demonstrated in symptomatic improvement between placebo, domperidone, and metoclopramide in this study.