Ten per cent of patients with angina pectoris have normal coronary arteries and cardiac function and, despite this reassurance, continue to have chest pain. Since pain of cardiac or esophageal origin is clinically difficult to differentiate, 50 patients with severe chest pain, normal cardiac function, and normal coronary arteriography with ergotamine provocation were evaluated with a symptomatic questionnaire and esophageal function test. On 24-hour esophageal pH monitoring, 23 patients had abnormal reflux, and 27 were normal. There was no difference in the incidence and severity of chest pain, esophageal symptoms, or medication taken between refluxers and nonrefluxers. Ten refluxers and ten nonrefluxers had chest pain on exercise electrocardiography. Thirteen refluxers documented chest pain during the pH monitoring period, and in 12 it coincided with a reflux episode. Fifteen nonrefluxers documented chest pain during the monitoring period, and in only one did it coincide with a reflux episode. Of the 23 refluxers, 12 were treated with medical therapy and 11 by a surgical antireflux procedure, and all followed for two to three years. Ten (91%) of the 11 surgically treated patients are totally free of chest pain compared with five (42%) of the 12 medically treated patients. All 12 patients who had chest pain coincide with a documented reflux episode responded positively to antireflux therapy, eight surgical and four medical. It is concluded that 46% of patients complaining of angina pectoris with normal cardiac function and coronary arteriography have gastroesophageal reflux as a possible etiology. Seventy-three per cent of these patients have total abolition of chest pain by either surgical or medical antireflux therapy. Patients whose experience of chest pain coincided with a documented reflux episode on 24-hour esophageal pH monitoring had a 100% response to medical or surgical therapy. Overall, surgical therapy gave better results (91%) but was associated with an 18% temporary morbidity. Objective evaluation of reflux status and its correlation to the symptom of chest pain by 24-hour pH monitoring allows for selective therapy in these difficult to manage patients.