The acid perfusion (Bernstein) test and esophageal pH monitoring are the two most popular tests for identifying esophageal acid sensitivity in difficult cases of reflux disease. Therefore, we prospectively compared these test results in 75 consecutive noncardiac chest pain patients who had both an acid perfusion test and chest pain during 24-hr pH testing. A positive acid perfusion test was defined by the replication of the patient's typical chest pain twice by the acid infusion. Esophageal pH testing identified abnormal amounts of acid reflux and correlated symptoms with acid reflux--the "symptom index." Fifteen patients (20%) had a positive acid perfusion test while 45 patients (59%) had a positive symptom index (range 6-100%). Only 9/34 (26%) patients with abnormal reflux had a positive acid perfusion test. Although it had excellent specificity (83-94%), the acid perfusion test had poor sensitivity (32-46%) when compared to the symptom index regardless of the percent positive cutoff level. The best positive predictive value for the acid perfusion test was 87%, but this occurred when the test sensitivity was 32%. Modifying the end point of a positive acid perfusion test to include heartburn improves the sensitivity (52-67%) while markedly compromising specificity and positive predictive value. Thus, esophageal pH monitoring correlating symptoms with acid reflux is superior to the acid perfusion test for identifying an acid sensitive esophagus in patients with noncardiac chest pain.