Long-term esophageal pH-metry has become the preferred test to quantify acid gastroesophageal reflux, but its accuracy in separating physiologic from pathological reflux is not well defined. To establish optimal thresholds of 24-h pH-metry, we studied 45 patients with clinically proven gastroesophageal reflux disease and 42 healthy volunteers. Twenty-four-hour esophageal pH was measured while the subject was ambulatory, using a combined glass electrode connected to a portable recorder. Percentage of time with esophageal pH less than 4, the number and mean duration of reflux episodes, the number of episodes lasting longer than 5 min, and the duration of the longest episode were calculated for periods of upright and supine body position, respectively. Discriminant analysis and receiver-operating-characteristic analysis were used to define optimal thresholds. A maximum of sensitivity (93.3%) and specificity (92.9%) was obtained using receiver-operating-characteristic analysis with the following criteria. Only percentage of time with esophageal pH less than 4 is considered. A subject is classified as "normal" if both values for the upright and supine body position are below the thresholds, otherwise the subject is classified as "pathological." The thresholds are 10.5% of time with esophageal pH less than 4 for the upright position and 6.0% for the supine position. Within the limits of this retrospective study, it is concluded that rather simple criteria yield a high diagnostic accuracy in the evaluation of long-term esophageal pH-metry. Nevertheless, a prospective and independent confirmation of these criteria is needed.