Abstract

A series of tests, including gastro-oesophageal sphincter pressure measurement, short-term pH tests, and 15-hour overnight oesophageal pH recording were applied to 42 normal subjects and 214 patients with typical reflux symptoms. The results were compared by multivariate discriminant analysis. Sphincter pressure measurements misclassified 32%, stressed provocative manoeuvres 14-5%, and the best single discriminator from the overnight pH study was time below pH 5, which misclassified 13%. However, a combination of the number of reflux episodes in 15 hours with their mean duration reduced misclassification to 8-8%. Using this function, a boundary between normal and reflux can be drawn, and the degree of abnormality can be expressed visually as well as numerically.