Lack of discriminant value of dyspepsia subgroups in patients referred for upper endoscopy

Abstract

Background: The subdivision of undiagnosed patients with dyspepsia into symptomatic subgroups (ulcerlike, dysmotilitylike, refluxlike, and nonspecific) may give a clue to the underlying cause; however, the value of this approach in practice is unclear. This study aimed to determine the discriminant value of dyspeptic symptoms and combinations of these symptoms.

Methods: A consecutive sample of 820 outpatients (median age, 62 years; 47% male) completed, before endoscopy, a validated questionnaire that measured 46 gastrointestinal symptoms.

Results: Of patients with functional dyspepsia (n = 162), 17% had ulcerlike, 9% dysmotilitylike, and 16% refluxlike dyspepsia alone; 31% fell into two or more symptom subgroups; and 27% had nonspecific symptoms. A similar distribution of the dyspepsia subgroups was observed in patients without functional dyspepsia. Younger age, female gender, frequent upper abdominal pain, no (or sometimes) pain relief with antacid use, and infrequent vomiting were predictive of functional dyspepsia vs. all other diagnoses (at a specificity of 80%, the sensitivity was 60%). The dyspepsia subgroups were poor discriminators in a separate model for functional dyspepsia vs. all other diagnoses (at a specificity of 80%, the sensitivity was only 43%).

Conclusions: The dyspepsia subgroups, as currently defined, appear to have little clinical utility and may be an inappropriate way of classifying dyspepsia.