Purpose: Eosinophilic esophagitis (EE) is an increasingly recognized cause of dysphagia. Information on the prevalence and factors predictive of EE is lacking. We aimed to assess:
1) The prevalence of EE in patients undergoing endoscopy for dysphagia.
2) The clinical and endoscopic factors predictive of EE.
Methods: Consecutive patients (18–60 years of age) presenting to the outpatient endoscopy unit at Mayo Clinic, Rochester for dysphagia between June 2005–6 were enrolled in this study. Patients completed the Mayo Dysphagia Questionnaire (MDQ), a validated tool. During endoscopy, mid-esophageal biopsies were obtained if there were endoscopic findings suggestive of EE, or if there was no endoscopically evident cause of dysphagia. All biopsies were read by 1of 5 experienced gastrointestinal pathologists. EE was defined as the presence of > 20 eosinophils/HPF.
Statistical analysis was performed using standard statistical tests for continuous and categorical variables. Univariate and stepwise multivariable logistic regression was performed using JMP software.
Results: 376 patients were included. Mean age was 45 years; 238 (63%) completed the MDQ. At endoscopy, changes suggestive of EE were seen in 21 (5.6%) patients. A total of 182 (48%) patients had no endoscopically evident cause for dysphagia. Overall, 222 (60%) patients had mid-esophageal biopsies, of whom 33 (15%) had EE by biopsy. 10 of 182 (9.8%) patients with normal endoscopy had EE by biopsy. 8 of 21 (38%) patients with endoscopic changes suggestive of EE (rings, furrows, mucosal fragility) had EE on biopsy. By univariate analysis, younger age, history of food impaction for more than 5 minutes and endoscopic features of EE were significant predictors of EE. Independent predictors following multivariate analysis are shown in table 1:
Conclusions: 1) Esophageal biopsies from normal appearing mucosa in patients with unexplained dysphagia will diagnose EE in a substantial proportion of patients.
2) A history of food impaction and specific endoscopic features point to EE as the diagnosis.
3) Increasing age and the use of a PPI indicate a lower risk of diagnosing EE.