Abstract
BACKGROUND: A proportion of patients with active eosinophilic esophagitis (EoE) have a normal-appearing esophagus on esophagogastroduodenoscopy (EGD). We aimed to determine the associations between the baseline clinical features and the endoscopically normal esophagus in EoE, as well as time trends in reporting. METHODS: In this retrospective study of active EoE cases from 2002 - 2018, patients with and without esophageal endoscopic abnormalities were compared. Multivariable logistic regression identified the independent predictors of a normal EGD. The proportion of patients with a normal EGD was determined per year, and before and after the introduction of the first EoE guidelines and the EoE Endoscopic Reference Score (EREFS). RESULTS: Of 878 EoE patients, 101 (11.5 %) had an endoscopically normal esophagus; they were younger (8.3 vs. 25.4 years), had shorter median symptom duration before diagnosis (2.8 vs. 5.0 years), were less likely to have dysphagia (40 % vs. 76 %) or food impaction (8 % vs. 33 %), and more likely to have abdominal pain (37 % vs. 19 %) (P < 0.01 for all). On multivariable logistic regression, independent predictors of a normal esophagus were younger age (odds ratio [OR] 0.96, 95 % confidence interval [CI] 0.94 - 0.98), abdominal pain (OR 2.03, 95 %CI 1.13 - 3.67), and lack of dysphagia (OR 0.49, 95 %CI 0.26 - 0.93). The proportion of patients with a normal esophagus decreased from 21 % before the first EoE guidelines to 7 % (P < 0.01) after introduction of the EREFS. CONCLUSIONS: An endoscopically normal esophagus is seen in ~10 % of active EoE patients and should not preclude biopsies; younger age, abdominal pain, and lack of dysphagia are independent predictors. The proportion of normal EGDs decreased over time, suggesting improved recognition of endoscopic findings.