Oesophageal narrowing is common and frequently under-appreciated at endoscopy in patients with oesophageal eosinophilia

Background: Estimation of the prevalence of oesophageal narrowing and its clinical relevance in patients with oesophageal eosinophilia is probably underestimated by endoscopy. Aims: To determine the sensitivity and specificity of oesophagogastroduodenoscopy (EGD) for oesophageal narrowing when compared to a structured oesophagram. Methods: We conducted a retrospective chart review at Mayo Clinic, Rochester of adult patients with eosinophilic oesophageal infiltration of >15 eosinophils/ hpf with symptomatic dysphagia, who underwent a structured barium oesophagram. The sensitivity and specificity of EGD were evaluated against the gold standard of oesophagram. Demographic and multiple clinical factors were evaluated as potential predictors of oesophageal narrowing. Results: Of the 58 patients identified, 34 (58.6%) had a narrowed oesophageal diameter (<21 mm). EGD had poor sensitivity (14.7%, 95% CI 5.0-31.1%) for detection of a narrowed oesophagus and only modest specificity (79.2%, 95% CI 57.8-92.9%). Even at a cut-off diameter of ED_max ≤ 15 mm, EGD had a sensitivity of only 25.0% (95% CI 5.5- 57.2%) for narrowed oesophagus. A history of >5 food impaction episodes, endoscopic rings, and female sex were the best predictors of oesophageal narrowing. 86% (6/7) patients with persistent dysphagia despite remission of histological eosinophilia responded to oesophageal dilation all of whom had radiological oesophageal narrowing and 71% of whom had no perceived oesophageal narrowing at EGD. Conclusions: Symptomatic oesophageal narrowing identified by barium oesophagography is common and under-recognised at endoscopy in patients with oesophageal eosinophilia.