The pattern of inflammation in eosinophilic esophagitis (EoE) is patchy, necessitating multiple biopsies to optimize diagnostic yield. Current consensus-based guidelines recommend 6 total biopsy samples at 2 sites, distal and either middle or proximal esophagus, although this recommendation is based on limited data. We aimed to determine whether this biopsy protocol sufficiently captures EoE diagnoses by evaluating the distribution of eosinophilia in a large EoE cohort.

This was a retrospective study of consecutive patients newly diagnosed with EoE with ≥2 esophageal segments biopsied. Demographic variables, clinical characteristics/history, endoscopic findings, and histologic results were manually reviewed. Distribution (proximal, middle, and/or distal) of eosinophilia (>15 eosinophils/high-power field [HPF]) was assessed. Predictors for non-distal disease (<15 eosinophils/HPF on distal biopsy samples) were evaluated by using multivariable logistic regression.

A total of 511 patients newly diagnosed with EoE with ≥2 segments biopsied were included. All patients underwent distal esophageal biopsy. Overall, 286 (56.0%) had ≥1 site with <15 eosinophils/HPF, including 51 (10%) with non-distal disease. Among patients with 3 segments biopsied (n = 60), 19 (31.7%) had eosinophilia at only 1 site, including 6 (10%) with isolated midesophageal disease and no isolated proximal eosinophilia. Discordant mid and proximal biopsy results were found in 18 (30%) patients, with 17 of 18 (94.4%) having mid esophageal eosinophilia. On multivariable analysis, increasing age (odds ratio, 1.02; 95% CI, 1.002-1.04; P = .03) and male sex (odds ratio, 1.89; 95% CI, 1.002-3.55; P = .049) independently predict non-distal disease.

Isolated segmental eosinophilia is common in EoE, including up to 10% non-distal disease. Discordant mid and proximal biopsy findings are prevalent, with no isolated proximal eosinophilia. Standard protocol should include routine biopsies of both distal and middle esophagus to maximize diagnostic yield.