BACKGROUND AND AIMS: Food impaction has been described in both eosinophilic esophagitis and proton pump inhibitor-responsive esophageal eosinophilia. The association between endoscopic/histologic features of esophageal eosinophilia and food impaction remains unclear. We aimed to identify clinical, endoscopic, and histologic findings associated with a history of food impaction in esophageal eosinophilia.

METHODS: This was a retrospective cohort study of adult esophageal eosinophilia patients at a tertiary center in 6/2005-10/2014. Only patients with ≥15 eosinophils/high-power field on mucosal biopsies were included. Demographics, comorbidities, symptoms, endoscopic/histologic findings on initial endoscopy, and history of food impaction were reviewed. Statistical analyses were performed using Fisher's exact test (univariate) and forward stepwise logistic regression (multivariate).

RESULTS: 400 patients (42 ± 14 years, 61 % male) were included, with 78 (20 %) having food impaction history. On univariate analyses, rings (62 vs 42 %, p = 0.003), erosions (12 vs 5 %, p = 0.03), eosinophil density on biopsy (40 [IQR = 30-50] vs 30 [IQR = 15-50], p = 0.004), and dysphagia (88 vs 62 %, p < 0.0001) were more prevalent among patients with food impaction history, while heartburn (10 vs 33 %, p < 0.0001) and abdominal pain (1 vs 12 %, p = 0.002) were less common. On multivariate analysis, rings (OR 2.6, p = 0.002), erosions (OR 3.2, p = 0.02), and eosinophil density (β-coefficient = 0.01, p = 0.04) remained associated with food impaction.

CONCLUSIONS: Findings of rings and erosions on endoscopy and increased eosinophil density on histology were independently associated with a history of food impaction in adult esophageal eosinophilia patients. Food impaction may result from both active inflammation (erosions and increased eosinophil density) and chronic fibrostenotic changes (rings).