Pilot Study of Radiological Assessment of Short-Term Corticosteroid Therapy on Esophageal Distensibility in Eosinophilic Esophagitis (EoE)

Abstract

Background: Esophageal foreign body impaction (EFBI) is a major presentation of eosinophilic esophagitis (EoE), but the impact of the increasing prevalence of EoE on the epidemiology of EFBI is not known. Aim: To determine whether the number of EFBI cases has increased with the increasing prevalence of EoE, and to assess characteristics of patients presenting with EFBI. Methods: Cases of EFBI from 2002 to 2009 were identified by querying the University of North Carolina Hospital's billing database (2005-2009), clinical database (2006-2009), and endoscopy database (2002-2009) for all records with ICD-9 code “935.1foreign body in the esophagus.” Charts were reviewed to confirm EFBI and to extract all pertinent data. Cases of EoE were defined per 2007 consensus guidelines and cross-referenced with our EoE database. Results: Of the 479 cases of EFBI identified (59% male, 67% Caucasian, mean age 35 ± 30 yrs), 289 (60%) were food impactions, 107 (22%) were coin impactions, and 83 (17%) were other objects (ex: golf tee, tooth, staple) or not specified. Overall, 45 EFBI patients (9%) had EoE, 39 of whom were diagnosed after the EFBI episode. The number of EGDs performed for food impaction nearly quadrupled between 2002 and 2009, however increasing diagnosis of EoE did not fully account for this trend (see figure). Only 97 (32%) of the 303 patients undergoing EGD had esophageal biopsies, and 31 (32%) of those patients biopsied had EoE (19 children, 12 adults). The type of impacted item (food vs. other) was associated with obtaining esophageal biopsies during the EGD (35% vs. 20%, p=0.012), but the proportion of cases with biopsies did not change over time. More subjects with food impaction had EoE compared to those with other types of EFBI (12% vs. 4%; p=0.003); only one patient with a coin impaction had EoE. On multivariate logistic regression analysis, the presence of EoE was themost significant predictor of recurrent EFBI (OR 3.6, 95%CI: 1.7-7.6). Conclusions: The number of EGDs performed for EFBI has increased dramatically at our center, and the increasing prevalence of EoE in this patient population only partially explains this trend. However, because only a minority of patients had biopsies at the time of EFBI, the incidence of EoE in this population is likely substantially underestimated. Physician education regarding the high prevalence of EoE in subjects with food impaction may increase the proportion of subjects biopsied and the number of EoE diagnoses made in this patient population.