Original article
Clinical endoscopy
Comparison of endoscopy and radiographic imaging for detection of esophageal inflammation and remodeling in adults with eosinophilic esophagitis

https://doi.org/10.1016/j.gie.2017.09.037Get rights and content

Background and Aims

Eosinophil predominant mucosal inflammation is central to the diagnosis and activity assessment of eosinophilic esophagitis (EoE). Esophageal mural remodeling is an important consequence of EoE that is responsible for adverse events of dysphagia, food impaction, and esophageal stenosis. The aim of this study was to compare upper endoscopy (EGD) with barium upper GI study (UGI) for the detection of esophageal inflammation and remodeling in adults with EoE.

Methods

A retrospective review on a single-center database of adults with confirmed EoE identified those with EGD and UGI performed within 6 months of each another. Studies were reviewed for mucosal inflammatory and remodeling abnormalities.

Results

Seventy patients were included. Initial UGI results were consistent with EoE in 10% and suggestive of EoE in 39%. Review of UGI by a senior GI radiologist increased detection of changes consistent with EoE (34%). EGD identified characteristic abnormalities in 93%, which was significantly greater than UGI (67%). Inflammatory features were more frequently appreciated on EGD (74%) compared with UGI (21%). There was no significant difference in fibrostenotic changes observed on EGD (84%) versus UGI (73%).

Conclusions

EGD and UGI have similar sensitivity for identifying the remodeling consequences of EoE; however, inflammatory features are better assessed on EGD. Inadequate sensitivity of UGI for composite features of EoE limits its capabilities as a diagnostic test, although radiologists’ awareness significantly increases the diagnostic yield of UGI. UGI and EGD may identify fibrostenotic changes unappreciated by its counterpart and thus provide complementary information in select patients.

Section snippets

Study design

This study examined a retrospective cohort of patients with EoE who presented between 2001 and 2014 at a single academic medical center. Patients were identified via review of a prospectively maintained EoE patient database. The study was approved by the institutional review board at Northwestern University.

Study population

Patients included fulfilled the 2011 EoE consensus recommendation criteria for EoE.1 All patients undergoing EGD and UGI over the study period were identified. Patients were included if the

Patient characteristics

Seventy patients fit inclusion criteria for analysis, with basic demographics outlined in Table 3. There was a male predominance and a mean age of 40 years and median age of 37 years. Some 63% had an atopic phenotype defined by the presence of asthma, allergic rhinitis, atopic dermatitis, or food allergy. Characteristics of EGD and UGI studies are detailed in Table 1. Most UGI studies were performed for dysphagia without a known diagnosis (n = 41, 59%), and approximately half (n = 36, 51%) were

Discussion

The diagnostic and therapeutic assessment of EoE currently relies on symptom reporting and histologic evaluation. Neither radiologic imaging nor endoscopic findings are required for the diagnosis or treatment monitoring, although both modalities provide the means of assessing esophageal remodeling.9, 10 Prior studies have demonstrated the importance of remodeling consequences in the risk of food impaction in EoE.9, 11 In the current study we demonstrated similar detection rates for overall

References (24)

  • I. Hirano et al.

    Endoscopic assessment of the oesophageal features of eosinophilic oesophagitis: validation of a novel classification and grading system

    Gut

    (2013)
  • N. Gentile et al.

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

    Aliment Pharmacol Therap

    (2014)
  • Cited by (0)

    DISCLOSURE: The following author disclosed financial relationships relevant to this publication: I. Hirano: Consultant and research support recipient from Shire, Regeneron, Receptos, and Adare. All other authors disclosed no financial relationships relevant to this publication.

    View full text