Correlation of endoscopic signs and mucosal alterations in children with eosinophilic esophagitis

Abstract

In children with eosinophilic esophagitis (EoE), the relationship among the endoscopic reference score (EREFS), the histology scoring system (EoEHSS), and the peak eosinophil count (PEC) is incompletely described. Our aim was to determine the relationship among EREFS, EoEHSS, and PEC and develop a predictive model using components of EREFS and EoEHSS for EoE activity. We analyzed 189 paired EREFSs, EoEHSSs, and PECs. Active EoE (aEoE; n= 98) was defined as≥15 eosinophils per high-power field and inactive EoE (iEoE; n= 91) as<15 eosinophils per high-power field. Spearman correlation (r) with Bonferroni correction was used to assess the relationship between EREFS, EoEHSS and PEC, and a back-transformed average Fisher test was used to determine the statistical significance of the differences. Receiver operating characteristic analysis was used to develop the predictive model. The relationship between total EREFS and EoEHSS was modest (r= 0.61) but significantly stronger than the correlation between total EREFS and PEC (r= 0.55; P= .04). The relationship between total EREFS and EoEHSS tended to be stronger in aEoE compared with iEoE (r= 0.41 vs 0.24; P= .09). Compared with EREFS, EoEHSS had a significantly higher area under the curve (0.78 vs 0.92; P= .04) to predict aEoE. A combination of furrows, eosinophilic inflammation, basal cell hyperplasia, eosinophilic abscess, and dilated intercellular spaces had an area under the curve of 0.97, accuracy of 98%, sensitivity of 97%, and specificity of 98% to predict aEoE. The endoscopy score modestly correlates with the histologic scoring system. Thus, the endoscopy score is not a reliable marker of tissue involvement in EoE. A panel of individual endoscopic and histologic signs hold promise to accurately predict EoE activity.