Determination of a treatment response threshold for the Eosinophilic Esophagitis Endoscopic Reference Score

Background Endoscopic features of eosinophilic esophagitis (EoE) are measured using the validated EoE Endoscopic Reference Score (EREFS); however, a threshold for treatment response has not been defined. We aimed to determine a cut-point for endoscopic response as measured by EREFS. Methods We performed a secondary analysis of a randomized clinical trial comparing budesonide slurry with swallowed fluticasone multidose inhaler for initial treatment of EoE. In the parent trial, EREFS was determined before and after treatment (score range 0–9), as were histologic findings and dysphagia symptoms. We performed tabular, flexible trend, and dependent mixture analyses of measures of treatment response to select the best clinical EREFS threshold. Results In the 111 included patients (mean age 39 years; 67 % male; 96 % white), an EREFS threshold of ≤ 2 was 80 % sensitive (95 % confidence interval [CI] 69 % to 88 %) and 83 % specific (95 %CI 67 % to 94 %) for histologic response (peak of < 15 eosinophils per high-power field). Flexible trend analysis and dependent mixture modeling similarly suggested that a threshold of ≤ 2 best captured the correlation of EREFS with histologic and symptomatic measures. Dependent mixture modeling found near-total membership in the response class at EREFS of 0 or 1 and > 75 % at EREFS of 2 or 3. Conclusions An EREFS of ≤ 2 was the best clinical threshold for endoscopic response to topical steroid treatment, and was consistent with clinical and histologic response. Therefore, future studies can report a binary outcome of endoscopic response when EREFS is 2 or less.