Inflammatory Markers in Gastric Fluids Differentiate Patients With Eosinophilic Gastritis: Search for a Disease Screener.

Introduction: Eosinophilic gastritis (EoG) is commonly missed because of limited biopsy collection and failure to request tissue eosinophil counts on gastric biopsies. Continued efforts to identify less invasive biomarkers to improve diagnosis and disease monitoring have resulted in little success. We studied gastric fluids and serum in EoG and control patients to determine whether less invasive inflammatory markers can predict active EoG disease.

Methods: By Luminex MAGPIX, we measured cytokines, chemokines, and matrix metalloproteinases biomarkers from gastric fluids collected during routine upper endoscopy and serum from patients with active EoG (n = 20), active eosinophilic esophagitis (EoE) (n = 21), and non-eosinophilic gastrointestinal disease controls (n = 19). Comparison of biomarker concentrations among patients and predictive modeling for EoG status were performed.

Results: Twenty-six biomarker in gastric fluids and 6 biomarkers in serum were significantly elevated in active EoG compared with active EoE and non-eosinophilic gastrointestinal disease controls. Tree-based model, eXtreme Gradient Boosting, identified important biomarkers in both gastric fluids and serum predictive of EoG.

Discussion: We successfully measured inflammatory markers in gastric secretions. Th2-mediated cytokines were elevated in gastric secretions in EoG, differentiating them from EoE and expanding our understanding of inflammation in EoG. Notably, our results are the first to implicate matrix metalloproteinases in the EoG inflammatory process. Importantly, we found that gastric secretions can discern patients with active gastric eosinophilia involvement. Modeling identified 9 markers that predicted EoG with an area under the curve of 0.86. With further validation, gastric fluids could be used as an easy test to screen EoG by identify patients who would benefit from histopathologic enumeration of eosinophils on biopsies.