Comparison of endoscopic, necropsy and histology scoring of equine gastric ulcers

Equine gastric ulcer syndrome (EGUS) represents a major health problem in performance horses. Much debate exists regarding endoscopic gastric ulcer scoring systems and their ability accurately to predict severity or depth of gastric ulcers. The purpose of this study was to evaluate the ability of an endoscopist to count gastric ulcers and predict gastric ulcer severity or depth using 2 endoscopic scoring systems and compare them to the same gastric ulcers see on necropsy and histopathology. Endoscopic examination of the stomach was performed under general anaesthesia on 23 mixed breed yearling horses, after feed was withheld for 24 h. Gastric ulcers were scored using 2 systems, number/severity-scoring (N/S) and practitioner simplified (PS) systems. After endoscopy, the horses were subjected to euthanasia and the stomach mucosa examined blindly and scored again at necropsy using above scoring systems. Representative gastric ulcers were then placed in 10% formalin and processed routinely for histopathology. The gastric ulcers were scored using a histopathology system (HSS) based on ulcer depth. Number scores in the N/S scoring system and PS on endoscopic and necropsy examinations were compared using Friedman 2 way analysis of variance. Where significant differences between variables were found a post hoc analysis was conducted using a Tukey's Studentised range (HSD) test. Severity scores using the N/S (ENGS) and PS scores recorded for the stomach via endoscopy and scores from HSS were evaluated for significant association using a Mantel-Haenszel Chi-square and Pearson moment correlation coefficient analysis. Significance was P<0.05. All horses had gastric ulcers in the nonglandular mucosa via endoscopic examination and at necropsy examination. Mean nonglandular ulcer number (ENGN) score was significantly (P = 0.0024) lower on endoscopic examination compared to the score at necropsy (NNGN); whereas PS scores were not significantly different on endoscopy when compared to necropsy examination. A significant but weak association was found between ENGS and HSS (3.89, P = 0.048; r = 0.453, P = 0.045) and no correlation was found between PS and HSS (1.2, P = 0.272; r = 0.117; P = 0.622). Only 1/23 horses had glandular ulcers observed via endoscopic examination whereas, 6/23 horses had glandular ulcers at necropsy and on histopathology. The prevalence of EGUS is high in stalled yearling horses. The endoscopist may underestimate the number of gastric ulcers and may not be able accurately to predict the severity or depth of those ulcers present in the nonglandular equine stomach. Furthermore, the endoscopist may miss glandular gastric ulcers.