Interobserver variation in the histopathological scoring of Helicobacter pylori related gastritis

Abstract

Aim - To test the reproducibility between two histopathologists of features of Helicobacter pylori gastritis, using the updated Sydney classification. Methods - 290 dyspeptic Dutch patients with biopsy proven H pylori infection were enrolled in the study. Gastric antral mucosal biopsy specimens were analysed before and after H pylori eradication treatment. The biopsies were scored semiquantitatively by two histopathologists, according to the updated Sydney classification system. Variables analysed included the density of H pylori infection, the degree of chronic inflammation, inflammatory activity, atrophy, intestinal metaplasia, and surface epithelial damage. Before grading biopsy specimens, both pathologists reached a consensus on the scoring of gastritis through interactive sessions using a multiheaded microscope. Subsequently all biopsy specimens were graded. Interobserver variability was also analysed using weighted κ scores. Results - For interobserver agreement on scoring the various gastritis features a high degree of reproducibility was reached overall. Agreement on grading of atrophy was the lowest; however, moderate to good reproducibility was achieved, with weighted κ values of 0.49 in the pretreatment biopsies and 0.52 in the post-treatment biopsies. Disagreement was most common in biopsy specimens with lesser degrees of atrophy. A high degree of agreement was obtained for intestinal metaplasia, with weighted κ values of 0.72 in the pretreatment biopsies and 0.73 in the post-treatment biopsies. The best agreement was reached in the assessment of the density of H pylori both before and after H pylori eradication treatment, with excellent weighted κ values of 0.76 and 0.95, respectively. The grade of reproducibility of inflammatory activity, superficial epithelial damage, and chronic inflammation was high, with weighted κ values varying from 0.60 to 0.76 and 0.62 to 0.83 before and after eradication, respectively. Conclusions - Reproducibility of grading H pylori related gastritis is high using the updated Sydney system. Despite the novel criteria for scoring atrophy, there was imperfect agreement on this feature between two independent histopathologists.