Scale for endoscopic detection of Helicobacter pylori infection in a high prevalence area

Colombia is a country with a high prevalence of Hp infection and gastric cancer (GC). The objective of this study was to evaluate the diagnostic yield of endoscopic findings for the diagnosis of Hp infection and, based on these findings, develop a new prediction score applicable to the local population.

Endoscopic findings were evaluated, and photographs were taken according to K. Yao's 22 stations in all individuals who underwent gastroscopy between January 1 and August 2023. All patients underwent biopsy with the Sydney protocol and hematoxylin, eosin, and Giemsa staining. Univariate and bivariate analysis was performed, with logistic regression evaluating a new score applicable to the cohort.

A total of 200 patients were included. The prevalence of Hp was 47.5%. The presence of nodular gastropathy and abnormalities in the regular arrangement of collecting venules (RAC) were the endoscopic findings with the highest PPV and NPV, respectively, while atrophy assessed with the Kimura-Takemoto score had a very poor correlation with Hp. The Kyoto score had an AUC of 0.71 and the modified Kyoto score had an AUC of 0.76, which is lower than the performance in the Asian population. With the findings with better correlation, a new endoscopic score called ROVIRA was created with a cut-off point ≥3 and an AUC of 0.91 for this cohort, which suggests that it could become a more accurate clinical tool to identify active infection.

Endoscopic findings are useful for predicting Hp infection, highlighting that the preserved RAC is sensitive to rule out Hp infection with a high NPV and on the other hand the nodularity or follicular pattern due to its good specificity. The new ROVIRA scale showed better performance than the Kyoto and modified Kyoto scores in this cohort, suggesting its potential as a clinical tool in Colombia or countries where, in addition to a high prevalence of gastric cancer, there is also a high prevalence of Hp. However, these data will require further validation.