Comparative Evaluation of the Glasgow-Blatchford, Pre-Endoscopic Rockall, and CANUKA Scores in Predicting Morbidity in Patients with Upper Gastrointestinal Bleeding

Abstract

Introduction and Objective: Upper gastrointestinal bleeding (UGIB) is a common reason for presentation in emergency departments. Various scoring systems have been developed to determine the need for endoscopic intervention and to predict the risk of adverse outcomes. The aim of this study was to evaluate the Glasgow-Blatchford score (GBS), CANUKA score, and pre-endoscopic Rockall score as predictors of a composite morbidity outcome in patients with non-variceal UGIB. Materials and Methods: This retrospective study included patients with non-variceal UGIB and compared the discriminative ability of three scoring systems to predict a composite morbidity outcome, defined as a hospital stay exceeding 72 hours, intensive care unit (ICU) admission, need for endoscopic hemostatic therapy, blood product transfusion, embolization, repeat endoscopic intervention, or surgical management. Areas under the curve (AUC) were compared using DeLong’s method. Results: A total of 496 patients were included (median age: 60 years; 62.3% male; 20.6% with cancer). The composite morbidity outcome occurred in 49.4% of patients, and 3.23% died. The CANUKA (AUC: 0.8101) and GBS (AUC: 0.8070) scores demonstrated superior discriminative ability (p = 0.0124). Individually, the GBS showed the best capacity to discriminate patients requiring ICU care (AUC: 0.8011; p = 0.0004) and blood product transfusion (AUC: 0.9123; p < 0.0001). Both CANUKA and GBS scores were superior in predicting the need for repeat endoscopic intervention, with AUROC values of 0.6601 and 0.6555, respectively. Conclusion: These findings suggest that the CANUKA and GBS scores are superior predictors of morbidity and should be preferred when making decisions regarding early intervention and close monitoring.