Comparative accuracy of Rockall and Glasgow–Blatchford scores in predicting outcomes of acute upper gastrointestinal bleeding: a prospective study at upper Egypt

Abstract Background Upper gastrointestinal bleeding (UGIB) is among the most urgent haemorrhagic emergencies encountered in clinical practice. It carries a substantial burden of morbidity and mortality, making rapid triage and evidence-based management indispensable. Validated scoring systems are essential tools to guide clinicians in identifying patients who are at the highest risk and who will benefit most from early intervention. Aim To assess and compare the efficiency of the Glasgow–Blatchford Score (GBS) with the Rockall Score (RS) in predicting mortality, morbidity and clinical outcomes among Egyptian patients presenting with acute UGIB. Methods We conducted a prospective observational study enrolling 150 consecutive patients admitted with confirmed UGIB. Detailed clinical, haematological, biochemical, and endoscopic data were recorded for each participant. Mortality, re-bleeding, and the need for haemostatic intervention were designated as primary endpoints. Both scoring systems were evaluated using receiver operating characteristic (ROC) curves, with optimal thresholds derived via the Youden Index. Regression modelling was applied to identify independent mortality predictors. Results Overall, the first outcome, which was in-hospital mortality, was found to be 5.3%. Mortality cases had significantly higher rates of adverse events including rebleeding, variceal bleeding, and haemodynamic instability. The GBS demonstrated high sensitivity for predicting the need for intervention (91.4%). ROC analysis identified a GBS cut-off value of > 12.5 for mortality prediction (AUC = 0.829), while an RS > 4.5 was the optimal threshold for mortality discrimination (AUC = 0.827). Rebleeding was the strongest independent predictor of mortality (OR = 17.4, p = 0.02). Conclusion Both GBS and RS are reliable tools for risk stratification in Egyptian patients with UGIB. A GBS >12.5 and RS >4.5 are critical thresholds for intensive care and urgent endoscopy. GBS is superior for identifying patients requiring intervention, whereas RS better predicts mortality. Routine implementation of both scores may improve early clinical decision-making and patient outcomes.