Evaluating the Prognostic Accuracy of New Scores for In-Hospital Outcomes in Cirrhotic Patients With Esophageal Variceal Bleeding.

Abstract

Background: Esophageal variceal bleeding (EVB) is a serious complication of cirrhosis and a major cause of upper gastrointestinal hemorrhage, carrying substantial risks of mortality and treatment failure. Prognostic scores are essential for guiding management. This study evaluated and compared the predictive accuracy of the ABC and MAP(ASH) scores with established models in cirrhotic patients with EVB.

Methods: We retrospectively analyzed 278 cirrhotic patients admitted for EVB at Da Nang Hospital, Vietnam, between January 2022 and January 2025 who underwent endoscopic variceal ligation. Data were collected for ABC, MAP(ASH), AIMS65, and Glasgow-Blatchford scores. Primary outcomes were in-hospital mortality and 5-day treatment failure. Predictive performance was assessed using AUROCs and statistical comparisons.

Results: The ABC score achieved the highest AUROC for predicting in-hospital mortality (0.88), significantly surpassing the MAP(ASH), GBS, and AIMS65 scores (p < 0.001 for all pairwise comparisons). A similar trend was observed for predicting 5-day treatment failure, where the ABC score again demonstrated the highest AUROC (0.79), outperforming both the GBS and AIMS65 scores; however, it showed comparable performance to MAP(ASH) (p = 0.19). In addition, the ABC score's risk stratification (low, medium, and high) accurately differentiated patients with varying mortality and treatment failure rates.

Conclusion: The ABC score is a highly effective and reliable tool for predicting in-hospital mortality and early treatment failure in cirrhotic patients with EVB. While the MAP(ASH) score remains valuable for predicting early treatment failure, the ABC score offers superior overall prognostic accuracy. These findings suggest that the ABC score can guide clinical decisions, particularly in resource-limited settings.