Comparative Performance Analysis of Aims65, Pre-Endoscopic Rockall, Glasgow-Blatchford, and Modified Age Blood Test Comorbidity Scores in Emergency Department Non-Variceal Upper Gastrointestinal Bleeding.

Background: Upper gastrointestinal bleeding (UGIB) can lead to significant morbidity and mortality. Assessing the need for endoscopic interventions and predicting in-hospital mortality are critical, particularly where access to endoscopy is limited.

Objective: This study evaluates the predictive accuracy of pre-endoscopy risk-scoring systems for identifying the need for endoscopic interventions and assessing in-hospital mortality in UGIB patients.

Methods: We conducted a retrospective cohort study in a tertiary emergency department with 450,000 annual visits. Patients aged ≥18 years presenting with UGIB and undergoing endoscopy between June 1, 2021, and July 31, 2022, were included. Glasgow-Blatchford, pre-endoscopic Rockall, AIMS65, and modified age blood test comorbidity (ABC) scores were calculated using clinical and laboratory data.

Results: Of 795 patients, 142 required endoscopic interventions and 653 did not. The Glasgow-Blatchford score predicted intervention needs (p = 0.007) but lacked clinical relevance. All scores predicted mortality (p < 0.001); however, AIMS65 and the modified ABC score had the highest predictive accuracy, especially in those requiring interventions (p < 0.001).

Conclusion: Pre-endoscopy scores have limited utility in predicting intervention needs. Male gender, hematemesis, melena, elevated urea, and low red cell distribution width (RDW) may indicate intervention requirements. The modified ABC score demonstrated strong predictive value for mortality, making it a clinically relevant tool for UGIB risk stratification.