Prognostic Value of BUN-Based Ratios for Mortality and Prolonged Hospitalization in Acute Upper Gastrointestinal Bleeding: Comparison with Established Risk Scores.

Background and

Objectives

This study investigated the prognostic value of two simple blood urea nitrogen (BUN)-based ratios, BUN/hemoglobin (Hb) and BUN/Albumin, for predicting in-hospital mortality and prolonged hospitalization in patients with acute upper gastrointestinal bleeding (UGIB). Their performance was compared with established risk scores, including the Glasgow-Blatchford score (GBS), AIMS-65, ABC and Rockall scores. Materials and

Methods

This retrospective cohort study included 486 patients evaluated for acute UGIB between March 2023 and February 2026. The diagnostic performance of BUN/Hb and BUN/Albumin ratios was assessed using receiver operating characteristic (ROC) analysis and compared with established risk scores. Associations with clinical outcomes were evaluated using logistic regression analyses.

Results

The median age was 67 years, and 292 patients (60.1%) were male. In-hospital mortality occurred in 17 patients (3.5%), while prolonged hospitalization was observed in 207 patients (42.6%). AIMS-65 showed the highest Area Under the Curve (AUC) for mortality prediction (0.799; 95% CI 0.696-0.902), followed by the ABC score (0.731) and the BUN/Albumin ratio (0.711). For prolonged hospitalization, BUN/Hb showed the highest AUC (0.706; 95% CI 0.660-0.752), although differences from established scores were not statistically significant. In multivariable analysis, BUN/Albumin remained associated with mortality, whereas BUN/Hb did not reach statistical significance for prolonged hospitalization. However, mortality-related findings should be interpreted with caution because only 17 in-hospital deaths occurred in the study cohort.

Conclusions

Simple BUN-based ratios may provide complementary prognostic information in acute UGIB. BUN/Albumin was associated with in-hospital mortality and showed modest discriminatory ability, but it did not demonstrate statistically significant superiority over established risk scores. BUN/Hb showed the numerically best discrimination for prolonged hospitalization, but without statistically significant superiority or persistent significance in multivariable analysis. Overall, these ratios may serve as supportive tools for early risk assessment rather than replacements for established risk scoring systems.