Abstract
Background: Early risk assessment in non-variceal upper gastrointestinal bleeding (NVUGIB) is essential for guiding clinical management. The neutrophil percentage-to-albumin ratio (NPAR) has recently been proposed as a marker reflecting both inflammatory response and physiological reserve. This study aimed to evaluate the prognostic value of NPAR for in-hospital mortality and its relationship with established risk scores in patients with NVUGIB. Methods: This retrospective observational study included 94 patients hospitalized with NVUGIB. NPAR was calculated using laboratory parameters obtained at admission. Patients were stratified according to AIMS65 (<2 vs. ≥2) and Rockall (<5 vs. ≥5) scores. In addition, inflammation-based indices, including neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and systemic immune-inflammation index (SII), were calculated. Predictive performance was evaluated using receiver operating characteristic (ROC) curve analysis, and associations with clinical outcomes were assessed. Results: The in-hospital mortality rate was 12.8%. NPAR values were significantly higher in patients with AIMS65 ≥ 2 and Rockall ≥ 5 (p < 0.001 for both). NPAR demonstrated good discriminative ability for AIMS65 ≥ 2 (AUC: 0.843) and moderate performance for Rockall ≥ 5 (AUC: 0.714). For mortality prediction, NPAR showed excellent performance (AUC: 0.900). A cut-off value of 27.4 yielded a sensitivity of 91.7% and a specificity of 75.6%. Higher NPAR values were associated with increased mortality risk (OR 31.9, 95% CI: 3.88–102.59, p < 0.001), while the negative predictive value was high (98.4%). In contrast, NLR, PLR, and SII showed limited predictive value for in-hospital mortality. Conclusions: NPAR shows promise as a potential prognostic biomarker for assessing disease severity and in-hospital mortality in NVUGIB. Its high negative predictive value and association with established risk scores suggest that it may complement current risk stratification approaches. However, these findings should be considered preliminary, given the retrospective design and limited sample size, and require validation in larger prospective studies.