Stress hyperglycemia ratio predicts adverse outcomes in emergency department patients with upper gastrointestinal bleeding

Abstract

Introduction Upper gastrointestinal bleeding (UGIB) is a common emergency condition with substantial morbidity. Early identification of patients at risk for adverse outcomes is essential for timely management. The stress hyperglycemia ratio (SHR) adjusts admission glucose for baseline glycemic control and may better reflect acute physiological stress than absolute glucose levels. We aimed to determine whether SHR predicts critical outcomes in emergency department (ED) patients with UGIB. Methods We retrospectively analyzed 345 adults with endoscopically confirmed UGIB at a tertiary medical center. SHR was computed as admission glucose divided by estimated average glucose from hemoglobin A1c. Multivariable logistic regression assessed associations between SHR and outcomes: intensive care unit (ICU) admission, blood transfusion, rebleeding, acute kidney injury (AKI), acute respiratory failure (ARF), in-hospital mortality, procedural intervention, and esophageal variceal (EV) bleeding. Predictive performance was compared with the complete Rockall score (CRS) and Glasgow-Blatchford score using receiver operating characteristic curves. Results Elevated SHR was independently associated with higher risks of ICU admission (adjusted odds ratio [aOR] = 2.10, P < 0.001), transfusion (aOR = 6.30, P < 0.001), rebleeding (aOR = 1.75, P = 0.04), AKI (aOR = 1.79, P < 0.001), and ARF (aOR = 1.96, P = 0.01). SHR moderately predicted transfusion (area under the curve [AUC] = 0.716) and ICU admission (AUC = 0.637), outperforming the CRS for both. Adding SHR to CRS improved transfusion prediction (ΔAUC = 7.2%, P = 0.02). Patients with SHR > 1.9 had significantly higher rates of ICU admission, transfusion, organ dysfunction, and EV bleeding. SHR remained predictive in both diabetic and nondiabetic subgroups. No significant association was observed between SHR and mortality or procedural intervention. Conclusion SHR was independently associated with adverse outcomes in UGIB, especially ICU admission and transfusion. As a simple, rapidly available marker that adjusts for baseline glycemic control, it may complement existing risk scores and support early, pre-endoscopic risk stratification in the ED, and warrants validation in prospective studies.