Risk stratifying patients with non-varicosic upper gastrointestinal hemorrhage using the Glasgow-Blatchford score: A case series of 91 patients

Introduction: Non-variceal upper gastrointestinal hemorrhage (NVUGIH) often leads to systematic hospitalization and emergency endoscopy. However, in most cases, it does not constitute an immediate life threat. This study aimed to evaluate the Glasgow-Blatchford Score (GBS) in predicting the need for transfusions, and/or endoscopic or surgical treatments. Materials and methods: We conducted a retrospective monocentric study including 91 patients admitted in the general surgery department of the Hospital La Rabta Tunis for a NVUGIH. Univariate analysis was performed with the Student t-test for continuous variables and with the Chi-square test for categorical variables. For a cut-off point of 9, we calculated the sensibility and the sensitivity of the GBS to predict the need for transfusions and/or hemostatic procedure. Results: During the study period, 91 patients were admitted for NVUGIH. Sixty-one patients (67%) were transfused. Seven patients (7.7%) underwent emergency surgery and two patients had endoscopic hemostasis. The predictive factors for the use of transfusion and/or hemostasic treatments were: Age >50 years, ASA score, HR ≥ 90 bpm, pallor, Hb ≤ 9,5 g/dl, Urea ≥9,7 mmol/L. For a cut-off of 9 points of the GBS, sensitivity was 85.71% and specificity 92.86%. The positive predictive value was 96%. The negative predictive value was 74%. Conclusion: The main interest of the GBS lies in dispatching the patients between intensive care units for therapeutic intervention (if GBS> = 9) and ordinary hospitalization for surveillance (if GBS <9). It then makes it possible to rationalize the management of patients with digestive hemorrhage to identify those requiring hospital treatments (transfusion, endoscopic treatment, or surgery).