External validation and comparison of the Glasgow-Blatchford bleeding score, the Rockall score and the AIMS65 score in upper gastro-intestinal hemorrhage: A cross-sectional observational study in Western Switzerland

Abstract Background: Upper gastro-intestinal bleeding presents a high incidence in Emergency department. This study aims to externally validate and determine the performance of the Rockall score, the Glasgow-Blatchford bleeding, the modified Glasgow Blatchford score (mGBS) and the AIMS65 score in an Emergency department.Methods: We performed a retrospective cross-sectional observational study between January 1, 2015 and December 31, 2019. We performed Receiver-Operating Characteristic curve and area under the curve (AUROC) to compare discrimination for each scores. The primary outcome was need for intervention or death, including transfusion, endoscopic or surgery intervention. The secondary outcome was in-hospital death.Results: We enrolled 1,521 patients with UIGB. Mean age was 68 [52 – 81] years old, 62 % were men. Melena and/or hematemesis were the most common complain at ED (73%). Primary outcome was positive for 422 patients (27.7%), 76 patients (5%) were positive for our secondary outcome. The Glasgow-Blatchford score and the modified Glasgow-Blatchford score shown the highest area AUROC, respectively 0.774 (95% CI=0.750-0.798) and 0.782 (95% CI=0.759-0.805). AIMS-65 and Pre-endoscopic Rockall score shown lower discrimination, respectively, 0.684 (95% CI=0.657-0.711) and 0.647 (95% CI=0.618-0.675).Conclusion: Regarding our primary outcome, the modified Glasgow-Blatchford score and the Glasgow-Blatchford score presented a good performance. A GBS or mGBS of 0 is safe to rule-out patients with UIGB from ED. Performance of AIMS-65 score and Pre-endoscopic Rockall score were moderate.