Acute Upper Gastrointestinal Bleed and Implication of Pre-endoscopy Score-Based Risk Stratification

Background: Acute upper gastrointestinal (GI) bleed is a medical emergency and need triaging, timely decision for endoscopy as well as prognostication. The aim was to compare the different pre-endoscopic risk assessment scores in patients of acute GI bleeding.

Methods: Patients with UGI bleeding presenting in the emergency were assessed for clinical, laboratory parameters and organ dysfunction. GBS (Glasgow–;Blatchford Score), AIMS65 (Albumin, INR, Mental Status, Systolic Blood Pressure, Age >65 years), and clinical Rockall analyzed in relation to need for urgent endoscopic or surgical interventions, re-bleeding, blood transfusion, complications, hospital stay, and 30-days survival.

Results: Hundred patients with mean age of 44.43 ± 14.49 years presented with hematemesis (40%), melena (37%), and both in (23%). Variceal bleeding was noted in 9%. At admission, mean GBS, AIMS65, and cRockall were 7.04, 0.93, and 1.99, respectively. All three equally predict the in-hospital mortality. GBS had higher sensitivity (96%) than cRockall (81.2%) and AIMS65 (61.9%) for predicting need of endoscopy. The sensitivity and specificity for need of blood transfusion were predicted in GBS score >6 (97.1%, 86.2%), AIMS65 score >2 (65.7%, 90%), and for cRockall score >3 (88.6%, 84.6%), respectively. Prediction for admission to intensive care unit (ICU) was comparable. Rebleeding was better predicted by GBS (AUROC-0.803) than AIMS65 (AUROC-0.745) and cRockall (AUROC-0.762).

Conclusions: Acute UGI bleeding in only one-10th is variceal. Need of blood transfusion and/or endoscopic intervention better predicted with a GBS score, whereas cRockall score can triage for early ICU admission. A low score suggests conservative treatment in absence of endoscopy. These scores could be part of management algorithm for particularly in resource poor setting.

Publication History

Article published online:
22 April 2024