Comparison of Glasgow Blatchford, pre-endoscopic Rockall, and modified early warning score systems to predict the clinical outcome of patients with upper gastrointestinal bleeding in the emergency.

INTRODUCTION: Upper gastrointestinal bleeding (UGIB) is a life-threatening presentation to the emergency department (ED). In a busy emergency department, emergency department, proper risk stratification is critical to better patients management for patients with variceal or nonvariceal bleeding. This study study was designed to the scoring systems (Modified Early Warning Score (MEW), Pre-endoscopic Rockall (PER), and Glasgow Blatchford Score (GBS) in predicting 15-day mortality, requirement of blood transfusion, probability of rebleeding, and patient outcome within 15-day period in ED. MATERIAL AND METHODS: This study was carried out in with 51 acute UGIB at the emergency department, Department (ED) and department of Medical gastroenterology (MGE) of Narayana Medical College and Hospital, Nellore, India, from February 2020 to June 2022. Clinical The clinical history, presenting signs and symptoms, comorbidities, vitals, laboratory variables, endoscopic diagnosis and treatment history of all patients were recorded and they were followed for 15 days to evaluate rebleeding and its outcome. The chi-square test was applied to qualitative variables. RESULTS: The study enroled 51 patients, of which 82.4% were male, with the majority between 51 and 60 years of age. The majority of cases were presented to the emergency department with haematemesis (60.8%). Non-vasriceal bleeding observed in 94.1% cases. In the 15-day follow- up, 7 patients (7.84%) died and 7 patients (13.7%) rebleed. Totally 26 (50.9%) had a MEWS score of 1, 13 (25.4%) had a score of 2, 9 (17.64%) had a score of >3, and 3 (5.88%) had a score of 0. A PER score of 1 was found in 17 (33.3%), 2 in 4 (7.84%), 3 in 3 (5.88%), 4 in 6 (11.76), 5 in 10 (19.6%) and 6 in 5 (9.80%) patients. 9 (17.6%) cases had a low-risk GBS score of 0-5. Compared to MEWS and GBS, the ROC curve for mortality calculated at 15 days for PERS was 0.96; 95% CI of 0.9 - 1.0, indicating good accuracy. The AUCROC curve for predicting rebleed by PERS score showed area under curve - 0.79, which is also better than the other 2 scoring systems. CONCLUSIONS: PER had a higher negative predictive value (90%) than GBS (80.7%) and MEWS (88.1%) for rebleed measurement. GBS had a higher negative predictive value (96.15%) than PER (52.5%) and MEWS (42.8%) to predict admission of a patient with UGIB. The GBS score >8.5, MEWS score >1.5, and the PER score 4.5 predicted rebleeding. The GBS predicted the need for packed red blood cell transfusions better than the MEWS score and the pre-endoscopic Rockall score. The MEWS score is better at predicting admission and type of bleeding.