Comparison of risk scoring systems in predicting clinical outcome at upper gastrointestinal bleeding patients in an emergency unit

Background: Admission Rockall score (RS), full RS, and Glasgow-Blatchford Bleeding Score (GBS) can all be used to stratify the risk in patients presenting with upper gastrointestinal bleeding (UGIB) in the emergency department (ED). The aim of our study was to compare both admission and full RS and GBS in predicting outcomes at UGIB patients in a Romanian ED. Patients and Methods: A total of 229 consecutive patients with UGIB were enrolled in the study. Patients were followed up 60 days after admission to ED because of UGIB episode to determine cases of rebleeding or death during this period. By using areas under the curve (AUCs), we compared the 3 scores in terms of identifying the most predictive score of unfavorable outcomes. Results: Rebleeding rate was 40.2% (92 patients), and mortality rate was 18.7% (43 patients). For the prediction of mortality, full RS was superior to GBS (AUC, 0.825 vs 0.723; P=.05) and similar to admission RS (AUC, 0.792). Glasgow-Blatchford Bleeding Score had the highest accuracy in detecting patients who needed transfusion (AUC, 0.888) and was superior to both the admission RS and full RS (AUC, 0.693 and 0.750, respectively) (P<.0001). In predicting the need for intervention, the GBS was superior to both the admission RS and full RS (AUC, 0.868, 0.674, and 0.785, respectively) (P<.0001 and P=.04, respectively). Conclusions: The GBS can be used to predict need for intervention and transfusion in patients with UGIB in our ED, whereas full RS can be successfully used to stratify the mortality risk in these patients. (C) 2013 Elsevier Inc. All rights reserved.