Comparison of risk scoring systems for critical care patients with upper gastrointestinal bleeding: predicting mortality and length of stay.

Upper gastrointestinal bleeding (UGIB) is a common reason for intensive care admission. While there exist a number of UGIB scoring systems which are used to predict mortality, there are limited studies assessing the discriminative value of these scores in intensive care unit (ICU) patients. The purpose of this study was to analyse five different UGIB scoring systems in predicting ICU mortality and length of stay and compare them to two commonly used ICU mortality scoring systems. We retrospectively identified all patients requiring ICU admission for UGIB to St James's Hospital over an 18-month period. We calculated their AIM65, Glasgow- Blatchford score, pre- and post-Rockall score, ABC, APACHE II and SOFA scores. We used area under the receiver operating characteristic curve (AUROC) to compare the predictive values of these six scoring systems for ICU and hospital mortality as well as ICU length of stay greater than seven days. APACHE II showed excellent discriminative value in predicting mortality in ICU patients (AUROC: 0.87; CI: 0.75-0.99) while the SOFA score showed good discriminative value (AUROC: 0.71; CI: 0.50-0.93). None of the UGIB scoring systems predicted mortality in these patients. All scoring systems showed poor discriminative value in predicting ICU length of stay. We were not able to validate any of these UGIB scoring systems for mortality or length of stay prediction in ICU patients. This study supports the validity of APACHE II as a clinical tool for predicting mortality in ICU patients with UGIB.