Safe discharge of patients with low-risk upper gastrointestinal bleeding (UGIB): Can the use of Glasgow-Blatchford Bleeding Score be extended?

Abstract

Introduction: Risk stratification of patients with suspected upper gastrointestinal bleeding (UGIB) using either Glasgow-Blatchford Bleeding Score (GBS) or preendoscopy Rockall score to facilitate early safe discharge (GBS=0, pre-Rockall=0) has been reported. This observational study compared score utility and considered the impact of extending the range of GBS or pre-Rockall scores permitting safe discharge. Methods: Consecutive adult patients presenting to acute medical admissions or the emergency department from September 2008-March 2009 with suspected UGIB had clinical history, vital signs, laboratory and endoscopy results prospectively recorded using electronic databases. GBS, pre-Rockall scores and a composite endpoint (blood transfusion, endoscopic therapy, interventional radiology, surgery or 30-day mortality) were calculated. Results: 388 patients with suspected UGIB were identified of which 92.3% were admitted (median (range) GBS=5 (0-19) and pre-Rockall=2 (0-11)) and 7.7% discharged (GBS=0 (0-4) and pre-Rockall=0 (0-4)). 186 (47.9%) underwent in-patient endoscopy. 151 (38.9%) were found to have the composite endpoint with 77.5% having transfusion, 45.7% endoscopic treatment and an 8.0% mortality within 30 days. AUROC (95% CI) for 30-day composite endpoint was 0.92 (0.89-0.94) using GBS and 0.75 (0.70-0.80) using pre-Rockall scores. Analysis using different GBS thresholds demonstrated that GBS=0, GBS ≤1 and GBS≤2 had superior utility in identifying freedom from an adverse clinical outcome at 30-days than pre-Rockall score 0. Conclusions: GBS is superior to pre-Rockall score in identifying patients with suspected UGIB who have a low likelihood of an adverse clinical outcome and can be considered for early discharge. Diagnostic performance at different thresholds suggests that patients with GBS≤2 could be considered for early discharge, doubling the number of eligible patients (15.2 to 32.5%). This has important patient safety and resource implications.