333 International Multicentre Study of Optimal Use of the Glasgow Blatchford Score in Identifying Low-Risk Patients With Upper Gastrointestinal Bleeding Suitable for out-Patient Management

Abstract

Background: Upper gastrointestinal haemorrhage (UGIH) is a common cause of hospital admission. The Glasgow Blatchford score (GBS) accurately identifies patients' risk of requiring hospital-based intervention or death. Patients with GBS of zero are at very low risk of poor outcome and could have out-patient management. Some authors suggest extending the definition of low-risk patients using a higher GBS cut-off value, possibly with an age adjustment. We compared three thresholds of the GBS and two age adjusted modifications to identify the optimal cut-off value or modification and assessed variation of performance across sites. Methods: Data were collected from consecutive patients presenting with UGIH at four centres in Scotland, England, Denmark, and New Zealand. The performance of each version was evaluated using sensitivity, specificity, proportion of identified low-risk patients, and outcome in patients classified as low-risk. Findings: 2305 patients were included. There were differences in age (p=0.0001), need for intervention (p<0.0001), mortality (p<0.015), and GBS (p=0.0001) between sites. The GBS at cut-off ≤1 and ≤2, and both modifications, had higher specificities (40-49%) than the GBS at cut-off zero (specificity 22%; p<0.001). The GBS at cut-off ≤2 had the highest specificity but 3% of classified low-risk patients developed adverse outcomes. There were no differences between the GBS at cut-off 0, ≤1, and the age-adjusted versions with respect to proportion of classified low-risk patients developing adverse outcomes. All scores had low specificity when used in New Zealand (2.5-11%). Interpretation: The GBS at cut-off ≤1, and the age-extended versions, are useful for clinical use and identify almost twice as many low-risk patients as the GBS at cut-off zero. By implementation of a protocol for non-admission of UGIH-patients based on one of these scores we expect that 15-20% of all admissions can be avoided safely. Further studies evaluating these scores outside Europe are needed.