High Glasgow Blatchford Score at admission is associated with recurrent bleeding after discharge for patients hospitalized with upper gastrointestinal bleeding

Background and study aims: Upper gastrointestinal bleeding (UGIB) is associated with significant morbidity. The Glasgow Blatchford Score (GBS) can predict endoscopic intervention and in-hospital death, but the ability to predict post-discharge outcomes is unknown. The aims of the study were to determine whether the admission GBS is associated with post-discharge rebleeding and 30-day readmission following hospitalization for UGIB. Patients and methods: In this prospective, observational, cohort study, consecutive patients who were hospitalized with UGIB were enrolled. Admission GBS scores were calculated, and patients with GBS > 7 were classified as high risk. Patients were contacted 30 days following discharge to determine: 1) rate of hospital readmission due to rebleeding, 2) all-cause readmissions, and 3) mortality. Multivariable Cox regression was used to determine associations between GBS, rebleeding, and readmission. Results: A total of 336 patients with UGIB were identified. Patients with high risk GBS were older (68 vs. 62 years; P = 0.01), and were more likely to receive blood (85% vs. 39%; P < 0.01) and require intensive care unit admission (64% vs. 50%; P = 0.02). Of the 309 patients who survived to discharge, 61 (20 %) were readmitted within 30 days, 25 (8 %) of whom had rebleeding. On multi-variable analysis adjusting for the need for endoscopic intervention, high risk GBS patients had higher rebleeding rates (hazard ratio [HR] 3.32, 95% confidence interval [CI] 1.26-11.4). On multivariable analysis, patients with more co-morbidities (HR 1.06, 95% CI 1.01-1.11) and cirrhosis (HR 2.23, 95% CI 1.19-4.04) had higher 30-day readmission rates. Conclusions: High GBS scores were associated with higher rebleeding rates following discharge. Patients with high GBS scores (> 7) should be monitored following discharge as they have a high risk of rebleeding.