RESUMO
BACKGROUND: Glasgow-Blatchford score (GBS) has been developed for risk stratification in management of acute upper gastrointestinal (GI) bleeding. However, the performance of GBS in patients with lower GI bleeding is unknown. AIM: To evaluate the performance of full or modified GBS and modified GBS in prediction of major clinical outcomes in patients with lower GI bleeding. METHODS: A retrospective study of patients admitted to a tertiary care center with either non-variceal upper GI bleeding or lower GI bleeding was conducted. The full and modified GBS were calculated for all patients. The primary outcome was a combined outcome of inpatient mortality, need for endoscopic, surgical, or radiologic procedure to control the bleed or treat the underlying source, and need for blood transfusion. RESULTS: A total of 1026 patients (562 cases for upper GI and 464 cases for lower GI) were included in the study. Hospital-based interventions and mortality were significantly higher in upper GI bleeding group. The performance of the full GBS in lower GI bleeding (area under the receiver operating curve (AUROC) 0.78, 95% CI 0.74-0.82) was comparable to full GBS in upper GI bleeding (AUROC 0.77, 95% CI 0.73-0.81) in predicting the primary outcome. Similarly, the performance of modified GBS in lower GI bleeding was shown to be comparable to modified GBS in upper GI bleeding (AUROC 0.78, 95% CI 0.74-0.83 vs. AUROC 0.76 95% CI 0.72-0.80). CONCLUSION: In patients with lower GI bleeding, both full GBS and modified GBS can predict the need for hospital-based interventions and mortality.