Abstract
Background. Blatchford and AIMS65 scores were developed to risk stratify patients with upper gastrointestinal bleed (UGIB). We sought to assess the performance of Blatchford and AIMS65 scores in predicting outcomes in elderly patients with nonvariceal UGIB. Methods. A retrospective cohort study of elderly patients (over 65 years of age) with nonvariceal UGIB admitted to a tertiary care center. Primary outcome was a combined outcome of in-hospital mortality, need for any therapeutic endoscopic, radiologic, or surgical intervention, rebleeding within 30 days, or blood transfusion. Secondary outcome was a combined outcome of in-hospital mortality or need for an intervention to control the bleed. Results. 164 patients were included. The primary outcome occurred in 119 (72.5%) patients. The secondary outcome occurred in 12 patients (7.2%). Blatchford score was superior to AIMS65 score in predicting the primary outcome (area under the receiver-operator curve (AUROC) 0.84 versus 0.68, resp., p < 0.001). Both scores performed poorly in predicting the secondary outcome (AUROC 0.56 versus 0.52, resp., p = 0.18). Conclusions. Blatchford score could be useful in predicting the need for hospital based interventions in elderly patients with nonvariceal UGIB. Blatchford and AIMS65 scores are poor predictors of the need for a therapeutic intervention to control bleeding.