BACKGROUND: The emergency department (ED) admission rate for elderly patients with non-variceal upper gastrointestinal bleeding (UGIB) is increasing. The AIMS65 and Glasgow-Blatchford score (GBS) are two distinct scoring systems proposed to predict in-hospital and post-discharge mortality, length of stay (LOS), and health-related costs in these patients. The objective of the present study is to evaluate the accuracy of these scoring systems, in conjunction with the Charlson comorbidity index (CCI), to predict 30-day mortality and LOS in UGIB patients who are 80 years of age or older METHODS: A retrospective analysis was undertaken of 182 patients with non-variceal UGIB who were admitted to the ED of Canakkale Onsekiz Mart University Hospital. The AIMS65, GBS, and CCI scores were calculated and adverse patient outcomes were assessed. RESULTS: The mean age of patients was 85.59 +/- 4.33 years, and 90 (49.5%) of the patients were males. The AIMS65 was superior to the GBS (area under the receiver operating characteristic curve [AUROC] 0.877 vs. 0.695, respectively) and CCI (AUROC 0.877 vs. 0.526, respectively) in predicting the 30-day mortality. All three scores performed poorly in predicting the LOS in hospital. The cutoff threshold that maximized sensitivity and specificity for mortality was three for the AIMS65 score (sensitivity, 0.87; specificity, 0.80; negative predictive values [NPV], 0.977; positive predictive values [PPV], 0.392), 14 for GBS (sensitivity, 0.83; specificity, 0.51; NPV, 0.923; PPV, 0.367), and 5 for CCI (sensitivity, 0.91; specificity, 0.22; NPV, 0.946; PPV, 0.145). CONCLUSION: The AIMS65 is a simple, accurate, and non-endoscopic scoring system that can be performed easily in ED settings. It is superior to GBS and CCI in predicting 30-day mortality in elderly patients with UGIB.