Abstract
Introduction: The aim of this study was to assess the power of clinical findings and scoring systems to predict mortality in patients over 65 years of age with non-variceal upper gastrointestinal bleeding. Materials and Method: Data on demographic profiles and risk estimation scores were retrospectively extracted from electronic hospital medical records and other electronic databases using a standard data extraction form. The AIMS65, pre-Rockall, modified Glasgow-Blatchford, T, and Baylor bleeding scores were calculated to estimate the 30-day mortality risk. The inclusion criteria were patients aged 65 and over who presented with active bleeding symptoms and had been diagnosed with acute upper gastrointestinal bleeding by the gastroenterology department. Results: The mean age was 75.23 years, and 23.6% of the patients died within 30 days. The 30-day mortality was associated with albumin levels, malignancy, and intensive care unit hospitalization. An inverse relationship was found between the albumin level and mortality, whereas the presence of cancer and the need for intensive care were associated with 2.8-fold and 2.2-fold increases in the risk of death, respectively. The AIMS65 score (AUC: 0.794) had the highest discriminative ability to predict 30-day mortality among all risk scores. Conclusion: Albumin levels, malignancy presence, and ICU admission were indicators of mortality risk in elderly patients with upper gastrointestinal bleeding. Calculating all the scores, excluding the Baylor Bleeding score, is beneficial for assessing the risk of mortality associated with upper gastrointestinal bleeding. The AIMS65 score demonstrates the highest discriminative ability. However, using these risk-scoring systems necessitates additional data. Keywords: Gastrointestinal Hemorrhage; Mortality; Aged.