Objective: Identify the best score that predicts each variable outcome (mortality, rebleeding and need for transfusion for more than 2 red blood cells pack) in patients with upper gastrointestinal bleeding until 30 days of the event. Material y methods: Patients included were those over 18 years, who had upper gastrointestinal bleeding between January 2014 to June 2015 in a general hospital of third level. The data was analyzed by the area under the curve ROC (Receiver Operating Characteristic).
Results: In total, there were 231 cases of upper gastrointestinal bleeding, 154 (66.7%) cases were male, the average age was 57.8 ± 20.02 years, the most common cause of bleeding was peptic ulcer: 111 (48.1%) cases, the mortality rate and rebleeding was 7.8% and 3.9% respectively. 5 patients were excluded from the analysis because they do not count with endoscopy study, the analysis was performed in 226 rest. In the evaluation of mortality, it was found an area under the curve ROC for Glasgow-Blatchford: 0.73, Rockall score: 0.86 and AIMS65 score: 0.90 (p<0.05) to predict rebleeding the Glasgow-Blatchford score: 0.73 Rockall score: 0.66 and AIMS65 score: 0.64 (p=0.41) and transfusion requirements of more than 2 globular packages the Glasgow-Blatchford score: 0.72, Rockall score: 0.67 and AIMS65 score: 0.77 (p=0.09).
Conclusions: AIMS65 score is a good predictor of mortality and is useful in predicting the need for more than 2 transfusions of red blood cells pack compared to score Glasgow-Blatchford and Rockall score.