[Validation of the Rockall score in elderly patients with non variceal upper gastrointestinal bleeding in a third level general hospital].

Objective: To validate the Rockall score in elderly patients with non variceal upper gastrointestinal bleeding, in terms of mortality and recurrent bleeding at 30 days follow-up.

Material And Methods: Patients older than 60 year-old, with non variceal upper gastrointestinal bleeding demonstrated by endoscopy, who were attended in a third level general hospital from June 2009 to June 2013, were included. Data was analyzed with the ROC (Receiver Operating Characteristic) curve. The area under curve was obtained to assess the Rockall score.

Results: One hundred ninety patients were included, 64.2% were males, with an average age of 74 years old. Overall mortality was 16.8% and respiratory failure with second bleeding episode was the most common cause of death (34.3% and 31.3% respectively). 5.52% of patients presented a second bleeding episode. A transfusion of more than 2 blood-packs was needed in the 24.7% of the patients. The area under ROC curve using the Rockall score was 0,76 (IC: 0.68-0.84) for mortality risk, 0.71 (IC: 0.55-0.88) for the risk of rebleeding and 0.66 (IC: 0.58-0.74) for needing a more than 2 blood-packs transfusion.

Conclusions: Rockall score is a good predictor for mortality and rebleeding during the 30 day-period after a non-variceal bleeding episode in elderly patients. The best sensibility and specificity was obtained with the scores of 5 and 6, respectively.