Value of the Rockall score in predicting mortality from Upper gastrointestinal bleeding

Introduction: Upper gastrointestinal bleeding (UGB) has a worldwide incidence between 50 and 140 patients for every 100 000 inhabitants a year; the mortality varies from a country to another, but during the last decades in Cuba, it fluctuates between 4 and 10% in those of non-varicose origin and reaches 30% in cases associated with portal hypertension.Objective: To evaluate the utility of the Rockall score in the risk stratification for mortality in patients with upper gastrointestinal bleeding.Material and Methods: A longitudinal prospective study was conducted in patients that came to “Dr. Carlos J. Finlay” Central Military Hospital with the diagnosis of upper gastrointestinal bleeding from November 2012 to March 2016, to whom the Rockall score was applied to determine its predictive capacity of estimating mortality.Results: 394 patients were studied; 48 of them died (12,19%). When stratifying the patients in risk groups according to Rockall score it was found that the greatest quantity belonged to the high risk group, which also included all the deceased during the study period. When applying the Receiver Operating Characteristic Curve (ROC) an area under the curve of 0,888 was obtained, which constitutes a good predictive ability for mortality.Conclusions: The Rockall score can be used to predict mortality in upper gastrointestinal bleeding during the evolution of patients suffering from it. Keywords: Rockall, upper gastrointestinal bleeding, mortality, worldwide incidence.