HEMORR2HAGES, ATRIA AND HAS estimation of hemorrhagic events in latin american patients with atrial fibrillation requiring anticoagulation.

Abstract

Aim: To compare and know the predictive efficacy of bleeding risk in hospitalized patients diagnosed with atrial fibrillation who require chronic anticoagulation using internationally available tools (HEMORR2HAGES, HAS-BLED and ATRIA). Results: A retrospective observational study was carried out involving 97 patients diagnosed with atrial fibrillation and treated with oral anticoagulants between January 2014 and March 2016. Taking the data without categorization, it is shown that the model used in the study  has correctly classied a 60,8% of the total cases correctly predicting the possibility of bleeding. By taking into account independet tools or variables, the prediction value increases by 10%, which indicates that the tools will be effective in the prediction in 70% of cases. Conclusion: Despite the fact that, when assessing the clinical superiority of the tools used in the prediction of events associated with bleeding, it was not possible to demonstrate a significant adjustment of the data obtained with the proposed statistical model, we can say that HEMORR2HAGES is the tool that could present a greater contribution to the prediction of bleeding according to the reality observed in Hospital.