Abstract 4262: Development and Validation of a Risk Score to Predict Serious Bleeding in Stable Outpatients with Atherothrombosis
Background: In clinical practice, it is often critical to assess the risk of bleeding related to chronic therapy with antiplatelet agents and anticoagulants. A risk score to quantify the bleeding risk for chronic therapy in outpatients would therefore be helpful for clinical decision-making.
Methods: We studied patients enrolled in the REduction of Atherothrombosis for Continued Health (REACH) Registry, an outpatient cohort of 68,375 patients at risk for or with established atherothrombosis. The outcome of interest (studied over a two-year period) was one or more episodes of serious bleeding, defined as bleeding requiring hospitalization and transfusion. Risk factors for bleeding were assessed using multiple regression on bootstrap re-samples. We constructed multiple potential scoring systems based on the least complex models. Competing scores were then compared on their discriminative ability via logistic regression.
Findings: 783 patients had serious bleeding. The final score contained age: (1– 6 points); peripheral artery disease (2 points); congestive heart failure (2 points); diabetes (1 point); hypertension (2 points); smoking: (3 points); aspirin (1 point); other antiplatelet agent (2 points); both antiplatelet agents (4 points) and oral anticoagulants (4 points). Risk stratification using our score was effective in classifying the risk level of subjects, with a more than eight-fold increase in risk between the highest and lowest quintiles (cf graphic).
Conclusion: A simple risk score using clinical variables is effective in predicting the risk of serious bleeding related to chronic antithrombotic therapy in outpatients with atherothrombosis.