Venous and Arterial Thrombosis in Patients Hospitalized for Acute Medical Illness

Abstract

Patients who are hospitalized for acute medical illness are at an increased risk of both venous and arterial thrombosis. Several large randomized clinical trials have established the efficacy of anticoagulation for the in-hospital duration, in the prevention of venous thromboembolism (VTE). These studies have led the American College of Chest Physicians to recommend administration of low molecular weight heparin or low dose unfractionated heparin for a duration of 6 to 14 days to acute medically ill patients who are at a high risk for VTE and are not at high risk of major bleeding. This guideline recommends using the Padua Prediction score to evaluate patient’s risk of VTE though it does acknowledge that this score lacks prospective validation and generalizability, among other limitations. Subsequent studies have demonstrated that these patients’ risk of thrombosis extends beyond the in-hospital period. Approximately 60% of VTE occurred after hospital discharge and 53% had occurred within 30 days of the index hospitalization. Five large randomized trials have since evaluated the efficacy and safety of extended-duration thromboprophylaxis (for 28 to 47 days) with enoxaparin, rivaroxaban, apixaban and betrixaban in hospitalized acutely medically ill patients. Of these agents, betrixaban, a novel factor Xa inhibitor, was the only one to demonstrate a significantly reduced risk of VTE coupled with no significant difference in major bleeding. Following the release of the results of the APEX trial, the FDA licensed betrixaban for extended duration thromboprophylaxis in acute medically ill patients at high risk for VTE. Much attention is given to the prevention of venous thrombosis in acutely medically ill patients; however, despite sharing common risk factors with arterial thrombotic events, little is done in the way of prophylaxis against ischemic cardiovascular events in these patients. Hypercoagulability and inflammation are known mechanisms that contribute to both pathologies; yet little has been done to explore the efficacy of anticoagulation in reducing the incidence of arterial thrombotic events in this patient population. We hypothesized that, by upstream inhibition of the prothrombinase complex with factor Xa inhibitors, acutely ill patients would not only experience less venous thrombosis but also reduced major adverse cardiovascular events. Despite successes in the development of novel therapeutic agents and regimens that have prevented thrombosis in patients enrolled in clinical trials, the question of how to identify high thrombotic risk (and low bleeding risk) patients in a clinical setting remains unanswered. Risk scores have known limitations that have prevented their widespread use in this setting. In fact, the most recent extended-duration thromboprophylaxis randomized trial enriched for high-risk patients using a popular risk score as a screening tool yet failed to demonstrate statistical significance in 12,000 patients despite a 23% relative risk reduction in the primary endpoint. This suggests a lack of sensitivity of the risk score in identifying truly high-risk patients. Machine learning algorithms are constructed to search for patterns in data that provide maximum predictive ability. These learning methods have demonstrated superiority to traditional diagnostic and prognostic tools in various domains. However, the performance of machine learning methods in the prediction of the occurrence of thrombosis has not been previously explored. We sought to explore the utility of this novel methodology in predicting thrombotic events in acutely medically ill patients enrolled in the APEX trial.