Contemporary utilization of antithrombotic therapy for stroke prevention in patients with atrial fibrillation: an audit in an Australian hospital setting

Background: To document antithrombotic utilization in patients with nonvalvular atrial fibrillation (NVAF), particularly, recently approved NOACs (nonvitamin K antagonist oral anticoagulants) and warfarin; and identify factors predicting the use of NOACs versus warfarin. Methods: A retrospective audit was conducted in an Australian hospital. Data pertaining to inpatients diagnosed with atrial fibrillation (AF) admitted between January and December 2014 were extracted. This included patient demographics, risk factors (stroke, bleeding), social history, medical conditions, medication history, medication safety issues, medication adherence, and antithrombotic prescribed at admission and discharge. Results: Among 199 patients reviewed, 84.0% were discharged on antithrombotics. Anticoagulants (+/- antiplatelets) were most frequently (52.0%) prescribed (two-thirds were prescribed warfarin, the remainder NOACs), followed by antiplatelets (33.0%). Among 41 patients receiving NOACs, 59.0% were prescribed rivaroxaban, 24.0% dabigatran, and 17.0% apixaban. Among patients aged 75 years and over, antiplatelets were most frequently used (37.0%), followed by warfarin (33.0%), then NOACs (14.0%). Compared with their younger counterparts, patients aged 75 years and over were significantly less likely to receive NOACs (14.0% versus 28.0%, p = 0.01). Among the ` most eligible' patients (Congestive Cardiac Failure, Hypertension (, Age >= 75 years, Age = 65-74 years, Diabetes Mellitus, Stroke/Transient Ischaemic Attack/Thromboembolism, Vascular disease, Sex female[CHA(2)DS(2)-VASc] score >= 2 and no bleeding risk factors), 46.0% were not anticoagulated on discharge. Patients with anaemia (68.0% versus 86.0%, p = 0.04) or a history of bleeding (65.0% versus 87.0%, p = 0.01) were less likely to receive antithrombotics compared with those without these risk factors. Warfarin therapy was less frequently prescribed among patients with cognitive impairment compared with patients with no cognitive issues (12.0% versus 23.0%, p = 0.01). Multivariate logistic regression modelling identified that patients with renal impairment were 3.6 times more likely to receive warfarin compared with NOACs (odds ratio = 3.6, 95% confidence interval = 0.08-0.90, p = 0.03, 60.0% correctly predicted; Cox and Snell R-2 = 0.51, Nagelkerke R-2 = 0.69). Conclusion: Despite the availability of NOACs, warfarin remains a preferred treatment option, particularly among patients with renal impairment. The high proportion of eligible patients still being prescribed antiplatelet therapy or ` no therapy' needs to be addressed.