Abstract 11915: Less Aspirin, Less Bleeding for Patients on Chronic Warfarin Anticoagulation for Atrial Fibrillation and/or Venous Thromboembolic Disease

Introduction: For some patients on chronic warfarin anticoagulation for atrial fibrillation and/or venous thromboembolic disease, the addition of aspirin (ASA) may increase bleeding with uncertain therapeutic benefit. Between 2017 and 2018, the six-center Michigan Anticoagulation Quality Improvement Initiative implemented a multi-center quality improvement project to reduce excess aspirin use. Hypothesis: We anticipated reductions in overall ASA use and bleeding. Methods: Warfarin treated patients taking ASA without a clear indication (eg. coronary artery disease) were identified. Their primary care provider was contacted to advise if ongoing ASA+warfarin was recommended. ASA was discontinued if suggested by their doctor. Data was abstracted by trained abstractors using predefined forms. In December 2019, we evaluated monthly rates of excess ASA use, bleeding, and thrombotic outcomes. Data were analyzed using piecewise linear regression. Results: We followed 6,738 patients for a median (interquartile range) of 6.7 months (3.2-19.3 months). Aspirin use declined significantly over the 24 months before our intervention (P<0.001) with an accelerated decline post intervention (p=0.001, Figure 1). Our intervention was associated with a non-significant decline in bleeding (p=0.12), emergency room (ER) visits for bleeding (p=0.35), admissions for bleeding (p=0.97), a significant decline in major bleeding (p=0.029), and no significant change in thrombotic outcomes (p=0.34). Including the decline in ASA use starting 24 months pre-intervention, reducing ASA use resulted in significant declines in any bleeding, major bleeding, and ER visits for bleeding (p=0.021, p=0.001, p=0.038 respectively). There remained no change in thrombotic outcomes (p=0.36). Conclusions: Our quality improvement intervention reduced aspirin use. Reductions in aspirin use were associated with reduced bleeding outcomes without increased thrombotic outcomes.