Abstract P143: Effects Of Oral Anticoagulant Choice On Mortality In The Primary Treatment Of Non-cancer Venous Thromboembolism: Medicare, 2011-2016

Abstract

Introduction: Oral anticoagulants (OAC) are the mainstay of venous thromboembolism (VTE) treatment, with newer direct oral anticoagulants (DOACs; rivaroxaban, apixaban, and dabigatran) approved for the treatment of VTE. Although clinical trials have evaluated DOACs versus warfarin individually, there are few direct comparisons of OAC treatment choice among VTE patients and risk of mortality. Hypothesis: Patients enrolled in Medicare who are prescribed DOACs for primary treatment of incident VTE will experience a lower rate of mortality compared to patients prescribed warfarin. Among different DOACs, there will be no difference in rate of mortality. Methods: We used the Medicare 20% sample to identify a cohort of 84,877 individuals with incident VTE from 2011-2016. We excluded VTE patients with evidence of prevalent cancer (N=27,603). Beneficiaries were then matched by OAC using patient characteristics at date of VTE diagnosis to create an analytic dataset of 35,922. All-cause mortality and date of death were ascertained from linkage with the National Death Index. We used Cox regression to estimate adjusted hazard ratios (HR) and 95% confidence intervals (CIs) for the association of OAC choice with mortality. All analyses were limited to 6-months of follow-up. Multiplicative interaction by age (≥75 years), sex, race and kidney disease were explored. Results: In our matched sample 21,320 individuals were prescribed warfarin, 10,504 rivaroxaban and 3,681 apixaban. The average age was 77 years (SD: 8 years), 63% were female, 82.0% were white, 10.4% black and 7.6% were other/unknown. There were 2,701 deaths within 6 months of VTE diagnosis, of which 1,726 were warfarin, 604 were rivaroxaban, and 272 were apixaban. In the adjusted models, individuals prescribed rivaroxaban had 0.85 times the rate of 6-month all-cause mortality compared to those prescribed warfarin (HR: 0.85; 95% CI: 0.77, 0.93). There was no difference in mortality rates comparing apixaban to warfarin (HR: 0.98; 95% CI: 0.85, 1.13) or apixaban to rivaroxaban (HR: 1.12; 95% CI: 0.95, 1.32). For the interaction analyses among those aged under 75, apixaban users had a higher risk of mortality compared to warfarin (HR: 1.37; 95% CI: 1.06, 1.77) and rivaroxaban users (HR: 1.32; 95% CI: 0.99, 1.77). No other interactions were observed. Conclusions: We found differences in mortality related to anticoagulation choice for the treatment of VTE, with rivaroxaban having a lower mortality risk than warfarin and a similar risk overall with apixaban. However, apixaban had a higher risk of mortality in secondary analyses stratified by age, with a higher risk in those <75. The age-dependent association of apixaban with mortality was not hypothesized and requires further confirmation.