Comparative effectiveness and safety of extended treatment vs. discontinuation of oral anticoagulants in patients with venous thromboembolism

Background: The effectiveness and safety of discontinuing vs. continuing oral anticoagulants (OACs) for the first-time venous thromboembolism (VTE) after initial 90-day anticoagulation in routine care remain unclear.

Objectives: To compare the health outcomes between OAC discontinuation vs. continuation after the initial 90-day treatment among patients with first-time VTE

Methods: Using the US Medicare data from 2009 to 2020, we identified individuals aged ≥ 65 years who newly initiated an OAC (warfarin, apixaban, rivaroxaban, dabigatran, or edoxaban) within 30 days following a first-time VTE admission and continued the OAC for at least 90 days. We defined OAC discontinuation as having a gap between OAC prescriptions any time after the initial 90-day OAC treatment and the end of the 30-day gap was set as the discontinuation date. For each discontinuer, we sampled a continuer using incidence-density sampling method (index date was “discontinuation date” for the discontinuer and “sampling date” for the continuer). We used 1:1 propensity score (PS) matching to balance 87 covariates. We ascertained the composite of hospitalization due to recurrent VTE, major bleeding, and mortality as the primary outcome and each individual outcome as the secondary outcome. We used Cox proportional hazards models to estimate hazard ratios (HR) and generalized linear models to estimate rate differences (RD) per 1,000 person-years.

Results: Our study cohort included 14,851 PS-matched warfarin discontinuer-continuer pairs (mean age [SD] 77.2 [7.5], female 63.7%; median follow-up (mFUP) time=220 days), 5,782 rivaroxaban PS-matched pairs (mean age 75.8 [7.5], female 60.3%; mFUP time=309 days), 5,266 apixaban PS-matched pairs (mean age 77.0 [7.5], female 64.0%, mFUP time=223 days; dabigatran and edoxaban cohorts had < 100 pairs). Compared to extended treatment, discontinuation of OAC was associated with an increased risk of the composite outcome (HR: 1.54 [95 % confidence intervals (CI): 1.47, 1.63], RD per 1000 person-years 30.99 [27.06, 34.92]), increased risk of recurrent VTE (HR: 5.42 [95% CI, 4.76, 6.17], RD: 25.92 [24.02, 27.83]), lower risk of major bleeding (HR: 0.44 [0.38, 0.50], RD:-10.86 [-12.50, -9.23]), and increased mortality (HR: 1.31 [1.23, 1.40], RD: 16.28 [13.16, 19.40]). This pattern was similar and consistent across the 3 OACs.

Conclusions: Compared to extended OAC treatment, discontinuation of warfarin, apixaban, or rivaroxaban after initial 90-day treatment was associated with an increased risk of recurrent VTE, lower risk of bleeding, and increased mortality.