Abstract Body: Background: ICH in the setting of FXa inhibitors is associated with a poor prognosis. Andexanet, an inactive recombinant form of human FXa, rapidly reverses the anticoagulant effect of FXa inhibitors. Compared to usual care, andexanet decreases the proportion of patients with hematoma expansion ≥35% by 12.1% but is associated with a 4.6% increase in the proportion with thrombotic events (TE) within 30 days. We sought to explore the effects of restarting anticoagulation on the risk of TE after anticoagulation reversal.
Methods: We pooled andexanet treated patients in the single arm ANNEXA-4 trial of patients with FXa inhibitor-associated major bleeding with those from the randomized controlled ANNEXA-I trial of patients with ICH to maximize the number of events. Baseline variables were collected at study entry, and thrombotic events were collected from study entry to day 30. Thrombotic events were characterized as arterial (myocardial infarction [MI], ischemic stroke, and systemic embolism) or all (arterial and venous thromboembolism [VTE]). The first use of an anticoagulant after entry was categorized as prophylactic or therapeutic. A stratified Cox model including anticoagulation start as a time-dependent variable was constructed to examine the association with thrombotic events.
Results: The pooled dataset included 741 patients, of whom 80% had an intracranial site of bleeding. There were 77 participants with thrombotic events, of whom 58 (75%) had arterial events including 40 with stroke and 21 with MI. The median (IQR) time to arterial events was 3.5 (2.0, 12.0) days from study entry, while the median (IQR) time between study entry and VTE was 14.0 (10.0, 20.0) days. Anticoagulation was started in 520 (70.2%). After adjusting for age, sex, prior MI, stroke, VTE, baseline anti-FXa level, high dose eligibility, and AF, restarting anticoagulation was associated with a reduced risk of all thrombotic events (HR 95% CI 0.53 [0.28, 0.99]) and arterial thrombotic events (HR 95% CI 0.34 [0.15, 0.76]). These findings were consistent in sensitivity analyses examining arterial thrombotic events while restricting the analysis to prophylactic anticoagulation (HR 95%CI 0.55 [0.24, 1.24]).
Conclusion: Treatment with prophylactic or therapeutic anticoagulation is associated with a reduced risk of thrombotic events in patients with FXa inhibitor-associated major bleeding receiving andexanet.