Direct oral anticoagulants (DOACs) are first line therapy for stroke prevention in NVAF patients. They were reported to be as effective as vitamin K antagonists (VKA) for stroke prevention in high-risk NVAF patients, with lower bleeding events. Compared to VKA, there is no need of regular anticoagulation monitoring. Yet, we still estimate 2 to 5% of major bleedings (MB) and up to 15% of clinically relevant non-major bleedings (CRNMB).

The goal of this study was to investigate specific anti-Xa activity (AXA) measurement as a mean to predict bleeding complications in NVAF patients. We also determined anti-Xa overdose risk factors.

We performed a prospective observational study from December 2016 to April 2020. We took blood samples on patients the day of inclusion to measure AXA. We considered as high responders, patients to whom AXA plasma levels were superior to the upper limit of the expected plasma level ranges from the European Heart Rhythm Association recommendations.

MB, CRNMB from time to inclusion to April 2020 were collected. We then compared group of high responders and non-high responders.

A total of 146 patients were analysed. Mean follow up was 20 months. Among them, 24 (16,4%) had MB or CRNMB during the follow up. 36 (24,6%) patients were high responders among which 10 had relevant bleedings. High responders showed a higher risk of bleeding in univariate analyses (P=0.034). After adjustment in multivariate analyses, only HASBLED had an increased risk of bleeding with an OR of 2.88 (1.74–5.12); probably due to a small number of patients and short time of follow up. We also showed that women and concomitant use of amiodarone were responsible for increasing AXA plasma level.

AXA plasma level might be predictive of bleeding. Further studies with higher number of patients should be carried out to prove it. Female and amiodarone are both overdose risk factors.