Oral Anticoagulant Use and the Development of New Cerebral Microbleeds in Cardioembolic Stroke Patients with Atrial Fibrillation


Cerebral microbleeds (CMBs) are a magnetic resonance imaging (MRI) marker for cerebral small vessel disease. Existing CMBs and those that newly develop are associated with the risks of stroke incidence and recurrence. The purpose of the present study was to investigate the association of oral anticoagulant (OAC) use and the development of new CMBs in cardioembolic stroke patients with atrial fibrillation.

Subjects and Methods

We prospectively followed cardioembolic stroke patients with atrial fibrillation who had been hospitalized in the stroke center of our hospital, had been prescribed anticoagulants at discharge, and underwent repeated brain MRI with an interval of at least one year from the baseline MRI. Assessing the presence, number and location of CMBs using T2*-weighted gradient-recalled echo MRI, we used logistic regression models to investigate the associations between OAC use and the incidence of new CMBs. We also examined associations of subsequent stroke with OACs and CMBs during the follow-up.


A total of 81 patients, consisting of 45 patients receiving direct oral anticoagulants (DOACs) and 36 patients receiving warfarin (WF), were analyzed in the present study. Baseline CMBs were observed in 19/81 patients (23.5%) and new CMBs in 18/81 patients (22.2%) on follow-up MRI (median interval, 34 months). Of the 31 new CMBs, 25 (80.6%) developed in the lobar location and 6 (19.4%) in the deep or infratentorial location. New CMBs occurred in 4 patients (10.0%) taking DOACs alone, in 10 patients (35.7%) taking WF alone, in 3 patients (37.5%) taking WF plus antiplatelet agents and in 1 patient (20.0%) taking DOAC plus antiplatelet agent. Regarding location, the new CMBs were the lobar type in 7 of the 10 patients taking WF alone, as well as in 3 of the 4 patients taking DOACs alone. In multivariate analysis, the presence of CMBs at baseline and WF use (vs. DOAC use) were associated with new CMBs (CMB presence at baseline: OR 4.16, 95% CI 1.19-14.44; WF use: OR 3.38, 95% CI 1.02-11.42). The presence of ≥ 2 CMBs at baseline was related to a higher risk of subsequent stroke (OR 7.25, 95% CI 1.01-52.35, P = 0.048).


Our findings suggest that DOAC compared with WF use at discharge is associated with a lower incidence of new CMBs in cardioembolic stroke patients with atrial fibrillation. Further prospective studies in the clinical setting are needed to confirm our exploratory data.