Anticoagulation and Risk of Stroke Recurrence in Patients with Embolic Stroke of Undetermined Source Having No Potential Source of Embolism.

BACKGROUND

This study aimed to determine whether use of oral anticoagulants (OACs) was associated with a reduced risk of recurrent stroke compared with use of antiplatelets (APs) in patients with embolic stroke of undetermined source (ESUS) having no potential source of embolism.

METHODS

Of 8,790 patients with acute ischemic stroke registered at 7 centers in the Fukuoka Stroke Registry from June 2007 to May 2017, we included 681 patients (mean age 69.7 [SD 14.1] years, 48.3% men) who experienced ESUS without a potential source of embolism and received OAC alone or AP alone. We estimated hazard ratios (HRs) and 95% confidential intervals (CIs) of recurrent ischemic stroke or any stroke after discharge using a Cox proportional hazards model and Fine and Gray model.

RESULTS

During a mean follow-up of 3.4 (SD 1.7) years, event rates of recurrent ischemic stroke were 4.4 per 100 person-years in 489 patients treated with AP and 2.0 per 100 person-years in 192 patients treated with OAC. OAC use was associated with a reduced risk of recurrent ischemic stroke, even after adjusting for potential confounding factors (multivariable-adjusted HR [95% CI], 0.42 [0.23-0.80]) and when additionally considering death as a competing risk (0.45 [0.24-0.85]). The reduced risk of recurrent ischemic stroke was still observed in patients treated with OAC (0.32 [0.15-0.67]) in reference to propensity score-matched patients treated with AP. These associations were maintained for all types of stroke, including ischemic and hemorrhagic stroke.

CONCLUSIONS

This nonrandomized observational study suggests that anticoagulation therapy might be associated with a reduced risk of recurrent stroke compared with antiplatelet therapy in patients with ESUS in whom no potential source of embolism was identified. Further study should be performed in consideration of a potential source of embolism even in patients with ESUS.