Meta-Analysis of Secondary Prevention of Cryptogenic Stroke

Background: Cryptogenic stroke and embolic stroke of undetermined source (ESUS) are a frequently encountered categories of ischemic stroke with an uncertain cause. Methods: We analyzed all randomized clinical trials (RCTs) that evaluated antithrombotic therapy and patent foramen ovale (PFO) closure in cryptogenic stroke and/or ESUS. We calculated aggregate hazard ratios (HRs) using direct and network meta-analysis. The primary outcome was recurrent ischemic stroke. Results: Ten RCTs with a total of 16,876 patients, randomizing 32,143 patient-years of follow-up (mean age 61.2 ± 13.5 with 39.2% female) were identified. Anticoagulation therapy was associated with significantly reduced recurrent ischemic stroke compared with antiplatelet therapy (HR = 0.66; [95% confidence interval (CI) = 0.47–0.94]). Meta-regression analysis showed significantly reduced recurrent stroke with longer duration of therapy, and significantly increased events with advanced age. Significant interactions were observed based on the presence of PFO, stroke type, and anticoagulant used. There were no significant differences with regard to the composite ischemic outcome, transient ischemic attack, any death, major bleeding, or intracranial bleeding. In the network meta-analysis, compared to antiplatelet therapy, warfarin (HR = 0.31; [95% credible interval (CrI) = 0.12–0.68]) and PFO closure (HR = 0.14; 95% CrI = 0.05–0.31]) were associated with significantly reduced recurrent ischemic stroke. In rank order, PFO closure was associated with the best outcome, followed by warfarin. Conclusions: Among patients with cryptogenic stroke, anticoagulation therapy, as compared with antiplatelet therapy is associated with lower rate of recurrent stroke. The small sample size and high heterogeneity with regards to bleeding outcomes require further large trials. In patients with PFO, closure and warfarin are associated with the lowest rates of recurrent stroke.