Use of anticoagulant therapy and cerebral microbleeds: a systematic review and meta-analysis (2783)

Objective: We aimed to perform a systematic review and meta-analysis to explore the potential association between anticoagulant use and occurrence of microbleeds (CMBs) in stroke and stroke-free individuals. Background: Anticoagulant therapy increases the risk that CMBs progress to intracerebral hemorrhage, but whether the therapy increases risk of CMB occurrence is unclear. Design/Methods: We searched observational studies in PubMed, Ovid EMBASE, and Cochrane Library from their inception until September 2019. We calculated the pooled odds ratio (OR) and 95% confidence interval (CI) for the prevalence and incidence of CMBs in anticoagulant users relative to non-anticoagulant users. Results: Forty-seven studies with 25,245 participants were included. The pooled analysis showed that anticoagulant use was associated with CMB prevalence (OR 1.54, 95% CI 1.26–1.88). The association was observed in subgroups stratified by type of participants: stroke-free, OR 1.86, 95% CI 1.25–2.77; ischemic stroke/transient ischemic attack, OR 1.33, 95% CI 1.06–1.67; and intracerebral hemorrhage, OR 2.26, 95% CI 1.06–4.83. Anticoagulant use was associated with increased prevalence of strictly lobar CMBs (OR 1.68, 95% CI 1.22–2.32) but not deep/infratentorial CMBs. Warfarin was associated with increased CMB prevalence (OR 1.64, 95% CI 1.23–2.18), but novel oral anticoagulants were not. Anticoagulant users showed higher incidence of CMBs during long-term follow-up (OR 1.72, 95% CI 1.22–2.44). Conclusions: Anticoagulant use is associated with higher prevalence and incidence of CMBs. This association appears to depend on location of CMBs and type of anticoagulants. More longitudinal investigations with adjustment for confounders are required to establish the causality. Disclosure: Dr. Cheng has nothing to disclose. Dr. Wang has nothing to disclose. Dr. Song has nothing to disclose. Dr. Qiu has nothing to disclose. Dr. Ming has nothing to disclose.