Borderzone Infarction Predicts Recurrence In Patients With Intracranial Atherosclerosis: A Meta-analysis

Background: Patients with symptomatic intracranial stenosis (ICAS) face elevated risks of recurrent cerebrovascular events (RCE) despite optimal medical therapy. Borderzone infarcts (BZI) indirectly correlate with impaired distal perfusion, a known mechanism of recurrence. Studies assessing associations between borderzone infarcts and recurrence rates are observational and have relatively small sample size. We therefore performed a meta-analysis of published studies investigating this association. Methods: We performed a Medline and Web of Science search using the key words (Intracranial Atherosclerosis OR Intracranial Stenosis) AND (Borderzone OR Infarct Pattern) to identify studies reporting associations between index infarct pattern and RCE, defined as recurrent ischemic stroke or neurological deterioration, or new infarct on follow up neuroimaging in patients with symptomatic ICAS. We included relevant papers and scientific abstracts with more than 20 patients included. For all outcome events we calculated the corresponding risk ratios (RRs) and 95% confidence intervals (95% CI). Results: We identified 178 studies using Web of Science and 384 studies using Medline with only 6 studies (591 patients) meeting our inclusion criteria (2 prospective and 4 retrospective). The weighted proportion of patients with BZI was 32.5% (28.7%-36.4%). During a follow-up period of 7-950 days, 33.1% (26.3%-40.5%) of patients with BZI had RCE and 63.6% (30.8%-89.1%) had new infarction on a 6-8 week follow up brain MRI. In meta-analysis, BZI was associated with increased rates of RCE (RR 2.40 95% CI 1.71-3.37) and new infarct(s) on follow up brain MRI (RR 2.55 95% CI 1.31-4.94). The findings were unchanged when the analysis was limited to 90-day RCE risks only (RR 2.22 95% 1.49-3.29). Conclusions: BZI are associated with over 2-fold increased risk of RCE and recurrent infarct in patients with symptomatic ICAS. Prospective studies are needed to validate these findings.