Abstract 100: Right Insular Infarction is an Independent Predictor of Mortality in Ischemic Stroke

Abstract

Introduction: The association between insular infarction and mortality has often been described. However, whether this is simply due to higher lesion volumes is still controversial. Hypothesis: We hypothesized that there is an association between insular infarction and mortality independent of lesion volume. Methods: We included consecutive stroke patients between 01.09.2008 and 11.11.2012 from the 1000Plus data base with an acute ischemic lesion on diffusion-weighted imaging on day one and a completed 90 days follow-up. Insular location of the infarction was determined using the Stroke Lesion Atlas (SLA). The SLA is an in-house developed spatial database for analyses of imaging studies (Figure). In multivariate regression analyses of mortality and insular infarcts we adjusted for age, lesion volume, atrial fibrillation, NIHSS, capsular infarcts, and left hemispheric infarcts. Results: We included 718 patients. Out of 165 patients with insular infarcts 20 patients died; among the 553 patients with lesions outside the Insula 16 patients died (mortality 12% vs. 3%; p<.001). In analyses adjusting for all mentioned confounders insular infarction was not an independent predictor of mortality. After exclusion of NIHSS insular infarcts were independently associated with mortality (OR=3.003, CI 1.41 - 6.38, p=.004). Right insular infarction was an independent mortality predictor adjusted for all confounders including the NIHSS (OR=2.793, CI 1.27 - 6.15, p=.011). Conclusion: Right insular involvement is an independent predictor of mortality in ischemic stroke and may be used to improve identification of patients at risk. Figure. Lesion overlap of patients who survived (A) or died (B). Red indicates voxels involved in at least 15% of patients. Deceased patients show a maximum lesion overlap in the right Insula.