Supplementary Material for: Long-Term Outcomes of Ischemic Stroke of Undetermined Mechanism: A Population-Based Prospective Cohort

<b><i>Background and Purpose:</i></b> Little is known about the short- and long-term outcomes of ischemic stroke of undetermined mechanism (ISUM). <b><i>Methods:</i></b> Subjects were recruited from the Mashhad Stroke Incidence Study. Ischemic stroke (IS) was classified on the basis of the TOAST criteria. We further categorized patients with ISUM into ISUM<sub>neg</sub> (negative clinical/test results for large artery, small artery) and ISUM<sub>inc</sub> (incomplete investigations). Cox proportional hazard models and the competing-risk regression model were used to compare 1 and 5 years mortality (all-causes) and recurrent rate among IS subtypes. <b><i>Results:</i></b> Overall, 1-year mortality was higher in those with ISUM<sub>inc</sub> than in ISUM<sub>neg</sub> (adjusted hazard ratio [aHR] 1.6, 95% CI 1.01–2.8; <i>p</i> = 0.04) and in other stroke subtypes. Cardioembolic stroke was associated with the greatest risk of stroke recurrence at one year (aHR 4.9, 95% CI 1.8–12.9; <i>p</i> = 0.001) and 5 years (HR 2.1, 95% CI 1.1–3.7; <i>p</i> = 0.01) as compared to ISUM<sub>neg</sub>. <b><i>Conclusions:</i></b> The classification of ISUM as a single group may lead to over- or underestimation of mortality and recurrence in this major category of IS. A better definition of ISUM is necessary to predict death and recurrence accurately.