Supplementary Material for: SPARKLE (Subtypes of Ischaemic Stroke Classification System), Incorporating Measurement of Carotid Plaque Burden: A New Validated Tool for the Classification of Ischemic Stroke Subtypes

<b><i>Background:</i></b> Previous classification systems of acute ischemic stroke (Causative Classification System, CCS, of acute ischemic stroke, Trial of Org 10172 in Acute Stroke Treatment, TOAST) established the diagnosis of large artery disease (LAD) based on the presence or absence of carotid stenosis. However, carotid plaque burden is a stronger predictor of cardiovascular risk than stenosis. Our objective was to update definitions of ischemic stroke subtypes to improve the detection of LAD and to assess the validity and reliability of a new classification system: SPARKLE (Subtypes of Ischaemic Stroke Classification System). <b><i>Methods:</i></b> In a retrospective review of clinical research data, we compared three stroke subtype classifications: CCS, TOAST and SPARKLE. We analyzed a random sample of 275 patients presenting with minor stroke or transient ischemic attack (TIA) in an Urgent TIA Clinic in London, Ont., Canada, between 2002 and 2012. <b><i>Results:</i></b> There was substantial overall agreement between SPARKLE and CCS (κ = 0.75), with significant differences in the rate of detection of LAD, cardioembolic and undetermined causes of stroke or TIA. The inter-rater reliability of SPARKLE was substantial (κ = 0.76) and the intra-rater reliability was excellent (κ = 0.91). <b><i>Conclusion:</i></b> SPARKLE is a valid and reliable classification system, providing advantages compared to CCS and TOAST. The incorporation of plaque burden into the classification of LAD increases the proportion of cases attributable to LAD and reduces the proportion classified as being of 'undetermined' etiology.