Ischemic Stroke Etiological Classification:the Agreement Analysis Of CISS,SPARKLE,TOAST

Background and Objective:Currently,there are several ischemic stroke etiological classification systems.One optimal stroke classification system needs to be validity,and high reliability.It can be easily applied among the different stroke research centers.Chinese Ischemic Stroke Sub-classification(CISS)and SPARKLE(Subtypes of Ischemic Stroke Classification System)has been published recently,however,both of them have not been tested for the agreement analysis.As a result,the objective of this study is to assess the level of their agreement among stroke subtype classifications made using the CISS,SPARKLE and TOAST(Trial of Org 10172 in Acute Stroke Treatment),and also analyze the inter-rater reliability of CISS.Methods: Six hundred and twenty three patients with acute ischemic stroke were consecutively recruited from Beijing Tiantan Hospital affiliated Capital Medical University between October 1st,2015 and April 19 th,2016.According to the inclusion criteria and exclusion criteria,four hundred and seven patients were the first acute ischemic stroke which satisfied the requirements for this study.Basing on the sample size of the relative trials,299 patients were consecutively took out as the sample size of this study,which were independently classified etiologies of ischemic stroke and these three etiological classification systems were compared further.Results: Overall,there were substantial agreements among these three classification systems: CISS versus SPARKLE(kappa value=0.684,P ? 0.001),CISS versus TOAST(kappa value=0.615,P ? 0.001),SPARKLE versus TOAST(kappa value=0.675,P ? 0.001).The inter-rater reliability of CISS was excellent(kappa value=0.857,P?0.001).Furthermore,among the three subtype classification systems,the findings of variance analysis of etiological subtypes were as the following.There was significant difference among TOAST,SPARKLE and CISS in Large-artery Atherosclerosis and Undetermined Etiology(P?0.05).No significant differences were found in patients with Cardioembolism and Other Etiology(P?0.05).For the Small Vessel Disease etiology,there was significant difference between CISS and SPARKLE,CISS and TOAST;however,no significant difference were found between SPARKLE and TOAST(P=0.824).Conclusion:There are generally substantial agreements among three ischemic stroke etiological classification systems.CISS is a valid and reliable classification system,which is suitable for different stroke research centers to use and compare their data.