Supplementary Material for: Validation of the Canadian TIA Score to Predict Subsequent Stroke Risk in Chinese TIA Patients

Background and Objectives: The Canadian TIA Score has been verified as a good predictive tool for subsequent stroke risk in Caucasian patients; however, it had insufficient external verification of other races. We aimed to validate the Canadian TIA Score in Chinese patients and compared it with ABCD2 for subsequent stroke risk after transient ischemic attack (TIA). Methods: The Third China National Stroke Registry (CNSR-III) was a nationwide, multicenter prospective registry recruiting consecutive patients with acute ischemic stroke or TIA within 7 days of the onset from August 2015 to March 2018. The Canadian TIA Score was verified in patients diagnosed with TIA from the CNSR-III (N = 1,184). The outcomes were subsequent stroke at 7 days, 14 days/discharge, 3 months, and 1 year. Outcomes were recorded by face-to-face assessment or telephone interview. The prognostic performance of the scoring system was assessed by the area under the receiver operator characteristic curve (AUC). Results: Of 1,184 TIA patients (mean [IQR] age, 61.00 [53.00–69.00] years; 413 women [34.88%]), there were 40 patients (3.38%) having subsequent stroke within 7 days, 45 (3·80%) within 14 days/hospitalization, 66 (5·57%) within 3 months, and 100 (8·45%) within 1 year. The Canadian TIA Score (AUC 0 63–0·68) seemed to be a better prognostic score of stroke risk than the ABCD2 score (AUC 0·61–0·62), although no significant differences were noted. In the subgroup of atypical TIA, the Canadian TIA Score showed significantly stronger predictivity than the ABCD2 score within 7 days (0.80 [0.62–0.98] vs. 0.52 [0.30–0.73]; difference in AUC, 0.28 [0.03–0.53]; p, 0.026), and marginal significantly stronger predictivity within 1 year (0.71 [0.61–0.80] vs. 0.58 [0.48–0.68]; difference in AUC, 0.12 [−0.01 to 0.25]; p, 0.06). Conclusion: The Canadian TIA Score might be a better prognostic score than the ABCD2 score for post-TIA stroke risk, especially in patients with atypical TIA.## 研究背景与研究目的 加拿大短暂性脑缺血发作评分（Canadian TIA Score）已被证实可作为高加索人群患者后续卒中风险的良好预测工具，但针对其他种族的外部验证仍较为不足。本研究旨在验证中国人群患者中的加拿大短暂性脑缺血发作评分，并将其与ABCD2评分对比，以评估短暂性脑缺血发作（transient ischemic attack, TIA）后患者的后续卒中风险。 ## 研究方法 第三次中国国家卒中登记研究（Third China National Stroke Registry, CNSR-III）是一项全国性多中心前瞻性登记研究，于2015年8月至2018年3月期间，纳入发病7天内的急性缺血性卒中或短暂性脑缺血发作的连续入组患者。本研究从CNSR-III数据库中筛选出确诊为短暂性脑缺血发作的患者共1184例，对加拿大短暂性脑缺血发作评分进行验证。研究结局为患者在卒中后7天、14天/出院时、3个月及1年时的新发卒中事件，通过面对面评估或电话访谈完成结局数据收集。采用受试者工作特征曲线下面积（area under the receiver operator characteristic curve, AUC）评估该评分系统的预后预测效能。 ## 研究结果 本研究共纳入1184例短暂性脑缺血发作患者，平均年龄（四分位数间距）为61.00[53.00~69.00]岁，其中女性413例，占比34.88%。7天内新发卒中患者40例（3.38%），14天/住院期间新发卒中患者45例（3.80%），3个月内新发卒中患者66例（5.57%），1年内新发卒中患者100例（8.45%）。加拿大短暂性脑缺血发作评分的受试者工作特征曲线下面积为0.63~0.68，其卒中风险预测效能似乎优于ABCD2评分（0.61~0.62），但组间差异未达到统计学显著性。在非典型短暂性脑缺血发作亚组中，加拿大短暂性脑缺血发作评分在7天内的预测效能显著优于ABCD2评分（0.80[0.62~0.98] vs 0.52[0.30~0.73]；曲线下面积差值为0.28[0.03~0.53]，p=0.026）；在1年随访期内，其预测效能呈边缘显著更优（0.71[0.61~0.80] vs 0.58[0.48~0.68]；曲线下面积差值为0.12[-0.01~0.25]，p=0.06）。 ## 研究结论 在短暂性脑缺血发作后卒中风险的预后评估中，加拿大短暂性脑缺血发作评分或许优于ABCD2评分，尤其适用于非典型短暂性脑缺血发作患者。