Abstract: Objective　To evaluate the value of TriAGe+ score in the diagnosis of stroke in patients with dizziness.Methods　A total of 1 437 patients with dizziness who received treatment from the Suzhou Hospital Affiliated to Anhui Medical University from February 2014 to August 2017 were enrolled.According to the final diagnosis，the patients were divided into two groups：stroke group （n=137，accounting for 9.5%） and non-stroke group （n=1 300）.The diagnostic value of TriAGe+ score for stroke in dizziness patients was determined by ROC curve and compared with ABCD2 score.According to the TriAGe+ score，those scored 10-17，8-9，5-7，0-4 were assigned to very high-risk group，high-risk group，medium-risk group and low-risk group，respectively.Comparisons were made between the 4 risk groups in terms of the incidence of stroke and between stroke and non-stroke groups in regard to clinical data.Stepwise Logistic regression analysis was performed for investigating the associated factors for stroke.Results　Patients with stroke and those without showed significant differences in the distribution of sex，blood pressure levels at admission，duration of dizziness，and type of dizziness，prevalence of functional limb weakness or speech impairment，prevalence of inducing factors for dizziness，prevalence of brainstem or cerebellar dysfunction，dizziness history percentages of labyrinth disorder or vestibular disease history，diabetes history，hypertension history，and atrial fibrillation history （P<0.05），except in the distribution of age （P>0.05）.Logistic regression analysis showed that sex，blood pressure levels at admission，functional limb weakness or speech impairment，duration of dizziness，inducing factors for dizziness，type of dizziness，brainstem or cerebellar dysfunction，dizziness history labyrinth disorder or vestibular disease history，diabetes history，hypertension history，atrial fibrillation history were associated factors for stroke in dizziness patients （P<0.05）.The incidence of stroke in very high-risk group，high-risk group，medium-risk group and low-risk group was 65.5% （57/87），27.0% （51/189），5.8% （23/394），0.8% （6/767），respectively，with significant differences （P<0.05），and the difference in terms of incidence of stroke between any two groups was significant（P<0.008 3）.For the diagnosis of stroke，the AUC of TriAGe+ score was larger than that of the ABCD2 score 0.890〔95%CI（0.866，0.914）〕vs 0.806〔95%CI（0.769，0.843）〕 （P<0.001）.When the TriAGe+ score was 10，its sensitivity was 0.834，specificity was 0.727，and Youden's index was 0.561.Conclusion　TriAGe+ score is contributive to quickly identifying stroke in dizziness patients.

Key words: Stroke, Dizziness, TriAGe+ score, ABCD2 score, Diagnosis

摘要： 目的　探讨TriAGe+评分对头晕患者脑卒中的诊断价值。方法　选取2014年2月—2017年8月于安徽医科大学附属宿州医院就诊的1 437例以头晕为主诉的患者为研究对象，根据最终是否确诊为脑卒中分为非脑卒中组1 300例和脑卒中组137例。根据TriAGe+评分结果分为4组：极高危组（10~17分）、高危组（8~9分）、中危组（5~7分）及低危组（0~4分），比较4组患者脑卒中发生率；比较脑卒中组和非脑卒中组临床资料的不同；采用Logistic逐步回归分析头晕患者发生脑卒中的影响因素；采用受试者工作特征（ROC）曲线比较TriAGe+和ABCD2评分对头晕患者脑卒中的诊断价值。结果　1 437例头晕患者最后确诊为脑卒中者137例（9.5%）。非脑卒中组和脑卒中组性别、入院时血压、有无肢体无力或言语障碍、头晕持续时间、头晕有无诱发因素、头晕类型、脑干或小脑功能障碍、头晕及迷路或前庭疾病病史、糖尿病病史、高血压病史、心房颤动病史比较，差异有统计学意义（P<0.05）；年龄比较，差异无统计学意义（P>0.05）；Logistic逐步回归分析结果显示，性别、入院时血压、有无肢体无力、有无言语障碍、头晕持续时间、头晕有无诱发因素、头晕类型、脑干或小脑功能障碍、头晕及迷路或前庭疾病病史、糖尿病病史、高血压病史、心房颤动病史是头晕患者发生脑卒中的影响因素（P<0.05）。极高危组（10~17分）、高危组（8~9分）、中危组（5~7分）及低危组（0~4分）脑卒中发生率分别为65.5%（57/87）、27.0%（51/189）、5.8%（23/394）及0.8%（6/767），4组间比较，差异有统计学意义（P<0.05）；4组间两两比较，差异均有统计学意义（P<0.008 3）。TriAGe+评分、ABCD2评分诊断头晕患者发生脑卒中的ROC曲线下面积分别为0.890〔95%CI（0.866，0.914）〕、0.806〔95%CI（0.769，0.843）〕，两者比较，差异有统计学意义（P<0.001）。TriAGe+评分为10分时，对应的灵敏度为0.834，特异度为0.727，约登指数为0.561。结论　TriAGe+评分有助于快速识别主诉为头晕的脑卒中患者。

关键词: 卒中, 头晕, TriAGe+评分, ABCD2评分, 诊断