3种评分工具对缺血性卒中伴心房颤动患者1年死亡风险评估的验证

Validity Comparison of Three Scores Tool for 1-year Death Risk Assessment in Chinese Ischemic Stroke Patients with Atrial Fibrillation

目的本文旨在通过对临床预后预测模型在中国伴心房颤动的缺血性脑血管病患者中的预测效度的评估,寻找能够准确预测预后结局的评分工具,更好地指导临床医疗决策。方法从中国国家卒中登记(China National Stroke Registry,CNSR)研究中选取连续入组的缺血性卒中住院患者为本研究的研究人群。收集患者人口学信息、临床特点及用药信息等基本资料,随访时间为1年,预后结局包括卒中复发及死亡。用χ2检验比较缺血性卒中伴或不伴心房颤动患者基线资料,P<0.01为差异具有统计学意义。采用C值表示的受试者工作特征曲线下面积(area under the curve,AUC)来评价缺血性卒中风险预测评分(ischemic stroke predictive risk score,IScore),住院前合并症、意识水平、年龄和局灶性神经功能缺损(preadmission comorbidities,level of consciousness,age,and neurologic deficit,PLAN)评分和洛桑卒中量表(acute stroke registry and analysis of Lausanne,ASTRAL)评分对1年死亡风险的预测效度。按照各个评分的分层标准计算各层人群死亡和卒中复发的事件发生率。结果从CNSR数据库中选取12 415例完成1年随访的缺血性卒中患者作为本研究的研究人群。其中,10 847例(87.37%)患者不伴心房颤动,1568例(12.63%)患者伴心房颤动。总研究人群1年全因死亡的事件发生率为13.4%,伴心房颤动患者的1年死亡率为34.6%,不伴心房颤动患者的1年死亡率为10.3%。随着各评分分数的增高,死亡事件发生率逐渐增加。IScore对于伴心房颤动的缺血性卒中患者预测死亡的C值为0.784,PLAN评分为0.769,ASTRAL评分为0.793。结论 IScore、PLAN评分和ASTRAL评分可针对1年死亡风险对缺血性卒中患者进行初步分层。各评分工具对于中国缺血性卒中伴心房颤动的患者1年死亡风险有较高的预测能力。

Objective To compare the predictive value of three scores tool for 1-year death risk in Chinese ischemic stroke patients with atrial fibrillation（AF）, to find an optimal scale so as to provide guidance for making clinical decisions.Methods Patients with acute ischemic stroke（AIS） selected from China National Stroke Registry（CNSR） were divided into AF group and non-AF group. Demographic information, clinical characteristics and drug information were collected. The primary outcome was 1-year stroke recurrence and death. Using chi-square test to compare the baseline data of two groups. The predictive value of ischemic stroke predictive risk score（IScore）, preadmission comorbidities, level of consciousness, age, and neurologic deficit（PLAN） score and acute stroke registry and analysis of Lausanne（ASTRAL） score for 1-year death risk of two groups were compared using C statistic（expressed by area under the curve）.Results A total of 12 415 AIS patients who completed 1-year follow-up from CNSR were included in this study. Of all 12 415 patients, 10 847（87.37%） were non-AF patients and 1568（12.63%） were AF patients. The total 1-year all-cause mortality of all subjects was 13.4%, while the 1-year mortality of AF group and non-AF group were 34.6% and 10.3%, respectively. The risk of death increased with elevation of the scores. The C statistic of IScore, PLAN and ASTRAL scores predicting 1-year death risk for AIS patients with AF were 0.784, 0.769 and 0.793, respectively.Conclusions IScore, PLAN and ASTRAL scores could preliminarily stratify the risk of 1-year death for AIS patients with AF, and all three scores tool had a high predictive value of 1-year risk of death for such patients.