基于血清生物标志物的卒中后痴呆发生风险评估模型研究  

Study on risk assessment model of post-stroke dementia based on serum biomarkers

作　　者：李小杏 焦冰 方桦 李坤彬 姚先丽 吴志远 孙平鸽 王留向 LI Xiao-xing;JIAO Bing;FANG Hua;LI Kun-bin;YAO Xian-li;WU Zhi-yuan;SUN Ping-ge;WANG Liu-xiang(Department of Neurorehabilitation,Zhengzhou Central Hospital(Zhengzhou Central Hospital Affiliated to Zhengzhou University),Zhengzhou 450000,Henan,China)

机构地区：[1]郑州市中心医院(郑州大学附属郑州市中心医院)神经康复科,河南郑州450000

出　　处：《广东医学》2023年第12期1548-1553,共6页Guangdong Medical Journal

基　　金：河南省医学科技攻关计划项目(2018020808)。

摘　　要：目的 基于血清生物标志物的卒中后痴呆发生风险评估模型研究。方法 这是一项纵向的预测模型开发研究,包括神经康复科病区2020年10月到2022年10月的数据。从患者入院后1 d内采集的外周静脉血样本中确定实验室检查结果。卒中发作后6个月的蒙特利尔认知评估(MoCA)评分<22确定为卒中后认知障碍(post-stroke cognitive impairment, PSCI)。基于多元logistic模型,生成列线图模型。结果 在完成随访的114例受试者中,61例受试者(53.5%)被确定为患有PSCI。PSCI(MoCA<22)组和非PSCI(MoCA≥22)组在年龄、性别、教育程度、高血压、美国国立卫生研究院卒中量表(NIHSS)评分、Fazekas评分、颅内动脉粥样硬化性狭窄(ICAS)≥50%、皮质梗死、尿酸、同型半胱氨酸(Hcy)、基质金属蛋白酶-9(MMP-9)、组织金属蛋白酶抑制剂-1(TIMP-1)、中性粒细胞/淋巴细胞比值(NLR)、淋巴细胞/单核细胞比值(LMR)差异有统计学意义(P<0.05),并被纳入初始回归模型。在二元logistic回归模型(LR方法)中,7个潜在预测因子被考虑用于模型开发:年龄、Fazekas评分、皮质梗死、MMP-9、TIMP-1、NLR和LMR。开发模型的受试者工作特征(ROC)曲线下面积(AUC)为0.837,敏感度为71.2%,特异度为84.5%。缺血性卒中患者PSCI预测概率列线图的校准曲线在该队列中表现出良好的一致性。1 000个bootstrap样本的结果估计AUC为0.814,表明该模型具有良好的区分能力。结论 本研究建立了一个易于使用的列线图模型来预测缺血性卒中后的认知能力。Objective To develop a risk assessment model for post-stroke dementia(PSD)based on serum biomarkers.Methods This was a longitudinal predictive model development study conducted in the Neurorehabilitation Department from October 2020 to October 2022.Laboratory test results were obtained from peripheral venous blood samples collected within one day of patient admission.Post-stroke cognitive impairment(PSCI)was defined as a Montreal Cognitive Assessment(MoCA)score of<22 at 6 months after stroke onset.A logistic model was generated using multivariate logistic regression.Results Among the 114 individuals who completed follow-up,61 subjects(53.5%)were identified as having PSCI.The PSCI(MoCA<22)group and the non-PSCI(MoCA≥22)group showed significant differences in age,gender,education level,hypertension,NIHSS score,Fazekas score,ICAS≥50%,cortical infarction,uric acid,homocysteine(Hcy),matrix metalloproteinase-9(MMP-9),tissue inhibitor of metalloproteinase-1(TIMP-1),neutrophil-to-lymphocyte ratio(NLR),and lymphocyte-to-monocyte ratio(LMR)(P<0.05),which were included in the initial regression model.In the binary logistic regression model(LR method),seven potential predictor variables were considered for model development:age,Fazekas score,cortical infarction,MMP-9,TIMP-1,NLR,and LMR.The area under the receiver operating characteristic(ROC)curve(AUC)of the developed model was 0.837,with a sensitivity of 71.2%and specificity of 84.5%.The calibration curve of the PSCI prediction probability nomogram for ischemic stroke patients demonstrated good consistency in this cohort.The estimated AUC from 1000 bootstrap samples was 0.814,indicating that the model had good discrimination ability.Conclusion This study established an easy-to-use nomogram model for predicting cognitive impairment after ischemic stroke.

关 键 词：卒中 血清生物标志物 卒中后认知障碍 列线图 

分 类 号：R743[医药卫生—神经病学与精神病学] R749[医药卫生—临床医学]