摘要
目的:开发脑卒中后血管性痴呆预测模型及风险评估工具。方法:采用双向队列研究设计,收集2019年1月-2021年10月,北京中医药大学东方医院医院信息系统中的230例脑卒中入院患者的临床资料,对脑卒中3个月后的患者进行随访,结局指标为血管性痴呆事件发生情况。采用Logistic回归方法进行预测变量的指标筛选,结合脑卒中领域专家意见确定模型变量,最终年龄、脑梗死史、便秘、中枢性言语障碍、BI(重度依赖)为脑卒中后血管性痴呆的危险因素(OR>1)。抗血小板聚集药物、降脂药物、NIHSS>4分为脑卒中后血管性痴呆的保护因素(OR<1)。使用多因素Logistic回归模型纳入8个影响因素构建预测模型,使用AUC值评价模型的预测能力。结果:156例患者患有血管性痴呆,多因素Logistic模型的预测工具计分(Score)算式:Score=2.3×年龄80岁以上(是=1,否=0)+1.7×年龄71~80岁(是=1,否=0)+1.5×年龄61~70岁(是=1,否=0)+0.8×脑梗死史(是=1,否=0)+0.8×便秘(是=1,否=0)+1.1×中枢性言语障碍(是=1,否=0)-0.7×抗栓药物使用(是=1,否=0)-1.4×降脂药物使用(是=1,否=0)+1.5×BI分类(重度依赖=2,非重度依赖=1)-0.1×NIHSS分类(非轻型卒中=2,轻型卒中=1),模型的AUC值为0.8347,模型截断值为0.669,Youden指数值为0.499,灵敏度为0.769,特异度为0.730。结论:建立的脑卒中后血管性痴呆预测模型,可较好区分脑卒中后血管性痴呆人群,为卒中后血管性痴呆预防及评估提供参考。
Objective:To develop a predictive model and risk assessment tool for vascular dementia in stroke patients.Methods:Using a two-way Cohort study design,from January 2019 to October 2021,we collected the clinical data of 230 stroke patients admitted to the Hospital information system of Dongfang Hospital of Beijing University of Chinese Medicine.Follow-up was performed three months after stroke,and the outcome was vascular dementia.Using Logistic regression method for the selection of predictive variable indicators,combined with expert opinions in the field of stroke,to determine the modeling variables.Age,history of cerebral infarction,constipation,Post-stroke aphasia,BI(heavy dependence)are the risk factors of vascular dementia due to stroke(OR>1).Anti platelet aggregation drugs,lipid-lowering drugs and NIHSS>4 were classified as the protective factors of vascular dementia in stroke(OR<1).The multivariate Logistic regression model was used to incorporate 8 influencing factors to construct a prediction model.The area under the Receiver operating characteristic(AUC value)is used to evaluate the early-warning capability of the model.Results:156 patients suffer from vascular dementia,and the predictive tool score of the multivariate logistic model is calculated as Score=2.3×Age 80+(Yes=1,No=0)+1.7×Age 71-80 years old(yes=1,no=0)+1.5×Age 61-70 years old(yes=1,no=0)+0.8×History of cerebral infarction(yes=1,no=0)+0.8×Constipation(Yes=1,No=0)+1.1×Post-stroke aphasia(Yes=1,No=0)-0.7×Antithrombotic drugs(yes=1,no=0)-1.4×Lipid-lowering drugs(yes=1,no=0)+1.5×BI classification(heavy dependence=2,non severe dependence=1)-0.1×NIHSS classification(non mild stroke=2,mild stroke=1),the AUC value of the model is 0.8347,the truncation value of the model is 0.669,the Youden index value is 0.499,the sensitivity is 0.769,and the specificity is 0.730.Conclusion:The predictive model for vascular dementia after stroke can effectively distinguish the population of vascular dementia after stroke,providing a certain reference ability for the prevention and evaluation of vascular dementia after stroke.
作者
高阳
贺立娟
王嘉麟
唐嫄
王耀华
孙冉冉
李悦
王祁
张文
张培
张华
张东
吕慧淼
Gao Yang;He Lijuan;Wang Jialin;Tang Yuan;Wang Yaohua;Sun Ranran;Li Yue;Wang Qi;Zhang Wen;Zhang Pei;Zhang Hua;Zhang Dong;Lyu Huimiao(Dongfang Hospital,Beijing University of Chinese Medicine,Beijing 100078,China)
出处
《亚太传统医药》
2025年第5期65-71,共7页
Asia-Pacific Traditional Medicine
基金
北京中医药大学基本科研业务费项目新教师启动基金项目(2021-JYB-XJSJJ-072)。
作者简介
高阳(1991-),男,北京中医药大学东方医院主治医师,研究方向为脑病康复;通讯作者:唐嫄(1987-),女,北京中医药大学东方医院主治医师,研究方向为脑病康复。E-mail:yirenrubing@126.com。
