缺血性脑卒中患者颈动脉钙化影响因素的列线图模型构建及验证

Construction and validation of a columnar line graph model based on factors influencing carotid artery calcification in patients with ischaemic stroke

[目的]探讨缺血性脑卒中患者颈动脉钙化影响因素,并构建列线图预测模型,为临床制定防治措施提供参考。[方法]选取500例缺血性脑卒中患者作为研究对象,按照7∶3比例随机分为建模组350例和验证组150例,统计颈动脉钙化发生情况。采用LASSO-Logistic回归方程分析颈动脉钙化影响因素,构建颈动脉钙化风险列线图预测模型。采用受试者工作特征(ROC)曲线、校准曲线分析列线图预测模型区分度、准确度。绘制临床决策曲线(DCA)评价预测模型有效性。[结果]建模组中,与无颈动脉钙化者相比,颈动脉钙化发生者年龄增高17.87%,吸烟史比例增高32.69%,空腹血糖水平增高22.47%,糖化血红蛋白水平增高0.69%,低密度脂蛋白胆固醇(LDLC)水平增高17.84%,尿酸水平增高22.42%,高敏C反应蛋白(hs-CRP)水平增高40.31%,估算的肾小球滤过率(eGFR)水平降低7.04%,差异有统计学意义(P<0.05)。验证组中,与无颈动脉钙化者相比,颈动脉钙化发生者年龄增高17.23%,吸烟史比例增高33.39%,空腹血糖水平增高22.37%,糖化血红蛋白水平增高0.75%,LDLC水平增高17.96%,尿酸水平增高24.44%,hs-CRP水平增高30.81%,eGFR水平降低6.46%,差异有统计学意义(P<0.05);建模组、验证组中以上述因素构建列线图预测模型,预测颈动脉钙化的AUC分别为0.953、0.972,且具有准确性及临床效用性。[结论]年龄增加、有吸烟史及空腹血糖、糖化血红蛋白、LDLC、尿酸、hs-CRP水平升高均为颈动脉钙化发生的独立危险因素,eGFR水平升高为独立保护因素,基于上述因素建立的列线图预测模型对颈动脉钙化发生具有一定预测价值。

Aim To investigate the factors influencing carotid artery calcification in patients with ischemic stroke and to construct a predictive model for columnar plots to provide reference for clinical formulation of prevention and control measures.Methods A total of 500 patients with ischemic stroke were randomly divided into modeling group(350 cases)and verification group(150 cases)according to a ratio of 7∶3,and the incidence of carotid artery calcification was analyzed.LASSO-Logistic regression equation was used to analyze the influencing factors of carotid artery calcification,and the prediction model of carotid artery calcification risk was built.Receiver operating characteristic(ROC)curve and calibration curve were used to analyze the nomogram to predict model differentiation and accuracy.Decision curve analysis(DCA)was drawn to evaluate the validity of the prediction model.Results In the modeling group,compared with those without carotid artery calcification,the age of those with carotid artery calcification increased by 17.87%,the proportion of smoking history increased by 32.69%,the level of fasting blood glucose increased by 22.47%,the level of glycated hemoglobin increased by 0.69%,and the level of low density lipoprotein cholesterol(LDLC)increased by 17.84%,the uric acid level increased by 22.42%,the high sensitivity C-reactive protein(hs-CRP)level increased by 40.31%,and the estimated glomerular filtration rate(eGFR)level decreased by 7.04%,with statistical significance(P<0.05).In the verification group,compared with those without carotid artery calcification,the age of those with carotid artery calcification increased by 17.23%,the proportion of smoking history increased by 33.39%,the level of fasting blood glucose increased by 22.37%,the level of glycated hemoglobin increased by 0.75%,the level of LDLC increased by 17.96%,and the level of uric acid increased by 24.44%,hs-CRP level increased by 30.81%,eGFR level decreased by 6.46%,the difference was statistically significant(P<0.05).In the modeling group and the verification group,the prediction models of carotid artery calcification were constructed based on the above factors,and the AUC for predicting carotid artery calcification was 0.953 and 0.972,respectively,which was accurate and clinically effective.Conclusion Increasing age,smoking history and increased fasting blood glucose,glycated hemoglobin,LDLC,uric acid and hs-CRP levels are independent risk factors for the occurrence of carotid artery calcification,and elevated eGFR levels are independent protective factor.The prediction model based on the above factors has certain predictive value for the occurrence of carotid artery calcification.