摘要
目的研究脐血pH值、胱抑素-C(Cys-C)、高迁移率族蛋白B1(HMGB1)与新生儿窒息(NA)病情转归的相关性。方法选取2020年9月至2022年6月四川大学华西第二医院眉山市妇女儿童医院收治的151例NA患儿,根据病情转归不同分为不良组、良好组,对患儿临床资料和脐血pH值、Cys-C、HMGB1进行单因素分析,运用随机森林算法对病情转归变量的重要性进行排序和筛选,多因素Logistic回归分析病情转归的相关影响因素,R语言绘制预测病情转归的列线图模型,受试者工作特征(ROC)曲线评价列线图的预测效能。结果不良组脐带绕颈、宫内窘迫、羊水粪染、窒息后肾损伤患儿占比及脐血Cys-C、HMGB1水平高于良好组,1 min、5 min Apgar评分,脐血pH值低于良好组(P<0.05);随机森林算法显示,变量重要性排序依次是1 min Apgar评分、pH值、HMGB1、Cys-C、宫内窘迫、羊水粪染、脐带绕颈、5 min Apgar评分、窒息后肾损伤;当变量数为6时,袋外数据错误率最低,故将重要性排序前6位变量作为NA患儿病情转归的预测因素纳入多因素Logistic回归分析,显示1 min Apgar评分、pH值、HMGB1、Cys-C、宫内窘迫、羊水粪染是NA患儿病情转归的相关影响因素(P<0.05);应用R语言绘制基于随机森林算法的Logistic回归模型的可视化列线图,其预测病情不良转归的预测风险能力指数(C-index)为0.887,显示出良好预测能力;ROC分析显示,其预测AUC为0.887,敏感度为89.47%,特异度80.36%。结论脐血pH值、Cys-C、HMGB1与NA患儿病情转归密切相关,基于三者构建的预测模型呈现出优良的预测性能,可作为早期预测病情转归的一个方案,为临床决策提供参考。
Objective To investigate the correlation between cord blood pH,cystatin-C(Cys-C),high mobility group protein B1(HMGB1)and the regression of neonatal asphyxia(NA).Methods 151 children with NA admitted to Meishan Women's and Childrens Hospital,West China Second Hospital,Sichuan University from September 2020 to June 2022 were selected and divided into poor and good groups according to their disease regression,univariate analysis was conducted on clinical data and umbilical cord blood pH,Cys-C,and HMGBl,random forest algorithm was used to rank and screen the importance of disease regression variables.Multivariate Logistic regression analysis was conducted on the relevant influencing factors of disease regression.R language was used to draw a column chart model for predicting disease regression,and the predictive efficacy of the column chart was evaluated using the receiver operating characteristic(ROC)curve.ResulstsThe proportion of children with cord winding,intrauterine distress,amniotic fluid fecal staining,post-asphyxia renal injury and cord blood Cys-C and HMGBI in the poor group were higher than those in the good group,and the 1 min Apgar score,5 min Apgar score and cord blood pH were lower than those in the good group(P<0.05).The random forest algorithm showed that the variables ranked in order of importance were 1 min Apgar score,pH,HMGB1,Cys-C,intrauterine distress,amniotic fluid fecal staining,cord winding,5 min Apgar score,and post-asphyxia renal injury;when the number of variables was 6,the error rate of out-of-bag data was the lowest,so the top 6 variables in order of importance were included as predictors of disease regression in children with NA in the multivariate Logistic regression analysis,showed that I min Apgar score,pH,HMCBI,Cys-C,intrauterine distress,and fecal staining of amniotic fluid were the relevant influencing factors for disease regression in children with NA(P<0.05).The visualized column line plot of the Logistic regression model based on the random forest algorithm was applied to R l
作者
阳梅
李明玉
许群芬
Yang Mei;Li Mingyu;Xu Qunfen(Department of Neonatology,Meishan Women's and Children's Hospital,West China Second Hospital,Sichuan University,Meishan Sichuan 620000;Department of Neonatology,West China Second Hospital,Sichuan University,Chengdu Sichuan 610000,P.R.China)
出处
《中国计划生育和妇产科》
2023年第11期83-87,95,I0001,共7页
Chinese Journal of Family Planning & Gynecotokology
