新生儿革兰阳性菌败血症发病风险预测模型的构建  

Construction of a risk prediction model for Gram-positive bacterial sepsis in neonates

作　　者：高正平 赵雪臻 寇晨[1] GAO Zhengping;ZHAO Xuezhen;KOU Chen(Department of Neonatology,Beijing Obstetrics and Gynecology Hospital,Capital Medical University,Beijing Maternal and Child Health Care Hospital,Beijing 100026,China)

机构地区：[1]首都医科大学附属北京妇产医院,北京妇幼保健院新生儿科,北京100026

出　　处：《山东医药》2024年第15期35-40,共6页Shandong Medical Journal

基　　金：北京市医院管理中心“青苗”计划(QML20211403)。

摘　　要：目的建立新生儿革兰阳性菌败血症发病风险预测模型,以早期预测新生儿革兰阳性菌败血症的发病风险。方法以是否发生革兰阳性菌败血症为结局变量,选取收集的149例新生儿革兰阳性菌败血症患儿26个临床指标(包括产前因素,感染前生命体征、症状及感染时血液指标等)为预测变量,通过最小绝对收缩和选择算子(Lasso)回归模型及多因素Logistic回归筛选新生儿革兰阳性菌败血症危险因素,基于筛选出的危险因素构建新生儿革兰阳性菌败血症发病风险预测模型,并可视化成列线图(列线图预测模型)。使用校准曲线、受试者工作特征曲线(ROC)和临床决策曲线(DCA)评估列线图预测模型的校准度、区分度和临床适用性。应用Bootstrap法对列线图预测模型进行内部验证。根据列线图中所有变量总分的四分位数将149例败血症患儿分为低危组(总评分0~73分)、中危组(总评分74~83分)、高危组(总评分84~95分)、极高危组(总评分>95分),采用单因素方差分析各组发生革兰阳性菌败血症的风险,以对列线图预测模型的预测效能进行临床验证。结果经过Lasso回归筛选13个预测变量,分别是患儿呼吸、胃潴留或腹胀、血糖水平、发病前3天体质量变化、患儿胎龄、分娩方式、患儿母亲孕期合并糖尿病、孕晚期感染、患儿的发病日龄、发病前中心静脉导管留置时间、发病前呼吸机应用时间、发病时血小板计数以及C-反应蛋白(P均<0.05)。进一步多因素Logistic回归分析筛选出5个危险因素,分别是呼吸增快、发病前3天体质量变化幅度大(>80 g/d)、患儿的发病日龄短(14 d内)、发病前中心静脉导管留置时间长(15~30 d)以及发病前呼吸机应用时间长(≥7 d,P均<0.05)。基于上述5个新生儿革兰阳性菌败血症危险因素构建了列线图预测模型。该列线图预测模型显示出较好的校准度;ROC曲线下面积为0.77,灵敏度69.Objective To develop a risk prediction model for Gram-positive bacterial sepsis in neonates and to predict the incidence risk of Gram-positive bacterial sepsis in newborns as early as possible.Methods The outcome variable was whether Gram-positive bacterial sepsis occurred or not.Data were collected from 149 neonates diagnosed with Gram-positive bacterial sepsis,including 26 clinical indicators(prenatal factors,vital signs and symptoms before infection,and blood indices at the time of infection).The least absolute shrinkage and selection operator(Lasso)regression model and multivariate Logistic regression were used to identify risk factors for Gram-positive bacterial sepsis in neonates.Based on these selected risk factors,a risk prediction model was constructed and visualized as a nomogram.The calibra tion,discrimination,and clinical utility of the nomogram were assessed using calibration curves,receiver operating characteristic(ROC)curves,and decision curve analysis(DCA).Internal validation of the nomogram was performed using the Bootstrap method.The 149 children with sepsis were categorized into low-risk(total score 0-73 points),medium-risk(total score 74-83 points),high-risk(total score 84-95 points),and very high-risk(total score>95 points)groups based on the quartiles of the total score in the nomogram.The risk of developing Gram-positive bacterial sepsis among these groups was evaluated using one-way ANOVA to validate the predictive performance of the nomogram.Results Thirteen predictive variables were identified through Lasso regression,including neonatal respiration,gastric retention or abdominal distension,blood glucose levels,body weight changes within 3 days before onset,gestational age,mode of delivery,maternal diabetes during pregnancy,late pregnancy infection,day of onset,duration of central venous catheterization before onset,duration of mechanical ventilation before onset,platelet count,and C-reactive protein(all P<0.05).Multivariate Logistic regression further identified five risk factors:rapid r

关 键 词：发病风险预测模型 列线图 败血症 革兰阳性菌败血症 

分 类 号：R72[医药卫生—儿科]