新生儿早发败血症合并持续性肺动脉高压风险的预测模型建立及验证  

Establishment and verification of a prediction model of neonatal early-onset sepsis with persistent pulmonary hypertension

作　　者：朱巧棉 张慧平 王晓林 Zhu Qiaomian;Zhang Huiping;Wang Xiaolin(Department of Pediatric Medicine,Shanxi Medical University,Taiyuan 030001,China;Department of Neonatal Intensive Care Medicine,Xi'an Jiaotong University Affiliated Children's Hospital,Xi'an 710002,China;Department of Neonatal Intensive Care Medicine,the Second Hospital of Shanxi Medical University,Taiyuan 030001,China)

机构地区：[1]山西医科大学儿科医学系,太原030001 [2]西安交通大学附属儿童医院新生儿重症医学科,西安710002 [3]山西医科大学第二医院新生儿重症医学科,太原030000

出　　处：《中华新生儿科杂志(中英文)》2025年第2期76-82,共7页Chinese Journal of Neonatology

基　　金：国家自然科学基金面上项目(82371547);山西省医学重点科研项目(2020XM29)。

摘　　要：目的探讨新生儿早发败血症(early-onset sepsis,EOS)合并持续性肺动脉高压(persistent pulmonary hypertension of the newborn,PPHN)的独立危险因素,并构建疾病预测模型。方法回顾性分析2019年1月至2023年7月山西医科大学第二医院新生儿重症医学科收治的新生儿早发败血症患儿的临床资料。采用R语言以7∶3比例将数据集随机分为训练集和测试集,根据是否存在PPHN分为PPHN组和无PPHN组。在训练集使用单因素分析、LASSO回归和多因素二元logistic回归分析筛选出新生儿早发败血症合并PPHN的独立危险因素并建立风险评估模型。使用受试者工作特征曲线和Hosmer-Lemeshow检验评估模型性能和拟合优度,并绘制模型列线图进行可视化风险预测。结果共纳入171例新生儿早发败血症患儿。训练集120例,其中败血症合并PPHN组40例,无PPHN组80例;测试集51例,其中败血症合并PPHN组16例,无PPHN组35例。结果显示妊娠期高血压(OR=4.894,95%CI 1.022~23.437)、男性(OR=5.259,95%CI 1.346~20.544)、白细胞介素6升高(OR=1.018,95%CI 1.011~1.026)、转化生长因子β1升高(OR=1.002,95%CI 1.001~1.004)是新生儿早发败血症合并PPHN的独立危险因素(P<0.05)。预测模型在训练集及测试集的曲线下面积分别为0.915(95%CI 0.857~0.973)和0.922(95%CI 0.844~0.995)。界值为0.388时,训练集的敏感度为85.0%,特异度为88.7%;测试集的敏感度为82.4%,特异度为88.2%。结论妊娠期高血压、男性、白细胞介素6升高、转化生长因子β1升高是新生儿早发败血症合并PPHN的独立危险因素。建立的预测模型有助于临床医生早期识别新生儿早发败血症发生PPHN的高危患儿。ObjectiveTo investigate the independent risk factors associated with neonatal early onset sepsis(EOS)in infants with persistent pulmonary hypertension of the newborn(PPHN),and to develop a predictive model for the disease.MethodsThe clinical data of EOS patients admitted to the neonatal intensive care unit of the Second Hospital of Shanxi Medical University from January 2019 to July 2023 were retrospectively analyzed.The data sets were randomly divided into a training set and a test set with a ratio of 7∶3 by using R language.According to the presence or absence of PPHN,the newborns were divided into PPHN group and non-PPHN group.In the training set,univariate analysis,LASSO regression analysis,and multivariate binary logistic regression were used to screen out the risk factors of EOS with PPHN and to develop a risk assessment model.The receiver operating characteristic area under curve(AUC)and the calibration curve were used to evaluate model performance and goodness of fit test,and the model nomogram was drawn for visual risk prediction.ResultsA total of 171 newborns with EOS were enrolled.The training set consisted of 120 cases,of which 40 involved sepsis with PPHN and 80 didn't involve PPHN.The test set consisted of 51 cases,including 16 cases with PPHN and 35 cases without PPHN.Pregnancy-induced hypertension syndrome(OR=4.894,95%CI 1.022-23.437),male sex(OR=5.259,95%CI 1.346-20.544),elevated IL-6(OR=1.018,95%CI 1.011-1.026),and TGF-β1(OR=1.002,95%CI 1.001-1.004)were independent risk factors for neonatal EOS complicated with PPHN(all P<0.05).The AUC of the prediction model in the training set and the test set were 0.915(95%CI 0.857-0.973)and 0.922(95%CI 0.844-0.995),respectively.Taking the prediction probability of 0.388 as the high-risk threshold,the sensitivity and specificity of the training test were 85.0%and 88.7%,and the sensitivity and specificity of the test set were 82.4%and 88.2%,respectively.ConclusionsPregnancy-induced hypertension syndrome,male sex,elevated IL-6 and TGF-β1 are independent ris

关 键 词：新生儿败血症 PPHN 危险因素 预测模型 

分 类 号：R722.1[医药卫生—儿科]