未足月胎膜早破新生儿早发感染预测模型的构建OA

Objective To construct a nomogram prediction model for early-onset neonatal infection(EONI)in preterm pre-mature rupture of membranes(PPROM),so as to provide evidence for early clinical identification of neonates at high risk of infection and formulation of precise intervention strategies.Methods A retrospective analysis was conducted on the clinical data of 762 singleton pregnant women with PPROM who delivered in the Department of Obstetrics,the First Affiliated Hospital of Guangxi Medical University from January 2014 to December 2023,among whom 188 cases developed EONI.Independent variables were screened through univariate and multivariate logistic regression analysis and LASSO algorithm,and a predictive model was constructed by combining expert opinions.The area under the receiver operating characteristic(ROC)curve(AUC)was used to evaluate the discriminative ability of the model.Hosmer-Lemeshow test,Brier score and calibration curve were ap-plied to verify the calibration degree of the model.Decision curve analysis(DCA)was performed to assess the clinical net ben-efit.In addition,internal validation was finally completed through 1000 Bootstrap resamplings.Results Variables finally in-cluded in the model were gestational age at delivery(days),time from membrane rupture to delivery(latency:18-50 h,>50 h),intrapartum fever(38℃-38.9℃),intra-amniotic infection,and neonatal asphyxia(1-minute Apgar score≤7).The prediction equation was:Logit(P)=17.919-0.081×gestational age at delivery(days)+0.317×time from membrane rupture to delivery(18-50 h,yes=1/no=0)+0.596×time from membrane rupture to delivery(>50 h,yes=1/no=0)+1.285×intrapartum fever(yes=1/no=0)+0.690×intra-amniotic infection(yes=1/no=0)+1.248×neonatal asphyxia(yes=1/no=0).The AUC of the model was 0.866(95%CI:0.834-0.893),with a sensitivity of 86.8%and specificity of 70.2%.The H-L test showed P=0.232(χ2=3.43,df=8),and the Brier score was 0.117.The calibration curve indicated good consistency between predicted and actual values.The decision curve showed that the model had significant clinical net benefit within the threshold probability range of 10%-95%.After internal validation in Bootstrap,the corrected C-index was 0.867(95%CI:0.833-0.895),which confirmed the stability of the model.Conclusion A longer gestational age(days)during delivery is a protective factor against premature infection in PPROM newborns.Time from membrane rupture to delivery>50 h and neonatal asphyxia(1-minute Apgar score≤7 points)are independent risk factors.Intrapartum fever(38℃-38.9℃),intra-amniotic infection,and time from membrane rupture to delivery of 18-50 h can increase the risk of infec-tion.The prediction model constructed in this study has good discriminative ability,calibration degree and clinical application value,which can provide reference for early clinical risk stratification and intervention.