Early Dynamic Prediction Of Moderate-to-severe Bronchopulmonary Dysplasia In Very Preterm Infants

Objective:Bronchopulmonary dysplasia（BPD）is a chronic lung disease of preterm infants affected by prenatal,perinatal,and postnatal factors.It has a high mortality and disability,and its adverse effects can last for a whole life.This study aims to determine the independent risk factors of moderate-to-severe BPD in very preterm infants based on the diagnostic and grading criteria for BPD proposed by the Neonatal Research Network（NRN）in 2019,to establish risk nomograms on different postnatal days,and to guide clinical dynamic and quantitative assessment.Methods:Clinical data were retrospectively collected from January 1,2019 to June 26,2022for very preterm infants admitted to the Neonatal Intensive Care Unit（NICU）of the First Hospital of Jilin University within 24 hours after birth.The clinical data collected mainly included the basic information of the infants,maternal pregnancy history and delivery information.Adverse events and respiratory support data were collected on postnatal days 1,3,7,and 14.The diagnosis and severity of BPD were performed based on the definition proposed by the Neonatal Research Network（NRN）in 2019,in which grade II～III BPD was the moderate-to-severe BPD group（case group）,and no BPD and grade I BPD were the control group.IBM SPSS 25.0 and 4.2.2 version of R language rms package were used to analyze the data between the two groups（P<0.05 was considered statistically significance）,to identify the independent risk factors of moderate-to-severe BPD in very preterm infants on different postnatal days,and construct nomograms.Through the receiver operating characteristics（ROC）curve and internal validation method to assess the predictive performance of the models.Results:1、A total of 400 very preterm infants were included in this study,including 140infants（35.00%）with moderate-to-severe BPD.Univariate analysis results:among the basic information,gestational age（GA）,birth weight（BW）,Apgar1-min,acute chorioamnionitis,test-tube baby,fetal distress,delivery intubation,RDS,delivery PS and using prenatal antibiotic in the two groups were statistically significant（P<0.05）.PS≥2 times,ventricular hemorrhage（gradeⅢ～Ⅳ）,respiratory support type,Fi O₂,invasive ventilation days between two groups were statistically significant on postnatal days 1,3,7 and 14（P<0.001）.Pneumothorax was statistically significant difference on postnatal days 7 and 14（P<0.05）.Neonatal sepsis was statistically significant difference on postnatal day 7（P<0.05）.Hemodynamically significant patent ductus arteriosus（hs PDA）were significant differences on postnatal days 3,7 and 14（P<0.05）.2、Logistic regression analysis results:GA（P<0.001）and PS≥2 times（P=0.004）,fetal distress（P=0.002）and delivery intubation（P<0.001）were independently correlated with the risk of moderate-to-severe BPD on postnatal days 1.GA（P<0.001）,PS≥2 times（P=0.012）,fetal distress（P=0.005）,delivery room intubation（P=0.001）,and respiratory support type（P=0.001）were independently associated with the risk of moderate-to-severe BPD on postnatal day 3.GA（P<0.001）,PS≥2 times（P=0.028）,fetal distress（P=0.005）,delivery room intubation（P=0.003）,neonatal sepsis（P=0.042）,and invasive ventilation days（P<0.001）was independently associated with the risk of moderate-to-severe BPD on postnatal day 7.GA（P=0.008）,PS≥2 times（P=0.013）,fetal distress（P<0.001）,neonatal pneumonia（P=0.013）,respiratory support type（P=0.029）,Fi O₂（P=0.022）,and invasive ventilation days（P=0.003）were independently associated with the risk of moderate-to-severe BPD on postnatal day 14.3、Performance of nomograms on different postnatal days:The areas under the ROC curve of the prediction models on postnatal days 1,3,7 and 14 were 0.801,0.822,0.830 and 0.856.The sensitivity and specificity of predicting moderate-to-severe BPD were 73.57%,72.69%;75.00%,75.77%;80.00%,70.00%;77.85%,78.85%;respectively.Bootstrap method was used for internal validation,and the calibration curve fitted well with the ideal curve.The area under the ROC curve after bootstrap validation were 0.806,0.819,0.826,and 0.853,respectively.Conclusion:1.GA,PS≥2 times,and fetal distress on different postnatal days were independently associated with moderate-to-severe BPD.2.On different postnatal days,there are differences in the risk factors that need to be evaluated.Intubation in the delivery room,neonatal sepsis,neonatal pneumonia,invasive ventilation days,respiratory support type and Fi O₂ only had independent predictive value for moderate-to-severe BPD on some postnatal days.3.The prediction models of moderate-to-severe BPD established on different postnatal days in the early life based on NRN 2019 diagnosis and classification criteria had good performance.