Construction Of A Predictive Model For The Onset Of Neonatal Respiratory Distress Syndrome In Patients With Hypertensive Disorders In Pregnancy

Objectives:1.To investigate the potential risk factors for the development of neonatal respiratory distress syndrome(NRDS)in patients with hypertensive disorder of pregnancy(HDP)and the prenatal fetal ultrasound tests associated with the development of NRDS The study was conducted to find out the independent factors influencing the development of NRDS in HDP patients.2.To develop a non-invasive,simple,economical and efficient model for predicting the onset of NRDS in HDP patients by combining multifactorial analysis,and to investigate its feasibility,so as to provide clinicians with evidence for early identification of high-risk newborns,guide the precise application of glucocorticoids,select the appropriate timing of delivery,and adequately prepare high-risk newborns before and after birth,and improve perinatal survival.Methods:1.Study population:A total of 754 pregnant women with HDP at 28-38+6 weeks of gestation who delivered at Qingdao Municipal Hospital from October 2020 to October 2022 were included,and pregnancy outcomes were followed up,with a total of 74 cases in the group with NRDS(case group)and 680 cases in the group without NRDS(control group).2.Data collection and study methods:(1)Fetal ultrasound Doppler examinations,including general fetal growth indicators,fetal lung volume measurements and fetal pulmonary aortic blood flow parameters,were performed 48 hours before delivery in all enrolled cases,and information on general maternal data and potential prenatal risk factors for the development of NRDS was collected,pregnancy outcomes were followed up,clinical data of the newborns were counted,and samples that did not meet the enrollment criteria were excluded.(2)Predictive models for the development of NRDS were constructed by combining the findings of previous studies,clinical experience and statistical results.The variables included in the analysis were: BPD,HL,fetal pulmonary aortic flow(AT,ET,AT/ET,EDV,PSV,MV,RI,PI),total fetal lung volume,maternal age,gestational age,number of pregnancies,number of deliveries,week of gestation,mode of delivery,whether or not there was a combination of(gestational diabetes,premature rupture of membranes,intrauterine growth retardation,placental abruption,intrauterine distress,in vitro fertilisation,other medical and surgical conditions).3.Statistical methods:SPSS25.0 statistical software was used for statistical analysis,(1)statistical comparison of basic information;(2)single-factor analysis of all variables included and elimination of non-statistically significant variables;(3)multi-factor analysis of variables with significance and search for independent influencing factors;(4)establishment of binary logistic regression prediction model with independent influencing factors as independent variables;(5)using ROC curve to test its discrimination and Hosmer-Lemeshow to test its calibration;using R software(R4.2.2)rms program package to plot Nomogram and calibration curve.Results:1.Single factor analysis:(1)Analysis of fetal prenatal ultrasound test indicators showed that in the case group compared to the control group,the former fetuses had BPD(t=12.698,P<0.001),HL(t=11.596,P<0.001),total lung volume(t=15.136,P<0.001),AT(t=9.533,P<0.001),AT/ET(t= 14.501,P<0.001),EDV(t=4.453,P<0.001),PSV(t=4.732,P<0.001),and MV(t=5.311,P<0.001)values were significantly smaller and PI(t=5.453,P<0.001)values were significantly larger.The most significant differences were found in total lung volume,AT and AT/ET.(2)When maternal antenatal factors were analysed,the case group compared to the control group showed significantly older maternal age(t=3.180,P=0.00)2,significantly smaller gestational weeks(t=12.608,P<0.001),caesarean section(2=17.098,P<0.001),placental abruption(2=5.638,P=0.018),intrauterine fetal distress(2= 29.358,P<0.001),fetal growth restriction(2=31.612,P<0.001)and gestational diabetes mellitus(2=59.111,P<0.001)had significantly higher combined rates.(3)The differences between the two groups of cases in pulmonary aortic flow(ET,RI)and maternal factors(number of pregnancies,number of deliveries,whether combined with IVF,whether combined with other medical or surgical conditions,whether combined with premature rupture of membranes)were not statistically significant(P>0.05).2.Multi-factor analysis:Multi-factor analysis was performed using variables that differed significantly in the univariate results and finally determined gestational week(P=0.017,OR=0.590,95%CI=0.383-0.909),total fetal lung volume(P=0.005,OR=0.094,95%CI=0.907-0.983),AT/ET(P=0.002,OR= 0.000,95% CI=0.000-0.000)and BPD(P=0.032,OR=0.822,95% CI=0.688-0.983)were protective factors for the development of NRDS,and fetal growth restriction(P=0.011,OR=6.742,95%CI=1.559-29.154),gestational diabetes mellitus(P =0.001,OR=6.821,95%CI=2.272-20.479)and intrauterine distress(P=0.001,OR=7.136,95%CI=2.206-23.084)were risk factors for the development of NRDS.Other included variables,including age,mode of delivery,placental abruption,PI,RI,EDV,PSV,and MV were not statistically significant in the multifactorial analysis(P>0.05).3.Predictive modeling:(1)Using gestational week,gestational diabetes,fetal growth restriction,intrauterine distress,total fetal lung volume,AT/ET,and BPD as independent variables,logistic regression analysis was performed to determine the final equation: logit P = 41.214 +(-0.415 x gestational week)+(1.652 x fetal growth restriction)+(1.904 x gestational diabetes)+(2.133 x fetal intrauterine distress)+(-0.174 x BPD)+(-30.721 x AT/ET)+(-0.052 x total fetal lung volume),P = exp(logit P)/1 + exp(logit P).(2)Nomogram plots of the prediction model were plotted to visualise the equations.The regression coefficients for gestational week(OR=0.660,95% CI=0.478-0.912),total fetal lung volume(OR=0.949,95% CI=0.913-0.987),AT/ET(OR=0.000,95%CI=0.000-0.000)and BPD(OR=0.841,95% CI=0.726-0.973)had negative regression coefficients,indicating that higher values were associated with a lower probability of NRDS;fetal growth restriction(OR=5.219,95% CI=1.401-19.440),gestational diabetes(OR=6.711,95% CI=2.536-17.758),and intrauterine distress(OR=8.442,95% CI= A positive regression coefficient for(OR=2.768-25.745)indicates a higher probability of NRDS in the presence of this factor.4.Model calibration:The model Hosmer-Lemeshow goodness-of-fit test was 2 = 9.198,p = 0.326,area under the ROC curve AUC = 0.801(p < 0.001,95% CI = 0.759-0.842),cut off = 0.384,Jorden index 0.489,sensitivity = 0.852,specificity = 0.637,and The calibration curve of the Nomogram plot is a straight line with a slope close to 1.Conclusions:1.the predictive model for the development of NRDS in pregnant women with HDP established by combining gestational week,fetal growth restriction,gestational diabetes,intrauterine distress,total fetal lung volume,AT/ET,and BPD has good efficacy and visualizes risk values using Nomogram plots.2.By predicting the risk of NRDS onset,early identification of high-risk infants provides scientific guidance for physicians’ clinical management work,and has great clinical significance in the precise application of glucocorticoids,selection of appropriate timing for termination of pregnancy,and adequate preparation for high-risk newborns before and after birth,which can reduce the occurrence of adverse pregnancy outcomes and improve the survival rate of perinatal infants.