Construction Of A Predictive Model For The Risk Of Severe Postpartum Hemorrhage In Placenta Accreta Spectrum

Objective:To investigate the risk factors of severe postpartum hemorrhage in p regnant women with placenta accreta spectrum(PAS).To establish a risk assessm ent model for prenatal prediction of severe postpartum hemorrhage in pregnant women with placenta accreta spectrum(PAS).Methods:All PAS cases from a single tertiary medical center from January 2011 to December 2021 were retrospectively analyzed.The patients were divided into severe postpartum hemorrhage group and non-severe postpartum hemorrhage group with the postpartum blood loss of 1500 ml as the boundary.The clinical information and placental imaging information of all cases were collected.The high risk factors and independent risk factors of severe postpartum hemorrhage associated with PAS were statistically analyzed,and then the risk warning model of severe postpartum hemorrhage was constructed.Results:A total of 463 cases of PAS were included in this study,including 213 cases of severe postpartum hemorrhage and 250 cases of non-severe postpartum hemorrhage,and the incidence of severe postpartum hemorrhage was 46.00%.The univariate analysis revealed that the cases in both groups had statistics difference(P<0.05)of severe postpartum hemorrhage in terms of gestation,delivery,history of vaginal bleeding during pregnancy,assisted reproduction,type of placenta previa,aggressive placenta previa,history of cesarean section,scarred uterus combined with placenta previa,ultrasound imaging features(hypoechoic zone in the myometrium of the lower uterine placenta section,relationship between the myometrium of the lower uterine segment and the placenta,vascularity of the placenta attachment in the uterine wall,strong echogenic lines in the bladder wall of the anterior uterine wall and between the uterine wall and the placenta,magnetic resonance imaging(MRI)for placental implantation,preoperative platelet count,gestational week of termination,hemostatic procedure,and intraoperative determination of placenta-myometrium relationship;logistic multivariate regression analysis showed that there was a statistical difference in the relationship between the placenta and the myometrium of the uterine wall in the case of fatal placenta previa,uterine wall and placenta.Logistic multivariate regression analysis showed that the independent risk factors for severe postpartum hemorrhage in PAS were aggressive placenta previa,strong echogenic metaplastic interface between the uterine wall and placenta,MRI diagnosis of placental implantation,preoperative platelet count,terminating gestational week,and intraoperative determination of placenta-myometrial wall relationship.Based on the analysis of independent high-risk factors,a model was developed to predict the risk of severe postpartum hemorrhage in PAS including clinical factors and placental imaging features:Ln(π/1-π)=Y,Y=-0.868-0.854 X4(platelets ≥150×109/L)+1.204 X10(aggressive placenta praevia)-0.754 X22(termination gestational week≤37 weeks)+1.042 X27(disappearance of strong echogenic metaphase interface between uterine wall and placenta on ultrasound)+1.191 X28(MRI diagnosis of placental implantation)+1.648 X29(intraoperative confirmation of placental penetration),variables in this equation:no takes the value 0,yes takes the value 1;sensitivity=0.8077;specificity=0.7049 for prediction of severe postpartum hemorrhage for this equation,with an overall correct rate of 0.7626;the area under the curve of the ROC curve was 0.8200.Conclusions:PAS related severe postpartum hemorrhage is associated with dan gerous placenta previa,strong echo decidual interface between uterine wall and placenta,MRI diagnosis of placental implantation,preoperative platelet value,termination of pregnancy weeks,and the relationship between placenta and uter ine muscle wall determined during operation.The risk early-warning model of PAS related severe postpartum hemorrhage based on the above independent hig h risk factors has high prediction accuracy,which can provide an important ref erence for the formulation of clinical diagnosis and treatment strategies.