Risk factors and prediction model for massive transfusion during cesarean section in singleton pregnancies with anterior placenta previa, prior caesarean and prenatal suspicion of placenta accreta spectrum: a retrospective case-control study.

BACKGROUND

Severe postpartum hemorrhage is frequently encountered during cesarean sections (CSs) in individuals with placenta accreta spectrum (PAS). The prompt and adequate administration of blood transfusions has emerged as a critical intervention in managing severe postpartum hemorrhage. Considering that anterior placenta previa (APP) and a prior CS constitute key risk factors for PAS, this investigation aims to investigate these risk factors and develop a prediction model for massive transfusion (MT) during CS in cases of singleton pregnancies with APP, a prior CS, and prenatal suspicion of PAS, utilizing the placenta accreta spectrum ultrasound scoring system (PASUSS).

METHODS

A cohort of 430 individuals with APP, having undergone prior CS, and with prenatal suspicion of PAS, as determined by PASUSS, were retrospectively examined at Shengjing Hospital, affiliated with China Medical University, between January 2018 and December 2021. These patients were divided into cohorts of MT (168 cases) and non-MT (262 cases) according to the volume of packed red blood cells transfused intraoperatively. The cohort was arbitrarily subdivided into training and validation cohorts in a 7:3 proportion. LASSO and multivariate logistic regression analyses were employed to ascertain independent risk factors for MT. A prediction model was developed, and its predictive efficacy was evaluated through the use of receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

RESULTS

Out of the 430 patients, 168 underwent MT during CS, reflecting an incidence rate of 39.07%. Independent risk factors for MT in cases of singleton pregnancies with APP, prior CS, and prenatal suspicion of PAS included emergency surgery, PASUSS score, preoperative hemoglobin level, clinical staging of PAS, and one-step conservative surgery. A nomogram was subsequently developed utilizing these identified factors. The areas under the ROC curves for the training and validation cohorts were 0.908 (0.875-0.940) and 0.925 (0.882-0.968), respectively. Both calibration curves and DCA demonstrated that this nomogram possessed a strong predictive value.

CONCLUSIONS

Independent risk factors for MT included emergency surgery, PASUSS score, preoperative hemoglobin level, clinical staging of PAS, and one-step conservative surgery. The nomogram constructed from these variables serves as an effective prediction model for identifying MT in singleton pregnancies characterized by APP, prior CS, and prenatal suspicion of PAS.