Nomogram to predict postpartum hemorrhage in cesarean delivery for women with scarred uterus: A retrospective cohort study in China

Aim To develop nomograms predicting the risk of postpartum hemorrhage (PPH) in cesarean delivery for singleton pregnant women with a scarred uterus in the north of China. Methods A retrospective cohort study was conducted. Totally 3722 singleton pregnant women with a scarred uterus who underwent a cesarean delivery in a large teaching hospital of north China between January 2013 and December 2017 were enrolled. Nomograms, a kind of user-friendly tool, were developed to predict PPH (blood loss ≥1000 mL or accompanied by signs or symptoms of hypovolemia within 24 h after the birth process) based on the model generated by logistic regression analysis. The discrimination and calibration of models were evaluated, and decision curve analysis was developed. Results Among 3722 enrolled women, 243 (6.53%) had PPH. There are six identified factors associated with PPH: maternal age, placental location, placenta previa, hypertensive disorders of pregnancy, fetal position and placenta accreta spectrum (PAS). The model achieved a good calibration (Hosmer-Lemeshow test P value 0.77) and discrimination (area under curve c-statistics 0.90, 95% confidence interval 0.86-0.93). Decision curve analysis showed the threshold probability by using our model is between 1.0% and 85.7%. A nomogram was developed accordingly. And another nomogram for women without placenta previa and PAS was also developed. Conclusion Two nomograms were first generated to predict PPH, respectively, for women with a scarred uterus and for women with a scarred uterus who do not have placenta previa or PAS. Placental location and fetal position were found to be closely related to PPH.