Prediction nomogram for antepartum hemorrhage in placenta previa women.

Abstract

Background: Placenta previa with antepartum hemorrhage (APH) is common and closely related to maternal and fetal morbidity and mortality. It is of the utmost importance to prepare for the possibility of APH using perinatal factors prior to delivery.

Objective: To develop and validate a nomogram predicting APH in women with placenta previa based on the perinatal factors.

Design: This was a retrospective and prospective cohort study of pregnant women with placenta previa.

Methods: The model was developed in the retrospective and validated in the prospective cohort study. Multivariate logistic regression was applied to discover independent variables and develop a nomogram to predict the possibility of APH. An Excel form computer interface was constructed to use the model.

Results: There are 1601 and 693 participants in the retrospective and prospective cohort study. Maternal age (odds ratio 0.950, 95% confidence interval 0.918-0.984), married (0.533, 0.309-0.920), parity (1.240, 1.024-1.502), threatened abortion (5.059, 3.648-7.014), and complete placenta previa (1.833, 95% CI 1.469-2.289) were independent variables for APH in placenta previa women. The area under the curve and concordance index were 0.828 and 0.676, respectively. The model was a good fit by the Hosmer-Lemeshow test (p = 0.352). The prospective validation proved the reliability of the prediction nomogram. The Excel form computer interface was practical.

Conclusion: A nomogram based on perinatal factors was developed and validated to predict APH in women with placenta previa. The reliable tool may thereafter offer important assistance for decision-making processes.