Prediction of obstetrical risk using maternal serum markers and clinical risk factors

Abstract

Abnormal maternal serum analytes (pregnancy associated plasma protein A, total human chorionic gonadotropin, alpha fetoprotein, Inhibin A, and unconjugated estriol) measured as part of aneuploidy screening programs have been associated with adverse obstetrical outcomes in euploid pregnancies. This study aimed to determine if their predictive ability could be enhanced with additional information on obstetrical history. Forty-five thousand two hundred eighty-seven women participated in the screening program and delivered euploid singleton infants between 2010 and 2012 in British Columbia, Canada. A split-sample design was used to develop and validate prognostic models for serious perinatal events (stillbirth, preterm birth <32 weeks, or HELLP syndrome) and severe pre-eclampsia [pre-eclampsia with preterm birth <34 weeks or small for gestational age <10th percentile] using logistic regression. Three thousand five hundred four women (7.7%) had at least one abnormal marker using standard cut-off values. The combination of serum markers and clinical risk factors improved the ability of statistical models to predict a serious perinatal event [area under the curve (AUC) = 0.62] and severe pre-eclampsia (AUC = 0.78) compared with serum markers or clinical risk factors alone. While detection rates are low, the combination of maternal serum markers and obstetrical history helps to identify a small subset of women at higher risk for serious perinatal events and severe pre-eclampsia.