Predictive Value of Second-Trimester Biomarkers and Maternal Features for Adverse Pregnancy Outcomes

Objective: To establish the predictive probability for placenta-associated morbidities using second-trimester α-fetoprotein (AFP), human chorionic gonadotropin (HCG), and maternal features. Patients and Methods: A retrospective database of all singleton deliveries with available second-trimester HCG and AFP results from 2005 to 2012 was built and divided into 0, 1, or 2 elevated markers (defined as ≥2 multiples of the median [MoM]). For each group, we analyzed the risk for adverse obstetric outcome - comprising preeclampsia, placental abruption, and birth weight below the 10th percentile - and the time of delivery in those pregnancies. Additionally, prediction models for adverse obstetric outcome, using logistic regression incorporating AFP, HCG, and other maternal characteristics, were calculated. Results: Among 22,124 women who delivered, 16,197 (73%) had AFP and HCG results. Compared with the group with normal markers, the adverse obstetric outcome rate was mildly increased with elevated HCG or AFP, but it was markedly increased when both markers were elevated (13 vs. 31%, OR 2.9, 95% CI 2.0-4.3). Delivery of newborns with adverse obstetric outcome was earlier with each additional elevated marker. The accuracy of predicting adverse obstetric outcome was improved by using prediction models for women with HCG or AFP ≥1.2 MoM that incorporated maternal age, BMI, parity, and chronic hypertension (C-statistic 61-75%). Conclusion: HCG and AFP combined with other maternal characteristics are useful tools for predicting the risk for adverse obstetric outcome.