Objective: Abnormal serum screening markers have been associated with adverse pregnancy outcomes. We sought to review the performance of combined abnormal first and/or second trimester maternal serum markers used in prenatal screening for aneuploidy and open neural tube defects for predicting preeclampsia (PET), small for gestational age (SGA), and stillbirth beyond 24 weeks' gestation.
Data sources and study selection: Medline, EMBASE, and Cochrane Library databases were searched for studies from 1970 to May 2010 that analyzed predictive abilities of combined serum markers for defined outcomes.
Data extraction and synthesis: Data were extracted independently by two authors, and 15 studies were included. Eight studies of 115,290 pregnancies, 11 studies of 144 853 pregnancies, and seven studies of 80 274 pregnancies examined PET, SGA, and stillbirth respectively. Because of the heterogeneity of marker combinations and thresholds, outcome definitions, and analytic methods, limited meta-analysis was possible for the outcomes of PET and SGA only. Three relatively homogeneous studies on prediction of PET, and two on prediction of SGA were meta-analyzed. Several single studies demonstrated utility in combining markers to predict adverse outcome; however, this effect was not confirmed after meta-analysis. The most common combination of markers evaluated was alpha fetoprotein and human chorionic gonadotrophin for all outcomes. The highest positive likelihood ratios for predicting PET (5.68; 95% CI 0.73 to 43.97) and SGA (6.18; 95% CI 1.84 to 20.85) were seen with combined alpha fetoprotein and human chorionic gonadotrophin (> 2.5 multiples of the median).
Conclusions: Currently, no identifiable combination of serum markers performs well as a screening test for preeclampsia, small for gestational age, and stillbirth beyond 24 weeks. Large cohort studies with standardized screening test parameters and outcomes are needed.