Competing risks model in screening for preeclampsia by maternal factors and biomarkers at 30-34 weeks' gestation

Abstract

Background: Preeclampsia (PE) affects 2-3% of all pregnancies and is a major cause of maternal and perinatal morbidity and mortality. We have proposed a 2-stage strategy for the identification of pregnancies at high risk of developing PE. The objective of the first stage, at 11-13 weeks' gestation, is a reduction in the prevalence of the disease through pharmacological intervention in the high-risk group. The objective of the second stage, during the second and/or third trimesters, is to improve perinatal outcome through close monitoring of the high-risk group for earlier diagnosis of the clinical signs of the disease and selection of the appropriate, time, place, and method of delivery.

Objective: The objective of the study was to examine the performance of screening for PE by a combination of maternal factors with early third-trimester biomarkers.

Study design: This was a cohort study and data were derived from consecutive women with singleton pregnancies attending for their routine hospital visit at 30-34 weeks' gestation in 3 maternity hospitals in England between March 2011 and December 2014. In the first phase of the study, only uterine artery pulsatility index (UTPI) was measured and then measurement of mean arterial pressure (MAP) was added, and in the final phase, the serum concentration of placental growth factor (PLGF) was measured and then soluble fms-like tyrosine kinase-1 (SFLT) was added. We had data on UTPI, MAP, PLGF, and SFLT from 30,935, 29,042, 10,123, and 8,264 pregnancies, respectively. The Bayes theorem was used to combine the a priori risk from maternal factors with various combinations of biomarker multiple of the median values. Ten-fold cross-validation was used to estimate the performance of screening for PE requiring delivery at < 37 weeks' gestation (preterm-PE) and those delivering at ≥ 37 weeks (term-PE). The empirical performance was compared with model predictions.

Results: In pregnancies that developed PE, the values of MAP, UTPI, and SFLT were increased and PLGF was decreased. For all biomarkers the deviation from normal was greater for preterm-PE than term-PE, and therefore, the performance of screening was inversely related to the gestational age at which delivery become necessary for maternal and/or fetal indications. Combined screening by maternal factors, MAP, UTPI, PLGF, and SFLT predicted 98% (95% confidence interval, 88-100%) of preterm-PE and 49% (95% confidence interval, 42-57%) of term-PE, at a false-positive rate of 5%. These empirical detection rates are compatible with the respective model-based rates of 98% and 54%, but the latter were optimistically biased.

Conclusion: Combination of maternal factors and biomarkers in the early third trimester could predict nearly all cases of preterm-PE and half of those with term-PE, at 5% false-positive rate.

Keywords: Bayes theorem; mean arterial pressure; placental growth factor; preeclampsia; pyramid of pregnancy care; soluble fms-like tyrosine kinase-1; survival model; third-trimester screening; uterine artery Doppler.