INTRODUCTION
Development of preventive strategies for stillbirth necessitate recognition that first, the aetiology is heterogeneous and often unknown, and second, the majority of stillbirths are related to placental dysfunction, reflected in the coexistence of small for gestational age (SGA) fetuses and / or preeclampsia. In a prospective study on screening for adverse obstetric outcomes involving 131,514 women with singleton pregnancies attending for routine pregnancy care at 19-24 weeks’ gestation, there were 477 (0.36%) stillbirths, 92.5% of which were antepartum and 7.5% intrapartum; placental dysfunction related stillbirths accounted for 59% of all antepartum stillbirths1. The dataset was used to develop and validate a logistic regression model for prediction of placental dysfunction related stillbirth; a combination of maternal risk factors, sonographic estimated fetal weight (EFW) and uterine artery pulsatility index (UtA-PI) predicted, at 10% false positive rate (FPR), 62% of all cases of placental dysfunction related stillbirths, 70% of those at <37 weeks’ gestation and 29% of those at ≥37 weeks.
In 93% of the placental dysfunction related stillbirths the birth weight was below the 10th percentile of The Fetal Medicine Foundation population charts1,2. It may therefore be preferable to use one model for prediction of both SGA and stillbirth, rather than two separate models; the management of pregnancies at high risk for these conditions is essentially the same and involves serial ultrasound examinations for early diagnosis of SGA and then Doppler assessment of the fetal circulation to determine the best time and mode of delivery. The traditional approach to identify a high risk group for SGA is the application of a scoring system. For example, in the UK, according to guidelines by the Royal College of Obstetricians and Gynaecologists (RCOG), a scoring system is applied to identify a high risk group for SGA in need of serial ultrasound scans from 26 weeks onwards3. An alternative method is provided by our novel two dimensional continuous competing risks model in which SGA is considered as a spectrum disorder whose severity is continuously reflected in both the gestational age at delivery (GA) and Z score in birth weight for gestational age (Z)4-8. The building block of this model is a patient-specific joint distribution of Z and GA, that is obtained by combining a history model with multivariate likelihood of biomarkers according to Bayes theorem4-8. Risk computation is feasible for any chosen cut-off in GA and Z, at any stage of pregnancy by adding any desired biomarker in the same model.
The objective of this study was to examine the predictive performance for placental dysfunction related stillbirths by the competing risks model for SGA based on a combination of maternal risk factors, EFW and UtA-PI and compare the performance to that of our stillbirth-specific logistic regression model using the same biomarkers 1and to the RCOG guideline for the investigation and management of the SGA fetus 3.