skip to primary navigationskip to content
 

Multivariate statistical modelling to characterise phenotypes of human placental dysfunction.

Placentally-related complications of human pregnancy are a major determinant of the global burden of disease. Placental dysfunction can be characterised at multiple levels, including (1) antenatal markers, such as ultrasonic and biochemical indictors of placental function, and (2) analysis of the placenta following delivery, including its size, shape and histopathological appearance. Each of these are independently related to the risk of preeclampsia, fetal growth restriction and spontaneous preterm birth. However, basic questions about the inter-relationships between these factors are unknown. For example, chronic placental inflammation is associated with the risk of preeclampsia and fetal growth restriction. Maternal serum levels of the angiogenic biomarker ratio sFLT1:PlGF are predictive of the same conditions. However, it is unknown whether the predictive associations of the ratio differ according to the placental phenotype. In the Pregnancy Outcome Prediction study, we followed 4,212 women having first pregnancies from their booking visit through to delivery. We obtained serial blood samples and performed serial, blinded ultrasonic fetal biometry and utero-placental Doppler. Placenta from all cases of preeclampsia and FGR were reported blind by two experienced perinatal pathologists. The current project has two aims: (1) to use multivariable statistical methods to determine the inter-relationships between antenatal markers of placentation (biochemical, metabolic, and ultrasonic) and different sub-types of placentally-related complications of human pregnancy, (2) to perform unsupervised clustering analysis to identify patterns of association of maternal characteristics and both biochemical and ultrasonic markers of placentation; we will then assess the distribution of clinical manifestation of disease across different clusters.