Implementing Mendelian Randomization to assess foetal risk from intrauterine prescriptive drug exposure for the treatment of diabetes, hypertension and thyroid related conditions in pregnancy

INTRODUCTION
Pregnant women are excluded from clinical trials for ethical and practical reasons; thus little is known about the causal impact of intrauterine exposure of prescribed medications on the developing foetus. Yet many chronic conditions necessitate maternal prescriptive medication such as diabetes, hypertension, and thyroid disorders. Untreated, these conditions are well-established risk factors for adverse neonatal outcomes such as preterm birth, impaired neurodevelopment, or stillbirth. However, the treatments themselves may also be associated with additional neonatal risks. Mendelian Randomization (MR) is an instrumental variable analysis using genetic variants as instruments to infer causality, whilst circumventing common limitations of observational epidemiological studies. Genetic variants that are related to the activity or expression to drug targets may be used to proxy intrauterine drug exposure and determine the potential foetal risk as an individual having, or not having, the genetic variant is equivalent to a lifetime exposure to the pharmacological intervention.
OBJECTIVES
The objective of this study was to discern potential adverse effects of maternal diabetic, hypertensive or thyroid disorder treatment on the foetus through MR analysis of maternal genetic drug target variants on a range of neonatal outcomes.
METHOD
We will conduct an MR analysis of maternal genetic drug targets on neonatal outcomes within The Norwegian Mother, Father and Child Cohort Study (MoBa). Using a derived list of potential drug subclasses, the drug targeted region of the genome will be identified using DrugBank. Genetic variants associated with the variation of expression will be identified. These variants will be extracted from MoBa and used to proxy intrauterine drug exposure. A one-sample within family MR analysis will be implemented to explore the potential for prediction of neonate adverse outcomes, such as gestational age, birthweight, mode of delivery and apgar score, using genetic data to proxy maternal drug exposure within the MoBa cohort.
RESULTS
Statistical analyses of these women and their neonates are ongoing. MoBa contains phenotypic data on approximately 114,500 children and 95,200 mothers, and genotypic data on 98,100 individuals. Preliminary summary statistics indicate analysis should be sufficiently powered, with cleaned phenotypic outcome data available for up to 113,627 (gestational age), 113,866 (birthweight), 114,138 (mode of delivery) and 113,646 (apgar score) offspring.
CONCLUSION
The risks to neonates of prescribing during pregnancy necessitates a method to safely and accurately discern whether there are adverse offspring outcomes associated with these treatments, alongside evaluating the extent of the maternal benefit. Triangulation of inferences across pharmacoepidemiological and genetic data is perhaps the best way to evaluate the potential risks currently associated with maternal prescriptive drug use. Evidence established in this study may be used in conjunction with existing literature, clinical trials and alternative study types provides guidance to physicians and mothers during pregnancy.