Smoking and Caffeine Consumption during Pregnancy and Offspring Mental Health

: Disentangling Causal from Confounded Associations using Genetic and Epigenetic Methods

Abstract

It is unclear whether observational associations between smoking and caffeine consumption during pregnancy and offspring mental health problems are causal or confounded. More robust evidence is needed to enable parents to make an informed choice about behaviours during pregnancy. Specifically, for caffeine consumption during pregnancy – a behaviour that is strongly correlated with smoking – pregnancy guidelines are unclear, leading to confusion and uncertainty for expecting parents. Understanding whether smoking and caffeine consumption during pregnancy are causally related to offspring mental health would strengthen the causal evidence base.
The identification of a molecular pathway that could explain how prenatal exposures become biologically embedded to increase the risk for offspring health outcomes could add to the evidence for a causal relationship. Despite a hypothesised role for DNA methylation in mediating the link between prenatal exposures and mental health outcomes, very few studies have tested this hypothesis. In this thesis, I applied genetically informed analyses, as well as examined DNA methylation as a potential biological mechanism, to understand whether prenatal smoking and caffeine exposure are causally related to mental health outcomes.
The results of this thesis indicate that observational associations between prenatal smoking exposure and externalising problems are likely to be confounded by a shared genetic liability between (maternal) smoking and (offspring) risk-seeking personality traits, in turn affecting offspring mental health outcomes. I found little evidence for associations between prenatal caffeine exposure and mental health outcomes. In line with these results, I found no evidence for a causal contribution of prenatal smoking- and caffeine-associated DNA methylation to offspring internalising problems. These results are in accordance with recent research, which indicates that, given the currently available technologies, DNA methylation may be more valuable as a biomarker for prenatal exposures and mental health outcomes than for understanding pathways to mental health problems.

Date of Award	2 Dec 2021
Original language	English
Awarding Institution	University of Bristol
Supervisor	Gemma C Sharp (Supervisor), Marcus R Munafo (Supervisor) & Luisa Zuccolo (Supervisor)