Abstract
Background and aimsThe incidence of obesity in pregnancy is growing worldwide and has become a large concern to public health due to the higher risks of adverse outcomes for both the woman and her offspring. The causal pathways associating maternal obesity in pregnancy and higher risks of neurodevelopmental disorders in the child and the roles of maternal depression and gestational diabetes (GDM) in these associations are not well understood. Further, these children are also at higher risk of metabolic disorders and obesity. The etiology of these disorders may have a uterine origin through the disruption of neurodevelopment, however which brain networks are implicated remains to be clarified and these have not been explored in the neonatal brain where postnatal effects are reduced. The availability of multidimensional datasets which have longitudinal measures of health through pregnancy and the advances of neuroimaging in the neonate offer unprecedented opportunities to study the associations above. When several aspects of health and antenatal exposures are better understood, the potential for early detection of vulnerabilities improves and the strategies for intervention become better targeted.
This thesis addresses multiple concepts in the relationship between maternal obesity and child health outcomes and employs various methods within the structural equation modeling toolbox which support the conceptual framework of the Developmental Origins of Health and Disease (DOHaD). It contains several aims:
Aim 1) Explore the concomitant exposures potentially adverse to normal neurodevelopment in relation to maternal antenatal depressive symptoms in pregnancies with obesity. Aim 2) Estimate the risks of psychopathology in 3-year old children born of obese mothers and the relationship between their psychological phenotypes and adiposity. Aim 3) Formulate causal models for the relationships between antenatal exposures and risks of psychopathology in the offspring of obese mothers. Aim 4) Identify the potential neurological underpinning for adverse child psychological and metabolic outcomes by studying the neonatal brain using MRI.
Methods
The thesis starts by exploring and reviewing the multiple facets of the obese pregnancy through physiology and maternal mental health (Chapter 1). Then, the potential pathways to how these facets may impact on neurodevelopment are described in Chapter 2, along with an epidemiological framework. Chapter 3 introduces the methodology applied in the subsequent chapters/studies which answer the aims of the thesis.
The study in chapter 4 involves data collected from the UPBEAT study of 1554 pregnant women with obesity. Latent class growth modeling of antenatal depressive symptom was performed based on scores from the Edinburgh Postnatal Depression Scale (EPDS). Classes were then compared (3-step approach) on maternal diet, blood biomarkers, infection rate, pregnancy and birth outcomes (Aim 1).
In chapter 5, latent class analysis based on the the Strength and Difficulty Questionnaire (SDQ) subscales was used to assess the psychological outcomes in a sample of 462 3-year old offspring of women from UPBEAT. Psychological phenotypes were compared on anthropometric measurements (Aim 2). The third aim was addressed by employing structural equation models (SEMs) using the Bayesian statistical framework to characterize the likely contribution of maternal antenatal GDM/glycaemic status, adiposity, inflammation and depression towards psychological outcomes in the children of UPBEAT participants.
Chapter 6 and 7 fulfill the fourth aim and relied on structural and diffusion MRI data acquired through the developing Human Connectome Project (dHCP). First, it involved segmenting the neonatal hypothalamus, nucleus accumbens and ventral tegmental area (Chapter 6). Then, tractography protocols for white-matter bundles were developed to generate 7 tracts implicated in energy homeostasis, the reward and limbic systems, in neonates of 37 to 44 weeks postmenstrual age. This was followed by a fixel-based analysis and SEMs to understand the development of tract-wise macro- and microstructure after normal-weight pregnancy. Subsequently, neonates born of normal-weight mothers (n=137) were compared to obesity-exposed newborns (n=28).
Results
In Chapter 4, results suggest there were four latent depressive symptom trajectories present in the UPBEAT cohort. Compared to women with very low risk of depression, the women with the highest longitudinal depressive profile had worse diet, higher rates of infection, higher inflammation (interleukin-6 and glycoprotein acteyls), lower placental growth factor, and 3 times the risk of preterm birth. There were further heterogeneities found across the other three classes on these measures. Infection rate was the lowest in the Not depressed class. There was no difference in the incidence of preeclampsia or GDM diagnoses between the classes but the Depressed class was more likely to have missed their OGTT and had higher glucose levels in the last trimester. There were socio-economic disparities and ethnic differences between the classes.
In Chapter 5, 11.3% (52/462) of the children in UPBEAT were at risk of psychopathology based on the total SDQ score. The LCA retrieved 3 classes: internalising phenotype (n=19,4.1%), externalising phenotype (n=37, 8%) and a well-adjusted phenotype (n=406, 87.9%). Compared to the well-adjusted children, those presenting the externalising profile were of higher BMI-for-age and heavier-for age (by World Health Organisation standards) and with larger waist circumferences. SEMs based statistics provided evidence that maternal antenatal depression is a significant predictor for risk of psychopathology at 3 years which was independent of GDM, maternal adiposity and postnatal depressive symptoms.
Chapter 6 demonstrated that the quality of the reconstructed structural data from the dHCP enabled the neonatal hypothalamus, nucleus accumbens and ventral tegmental areas to be segmented following established protocols in the adult literature.
Chapter 7 presents, to my knowledge, the first time in the human neonate, tractography of the stria terminalis, medial forebrain bundle, dorsal longitudinal fasciculus, amygdalo-accumbens fasciculus and ventral amygdalofugal pathway. Results indicate that there is heterogeneity in the development of these white-matter tracts. Fiber density increases across time in the short period of 37 to 44 weeks postmenstrual age but not equally across the bundles. Furthermore, male neonate show larger cross-section than female in the bundles associated with the temporal pole and amygdala. There was an interaction of exposure group with PMA on the fibre density in the bilateral amygdaloaccumbens fasciculus and right uncinate fasciculus so that obesity-exposed neonate did not show the same increase in fibre density with increasing age as the controls.
Conclusion
In conclusion, the study of associations between maternal obesity and child health outcomes need to take a holistic approach which encompasses the multiple exposures potentially adverse to the fetus in utero. Although a genetic contribution was beyond the scope of this thesis, the role of maternal mental health, socio-economic disparities and medical determinants are likely to be significant determinants in explaining health outcomes and some aspects in the transgenerational transfer of disease risk. The prioritization of pre-conceptual health advocated through the DOHaD framework should promote strong policies addressing not only metabolism but also mental health and social disparities into any future initiatives.
| Date of Award | 1 Nov 2022 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Mary Rutherford (Supervisor) & Lucilla Poston (Supervisor) |