Maternal exposure to air pollution in pregnancy and adverse pregnancy and birth outcomes in Victoria, Australia

Background Air pollution is a leading risk factor for global burden of disease. Increasingly, harms associated with maternal exposure to air pollution in pregnancy, both background and episodic, have been described. The gestational period represents a critical developmental window in determining health of the infant, as well as health into childhood and adulthood. However, the relationship between maternal exposure to air pollution in pregnancy and maternal and perinatal outcomes is not well characterised. The need to understand this relationship is becoming increasingly urgent to guide the public health response to air pollution, especially as the threat of severe outdoor air pollution events caused by landscape fires are projected to increase with a changing climate. One such severe smoke event was the 2014 Hazelwood coal mine fire; a six-week air pollution event that affected the Latrobe Valley in south eastern Australia. The Latrobe Early Life Follow-up Study was established to specifically address the potential health harms associated with the coal mine fire for pregnant women and young children. Additionally, understanding this relationship regarding background air pollution is necessary, as many studies are based in relatively polluted cities or countries and not easily generalisable to the Australian context. Aims of the thesis 1. To synthesise all published literature exploring the obstetric and neonatal outcomes associated with short- to medium-term outdoor air pollution events (Chapter 2) 2. To determine whether maternal exposure to fine particulate matter (PM2.5) from a coal mine fire in Victoria, Australia was associated with selected obstetric complications (Chapter 3) 3. To determine whether maternal exposure to PM2.5 from a coal mine fire in Victoria, Australia was associated with selected adverse birth outcomes, including fetal growth restriction and preterm birth (Chapter 4) 4. To determine whether maternal exposure to low-level ambient air pollution in Victoria was associated with selected obstetric complications (Chapter 5) 5. To determine whether maternal exposure to low-level ambient air pollution in Victoria was associated with selected adverse birth outcomes, including fetal growth restriction and preterm birth (Chapter 6) Methods Maternal and fetal harms associated with air pollution, both background and episodic severe smoke events, was explored in the Australian state of Victoria using the following methods: Aim One was addressed with a systematic review of the literature exploring the pregnancy and neonatal outcomes associated with short-to medium-term exposure to outdoor air pollution. Aims Two and Three were addressed through cohort studies that examined obstetric and perinatal outcomes for women who conceived before, during and after the Hazelwood coal mine fire utilising administrative perinatal datasets of births in the affected area. Maternal exposure to average and peak coal mine fire attributable PM2.5 were derived from chemical transport models and assigned to maternal residence at time of birth. Associations between maternal exposure to coal mine fireattributable PM2.5 was assessed with multivariable linear and log-binomial regression models adjusting for potential confounders. Aim three was further explored with use of identified data collected as part of the Latrobe Early Life Study, which enabled assignment of a more geographically resolved exposure estimates that accounted for maternal mobility during the smoke event. Aims Four and Five were investigated utilising state-wide administrative perinatal data. The association between maternal nitrogen dioxide (NO2) and PM2.5 exposure and obstetric and perinatal outcomes were assessed with multivariable linear and log-binomial regression models. Annual ambient air pollution data was derived from state-wide land use regression models estimating PM2.5 and NO2 for maternal residence at birth. Results and Discussion The systematic review of literature presented in Chapter Two demonstrated there is relatively limited research exploring the association between abrupt, time-limited smoke events and both obstetric and perinatal outcomes. Additionally, studies are highly heterogenous in their methodology. Overall, there is some evidence that maternal exposure to acute changes in air quality of short- to medium- term duration increases the risk of fetal growth restriction and preterm birth. The relationship for other adverse obstetric or neonatal outcomes is less clear. To address these gaps in the evidence, through Chapters Three and Four we found maternal exposure to coal mine fire attributable PM2.5 was associated with an increased risk of gestational diabetes mellitus (GDM); second trimester exposure was of critical importance. No association for hypertensive disorders or conditions such as placental abruption was observed. At odds with the literature, we found no association for adverse birth outcomes, including fetal growth restriction or prematurity. However, we did observe that GDM was an effect modifier, whereby babies born to exposed gestational diabetic mothers had greater birth weights with increasing PM2.5 exposure. In Chapters Five and Six, we contribute to the relatively sparse evidence base concerning ambient air pollution and obstetric and perinatal health outcomes. We found mixed outcomes between maternal exposure to annual ambient NO2 and PM2.5 and obstetric outcomes, including hypertensive disorders of pregnancy, GDM and placental abruption. There was evidence that even low-level ambient air pollution was associated with fetal growth restriction. This association was stronger for women with GDM and hypertensive disorders of pregnancy. These findings demonstrate that maternal exposure to low-level ambient air pollution, below Australian Standards and WHO Air Quality Guidelines, may be associated with important adverse fetal outcomes, especially fetal growth restriction, which carries important public health and policy implications. Conclusions This thesis presents a range of studies to address this increasingly important gap in knowledge concerning maternal and perinatal health harms associated with air pollution exposure in pregnancy that is relevant to the Australian context. Novel findings were identified that have since been supported by other studies, namely the association between maternal exposure to a severe smoke event and GDM. Additionally, findings suggested that women with GDM and hypertensive disorders may be a particularly susceptible subgroup of the pregnant population. The findings of this thesis have been shared with the affected community, government and the scientific community, as well as informed the development of a resource for pregnant women. These findings, along with ongoing similar research, will inform robust provision of health protection advice and the public health and emergency response to air pollution for pregnant women.