Examining the association between air pollution exposure (PM10, PM2.5 and NO2) across the lifecourse on behavioural and mental health outcomes in adulthood in the Christchurch Health and Development study: a prospective birth cohort study

The adverse effects of air pollution on human physical health are well established, but potential links with behavioural and mental health problems are also suggested. However, evidence is limited, with few studies employing spatial life course methods or considering the effects of unmeasured and residual confounding factors. This study addresses the identified gap by using prospective longitudinal data from the Christchurch Health and Development Study to examine associations between cumulative air pollution exposure and subsequent behaviour and mental health outcomes in adulthood. The study sample consists of a birth cohort of 1265 individuals born during 1977 in Christchurch, New Zealand. Individual-level residential address data were linked to air pollution estimates, including annual black smoke levels (micrograms per cubic meter of air or μg/m 3 ) from during pregnancy to age 10 years (1976–1987), and PM 10 (21–40 years), PM 2.5 (30–40 years), and NO 2 (40 years) in adulthood. Behavioural (i.e. nicotine, alcohol, cannabis, and other illicit substance use disorders) and mental health outcomes (i.e. depressive symptoms, anxiety disorders, and suicidal ideation) were prospectively assessed from age 21–40 years. Findings from unadjusted population-averaged GEE logistic regression models indicated only weak associations between cumulative air pollution exposure and some behavioural and mental health outcomes. In fully adjusted models, associations were fully attenuated, other than PM 2.5 exposure and nicotine abuse/dependence (OR = 1.074 [1.0156, 1.136]). While further research is needed to confirm our findings, air pollution exposure may be less important than more proximal factors for behaviour and mental health outcomes in adulthood. • Employed a spatial lifecourse epidemiological framework. • Historical and contemporary air pollution were estimated for a large prospective birth cohort study. • While air pollution exposure was related to some outcomes, any effects were often small. • Associations were largely attenuated in fully adjusted models. • More robust data are needed to confirm our findings.