Abstract
Background Few cohort studies of the health effects of urban air pollution have been published. There is evidence, most consistently in studies with individual measurement of social factors, that more deprived populations are particularly sensitive to air pollution effects.
Methods Records from the 1996 New Zealand census were anonymously and probabilistically linked to mortality data, creating a cohort study of the New Zealand population followed up for 3 years. There were 1.06 million adults living in urban areas for which data were available on all covariates. Estimates of exposure to air pollution (measured as particulate matter with an aerodynamic diameter less than 10 mu m, PM10) were available for census area units from a previous land use regression study. Logistic regression analyses were conducted to investigate associations between cause-specific mortality rates and average exposure to PM10 in urban areas, with control for confounding by age, sex, ethnicity, social deprivation, income, education, smoking history and ambient temperature.
Results The odds of all-cause mortality in adults (aged 30-74 years at census) increased by 7% per 10 mu g/m(3) increase in average PM10 exposure (95% CI 3% to 10%) and 20% per 10 mu g/m(3) among Maori, but with wide CI (7% to 33%). Associations were stronger for respiratory and lung cancer deaths.
Conclusions An association of PM10 with mortality is reported in a country with relatively low levels of air pollution. The major limitation of the study is the probable misclassification of PM10 exposure. On balance, this means the strength of association was probably underestimated. The apparently greater association among Maori might be due to different levels of co-morbidity.