BackgroundLarge cohort studies have been used to characterise the association between long-term exposure to fine particulate matter (PM2.5) air pollution with non-accidental, and cause-specific mortality. However, there has been no consensus as to the shape of the association between concentration and response. MethodsTo examine the shape of this association, we developed a new cohort based on respondents to the 2001 Canadian census long-form. We applied new annual PM2.5 concentration estimates based on remote sensing and ground measurements for Canada at a 1km spatial scale from 1998 to 2011. We followed 2.4 million respondents who were non-immigrants aged 25–90 years and did not reside in an institution over a 10 year period for mortality. Exposures were assigned as a 3-year mean prior to the follow-up year. Income tax files were used to account for residential mobility among respondents using postal codes, with probabilistic imputation used for missing postal codes in the tax data. We used Cox survival models to determine hazard ratios (HRs) for cause-specific mortality. We also estimated Shape Constrained Health Impact Functions (a concentration-response function) for selected causes of death. ResultsIn models stratified by age, sex, airshed, and population centre size, and adjusted for individual and neighbourhood socioeconomic variables, HR estimates for non-accidental mortality were HR = 1.18 (95% CI: 1.15–1.21) per 10μg/m3 increase in concentration. We observed higher HRs for cardiovascular disease (HR=1.25; 95% CI: 1.19–1.31), cardio-metabolic disease (HR = 1.27; 95% CI: 1.21–1.33), ischemic heart disease (HR = 1.36; 95% CI: 1.28–1.44) and chronic obstructive pulmonary disease (COPD) mortality (HR = 1.24; 95% CI: 1.11–1.39) compared to HR for all non-accidental causes of death. For non-accidental, cardio-metabolic, ischemic heart disease, respiratory and COPD mortality, the shape of the concentration-response curve was supra-linear, with larger differences in relative risk for lower concentrations. For both pneumonia and lung cancer, there was some suggestion that the curves were sub-linear. ConclusionsAssociations between ambient concentrations of fine particulate matter and several causes of death were non-linear for each cause of death examined.