Air pollution impacts at very low levels: Shape of the concentration-mortality relationship in a large population-based Canadian cohort

Background Mortality is associated with long-term exposure to fine particulate matter, although the form of these associations remain poorly understood at lower concentrations. We applied novel 1 km satellite-derived estimates of PM2.5 exposure to a population-based cohort of 7.1 million Canadians. Methods Cox proportional hazard ratios (HRs) were estimated for nonaccidental and cause-specific mortality and 10-year moving average exposure in models adjusted for a wide array of individual and contextual covariates. The shape of the concentration–response function was investigated using restricted cubic splines, threshold models and an extension of the Shape Constrained Health Impact Function (eSCHIF). Analyses examined sensitivity to co-pollutants, concentration thresholds and regional variation. Results Each 10-µg/m3 increase in PM2.5 corresponded with a nonaccidental mortality HR of 1.084 (1.073 -1.096). PM2.5 was associated mortality from ischemic heart disease, pneumonia, COPD, diabetes, and cerebrovascular disease, but not with heart failure, lung cancer, or kidney failure mortality. HR predictions steeply increased from the minimum concentration of 2.5 µg/m3 to 4.5 µg/m3, flattened from 4.5 µg/m3 to 8.0 µg/m3, then increased for concentrations above 8.0 µg/m3, a pattern also reflected in threshold and eSCHIF results. When restricting to those with exposures <10 µg/m3 shapes indicated positive associations for concentrations >9 µg/m3 with indications of adverse effects on mortality at concentrations as low as 2.5 µg/m3. In sensitivity analyses, PM2.5 - mortality associations were only observed in the highest tertile of oxidant gases with shapes varying across airsheds. These differences, were not related to variation in cohort composition or its access to healthcare, suggesting a role of spatially varying pollutant mixtures not sufficiently characterized by PM2.5 mass concentrations. Conclusions In a large Canadian cohort, associations were observed between exposure to PM2.5 with mortality at concentrations as low as 2.5 µg/m3, with no clear evidence of a threshold.