Background

Bushfire smoke is a major ongoing environmental hazard in Australia. In the summer of 2019-2020 smoke from an extreme bushfire event exposed large populations to high concentrations of particulate matter (PM) pollution. In this study we aimed to estimate the effect of bushfire-related PM of less than 2.5 um in diameter (PM2.5) on the risk of mortality in Sydney, Australia from 2010 to 2020.

Methods

We estimated concentrations of PM2.5 for three subregions of Sydney from measurements at monitoring stations using inverse-distance weighting and cross-referenced extreme days (95th percentile or above) with satellite imagery to determine if bushfire smoke was present. We then used a seasonal and trend decomposition method to estimate the Non-bushfire PM2.5 concentrations on those days. Daily PM2.5 concentrations above the Non-bushfire concentrations on bushfire smoke days were deemed to be Bushfire PM2.5. We used distributed-lag non-linear models to estimate the effect of Bushfire and Non-bushfire PM2.5 on daily counts of mortality with sub-analyses by age. These models controlled for seasonal trends in mortality as well as daily temperature, day of week and public holidays.

Results

Within the three subregions, between 110 and 134 days were identified as extreme bushfire smoke days within the subregions of Sydney. Bushfire-related PM2.5 ranged from 6.3 to 115.4 µg/m3. A 0 to 10 µg/m3 increase in Bushfire PM2.5 was associated with a 3.2% (95% CI 0.3, 6.2%) increase in risk of all-cause death, cumulatively, in the 3 days following exposure. These effects were present in those aged 65 years and over, while no effect was observed in people under 65 years.

Conclusion

Bushfire PM2.5 exposure is associated with an increased risk of mortality, particularly in those over 65 years of age. This increase in risk was clearest at Bushfire PM2.5 concentrations up to 30µg/m3 above background (Non-bushfire), with possible plateauing at higher concentrations of Bushfire PM2.5.