Wan, LeiBambrick, HilaryTong, Michael2026-06-072026-06-070195-9255WOS:001419592900001ORCID:/0000-0002-9694-9207/work/216711842ORCID:/0000-0001-5361-950X/work/216715357https://hdl.handle.net/1885/733810059Background: As China's population is aging rapidly, understanding the shifts in PM2.5 attributable mortality within this context is crucial for informing future clean air policies. Methods: We adopted a Fusion relative risk model, combined with 100 m resolution age structure data, to better estimate China's age- and cause-specific mortality attributable to PM2.5 over 2010–2019, and projected attributable mortality in 2030 and 2060 at a 0.1° spatial resolution under three scenarios: Baseline, Carbon-peak and Carbon-neutral. We also assessed the impact of population aging using a decomposition method at the grid level. Results: PM2.5 attributable deaths declined by 9.2 % from 2010 (1.95 million, 95 % UI: 1.80–2.09) to 2019 (1.77 million, 95 % UI: 1.64–1.90), with population aging contributing an increase of 0.48 million deaths. The elderly population constituted over 70 % of total attributable mortality during 2010–2019, and this share is expected to increase to over 90 % in 2060 under three future scenarios. Under Baseline scenario, attributable deaths are expected to increase, with population aging as the major contributor. Under Carbon-peak scenario, the projected mortality declines over 2019–2030 and 2030–2060 will be partly offset by population aging. Under Carbon-neutral scenario, population aging is projected to increase attributable deaths by 0.57 million and 1.27 million over the two periods, largely offsetting the reductions achieved by the declines in PM2.5 concentrations and cause-specific baseline mortality rates. Conclusions: Population aging is the main factor that increases PM2.5 attributable mortality. Specific measures considering the vulnerability of the elderly are needed to further alleviate future health burden from air pollution.This study is supported by the Australian Government Research Training Program Scholarship. We would like to thank Prof Richard Burnett for providing the parameters for the Fusion relative risk model. We would also like to thank Prof Qiang Zhang and Dr. Yang Liu at Tsinghua University for providing simulated PM2.5 data under future scenarios.11en© 2025 The AuthorsCarbon-neutralCarbon-peakFusion relative risk modelPM attributable mortalityPopulation aging202510.1016/j.eiar.2025.10782310.1016/j.eiar.2025.10782385214697505