Long-term exposure to wildfire-related PM2·5 and overall survival of older patients with lung cancer: a retrospective, registry-based cohort study.

Abstract

Background: Fine particulate matter (PM_2·5) has been identified as a key risk factor for lung cancer incidence and survival. PM_2·5 from wildfire smoke is an increasingly important contributor to total PM_2·5 in the USA, driven by climate change, and potentially more toxic than non-wildfire PM_2·5 due to its smaller particle size and higher toxic contents. However, the effects of wildfire-related PM_2·5 on the overall survival of patients with lung cancer remain unclear.

Methods: This retrospective cohort study included patients aged 65 years or older with histologically confirmed primary small-cell lung cancer or non-small-cell lung cancer (as defined by the International Classification of Diseases for Oncology, Third Edition) from the Surveillance, Epidemiology, and End Results (SEER)-Medicare database between Jan 1, 2006, and Dec 31, 2019, for whom residential zip code-based estimates of wildfire-related PM_2·5 and non-wildfire PM_2·5 were available. Individuals diagnosed at the in situ stage, without continuous enrolment in Medicare Parts A and B, or with missing exposure or covariate data were excluded. Participants were followed up annually from diagnosis until death, loss to follow-up, or the end of the study in 2019, whichever came first. Daily estimates of ambient wildfire-related PM_2·5 and total PM_2·5 across the contiguous USA (from 2006 to 2019) were obtained from previously developed and validated models and spatially aggregated to zip code level. Non-wildfire PM_2·5 was calculated as the difference between total PM_2·5 and wildfire-related PM_2·5, and negative values were excluded. The main outcome was all-cause mortality. A time-varying Cox proportional hazards model was applied to estimate the primary exposure of interest of the long-term effects of 3-year moving averages of wildfire-related PM_2·5 and non-wildfire PM_2·5 concentrations on overall survival.

Findings: Between Jan 1, 2008, and Dec 31, 2019 (the analysis period), 414 016 patients with lung cancer contributed 1 211 490 person-years of follow-up, with a median follow-up of 2 years (IQR 1-4). Of these, 208 468 (50·4%) patients were female, 205 548 (49·6%) were male, 356 356 (86·1%) were White, 36 914 (8·9%) were Black, 18 982 (4·6%) were Asian or Pacific Islander, and 1269 (0·3%) were American Indian or Alaska Native; race or ethnicity was unknown for 495 patients (0·1%). The median concentrations of PM_2·5 were 0·28 μg/m³ (IQR 0·21-0·39) for wildfire-related PM_2·5 and 8·11 μg/m³ (6·84-9·47) for non-wildfire PM_2·5. Each 1-SD increase in wildfire-related PM_2·5 or non-wildfire PM_2·5 was associated with an increased risk of death for patients with lung cancer (HR 1·0229, 95% CI 1·0189-1·0270; p<0·0001 for wildfire-related PM_2·5 and HR 1·0384, 1·0335-1·0432; p<0·0001 for non-wildfire PM_2·5). The HR per 1-μg/m³ increase in PM_2·5 concentration was 1·0782 (1·0641-1·0924; p<0·0001) for wildfire-related PM_2·5 and 1·0171 (1·0149-1·0192; p<0·0001) for non-wildfire PM_2·5. Overall, 696 (95% CI 576-814) average annual excess deaths in this cohort were attributable to wildfire-related PM_2·5 and 3465 (3061-3864) were attributable to non-wildfire PM_2·5.

Interpretation: Wildfire-related PM_2·5 showed greater toxicity per unit exposure than non-wildfire PM_2·5. As wildfire-related PM_2·5 is the fastest-growing contributor to ambient air pollution, there is an urgent need for updated strategies to manage wildfires and mitigate their effects.

Funding: National Institute of Environmental Health Sciences, National Cancer Institute, National Institutes of Health, National Center for Advancing Translational Sciences, Centers for Disease Control and Prevention's National Program of Cancer Registries, and Stony Brook University.