B97-06 Seasonal Wildfire Smoke Exposure, Circulating Proteomics, and Health Outcomes: A Molecular Epidemiological Study in Spiromics

Rationale North Americans are being exposed to increasing poor air quality due to extended wildfire smoke (WFS) seasons, with uncertain health consequences. Methods We analyzed 2,159 SPIROMICS participants with plasma proteomics (SomaScan 7K platform) across visits from 2010 to 2020. WFS-related fine particulate matter (PM2.5) and black carbon (BC) were estimated by clinical sites (n = 12) as the 120-day average preceding each visit. This window was chosen based on enumeration of days affected by smoke plumes in the area of each site. We next derived temporal trajectory patterns of WFS exposure using extended mixed models with latent classes. Baseline proteins were residualized for age, sex, race/ethnicity, obesity, smoking status, and COPD, and we applied elastic net regression to select a WFS PM2.5 protein signature, yielding individual composite protein scores as the weighted sum of selected proteins. Score performance was validated at later visits via linear regression. Finally, using linear mixed-effects models, we related the WFS-related protein score to longitudinal outcomes, including St. George’s Respiratory Questionnaire (SGRQ) scores, the Pittsburgh Sleep Quality Index, Functional Assessment of Chronic Illness Therapy subscales, and six-minute walk distance. Results The WFS PM2.5 protein signature comprised 57 upregulated and 68 downregulated proteins. Higher WFS PM2.5 and BC levels significantly associated with higher scores at visits 1, 2, and 4, regardless of diverse clinic sites or seasons (all P < 0.001), but only in the warm season at visit 5. We identified three WFS exposure patterns: (1) low/stable, (2) recent-high (higher exposure near the visit), and (3) prior-high (higher exposure earlier in the window). Both recent-high and prior-high groups had higher WFS PM2.5 and higher protein scores than the low/stable group, with no difference between each other. Per standard deviation (0.027) increase in WFS-related protein scores was associated with higher SGRQ symptom, activity, impact, and total scores (beta=1.31-2.22, all P < 0.001), worse sleep quality (beta=0.24, P = 0.003), lower assessments in fatigue, physical, emotional, and functional well-being (all P < 0.05), and a shorter six-minute walk distance (beta= -16.68, P < 0.001). Conclusion Using a deeply phenotyped cohort and estimated incremental air pollution due to WFS, we defined a plasma proteomic score of seasonal WFS exposure that correlates with multidomain quality of life and exercise capacity, which requires external validation. Our exposure trajectory results suggest that cumulative WFS exposure may be more influential than exposure timing within a WFS season.