Long-term ambient air pollution exposure, metabolomic signatures, and incident pulmonary hypertension: A large prospective cohort study.
Evidence linking long-term multipollutant exposure with incident pulmonary hypertension (PH) and circulating metabolic profiles is limited. We investigated individual and joint air pollution exposures in relation to incident PH, characterized pollution-related circulating metabolic signatures, and evaluated their potential mediating roles. This prospective cohort included 444,346 participants without PH at baseline who had air pollution exposure estimates and nuclear magnetic resonance metabolomics data. PM2.5, PM10, NO2, and NOx were assessed as individual pollutants and jointly summarized using an air pollution score (APS). Exposure-related metabolic signatures were derived by elastic-net regression, and their associations with incident PH and potential mediating contributions were evaluated using Cox proportional hazards models and mediation analyses. During 13.58 years of median follow-up, 2328 participants developed PH. Higher APS was associated with increased PH risk (hazard ratio [HR] per 1-SD increment, 1.12; 95% confidence interval [CI], 1.08-1.17). Individual pollutants were also positively associated with PH, with HRs ranging from 1.06 to 1.93 per 10-μg/m3 increment. Elastic-net regression yielded 105 metabolites for the APS signature and 52-123 metabolites for pollutant-specific signatures, mainly involving lipoprotein-related measures, fatty acids, and amino acids. These signatures were also associated with incident PH (HRs per 1-SD increment, 1.12-1.24). The overall metabolic signature mediated 11.74% (95% CI, 7.91%-18.95%) of the APS-PH association, whereas pollutant-specific signatures mediated 11.38%-26.00% of the corresponding pollutant-PH associations. Both individual pollutants and the weighted APS were positively associated with incident PH, with circulating metabolic alterations potentially contributing to these associations.