Investigating associations between anti-nuclear antibody positivity and combined long-term exposures to NO2, O3, and PM2.5 using a Bayesian kernel machine regression approach

Abstract

BackgroundAir pollution has many adverse health effects, but the combined or synergistic effects of multiple ambient air pollutants on anti-nuclear antibodies (ANA, a serologic marker of systemic autoimmune rheumatic disease, SARDs) have never been assessed. ObjectiveTo flexibly model ANA and individual and joint associations of long-term exposures to nitrogen dioxide (NO2), ozone (O3), and fine particles matter (PM2.5) using a Bayesian Kernel machine regression (BKMR) approach and to compare the results to those from individual logistic regressions. MethodsSerum ANA positivity was determined for randomly selected CARTaGENE general population subjects in Quebec, Canada. CARTaGENE is a public research platform created for investigating the associations of environmental, genomic, and lifestyle factors on chronic diseases. Ambient NO2, O3, and PM2.5 estimates, derived from ground-measurement and chemical-transport-model simulated concentrations, were assigned to subjects based on residential postal codes at the time of blood collection. Our models adjusted for age, sex, French Canadian origin, smoking, and family income. ResultsConcentrations of NO2, O3, and PM2.5 were closely correlated in space. In the 5485 CARTaGENE subjects studied, we did not see clear associations between NO2, PM2.5 or O3 and ANA positivity, with either the BKMR or logistic models. ConclusionsBKMR did not uncover associations between ANA positivity and individual levels or combined exposures of NO2, O3, and PM2.5; neither did simpler logistic models. Additional studies (in younger populations, in distinct race/ethnicity groups, and/or in jurisdictions with high air pollution levels) would be helpful to reinforce current findings.