Fine particulate matter constituents, ozone, and systemic lupus erythematosus: A large general-population cohort analysis with extended quantile g-computation.

Abstract

To investigate whether exposure to mixture of individual fine particulate matter (PM_2.5) chemical constituents was associated with incident systemic lupus erythematosus (SLE) and if ozone modifies this association and/or is associated with SLE onset. We used MarketScan claims data to assemble a cohort of adults with ≥2 years insurance coverage and no prior SLE diagnosis, with cohort entry defined as first qualifying date. From this full cohort, we identified all incident SLE cases and a 20% random sample, yielding a case-cohort of 8.3 million people. Concentrations of ozone and seven PM_2.5 constituents (ammonium, black carbon, mineral dust, sulfate, nitrate, organic matter, sea salt) were estimated at core-based statistical area/metropolitan-division level. Mean annual exposures for the 2 years prior to cohort entry were assigned to each subject. Extended quantile g-computation produced hazard ratios (HR) for the PM_2.5 mixture and weights (i.e. relative contributions to SLE onset) for individual constituents, adjusting for sex, age, chronic obstructive pulmonary disease, geographical region, and calendar year. We identified 21,485 new SLE cases. At the median ozone level (36.15 parts per billion, ppb), each quartile increase of PM_2.5 mixture was associated with a hazard ratio (HR) of 1.14 (95% confidence interval, CI 1.11-1.18) for SLE risk; mineral dust dominated this effect. Ozone was associated with SLE risk (HR 1.009 per ppb, 95% CI 1.001-1.017). PM_2.5's effect was particularly strong in settings of low ozone. This study demonstrated how PM_2.5 constituents (especially mineral dust) and ozone are associated with SLE incidence.