High traffic pollution predicts rapid decline in lung function among children with cystic fibrosis

Abstract

Background: Environmental exposures and community characteristics (i.e. geomarkers) have been shown to influence pulmonary outcomes; however, the ability for geomarkers to predict rapid lung-function decline among cystic fibrosis (CF) patients has not been studied.Methods: We applied a longitudinal model to predict rapid lung function decline, defined using forced expiratory volume in 1 second (FEV1) relative to patient and center specific norms, based on registry data for 185 CF patients aged 6-20 years receiving care at Cincinnati Children’s Hospital CF Center between 2012 and 2017. Clinical and demographic characteristics (F508del alleles, sex, Medicaid insurance use and pancreatic enzymes) were used as predictors in our initial model. Residential addresses were geocoded and average exposure to elemental carbon attributable to traffic sources (ECAT), a marker of traffic-related air pollution (TRAP), was estimated for the three months prior to each clinical encounter using a previously validated spatiotemporal land-use model. Neighborhood deprivation was derived at the census-tract level using an index comprised of American Community Survey measures related to poverty, education, housing, and access to healthcare. Greenness near the home was derived from the normalized differential vegetation index. We implemented real-time prediction after including geomarkers with the clinical and demographic predictors. Covariates were selected with the likelihood ratio test (LRT). Model fit including/excluding geomarkers was assessed using Akaike information criteria (delta-AIC).Results: Including geomarkers yielded a significantly better fit and improved prediction of FEV1 compared to a model with only clinical/demographic characteristics (LRT statistic: 45.4, P<0.0001; delta-AIC=27.4). Specifically, an increase of 1 ug/m3 of ECAT was associated with a1.13% predicted/yr (95% CI: 0.33, 1.93% predicted/yr) more rapid decline.Conclusion: Exposure to TRAP is an important predictor of pulmonary decline in CF that may be used to enhance clinician assessments of prognosis and enable personalized environmental health interventions.