County-level air quality and the prevalence of diagnosed chronic kidney disease in the US Medicare population

Background Considerable geographic variation exists in the prevalence of chronic kidney disease across the United States. While some of this variability can be explained by differences in patient-level risk factors, substantial variability still exists. We hypothesize this may be due to understudied environmental exposures such as air pollution. Methods Using data on 1.1 million persons from the 2010 5% Medicare sample and Environmental Protection Agency air-quality measures, we examined the association between county-level particulate matter <= 2.5 mu m (PM2.5) and the prevalence of diagnosed CKD, based on claims. Modified Poisson regression was used to estimate associations (prevalence ratios [PR]) between county PM2.5 concentration and individual-level diagnosis of CKD, adjusting for age, sex, race/ethnicity, hypertension, diabetes, and urban/rural status. Results Prevalence of diagnosed CKD ranged from 0% to 60% by county (median = 16%). As a continuous variable, PM2.5 concentration shows adjusted PR of diagnosed CKD = 1.03 (95% CI: 1.02-1.05; p<0.001) for an increase of 4 mu g/m(3) in PM2.5. Investigation by quartiles shows an elevated prevalence of diagnosed CKD for mean PM2.5 levels >= 14 mu g/m(3) (highest quartile: PR = 1.05, 95% CI: 1.03-1.07), which is consistent with current ambient air quality standard of 12 mu g/m 3 , but much lower than the level typically considered healthy for sensitive groups (similar to 40 mu g/m(3)). Conclusion A positive association was observed between county-level PM2.5 concentration and diagnosed CKD. The reliance on CKD diagnostic codes likely identified associations with the most severe CKD cases. These results can be strengthened by exploring laboratory-based diagnosis of CKD, individual measures of exposure to multiple pollutants, and more control of confounding.