Causal inference for the effect of environmental chemicals on chronic kidney disease

Abstract

The impacts of environmental chemicals on the decline of kidney function have been suggested by a limited number of statistical and animal studies. Thus, those exposures may be modifiable risk factors for chronic kidney disease. Some of the chemicals, such as Perfluoroalkyl acid (PFA), are pervasive throughout our environment, determining their health effects is an important public health concern. In this study, we examined cross-sectional data from the 2009–2010 cycle of the National Health and Nutrition Examination Survey (NHANES) using a statistical causal inference method-generalized propensity score method, to determine the links between concentrations of several major environmental chemicals and kidney function measured by the estimated glomerular filtration rate (eGFR). Various generalized propensity score estimation methods including Hirano-Imbens, additive spline, and a generalized additive model were compared. Among the examined environmental chemicals, each of the statistical models used associated an increase in PFA concentration with a decline in eGFR. We conclude that PFA is a modifiable risk factor for chronic kidney disease and the statistical causal method produces credible results in estimating the effect of chemical exposures on a continuous measure of kidney functions with an observational dataset.