Air pollution and gastrointestinal diseases in Utah

Abstract

The valleys of northern Utah, where most of Utah’s population resides, experience episodic air pollution events well in excess of the National Ambient Air Quality Standards. Most of the events are due to an accumulation of particulate matter during persistent cold air pools in winter from both direct emissions and secondary chemical reactions in the atmosphere. High wintertime ozone concentrations are occasionally observed in the Uintah Basin, in addition to particulate matter. At other times of the year, blowing dust, wildland fires, fireworks, and summertime ozone formation contribute to local air pollution. The objective of this dissertation is to investigate one facet of the health effects of Utah’s air pollution on its residents: the acute impacts of air pollution on gastrointestinal (GI) disease.

To study the health effects of these episodic pollution events, some measure of air pollution exposure must be matched to the health data. Time and place are used to link the health data for a person with the pollution data. This dissertation describes the method of kriging data from the sparse pollution monitoring network to estimate personal air pollution history based on the zip code of residence. This dissertation then describes the application of these exposure estimates to a health study on GI disease.

The purpose of the GI study is to retrospectively look at two groups of patients during 2000-2014: those with autoimmune disease of the GI tract (inflammatory bowel disease, IBD) and those with allergic disease of the GI tract (eosinophilic esophagitis, EoE) to determine whether disease exacerbations occur more commonly during and following periods of poor air quality compared to periods of good air quality. The primary analysis method is case crossover design. In addition to using the kriged air pollution estimates, the analysis was repeated using simpler empirical estimation methods to assess whether the odds ratios are sensitive to the air pollution estimation method.

The data suggests an association between particulate matter smaller than 2.5 microns and prednisone prescriptions, gastrointestinal infections in general, clostridium difficile infections specifically, and hospitalizations among people who have at least five entries of IBD diagnosis codes in their medical records. EoE exacerbations appear to be associated with high concentrations of particulate matter as well as ozone.