Effects of long-term exposure of particle components on inpatient stroke hospitalizations in U.S. adults ages 40 years and up using Bayesian kernel machine regression (BKMR)

Abstract

BACKGROUND: Air pollutants, including PM2.5, have been shown to adversely affect health; however, few studies have investigated the long-term exposure of particle components in conjunction with PM2.5 and ozone to assess their individual and additive effects on cerebrovascular incidents such as stroke. AIM: We aim to utilize a Bayesian Kernel machine regression (BKMR) to assess the individual and join effects of air pollutants including 15 different particle components such as organic carbon (OC), elemental carbon (EC), copper (Cu), and zinc (Z), along with PM2.5 and ozone, on counts of inpatient Parkinson’s hospitalizations for adults ages 40 years and up. METHODS: Inpatient records were collected from the State Inpatient Databases which included hospitals from 12 U.S. states ranging in years from 2000 through 2016. We also included temperature from Daymet and variables from the U.S. census to control for socio-economic status. All variables were aggregated to the annual level. RESULTS We observed a significant increase of 2.00 (95%CI: 1.72,2.29), 5.87 (95%CI: 5.57,6.16), and 9.81 (95%CI: 9.51,10.12) in the number of inpatient stroke hospitalizations each year at the 25th, 50th, and 75th percentiles of pollutant mixture, respectively. CONCLUSIONS: Our results indicate that the mixture of pollutants greatly contribute to the increase in the number of stroke hospitalizations each year and that the effects of short-term exposures of particle components on stroke hospitalizations should be assessed next. KEYWORDS: Stroke, PM Components, PM2.5, Ozone, BKMR