Combined health effects of PM2.5 components on respiratory mortality in short-term exposure using BKMR: A case study in Sichuan, China

Siwei Zhai; Jing Zeng; Yi Zhang; Jingfei Huang; Xuelin Li; Wei Wang; Tao Zhang; Ying Deng; Fei Yin; Yue Ma

doi:10.1016/j.scitotenv.2023.165365

Combined health effects of PM_2.5 components on respiratory mortality in short-term exposure using BKMR: A case study in Sichuan, China

Sci Total Environ. 2023 Nov 1:897:165365. doi: 10.1016/j.scitotenv.2023.165365. Epub 2023 Jul 10.

Authors

Siwei Zhai¹, Jing Zeng², Yi Zhang¹, Jingfei Huang¹, Xuelin Li¹, Wei Wang¹, Tao Zhang¹, Ying Deng², Fei Yin¹, Yue Ma³

Affiliations

¹ West China School of Public Health and West China Fourth Hospital, Sichuan University, China.
² Sichuan Provincial Disease Prevention and Control Center, China.
³ West China School of Public Health and West China Fourth Hospital, Sichuan University, China. Electronic address: gordonrozen@qq.com.

PMID: 37437633
DOI: 10.1016/j.scitotenv.2023.165365

Abstract

One of the major causes of global mortality is respiratory diseases. Fine particulate matter (PM_2.5) increased the risk of respiratory death in short-term exposure. PM_2.5 is the chemical mixture of components with different health effects. The combined health effects of PM_2.5 are determined by the role of each component and the potential interaction between components, but they have not been studied in short-term exposure. Sichuan Province (SC), with high respiratory mortality and heavy PM_2.5 pollution, had distinctive regional differences in four regions in sources and proportions of PM_2.5, so it was divided into four regions to explore the combined health effects of PM_2.5 components on respiratory mortality in short-term exposure and to identify the main hazardous components. Due to the multicollinear, interactive, and nonlinear characteristics of the associations between PM_2.5 components and respiratory mortality, Bayesian kernel machine regression (BKMR) was used to characterize the combined health effects, along with quantile-based g-computation (QGC) as a reference. Positive combined effects of PM_2.5 were found in all four regions of Sichuan using BKMR with excess risks (ER) of 0.0101-0.0132 (95 % CI: 0.0093-0.0158) and in the central basin and northwest basin using QGC with relative risks (RR) of 1.0064 (95 % CI: 1.0039, 1.0089) and 1.0044 (95 % CI: 1.0022, 1.0066), respectively. In addition, the adverse health effect was larger in cold seasons than that in warm seasons, so vulnerable people should reduce outdoor activities in heavily polluted days, especially in the cold season. For the components of PM_2.5, the BC and OM mainly from traffic, dominated the adverse health effects on respiratory mortality. Furthermore, NO₃^- might aggravate the adverse health effects of BC/OM. Therefore, BC/OM and NO₃^- should be focused together in air pollution control.

Keywords: Bayesian kernel machine regression (BKMR); PM(2.5) components; Quantile-based g-computation (QGC); Respiratory mortality; Sichuan.

MeSH terms

Air Pollutants* / analysis
Air Pollutants* / toxicity
Air Pollution* / adverse effects
Air Pollution* / analysis
Bayes Theorem
China / epidemiology
Environmental Exposure
Humans
Particulate Matter / analysis
Respiratory Tract Diseases* / chemically induced

Substances

Air Pollutants
Particulate Matter