Estimation of PM_2.5 mortality burden in China with new exposure estimation and local concentration-response function

Highlights

• Developing LUR model for PM_2.5 combined satellite and surface observation.

• The spatial resolution of our study is 1 km × 1 km, more elaborate than previous studies.

• Updating PM_2.5-related premature deaths using local concentration–response function.

• Our estimation of premature deaths shows higher results than previous understanding.

Abstract

The estimation of PM_2.5-related mortality is becoming increasingly important. The accuracy of results is largely dependent on the selection of methods for PM_2.5 exposure assessment and Concentration-Response (C-R) function. In this study, PM_2.5 observed data from the China National Environmental Monitoring Center, satellite-derived estimation, widely collected geographic and socioeconomic information variables were applied to develop a national satellite-based Land Use Regression model and evaluate PM_2.5 exposure concentrations within 2013–2015 with the resolution of 1 km × 1 km. Population weighted concentration declined from 72.52 μg/m³ in 2013 to 57.18 μg/m³ in 2015. C-R function is another important section of health effect assessment, but most previous studies used the Integrated Exposure Regression (IER) function which may currently underestimate the excess relative risk of exceeding the exposure range in China. A new Shape Constrained Health Impact Function (SCHIF) method, which was developed from a national cohort of 189,793 Chinese men, was adopted to estimate the PM_2.5-related premature deaths in China. Results showed that 2.19 million (2013), 1.94 million (2014), 1.65 million (2015) premature deaths were attributed to PM_2.5 long-term exposure, different from previous understanding around 1.1–1.7 million. The top three provinces of the highest premature deaths were Henan, Shandong, Sichuan, while the least ones were Tibet, Hainan, Qinghai. The proportions of premature deaths caused by specific diseases were 53.2% for stroke, 20.5% for ischemic heart disease, 16.8% for chronic obstructive pulmonary disease and 9.5% for lung cancer. IER function was also used to calculate PM_2.5-related premature deaths with the same exposed level used in SCHIF method, and the comparison of results indicated that IER had made a much lower estimation with less annual amounts around 0.15–0.5 million premature deaths within 2013–2015.

Introduction

PM_2.5 pollution has caused people's growing concern about its potential adverse health effects. The Global Burden of Disease (GBD) 2015 assessment indicated that PM_2.5 contributed 4.24 million premature deaths around the world (Forouzanfar et al., 2016; IHME and HEI, 2017) and a number of epidemiological studies suggested that long-term exposure to high concentrations of PM_2.5 is positively associated with deaths from stroke, lung cancer (LC), chronic obstructive pulmonary disease (COPD) and ischemic heart disease (IHD) (Anderson et al., 2012; Dockery et al., 1993; Ren et al., 2006). For China, rapid economic development and urbanization over the past years have greatly damaged the environment (Lin et al., 2016b; Richter et al., 2005; Liu et al., 2016a, 2018) and Lin et al., 2016a revealed a significant association between PM_2.5 and mortality and estimated mortality burden of PM_2.5 in six Chinese cities. About 98% people are exposed to the severe PM_2.5 pollution that could not reach the World Health Organization (WHO) Air Quality Guidelines (AQG) of 10 μg/m³ for annual mean (Apte et al., 2015). Therefore, it is extremely important to estimate China's premature deaths related by PM_2.5 for improving the control policy against air pollution.

Prior studies have suggested diverse results of premature deaths associated with PM_2.5 in China. For the year of 2010, 1.1 million (HEI, 2013), 1.28 million (OECD, 2014), 1.36 million (Lelieveld et al., 2015), 1.26 million (Xie et al., 2016) and 1.6 million (Rohde and Muller, 2015) of total premature deaths were estimated respectively. These distinctions are mainly caused by different selections of exposure assessment methods and concentration-response (C-R) functions, which are also the most significant parts of health effect assessments (Maji et al., 2017).

Most previous PM_2.5 ambient exposure estimation mainly depended on station monitoring interpolation method, satellite-based aerosol optical depth (AOD) model, air quality model and land use regression (LUR) model. However, each approach has somewhat limitations such as the mis-classification of interpolation method (Rohde and Muller, 2015), limited resolution of satellite device (Ma et al., 2016), huge computational cost of air quality model (Marshall et al., 2008) and space limitations of LUR (Johnson et al., 2010), restricting its application for predicting fine-scale air pollution concentrations over large geographical areas. More recently, a small number of LUR models combined with satellite estimation have been presented with broader spatial coverage, throughout United States (Novotny et al., 2011; Beckerman et al., 2013; Bechle et al., 2015), Europe (Beelen et al., 2009) and Australia (Knibbs et al., 2014). Knibbs et al., 2016 performed independent evaluations of national satellite-based LUR model in Australia and demonstrated its reliability. Combining the advantages of the satellite estimation and LUR models, national satellite-based LUR model for PM_2.5 is considered as a powerful tool to provide the accurate estimation of PM_2.5 exposure levels and its long-term health effect at a fine resolution in a large area (Novotny et al., 2011), but as far what we know, it has not been reported in China.

Another significant section of health effect assessment is the selection of C-R function. Previous studies have estimated premature mortality related by PM_2.5 with linear extrapolation of the C-R function which was announced in cohort studies (Lelieveld et al., 2013; Zhang et al., 2008). Later, the Integrated Exposure Regression (IER) was developed by Burnett et al. (2014) to produce a more reasonable prediction on relative risks by the integration of available information, including ambient air pollution (AAP) and second-hand smoke (SHS), and IER has been applied in a number of recent estimation of premature deaths related by PM2.5 (Liu et al., 2016b; Liu et al., 2017; Song et al., 2017; Xie et al., 2016; Wang et al., 2017). More recently, based on a large national cohort of 189,793 Chinese men, Yin et al., 2017 developed a Shape Constrained Health Impact Function (SCHIF), which could provide the exposure-response relationship between PM_2.5 and mortality over a much broader range of exposure than previously studied specifically at high PM2.5 concentrations. Yin et al., 2017 also showed that IER may underestimate the excess relative risk of PM_2.5 that exceeds the exposure range in China and other developing countries. Therefore, it is necessary to update the health burden attributable to PM_2.5 in China as well as quantitatively examine the impacts caused by various selections of C-R function.

In this study, Chinese national air monitoring network, satellite-derived estimation, wide collection of geographic and socioeconomic information were utilized to develop a national satellite-based LUR model for PM_2.5. Exposed levels across China within 2013–2015 were quantified with the fine resolution of 1 km × 1 km. Based on this long-term exposure, the estimation on premature mortality attributable to PM_2.5 in China was updated by using new SCHIF method, whose differences were analyzed quantitatively with the widely used IER method.