Association Between Nitrogen Dioxide Pollution and Cause-Specific Mortality in China: Cross-Sectional Time Series Study (Preprint)

Abstract

BACKGROUND Nitrogen dioxide (NO<sub>2</sub>) has been frequently linked to a range of diseases and associated with high rates of mortality and morbidity worldwide. However, there is limited evidence regarding the risk of NO<sub>2</sub> on a spectrum of causes of mortality. Moreover, adjustment for potential confounders in NO<sub>2</sub> analysis has been insufficient, and the spatial resolution of exposure assessment has been limited. OBJECTIVE This study aimed to quantitatively assess the relationship between short-term NO<sub>2</sub> exposure and death from a range of causes by adjusting for potential confounders in Guangzhou, China, and determine the modifying effect of gender and age. METHODS A time series study was conducted on 413,703 deaths that occurred in Guangzhou during the period of 2010 to 2018. The causes of death were classified into 10 categories and 26 subcategories. We utilized a generalized additive model with quasi-Poisson regression analysis using a natural cubic splines function with lag structure of 0 to 4 days to estimate the potential lag effect of NO<sub>2</sub> on cause-specific mortality. We estimated the percentage change in cause-specific mortality rates per 10 μg/m<sup>3</sup> increase in NO<sub>2</sub> levels. We stratified meteorological factors such as temperature, humidity, wind speed, and air pressure into high and low levels with the median as the critical value and analyzed the effects of NO<sub>2</sub> on various death-causing diseases at those high and low levels. To further identify potentially vulnerable subpopulations, we analyzed groups stratified by gender and age. RESULTS A significant association existed between NO<sub>2</sub> exposure and deaths from multiple causes. Each 10 μg/m<sup>3</sup> increment in NO<sub>2</sub> density at a lag of 0 to 4 days increased the risks of all-cause mortality by 1.73% (95% CI 1.36%-2.09%) and mortality due to nonaccidental causes, cardiovascular disease, respiratory disease, endocrine disease, and neoplasms by 1.75% (95% CI 1.38%-2.12%), 2.06% (95% CI 1.54%-2.59%), 2.32% (95% CI 1.51%-3.13%), 2.40% (95% CI 0.84%-3.98%), and 1.18% (95% CI 0.59%-1.78%), respectively. Among the 26 subcategories, mortality risk was associated with 16, including intentional self-harm, hypertensive disease, and ischemic stroke disease. Relatively higher effect estimates of NO<sub>2</sub> on mortality existed for low levels of temperature, relative humidity, wind speed, and air pressure than with high levels, except a relatively higher effect estimate was present for endocrine disease at a high air pressure level. Most of the differences between subgroups were not statistically significant. The effect estimates for NO<sub>2</sub> were similar by gender. There were significant differences between the age groups for mortality due to all causes, nonaccidental causes, and cardiovascular disease. CONCLUSIONS Short-term NO<sub>2</sub> exposure may increase the risk of mortality due to a spectrum of causes, especially in potentially vulnerable populations. These findings may be important for predicting and modifying guidelines for NO<sub>2</sub> exposure in China.