This paper is in the following e-collection/theme issue:

Environmental Health (66) Prevention and Health Promotion (879) Cross-Sectional Studies in Public Health (203) Public Health (158) Statistical Methods for Surveillance and Population Health (146) Innovative Methods in Public Health and Surveillance (305)

Published on in Vol 10 (2024)

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/44648, first published .
Association Between Nitrogen Dioxide Pollution and Cause-Specific Mortality in China: Cross-Sectional Time Series Study

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

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

Authors of this article:

Jie Zeng1 Author Orcid Image ;   Guozhen Lin2, 3 Author Orcid Image ;   Hang Dong2, 3 Author Orcid Image ;   Mengmeng Li4 Author Orcid Image ;   Honglian Ruan5 Author Orcid Image ;   Jun Yang5 Author Orcid Image
Article Authors Cited by (1) Tweetations (2) Metrics

    Original Paper

    1Department of Internet Medical Center, Guangdong Second Provincial General Hospital, Guangzhou, China

    2Guangzhou Center for Disease Control and Prevention, Guangzhou, China

    3Institute of Public Health, Guangzhou Medical University and Guangzhou Center for Disease Control and Prevention, Guangzhou, China

    4State Key Laboratory of Oncology in South China, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangzhou, China

    5School of Public Health, Guangzhou Medical University, Guangzhou, China

    *these authors contributed equally

    Corresponding Author:

    Jun Yang, PhD

    School of Public Health

    Guangzhou Medical University

    No. 1 Xinzao Road

    Xinzao Town, Panyu District

    Guangzhou, 511436

    China

    Phone: 86 020 37103532

    Email: yangjun_eci@jnu.edu.cn


    Background: Nitrogen dioxide (NO2) 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 NO2 on a spectrum of causes of mortality. Moreover, adjustment for potential confounders in NO2 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 NO2 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 NO2 on cause-specific mortality. We estimated the percentage change in cause-specific mortality rates per 10 μg/m3 increase in NO2 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 NO2 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 NO2 exposure and deaths from multiple causes. Each 10 μg/m3 increment in NO2 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 NO2 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 NO2 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 NO2 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 NO2 exposure in China.

    JMIR Public Health Surveill 2024;10:e44648

    doi:10.2196/44648

    Keywords

    nitrogen dioxide; cause-specific mortality; stratification effect; vulnerable subpopulations; China 


    Ambient air pollution is one of the greatest environmental risks to human health, with 8 million deaths reported worldwide every year [1]. Rapid industrialization in China over the last decades has significantly increased the emission of pollutants. Both short and long-term exposure to ambient air pollution increases the risk of death, years of life lost, and years of disability, thereby increasing the burden of disease [2-6]. Exposure to air pollution, including particulate matter (PM), nitrogen oxides (NOx), ozone, and sulfur dioxide, may increase the risk of respiratory disease through oxidative damage, which occurs via inflammatory injury and the production of reactive oxygen species [7].

    Nitrogen dioxide (NO2) is a toxic gas that is present in traffic emissions, and chronic exposure to NO2 is associated with respiratory inflammation, allergies, infections, and other symptoms. The current acceptable limits of NO2 exposure in China are 105 ppb (200 μg/m3) over 1 hour and an average of 21 ppb (40 μg/m3) annually [8]. In recent decades, however, the concentration of NO2 in China has exceeded the global average level [9]. Guangzhou is the capital and largest city of Guangdong province in China, with a population of 13,964,637 [10]. According to data released by the World Health Organization [11] in 2018, the annual average concentration of NO2 reached 50 μg/m3, which exceeds air quality standards (25 μg/m3).

    Evidence clearly demonstrates that short-term exposure to NO2 is related to an increased risk of mortality from different diseases, showing independent and linear trends [12-16]. For instance, a multicity analysis showed NO2 exposure increased the all-cause mortality rates, as well as mortality due to cardiovascular and respiratory diseases [12]. However, other studies have reported a lower mortality risk due to chronic diseases in the event of NO2 exposure [17]. Research on the mechanisms by which NO2 affects mortality has come to limited conclusions. For example, it is not yet known whether NO2 exposure can cause endocrine disease and neoplasms.

    Regarding the effect of NO2 on diseases, there is less evidence for the different effects caused by stratification of meteorological factors. The significance of other meteorological risk factors, such as temperature, humidity, air pressure, and wind speed, has been less frequently explored. In addition, most of the studied disease outcomes affected by NO2 have included only cardiovascular and respiratory diseases. Moreover, the influence of age and gender on the association between NO2 and mortality is still unclear. Therefore, it is essential to fully explore predisposing diseases associated with NO2 exposure.

    To this end, we aimed to examine the relationship between short-term NO2 exposure and a spectrum of causes of mortality in Guangzhou to identify the influencing factors. Our research may offer unique insights into the health effects of NO2 and the underlying factors and help formulate public policies to protect potentially vulnerable subgroups [18].


    Sample and Data

    The daily death count in Guangzhou during the period between 2010 and 2018 was obtained from the city registry, which included all permanent residents and is published by the Guangzhou Center for Disease Control and Prevention [19]. Using the 10th Revision of the International Classification of Diseases (ICD-10), the causes of the deaths were classified as all causes (A00-Z99), nonaccidental causes (A00-R99), cardiovascular diseases (I00-I99), neoplasms (C00-D48), endocrine diseases (D50-D89, E00-E90), nervous system diseases (G00-G99), respiratory diseases (J00-J99), digestive diseases (K00-K93), genitourinary diseases (N00-N99), and external causes (V01-Y89). These 10 groups were divided further into 26 subgroups, with each group containing at least an average number of deaths per day, ensuring that the model was convergent.

    Measures of Variables

    Data on the average daily level of NO2 from 2010 to 2018 were obtained from the Guangzhou Bureau of Environmental Protection. The data were collected from 11 fixed-point air pollution monitoring stations across the city (Figure 1). At the same time, the daily concentration of atmospheric dynamic diameter particulate matter (PM10 or PM2.5), ozone (O3), sulfur dioxide (SO2), carbon monoxide (CO), and other gaseous pollutants were obtained at the same sites and were used to adjust for potential hybrid co-pollutants during model building. Exposure levels were determined by calculating the average daily concentration level at each monitoring station. General data on the mean temperature (°C), mean relative humidity (%), mean air pressure (hPa), sunshine duration (h), wind speed (km/h), and daily air quality index were obtained from the China Meteorological Data Sharing Service System [20].

    ‎
    Figure 1. Geographic map showing the distribution of the 11 environmental monitoring stations in Guangzhou, Guangdong Province, China.

    Models and Data Analysis Procedure

    A generalized additive model with Poisson regression analysis was used to determine the relationship between NO2 exposure and each cause of mortality after adjusting for potential covariates. A natural cubic spline function (NS) of 7 degrees of freedom (df) per year was used to control for long-term and seasonal trends in mortality. In the df adjustment of other weather factors, the NS functions of 6, 3, and 3 dfs were used to adjust the dfs of temperature, relative humidity, wind speed, and atmospheric pressure, respectively. The day of the week and public holidays were added to the model as indicator variables. The method and selection of model parameters have been described in a previous research analysis [21,22]. After establishing the basic model, we used an NS of 2 to 6 dfs to investigate the relationship between NO2 concentration and mortality as a result of each cause of death. Based on the most likely days of lag, a moving average method was used to correlate the cumulative effects of pollutants over the defined lag period. This is consistent with a previous study [15,23] that selected an empirical maximum of 4 dfs for NO2 and used a lag model structure of 0 to 4 days to obtain an estimate of the cumulative risk of NO2 exposure during the first 4 days.

    To identify potentially vulnerable subpopulations at a particular risk of mortality due to NO2 exposure on a spectrum of mortality causes, stratified analyses were conducted based on age groups (0-64 and ≥65 years) and gender (men and women) based on the causes of death that showed a significant association with NO2. In addition, in order to analyze the interactive health effects between meteorological factors and NO2, the stratification analyses were conducted based on the levels of temperature, humidity, wind speed, and air pressure, with their median values as the cutoff. The difference between the 2 groups was analyzed using a z test, as follows [24]:

    where E1 and E2 are the logarithms of risk ratios for 2 categories, such as men and women or 0-64 years and ≥65 years, and SE (E1) and SE (E2) were used to indicate the relevant standard error [25,26].

    Multiple sensitivity analyses were conducted to evaluate the robustness of the model and parameter validation estimates. The dfs for the time variable were set as 3 years to 7 years, and the dfs for the weather variable were set as 3 to 6. To adjust for the mixing effect of temperature during longer lag days, 2-dimensional, cross-basis NSs and 4 dfs were used for temperature and lag dimensions, respectively, and a maximum lag of 21 days was used for the distributed lag nonlinear model (DLNM) [27]. Dual pollutant models were constructed to verify the robustness of the main model after adjusting for confounders by adding another air pollutant (O3, PM2.5, PM10, CO, SO2) to the main model.

    All statistical analyses were performed using the mgcv and dlnm packages of R software (version 3.4.1) by major models [28]. The change in mortality was calculated for every 10 μg/m3 increase in NO2 grade. All statistical results are presented as 2-sided values, while a P value <.05 was considered to indicate statistical significance.

    Ethics Approval

    Ethics approval was not required for secondary analysis of the anonymous data in this study.


    The average annual temperature, humidity, daily NO2 concentrations, and number of cause-specific deaths are summarized in Table 1. The Spearman correlation between air pollution and weather conditions is shown in Figure S1 in Multimedia Appendix 1. The correlation coefficients between NO2 and other constituents ranged from 0.1 to 0.94. Temperature and wind speed were negatively correlated with NO2, while NO2 was negatively correlated with other meteorological factors and air pollutants.

    Table 1. Summary statistics of environmental monitoring and daily death counts during 2010 to 2018 in Guangzhou, China.
    VariableMean (SD)
    Air pollutant concentrations

    		NO2 (μg/m3)47 (19)

    		Particulate matter (PM)2.5 (μg/m3)38 (22)

    		SO2 (μg/m3)18 (12)

    		O3 (μg/m3)82 (47)

    		CO (mg/m3)0.9 (0.2)

    		Air quality index72 (31)
    Weather conditions

    		Temperature (°C)22 (6)

    		Relative humidity (%)79 (11)

    		Air pressure (hPa)1007 (56)

    		Wind speed (km/h)22 (11)
    Daily numbers of deaths

    		All causes126 (27)

    		Nonaccidental causes119 (26)

    		Circulatory disease49 (14)

    		Respiratory disease19 (7)

    		Digestive disease4 (2)

    		Nervous disease1 (1)

    		Genitourinary disease2 (1)

    		External causes7 (3)

    		Endocrine disease5 (3)

    		Neoplasms35 (8)

    The average annual NO2 level was 47 μg/m3. During the study period of 2010 to 2018, 413,703 all-cause deaths were recorded in Guangzhou, with 126 deaths occurring daily, on average. The majority (391,543/413,703, 94.6%) of the deaths were due to nonaccidental causes. Other causes included cardiovascular diseases (160,226/413,703, 38.7%), respiratory diseases (63,290/413,703, 15.3%), digestive diseases (12,787/413,703, 3.1%), nervous system diseases (3659/413,703, 0.9%), genitourinary diseases (5314/413,703, 1.3%), external causes (22,147/413,703, 5.4%), endocrine diseases (14,972/413,703, 3.6%), and neoplasms (115,703/413,703, 28%; Table S1 in Multimedia Appendix 1).

    The association between NO2 exposure and mortality due to specific causes over 6 days (lags 0 to 6) and stratified by gender and age are shown in Table 2 and Figure S2 in Multimedia Appendix 1. For all-cause mortality, the highest risk of death due to NO2 exposure was observed at the lag of 2 or 3 days and was usually limited to 4 days. Effect estimates of NO2 were stronger for women and older subgroups for different diseases before the lag of 4 days, and the effects stabilized thereafter. The mortality risk increased linearly with NO2 concentrations at a lag of 0 to 4 days (Figure 2). Finally, we examined the stability of the model by using lag 0 days for NO2 and observed a similar linear relationship between NO2 and mortality (Figure S4 in Multimedia Appendix 1). The results of the generalized Poisson regression are shown in Table S2 in Multimedia Appendix 1. The goodness of fit index R2 of the model was 0.69, which meant that the model was well constructed.

    Table 2. The relative risk of death resulting from the different causes associated with an increase of 10 μg/m3 in NO2 levels with a 1-day delay (lag days of 0, 1, 2, 3, 4, 5, and 6) for each gender and age group.
    Causes of deathAll, % change (95% CI)Gender, % change (95% CI)Age (years), % change (95% CI)


    		MaleFemale0-64≥65
    All causes

    		Lag 01.009 (1.006-1.011)1.009 (1.005-1.012)1.009 (1.005-1.012)1.007 (1.004-1.010)1.010 (1.006-1.013)

    		Lag 11.011 (1.008-1.014)1.010 (1.007-1.013)1.013 (1.010-1.017)1.007 (1.004-1.010)1.014 (1.011-1.017)

    		Lag 21.010 (1.007-1.013)1.009 (1.006-1.012)1.012 (1.009-1.016)1.006 (1.002-1.009)1.014 (1.010-1.017)

    		Lag 31.007 (1.005-1.010)1.007 (1.003-1.010)1.008 (1.004-1.011)1.005 (1.001-1.008)1.009 (1.006-1.012)

    		Lag 41.003 (1.001-1.007)1.004 (1.001-1.007)1.004 (1.000-1.007)1.003 (1.000-1.006)1.005 (1.001-1.008)

    		Lag 51.001 (0.998-1.003)1.001 (0.998-1.005)0.999 (0.996-1.003)1.000 (0.996-1.003)1.001 (0.998-1.004)

    		Lag 60.997 (0.995-1.999)0.997 (0.994-1.001)0.997 (0.993-1.000)0.997 (0.994-1.000)0.997 (0.994-1.001)
    Nonaccidental causes

    		Lag 01.008 (1.006-1.011)1.009 (1.005-1.012)1.008 (1.004-1.012)1.007 (1.003-1.010)1.009 (1.006-1.013)

    		Lag 11.011 (1.009-1.014)1.010 (1.007-1.013)1.013 (1.010-1.017)1.007 (1.003-1.010)1.014 (1.011-1.017)

    		Lag 21.010 (1.008-1.013)1.008 (1.005-1.012)1.013 (1.010-1.017)1.005 (1.002-1.009)1.014 (1.011-1.017)

    		Lag 31.007 (1.005-1.010)1.007 (1.003-1.010)1.008 (1.005-1.012)1.005 (1.001-1.008)1.009 (1.006-1.012)

    		Lag 41.004 (1.002-1.007)1.004 (1.001-1.007)1.004 (1.000-1.008)1.003 (1.000-1.007)1.005 (1.001-1.008)

    		Lag 51.001 (0.998-1.003)1.001 (0.998-1.005)1.000 (0.996-1.003)1.000 (0.996-1.004)1.001 (0.998-1.005)

    		Lag 60.997 (0.995-1.000)0.997 (0.994-1.001)0.997 (0.993-1.000)0.997 (0.994-1.000)0.997 (0.994-1.000)
    Circulatory diseases

    		Lag 01.009 (1.005-1.012)1.008 (1.004-1.013)1.009 (1.004-1.014)1.008 (1.002-1.014)1.009 (1.004-1.013)

    		Lag 11.013 (1.009-1.016)1.011 (1.004-1.013)1.014 (1.004-1.014)1.006 (1.000-1.012)1.016 (1.004-1.013)

    		Lag 21.012 (1.008-1.016)1.009 (1.004-1.014)1.015 (1.004-1.014)1.005 (1.000-1.012)1.015 (1.011-1.020)

    		Lag 31.009 (1.006-1.013)1.010 (1.005-1.014)1.009 (1.004-1.014)1.004 (0.998-1.010)1.012 (1.007-1.016)

    		Lag 41.007 (1.003-1.011)1.009 (1.004-1.014)1.005 (1.000-1.010)1.004 (0.998-1.010)1.008 (1.004-1.013)

    		Lag 51.003 (0.999-1.007)1.004 (1.004-1.014)1.001 (0.996-1.006)1.002 (0.996-1.008)1.003 (0.999-1.008)

    		Lag 60.997 (0.993-1.000)0.997 (0.992-1.001)0.997 (0.996-1.006)0.996 (0.991-1.002)0.997 (0.993-1.001)
    Respiratory diseases

    		Lag 01.010 (1.004-1.015)1.012 (1.005-1.020)1.006 (0.997-1.014)1.017 (1.006-1.030)1.008 (1.002-1.014)

    		Lag 11.015 (1.010-1.021)1.015 (1.008-1.022)1.017 (1.008-1.025)1.013 (1.001-1.025)1.016 (1.010-1.022)

    		Lag 21.017 (1.011-1.022)1.015 (1.008-1.022)1.019 (1.011-1.028)1.010 (0.998-1.022)1.018 (1.012-1.024)

    		Lag 31.008 (1.003-1.014)1.010 (1.003-1.017)1.005 (0.997-1.014)1.003 (0.991-1.015)1.009 (1.003-1.016)

    		Lag 41.005 (0.999-1.011)1.008 (1.001-1.015)1.001 (0.992-1.009)1.010 (0.998-1.022)1.004 (0.998-1.010)

    		Lag 51.001 (0.996-1.007)1.001 (0.994-1.008)1.002 (0.993-1.010)1.010 (0.997-1.022)0.999 (0.993-1.005)

    		Lag 60.997 (0.992-1.003)0.996 (0.989-1.003)1.000 (0.991-1.008)1.001 (0.989-1.013)0.996 (0.990-1.003)
    Digestive diseases

    		Lag 01.004 (0.992-1.015)0.997 (0.982-1.012)1.014 (0.996-1.033)0.996 (0.979-1.014)1.010 (0.995-1.025)

    		Lag 11.003 (0.992-1.015)0.997 (0.983-1.012)1.010 (0.992-1.029)0.996 (0.979-1.013)1.008 (0.993-1.023)

    		Lag 20.999 (0.988-1.010)0.996 (0.981-1.010)1.003 (0.985-1.022)0.993 (0.976-1.011)1.003 (0.988-1.018)

    		Lag 30.998 (0.987-1.009)0.997 (0.982-1.011)1.000 (0.981-1.019)1.001 (0.984-1.019)0.996 (0.981-1.011)

    		Lag 40.997 (0.985-1.008)0.990 (0.975-1.005)1.008 (0.989-1.027)0.999 (0.982-1.017)0.995 (0.980-1.011)

    		Lag 50.991 (0.980-1.003)0.988 (0.973-1.003)0.997 (0.978-1.016)0.991 (0.974-1.009)0.992 (0.976-1.008)

    		Lag 60.995 (0.980-1.003)0.995 (0.981-1.010)0.995 (0.976-1.013)0.997 (0.980-1.014)0.994 (0.979-1.009)
    Nervous diseases

    		Lag 00.995 (0.973-1.017)0.986 (0.957-1.016)1.005 (0.972-1.038)0.998 (0.969-1.028)0.990 (0.958-1.023)

    		Lag 10.992 (0.971-1.014)0.982 (0.953-1.012)1.004 (0.973-1.037)0.999 (0.971-1.028)0.983 (0.952-1.016)

    		Lag 20.999 (0.978-1.021)0.995 (0.966-1.025)1.004 (0.973-1.037)0.995 (0.967-1.024)1.004 (0.972-1.037)

    		Lag 31.010 (0.988-1.032)1.015 (0.985-1.045)1.005 (0.973-1.038)1.006 (0.978-1.036)1.014 (0.981-1.047)

    		Lag 41.009 (0.987-1.032)1.025 (0.995-1.057)0.991 (0.958-1.025)1.027 (0.997-1.057)0.985 (0.952-1.020)

    		Lag 50.985 (0.963-1.007)0.995 (0.965-1.026)0.973 (0.940-1.006)1.010 (0.981-1.039)0.951 (0.918-0.984)

    		Lag 60.980 (0.959-1.002)0.977 (0.949-1.007)0.983 (0.951-1.016)1.004 (0.976-1.033)0.947 (0.916-0.980)
    Genitourinary diseases

    		Lag 01.009 (0.991-1.028)1.008 (0.984-1.032)1.012 (0.985-1.040)1.011 (0.984-1.038)1.008 (0.984-1.033)

    		Lag 11.007 (0.989-1.025)1.010 (0.986-1.034)1.004 (0.977-1.032)1.005 (0.979-1.032)1.008 (0.984-1.033)

    		Lag 21.004 (0.986-1.022)1.018 (0.995-1.043)0.986 (0.959-1.013)1.006 (0.979-1.033)1.002 (0.977-1.027)

    		Lag 30.998 (0.979-1.016)1.005 (0.981-1.030)0.989 (0.961-1.017)0.995 (0.968-1.022)0.999 (0.975-1.025)

    		Lag 40.994 (0.976-1.013)0.986 (0.961-1.011)1.005 (0.977-1.034)0.976 (0.949-1.004)1.009 (0.983-1.035)

    		Lag 50.991 (0.973-1.010)0.986 (0.962-1.011)0.998 (0.970-1.027)0.978 (0.951-1.006)1.003 (0.977-1.029)

    		Lag 60.988 (0.970-1.007)0.995 (0.971-1.019)0.980 (0.953-1.008)0.989 (0.962-1.016)0.987 (0.963-1.013)
    External causes

    		Lag 01.013 (1.004-1.022)1.008 (0.996-1.020)1.020 (1.006-1.034)1.012 (1.000-1.024)1.014 (1.000-1.029)

    		Lag 11.010 (1.001-1.020)1.008 (0.996-1.019)1.014 (1.000-1.029)1.009 (0.998-1.021)1.012 (0.998-1.027)

    		Lag 21.006 (0.996-1.015)1.010 (0.998-1.022)0.999 (0.985-1.013)1.009 (0.997-1.021)1.001 (0.986-1.015)

    		Lag 31.002 (0.993-1.012)1.005 (0.993-1.017)0.998 (0.984-1.012)1.004 (0.992-1.016)1.000 (0.985-1.014)

    		Lag 41.001 (0.991-1.010)1.001 (0.989-1.014)1.000 (0.985-1.014)1.000 (0.988-1.012)1.003 (0.988-1.018)

    		Lag 50.998 (0.988-1.007)1.002 (0.990-1.014)0.991 (0.977-1.006)0.999 (0.987-1.011)0.996 (0.981-1.011)

    		Lag 60.998 (0.989-1.007)0.999 (0.987-1.011)0.996 (0.982-1.010)0.996 (0.984-1.007)1.001 (0.986-1.015)
    Endocrine diseases

    		Lag 01.011 (1.000-1.022)1.006 (0.991-1.022)1.015 (1.000-1.030)1.003 (0.986-1.019)1.017 (1.003-1.031)

    		Lag 11.020 (1.009-1.030)1.017 (1.002-1.032)1.022 (1.008-1.037)1.010 (0.994-1.026)1.027 (1.013-1.041)

    		Lag 21.020 (1.009-1.031)1.019 (1.003-1.034)1.021 (1.007-1.036)1.012 (0.996-1.029)1.026 (1.011-1.040)

    		Lag 31.007 (0.997-1.018)1.003 (0.987-1.019)1.011 (0.996-1.026)1.003 (0.987-1.020)1.010 (0.996-1.025)

    		Lag 40.998 (0.987-1.010)0.996 (0.980-1.012)1.001 (0.986-1.016)0.995 (0.978-1.012)1.001 (0.986-1.016)

    		Lag 50.997 (0.986-1.008)0.996 (0.980-1.013)0.997 (0.982-1.012)0.992 (0.975-1.009)1.000 (0.986-1.015)

    		Lag 60.988 (0.977-0.999)0.981 (0.966-0.997)0.994 (0.979-1.008)0.992 (0.976-1.009)0.985 (0.971-1.000)
    Neoplasms

    		Lag 01.006 (1.002-1.010)1.008 (1.002-1.013)1.003 (0.996-1.010)1.003 (0.998-1.008)1.010 (1.004-1.017)

    		Lag 11.007 (1.003-1.011)1.007 (1.002-1.012)1.007 (1.000-1.013)1.007 (1.002-1.012)1.008 (1.001-1.015)

    		Lag 21.007 (1.003-1.011)1.007 (1.002-1.012)1.007 (1.000-1.013)1.006 (1.001-1.011)1.008 (1.001-1.015)

    		Lag 31.006 (1.002-1.010)1.003 (0.998-1.008)1.011 (1.004-1.018)1.006 (1.001-1.011)1.006 (0.999-1.013)

    		Lag 41.002 (0.998-1.006)0.999 (0.994-1.005)1.006 (1.000-1.013)1.003 (0.998-1.008)1.001 (0.994-1.008)

    		Lag 51.000 (0.996-1.005)1.001 (0.996-1.006)0.999 (0.992-1.006)1.000 (0.995-1.005)1.001 (0.994-1.009)

    		Lag 61.000 (0.996-1.004)1.001 (0.996-1.007)0.999 (0.992-1.005)0.999 (0.994-1.004)1.004 (0.997-1.010)
    ‎
    Figure 2. The dose-effect of the curves of NO2 and mortality as a result of (A) all causes, (B) nonaccidental causes, (C) cardiovascular diseases, (D) respiratory diseases, (E) digestive diseases, (F) nervous system diseases, (G) genitourinary diseases, (H) external causes, (I) endocrine diseases, and (J) neoplasms at 7 degrees of freedom with a lag of 0 to 4 days, stratified by gender and age (years).

    Among 10 categories, for every 10 μg/m3 increment in NO2 at a lag of 0 to 4 days, the risks of mortality increased due to all causes by 1.73% (95% CI 1.36%-2.09%) and due to nonaccidental causes, cardiovascular diseases, respiratory diseases, endocrine diseases, 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. On the other hand, NO2 did not significantly impact deaths due to digestive diseases, genitourinary diseases, nervous diseases, and external causes (Table 3 and Figure S3 in Multimedia Appendix 1).

    Table 3. Percentage increase in mortality as a result of different diseases per 10 μg/m3 increase in NO2 with a lag of 0 to 4 days in Guangzhou, China, stratified by gender and age groups.
    Causes of deathAll, % change (95% CI)Gender, % change (95% CI)Age (years), % change (95% CI)


    		MaleFemale0-64≥65
    All causes1.73 (1.36 to 2.09)1.56 (1.11 to 2.00)1.95 (1.45 to 2.46)1.15 (0.66 to 1.63)2.15 (1.68 to 2.62)
    Nonaccidental causes1.75 (1.38 to 2.12)1.57 (1.12 to 2.03)1.98 (1.47 to 2.50)1.13 (0.62 to 1.63)2.18 (1.71 to 2.66)
    Cardiovascular diseases2.06 (1.54 to 2.59)1.97 (1.29 to 2.65)2.17 (1.46 to 2.90)1.13 (0.29 to 1.98)2.50 (1.87 to 3.12)
    Chronic rheumatic heart diseases2.32 (1.56 to 3.08)2.07 (1.05 to 3.10)2.62 (1.55 to 3.71)1.62 (0.37 to 2.88)2.68 (1.76 to 3.60)
    Hypertensive diseases3.39 (1.80 to 4.99)2.79 (0.50 to 5.13)3.90 (1.75 to 6.10)1.88 (–1.15 to 5.01)3.88 (2.08 to 5.71)
    Ischemic heart diseases1.46 (0.70 to 2.23)1.61 (0.56 to 2.66)1.32 (0.26 to 2.40)0.59 (–0.74 to 1.93)1.82 (0.91 to 2.73)
    Acute ischemic heart disease1.27 (0.20 to 2.34)1.71 (0.30 to 3.13)0.71 (–0.85 to 2.30)0.84 (–0.84 to 2.56)1.51 (0.18 to 2.87)
    Acute myocardial infarction1.23 (0.13 to 2.34)1.69 (0.24 to 3.16)0.65 (–0.98 to 2.30)1.05 (–0.70 to 2.82)1.33 (–0.05 to 2.73)
    Myocardial infarction1.23 (0.12 to 2.34)1.67 (0.22 to 3.14)0.66 (–0.97 to 2.31)1.04 (–0.70 to 2.82)1.33 (–0.05 to 2.73)
    Chronic ischemic heart disease1.65 (0.60 to 2.72)1.54 (0.02 to 3.07)1.77 (0.33 to 3.23)0.21 (–1.99 to 2.45)2.05 (0.85 to 3.26)
    Other forms of heart disease–0.61 (–3.09 to 1.93)–0.66 (–3.97 to 2.76)–0.59 (–4.34 to 3.30)–1.28 (–4.81 to 2.39)0.04 (–3.41 to 3.62)
    Cerebrovascular diseases2.48 (1.70 to 3.27)2.21 (1.16 to 3.27)2.80 (1.69 to 3.91)1.64 (0.35 to 2.94)2.89 (1.95 to 3.84)
    Stroke2.48 (1.55 to 3.42)2.30 (1.03 to 3.58)2.71 (1.35 to 4.08)1.47 (–0.01 to 2.97)3.07 (1.90 to 4.25)
    Intracerebral hemorrhagic stroke1.75 (0.21 to 3.30)2.18 (0.21 to 4.18)1.19 (–1.18 to 3.61)1.23 (–0.90 to 3.40)2.18 (0.06 to 4.35)
    Ischemic stroke2.84 (1.35 to 4.35)2.35 (0.30 to 4.43)3.35 (1.24 to 5.50)0.95 (–1.72 to 3.69)3.60 (1.85 to 5.39)
    Arteries, arterioles, and capillaries0.27 (–3.75 to 4.45)0.44 (–4.40 to 5.53)–0.29 (–7.17 to 7.09)2.04 (–3.73 to 8.16)–1.24 (–6.66 to 4.50)
    Respiratory diseases2.32 (1.51 to 3.13)2.53 (1.50 to 3.57)2.03 (0.82 to 3.27)2.15 (0.41 to 3.92)2.36 (1.46 to 3.27)
    Influenza and pneumonia2.47 (1.35 to 3.60)2.72 (1.20 to 4.27)2.19 (0.62 to 3.80)2.50 (0.06 to 5.00)2.45 (1.21 to 3.71)
    Chronic lower respiratory disease2.33 (1.16 to 3.52)2.57 (1.17 to 3.99)1.90 (–0.04 to 3.87)2.38 (–0.14 to 4.96)2.32 (1.02 to 3.63)
    Chronic obstructive pulmonary disease2.33 (1.12 to 3.56)2.47 (1.03 to 3.93)2.10 (0.08 to 4.16)2.71 (0.01 to 5.48)2.25 (0.93 to 3.58)
    Other respiratory disease1.52 (–0.69 to 3.77)1.31 (–1.81 to 4.53)1.68 (–1.41 to 4.87)3.19 (–1.64 to 8.27)1.09 (–1.34 to 3.59)
    Digestive diseases–0.09 (–1.71 to 1.55)–1.18 (–3.26 to 0.94)1.61 (–1.07 to 4.35)–0.71 (–3.19 to 1.83)0.42 (–1.73 to 2.63)
    Esophagus, stomach, and duodenum disease0.34 (–3.10 to 3.89)–0.95 (–5.55 to 3.89)1.90 (–3.14 to 7.21)1.25 (–5.01 to 7.93)–0.03 (–4.08 to 4.20)
    Liver disease0.28 (–2.63 to 3.27)–0.42 (–3.71 to 2.98)2.44 (–3.58 to 8.84)–1.34 (–4.64 to 2.08)5.14 (–0.77 to 11.39)
    Other digestive disease–0.07 (–3.72 to 3.71)–1.16 (–5.95 to 3.88)1.39 (–4.29 to 7.41)0.49 (–6.45 to 7.94)–0.34 (–4.55 to 4.06)
    Nervous diseases0.32 (–2.78 to 3.51)0.03 (–4.20 to 4.44)0.64 (–3.93 to 5.42)1.18 (–2.92 to 5.46)–0.95 (–5.54 to 3.86)
    Genitourinary diseases0.57 (–2.04 to 3.24)1.12 (–2.31 to 4.67)–0.07 (–3.97 to 3.98)–0.29 (–4.12 to 3.71)1.17 (–2.35 to 4.81)
    Urinary diseases0.57 (–2.04 to 3.24)1.12 (–2.31 to 4.67)–0.07 (–3.97 to 3.98)–0.29 (–4.12 to 3.71)1.17 (–2.35 to 4.81)
    Renal failure0.82 (–2.75 to 4.52)2.22 (–2.51 to 7.16)–0.87 (–6.17 to 4.74)–0.20 (–5.41 to 5.30)1.68 (–3.08 to 6.67)
    External causes1.32 (–0.01 to 2.67)1.36 (–0.37 to 3.11)1.26 (–0.76 to 3.33)1.38 (–0.33 to 3.12)1.27 (–0.80 to 3.38)
    Road traffic injury0.28 (–2.89 to 3.55)0.68 (–2.85 to 4.34)–1.10 (–7.86 to 6.17)–0.40 (–3.66 to 2.97)13.02 (–0.64 to 28.55)
    Intentional self-harm3.99 (0.22 to 7.90)2.08 (–2.59 to 6.97)7.09 (1.06 to 13.48)3.05 (–0.93 to 7.20)9.85 (–0.31 to 21.03)
    Endocrine diseases2.40 (0.84 to 3.98)1.76 (–0.48 to 4.06)2.95 (0.81 to 5.14)1.01 (–1.36 to 3.43)3.41 (1.35 to 5.51)
    Diabetes2.04 (0.12 to 3.99)1.15 (–1.54 to 3.91)2.97 (0.25 to 5.76)1.02 (–1.72 to 3.83)3.00 (0.31 to 5.77)
    Neoplasm1.18 (0.59 to 1.78)1.03 (0.28 to 1.77)1.45 (0.48 to 2.42)1.06 (0.34 to 1.79)1.41 (0.41 to 2.42)
    Pancreas disease0.22 (–3.12 to 3.69)2.35 (–2.30 to 7.24)–2.11 (–6.76 to 2.78)0.29 (–3.86 to 4.61)0.24 (–5.18 to 5.96)

    Among the 26 disease subcategories, after NO2 exposure, there were significant increases in the top 10 causes with the highest risk of mortality: intentional self-harm (3.99%, 95% CI 0.22%-7.90%), hypertensive diseases (3.39%, 95% CI 1.80%-4.99%), ischemic stroke (2.84%, 95% CI 1.35%-4.35%), cerebrovascular diseases (2.48%, 95% CI 1.70%-3.27%), stroke (2.48%, 95% CI 1.55%-3.42%), influenza and pneumonia (2.47%, 95% CI 1.35%-3.60%), chronic obstructive pulmonary disease (2.33%, 95% CI 1.12%-3.56%), chronic lower respiratory disease (2.33%, 95% CI 1.16%-3.52%), chronic rheumatic heart disease (2.32%, 95% CI 1.56%-3.08%), and diabetes (2.04%, 95% CI 0.12%-3.99%; Table 3 and Figure S3 in Multimedia Appendix 1).

    The results of the stratified analysis are summarized in Table 3. For all-cause mortality, the impact of NO2 was relatively higher among women than men, but the difference was not statistically significant (P=.32). Older adults were at a higher risk than the younger age groups (P=.02), and their death rates due to NO2 exposure increased by 2.15% (95% CI 1.68%-2.62%) and 1.15% (95% CI 0.66%-1.63%), respectively. Age was a particularly significant factor for deaths due to nonaccidental causes and cardiovascular diseases.

    Stratification analyses by meteorological factor levels were further conducted (Table 4). We observed relatively higher effect estimates of NO2 on mortality for low levels of temperature, relative humidity, and wind speed than high levels, although most of the differences between the subgroups were not statistically significant. For air pressure, relatively higher effect estimates were found for all causes, nonaccidental causes, cardiovascular diseases, respiratory diseases, and neoplasms at low levels of air pressure than at high levels of air pressure. However, the opposite was true for endocrine diseases.

    Table 4. Stratified analysis of the percent increase in mortality as a result of different diseases per 10 μg/m3 increase in NO2 at a lag of 0 to 4 days during 2010 to 2018 in Guangzhou, China.
    CausesTotal, % change (95% CI)Temperature, % change (95% CI)Humidity, % change (95% CI)Wind speed, % change (95% CI)Air pressure, % change (95% CI)


    		LowHighP
    value    		LowHighP
    value    		LowHighP
    value    		LowHighP
    value
    All causes1.73 (1.36 to 2.09)2.14 (1.71 to 2.57)1.42 (0.53 to 2.33).161.95 (1.45 to 2.45)1.41 (0.85 to 1.97).161.70 (1.20 to 2.21)1.65 (1.06 to 2.24).891.78 (0.91 to 2.65)2.19 (1.75 to 2.63).41
    Nonaccidental causes1.75 (1.38 to 2.12)2.21 (1.77 to 2.65)1.28 (0.37 to 2.20).072.01 (1.50 to 2.53)1.37 (0.80 to 1.94).101.71 (1.19 to 2.23)1.70 (1.10 to 2.31).991.74 (0.84 to 2.64)2.24 (1.80 to 2.69).32
    Cardiovascular diseases2.06 (1.54 to 2.59)2.63 (2.01 to 3.24)1.79 (0.45 to 3.15).272.44 (1.73 to 3.16)1.31 (0.51 to 2.12).042.14 (1.40 to 2.88)1.77 (0.95 to 2.60).512.10 (0.82 to 3.41)2.86 (2.23 to 3.49).31
    Respiratory diseases2.32 (1.51 to 3.13)2.88 (1.90 to 3.86)2.68 (0.58 to 4.83).872.58 (1.46 to 3.71)1.93 (0.69 to 3.18).452.53 (1.42 to 3.64)2.15 (0.82 to 3.51).672.45 (0.52 to 4.41)2.97 (1.96 to 4.00).64
    Endocrine diseases2.40 (0.84 to 3.98)3.04 (1.15 to 4.97)0.88 (–3.15 to 5.07).352.63 (0.43 to 4.87)1.82 (–0.50 to 4.19).621.94 (–0.20 to 4.13)2.99 (0.46 to 5.59).544.57 (0.68 to 8.62)2.59 (0.68 to 4.54).38
    Neoplasms1.18 (0.59 to 1.78)1.38 (0.66 to 2.09)0.27 (–1.16 to 1.73).181.38 (0.56 to 2.21)1.05 (0.13 to 1.97).601.19 (0.38 to 2.02)0.88 (–0.08 to 1.85).630.90 (–0.46 to 2.28)1.10 (0.36 to 1.83).81

    The sensitivity analysis showed that each of the different modeling strategies provided roughly similar results. First, the calendar date and weather variables were varied using 3 to 7 dfs and 3 to 6 dfs, respectively, year-on-year. Second, we used DLNM to adjust for temperature and found that the estimated effect of NO2 was slightly attenuated but still statistically significant. Third, the main findings were stable with additional adjustment for SO2, PM2.5, O3, PM10, and CO in the 2-pollutant model. As shown in Figure S5 in Multimedia Appendix 1, the model demonstrated stability and reliability after using alternative dfs to control for the meteorological confounders.


    Principal Findings

    To the best of our knowledge, this is one of few studies to investigate the relationship between NO2 exposure and deaths due to multiple causes. An increment of 10 μg/m3 in NO2 increased the risk of all-cause mortality by 1.73% and had a similar impact on deaths due to nonaccidental causes (1.75% increase), cardiovascular diseases (2.06% increase), respiratory diseases (2.32% increase), endocrine diseases (2.40% increase), and neoplasms (1.18% increase). Following NO2 exposure, older adults were at a higher risk of death due to all causes, nonaccidental causes, and cardiovascular diseases. Furthermore, the impact of NO2 was higher in warm temperatures than in cold temperatures.

    Multiple studies have found that low wind speeds lead to high atmospheric stability and high levels of air pollution, which in turn have led to higher numbers of COVID-19–related infections and deaths [29,30]. Atmospheric stability based on low wind speeds reduces the diffusion of gases and air pollution particles [29]. The results of our analyses were similar. However, several studies found that wind speed and COVID-19 are positively correlated. An inverted U-shaped dose-response curve was found for wind speed and COVID-19 [30]. A Japanese study involving 74 participants reported that low humidity and high air pressure contribute to brain hemorrhage [31]. In contrast, a hospital-based study in Mexico showed no significant relationship between barometric pressure and stroke [32]. However, the influence of humidity and air pressure may be not as strong as other weather conditions such as temperature. Therefore, we included humidity, wind speed, and air pressure as confounding factors in the model to avoid potential bias.

    Previous studies on the association between NO2 and mortality mainly focused on overall causes, cardiovascular diseases, and respiratory diseases [13,15,16]. Our study also showed a statistically significant association between NO2 and endocrine system diseases. We found that the impact of NO2 exposure on mortality rates associated with endocrine, respiratory, and cardiovascular diseases was higher than that reported in a previous meta-analysis [33]. Studies previously conducted in China also reported that, for every 10 μg/m3 increase in NO2 concentration, mortality rates related to all causes, endocrine diseases, cardiovascular diseases, and respiratory diseases increased by 0.96%, 1.13%, 1.01%, and 1.22%, respectively [34,35]. In western countries, available data indicate that NO2 exposure leads to an increase in total mortality as well as mortality associated with endocrine diseases, cardiovascular diseases, and respiratory diseases, by 0.33%, 0.38%, 0.40%, and 0.38%, respectively [13,36]. These differences could be attributed to racial factors and the varying analytical techniques among the different studies.

    NO2 can affect disease-related mortality through various mechanisms. For instance, altered immune responses and inflammatory reactions following NO2 exposure may lead to endocrinological disorders. Inhaled NO2 triggers production of proinflammatory cytokines by alveolar macrophages, resulting in local oxidative stress. Furthermore, the cytokines enter systemic circulation and affect distant tissues, contributing to autoimmune responses and metabolic dysfunction [37,38]. NO2 is also a respiratory irritant that generates highly reactive, free radicals [39], which can cause severe lung injury, and death can occur depending on the dose and duration of exposure [40]. In addition, for individuals exposed to NO2, levels of inflammatory markers such as IL-12 and C-reactive protein in the blood are elevated, which may destabilize atherosclerotic plaques and result in their rupture, leading to increased blood pressure and enhanced thrombosis formation [41].

    Gender and age affect the level of sensitivity to air pollution [42,43]. The impact of the aforementioned factors on the effects generated by NO2 were largely inconclusive in previous studies. We found that gender was not a determinant of mortality risk due to NO2 exposure, which is consistent with previous studies [15,44-46]. However, one study showed that men are more susceptible to the detrimental effects of NO2 [14], which can be attributed to occupational and physiological differences. Furthermore, several studies have shown that older adults are predisposed to NO2-related mortality risk [13,15,23,47]. Consistent with this, the older adults in our cohort showed a higher risk of death due to all causes, nonaccidental causes, and cardiovascular diseases following NO2 exposure. A possible explanation is that older people have higher rates of chronic diseases and weaker immune systems, which may exacerbate the pathological effects of environmental NO2 pollution [48]. However, one study showed a link between NO2 exposure and the first episode of atrial fibrillation in young adults but not older adults [49]. NO2 is a significant risk factor for sudden death and melancholia among younger individuals, which may explain these results. Considering the rapidly aging population and changes in disease patterns, public policymaking, risk assessment, and air pollution standards should be modified to reduce the impact of NO2 on public health in China.

    The highest risk of death due to NO2 exposure was observed at a lag of 1 or 2 days and persisted for 4 days. However, some studies have observed the strongest link between NO2 levels and the number of hospital admissions on day 0 [46]. Similar trends in the effects of a lag of NO2 exposure have been reported for the mortality due to all causes, cardiovascular diseases, and respiratory diseases [13]. Moreover, the shape of the lag effects can be described by 2 different patterns. The first is that the NO2 levels recorded on the previous 2 days have a greater effect on mortality due to all causes, nonaccidental causes, circulatory diseases, respiratory diseases, endocrine diseases, and tumors. The second is that mortality resulting from digestive, nervous, and urogenital diseases as well as external systems tends to be more evenly distributed during the first 6 days. These inconsistencies may be a result of differences in the biological mechanisms that underlie the biological effects of NO2.

    We also found that the impact of NO2 exposure was higher on cold days, which is consistent with the findings of previous studies. For instance, studies conducted in Shanghai and Wuhan have shown that the effects of NO2 are highest in the winter [50,51]. Moreover, other research groups in China reported no substantial variation in the effects of NO2 on mortality due to nonaccidental causes and cardiovascular diseases across different seasons [23,52,53]. However, a large time series analysis conducted in the northeast United States reported a peak mortality risk from NO2 exposure in the summer [54]. Likewise, an Italian study conducted between April and September also found a stronger association between NO2 and mortality due to natural causes, cardiac diseases, and respiratory diseases [14]. A meta-analysis further demonstrated that NO2 had a greater impact on hospital admissions and mortality in hot weather (May 1 to September 30) [55]. One possible explanation is that people are exposed to higher levels of NO2 during the warmer periods of the year due to greater involvement in outdoor activities [14]. Furthermore, the production of NO2 also increases at higher temperatures [56]. Another reason is that heat may promote thrombosis by increasing blood viscosity and secondary cholesterol levels [57]. Finally, individual susceptibility to air pollutants may increase in the summer [58]. These inconsistent results may be due to the influence of concentrations of NO2 components, regional gaseous pollutants, climatic conditions, resident exposure patterns, socioeconomic characteristics, and different data analysis methods.

    This study has several salient points. First, the study showed that NO2 was related not only to all-cause mortality but also to cause-specific mortality. Second, the relationship between NO2 and mortality due to different causes and the potential modifying effects of gender, age, and season were analyzed. These findings may have important implications for developing targeted strategies to protect vulnerable populations from the harmful effect of NO2. The detrimental effects of NO2 can be minimized by reducing time spent outdoors and physical exertion, wearing a mask when going out, avoiding heavily trafficked roads, and increasing immunity [59].

    Nevertheless, there are several limitations that need to be considered. First, our study was limited to the Guangzhou area, and the results cannot be generalized to other regions due to differences in the population structure and environmental characteristics. Second, environmental monitoring data may only indicate population averages rather than individual levels, which may have introduced inevitable measurement bias [60]. Third, the classification of the cause of death was based only on ICD-10 codes and does not detail the actual cause of death of the patient, which may also lead to bias. Finally, causal relationships and the underlying pathophysiology of our findings could not be ascertained and warrant further research.

    Conclusions

    In summary, our study comprehensively examined the association between NO2 exposure and a spectrum of causes for the risk of mortality. Through subgroup analysis, we found that gender did not significantly modify the NO2-related mortality risks, while older adults were more susceptible to death due to all causes, nonaccidental causes, and cardiovascular diseases when exposed to NO2. Furthermore, the association between NO2 and cause-specific mortality is stronger in the warm season. We found a relatively higher effect of NO2 on health on days with higher temperatures, relative humidity, and wind speed but lower pressure levels, though the between-group differences were not statistically significant. Considering heterogeneity in socioeconomic characteristics and population structure among different regions, as well as the limitations of the inaccuracy of monitoring data and measurement error of ICD classifications, our findings still need to be confirmed in other regions. Our study provides new evidence to develop prevention-oriented health policies, highlighting that it is necessary to strengthen air quality standards to protect public health from NO2 pollution.

    Acknowledgments

    The study was supported by Guangzhou Science and Technology Program Project (No. 2023A03J0284) and Guangdong Province Medical Science and Technology Research Project (No. A2023042).

    Authors' Contributions

    JY conceptualized the study, acquired the funding, and supervised the study. GL and HD provided the resources for the study. GL, HD, and HR performed the investigation. JZ performed the formal analysis. JZ created the visualizations. JZ wrote the original manuscript draft. GL, HD, ML, HR, and JY reviewed and edited the manuscript draft.

    Conflicts of Interest

    None declared.

    Multimedia Appendix 1

    Supplementary tables and figures.

    DOCX File , 1482 KB
    1. Almetwally AA, Bin-Jumah M, Allam AA. Ambient air pollution and its influence on human health and welfare: an overview. Environ Sci Pollut Res Int. Jul 03, 2020;27(20):24815-24830. [CrossRef] [Medline]
    2. Rajarathnam U, Sehgal M, Nairy S, Patnayak RC, Chhabra SK, Ragavan KVS, et al. HEI Health Review Committee. Part 2. Time-series study on air pollution and mortality in Delhi. Res Rep Health Eff Inst. Mar 2011(157):47-74. [Medline]
    3. Liu Y, Pan J, Zhang H, Shi C, Li G, Peng Z, et al. Short-term exposure to ambient air pollution and asthma mortality. Am J Respir Crit Care Med. Jul 01, 2019;200(1):24-32. [CrossRef] [Medline]
    4. Stafoggia M, Bellander T. Short-term effects of air pollutants on daily mortality in the Stockholm county - A spatiotemporal analysis. Environ Res. Sep 2020;188:109854. [FREE Full text] [CrossRef] [Medline]
    5. Liu T, Zhou Y, Wei J, Chen Q, Xu R, Pan J, et al. Association between short-term exposure to ambient air pollution and dementia mortality in Chinese adults. Sci Total Environ. Nov 25, 2022;849:157860. [CrossRef] [Medline]
    6. Zhou P, Hu J, Yu C, Bao J, Luo S, Shi Z, et al. Short-term exposure to fine particulate matter constituents and mortality: case-crossover evidence from 32 counties in China. Sci China Life Sci. Dec 14, 2022;65(12):2527-2538. [CrossRef] [Medline]
    7. Huang Y, Zhu M, Ji M, Fan J, Xie J, Wei X, et al. Air pollution, genetic factors, and the risk of lung cancer: a prospective study in the UK Biobank. Am J Respir Crit Care Med. Oct 01, 2021;204(7):817-825. [CrossRef] [Medline]
    8. Air quality guidelines for Europe, 2nd edition. World Health Organization. 2000. URL: https://www.who.int/publications/i/item/9789289013581 [accessed 2024-01-21]
    9. Brauer M, Freedman G, Frostad J, van Donkelaar A, Martin RV, Dentener F, et al. Ambient air pollution exposure estimation for the global burden of disease 2013. Environ Sci Technol. Jan 05, 2016;50(1):79-88. [CrossRef] [Medline]
    10. Guangzhou Statistical Yearbook 2018. Gangzhou Municipal People's Governmen. URL: http://www.gz.gov.cn/zfjgzy/gzstjj/xxgk/tjxx/tjnj/content/post_2993872.html [accessed 2024-01-21]
    11. Ambient (outdoor) air pollution. World Health Organization. Dec 19, 2022. URL: https://www.who.int/en/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health/ [accessed 2024-01-21]
    12. Burnett RT, Stieb D, Brook JR, Cakmak S, Dales R, Raizenne M, et al. Associations between short-term changes in nitrogen dioxide and mortality in Canadian cities. Archives of Environmental Health: An International Journal. May 2004;59(5):228-236. [CrossRef] [Medline]
    13. Samoli E, Aga E, Touloumi G, Nisiotis K, Forsberg B, Lefranc A, et al. Short-term effects of nitrogen dioxide on mortality: an analysis within the APHEA project. Eur Respir J. Jun 01, 2006;27(6):1129-1138. [FREE Full text] [CrossRef] [Medline]
    14. Chiusolo M, Cadum E, Stafoggia M, Galassi C, Berti G, Faustini A, et al. EpiAir Collaborative Group. Short-term effects of nitrogen dioxide on mortality and susceptibility factors in 10 Italian cities: the EpiAir study. Environ Health Perspect. Sep 2011;119(9):1233-1238. [FREE Full text] [CrossRef] [Medline]
    15. Chen R, Yin P, Meng X, Wang L, Liu C, Niu Y, et al. Associations between ambient nitrogen dioxide and daily cause-specific mortality: evidence from 272 Chinese cities. Epidemiology. Jul 2018;29(4):482-489. [CrossRef] [Medline]
    16. Meng X, Liu C, Chen R, Sera F, Vicedo-Cabrera AM, Milojevic A, Garcia; et al. Short term associations of ambient nitrogen dioxide with daily total, cardiovascular, and respiratory mortality: multilocation analysis in 398 cities. BMJ. Mar 24, 2021;372:n534. [FREE Full text] [CrossRef] [Medline]
    17. Seaton A, Dennekamp M. Hypothesis: ill health associated with low concentrations of nitrogen dioxide--an effect of ultrafine particles? Thorax. Dec 01, 2003;58(12):1012-1015. [FREE Full text] [CrossRef] [Medline]
    18. Künzli N. The public health relevance of air pollution abatement. Eur Respir J. Jul 01, 2002;20(1):198-209. [FREE Full text] [CrossRef] [Medline]
    19. Guangzhou Center for Disease Control and Prevention. URL: http://www.gzcdc.org.cn/ [accessed 2024-01-21]
    20. China Meteorological Data Sharing Service System. URL: http://data.cma.cn/ [accessed 2024-01-21]
    21. Yang J, Ou C, Ding Y, Zhou Y, Chen P. Daily temperature and mortality: a study of distributed lag non-linear effect and effect modification in Guangzhou. Environ Health. Sep 14, 2012;11(1):1. [CrossRef] [Medline]
    22. Yang J, Zhou M, Zhang F, Yin P, Wang B, Guo Y, et al. Diabetes mortality burden attributable to short-term effect of PM in China. Environ Sci Pollut Res Int. May 23, 2020;27(15):18784-18792. [CrossRef] [Medline]
    23. Chen R, Samoli E, Wong C, Huang W, Wang Z, Chen B, et al. CAPES Collaborative Group. Associations between short-term exposure to nitrogen dioxide and mortality in 17 Chinese cities: the China Air Pollution and Health Effects Study (CAPES). Environ Int. Sep 15, 2012;45:32-38. [CrossRef] [Medline]
    24. Yang J, Zhou M, Li M, Yin P, Hu J, Zhang C, et al. Fine particulate matter constituents and cause-specific mortality in China: A nationwide modelling study. Environ Int. Oct 2020;143:105927. [FREE Full text] [CrossRef] [Medline]
    25. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ. Jan 25, 2003;326(7382):219. [FREE Full text] [CrossRef] [Medline]
    26. Yang J, Yin P, Sun J, Wang B, Zhou M, Li M, et al. Heatwave and mortality in 31 major Chinese cities: Definition, vulnerability and implications. Sci Total Environ. Feb 01, 2019;649:695-702. [CrossRef] [Medline]
    27. Gasparrini A, Armstrong B, Kenward MG. Distributed lag non-linear models. Stat Med. Sep 20, 2010;29(21):2224-2234. [FREE Full text] [CrossRef] [Medline]
    28. Gasparrini A. Distributed lag linear and non-linear models in R: the package dlnm. J. Stat. Soft. 2011;43(8):1-20. [CrossRef] [Medline]
    29. Coccia M. The effects of atmospheric stability with low wind speed and of air pollution on the accelerated transmission dynamics of COVID-19. International Journal of Environmental Studies. Sep 15, 2020;78(1):1-27. [CrossRef]
    30. Ali Q, Raza A, Saghir S, Khan MTI. Impact of wind speed and air pollution on COVID-19 transmission in Pakistan. Int J Environ Sci Technol (Tehran). Mar 13, 2021;18(5):1287-1298. [FREE Full text] [CrossRef] [Medline]
    31. Ono Y, Horibe H, Hayakawa N, Aoki N, Okada H. Biometeorologie studies on cerebrovascular diseases. (IV). Evaluation of meterorologic factors, their changes or combinations on the occurrence of cerebrovascular accident. Jpn Circ J. Feb 1970;34(2):69-75. [CrossRef] [Medline]
    32. Olivares L, Castańeda E, Grifé A, Alter M. Risk factors in stroke: a clinical study in Mexican patients. Stroke. Sep 1973;4(5):773-781. [CrossRef] [Medline]
    33. Li C, Fang D, Xu D, Wang B, Zhao S, Yan S, et al. Main air pollutants and diabetes-associated mortality: a systematic review and meta-analysis. Eur J Endocrinol. Nov 2014;171(5):R183-R190. [CrossRef] [Medline]
    34. Chen G, Song G, Jiang L, Zhang Y, Zhao N, Chen B, et al. Short-term effects of ambient gaseous pollutants and particulate matter on daily mortality in Shanghai, China. J Occup Health. Feb 16, 2008;50(1):41-47. [FREE Full text] [CrossRef] [Medline]
    35. Zhang Q, Liu C, Wang Y, Gong J, Wang G, Ge W, et al. Associations of long-term exposure to ambient nitrogen dioxide with indicators of diabetes and dyslipidemia in China: A nationwide analysis. Chemosphere. Apr 2021;269:128724. [FREE Full text] [CrossRef] [Medline]
    36. Darbre P. Overview of air pollution and endocrine disorders. IJGM. May 2018;Volume 11:191-207. [CrossRef] [Medline]
    37. Ritz SA. Air pollution as a potential contributor to the 'epidemic' of autoimmune disease. Med Hypotheses. Jan 2010;74(1):110-117. [CrossRef] [Medline]
    38. Mousavi SE, Delgado-Saborit JM, Adivi A, Pauwels S, Godderis L. Air pollution and endocrine disruptors induce human microbiome imbalances: A systematic review of recent evidence and possible biological mechanisms. Sci Total Environ. Apr 10, 2022;816:151654. [FREE Full text] [CrossRef] [Medline]
    39. Kelly FJ, Blomberg A, Frew A, Holgate ST, Sandstrom T. Antioxidant kinetics in lung lavage fluid following exposure of humans to nitrogen dioxide. Am J Respir Crit Care Med. Dec 1996;154(6 Pt 1):1700-1705. [CrossRef] [Medline]
    40. Chen T, Gokhale J, Shofer S, Kuschner WG. Outdoor air pollution: nitrogen dioxide, sulfur dioxide, and carbon monoxide health effects. Am J Med Sci. Apr 2007;333(4):249-256. [CrossRef] [Medline]
    41. Ruidavets J, Cassadou S, Cournot M, Bataille V, Meybeck M, Ferrières J. Increased resting heart rate with pollutants in a population based study. J Epidemiol Community Health. Aug 01, 2005;59(8):685-693. [FREE Full text] [CrossRef] [Medline]
    42. Felber Dietrich D, Gemperli A, Gaspoz J, Schindler C, Liu LS, Gold DR, et al. SAPALDIA Team. Differences in heart rate variability associated with long-term exposure to NO2. Environ Health Perspect. Oct 2008;116(10):1357-1361. [FREE Full text] [CrossRef] [Medline]
    43. Liu S, Krewski D, Shi Y, Chen Y, Burnett RT. Association between maternal exposure to ambient air pollutants during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol. Aug 31, 2007;17(5):426-432. [CrossRef] [Medline]
    44. Aga E, Samoli E, Touloumi G, Anderson H, Cadum E, Forsberg B, et al. Short-term effects of ambient particles on mortality in the elderly: results from 28 cities in the APHEA2 project. Eur Respir J Suppl. May 01, 2003;40(Supplement 40):28s-33s. [FREE Full text] [CrossRef] [Medline]
    45. Cakmak S, Dales RE, Judek S. Do gender, education, and income modify the effect of air pollution gases on cardiac disease? J Occup Environ Med. Jan 2006;48(1):89-94. [CrossRef] [Medline]
    46. Collart P, Dubourg D, Levêque A, Sierra NB, Coppieters Y. Short-term effects of nitrogen dioxide on hospital admissions for cardiovascular disease in Wallonia, Belgium. Int J Cardiol. Mar 15, 2018;255:231-236. [CrossRef] [Medline]
    47. Gouveia N, Fletcher T. Time series analysis of air pollution and mortality: effects by cause, age and socioeconomic status. J Epidemiol Community Health. Oct 2000;54(10):750-755. [FREE Full text] [CrossRef] [Medline]
    48. Simoni M, Baldacci S, Maio S, Cerrai S, Sarno G, Viegi G. Adverse effects of outdoor pollution in the elderly. J Thorac Dis. Jan 2015;7(1):34-45. [FREE Full text] [CrossRef] [Medline]
    49. Gerber Y. Can atrial fibrillation be caused or triggered by air pollution? An epidemiological perspective. Eur J Prev Cardiol. Jul 11, 2019;26(11):1205-1207. [CrossRef] [Medline]
    50. Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N, et al. Season, sex, age, and education as modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China: The Public Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect. Sep 2008;116(9):1183-1188. [FREE Full text] [CrossRef] [Medline]
    51. Qian Z, Lin H, Stewart WF, Kong L, Xu F, Zhou D, et al. Seasonal pattern of the acute mortality effects of air pollution. J Air Waste Manag Assoc. Apr 22, 2010;60(4):481-488. [CrossRef] [Medline]
    52. Li J, Woodward A, Hou X, Zhu T, Zhang J, Brown H, et al. Modification of the effects of air pollutants on mortality by temperature: A systematic review and meta-analysis. Sci Total Environ. Jan 01, 2017;575:1556-1570. [CrossRef] [Medline]
    53. Lin C, Liao C. Temperature-dependent association between mortality rate and carbon monoxide level in a subtropical city: Kaohsiung, Taiwan. Int J Environ Health Res. Jun 2009;19(3):163-174. [CrossRef] [Medline]
    54. Peng RD, Dominici F, Pastor-Barriuso R, Zeger SL, Samet JM. Seasonal analyses of air pollution and mortality in 100 US cities. Am J Epidemiol. Mar 15, 2005;161(6):585-594. [CrossRef] [Medline]
    55. Biggeri A, Baccini M, Bellini P, Terracini B. Meta-analysis of the Italian studies of short-term effects of air pollution (MISA), 1990-1999. Int J Occup Environ Health. Jul 19, 2005;11(1):107-122. [CrossRef] [Medline]
    56. Analitis A, Michelozzi P, D’Ippoliti D, de’Donato F, Menne B, Matthies F, et al. Effects of heat waves on mortality. Epidemiology. 2014;25(1):15-22. [CrossRef] [Medline]
    57. McGeehin MA, Mirabelli M. The potential impacts of climate variability and change on temperature-related morbidity and mortality in the United States. Environ Health Perspect. May 2001;109 Suppl 2(Suppl 2):185-189. [FREE Full text] [CrossRef] [Medline]
    58. Pekkanen J, Brunner EJ, Anderson HR, Tiittanen P, Atkinson RW. Daily concentrations of air pollution and plasma fibrinogen in London. Occup Environ Med. Dec 2000;57(12):818-822. [FREE Full text] [CrossRef] [Medline]
    59. Nambu Z, Yokoyama E. Protection with cycloheximide or emetine against pulmonary edema induced by ozone or nitrogen dioxide. Sangyo Igaku. Mar 1982;24(2):133-140. [CrossRef] [Medline]
    60. Zanobetti A, Schwartz J. The effect of fine and coarse particulate air pollution on mortality: a national analysis. Environ Health Perspect. Jun 2009;117(6):898-903. [FREE Full text] [CrossRef] [Medline]

    ‎
    DLNM: distributed lag nonlinear model
    ICD-10: 10th Revision of the International Classification of Diseases
    NS: natural cubic spline function
    PM: particulate matter

    Edited by A Mavragani, T Sanchez; submitted 29.11.22; peer-reviewed by Z Sun, C Ou, M Coccia, M Raimi; comments to author 28.07.23; revised version received 18.09.23; accepted 07.01.24; published 05.02.24.

    Copyright

    ©Jie Zeng, Guozhen Lin, Hang Dong, Mengmeng Li, Honglian Ruan, Jun Yang. Originally published in JMIR Public Health and Surveillance (https://publichealth.jmir.org), 05.02.2024.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Public Health and Surveillance, is properly cited. The complete bibliographic information, a link to the original publication on https://publichealth.jmir.org, as well as this copyright and license information must be included.