Skip to main content
Download PDF
Download ePub

Urban–rural disparities in the association of nitrogen dioxide exposure with cardiovascular disease risk in China: effect size and economic burden

International Journal for Equity in Health volume 23, Article number: 22 (2024) Cite this article

Abstract

Background

Together with rapid urbanization, ambient nitrogen dioxide (NO2) exposure has become a growing health threat. However, little is known about the urban–rural disparities in the health implications of short-term NO2 exposure. This study aimed to compare the association between short-term NO2 exposure and hospitalization for cardiovascular disease (CVD) among urban and rural residents in Shandong Province, China. Then, this study further explored the urban–rural disparities in the economic burden attributed to NO2 and the explanation for the disparities.

Methods

Daily hospitalization data were obtained from an electronic medical records dataset covering a population of 5 million. In total, 303,217 hospital admissions for CVD were analyzed. A three-stage time-series analytic approach was used to estimate the county-level association and the attributed economic burden.

Results

For every 10-μg/m3 increase in NO2 concentrations, this study observed a significant percentage increase in hospital admissions on the day of exposure of 1.42% (95% CI 0.92 to 1.92%) for CVD. The effect size was slightly higher in urban areas, while the urban–rural difference was not significant. However, a more pronounced displacement phenomenon was found in rural areas, and the economic burden attributed to NO2 was significantly higher in urban areas. At an annual average NO2 concentration of 10 μg/m3, total hospital days and expenses in urban areas were reduced by 81,801 (44,831 to 118,191) days and 60,121 (33,002 to 86,729) thousand CNY, respectively, almost twice as much as in rural areas. Due to disadvantages in socioeconomic status and medical resources, despite similar air pollution levels in the urban and rural areas of our sample sites, the rural population tended to spend less on hospitalization services.

Conclusions

Short-term exposure to ambient NO2 could lead to considerable health impacts in either urban or rural areas of Shandong Province, China. Moreover, urban–rural differences in socioeconomic status and medical resources contributed to the urban–rural disparities in the economic burden attributed to NO2 exposure. The health implications of NO2 exposure are a social problem in addition to an environmental problem. Thus, this study suggests a coordinated intervention system that targets environmental and social inequality factors simultaneously.

Introduction

Under the circumstances of rapid urbanization, ambient nitrogen dioxide (NO2) exposure has become a growing health threat in China [1]. The combined pace of economic growth and urbanization has led to tremendous increases in energy consumption. Consequently, NO2 is a major ambient pollutant [2]. Ambient NO2 exposure is considered to lead to cardiovascular disease (CVD) via dozens of molecular alterations, including but not limited to systemic inflammation, oxidative stress, endothelial dysfunction, coagulation, lipid metabolism, and vascular smooth muscle cell proliferation [3]. A series of epidemiological studies have linked short-term NO2 exposure to various CVD risks, such as hypertension, coronary heart disease, stroke, arrhythmia, and dyslipidemia [4,5,6,7,8]. Notably, the urban–rural disparities in CVD incidence and morbidity are prominent in China. The most recent national survey revealed that the prevalence rates of CVD among rural and urban residents were 139.3 ‰ and 168.0 ‰, respectively, while the mortality rates were 323.29 and 277.92 per 100,000 rural and urban residents, respectively, and the mortality rate from CVD in rural regions had consistently surpassed and remained higher than that in urban areas since 2009. Thus, the extent to which short-term NO2 exposure may lead to disparities should be determined.

Due to the difficulty in obtaining air pollution data and resident health data in rural areas, most studies on the health implications of NO2 have been carried out in urban areas. Only a few studies have compared the association of NO2 with CVD risk between urban and rural areas, and the available results are mixed. A study identified that the health risks from short-term NO2 exposure increased with the urbanization process based on data from the Pearl River Delta region [9]. A comparative study of urban and rural areas in Guangxi Province, China, also suggested that the effects of NO2 on CVD hospitalizations were not significant in rural areas, whereas urban residents were significantly and negatively affected by NO2 exposure [10]. In contrast, Li et al. found that rural residents were more sensitive to short-term NO2 exposure than urban residents in terms of CVD mortality, but the differences were not significant [11]. Thus, additional research is warranted to better understand the urban–rural disparities in this association.

Social inequality between urban and rural residents may further contribute to urban–rural disparities in the association of nitrogen dioxide exposure with CVD risk in terms of effect size and economic burden. Although rural–urban differences in access to basic health care have narrowed in China, access to high-quality health care services persist due to dramatic urban–rural differences in socioeconomic status. Rural residents usually have limited access to high-quality health care services, which may cause delayed and even reduced treatment of CVD attributed to short-term NO2 exposure. Compared to their urban counterparts, the rural population ranks lower in socioeconomic status [12], which further constrains rural residents in affording high-quality health care and leads to urban–rural disparities in the economic burden attributed to NO2 exposure. As Mohai et al. reported, environmental health issues are not only an environmental problem but also connected to social inequality [13]. More studies are also needed to corroborate the extent to which to what extent social inequality in socioeconomic status and access to health care may influence the urban–rural disparities in the effects of NO2 exposure.

Therefore, the objective of this study was to compare the association between short-term NO2 exposure and hospitalization for CVD among urban and rural residents in Shandong Province, China. Then, this study further explored the urban–rural disparities in the economic burden attributed to NO2 and the potential explanations from the perspective of social inequalities. To our knowledge, this is the first study to assess the urban–rural disparities in the economic burden attributed to NO2 exposure. A better understanding of the urban–rural disparities in short-term NO2-related health risks is of great importance for effective and timely decision-making in designing spatially targeted health interventions and developing NO2-resilient health systems. Furthermore, this study provides evidence to effectively allocate urban and rural medical resources from the perspectives of environmental justice and social equality.

Chinese urban–rural background

China has experienced rapid urbanization in the past five decades, which took approximately a hundred years for Western societies [14]. Along with the urbanization process, urban–rural health and social disparities are becoming dramatic [15]. These disparities are rooted in development patterns and policy systems [16]. In the process of urbanization beginning in the 1970s, a series of urban-biased policies were implemented in China to invest resources in urban areas, contributing to rapid development in these areas [17]. However, the urban control and exploitation of rural areas have led to the decline of rural areas and the formation of an urban–rural dual structure in terms of economic levels and access to welfare [15]. For example, the per capita disposable income of urban residents increased from CNY 343.4 in 1978 to CNY 6280.0 in 2000, while that of rural residents increased from CNY 133.6 to CNY 2253.4 during the same period [18]. Moreover, medical resources are mostly concentrated in urban areas. In 2000, the numbers of health technicians and beds per thousand people in rural areas were 2.41 and 1.50, respectively. The numbers of health technicians and beds in urban areas were 2.15 and 2.33 times those in rural areas, respectively [18]. However, air quality was much better in rural areas due to the slow industrialization process in this period [19].

Since 2000, urban–rural relations and the dependence of urban areas on rural areas have begun to change [14]. New types of urbanization and rural revitalization strategies have been put forward in succession, and urban and rural areas have exhibited a new trend of integrated development. Specifically, in rural areas, targeted poverty alleviation aims to lift all destitute households out of poverty and underdevelopment. Then, the Beautiful Countryside Plan can allow the improvement of infrastructure and social welfare in rural areas. Specifically, China has endeavored to achieve universal health coverage (UHC) and grant access for every citizen to equitable, accessible, and reliable health services and protection. To ensure that urban and rural residents enjoy equal access to basic medical care, China has established a medical security system covering 1.3 billion people, with a participation rate of over 95% [20]. Combined with the reform of public hospitals, a hierarchical medical system, the development of contracts with family doctors, and the basic drug system, China’s health care reform has achieved population coverage, service coverage, and cost coverage, which are the three dimensions of UHC realization [21].

However, due to industrialization and solid fuel emissions, air pollution is becoming a major problem in a growing number of rural areas. Moreover, the imbalance between rural and urban development continues to be prominent in many provinces [22]. Urban areas usually hold an advantage in socioeconomic status. According to the 2021 yearbook [23], the per capita disposable income of urban and rural residents was CNY 43833.8 and CNY 17131.5 in 2020, respectively. In addition to income inequality, urban–rural health inequality remains problematic. Despite similar access to basic medical care, high-quality health care resources are distributed mostly in urban areas [24, 25]. Compared with urban residents, the utilization of health services and annual health and hospitalization expenses are poorer for rural residents [12]. A higher mortality rate attributed to CVD has been documented in rural areas despite of a lower prevalence of CVD risk. Notably, spatial differences are dramatic in terms of urban–rural characteristics. For example, urban–rural differences in the NO2 annual average concentration exhibit a great gap in Chongqing and Guangxi but not in Shandong or Beijing [26]. A similar phenomenon was found regarding socioeconomic status and access to health care. Residents in rural areas of Guangdong and Hainan provinces are entitled to equal and even better health care than the average level. Thus, representative quantitative evidence is needed to reveal the associations between NO2 exposure and CVD risk in urban and rural areas.

Methods

Study population

This study derived data from the Cheeloo Lifespan Electronic Health Research Data Library (Cheeloo LEAD) using a three-stage cluster random sampling method. Cohorts from 39 counties were obtained, totaling 5 million individuals, sampled from 136 counties in Shandong Province, which has a total population of 101 million. The specific sampling process and the demographic characteristics of this population are provided in Supplementary Materials Fig. S1. A more detailed description of the study design and sampling procedure can be found at http://www.mhdata.sdu.edu.cn/cheeloolead.htm and in previously published studies [27,28,29]. Urban and rural areas are delineated based on the urban–rural categorization code established by the National Bureau of Statistics of China in 2015 (http://www.stats.gov.cn/sj/tjbz/qhdm/). This classification code comprises three numbers, where the initial digit being 1 indicates an urban area, and a first digit of 2 signifies a rural area. The 39 county-level units included 21 rural counties and 19 urban counties. The county names, county codes, and sample sizes of the sampled counties are shown in Table S1. Of them, Tengzhou city included an urban county and a rural county; thus, there were 39 county-level units.

CVD hospital admission identification

Electronic medical records and medical insurance data were extracted for the sampled residents, and individual identification numbers and admission times were used as indexes to merge the information from the two data set. The essential hospital records of the study population included the names and codes of the discharge diagnosis, the length of hospitalization, and the names and expenses of prescriptions during hospitalization. A total of 1.7 million hospitalizations were extracted. Among them, 336,621 hospitalizations for CVD were screened according to International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes. After excluding a portion of missing data points, the final sample size was 303,127. Subsamples were further screened according to the ICD-10 (coronary heart disease: I20-I25; ischemic stroke: I63; hypertension: I10-I12). For a specific disease in a specific county on a given day, the county-level hospital admissions were obtained by summing the total number of hospital admissions on that day from the sampled population of that county. When calculating attributable hospital days and the burden of hospitalization costs, the total was derived by multiplying the average expenses and average length of stay for all admissions in a specific county during the study period by the attributable number of admissions for that county.

Air pollution and meteorological data

Daily ambient NO2, particulate matter with an aerodynamic diameter of 2.5 μm or less (PM2.5), particulate matter with an aerodynamic diameter of 10 μm or less (PM10), sulfur dioxide (SO2), carbon monoxide (CO) and ozone (O3) data covering Shandong Province from 2015 to 2017 at a spatial resolution of 0.1° (≈10 km2) were collected from ChinaHighAirPollutants (CHAP, available at https://weijing-rs.github.io/product.html). These data are estimated by a space-time extremely randomized tree (STET) model. This model was developed to integrate satellite remote sensing products, atmospheric reanalysis, and ground-based measurements to complete model simulations. The pollutant estimations were reliable since they exhibited high R2 values of 0.80–0.91, with reference to surface observations obtained by adopting the independent 10-fold cross-validation approach. County-level data were extracted by averaging the grid values.

The daily mean temperature and relative humidity were based on the daily source data from a total of 131 meteorological monitoring stations in Shandong and adjacent provinces from the China Meteorological Data Sharing Service (http://data.cma.cn/). A thin-plate smooth spline function, with longitude and latitude as independent spline variables and elevation as a covariate considered in the function, was applied to interpolate the daily mean temperature and relative humidity grid at 0.01° * 0.01° resolution for the whole of Shandong Province from 2015 to 2017. County-level data were extracted by averaging the grid values.

Statistical analysis

Effect size and economic burden estimation

A three-stage time series design was used to estimate the association of short-term NO2 exposure with cardiovascular disease, coronary heart disease, stroke, and hypertension and the corresponding attributable hospital admissions, hospital days, and total hospital expenses.

In the first stage, a time series of a quasi-Poisson generalized linear regression model allowing for overdispersed admission counts was used to estimate county-specific associations [30].

$$Log\left(E\left({Y}_t\right)\right)=\alpha +\beta \left({NO}_2\right)+ Day\ of\ the\ week+ Holiday+ ns\left( calendar\ time, df=7\ per\ year\right)+ ns\left( temperature, df=6\right)+ ns\left( relative\ humidity, df=3\right)$$
(1)

In the model, several confounding covariates were incorporated, including daily mean temperature, relative humidity, calendar time, holiday, and day of the week, which were predefined according to previously published studies [31, 32]. In the equation, E(Yt) is the expected count of admissions in the analyzed county on day t, and β(NO2) is the log relative risk of hospital admissions associated with a 10-μg/m3 increase in NO2. Following Tian et al., the day of the week and holiday are the indicator variables to account for possible differences between weekdays and weekends and holidays and nonholidays. Previous studies have found variations in healthcare service utilization between weekdays and weekends [33], holidays and non-holidays [34]. In Chinese cultural context, people prefer not to receive health services on holidays (such as Spring Festival and Mid-Autumn Festival, which are usual days for family reunion). Moreover, in China, there could be overlap between the two variables, but few days on both holidays and weekends. ns (temperature) and ns (relative humidity) are natural cubic splines with 6 df for the 3-day moving average temperature and 3 df for the 3-day moving average relative humidity to adjust for potential lag and nonlinear effects effect of temperature and relative humidity; ns (calendar time) is a natural cubic spline function of time with seven degrees of freedom (df) per year to adjust for seasonality and time trends. Confounding effects of time-invariant or slowly varying risk factors at the individual level (e.g., sex, age, and comorbidities) could be naturally controlled for in the model [32].

This study modeled the association between NO2 and hospital admissions using a distributed lag model with a linear lag response function, inspecting the lag structure on a single lag day of 0 to 4 and moving average of the present and previous days (lag 0–4), respectively, to identify the optimal lag choices. In this model, lag 0 corresponded to the present day, lag 1 to the previous day, lag 2 to the day before lag 1, lag 3 to the day before lag 2, and lag 0–1 represented the two-day moving average of the present and previous day. The single-day and cumulative exposure effects were calculated based on exposures defined by these two different lag structures, and the estimated cumulative exposure effect was similar to the sum of the coefficients for single-day exposure effects in the distributed lag model [35].

In the second stage, random effects meta-analyses were applied to pool the county-specific associations to obtain urban, rural and overall estimates [36, 37]. The associations were calculated and expressed as the percentage change (95% CI) for each NO2 increase of 10 μg/m3.

$$RR={e}^{\beta }$$
(2)
$$Percentage\ change\ \left(\%\right)=\left( RR-1\right)\ast 100$$
(3)

A two-sample test was implemented to assess statistically significant differences in the estimates (E) between urban and rural areas based on the point estimate and standard error (SE) [38].

$$Z=\frac{E_{urban}-{E}_{rural}}{\sqrt{SE{\left({E}_{urban}\right)}^2+ SE{\left({E}_{rural}\right)}^2}}$$
(4)

In the third stage, the urban and rural effect estimates from the second stage were used to calculate the attributable number (AN) and attributable fraction (AF) [39] corresponding to the reduction in hospital admissions, length of hospital stays and total hospital expenses at the optimal lag choice period when NO2 concentrations reached the 2005 World Health Organization Global Air Quality Guidelines (WHO 2005 AQG) and WHO 2021 AQG, respectively.

$$A{N}_ic= N_ic\ast \frac{RR_ic-1}{RR_ic}, with\ {RR}_ic={e}^{\left(\beta \ast \frac{D_ic- AQG}{10}\right)}$$
(5)

where ANic is the county-specific attributable number of hospital admissions; Nic is the annual total hospital admissions in year i for urban or rural county c; β is the coefficient derived from the second stage; Dic is the annual average concentration of NO2 in year i for county c; and AQG is the World Health Organization Air Quality Guidelines annual average concentration of NO2, which is 40 μg/m3 for the WHO 2005 AQG and 10 μg/m3 for the WHO 2021 AQG. Total AN is summed by ANic. AF is calculated by dividing the AN by the sum of Nic.

In addition, the AFs and ANs of total hospital stays and expenses were estimated using the following formula [40, 41]:

$$ANec= AEc\ast \sum A{N}_ic, ANdc= ADc\ast \sum A{N}_ic$$
(6)

where ANec and ANdc are the county-specific attributable number of hospital expenses and hospital stays, respectively. ANic is the AN of hospital admissions for county c during the study period. AEc and ADc are the average expenses and average length of stay for all admissions in county c during the study period. AF was calculated by dividing the AN by the sum of the total expenses.

Potential reasons for urban–rural disparities

Finally, meta-regression models with county-level social characteristics (such as access to health care and GDP per capita) as independent variables were employed to check the role of social inequality in the urban–rural association. County-level association estimation could not adjust for individual-level risk factors for CVD, such as lifestyle factors and obesity. Instead, given the relationship of GDP with lifestyle and obesity, we attempted to explore the role of economic development in the association between NO2 exposure and CVD risk at lag day of 0. These methods expanded into multivariate meta-regression models with specific predictors to explain the potential heterogeneity, representing a refined parameterization within the linear mixed effects meta-analytic framework [42].

Moreover, health service utilization between rural and urban residents was compared to explore the potential explanation for urban–rural disparities in the economic burden attributed to NO2 exposure. In addition, this study followed up on the death outcome of the participants up to October 2020, and the survival curve of CVD between urban and rural residents was evaluated using Kaplan–Meier curves [43].

Sensitivity analysis

Since older people are less likely to move between rural and urban areas, the association among older people was also evaluated to reduce the bias resulting from the dynamic movement of citizens between rural and urban areas. In addition, five other co-pollutants (PM10, PM2.5, SO2, CO, and O3) were added to fit the two-pollutant model. By doing so, this study could determine the independent effects of short-term exposure to NO2 on CVD admissions. Based on previous studies, the association between short-term exposure to NO2 and increased risk of hospitalization for CVD was assumed to be linear in the main model [7, 44]. To explore the potential nonlinear correlation, NO2 was adjusted for using natural cubic splines, two knots were set at concentrations of 20 μg/m3 and 40 μg/m3, and the meta-smoothing method was used to summarize the concentration–response relationship curves.

All statistical analyses were conducted in R software (version 4.2.0) using the tsModel and dlnm packages for fitting first-stage models, the mixmeta package for performing meta-analyses, and the survminer package for performing survival analysis.

Results

Descriptive statistics

Table 1 shows the summary statistics for total hospital admissions, lengths of stay, and total expenses for each estimated disease during the study period in both urban and rural areas. Among the 303,217 hospital admissions for CVD, coronary heart disease had the largest number of hospital admissions in total (85,168) and urban areas (50,487), while ischemic stroke had the largest number of hospital admissions in rural areas (37,276). The total hospital stays and hospital expenses reached 4,302,498 days and 2,750,867 CNY for CVD, respectively. In urban areas, the total number of hospital days (3,032,005 days) and hospital costs (175,769 thousand CNY) for CVD were higher than those in rural areas (127,493 days and 1,000,098 thousand CNY).

Table 1 Characteristics of hospital admission cases in the included Shandong counties
Full size table

Table 2 shows the statistical description of NO2 concentrations, temperature, and relative humidity in the study areas. The daily average NO2 concentration during the study period was 36.1 μg/m3, which was slightly higher in urban areas (37.2 μg/m3) than in rural areas (35.1 μg/m3). The daily average concentration of NO2 ranged from 14.74 to 93.78 μg/m3 in urban areas. This value ranged from 14.49 to 87.00 μg/m3 in rural areas. Figure 1 displays the geographical distribution of the sampled urban and rural areas within Shandong Province, along with the distribution of average NO2 concentrations. This shows that the distribution of the sampling points was quite dispersed. In addition, ambient NO2 pollution intensified from east to west, a pattern that is consistent with the distribution characteristics observed across China. Among our sample sites, daily NO2 concentrations were similar in urban and adjacent rural areas of coastal cities such as Qingdao, Yantai, and Weihai, whereas compared to adjacent rural areas, the air pollution levels of urban areas were obviously higher in some inland cities such as Jinan, Zaozhuang and Linyi. Furthermore, county-level hospital admissions, along with air pollution levels, socioeconomic factors, and health care resource indicators, are shown in Table S2. Overall, GDP per capita and number of beds per thousand people were higher in urban areas.

Table 2 Description of daily air pollutant concentrations, temperature and relative humidity in the included Shandong counties
Full size table
Fig. 1
figure 1

The distribution of 39 Shandong counties and their average nitrogen dioxide concentrations

Full size image

Associations between NO2 exposure and hospital admissions for CVD

Figure 2  shows the total, urban, and rural estimates of the associations between NO2 and hospital admissions for CVD as well as coronary heart disease, ischemic disease, and hypertension on different lag days (including single-day lags and cumulative lags from 0 to 4 days). For the total effect estimates, a similar lagged pattern, namely, a significant and almost highest estimate at lag 0, was exhibited for total CVD and for the other three cause-specific diseases. For a 10-μg/m3 increase in NO2 concentrations, this study observed a significant percentage increase in hospital admissions on the day of exposure of 1.42% (95% confidence interval 0.92 to 1.92%) for CVD, 1.47% (0.59 to 2.35%) for coronary heart disease, 1.57% (0.64 to 2.51%) for ischemic stroke, and 2.54% (1.21 to 3.88%) for hypertension. However, as the lag day length increased, the single-day effect of cardiovascular disease began to show a protective effect, and the cumulative effect gradually decreased and was no longer significant. This is referred to as the ‘displacement’ phenomenon by Schwartz [45].

Fig. 2
figure 2

Percentage increase in cardiovascular disease, coronary heart disease, ischemic stroke, and hypertension hospital admissions per 10-μg/m3 increase in short-term ambient nitrogen dioxide exposure for overall (a) rural and urban counties (b). Note: The results were controlled for risk factors, including daily mean temperature, relative humidity, calendar time, public holidays, and day of the week. A distributed lag model was applied to estimate the county-specific associations, while random effects meta-analyses were used to pool the county-specific associations. Lag 0 corresponded to the present day, lag 1 to the previous day, lag 2 to the day before lag 1, lag 3 to the day before lag 2, and lag 0–1 represented the two-day moving average of the present and previous day

Full size image

Urban–rural disparities in the association between NO2 exposure and hospital admissions for CVD

The comparison of the urban and rural estimates is shown in Fig. 2(b). Separate estimates in urban and rural areas also suggested a significant effect on hospital admissions for CVD at lag 0, with 1.51% (0.82 to 2.21%) and 1.34% (0.58 to 2.10%), respectively. The estimated effects were nonsignificantly different for urban and rural counties (Table 3) at lag 0 (P value = 0.738). Similarly, the ‘displacement’ phenomenon was observed in the urban and rural estimates. Nevertheless, a more pronounced displacement phenomenon was found in rural areas, with a statistically significant difference (urban vs. rural: 1.43% (0.26 to 2.62%) vs. − 0.52% (− 1.89 to 0.88%), P value = 0.035). In addition, the results for the three cause-specific CVDs differed between urban and rural areas. For a 10-μg/m3 increase in NO2 concentrations, there was a significant percentage increase in hospital admissions on the day of exposure of 1.92% (0.56 to 3.29%) for ischemic stroke and 2.49% (0.50 to 4.53%) for hypertension among urban residents, while there was a significant increase of 2.18% (0.86 to 3.51%) for coronary heart disease and 2.45% (0.61 to 4.32%) for hypertension among rural residents.

Table 3 Percentage changes in hospital admissions associated with a 10-μg/m3 increase in NO2 on lag 0 days and lag 04 days
Full size table

The pooled concentration–response curve (Fig. S2) for the association between NO2 and hospital admissions showed positive and nearly linear concentration–response curves, with no discernible thresholds. In the sensitivity analysis, models with two pollutants were used. The associations of NO2 with total and cause-specific CVDs were still robust after adjustment for co-pollutants (Table 4). However, the effect size of NO2 increased slightly after adjustment for PM2.5, PM10, SO2, O3 and CO. Fig. S3 shows the comparison between the results of subsample analysis in the aging population (aged above 60, n = 214,792) and of all study subjects. Since older people are less likely to move between rural and urban areas, the results for the total population were very similar to those in the aging population, suggesting that the dynamic movement of citizens between rural and urban areas did not bias our estimations.

Table 4 Percentage changes in cardiovascular disease, coronary heart disease, ischemic stroke, and hypertension hospital admissions associated with a 10-μg/m3 increase in NO2 on lag 0 days, with and without adjustment for co-pollutants
Full size table

Economic burden attributed to NO2 exposure in terms of CVD

Table 5 shows the AFs and ANs of hospital admissions, total hospital days, and total expenses that could be reduced if annual NO2 concentrations reached the WHO 2021 AQG. By doing so, this study could reflect the disease burden and economic burden of CVD caused by NO2. At an annual average NO2 concentration of 10 μg/m3 (WHO 2021 AQG), the reduced AN of CVD hospital admissions would be 5447 (2990 to 7859) in urban areas, with an AF of 3.39% (1.86 to 4.89%). The results in rural areas were very similar to those in urban areas, with AN and AF values of 4765 (2099 to 7376) and 3.34 (1.47 to 5.17), respectively. However, there were apparent urban–rural differences in length of stay and hospital expenses. Total hospital days and expenses in urban areas would be reduced by 81,801 (44,831 to 118,191) days and 60,121 (33,002 to 86,729) thousand CNY, respectively, almost twice as much as in rural areas, with 42,131 (18,562 to 65,212) days and 32,875 (14,483 to 50,886) thousand CNY, respectively. Hypertension showed similar results: the burden of disease in urban areas was similar to that in rural areas, but the economic burden was almost twice as high. The previous WHO 2005 AQG standard set an annual average value of 40 μg/m3 for NO2, and this study also calculated the ANs and AFs of hospital admissions, hospital days, and hospital expenses for this guideline in both urban and rural areas (Table S4). Since many of our study areas already met this level, this guideline value would have resulted in a smaller reduction in both the disease and economic burden from NO2. The AN and AF were 327 (179 to 475) and 0.46% (0.25 to 0.66%) for hospital admissions for CVD in urban areas and 185 (81 to 289) and 0.26% (0.11 to 0.40%) in rural areas, respectively.

Table 5 Attributable numbers and fractions of hospital admissions, total hospital stays and total expenses (thousand CNY) that can be reduced when the annual NO2 concentration reaches the WHO 2021 AQG
Full size table

Exploration of the reasons behind urban–rural disparities

Table S3 shows the results of the meta-regression model adjusting for county-level access to health care and economic development. The estimated heterogeneity (I2) in the overall exposure–response associations for CVD hospital admissions at lag day of 0 between counties was 13.8%. Adding the GDP per capita and hospital beds per thousand people indicators to the model increased the model’s heterogeneity to 16.0% and decreased it to 12.8% respectively, with the P values of the coefficient estimates not being significant. The results suggest that the association between NO2 exposure and CVD risks at lag day of 0, when the effects are most pronounced, was not shaped by access to health care or economic development.

However, urban–rural differences in access to health care contributed to urban–rural disparities in the economic burden attributed to NO2 exposure. Table 6 displays the comparison of health service utilization between rural and urban residents. Rural residents tended to receive health care services in primary and secondary institutions and spent less per admission. In this study, 34.05 and 33.88% of rural patients received treatment in primary and secondary medical institutions, respectively, compared with 19.30 and 25.98% of urban patients. On average, total expenses per admission were lower among rural residents (urban vs. rural: 6826 CNY vs. 3854 CNY); when stratified by facility level, the differences became significantly smaller. Usually, tertiary institutions are more likely to provide high-quality health services, with a higher expense [46]. Thus, the urban–rural disparities in the economic burden attributed to NO2 exposure might be due to differences in access to high-quality health services between urban and rural residents. This study further tracked the mortality outcomes to support that urban residents had a high access to high-quality health services. The mortality outcomes of the study subjects was tracked until October 2020, and the Kaplan–Meier survival curve showed that there was a significant difference in the CVD survival time between urban and rural residents (Fig. 3). Urban residents showed a higher CVD survival probability, suggesting urban residents had a high access to high-quality health services. Since urban-rural difference in access to health services contributed to urban–rural disparities in the economic burden attributed to NO2 exposure, urban–rural disparities in the effects of NO2 exposure are a social problem in addition to environmental justice.

Table 6 Comparison of health service utilization between rural and urban residents
Full size table
Fig. 3
figure 3

Kaplan–Meier survival curve of CVD between urban and rural residents

Full size image

Discussion

The association between short-term NO2 exposure and hospitalization for CVD

Based on data on 303,217 hospital admissions for CVD, this study estimated the urban–rural disparities in the association between short-term NO2 exposure and hospitalization for CVD in Shandong Province, China. As expected, this study found that NO2 was positively and significantly associated with hospitalization for CVD. For every 10-μg/m3 increase in NO2 concentrations, there were significant percentage increases in hospital admissions on the day of exposure of 1.42% (95% confidence interval 0.92 to 1.92%) for CVD, 1.47% (0.59 to 2.35%) for coronary heart disease, 1.57% (0.64 to 2.51%) for ischemic stroke, and 2.54% (1.21 to 3.88%) for hypertension. The estimations were on par with those of studies conducted in China [4, 6, 47,48,49] but greater than those in a global meta-analysis [8]. The meta-analysis revealed that a 10-μg/m3 increase in 24-h NO2 exposure was associated with a 0.66% (0.32 to 1.01%) increase in CVD hospital admissions. Nevertheless, the meta-analysis included 204 studies up to 2011, of which there were only 2 studies from East Asia.

This study further conducted estimations on the economic burden attributed to NO2 exposure. If the concentration of NO2 could be reduced to the WHO 2021 AQG standard (10 μg/m3), the AN of CVD hospital admissions that could be reduced would be 10,162 (6638 to 13,637), with attributed hospital expenses of 91,269 (59,627 to 122,479) thousand CNY. Compared to the WHO AQG for 2021, the reduced ANs and economic burden according to the WHO AQG for 2005 were obviously smaller. This could be because more than half of our sample sites met the WHO 2005 AQG criteria. In 2017, there were only three counties that did not meet the criteria. However, our study continued to observe the health hazards resulting from NO2 exposure, which justifies the importance of the WHO 2021 AQG.

Urban–rural disparities in the association between short-term NO2 exposure and CVD hospital admissions

Separate estimations for urban and rural areas also suggested a significant effect of short-term NO2 exposure on CVD hospital admissions. A 10-μg/m3 increase in 24-h NO2 exposure was associated with 1.51% (0.82 to 2.21%) and 1.34% (0.58 to 2.10%) increases in CVD hospital admissions in rural and urban areas, respectively. However, the estimated effects were not significantly different for urban and rural counties. The results were consistent with those of Li et al. [13], who assessed urban–rural disparities in Beijing. In addition, the results were in agreement with those of Lin et al. [50] and Liu et al. [51], who compared the associations of short-term PM2.5/ozone with mortality (total, CVD, CED, and RESP) between urban cities and rural areas in Jiangsu and all of China. Conversely, no associations were found between short-term NO2 exposure and CVD mortality in urban or rural areas of Italy [52]. Moreover, a positive association between NO2 exposure and cardiovascular hospitalizations at lag 0 and lag 1 was found in urban areas of Guangxi Province [10], whereas the effect of NO2 was not significant in rural areas. The study populations had different urban–rural characteristics in terms of NO2 exposure across studies, which may have led to the differences in the results. For example, the rural areas of Italy and Guangxi were both characterized by low-level NO2 exposure, whereas the NO2 concentrations in the rural and urban areas of Shandong, Beijing, and Jiangsu were very similar.

However, a more pronounced displacement phenomenon was found in rural areas. The displacement phenomenon, referring to an increased risk ratio at short lags followed by an apparently protective effect at longer lags. This suggests that highly vulnerable people who are admitted to the hospital due to CVD may have simply had their problem brought forward by a few days as a result of NO2 exposure. Rural residents tended to have poorer overall health condition with lower health awareness compared to their urban counterparts, thus the displacement phenomenon might be greater among rural residents. This phenomenon has been observed for air pollution and temperature-related deaths [35, 45].

Urban–rural differences in the economic burden attributed to NO2 exposure

Despite similar effect sizes, hospital days and expenses resulting from NO2 exposure in urban areas were almost twice as high as those in rural areas. This fact may demonstrate that social inequality and environmental justice may be interrelated. First high-quality medical resources are mainly distributed in urban areas. Rural residents have limited access to high-quality health care compared to urban residents and tend to receive health care in primary and secondary medical institutions. Correspondingly, expenses are generally lower for primary and secondary medical institutions than for tertiary institutions. In addition, a low socioeconomic status constrains rural residents’ health investments. To avoid catastrophic medical expenditures, rural residents may reduce hospital days and expenses [53]. Of course, our study could not clarify the potential explanation. Thus, additional research is warranted to better understand precisely how these differences may contribute to health disparities between urban and rural areas. However, based on our results, urban–rural disparities in the effects of NO2 exposure are a social problem in addition to environmental justice.

Contributions and limitations

Based on the effects of short-term NO2 exposure on CVD hospitalization, this study also included the urban–rural disparities in the economic burden attributed to NO2 exposure. Many studies have evaluated the range of economic burdens associated with health problems caused by air pollution and made specific estimates [54,55,56,57,58,59]. Nevertheless, to our knowledge, no studies have compared these estimates between urban and rural areas. In addition, this study linked urban–rural disparities in environmental health with social inequality, which may enhance our understanding of the urban–rural disparities and the differences in previous studies. Our findings can help to provide justification for a coordinated intervention system that targets environmental factors and socioeconomic inequality simultaneously.

However, our study has several limitations. First, exposure misclassification could have occurred in our time series study design. Specifically, (i) county-level daily ambient air pollution could not exactly reflect personal exposure, and (ii) due to limited available data, we could not identify the divergence resulting from indoor-outdoor exposure. Since rural residents are more likely to be exposed to indoor NO2, the effect size of ambient NO2 exposure may be overestimated in rural areas, and urban-rural disparities in the effects of ambient NO2 exposure may be underestimated. However, disparities in the economic burden may remain. Second, the study area covered only Shandong Province, China. However, from east to west in Shandong Province, the environmental and sociodemographic characteristics are similar to those in China as a whole. Data spanning a wide range of economies and environments provide a good sample for analyzing the association. Third, our study sample was selected separately from urban community populations in municipal districts and rural villagers in rural counties to better compare typical rural–urban differences; however, this may have led to an overestimation of urban–rural disparities. Finally, our study was a time series observational study and could not control for sufficient variables; thus, this study could not explain the pathways of this phenomenon. However, we stratified our sample according to age and employed meta-regression models with the inclusion of access to health services and GDP to capture more factors influencing CVD risks. Due to the above limitations, our results should be interpreted with caution.

Policy implications

Our investigation can provide guidance for the creation and implementation of prevention and intervention programs to mitigate the negative effects of short-term NO2 exposure. We need to attach importance to the early warning of NO2 pollution in the cardiovascular system. NO2 pollution in rural areas, which has often been neglected in the past, should receive special attention. In addition, a coordinated environmental health policy linking environmental and social factors with health is highly recommended to reduce the interaction between environmental inequities and social inequality. Mitigation and adaptation such as enhancing equity in access to health services and socioeconomic status, should be implemented in an integrated way. Finally, considering that dozens of counties meet the 2005 AQG standard but continue to be affected by NO2 exposure in terms of health, the WHO 2021 AQG should be widely adopted, and the guidelines should be dynamically adjusted.

Conclusion

Based on data on 303,217 hospital admissions for CVD in Shandong Province, China, this study applied a distributed lag model and random effects meta-analyses to estimate the short-term association between NO2 exposure and hospitalization for CVD in both rural and urban areas. Afterward, urban–rural differences in the AN and AF attributed to NO2 were compared to reveal urban–rural disparities in the economic burden of CVD attributed to NO2 exposure. This study found a positive and significant association. Although the effect size was slightly higher in urban areas, the urban–rural difference was not significant. Nevertheless, a more pronounced displacement phenomenon was found in rural areas, and hospitalization expenses were significantly higher in urban areas. Differences in access to high-quality health care and socioeconomic status may partly explain the urban–rural disparities in the economic burden. Urban–rural disparities in the health implications of short-term NO2 exposure are a social problem in addition to an environmental problem. Thus, we may need to pay special attention to these rural areas in terms of the health implications of NO2 exposure, and a coordinated environmental health policy linking environmental and social factors with health should be implemented. However, our results should be applied with caution because of potential measurement error of exposure assessment, and lack of risk factors.

Availability of data and materials

The datasets generated and/or analyzed during the current study are not publicly available to preserve individuals’ privacy.

Abbreviations

AF:

Attributable fraction

AN:

Attributable number

CHAP:

China high air pollutants dataset

95% CI:

95% Confidence interval

CNY:

Chinese yuan

CO:

Carbon monoxide

CVD:

Cardiovascular disease

ICD-10:

International statistical classification of diseases and related health problems 10th revision

NO2 :

Nitrogen dioxide

O3 :

Ozone

PM2.5 :

Particulate matter with an aerodynamic diameter of 2.5 μm or less

PM10 :

Particulate matter with an aerodynamic diameter of 10 μm or less

PPS:

Probability proportional to size

STET:

Space-time extremely randomized tree

SO2 :

Sulfur dioxide

UHC:

Universal health coverage

WHO 2005 AQG:

2005 World health organization global air quality guidelines

WHO 2021 AQG:

2021 World health organization global air quality guidelines

References

  1. Song CB, Wu L, Xie YC, He JJ, Chen X, Wang T, Lin YC, Jin TS, Wang AX, Liu Y, et al. Air pollution in China: status and spatiotemporal variations. Environ Pollut. 2017;227:334–47. https://doi.org/10.1016/j.envpol.2017.04.075.

    Article  CAS  PubMed  Google Scholar 

  2. Zhang K, Batterman S. Air pollution and health risks due to vehicle traffic. Sci Total Environ. 2013;450:307–16. https://doi.org/10.1016/j.scitotenv.2013.01.074.

    Article  CAS  PubMed  ADS  Google Scholar 

  3. Zhang QL, Du XH, Li HC, Jiang YX, Zhu XL, Zhang Y, et al. Cardiovascular effects of traffic-related air pollution: a multi-omics analysis from a randomized, crossover trial. J Hazard Mater. 2022;435 https://doi.org/10.1016/j.jhazmat.2022.129031.

  4. Li L, Huang SL, Tian YC, Ji JJ, Zhang Y, Hu J, et al. Short-term exposure to nitrogen dioxide and ischemic stroke incidence in Shenzhen, China: modification effects by season and temperature. Ecotox Environ Safe. 2022;239 https://doi.org/10.1016/j.ecoenv.2022.113644.

  5. Liu YR, Dong JY, Zhai GY. Association between air pollution and hospital admissions for hypertension in Lanzhou, China. Environ Sci Pollut Res. 2022;29:11976–89. https://doi.org/10.1007/s11356-021-16577-5.

    Article  CAS  Google Scholar 

  6. Ma YX, Zhang HP, Zhao YX, Zhou JD, Yang SX, Zheng XD, Wang SG. Short-term effects of air pollution on daily hospital admissions for cardiovascular diseases in western China. Environ Sci Pollut Res. 2017;24:14071–9. https://doi.org/10.1007/s11356-017-8971-z.

    Article  CAS  Google Scholar 

  7. Meng X, Liu C, Chen RJ, Sera F, Vicedo-Cabrera AM, Milojevic A, et al. Short term associations of ambient nitrogen dioxide with daily total, cardiovascular, and respiratory mortality: multilocation analysis in 398 cities. BMJ. 2021:372. https://doi.org/10.1136/bmj.n534.

  8. Mills IC, Atkinson RW, Kang S, Walton H, Anderson HR. Quantitative systematic review of the associations between short-term exposure to nitrogen dioxide and mortality and hospital admissions. BMJ Open. 2015;5 https://doi.org/10.1136/bmjopen-2014-006946.

  9. Lu XC, Yao T, Fung JCH, Lin CQ. Estimation of health and economic costs of air pollution over the Pearl River Delta region in China. Sci Total Environ. 2016;566:134–43. https://doi.org/10.1016/j.scitotenv.2016.05.060.

    Article  CAS  PubMed  ADS  Google Scholar 

  10. Zhang GL, Liu XX, Zhai SY, Song GX, Song HQ, Liang LZ, Kong YF, Ma R, He XX. Rural-urban differences in associations between air pollution and cardiovascular hospital admissions in Guangxi, Southwest China. Environ Sci Pollut Res. 2022;29:40711–23. https://doi.org/10.1007/s11356-021-18196-6.

    Article  CAS  Google Scholar 

  11. Zhao S, Liu SL, Hou XY, Sun YX, Beazley R. Air pollution and cause-specific mortality: a comparative study of urban and rural areas in China. Chemosphere. 2021;262 https://doi.org/10.1016/j.chemosphere.2020.127884.

  12. Ma C, Song Z, Zong QQ. Urban-rural inequality of opportunity in health care: evidence from China. Int J Environ Res Public Health. 2021;18 https://doi.org/10.3390/ijerph18157792.

  13. Mohai P, Saha R. Which came first, people or pollution? A review of theory and evidence from longitudinal environmental justice studies. Environ Res Lett. 2015;10 https://doi.org/10.1088/1748-9326/10/12/125011.

  14. Gong P, Liang S, Carlton EJ, Jiang QW, Wu JY, Wang L, Remais JV. Urbanisation and health in China. Lancet. 2012;379:843–52. https://doi.org/10.1016/S0140-6736(11)61878-3.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Xu JQ, Zheng J, Xu LZ, Wu HT. Equity of health services utilisation and expenditure among urban and rural residents under universal health coverage. Int J Environ Res Public Health. 2021;18 https://doi.org/10.3390/ijerph18020593.

  16. Ma LJC. Urban transformation in China, 1949-2000: a review and research agenda. Environ Plan A. 2002;34:1545–69. https://doi.org/10.1068/a34192.

    Article  Google Scholar 

  17. Chen MX, Zhou Y, Huang XR, Ye C. The integration of new-type urbanization and rural revitalization strategies in China: origin, reality and future trends. Land. 2021;10 https://doi.org/10.3390/land10020207.

  18. China Statistical Yearbook. National Bureau of Statistics of China. 2001. http://www.stats.gov.cn/sj/ndsj/2001c/mulu.htm. Accessed 23 Jul 2023 .

    Google Scholar 

  19. Liang LW, Wang ZB, Li JX. The effect of urbanization on environmental pollution in rapidly developing urban agglomerations. J Clean Prod. 2019;237 https://doi.org/10.1016/j.jclepro.2019.117649.

  20. Fang H, Eggleston K, Hanson K, Wu M. Enhancing financial protection under China's social health insurance to achieve universal health coverage. BMJ. 2019;365 https://doi.org/10.1136/bmj.l2378.

  21. Chen W, Zhang Q, Renzaho AMN, Zhou FJ, Zhang H, Ling L. Social health insurance coverage and financial protection among rural-to-urban internal migrants in China: evidence from a nationally representative cross-sectional study. BMJ Glob Health. 2017;2 https://doi.org/10.1136/bmjgh-2017-000477.

  22. Wu XM, Perloff JM. China’s income distribution, 1985-2001. Rev Econ Stat. 2005;87:763–75. https://doi.org/10.1162/003465305775098206.

    Article  Google Scholar 

  23. China Statistical Yearbook. National Bureau of Statistics of China. 2021. http://www.stats.gov.cn/sj/ndsj/2021/indexch.htm. Accessed 23 Jul 2023 .

    Google Scholar 

  24. Ying ML, Wang SJ, Bai C, Li Y. Rural-urban differences in health outcomes, healthcare use, and expenditures among older adults under universal health insurance in China. PLoS One. 2020;15 https://doi.org/10.1371/journal.pone.0240194.

  25. Jia P, Wang YF, Yang M, Wang LM, Yang XC, Shi XY, et al. Inequalities of spatial primary healthcare accessibility in China. Soc Sci Med. 2022;314 https://doi.org/10.1016/j.socscimed.2022.115458.

  26. Zhou NY, Cui ZH, Yang SM, Han X, Chen GC, Zhou ZY, Zhai CZ, Ma MF, Li LB, Cai M, et al. Air pollution and decreased semen quality: a comparative study of Chongqing urban and rural areas. Environ Pollut. 2014;187:145–52. https://doi.org/10.1016/j.envpol.2013.12.030.

    Article  CAS  PubMed  Google Scholar 

  27. Du WY, Yin CN, Wang HT, Li ZW, Wang WJ, Xue FZ, et al. Infectious diseases among elderly persons: Results from a population-based observational study in Shandong province, China, 2013-2017. J Glob Health. 2021;11 https://doi.org/10.7189/jogh.11.08010.

  28. Zhao L, Wang H-T, Ye R-Z, Li Z-W, Wang W-J, Wei J-T, et al. Profile and dynamics of infectious diseases: a population-based observational study using multi-source big data. BMC Infect Dis. 2022;22 https://doi.org/10.1186/s12879-022-07313-6.

  29. Li Z-W, Yin C-N, Wang H-T, Du W-Y, Wang W-J, Xue F-Z, Zhao L, Cao W-C, Cheeloo Ecohlth Consortium C. Incidence trend and disease burden of seven vaccine-preventable diseases in Shandong province, China, 2013-2017: Findings from a population-based observational study. Vaccine: X 2022, 10. https://doi.org/10.1016/j.jvacx.2022.100145.

  30. Gasparrini A, Armstrong B, Kenward MG. Distributed lag non-linear models. Stat Med. 2010;29:2224–34. https://doi.org/10.1002/sim.3940.

    Article  MathSciNet  CAS  PubMed  PubMed Central  Google Scholar 

  31. Yin P, He GJ, Fan MY, Chiu KY, Fan MR, Liu C, et al. Particulate air pollution and mortality in 38 of China's largest cities: time series analysis. BMJ. 2017:356. https://doi.org/10.1136/bmj.j667.

  32. Tian YH, Liu H, Wu YQ, Si YQ, Song J, Cao YY, et al. Association between ambient fine particulate pollution and hospital admissions for cause specific cardiovascular disease: time series study in 184 major Chinese cities. BMJ. 2019:367. https://doi.org/10.1136/bmj.l6572.

  33. Kostis WJ, Demissie K, Marcella SW, Shao YH, Wilson AC, Moreyra AE, Grp MS. Weekend versus weekday admission and mortality from myocardial infarction. N Engl J Med. 2007;356:1099–109. https://doi.org/10.1056/NEJMoa063355.

    Article  CAS  PubMed  Google Scholar 

  34. Lenti MV, Klersy C, Brera AS, Musella V, Benedetti I, Padovini L, et al. Clinical complexity and hospital admissions in the December holiday period. PLoS One. 2020:15. https://doi.org/10.1371/journal.pone.0234112.

  35. Bhaskaran K, Gasparrini A, Hajat S, Smeeth L, Armstrong B. Time series regression studies in environmental epidemiology. Int J Epidemiol. 2013;42:1187–95. https://doi.org/10.1093/ije/dyt092.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Sera F, Armstrong B, Blangiardo M, Gasparrini A. An extended mixed-effects framework for meta-analysis. Stat Med. 2019;38:5429–44. https://doi.org/10.1002/sim.8362.

    Article  MathSciNet  PubMed  Google Scholar 

  37. Viechtbauer W. Conducting Meta-analyses in R with the metafor package. J Stat Softw. 2010;36(1–48) https://doi.org/10.18637/jss.v036.i03.

  38. Altman DG, Bland JM. Statistics notes - interaction revisited: the difference between two estimates. BMJ. 2003;326:219–219. https://doi.org/10.1136/bmj.326.7382.219.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Xie Y, Li Z, Zhong H, Feng XL, Lu P, Xu Z, Guo T, Si Y, Wang J, Chen L, et al. Short-term ambient particulate air pollution and hospitalization expenditures of cause-specific cardiorespiratory diseases in China: a multicity analysis. Lancet Reg Health-W Pac. 2021;15:100232–100232. https://doi.org/10.1016/j.lanwpc.2021.100232.

    Article  Google Scholar 

  40. Chen GB, Zhang YM, Zhang WY, Li SS, Williams G, Marks GB, Jalaludin B, Abramson MJ, Luo FM, Yang D, et al. Attributable risks of emergency hospital visits due to air pollutants in China: a multi-city study. Environ Pollut. 2017;228:43–9. https://doi.org/10.1016/j.envpol.2017.05.026.

    Article  CAS  PubMed  Google Scholar 

  41. Wei YG, Wang Y, Di Q, Choirat C, Wang Y, Koutrakis P, et al. Short term exposure to fine particulate matter and hospital admission risks and costs in the Medicare population: time stratified, case crossover study. BMJ. 2019:367. https://doi.org/10.1136/bmj.l6258.

  42. Gasparrini A, Armstrong B, Kenward MG. Multivariate meta-analysis for non-linear and other multi-parameter associations. Stat Med. 2012;31:3821–39. https://doi.org/10.1002/sim.5471.

    Article  MathSciNet  CAS  PubMed  PubMed Central  Google Scholar 

  43. Kaplan EL, Meier P. Nonparametric-estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–81. https://doi.org/10.2307/2281868.

    Article  MathSciNet  Google Scholar 

  44. Liu C, Chen R, Sera F, Vicedo-Cabrera AM, Guo YM, Tong SL, Coelho M, Saldiva PHN, Lavigne E, Matus P, et al. Ambient particulate air pollution and daily mortality in 652 cities. N Engl J Med. 2019;381:705–15. https://doi.org/10.1056/NEJMoa1817364.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Schwartz J. Harvesting and long term exposure effects in the relation between air pollution and mortality. Am J Epidemiol. 2000;151:440–8. https://doi.org/10.1093/oxfordjournals.aje.a010228.

    Article  CAS  PubMed  Google Scholar 

  46. Xu HY, Yang YJ, Wang CS, Yang JG, Li W, Zhang X, et al. Association of Hospital-Level Differences in care with outcomes among patients with acute ST-segment elevation myocardial infarction in China. JAMA Netw Open. 2020:3. https://doi.org/10.1001/jamanetworkopen.2020.21677.

  47. Feng YT, Lang CF, Chen C, Asena MH, Fang Y, Zhang RD, et al. Association between air pollution exposure and coronary heart disease hospitalization in a humid sub-tropical region of China: a time-series study. Front. Public Health. 2023:10. https://doi.org/10.3389/fpubh.2022.1090443.

  48. Ren QQ, Li SY, Xiao CL, Zhang JZ, Lin H, Wang S. The impact of air pollution on hospitalization for cardiovascular and cerebrovascular disease in Shenyang, China. Iran J Public Health. 2020;49(1476–1484) https://doi.org/10.18502/ijph.v49i8.3891.

  49. Song J, Lu MX, Lu JG, Chao L, An Z, Liu Y, Xu DQ, Wu WD. Acute effect of ambient air pollution on hospitalization in patients with hypertension: a time-series study in Shijiazhuang. China Ecotox Environ Safe. 2019;170:286–92. https://doi.org/10.1016/j.ecoenv.2018.11.125.

    Article  CAS  Google Scholar 

  50. Lin CY, Ma YQ, Liu RY, Shao YC, Ma ZW, Zhou L, et al. Associations between short-term ambient ozone exposure and cause-specific mortality in rural and urban areas of Jiangsu, China. Environ Res. 2022:211. https://doi.org/10.1016/j.envres.2022.113098.

  51. Liu T, Meng HR, Yu M, Xiao YZ, Huang BA, Lin LF, et al. Urban-rural disparity of the short-term association of PM2.5 with mortality and its attributable burden. Innovation. 2021:2. https://doi.org/10.1016/j.xinn.2021.100171.

  52. Gariazzo C, Renzi M, Marinaccio A, Michelozzi P, Massari S, Silibello C, et al. Association between short-term exposure to air pollutants and cause-specific daily mortality in Italy. A nationwide analysis. Environ Res. 2023:216. https://doi.org/10.1016/j.envres.2022.114676.

  53. Yap A, Cheung MJ, Kakembo N, Kisa P, Muzira A, Sekabira J, Ozgediz D. From procedure to poverty: out-of-pocket and catastrophic expenditure for pediatric surgery in Uganda. J Surg Res. 2018;232:484–91. https://doi.org/10.1016/j.jss.2018.05.077.

    Article  PubMed  Google Scholar 

  54. Ban J, Ma RM, Zhang Y, Li TT. PM2.5-associated risk for cardiovascular hospital admission and related economic burdens in Beijing, China. Sci Total Environ. 2021:799. https://doi.org/10.1016/j.scitotenv.2021.149445.

  55. Ji YH, Liu B, Song J, Pan RB, Cheng J, Wang H, Su H. Short-term effects and economic burden assessment of ambient air pollution on hospitalizations for schizophrenia. Environ Sci Pollut Res. 2022;29:45449–60. https://doi.org/10.1007/s11356-022-19026-z.

    Article  CAS  Google Scholar 

  56. Jiang WYH, Chen H, Liao JQ, Yang X, Yang B, Zhang YQ, et al. The short-term effects and burden of particle air pollution on hospitalization for coronary heart disease: a time-stratified case-crossover study in Sichuan, China. Environ Health. 2022:21. https://doi.org/10.1186/s12940-022-00832-4.

  57. Wen B, Wu Y, Ye TT, Xu RB, Yu WH, Yu P, et al. Short-term exposure to ozone and economic burden of premature mortality in Italy: a nationwide observation study. Ecotox Environ Safe. 2022:241. https://doi.org/10.1016/j.ecoenv.2022.113781.

  58. Zhang P, Zhou XY. Pricing air pollution: evidence from short-term exposure to air pollution on hospitalization of acute bronchitis and chronic obstructive pulmonary disease in southwestern China. Int Health. 2022;14:572–9. https://doi.org/10.1093/inthealth/ihab071.

    Article  PubMed  Google Scholar 

  59. Zhang ZH, Yao MH, Wu WJ, Zhao X, Zhang JY. Spatiotemporal assessment of health burden and economic losses attributable to short-term exposure to ground-level ozone during 2015-2018 in China. BMC Public Health. 2021:21. https://doi.org/10.1186/s12889-021-10751-7.

Download references

Acknowledgements

Not applicable.

Funding

Shandong Natural Science Foundation General Project [ZR202211080088] and grants from Yellow River National Strategic Research Institute, Shandong University.

Author information

Author notes

  1. Yike Zhang and Mengxiao Hu contributed equally to this work.

Authors and Affiliations

  1. Department of Biostatistics, School of Public Health, Cheeloo College of Medicine, Shandong University, Jinan, Shandong, 250012, China

    Yike Zhang, Mengxiao Hu, Bowen Xiang, Haiyang Yu & Qing Wang

  2. National Institute of Health Data Science of China, Shandong University, Jinan, China

    Yike Zhang, Mengxiao Hu, Bowen Xiang, Haiyang Yu & Qing Wang

Authors
  1. Yike Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Mengxiao Hu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Bowen Xiang
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Haiyang Yu
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Qing Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

YZ: Conceptualization, Methodology, Software, Writing-Original draft. MH: Data curation, Writing-Original draft. BX: Visualization, Investigation. HY: Formal analysis. QW: Writing-Reviewing and Editing, Supervision. All the authors have read and approved the final manuscript.

Corresponding author

Correspondence to Qing Wang.

Ethics declarations

Ethics approval and consent to participate

This study was approved by the School of Public Health of Shandong University (approval number: LL2303021).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Table S1.

County names, county codes and sample sizes of the included Shandong counties. Table S2. Hospital admissions, socioeconomic and health care accessibility indicators, and ambient NO2 concentrations of each area during the study period. Table S3. Meta-regression results for per capita GDP and hospital beds per thousand people of 39 counties in two-stage model analysis. Table S4. Attributable numbers and fractions of hospital admissions, total hospital stays and total expenses (thousand CNY) that can be reduced when the annual NO2 concentration reaches the WHO 2005 AQG. Fig. S1. Three-stage cluster sampling process, demographic characteristics of the sample population, and sample selection flowchart. Fig. S2. Concentration–response curve between NO2 concentrations (lag 0) and cardiovascular disease (a), coronary heart disease (b), ischemic stroke (c), and hypertension (d) hospital admissions. The vertical scale can be interpreted as the relative change in the mean effect of NO2 on mortality; the fraction of the curve below zero denotes a smaller estimate than the mean effect. Fig. S3. Comparison between the results of (a) subsample analysis in the aging population (aged above 60) and among (b) all study subjects.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhang, Y., Hu, M., Xiang, B. et al. Urban–rural disparities in the association of nitrogen dioxide exposure with cardiovascular disease risk in China: effect size and economic burden. Int J Equity Health 23, 22 (2024). https://doi.org/10.1186/s12939-024-02117-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12939-024-02117-3

Keywords

International Journal for Equity in Health

ISSN: 1475-9276

Contact us