Advertisement
  • Loading metrics

Ambient air pollution and cause-specific risk of hospital admission in China: A nationwide time-series study

  • Jiangshao Gu ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

    gujs15@mails.tsinghua.edu.cn (JG); tingchen@tsinghua.edu.cn (TC)

    Affiliations Center for Big Data Research in Health and Medicine, Institute for Data Sciences, Tsinghua University, Beijing, China, Tsinghua-Fuzhou Institute of Digital Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China, Institute for Artificial Intelligence, State Key Lab of Intelligent Technology and Systems, Department of Computer Science and Technology, Tsinghua University, Beijing, China

  • Ying Shi,

    Roles Data curation, Writing – review & editing

    Affiliation China Standard Medical Information Research Center, Shenzhen, China

  • Yifang Zhu,

    Roles Data curation, Writing – review & editing

    Affiliations Department of Environmental Health Science, University of California at Los Angeles, Los Angeles, California, United States of America, Institute of the Environment and Sustainability, University of California at Los Angeles, Los Angeles, California, United States of America

  • Ning Chen,

    Roles Data curation, Funding acquisition, Writing – review & editing

    Affiliations Center for Big Data Research in Health and Medicine, Institute for Data Sciences, Tsinghua University, Beijing, China, Tsinghua-Fuzhou Institute of Digital Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China, Institute for Artificial Intelligence, State Key Lab of Intelligent Technology and Systems, Department of Computer Science and Technology, Tsinghua University, Beijing, China

  • Haibo Wang,

    Roles Data curation, Supervision, Writing – review & editing

    Affiliations China Standard Medical Information Research Center, Shenzhen, China, Clinical Trial Unit, First Affiliated Hospital of Sun Yat-Sen University, Guangzhou, China

  • Zongjiu Zhang ,

    Roles Data curation, Supervision, Writing – review & editing

    ‡ These authors are joint senior authors on this work.

    Affiliation Bureau of Medical Administration, National Health Commission of the People’s Republic of China, Beijing, China

  • Ting Chen

    Roles Conceptualization, Data curation, Funding acquisition, Supervision, Writing – review & editing

    gujs15@mails.tsinghua.edu.cn (JG); tingchen@tsinghua.edu.cn (TC)

    ‡ These authors are joint senior authors on this work.

    Affiliations Center for Big Data Research in Health and Medicine, Institute for Data Sciences, Tsinghua University, Beijing, China, Tsinghua-Fuzhou Institute of Digital Technology, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China, Institute for Artificial Intelligence, State Key Lab of Intelligent Technology and Systems, Department of Computer Science and Technology, Tsinghua University, Beijing, China

Ambient air pollution and cause-specific risk of hospital admission in China: A nationwide time-series study

  • Jiangshao Gu, 
  • Ying Shi, 
  • Yifang Zhu, 
  • Ning Chen, 
  • Haibo Wang, 
  • Zongjiu Zhang, 
  • Ting Chen
PLOS
x

Abstract

Background

The impacts of air pollution on circulatory and respiratory systems have been extensively studied. However, the associations between air pollution exposure and the risk of noncommunicable diseases of other organ systems, including diseases of the digestive, musculoskeletal, and genitourinary systems, remain unclear or inconclusive. We aimed to systematically assess the associations between short-term exposure to main air pollutants (fine particulate matter [PM2.5] and ozone) and cause-specific risk of hospital admission in China over a wide spectrum of human diseases.

Methods and findings

Daily data on hospital admissions for primary diagnosis of 14 major and 188 minor disease categories in 252 Chinese cities (107 cities in North China and 145 cities in South China) from January 1, 2013, to December 31, 2017, were obtained from the Hospital Quality Monitoring System of China (covering 387 hospitals in North China and 614 hospitals in South China). We applied a 2-stage analytic approach to assess the associations between air pollution and daily hospital admissions. City-specific associations were estimated with quasi-Poisson regression models and then pooled by random-effects meta-analyses. Each disease category was analyzed separately, and the P values were adjusted for multiple comparisons. A total of 117,338,867 hospital admissions were recorded in the study period. Overall, 51.7% of the hospitalized cases were male, and 71.3% were aged <65 years. Robust positive associations were found between short-term PM2.5 exposure and hospital admissions for 7 major disease categories: (1) endocrine, nutritional, and metabolic diseases; (2) nervous diseases; (3) circulatory diseases; (4) respiratory diseases; (5) digestive diseases; (6) musculoskeletal and connective tissue diseases; and (7) genitourinary diseases. For example, a 10-μg/m3 increase in PM2.5 was associated with a 0.21% (95% CI 0.15% to 0.27%; adjusted P < 0.001) increase in hospital admissions for diseases of the digestive system on the same day in 2-pollutant models (adjusting for ozone). There were 35 minor disease categories significantly positively associated with same-day PM2.5 in both single- and 2-pollutant models, including diabetes mellitus, anemia, intestinal infection, liver diseases, gastrointestinal hemorrhage, renal failure, urinary tract calculus, chronic ulcer of skin, and back problems. The association between short-term ozone exposure and respiratory diseases was robust. No safety threshold in the exposure–response relationships between PM2.5 and hospital admissions was observed. The main limitations of the present study included the unavailability of data on personal air pollution exposures.

Conclusions

In the Chinese population during 2013–2017, short-term exposure to air pollution, especially PM2.5, was associated with increased risk of hospitalization for diseases of multiple organ systems, including certain diseases of the digestive, musculoskeletal, and genitourinary systems; many of these associations are important but still not fully recognized. The effect estimates and exposure–response relationships can inform policy making aimed at protecting public health from air pollution in China.

Author summary

Why was this study done?

  • Besides the well-known cardiorespiratory effects of air pollution, an increasing number of studies suggest that air pollution might be associated with certain non-cardiorespiratory diseases.
  • Evidence of associations between air pollution exposure and the risk of many non-cardiorespiratory diseases is still scarce and inconclusive.
  • Few studies have characterized the acute health effects of air pollution on multiple organ systems using uniform methodology and databases.

What did the researchers do and find?

  • We conducted a national time-series study using data of 117,338,867 hospital admissions for 14 major and 188 minor disease categories in 252 Chinese cities from 2013 to 2017, to assess the associations between short-term exposure to fine particulate matter (PM2.5) and ozone and cause-specific risk of hospital admission on a national scale.
  • City-specific associations were estimated with quasi-Poisson regression models and then pooled by random-effects meta-analyses. Each disease category was analyzed separately, and the P values were adjusted for multiple comparisons.
  • Short-term PM2.5 exposure was significantly positively associated with hospital admissions for 13 major disease categories (of which 7 associations were considered robust, including for diseases of the digestive, musculoskeletal, and genitourinary systems) and 35 minor disease categories, whether adjusted for ozone or not. The association between short-term ozone exposure and respiratory diseases was robust.

What do these findings mean?

  • To our knowledge, this is the first national study in China aimed at systematically investigating possible ways in which short-term air pollution exposure may be associated with severe illnesses requiring hospitalization.
  • Our findings highlight the extensive adverse impacts of air pollution on human health, and indicate the significant social benefits of effective mitigation measures.

Introduction

Air pollution has become a public health concern worldwide, with exposure linked to increased morbidity and mortality [14]. Criteria air pollutants include fine particulate matter of 2.5 μm or less in aerodynamic diameter (PM2.5), tropospheric ozone (O3), sulfur dioxide (SO2), nitrogen dioxide (NO2), and carbon monoxide (CO). In particular, PM2.5 and O3 were used by the Global Burden of Disease Study as the 2 indicators to quantify population exposure to air pollution when estimating the global burden of disease attributable to ambient air pollution, since they were the most consistent and robust predictors of adverse outcomes in previous studies [5].

The health effects of air pollution on the circulatory and respiratory systems are well documented [6,7]. There is emerging evidence for increased risk of some non-cardiorespiratory diseases related to air pollution, e.g., diabetes, autism in children, dementia in the elderly, and premature birth and low birthweight [8,9]. Several recent studies also revealed positive associations between air pollution and certain diseases of the digestive, skeletal, and urinary systems, e.g., peptic ulcer bleeding [10], bone loss over time and bone fracture [11], and incident chronic kidney disease and progression to end-stage renal disease [1214]. There are at least 3 hypotheses to explain the extrapulmonary effects of air pollution. First, inhaled pollutants may induce pulmonary inflammation and oxidative stress, which is sufficient to cause systemic inflammation and oxidative stress [6,15]. Second, the lung autonomic nervous system may be provoked by pulmonary exposure, which could then result in autonomic nervous system imbalance [15]; the levels of stress hormones may also be altered [16]. Third, the pollutants may reach and directly interact with remote organs, e.g., by penetrating the alveoli and entering the circulation [17,18]. However, the epidemiological evidence is still scarce and inconclusive for many diseases, and is usually subject to small data size, limited representativeness, and possible publication bias. There are few studies that have characterized the health effects of air pollution on multiple organ systems using uniform methodology and databases.

China has a population of more than 1.3 billion being exposed to relatively high levels of air pollution. In recent years, China has gradually built up high-quality national databases for inpatient discharge registration, which, combined with national monitoring networks of ambient air quality, provide an opportunity to systematically investigate possible ways in which air pollution exposure may be associated with severe illnesses requiring hospitalization. In this study, we performed a nationwide time-series analysis based on hospital admissions in 252 Chinese cities from 2013 to 2017, to assess the associations between short-term exposure to main air pollutants (PM2.5 and O3) and cause-specific risk of hospital admission on a national scale. Diseases were classified by 2 approaches at different granularities. Potential effect modifiers and exposure–response relationships were also evaluated.

Methods

Data collection

The daily hospital admission data during the period January 1, 2013, to December 31, 2017, came from the Hospital Quality Monitoring System (HQMS), a national registration database of electronic inpatient discharge records of class 3 hospitals in China, under the administration of the National Health Commission of the People’s Republic of China. Class 3 hospitals are the highest ranked medical institutions in China’s healthcare system, corresponding to tertiary hospitals in the US and Europe but also providing primary and secondary care to less serious patients. In Western countries, healthcare resource utilization is generally scheduled by appointment. However, there is no general-practitioner-based referral system in China [19,20]. Regular outpatient visits and hospital admissions are generally unscheduled and are on a first-come, first-served basis. People usually go to hospital promptly when they develop symptoms and will be admitted immediately if necessary. Ninety-five percent of the total Chinese population was covered by social health insurance schemes by the end of 2017 [21]. Therefore, hospital admission records can provide reliable and timely information on the health status of a geographically defined population in China. Since January 1, 2013, class 3 hospitals in China have been mandated to automatically submit inpatient discharge records to the HQMS on a daily basis, in a nationally standardized format. Each record describes a hospital stay using 346 information items, from which we extracted the date of admission, sex, age, and the primary discharge diagnosis. The diagnosis is coded using the International Classification of Diseases, 10th Revision (ICD-10) [22] by certified professional medical coders at each hospital. Quality control is automatically performed at the time of submission to guarantee the completeness, consistency, and accuracy of data. By December 31, 2017, the HQMS covered 1,001 (67.0%) class 3 hospitals and 1,011,375 (78.4%) hospital beds in 252 Chinese cities. The population, the total number of class 3 hospitals, the number of class 3 hospitals covered by the HQMS, the total number of hospital beds, the number of hospital beds covered by the HQMS, and the coverage rates of class 3 hospitals and hospital beds by the HQMS in each of the 252 cities are presented in S1 Table. This database has been used to track the epidemiological situation and trends of several diseases in China [2325].

To clearly characterize the health effects of air pollution over a wide spectrum of human diseases, we defined the disease categories at 2 levels. Major disease categories are based on the chapter division (first-level classification) of the ICD-10 diagnostic coding system [22]. Specifically, we treated each ICD-10 chapter as a candidate category and assigned each diagnosis code to the ICD-10 chapter that it belongs to. There are 22 ICD-10 chapters in total, of which the first 14 were included in the scope of this study (14 major disease categories). Minor disease categories are based on the Clinical Classifications Software (CCS), a validated approach developed by the Agency for Healthcare Research and Quality [26]. The CCS is a comprehensive categorization scheme for aggregating diagnosis codes into a manageable number of disease categories on the basis of clinical homogeneity, and has been widely used to examine patterns of specific health conditions [27,28]. All diagnosis codes were collapsed into 260 mutually exclusive CCS categories, of which we selected as candidates 191 that conceptually belonged to the 14 major disease categories. We excluded 3 categories (CCS codes 56, 61, and 75) that had too few data, leaving 188 as the final set of minor disease categories. For each cause (major or minor disease category), we obtained the daily counts of citywide hospital admissions by summing the daily number of admissions for primary diagnosis of this cause in each hospital in a city. All data used were anonymized and de-identified prior to analysis, under the supervision of Bureau of Medical Administration, National Health Commission of the People’s Republic of China. Because the data were analyzed at the aggregate level with no individual identifiers involved, institutional review board approval and participant written consent were not required for this study.

The daily air pollution data came from the National Air Pollution Monitoring System (http://106.37.208.233:20035/), administered by China’s Ministry of Ecology and Environment. The number of monitoring stations in a city ranges from 1 to 17, with a median of 4. These fixed-site stations are mandated not to be situated in the neighborhood of distinct emission sources (including but not limited to traffic, industry, and open burning). All measurement procedures are in accordance with China’s Ambient Air Quality Standards (GB3095-2012). For each city, we derived daily 24-hour average concentrations of PM2.5, SO2, NO2, and CO and maximum 8-hour average concentrations of O3, averaged across all valid monitoring sites, to represent the population exposure to ambient air pollution. We also collected daily mean temperature and mean relative humidity for each city from the China Meteorological Data Sharing Service System (http://data.cma.cn/). During the study period, the missing data rates were 1.69% for PM2.5, 1.72% for O3, 1.68% for SO2, 1.67% for NO2, 1.72% for CO, 0.70% for temperature, and 0.70% for relative humidity. Days with missing monitoring measurements were excluded from analysis.

Statistical analysis

We applied a 2-stage analytic approach to assess the associations between short-term exposure to air pollutants and daily hospital admissions; this approach has been widely used in previous multisite time-series studies [1,2933]. The methods and models used in this study were determined before the analyses were conducted. In the first stage, we fit quasi-Poisson regression models separately to the daily time-series data of air pollutants and hospital admissions in each city for each cause, linked by date, to estimate the city- and cause-specific relative risk (RR) of hospital admission associated with air pollutants. The potential confounding effects of weather, seasonality, and long-term patterns were controlled for by smoothing functions (natural cubic splines). Specifically, we introduced the following covariates in the models, which were prespecified according to previously published studies [1,19,20,2938]: (1) a natural cubic spline smoother of calendar day with 10 degrees of freedom (df) per year; (2) natural cubic spline smoothers of the temperature on the same day as admission (lag 0) and the average temperature over the 3 days before admission (lag 1–3), both with 6 df; (3) natural cubic spline smoothers of the relative humidity at lag 0 and lag 1–3, both with 3 df; (4) indicator variables for the day of the week and public holidays. The df values for calendar day, temperature, and relative humidity were selected based on the parameters used in previously published studies [1,19,20,2938], and were further examined by sensitivity analyses described below. In the second stage, the city-specific RR estimates were pooled by random-effects meta-analyses to generate the national average RR estimates and P values testing the null hypothesis of no association [36,39].

For PM2.5 and O3, we estimated their effects on daily hospital admissions using single-pollutant models and 2-pollutant models (adjusting for each other), as well as 3-pollutant models additionally adjusting for 1 of SO2, NO2, and CO to examine the robustness of our results. Multiple lag structures of PM2.5 and O3 (single-day and moving average exposures up to 3 days before admission) were used to explore the lag patterns in the acute effects of air pollution exposure. We always used the 2-day moving average exposure (lag 0–1) as the exposure metric of co-pollutants in 2- and 3-pollutant models [32,33]. To address the multiple testing problem, we applied the Benjamini–Hochberg procedure to adjust the P values [40]. Specifically, given a pollutant and a model specification (including the lag of exposure, adjustment for co-pollutants, and df for smoothing functions), we could derive 14 and 188 P values for major and minor disease categories, respectively. The 14 or 188 P values were treated as a batch and adjusted together through the Benjamini–Hochberg procedure. The associations with adjusted P < 0.05 were considered statistically significant, keeping the false discovery rate < 5%.

To explore potential effect modifiers, we conducted subgroup analyses by sex (male and female), age (<65, 65–74, and ≥75 years), season (cool season, from October to March; warm season, from April to September) [3,4], and region (North China and South China, divided by the Huai River–Qinling Mountain line [the 33th parallel of north latitude, in practice]) (S1 Fig) [41]. The statistical significance of effect modification (difference in effect estimates between subgroups) was tested by a 2-sample Z-test [3,4]. We did not adjust the P values for association or effect modification in the subgroup analyses due to the exploratory nature of these analyses. Following the suggestion of reviewers, we further divided China into 6 regions based on geography, climate, and culture (middle north, northeast, east, middle south, southwest, and northwest) (S1 Fig), and repeated the main analyses in each region to obtain the regional average estimates. We developed the exposure–response curves of the relationship between air pollution and hospitalization at the national level as was done in previous studies [42,43]. Briefly, we replaced the linear term for air pollutants in the first-stage regression model with a natural cubic spline smoother with 3 df (technically, the coefficient of log RR was replaced by 3 coefficients used to represent the spline, and the variance of the coefficient was replaced by a 3 × 3 variance–covariance matrix). We then conducted multivariate random-effects meta-analyses to generate the national average estimates of the exposure–response curves.

As sensitivity analyses, we changed the df for the smoothing function of calendar day over the range of 6 to 14 per year, for temperature over the range of 3 to 9, and for relative humidity over the range of 3 to 9. These ranges were chosen to be consistent with previous nationwide studies [19,20,31,3638]; we specially tested if the associations remained after more aggressive adjustment for potential confounding effects of time trends and weather conditions (i.e., using more df). Following the suggestion of reviewers, we included 8 outcomes (the CCS categories birth trauma; joint disorders and dislocations, trauma-related; spinal cord injury; sprains and strains; poisoning by psychotropic agents; poisoning by other medications and drugs; rehabilitation care, fitting of prostheses, and adjustment of devices; and medical examination/evaluation) as negative controls [44], for which no biological or clinical evidence supports an association with air pollution. We repeated the main analyses on these negative control outcomes to examine if our results were substantially biased owing to residual confounding.

Finally, to gauge the potential public health benefits of air pollution control measures, we calculated the annual reduction in hospital admissions (H) and hospitalization expenses (E) attributable to a 10-μg/m3 reduction in the daily PM2.5 level in China [1,20,30,31,37]. H is defined as H = (exp(β × Δx) − 1) × N, where β is the national average effect estimate (log RR) for a 1-μg/m3 change in PM2.5 from the main analyses, Δx is 10 μg/m3, and N is the total number of hospital admissions in China in 2016, collected from China Health and Family Planning Statistical Yearbook 2017 (S2 Table). E is defined as E = c × H, where c is the average cost for each hospitalization in China in 2016, collected from China Health and Family Planning Statistical Yearbook 2017 (S2 Table). H and E were calculated for each cause and then added up to get an overall estimate.

Results are presented as point estimates and 95% confidence intervals (CIs) of the percentage increase in daily hospital admissions associated with a 10-μg/m3 increase in PM2.5 or O3. P values are always 2-sided. All analyses were done in R software version 3.4.4 (R Foundation for Statistical Computing), with package mgcv for fitting regression models and package mvmeta for conducting random-effects meta-analyses.

Results

A total of 117,338,867 hospital admissions in 252 Chinese cities (107 cities in North China and 145 cities in South China) (S1 Fig) from 2013 to 2017 were included in this study, collected from 1,001 class 3 hospitals covered by the HQMS (387 hospitals in North China and 614 hospitals in South China) (S1 Table). The demographic characteristics of the health data are provided in Table 1 and S3 Table. Overall, 51.7% of the hospitalized cases were male, and 71.3% were aged <65 years. The distributions of daily counts of citywide hospital admissions are provided in Table 2. On average per city, there were 273.3 hospital admissions per day for 14 major disease categories and 188 minor disease categories. The national average levels of ambient air pollutants were 50.6 μg/m3 for PM2.5, 87.2 μg/m3 for O3, 24.0 μg/m3 for SO2, 31.4 μg/m3 for NO2, and 1.1 mg/m3 for CO during the study period. More statistics of the environmental data (distributions of citywide annual-average levels and national average Pearson correlation coefficients) are provided in S4 and S5 Tables.

thumbnail
Table 1. Demographic characteristics of hospital admissions for 14 major disease categories in 252 Chinese cities, 2013–2017.

https://doi.org/10.1371/journal.pmed.1003188.t001

thumbnail
Table 2. Summary statistics of annual-average daily counts of citywide hospital admissions in 252 Chinese cities, 2013–2017.

https://doi.org/10.1371/journal.pmed.1003188.t002

Generally, the effect estimates for PM2.5 for most disease categories were largest on the same day (lag 0 days), and decreased sharply in the subsequent days (S2 Fig; S6S12 Tables). The lag patterns of the effect estimates for O3 varied by disease category, with same-day exposure (lag 0 days) leading to the largest effect estimates for some diseases (e.g., diseases of the circulatory system) and previous-day exposure (lag 1 day) leading to the largest effect estimates for other diseases (e.g., diseases of the respiratory system) (S3 Fig; S13S19 Tables). At the national level, the same-day concentration of PM2.5 (lag 0 days) was significantly positively associated with hospital admissions for 13 major disease categories in both single- and 2-pollutant models (Fig 1). For example, each 10-μg/m3 increase in same-day PM2.5 was associated with a 0.21% (95% CI 0.15% to 0.27%; adjusted P < 0.001) increase in hospital admissions for diseases of the digestive system when adjusted for O3. In particular, the associations of PM2.5 with 7 major disease categories were robust to the lag of exposure and further adjustment for co-pollutants (S2 and S4 Figs): (1) endocrine, nutritional, and metabolic diseases; (2) diseases of the nervous system; (3) diseases of the circulatory system; (4) diseases of the respiratory system; (5) diseases of the digestive system; (6) diseases of the musculoskeletal system and connective tissue; and (7) diseases of the genitourinary system. At the national level, the 2-day moving average concentration of O3 (lag 0–1 days) was significantly positively associated with hospital admissions for diseases of the respiratory system in both single- and 2-pollutant models, and the association was robust to the lag of exposure and further adjustment for co-pollutants (Fig 1, S3 and S5 Figs).

thumbnail
Fig 1. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 and O3 by major disease category, on average across all cities.

Results are presented as point estimates and 95% CIs of the percentage increase in daily hospital admissions associated with a 10-μg/m3 increase in PM2.5 or O3. Major disease categories are based on the chapter division of the ICD-10 diagnostic coding system. Single-day exposure on the same day (lag 0) was used as the exposure metric of PM2.5. Two-day moving average exposure (lag 0–1) was used as the exposure metric of O3. In single-pollutant models, the effects of PM2.5 and O3 were estimated without adjustment for co-pollutants; in 2-pollutant models, the effects of PM2.5 were estimated after adjustment for O3, and the effects of O3 were estimated after adjustment for PM2.5. The Benjamini–Hochberg procedure was applied to adjust the P values across the 14 major disease categories. *Statistically significant estimate (P < 0.05).

https://doi.org/10.1371/journal.pmed.1003188.g001

The national average RR estimates for PM2.5 and O3 by minor disease category are shown in Table 3 and S6S19 Tables. The same-day concentration of PM2.5 (lag 0 days) was significantly positively associated with hospital admissions for 35 minor disease categories in both single- and 2-pollutant models, 6 categories in single-pollutant models only, and 4 categories in 2-pollutant models only. Among the 35 minor disease categories significantly associated with PM2.5 in both single- and 2-pollutant models, there were 9 categories of circulatory diseases including essential hypertension, acute myocardial infarction, ischemic heart diseases, pulmonary heart disease, cardiac dysrhythmias, congestive heart failure, acute cerebrovascular disease, and transient cerebral ischemia; 6 categories of respiratory diseases, including pneumonia, acute bronchitis, upper respiratory infections, and chronic obstructive pulmonary disease (COPD) and bronchiectasis; 7 categories of digestive diseases, including intestinal infection, esophageal disorders, gastritis and duodenitis, appendiceal conditions, liver diseases, gastrointestinal hemorrhage, and noninfectious gastroenteritis; 3 categories of genitourinary diseases, including chronic renal failure, calculus of urinary tract, and nephritis, nephrosis, and renal sclerosis; and 10 categories of other diseases, including bacterial infection, cancer of bronchus and lung, cancer of breast, diabetes mellitus without/with complications, crystal arthropathies, anemia, Parkinson disease, chronic ulcer of skin, and spondylosis, intervertebral disc disorders, and other back problems. For example, each 10-μg/m3 increase in same-day PM2.5 was associated with a 0.34% (95% CI 0.14% to 0.53%; adjusted P = 0.007) increase in hospital admissions for gastrointestinal hemorrhage when adjusted for O3. The 2-day moving average concentration of O3 (lag 0–1 days) was significantly positively associated with hospital admissions for pneumonia, upper respiratory infections, and asthma in both single- and 2-pollutant models, as well as acute bronchitis, COPD and bronchiectasis, and other upper respiratory disease in single-pollutant models only.

thumbnail
Table 3. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 and O3 by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.t003

The results of the subgroup analyses are shown in Fig 2, S6 and S7 Figs. We found stronger associations between PM2.5 and most disease categories in people aged 65–74 years or ≥75 years than in people aged <65 years; for several disease categories, the effect modification was statistically significant (in an exploratory analysis without adjustment for multiple comparisons) (Fig 2 and S6 Fig). However, we did not observe the same pattern for O3 (S7 Fig). The evidence of effect modification by sex, season, and region (using both 2- and 6-region divisions) was mixed and uncertain (S6, S7, S12 and S13 Figs). The results of the exposure–response analyses are shown in Fig 3, S8 and S9 Figs. The exposure–response curves between PM2.5 and hospital admissions generally showed steeper slopes at lower concentrations than at higher concentrations, with no safety threshold below which PM2.5 exposure had no health effect (Fig 3 and S8 Fig). A threshold of O3 in the range of 40 to 50 μg/m3 (near the background level of O3) was observed in the exposure–response curve between O3 and hospital admissions for diseases of the respiratory system (S9 Fig).

thumbnail
Fig 2. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 by major disease category for study subgroups, on average across all cities.

(A) Endocrine, nutritional, and metabolic diseases; (B) nervous diseases; (C) circulatory diseases; (D) respiratory diseases; (E) digestive diseases; (F) genitourinary diseases. Results are presented as point estimates and 95% CIs of the percentage increase in daily hospital admissions associated with a 10-μg/m3 increase in PM2.5. Major disease categories are based on the chapter division of the ICD-10 diagnostic coding system. Single-day exposure on the same day (lag 0) was used as the exposure metric of PM2.5. The effects of PM2.5 were estimated after adjustment for O3. The cool season is from October to March; the warm season is from April to September. The 2 regions of China (North and South) are divided by the Huai River–Qinling Mountain line. The P values were not adjusted for multiple comparisons. Results for all disease categories and for O3 are shown in S6 and S7 Figs. *Statistically significant estimate (P < 0.05).

https://doi.org/10.1371/journal.pmed.1003188.g002

thumbnail
Fig 3. Exposure–response curves of the effects of PM2.5 on hospital admissions by major disease category, on average across all cities.

(A) Endocrine, nutritional, and metabolic diseases; (B) nervous diseases; (C) circulatory diseases; (D) respiratory diseases; (E) digestive diseases; (F) genitourinary diseases. The vertical scale represents the relative change from the mean effect of PM2.5 on daily hospital admissions, with solid lines indicating point estimates and dashed lines indicating 95% CIs. Major disease categories are based on the chapter division of the ICD-10 diagnostic coding system. Single-day exposure on the same day (lag 0) was used as the exposure metric of PM2.5. The national average exposure–response curves were developed using multivariate meta-analysis approaches. Results for all disease categories and for O3 are shown in S8 and S9 Figs.

https://doi.org/10.1371/journal.pmed.1003188.g003

The results of the sensitivity analyses are shown in S10 and S11 Figs. Our results remained stable using alternative df for the smoothing functions of temperature and relative humidity. The results for PM2.5 remained stable using alternative df for the smoothing function of calendar day. The results for O3 were substantially biased when using too few df for the smoothing function of calendar day (e.g., the association between O3 and hospital admissions for diseases of the respiratory system even became negative when df = 6 per year), indicating that the confounding effects of seasonality and/or long-term patterns were not controlled adequately in that case; however, the bias disappeared with df ≥ 8 per year, and the results remained stable with increasing df up to 14 per year, demonstrating that the models used in the main analyses had successfully removed the confounding effects. None of the 8 negative control outcomes were associated with PM2.5 or O3 in single- or 2-pollutant models, which provided further reassurance that our findings were not substantially confounded or biased (S20 Table).

Table 4 provides the estimated annual reduction in cause-specific hospital admissions and hospitalization expenses attributable to a 10-μg/m3 reduction in daily PM2.5 level in China. Based on hospital admissions and costs in 2016, a 10-μg/m3 reduction in PM2.5 would reduce the number of hospital admissions by 120,145 (95% CI 107,698 to 132,592) and hospitalization expenses by 1,011.9 (95% CI 903.4 to 1,120.4) million yuan nationwide, of which 52.4% of reduced cases and 50.1% of reduced costs would be attributable to reductions in diseases of the circulatory and respiratory systems, with reductions in non-cardiorespiratory diseases accounting for the other 47.6% of reduced cases and 49.9% of reduced costs.

thumbnail
Table 4. Annual reduction in hospital admissions and hospitalization expenses attributable to a 10-μg/m3 reduction in daily PM2.5 level in China.

https://doi.org/10.1371/journal.pmed.1003188.t004

Discussion

This national study systematically assessed the acute effects of main air pollutants on hospital admissions for 14 major and 188 minor disease categories using standardized analytic approaches, providing a unified perspective on the health effects of air pollution on multiple organ systems, with minimal publication bias. We found robust positive associations between PM2.5 and 7 major disease categories and between O3 and respiratory diseases, as well as significant evidence that diseases in dozens of minor disease categories—many are non-cardiorespiratory diseases—can be triggered and/or exacerbated by short-term exposure to air pollution. Consistent with prior research [4,20,31,36,37], we generally observed higher risks related to PM2.5 in older people (65–74 or ≥75 years) than in younger people (<65 years), and steeper slopes of exposure–response curves between PM2.5 and hospital admissions at lower concentrations than at higher concentrations.

Our results regarding cardiorespiratory diseases are in line with the current understanding of the deleterious impacts of air pollution on the circulatory and respiratory systems. The well-known outcomes associated with air pollutants serve as positive controls for this study. The size of effects in this analysis is generally similar to those of recent multisite studies in China [20,31,36,37], and smaller than those from North America and Europe [1,2,4]. The lower effect estimates of this study than those from high-income countries may have at least 2 possible explanations. First, the levels of particulate air pollution were much higher in China than in the high-income countries. As observed in our study and previous multisite studies [4,20,31,37], the slopes of exposure–response relationships were generally steeper at lower concentrations than at higher concentrations, which may be partly attributable to larger exposure measurement errors on heavily polluted days, when public health interventions would be reinforced and people would more actively take self-protection measures. Second, the chemical profile of PM2.5 was different between China and the high-income countries. PM2.5 in China’s air had a higher proportion of crustal materials, which may have lower toxicity than components that originate from fossil fuel combustion [31,37]. We demonstrated that ambient exposure to PM2.5 was robustly associated with hospital admissions for endocrine, nutritional, and metabolic diseases and diseases of the nervous system in the Chinese population. We also identified diabetes mellitus without/with complications and Parkinson disease as typical minor disease categories that were significantly associated with same-day PM2.5 in both single- and 2-pollutant models. These results support the growing evidence that acute exposure to air pollution can increase the risk of certain diseases of these organ systems [4547].

More importantly, by inspecting the associations of air pollution with a broad spectrum of diseases, this study provides new insights into the health effects of air pollution that were scarcely or never reported before. We found a robust positive association between PM2.5 and hospitalizations for diseases of the digestive system. Previous studies have suggested that air pollution exposure is associated with the risk of some gastrointestinal diseases, e.g., gastroenteric disorders in young children [48], acute appendicitis [49], inflammatory bowel diseases [50], and peptic ulcer bleeding [10]. For example, Tian et al. [10] reported a positive association of daily air pollution and emergency admissions for peptic ulcer bleeding using the data of 8,566 recorded cases in Hong Kong’s elderly population during 2005–2010, which was one of the largest studies to date on this topic. Exposure to PM2.5 was found to induce a nonalcoholic steatohepatitis-like phenotype and hepatic fibrosis in animal models [51,52]; evidence from human data is limited. In this analysis, we observed consistent associations between same-day PM2.5 and various digestive diseases, including intestinal infection, esophageal disorders, gastritis and duodenitis, appendiceal conditions, liver diseases (n = 1,398,024), gastrointestinal hemorrhage (n = 811,588), and noninfectious gastroenteritis. Our findings strengthen the hypothesis that air pollution exposure adversely affects certain diseases of the digestive system, which should be fully considered in policy making aimed at protecting public health from air pollution.

We also found a robust positive association between PM2.5 and hospitalizations for diseases of the genitourinary system. A Chinese biopsy series study found that long-term exposure to PM2.5 was associated with the odds for membranous nephropathy [53]. Recent cohort studies in US veterans [12,13] and Taiwanese residents [14] reported that elevated levels of PM2.5 and other pollutants were associated with increased risk of incident chronic kidney disease, chronic kidney disease progression, and end-stage renal disease. In this analysis, we observed consistent associations between same-day PM2.5 and several genitourinary diseases, including nephritis, nephrosis, and renal sclerosis; chronic renal failure; and calculus of urinary tract. Our findings indicate that even being exposed to air pollution for a short time can have adverse effects on one’s genitourinary system, highlighting the importance of prompt public health intervention, especially for susceptible populations.

In addition to the above, we found a robust positive association between PM2.5 and hospitalizations for diseases of the musculoskeletal system and connective tissue. Also noteworthy is that higher PM2.5 exposure was significantly associated with increased risk of hospitalization for crystal arthropathies, anemia, chronic ulcer of skin, and spondylosis, intervertebral disc disorders, and other back problems in both single- and 2-pollutant models, as well as anxiety, somatoform, dissociative, and personality disorders and conditions associated with dizziness or vertigo in single-pollutant models. Chronic ulcer of skin, which had a relatively small sample size (n = 61,605) among all minor disease categories, had one of the largest effect estimates of PM2.5 among all minor disease categories. Overall, these findings are novel—some associations are supported by previous studies, e.g., PM2.5 and anxiety disorders [5456], while others are reported to our knowledge for the first time. Certain musculoskeletal diseases and mental disorders that were associated with PM2.5 in this study contribute to the global burden of disease especially in terms of nonfatal consequences. For example, low back pain has prevailed as the top cause of years lived with disability worldwide for nearly 3 decades [57,58]. Our findings indicate that air pollution may be a novel risk factor for the diseases that were associated with particulate matter in this study, and has the potential to account for the unattributed disease burden [59].

We estimated a substantial annual reduction in the number of hospital admissions and hospitalization expenses that would be attributable to an improvement in daily PM2.5 in China, indicating that the government should develop effective mitigation policies to reduce the burden of disease attributable to air pollution. According to the cause-specific estimates, 47.6% of reduced hospital admissions and 49.9% of reduced hospitalization expenses would be attributable to non-cardiorespiratory diseases. In contrast, a time-series study using mortality data in 272 Chinese cities reported that only 15.2% of premature deaths prevented by PM2.5 reduction would be attributable to non-cardiorespiratory diseases [31]. It should be noted that our estimates are based on short-term associations between air pollution and daily hospital admissions, and thus are not comparable to estimates based on long-term associations, e.g., in the Global Burden of Disease Study [5]; however, our estimates can be compared with those from studies of the same design [31]. Therefore, the precise estimation of the morbidity burden attributable to air pollution in China and worldwide deserves further research.

The biological mechanisms underlying the associations between air pollution and cardiorespiratory diseases are relatively well understood. Epidemiological and toxicological studies have demonstrated that exposure to particulate air pollution may induce systemic inflammation and oxidative stress, cause autonomic imbalance favoring sympathetic tone, and activate the hypothalamic–pituitary–adrenal axis, leading to vasoconstriction, hypertension, tachycardia, reduced heart rate variability, increased plasma viscosity, dyslipidemia, vascular endothelial dysfunction, atherosclerosis progression, platelet aggregation, hypercoagulability, and increased thrombogenesis [6,7,15]. One or more of these mechanistic pathways may be involved in the extrapulmonary effects of air pollution [60,61]. Moreover, there are a series of hypotheses to explain the associations between air pollution and specific non-cardiorespiratory diseases, some of which have been supported by emerging clinical and experimental evidence. Short- and long-term exposures to PM2.5 and other air pollutants have been suggested to adversely affect glucose and insulin homeostasis, leading to glucose intolerance, decreased insulin sensitivity, increased serum lipid levels, and, finally, a higher risk of diabetes mellitus as well as other metabolic diseases [16,45,62]. Air pollution exposure has been reported to be associated with several key factors in the pathophysiology of central nervous system diseases, including neuro-inflammation and oxidative stress, microglial activation, dopaminergic neuron damage, and cerebrovascular impairment [46,47]. These pathways may also play an important role in the pathogenesis of mental and behavioral disorders [5456]. Air pollutants can be transported to the gastrointestinal tract by either mucociliary clearance from the lungs or ingestion of polluted food and water, and may exert direct toxic effects on gastrointestinal epithelial cells and influence gut microbial composition, leading to increased intestinal permeability, altered intestinal immunity, and oxidative stress and inflammatory response [10,63,64]. The skin, the largest and outermost organ of the body, may be adversely affected by pollutants via direct exposure from the air or indirect exposure from the systemic circulation [65]. The kidney, a highly vascularized organ, is vulnerable to both macro- and microvascular dysfunction and thus is likely to be affected through the pathways of air-pollution-induced cardiovascular toxicity [12,66]. Finally, we postulate that air pollution exposure may increase the pain of various conditions (e.g., renal colic caused by kidney stones and chronic back pain) through prostaglandin pathways. Prostaglandins are major proinflammatory mediators that can sensitize pain receptors [67,68]. Increased levels of prostaglandins cause more pain. Recent animal and human studies have linked excessive prostaglandins with short- and long-term exposures to PM2.5 and other air pollutants [6770]. Nevertheless, the exact biological mechanisms underlying the reported associations remain elusive and deserve further research.

This study has some strengths. First, by analyzing all causes of hospitalization rather than a few prespecified outcomes, and reporting all results, this study provides evidence for health effects of air pollution that may have been neglected due to lack of prior knowledge. Second, this study defined disease categories at 2 levels. The major disease categories were used to evaluate the overall health effects of air pollution on each organ system, while the minor disease categories were used to identify the specific diseases that can be triggered and/or exacerbated by air pollution exposure. Third, this study has the advantages of a large dataset with good internal consistency in data collection, good timeliness (2013–2017), and extensive geographical and population coverage (252 cities in China; all ages), ensuring the generalizability of our results to the Chinese population or other populations in similar settings. Fourth, this study also provides the exposure–response relationships between air pollution and cause-specific hospital admissions, which support the hypothesis that there is no safety threshold for the health effects of PM2.5 exposure, but there may be a safety threshold close to the background concentration of O3 for the effect of O3 exposure on respiratory diseases. The possible existence of a threshold for O3 has been suggested by several studies conducted in high-income countries [4,43], and should be further investigated in low- and middle-income countries, where the evidence is still limited.

This study has some limitations. First, we used monitored ambient concentrations of air pollutants as the proxy for personal exposures, leading to exposure measurement errors that are expected to bias the effect estimates toward null [71]. Second, we cannot rule out the possibility of misclassifications resulting from diagnostic or coding errors in this large-scale nationwide dataset. Generally, the more specific the diagnostic categories that are used, the greater the likelihood of outcome misclassification. The diagnosis coding accuracy for the minor disease categories may thus be lower than that for the major disease categories. Therefore, our results concerning minor disease categories should be interpreted with caution. Future studies based on medically reviewed outcome data are warranted to confirm our findings. Third, our data cannot distinguish between scheduled and unscheduled hospitalizations. Including scheduled hospital admissions (e.g., planned surgeries), which are inherently unrelated to air pollution, is expected to reduce the estimation precision. However, a general-practitioner-based referral system is not available in China [19,20]. Hospital visits and admissions are generally unscheduled and are on a first-come, first-served basis. Therefore, the impact of scheduled hospitalizations on our results is expected to be minor. Fourth, we did not have data on the chemical composition or emission sources of particulate matter, hindering us from further research. Fifth, we cannot estimate the long-term risk related to air pollution under the time-series analysis design. Studies of other designs (e.g., cohort studies) should be conducted to entirely understand the hazardous effects of air pollution, as well as replications of this work in other populations to confirm the external validity of our results.

Conclusions

This study provides a comprehensive picture of the associations between short-term exposure to main air pollutants and cause-specific risk of hospital admission in China over a wide spectrum of human diseases. Our analysis showed that air pollution exposure was associated with increased risk of hospitalization for diseases of multiple organ systems, including certain diseases of the digestive, musculoskeletal, and genitourinary systems; many of these associations are important but still not fully recognized. This study also evaluated potential effect modifiers and exposure–response relationships between air pollution and hospitalization. These results can inform policy making aimed at protecting public health from air pollution in China.

Supporting information

S1 Fig. Locations of the 252 Chinese cities in this study.

The base map was created using the open-source R package maps.

https://doi.org/10.1371/journal.pmed.1003188.s001

(TIF)

S2 Fig. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 by major disease category using different lag structures, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s002

(DOCX)

S3 Fig. Percent change in hospital admissions per 10-μg/m3 increase in O3 by major disease category using different lag structures, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s003

(DOCX)

S4 Fig. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 by major disease category using 3-pollutant models, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s004

(DOCX)

S5 Fig. Percent change in hospital admissions per 10-μg/m3 increase in O3 by major disease category using 3-pollutant models, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s005

(DOCX)

S6 Fig. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 by major disease category for study subgroups, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s006

(DOCX)

S7 Fig. Percent change in hospital admissions per 10-μg/m3 increase in O3 by major disease category for study subgroups, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s007

(DOCX)

S8 Fig. Exposure–response curves of the effects of PM2.5 on hospital admissions by major disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s008

(DOCX)

S9 Fig. Exposure–response curves of the effects of O3 on hospital admissions by major disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s009

(DOCX)

S10 Fig. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 by major disease category using alternative model specifications, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s010

(DOCX)

S11 Fig. Percent change in hospital admissions per 10-μg/m3 increase in O3 by major disease category using alternative model specifications, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s011

(DOCX)

S12 Fig. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 by major disease category in 6 regions of China, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s012

(DOCX)

S13 Fig. Percent change in hospital admissions per 10-μg/m3 increase in O3 by major disease category in 6 regions of China, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s013

(DOCX)

S1 Table. City-specific population and coverage rates of class 3 hospitals and hospital beds by the Hospital Quality Monitoring System (HQMS) in 252 Chinese cities.

https://doi.org/10.1371/journal.pmed.1003188.s015

(DOCX)

S2 Table. Annual number of hospital admissions and average cost per hospitalization in China.

https://doi.org/10.1371/journal.pmed.1003188.s016

(DOCX)

S3 Table. Demographic characteristics of cause-specific hospital admissions in 252 Chinese cities, 2013–2017.

https://doi.org/10.1371/journal.pmed.1003188.s017

(DOCX)

S4 Table. Summary statistics of annual-average levels of ambient air pollutants and weather conditions in 252 Chinese cities, 2013–2017.

https://doi.org/10.1371/journal.pmed.1003188.s018

(DOCX)

S5 Table. Pearson correlation coefficients among daily levels of ambient air pollutants and weather conditions, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s019

(DOCX)

S6 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 0 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s020

(DOCX)

S7 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 1 day) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s021

(DOCX)

S8 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 2 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s022

(DOCX)

S9 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 3 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s023

(DOCX)

S10 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 0–1 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s024

(DOCX)

S11 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 0–2 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s025

(DOCX)

S12 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 (lag 0–3 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s026

(DOCX)

S13 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 0 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s027

(DOCX)

S14 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 1 day) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s028

(DOCX)

S15 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 2 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s029

(DOCX)

S16 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 3 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s030

(DOCX)

S17 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 0–1 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s031

(DOCX)

S18 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 0–2 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s032

(DOCX)

S19 Table. Percent change in hospital admissions per 10-μg/m3 increase in O3 (lag 0–3 days) by minor disease category, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s033

(DOCX)

S20 Table. Percent change in hospital admissions per 10-μg/m3 increase in PM2.5 and O3 for 8 negative control outcomes, on average across all cities.

https://doi.org/10.1371/journal.pmed.1003188.s034

(DOCX)

References

  1. 1. Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, et al. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA. 2006;295(10):1127–34. pmid:16522832
  2. 2. Lanzinger S, Schneider A, Breitner S, Stafoggia M, Erzen I, Dostal M, et al. Ultrafine and fine particles and hospital admissions in central Europe. Results from the UFIREG study. Am J Respir Crit Care Med. 2016;194(10):1233–41. pmid:27224452
  3. 3. Di Q, Wang Y, Zanobetti A, Wang Y, Koutrakis P, Choirat C, et al. Air pollution and mortality in the Medicare population. N Engl J Med. 2017;376(26):2513–22. pmid:28657878
  4. 4. Di Q, Dai L, Wang Y, Zanobetti A, Choirat C, Schwartz JD, et al. Association of short-term exposure to air pollution with mortality in older adults. JAMA. 2017;318(24):2446–56. pmid:29279932
  5. 5. Cohen AJ, Brauer M, Burnett R, Anderson HR, Frostad J, Estep K, et al. Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study 2015. Lancet. 2017;389(10082):1907–18. pmid:28408086
  6. 6. Rajagopalan S, Al-Kindi SG, Brook RD. Air pollution and cardiovascular disease: JACC state-of-the-art review. J Am Coll Cardiol. 2018;72(17):2054–70. pmid:30336830
  7. 7. North CM, Rice MB, Ferkol T, Gozal D, Hui C, Jung SH, et al. Air pollution in the Asia-Pacific region. A joint Asian Pacific Society of Respirology/American Thoracic Society perspective. Am J Respir Crit Care Med. 2019;199(6):693–700. pmid:30508489
  8. 8. Bowe B, Xie Y, Li T, Yan Y, Xian H, Al-Aly Z. The 2016 global and national burden of diabetes mellitus attributable to PM2.5 air pollution. Lancet Planet Health. 2018;2(7):e301–12. pmid:30074893
  9. 9. Landrigan PJ, Fuller R, Acosta NJR, Adeyi O, Arnold R, Basu NN, et al. The Lancet Commission on pollution and health. Lancet. 2018;391(10119):462–512. pmid:29056410
  10. 10. Tian L, Qiu H, Sun S, Tsang H, Chan KP, Leung WK. Association between emergency admission for peptic ulcer bleeding and air pollution: a case-crossover analysis in Hong Kong’s elderly population. Lancet Planet Health. 2017;1(2):e74–81. pmid:29851584
  11. 11. Prada D, Zhong J, Colicino E, Zanobetti A, Schwartz J, Dagincourt N, et al. Association of air particulate pollution with bone loss over time and bone fracture risk: analysis of data from two independent studies. Lancet Planet Health. 2017;1(8):e337–47. pmid:29527596
  12. 12. Bowe B, Xie Y, Li T, Yan Y, Xian H, Al-Aly Z. Particulate matter air pollution and the risk of incident CKD and progression to ESRD. J Am Soc Nephrol. 2018;29(1):218–30. pmid:28935655
  13. 13. Bowe B, Xie Y, Li T, Yan Y, Xian H, Al-Aly Z. Associations of ambient coarse particulate matter, nitrogen dioxide, and carbon monoxide with the risk of kidney disease: a cohort study. Lancet Planet Health. 2017;1(7):e267–76. pmid:29851625
  14. 14. Chan TC, Zhang Z, Lin BC, Lin C, Deng HB, Chuang YC, et al. Long-term exposure to ambient fine particulate matter and chronic kidney disease: a cohort study. Environ Health Perspect. 2018;126(10):107002. pmid:30392394
  15. 15. Chin MT. Basic mechanisms for adverse cardiovascular events associated with air pollution. Heart. 2015;101(4):253–6. pmid:25552258
  16. 16. Li H, Cai J, Chen R, Zhao Z, Ying Z, Wang L, et al. Particulate matter exposure and stress hormone levels: a randomized, double-blind, crossover trial of air purification. circulation. 2017;136(7):618–27. pmid:28808144
  17. 17. Miller MR, Raftis JB, Langrish JP, McLean SG, Samutrtai P, Connell SP, et al. Inhaled nanoparticles accumulate at sites of vascular disease. ACS Nano. 2017;11(5):4542–52. pmid:28443337
  18. 18. Maher BA, Ahmed IA, Karloukovski V, MacLaren DA, Foulds PG, Allsop D, et al. Magnetite pollution nanoparticles in the human brain. Proc Natl Acad Sci U S A. 2016;113(39):10797–801. pmid:27601646
  19. 19. Tian Y, Xiang X, Juan J, Sun K, Song J, Cao Y, et al. Fine particulate air pollution and hospital visits for asthma in Beijing, China. Environ Pollut. 2017;230:227–33. pmid:28654880
  20. 20. Tian Y, Liu H, Liang T, Xiang X, Li M, Juan J, et al. Fine particulate air pollution and adult hospital admissions in 200 Chinese cities: a time-series analysis. Int J Epidemiol. 2019;48(4):1142–51. pmid:31157384
  21. 21. Meng Q, Mills A, Wang L, Han Q. What can we learn from China’s health system reform? BMJ. 2019;365:l2349. pmid:31217222
  22. 22. World Health Organization. International statistical classification of diseases and related health problems, 10th revision. Geneva: World Health Organization; 2016 [cited 2020 Mar 15]. https://icd.who.int/browse10/2016/en.
  23. 23. Zhang L, Long J, Jiang W, Shi Y, He X, Zhou Z, et al. Trends in chronic kidney disease in China. N Engl J Med. 2016;375(9):905–6. pmid:27579659
  24. 24. Li J, Cui Z, Long J, Huang W, Wang J, Zhang H, et al. Primary glomerular nephropathy among hospitalized patients in a national database in China. Nephrol Dial Transplant. 2018;33(12):2173–81. pmid:29509919
  25. 25. Li J, Cui Z, Long JY, Huang W, Wang JW, Wang H, et al. The frequency of ANCA-associated vasculitis in a national database of hospitalized patients in China. Arthritis Res Ther. 2018;20(1):226. pmid:30286799
  26. 26. Healthcare Cost and Utilization Project. Clinical Classifications Software (CCS) for ICD-9-CM. Rockville (MD): Healthcare Cost and Utilization Project; 2017 [cited 2020 Mar 15]. https://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp.
  27. 27. Radley DC, Gottlieb DJ, Fisher ES, Tosteson AN. Comorbidity risk-adjustment strategies are comparable among persons with hip fracture. J Clin Epidemiol. 2008;61(6):580–7. pmid:18471662
  28. 28. Bobb JF, Obermeyer Z, Wang Y, Dominici F. Cause-specific risk of hospital admission related to extreme heat in older adults. JAMA. 2014;312(24):2659–67. pmid:25536257
  29. 29. Samet JM, Dominici F, Curriero FC, Coursac I, Zeger SL. Fine particulate air pollution and mortality in 20 U.S. cities, 1987–1994. N Engl J Med. 2000;343(24):1742–9. pmid:11114312
  30. 30. Peng RD, Chang HH, Bell ML, McDermott A, Zeger SL, Samet JM, et al. Coarse particulate matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients. JAMA. 2008;299(18):2172–9. pmid:18477784
  31. 31. Chen R, Yin P, Meng X, Liu C, Wang L, Xu X, et al. Fine particulate air pollution and daily mortality. a nationwide analysis in 272 Chinese cities. Am J Respir Crit Care Med. 2017;196(1):73–81. pmid:28248546
  32. 32. Liu C, Yin P, Chen R, Meng X, Wang L, Niu Y, et al. Ambient carbon monoxide and cardiovascular mortality: a nationwide time-series analysis in 272 cities in China. Lancet Planet Health. 2018;2(1):e12–8. pmid:29615203
  33. 33. Wang L, Liu C, Meng X, Niu Y, Lin Z, Liu Y, et al. Associations between short-term exposure to ambient sulfur dioxide and increased cause-specific mortality in 272 Chinese cities. Environ Int. 2018;117:33–9. pmid:29715611
  34. 34. Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N, et al. Differentiating the effects of fine and coarse particles on daily mortality in Shanghai, China. Environ Int. 2007;33(3):376–84. pmid:17229464
  35. 35. Xie W, Li G, Zhao D, Xie X, Wei Z, Wang W, et al. Relationship between fine particulate air pollution and ischaemic heart disease morbidity and mortality. Heart. 2015;101(4):257–63. pmid:25341536
  36. 36. Tian Y, Liu H, Zhao Z, Xiang X, Li M, Juan J, et al. Association between ambient air pollution and daily hospital admissions for ischemic stroke: a nationwide time-series analysis. PLoS Med. 2018;15(10):e1002668. pmid:30286080
  37. 37. Tian Y, Liu H, Wu Y, Si Y, Song J, Cao Y, 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:l6572. pmid:31888884
  38. 38. Gu J, Shi Y, Chen N, Wang H, Chen T. Ambient fine particulate matter and hospital admissions for ischemic and hemorrhagic strokes and transient ischemic attack in 248 Chinese cities. Sci Total Environ. 2020;715:136896. pmid:32007884
  39. 39. Yin P, He G, Fan M, Chiu KY, Fan M, Liu C, et al. Particulate air pollution and mortality in 38 of China’s largest cities: time series analysis. BMJ. 2017;356:j667. pmid:28292780
  40. 40. Benjamini Y, Hochberg Y. Controlling the false discovery rate—a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 1995;57(1):289–300.
  41. 41. Ebenstein A, Fan M, Greenstone M, He G, Zhou M. New evidence on the impact of sustained exposure to air pollution on life expectancy from China’s Huai River Policy. Proc Natl Acad Sci U S A. 2017;114(39):10384–9. pmid:28893980
  42. 42. Samoli E, Analitis A, Touloumi G, Schwartz J, Anderson HR, Sunyer J, et al. Estimating the exposure-response relationships between particulate matter and mortality within the APHEA multicity project. Environ Health Perspect. 2005;113(1):88–95. pmid:15626653
  43. 43. Bell ML, Peng RD, Dominici F. The exposure-response curve for ozone and risk of mortality and the adequacy of current ozone regulations. Environ Health Perspect. 2006;114(4):532–6. pmid:16581541
  44. 44. Lipsitch M, Tchetgen Tchetgen E, Cohen T. Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology. 2010;21(3):383–8. pmid:20335814
  45. 45. Zanobetti A, Dominici F, Wang Y, Schwartz JD. A national case-crossover analysis of the short-term effect of PM2.5 on hospitalizations and mortality in subjects with diabetes and neurological disorders. Environ Health. 2014;13(1):38. pmid:24886318
  46. 46. Lee H, Myung W, Kim DK, Kim SE, Kim CT, Kim H. Short-term air pollution exposure aggravates Parkinson’s disease in a population-based cohort. Sci Rep. 2017;7:44741. pmid:28300224
  47. 47. Kasdagli MI, Katsouyanni K, Dimakopoulou K, Samoli E. Air pollution and Parkinson’s disease: a systematic review and meta-analysis up to 2018. Int J Hyg Environ Health. 2019;222(3):402–9. pmid:30606679
  48. 48. Orazzo F, Nespoli L, Ito K, Tassinari D, Giardina D, Funis M, et al. Air pollution, aeroallergens, and emergency room visits for acute respiratory diseases and gastroenteric disorders among young children in six Italian cities. Environ Health Perspect. 2009;117(11):1780–5. pmid:20049132
  49. 49. Kaplan GG, Tanyingoh D, Dixon E, Johnson M, Wheeler AJ, Myers RP, et al. Ambient ozone concentrations and the risk of perforated and nonperforated appendicitis: a multicity case-crossover study. Environ Health Perspect. 2013;121(8):939–43. pmid:23842601
  50. 50. Ananthakrishnan AN, Bernstein CN, Iliopoulos D, Macpherson A, Neurath MF, Ali RAR, et al. Environmental triggers in IBD: a review of progress and evidence. Nat Rev Gastroenterol Hepatol. 2018;15(1):39–49. pmid:29018271
  51. 51. Zheng Z, Xu X, Zhang X, Wang A, Zhang C, Huttemann M, et al. Exposure to ambient particulate matter induces a NASH-like phenotype and impairs hepatic glucose metabolism in an animal model. J Hepatol. 2013;58(1):148–54. pmid:22902548
  52. 52. Zheng Z, Zhang X, Wang J, Dandekar A, Kim H, Qiu Y, et al. Exposure to fine airborne particulate matters induces hepatic fibrosis in murine models. J Hepatol. 2015;63(6):1397–404. pmid:26220751
  53. 53. Xu X, Wang G, Chen N, Lu T, Nie S, Xu G, et al. Long-term exposure to air pollution and increased risk of membranous nephropathy in China. J Am Soc Nephrol. 2016;27(12):3739–46. pmid:27365535
  54. 54. Power MC, Kioumourtzoglou MA, Hart JE, Okereke OI, Laden F, Weisskopf MG. The relation between past exposure to fine particulate air pollution and prevalent anxiety: observational cohort study. BMJ. 2015;350:h1111. pmid:25810495
  55. 55. Pun VC, Manjourides J, Suh H. Association of ambient air pollution with depressive and anxiety symptoms in older adults: results from the NSHAP study. Environ Health Perspect. 2017;125(3):342–8. pmid:27517877
  56. 56. Braithwaite I, Zhang S, Kirkbride JB, Osborn DPJ, Hayes JF. Air pollution (particulate matter) exposure and associations with depression, anxiety, bipolar, psychosis and suicide risk: a systematic review and meta-analysis. Environ Health Perspect. 2019;127(12):126002. pmid:31850801
  57. 57. GBD 2017 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392(10159):1789–858. pmid:30496104
  58. 58. Yang G, Wang Y, Zeng Y, Gao GF, Liang X, Zhou M, et al. Rapid health transition in China, 1990–2010: findings from the Global Burden of Disease Study 2010. Lancet. 2013;381(9882):1987–2015. pmid:23746901
  59. 59. GBD 2017 Risk Factor Collaborators. Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392(10159):1923–94. pmid:30496105
  60. 60. Schraufnagel DE, Balmes JR, Cowl CT, De Matteis S, Jung SH, Mortimer K, et al. Air pollution and noncommunicable diseases: a review by the Forum of International Respiratory Societies’ Environmental Committee, part 1: the damaging effects of air pollution. Chest. 2019;155(2):409–16. pmid:30419235
  61. 61. Schraufnagel DE, Balmes JR, Cowl CT, De Matteis S, Jung SH, Mortimer K, et al. Air pollution and noncommunicable diseases: a review by the Forum of International Respiratory Societies’ Environmental Committee, part 2: air pollution and organ systems. Chest. 2019;155(2):417–26. pmid:30419237
  62. 62. Chen Z, Salam MT, Toledo-Corral C, Watanabe RM, Xiang AH, Buchanan TA, et al. Ambient air pollutants have adverse effects on insulin and glucose homeostasis in Mexican Americans. Diabetes Care. 2016;39(4):547–54. pmid:26868440
  63. 63. Mutlu EA, Engen PA, Soberanes S, Urich D, Forsyth CB, Nigdelioglu R, et al. Particulate matter air pollution causes oxidant-mediated increase in gut permeability in mice. Part Fibre Toxicol. 2011;8:19. pmid:21658250
  64. 64. Kish L, Hotte N, Kaplan GG, Vincent R, Tso R, Ganzle M, et al. Environmental particulate matter induces murine intestinal inflammatory responses and alters the gut microbiome. PLoS ONE. 2013;8(4):e62220. pmid:23638009
  65. 65. Wang C, Zhu G, Zhang L, Chen K. Particulate matter pollution and hospital outpatient visits for endocrine, digestive, urological, and dermatological diseases in Nanjing, China. Environ Pollut. 2020;261:114205. pmid:32113107
  66. 66. Yang YR, Chen YM, Chen SY, Chan CC. Associations between long-term particulate matter exposure and adult renal function in the Taipei metropolis. Environ Health Perspect. 2017;125(4):602–7. pmid:27713105
  67. 67. Li W, Wilker EH, Dorans KS, Rice MB, Schwartz J, Coull BA, et al. Short-term exposure to air pollution and biomarkers of oxidative stress: the Framingham Heart Study. J Am Heart Assoc. 2016;5(5):e002742. pmid:27126478
  68. 68. Yan Q, Liew Z, Uppal K, Cui X, Ling C, Heck JE, et al. Maternal serum metabolome and traffic-related air pollution exposure in pregnancy. Environ Int. 2019;130:104872. pmid:31228787
  69. 69. Huang W, Wang G, Lu SE, Kipen H, Wang Y, Hu M, et al. Inflammatory and oxidative stress responses of healthy young adults to changes in air quality during the Beijing Olympics. Am J Respir Crit Care Med. 2012;186(11):1150–9. pmid:22936356
  70. 70. Havet A, Zerimech F, Sanchez M, Siroux V, Le Moual N, Brunekreef B, et al. Outdoor air pollution, exhaled 8-isoprostane and current asthma in adults: the EGEA study. Eur Respir J. 2018;51(4):1702036. pmid:29618600
  71. 71. Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, et al. Exposure measurement error in time-series studies of air pollution: concepts and consequences. Environ Health Perspect. 2000;108(5):419–26. pmid:10811568