Effects of ambient particulate matter on fasting blood glucose among primary school children in Guangzhou, China

Environmental Research 176 (2019) 108541

Contents lists available at ScienceDirect

Environmental Research journal homepage: www.elsevier.com/locate/envres

Effects of ambient particulate matter on fasting blood glucose among primary school children in Guangzhou, China

T

Li Caia,1, Suhan Wangb,c,1, Peng Gaod, Xiaoting Shene, Bin Jalaludinf, Michael S. Bloomg, Qiong Wangb,c, Junzhe Baob,c, Xia Zenga, Zhaohuan Guia, Yajun Chena,∗∗, Cunrui Huangb,c,∗ a

Department of Maternal and Child Health, School of Public Health, Sun Yat-sen University, China Department of Health Policy and Management, School of Public Health, Sun Yat-sen University, China c Laboratory of Meteorology and Health, Shanghai Meteorological Service, China d Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, School of Atmospheric Sciences, Sun Yat-sen University, China e Center for Reproductive Medicine, The First Affiliated Hospital of Sun Yat-sen University, China f Population Health, South Western Sydney Local Health District, Liverpool, NSW, 2170, Australia g Department of Environmental Health Sciences and Epidemiology and Biostatistics, University at Albany, State University of New York, Rensselaer, NY, 12144, USA b

A R T I C LE I N FO

A B S T R A C T

Keywords: Air pollution Blood glucose Diet Children Effect modification

Background: Exposure to ambient particulate matter (PM) has been linked with diabetes and elevated blood glucose in adults. However, there are few reports on the effects of PM on fasting blood glucose (FBG) among children. Objectives: The study aimed to assess the associations between medium-term exposure of ambient particles with diameters ≤2.5 μm (PM2.5), and ≤10 μm (PM10) and FBG in a general population of children, and also to explore the modifying effects of diet. Methods: In this cross-sectional study, we enrolled 4234 children (aged 6–13 years) residing in Guangzhou, China, in 2017. Individual PM2.5 and PM10 exposures during the 186-day period before each physical examination were retrospectively estimated by an inverse distance weighting interpolation and time-weighted approach according to their home address, school address, and activity patterns. Linear mixed effect models were used to examine the relationships between PM2.5 and PM10 with FBG after adjusting for covariates. Results: We found that per 10 μg/m3 increase in PM2.5 and PM10 levels during the 186-day period were associated with 2.3% (95% CI: 1.0%, 3.8%) higher FBG and 0.9% (95% CI: 0.5%, 1.4%) higher FBG, respectively. Stronger effect estimates were observed among subgroups of children with a family history of diabetes, and higher intake of sugar-sweetened beverages (SSBs). Also, we found significant interactions between PM2.5 concentration and family history of diabetes and SSBs intake on FBG. Conclusions: Medium-term exposure to ambient PM2.5 and PM10 were associated with higher FBG levels in children, and that higher SSBs intake might modify these associations.

1. Introduction The adverse health effects of particulate matter (PM) exposure are a major global public health threat to public health. Both short- and longterm PM exposures can increase hospital admissions of respiratory and cardiovascular diseases, and increase mortality from these diseases (GDB, 2017; WHO, 2006). Emerging epidemiological evidence indicates that PM exposure also increases the risk of type 2 diabetes

mellitus (T2DM) in the general population (Balti et al., 2014; Yang et al., 2018a; Rao et al., 2015a; Eze et al., 2015; Thiering et al., 2013; Bowe et al., 2018). Although the mechanisms underlying the reported associations between PM exposure and blood glucose levels and T2DM remain unclear, one possible explanation is that PM inhalation initiates respiratory system oxidative stress and inflammatory responses, leading to systemic inflammation, and eventually insulin resistance. These intermediate factors may affect metabolism and lead to the early onset of

Abbreviations: FBG, fasting blood glucose; PM, particulate matter; SSBs, sugar-sweetened beverages ∗ Corresponding author. Department of Health Policy and Management, School of Public Health, Sun Yat-sen University, China. ∗∗ Corresponding author. E-mail addresses: [email protected] (Y. Chen), [email protected] (C. Huang). 1 Li Cai and Suhan Wang contributed equally to this work. https://doi.org/10.1016/j.envres.2019.108541 Received 2 March 2019; Received in revised form 6 June 2019; Accepted 17 June 2019 Available online 18 June 2019 0013-9351/ © 2019 Elsevier Inc. All rights reserved.

Environmental Research 176 (2019) 108541

L. Cai, et al.

Fig. 1. Sampling procedure for the study participants.

glucose levels. To address the pending evidence gap, we conducted a cross-sectional study of 4234 children aged 6–13 years and residing in Guangzhou, China. We explored the associations of PM exposure with fasting blood glucose (FBG) and investigate the potential modifying effects of diet.

T2DM (Franklin et al., 2015; Rajagopalan and Brook, 2012). Recent epidemiological studies have also investigated associations between air pollution exposure and blood glucose level, a potential indicator of elevated insulin resistance, in adult populations (Eze et al., 2015). These studies have found positive associations between greater levels of short-term (Peng et al., 2016), medium-term (i.e., weeks to months) (Peng et al., 2016; Lucht et al., 2018; Sade et al., 2015) and long-term (i.e., ≥1 year) (Cai et al., 2017; Chuang et al., 2011; Liu et al., 2016; Wolf et al., 2016) exposure to PM, nitrogen dioxide (NO2), sulfur dioxide (SO2) and ozone (O3), and higher blood glucose levels among adults, especially in aging populations. Other epidemiological studies have reported that intake of fruits and vegetables, dietary antioxidants and B vitamins, or omega-3 fatty acid supplementation can protect against the adverse effects of air pollution in adults (Guxens et al., 2012; Lin et al., 2018; Schulz et al., 2015; Tong et al., 2012; Zhong et al., 2017). Children are particularly vulnerable to adverse effects from PM exposure, since metabolic disorders in childhood can increase the risk of a range of metabolic diseases in adulthood (Koskinen et al., 2017). However, there are very few epidemiological studies and limited evidence on the relationship between air pollution and children's blood glucose level. One study of overweight/obese children found that greater long-term exposure to PM with diameter ≤2.5 μm (PM2.5) was associated with elevated blood glucose (Toledo-Corral et al., 2018). Another study reported a significant correlation between poorer air quality and higher blood glucose in children (Poursafa et al., 2014). These two studies focused on long-term air pollution exposure. However, to date, no published literature has reported the association between medium-term PM exposure and blood glucose levels in children. The association between PM exposure and blood glucose in Chinese children remains unknown. In the last two decades, the prevalence of overweight and obesity in Chinese children has increased rapidly (NCDRisC, 2017; Wang et al., 2017), likely due in part to environmental pollution, poorer diets and the lack of physical activity (Wang et al., 2017). One cohort study in Beijing, China, found that childhood obesity increased the risk for adult metabolic syndrome and diabetes (Liang et al., 2015). Clearly, more needs to be done to identify modifiable risk factors to prevent a diabetes epidemic among Chinese children. Furthermore, no studies to date have described the impact of dietary factors on associations between air pollution exposure and children's blood

2. Methods 2.1. Study area Guangzhou is the third largest city in China and is located within the Pearl River Delta, one of the most developed areas in southern China. Guangzhou has a typical oceanic monsoon climate, and the annual average temperature is between 21.5 °C and 22.2 °C (Bureau, 2018). Although air quality in Guangzhou is better than many northern Chinese cities, levels of air pollutants during 2016–2017 (i.e., annual mean concentration PM2.5: 36 μg/m³) still exceeded those recommended by Chinese National Ambient Air Quality Standards (NAAQS Guideline Values, GB3095-2012, e.g., 35 μg/m3 for PM2.5 and 50 μg/m3 for PM10). The main PM sources in Guangzhou are traffic emission, coal combustion, industrial sources, and non-point agricultural sources (Cui et al., 2015; Guangdong Environmental M, 2018). 2.2. Study design This study was conducted using data from the baseline examination of a prospective population-based cohort study (Registration number: NCT03582709). We used a multistage stratified random cluster sampling method to recruit study participants from March to May 2017. First, we randomly selected five districts from Guangzhou city. Second, we randomly selected one elementary school from each study district. All students of the five schools (n=8324) were invited to participate and completed the questionnaire. Of these students, we obtained written informed consent for blood sample collection from 4991 parents and students. Third, we excluded students with missing FBG data (n = 236; 4.7% of total), or those with missing residential address (n = 528; 10.6%). Finally, we enrolled 4234 individuals who had both FBG and residential address data as study subjects. We retrospectively assessed individual exposures to PM2.5 and PM10 during the 186-day 2

Environmental Research 176 (2019) 108541

L. Cai, et al.

‘average time activity patterns’ into consideration. For example, from Monday to Friday during the school semester, we considered that students stayed at school from 8 a.m. to 4 p.m., and stayed at home at other time of the day. We also assumed that school children stayed at home from Saturday to Sunday during the school semester and during the winter school vacation. To adjust for the confounding effects of meteorological conditions, we obtained data for daily average ambient temperature and humidity in Guangzhou between September 2016 and May 2017 from the Guangzhou Meteorological Bureau. The average temperature and relative humidity were calculated as the mean of daily average values over the corresponding PM exposure period (186 days).

period before blood collection for FBG measurement. A flow chart of the recruitment process is presented in Fig. 1. A standardized questionnaire was administered to all students to obtain data on their demographic characteristics, health status (including previous medical history and current symptoms), and family history. Information on physical activity and dietary intake was collected using a validated questionnaire, which was completed by the participants together with their parents. Anthropometric data, including body weight and height, were acquired by study staff according to a standardized protocol. The study protocol was approved by the Ethics and Human Subject Committee of Sun Yat-Sen Medical College and all study participants and their parents provided informed written consent.

2.5. FBG measurements 2.3. Definition of covariates Blood collection for FBG levels (mg/dL) was undertaken before the physical examination. Participants were asked to fast for at least 10 h before the physical examination (from 8 a.m. to 10 a.m.) during the period from March 2017 to May 2017. To ensure children's compliance with fasting, parents or guardians were notified by head teachers to ensure that their child fasted before the blood sample collection. As a further check, before the blood sampling, the head teachers asked each student whether they had breakfast that morning. We collected blood only from students who confirmed that they did not eat breakfast (we provided a breakfast after blood collection). A 5 mL blood sample was collected from each student and stored at 4 °C prior to analysis. Within half an hour of collection, whole blood samples were centrifuged at 3000 revolution/minute for 15 min to obtain serum samples. Within 2 h of separation, serum samples were analyzed enzymatically by the glucose oxidase method using a Hitachi 7180 automatic biochemical analyzer (Hitachi High Tech Solutions, Co., Tokyo, Japan). All analyses were performed in the Guangzhou Primary and Secondary School Hygiene and Health Promotion Center following a standard operating procedure.

Parental education was defined as the highest level of education completed by either parent (below senior high school/senior high school/junior college/college or above). Breastfeeding includes full and partial breastfeeding during the first 6 months of life. In the present study, 99.2% of mothers were never smokers (n=4198), and only 0.3% of mothers were current smokers (n=14), whilst 35.4% of fathers were current smokers (n=1500). Thus, we combined the smoking status of father and mother as parental smoking status. If both parents were never smokers, then parental smoking status was defined as ‘never smokers’. If either of the parents were former/current smoker, then parental smoking status was defined as ‘former/current smoker’. Children's weight status (underweight, overweight and obese) was defined using the World Health Organization (WHO) criteria of body mass index (BMI) for age chart (http://www.who.int/growthref/tools/ en/, accessed 29 September 2016). Dietary intake included the consumption of fruits, vegetables, fish, grains, milk, soybean products and sugar-sweetened beverages (SSBs; for example, Coca Cola, Pepsi). Children reported the frequency (days) and amount (servings) of dietary intake over the past seven days. Servings were described by using common food containers in our questionnaire in order to obtain quantitative measures. The average daily intake of a single food was estimated as follows: average daily intake = [days × (amount in each of those days)]/7. Dietary intake was classified into low or high levels by the medians of the daily intake. The dietary questionnaire was validated in an earlier pilot study that showed acceptable reliability and validity (data not reported).

2.6. Statistical analyses The distributions of continuous variables were examined using the Shapiro–Wilk normality test. After natural logarithm transformation, FBG followed a normal distribution. Effect estimates were back-transformed from the log scale using 100 × [exp (β) −1] and represented as percent differences with corresponding 95% confidence intervals (CI). We applied natural cubic splines with five knots in mixed effect models to test for non-linear PM effects. We did not find significant nonlinear effects for PM2.5 or PM10. We therefore conducted linear mixed effect models to investigate the main effects of 186-day PM2.5 and PM10 levels on FBG. We used a directed acyclic graph (DAG) to identify potentially confounding variables for inclusion in multiple regression models (Supplementary Fig. S1). Three models with increasing covariate adjustment were conducted. Model 1 was adjusted for age (continuous) and sex, and schools and residential districts were fitted as random effects. In model 2, we further adjusted for 186-day mean temperature (continuous), 186-day mean relative humidity (continuous), family income (categorical), and residuals from a regression model of 7-day average PM concentration. The 7-day average PM concentrations before the blood draw were used to adjust for the potential impact of short-term effects of PM on FBG. Where the mediumterm and short-term PM levels were highly correlated, they were regressed against each other and individual residuals were included in the models (Yang et al., 2018b). In model 3, we further adjusted for BMI (continuous), average outdoor physical activity time (categorical: < 1; 1–1.9; 2–4; > 4 h/day), breast feeding (yes vs. no), parental education, parental smoking status, and family history of diabetes (yes vs. no) to assess the impact. We also ran the main model analysis (model 2) using different exposure time windows before FBG measurement (45, 60, 75, 120, 180, and 186 days) for PM2.5 and PM10.

2.4. Air pollution exposures As there is no published literature reported the relationship between medium-term PM exposure and FBG in children, we retrospectively assessed individual medium-term exposures to ambient PM2.5 and PM10, during the 186-day interval before each FBG examination for each subject by inverse distance weighting interpolation (Wu et al., 2017) and time-weighted approach according to home address, school address and average time activity patterns. We obtained data for PM2.5, PM10, SO2, NO2, and O3 concentrations from 11 air quality monitoring stations maintained by the State Environmental Protection Administration of China in Guangzhou. We assessed air pollution exposure by using the following procedures. First, the residential and school addresses for all students at the date of physical examination and the 11 monitoring stations in Guangzhou were geocoded by Baidu Map Application Programming Interface (http://lbsyun.baidu.com/, accessed 29 September 2016) to obtain their longitude and latitude coordinates. Second, we used inverse distance weight modeling to predict PM2.5, PM10, SO2, NO2, and O3 exposure levels for each residence and school addresses on each single day employing hourly air pollutant concentrations from air quality monitoring data between September 2016 and May 2017 (Wu et al., 2017; Brauer et al., 2008; Kim et al., 2014). Third, we also took 3

Environmental Research 176 (2019) 108541

L. Cai, et al.

We also implemented two-pollutant models where gaseous air pollutants (SO2, NO2 or O3) were added individually to the model with PM to evaluate the conditional effects of two air pollutants on FBG in children. We did not run three pollutant models as SO2, NO2, and O3 were highly correlated with each other (Spearman correlation coefficients > 0.8; see Supplementary Table S1). We further stratified the study participants by sex, age, weight status, family income, family history of diabetes, and dietary intake (fruits, vegetables, fish, grains, milk, soybean products, and SSBs), and investigated the associations between PM2.5 and PM10 level with FBG in each subgroup. Dietary intake was classified into low or high levels by the medians of the daily intakes. Linear mixed effect models (model 2) with adjustment for the above confounders were used to investigate the associations. We also included interaction terms in the linear mixed effect models to test the interactions between PM2.5, and PM10 with sex, age, weight status, family income, family history of diabetes, and diet. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, N.C.). We defined statistical significance for main effects and interactions as P < 0.05 for a two-tailed test.

Table 1 General characteristics of the children in the study.

3. Results 3.1. Demographic characteristics and correlation of PM levels and meteorological variables Demographic characteristics of the students are shown in Table 1. The median (25th, 75th) FBG level was 90.2 (85.3, 95.2) mg/dL. The mean age of all students was 9.1 ± 1.8 years, and 2264 participants (53.5%) were males. The median, 25th, and 75th percentiles for dietary intake among all students is also shown in Table 1. In general, children in our study had higher parental education and family income levels compared to the general population according to the Guangzhou census (data not shown). 3.2. Main analysis: PM and FGB in children The median distance between students’ residential and school addresses and the nearest air quality monitoring stations was 1.4 km, ranging from 0.07 to 24.6 km; more than 95% of students lived within 3.16 km of the nearest air quality monitoring station (Fig. 2). The distribution of 186-day PM levels are shown in Table 2. Daily ambient PM2.5 and PM10 concentrations were 45.04 μg/m3 and 70.88 μg/m3, respectively, which exceeded the Grade I levels of 35 μg/m3 for PM2.5 and 50 μg/m3 for PM10 set forth by the Chinese NAAQS Guideline Values (GB3095-2012). Ambient PM2.5, PM10, temperature, and humidity were weakly to moderately inter-correlated (Table 2). For example, PM2.5 was positively correlated with PM10 (r = 0.56; P < 0.01) and temperature (r = 0.12; P < 0.01). We found similar results when including only those who lived within 3.16 km of the nearest air quality monitoring station (n=4028) (Supplementary Table S2). The time trend for PM2.5 and PM10 in Guangzhou during the study period is shown in Supplementary Fig. S2. In single-pollutant models (model 2), each 10 μg/m3 increase in 186-day mean PM2.5 and PM10 was associated with a 2.3% (95% CI: 1.0%, 3.8%) and a 0.9% (95% CI: 0.5%, 1.4%) higher FBG, respectively (Table 3). Further adjustment for other covariates (model 3) did not change the results meaningfully (Supplementary Table S3). To assess the robustness of the associations, we excluded those who lived > 3.16 km from the nearest air quality monitoring station and found that the associations of 186-day mean PM2.5 and PM10 with FBG were essentially unchanged (Supplementary Table S4). We also observed significant positive associations for 60, 75, 120, and 180-day mean exposures to PM2.5 and PM10. The 45-day mean exposure to PM10, but not PM2.5, was positively related to FBG (Supplementary Fig. S3). In two-pollutant models, similar associations were observed for both PM2.5 and PM10 when NO2 or SO2 were added to the models, while

Variables

All Participants (n=4234)

Age, year (mean ± SD) Males, n (%) BMI, kg/m2 (mean ± SD) Distribution of weight status, n (%) Underweight Normal weight Overweight Obese FBG, mg/dL Diabetes, n (%) Breast feeding, n (%) Family history of diabetes, n (%) Parental education level, n (%) Below senior high school Senior high school Junior college College or above Monthly family income, n (%) < 5000 RMB 5000–7999 RMB 8000–14999 RMB ≥15000RMB Refused to answer Average outdoor physical activity time, n (%) < 1 h/day 1–1.9 h/day 2–4 h/day > 4 h/day Parental smoking status, n (%) Never smokers Former smokers Current smokers Fruit intake (servings a/d) Vegetable intake (servings a/d) Milk intake (cup b/d) Fish intake (servings c/d) Grain intake (servings d/d) Soybean product intake (servings e/d) Sugar-sweetened beverage intake (cup b/d)

9.1 ± 1.8 2264 (53.5) 17.1 ± 3.2 222 (5.2) 2989 (70.6) 609 (14.4) 414 (9.8) 90.2 (85.3, 95.2) 3 (0.07) 3364 (82.2) 1224 (30.2) 266 (6.3) 817 (19.3) 1173 (27.7) 1978 (46.7) 894 (21.1) 1062 (25.1) 1051 (24.8) 546 (12.9) 681 (16.1) 1569 (37.1) 1823 (43.0) 648 (15.3) 194 (4.6) 2290 (54.1) 439 (10.4) 1505 (35.5) 2.0 (1.0, 3.0) 3.0 (1.8, 5.1) 0.7 (0.4, 1.0) 0.7 (0.3, 1.4) 0.4 (0.1, 1.0) 0.4 (0.1, 0.8) 0.1 (0, 0.3)

Note: Data were presented as mean ± SD, median (25th, 75th) or n (%). a A serving of fruits or vegetables is equivalent to 100 g. b A cup is equivalent to 250 mL. c A serving of fish is equivalent to 50 g. d A serving of grains is equivalent to 60 g. e A serving of soybean products (solid) is equivalent to 50 g.

these associations were attenuated when adjusted for O3 (Table 3). Overall, the associations of 186-day mean PM2.5 and PM10 with FBG in single-pollutant models and two-pollutant models were similar, with only minor differences. Therefore, in the following section, we present results from single-pollutant models only.

3.3. Effect modification We stratified the analysis by sex, age, weight status, family income, family history of diabetes, and dietary intake (Table 4). In general, we observed larger associations in subgroups with a family history of diabetes, and those with higher intake of SSBs. For instance, each 10 μg/m3 increase in 186-day mean PM2.5 was associated with 1.7% and 4.1% higher levels of FBG among subgroups with low and high intake of SSBs. We found statistically significant interactions between PM2.5 and family history of diabetes (Pinteraction = 0.017), and SSBs intake (Pinteraction = 0.018). However, there were no significant interaction between PM and sex, age, weight status, family income, and intake of fruit, vegetable, fish, grain, milk, and soybean products.

4

Environmental Research 176 (2019) 108541

L. Cai, et al.

Fig. 2. Spatial distribution of 11 fixed ambient air quality monitoring stations, 5 schools, and residential addresses of 4234 study participants in Guangzhou, China.

4. Discussion

PM2.5 exposure was positively associated with FBG (0.89 mg/dL per 3.12 μg/m3 PM2.5) (Peng et al., 2016). Observations from non-diabetic participants (n=7108) in the Heinz Nixdorf Recall Study also found a positive association between PM and FBG (28-d PM2.5: 0.91 mg/dL per 5.70 μg/m3) (Lucht et al., 2018). Similarly, a recent cohort study of 1023 Mexican American adults at high risk for diabetes (age range: 17.9–65.6 years) demonstrated a significant positive association between 7-day PM2.5 exposure and FBG (0.70 mg/dL per 5 μg/m3 PM2.5) (Chen et al., 2016). Results from another study conducted among 1023 elderly participants in Taiwan found both long-term PM2.5 and PM10 exposure were significantly positively associated with FBG (Chuang et al., 2011). A single study of 429 overweight and obese children in Los Angeles, California found that each 5.2 μg/m3 cumulative 12-month PM2.5 exposure was associated with 1.7% higher FBG (P < 0.001)

In this large cross-sectional sample of Chinese children, we found that a 10 μg/m3 increase in 186-day mean PM2.5 and PM10 level was associated with 2.3% and 0.9% higher levels of FBG, respectively. In addition, stronger associations between PM2.5 exposure with FBG were observed in children with a family history of diabetes or a higher intake of SSBs. To the best of our knowledge, this study is the first to evaluate the associations between ambient PM exposure and FBG in a general population of children. Although limited previous published reports are available for children, and none in Chinese children, several studies have investigated associations between PM and FBG in adults. The Normative Aging Study examined 551 non-diabetic participants and reported that 28-day

Table 2 Summary statistics and Spearman correlations of 186-day mean PM levels and meteorological variables. Exposure

Summary statistics Mean

PM2.5 (μg/m³) PM10 (μg/m³) Temperature (°C) Humidity (%)

45.0 70.9 19.0 77.7

Median 45.5 70.4 18.9 77.7

Spearman correlation coefficients Min 38.3 55.8 18.7 77.5

Max 48.4 83.4 19.5 77.9

Interquartile range 3.2 6.8 0.3 0.2

Note: Spearman correlation coefficients, ∗P < 0.01. 5

PM2.5 1.00

PM10 ∗

0.56 1.00

Temperature ∗

0.12 −0.41∗ 1.00

Humidity 0.14∗ 0.15∗ −0.78∗ 1.00

Environmental Research 176 (2019) 108541

L. Cai, et al.

(Toledo-Corral et al., 2018). The present study found a positive association between 186-day PM2.5 and PM10 exposure and FBG in a general population of children. We also found significant positive associations for 60, 75, 120, and 180-day PM2.5 and PM10 exposure. The PM effects found in our study are similar to those from previous studies, despite differences in the range of PM exposures (i.e., PM levels in Guangzhou tend to be higher than reported for U.S. cities (Toledo-Corral et al., 2018)), exposure duration, study population, and methodology. We specifically focused on children aged 6–13 years in this study as they are potentially more vulnerable to the adverse health effects of PM exposure than adults (Sly and Flack, 2008) and there is little relevant epidemiological data available in the literature for this age-group. The majority of our study participants (99.97%) were nondiabetic children. When we excluded children with diabetes (n=8), the associations of PM2.5 and PM10 exposure with FBG were essentially unchanged (data not shown). Nondiabetic children are an important group for investigating how PM exposure affects physiological pathways that result in abnormal glucose regulation and ultimately leading to T2DM in the general population. Three prior studies that included a high prevalence of diabetic adults in the study population (prevalence above 19.7%) also indicated positive associations between air pollution and FBG (Sade et al., 2015; Ward-Caviness et al., 2015). Though the effects of air pollution on FBG might be small, relative to the effects of medication,

Table 3 Associations between per 10 μg/m³ increment in 186-day mean PM levels and FBG in children (n=4234). Exposure

FBG (mg/dL) % changes (95% confidence interval)

PM2.5 Single-pollutant model Two-pollutant models PM2.5 adjusted for NO2 PM2.5 adjusted for SO2 PM2.5 adjusted for O3 PM10 Single-pollutant model Two-pollutant models PM10 adjusted for NO2 PM10 adjusted for SO2 PM10 adjusted for O3

Model 1

Model 2

−0.9 (−2.0, 0.2)

2.3 (1.0, 3.8)*

−0.8 (−2.0, 0.3) 1.0 (−0.2, 2.2) −0.6 (−1.7, 0.6)

2.5 (0.9, 4.2)* 2.7 (1.2, 4.3)* 1.0 (−1.1, 3.2)

1.7 (1.2, 2.1)*

0.9 (0.5, 1.4)*

0.6 (0.1, 1.1)* 1.2 (0.8, 1.6)* −0.1 (−0.6, 0.5)

1.0 (0.5, 1.6)* 1.0 (0.5, 1.4)* 1.0 (−0.2, 2.1)

Model 1: Linear mixed effect models with adjustment for age and sex. Schools and residential districts were fitted as random effects in models. Model 2: Linear mixed effect models were further adjusted for 186-day mean temperature, 186-day mean relative humidity, family income, and residuals from regression model of 7-day average PM concentration.

Table 4 Associations between per 10 μg/m³ increment in 186-day mean PM levels and FBG among subgroups of children. Subgroups

Sex Males Females Age (year) 6-9 10-13 Weight status Underweight Normal weight Overweight Obese Family income < 8000 RMB/m ≥8000 RMB/m Refused to answer Family history of diabetes Yes No Fruit intake Low High Vegetable intake Low High Fish intake Low High Grain intake Low High Milk intake Low High Soybean products intake Low High SSBs intake Low High

PM2.5

PM10 a

% changes (95%CI)

P

2.6 (0.8, 4.4) 2.2 (0.2, 4.3)

0.005 0.031

2.4 (0.8, 4,0) 2.3 (0.0, 4.8)

0.003 0.055

2.3 (−3.3, 8.3) 1.9 (0.2, 3.6) 3.0 (0.0, 6.2) 3.0 (−0.5, 6.5)

0.431 0.030 0.051 0.092

3.4 (1.4, 5.4) 1.3 (−0.8, 3.5) 2.2 (−1.2, 5.7)

0.001 0.234 0.198

4.8 (2.1, 7.6) 1.4 (−0.1, 3.0)

< 0.001 0.071

3.0 (1.4, 4.7) 1.6 (−0.7, 4.0)

< 0.001 0.183

2.9 (1.3, 4.7) 2.3 (0.1, 4.6)

0.001 0.041

2.6 (0.8, 4.3) 2.3 (0.2, 4.5)

0.003 0.031

2.5 (0.7, 4.3) 2.4 (0.4, 4.5)

0.007 0.021

2.8 (1.0, 4.6) 1.3 (−0.8, 3.4)

0.002 0.222

2.8 (0.7, 4.8) 2.0 (0.2, 3.9)

0.007 0.033

1.7 (0.1, 3.4) 4.1 (1.7, 6.6)

0.039 0.001

Pinter

a

% changes (95%CI)

P

1.0 (0.4, 1.6) 0.8 (0.2, 1.5)

0.001 0.015

1.0 (0.5, 1.5) 0.8 (0.0, 1.6)

< 0.001 0.042

0.7 (−1.3, 2.7) 0.8 (0.2, 1.3) 1.2 (0.2, 2.2) 1.2 (0.0, 2.4)

0.491 0.009 0.023 0.045

1.2 (0.5, 1.9) 0.7 (−0.1, 1.4) 0.7 (−0.4, 1.9)

< 0.001 0.071 0.231

1.6 (0.7, 2.4) 0.7 (0.1, 1.2)

< 0.001 0.012

1.1 (0.6, 1.7) 0.7 (−0.1, 1.5)

< 0.001 0.073

1.1 (0.5, 1.6) 1.0 (0.2, 1.7)

< 0.001 0.013

1.0 (0.4, 1.5) 0.9 (0.2, 1.6)

0.001 0.011

1.0 (0.4, 1.6) 1.0 (0.3, 1.7)

0.001 0.006

0.9 (0.3, 1.5) 0.7 (0.0, 1.4)

0.003 0.048

1.1 (0.4, 1.7) 0.8 (0.2, 1.4)

0.002 0.014

0.7 (0.2, 1.3) 1.5 (0.7, 2.3)

0.010 < 0.001

0.853

Pinter 0.679

0.563

0.907

0.907

0.913

0.496

0.256

0.017

0.109

0.697

0.283

0.472

0.489

0.986

0.879

0.965

0.887

0.091

0.879

0.802

0.620

0.018

0.568

SSBs, sugar-sweetened beverages. a Linear mixed effect models with adjustment for age, sex, 186-day mean temperature, 186-day mean relative humidity, family income, and residuals from regression model of 7-day average PM concentration. Schools and residential districts were fitted as random effects in models. 6

Environmental Research 176 (2019) 108541

L. Cai, et al.

present study, PM exposure exceeded Chinese national health standards. According to a systematic review, even relatively small increases in FBG can accelerate the risk of cardio-metabolic diseases (Zhang et al., 2010), and lead to a greater frequency of cardiovascular events and mortality (Cohen et al., 2009; Schottker et al., 2016). Therefore, these small increases in FBG associated with PM exposure may have widereaching impacts on children's health. Future studies may also need to consider greenness. A recent study showed that time spent in green spaces was negatively associated with FBG levels in children (Dadvand et al., 2018). Higher air pollution exposure is often associated with lower level of greenness and both factors relate to socioeconomic status (Thiering et al., 2016). Furthermore, interventions to reduce intake of SSBs in PM-polluted areas and during PM-polluted periods may help mitigate the risk of cardio-metabolic disease from PM exposure. The limitations of our study need to be noted. Firstly, although we have adjusted for a number of individual-level confounders in the models, we cannot exclude the effects of other unmeasured potential confounding factors, including environmental tobacco smoke, indoor PM exposure, greenspace, traffic noise, and area-level socioeconomic status. Additionally, though we adjusted for average humidity and temperature over the 186 days preceding the FBG, we cannot rule out residual confounding from time varying effects. Secondly, as a crosssectional study, we were limited in our ability to ascertain the temporal relationship between ambient PM and FBG and effect modification by dietary factors. Furthermore, long-term dietary data were not available in our study and so we may have misclassified dietary exposure for some participants. A future investigation that captures long-term dietary data will be necessary to more definitively assess the modifying role of diet in air pollution exposure-FBG associations among children. Thirdly, the duration of residence at the home address was not considered in this study. The association we found between PM exposure and FBG may not be exclusively due to medium-term exposure. Fourthly, we calculated the individual PM exposure levels according to the available information (i.e., home and school addresses and average time activity patterns), which might lead to exposure misclassification. However, any misclassification is likely to have biased the effect estimates toward the null.

diet, and physical activity, our results underscore the importance of reducing PM pollution for the early prevention of diabetes in children. Previous epidemiological studies have reported that diet modifies the adverse health effects of air pollution. One recent study among 29,032 participants from the World Health Organization Study on global AGEing and adult health demonstrated that intake of fruits and vegetables might mitigate the adverse effects of PM2.5 on lung function (Lin et al., 2018). A previous U.S. study also reported that antioxidant dietary intake protects against the adverse effects of PM on blood pressure (Schulz et al., 2015). Additionally, a Spanish study reported stronger negative associations between air pollution exposure and infant mental development scores among those whose mothers had low fruit/vegetable intake during pregnancy (Guxens et al., 2012). Fish is rich in long chain poly-unsaturated fatty acids, such as omega-3 fatty acids. One U.S. study conducted among healthy middle-aged adults reported that omega-3 fatty acid supplementation could be protective against the adverse cardiac and lipid effects associated with air pollution exposure (Tong et al., 2012). Further, a pilot human intervention trial study demonstrated that B vitamins could attenuate the epigenetic effects of PM exposure (Zhong et al., 2017). Milk and grains are excellent sources of B vitamins. In the present study, results from a subgroup analysis suggested an association between PM2.5 and FBG in children with lower intake of fruit or milk, but not in their counterparts. Similar results were found for the association of PM10 by subgroup of fruit intake. However, the interaction terms between PM and diet were not statistically significant. More studies are needed before any firm conclusions can be drawn. Prospective cohort studies have reported that a greater daily intake of SSBs was significantly associated with a greater risk of developing T2DM (Malik et al., 2010). The present study also found that higher SSBs consumption could exacerbate the adverse associations of PM2.5 and PM10 with FBG in children. These results add to the evidence that multilevel strategies should be developed to reduce SSBs intake, since up to 66.6% of Chinese students consumed SSBs according to a nationwide survey (Gui et al., 2017). Ours is the first study to report the modifying effect of SSBs consumption on air pollution-associated health effects. Reducing SSBs intake, a modifiable factor, might be included as part of a comprehensive strategy to mitigate the adverse effects of ambient PM exposure on FBG in children. Further studies should be conducted to explore the mechanism. Several signal pathways described in previous studies might account for associations between PM exposure and blood glucose metabolism (Rajagopalan and Brook, 2012; Liu et al., 2013; Rao et al., 2015b). PM can induce oxidative stress and low-grade inflammation in the lungs, which results in chronic inflammation and even affects adipose tissue (Rao et al., 2015b; Sun et al., 2005, 2009). Systemic and adipose tissue inflammation contribute to impaired insulin signaling pathways, which play vital roles in moderating glucose metabolism (Rao et al., 2015b; Haberzettl et al., 2016). Dietary consumption of polyphenols, carotenoids, flavonoids, vitamin B and C, and unsaturated fatty acids might mitigate the harmful effects of PM exposure, with their antioxidant activities protecting against exogenous oxidative damage and inflammatory response, and DNA methylation changes (Tong et al., 2012; Zhong et al., 2017; Bowler and Crapo, 2002; Romieu et al., 1998). Several previous epidemiological studies have shown that mediumterm air pollution was associated not only with systemic inflammation (Rajagopalan and Brook, 2012) but also with insulin resistance (Brook et al., 2013; Kelishadi et al., 2009; Kim and Hong, 2012), with similar findings from studies of glucose regulation among pregnant women (Fleisch et al., 2014; Lu et al., 2017; Robledo et al., 2015). The abnormal glucose metabolism associated with medium-term air pollution exposure may help to explain the associations between long-term air pollution exposure and incident diabetes mellitus (Balti et al., 2014). PM exposure in Guangzhou is generally considered much less severe than in the northern or central parts of China. Nonetheless, in the

5. Conclusions In conclusion, our results suggest that medium-term exposure to ambient PM2.5 and PM10 was associated with higher blood glucose level in children, and SSBs intake might modify the magnitude of these effects.

Declarations of interest None.

Funding This study was supported by grants from the National Key R&D Program of China (2018YFA0606200); National Natural Science Foundation of China (81673193); Natural Science Foundation of Guangdong Province, China (2017A030310249 and 2018A030313948); and Guangdong Provincial Natural Science Foundation Team Project, China (2018B030312005).

Ethics The study protocol was approved by the Ethics and Human Subject Committee of Sun Yat-Sen Medical College and all study participants and their parents provided informed consent. 7

Environmental Research 176 (2019) 108541

L. Cai, et al.

Conflicts of interest

Guangdong Environmental Monitoring Center. . http://www.gdemc.gov.cn/, Accessed date: 4 January 2018. Gui, Z.H., Zhu, Y.N., Cai, L., Sun, F.H., Ma, Y.H., Jing, J., Chen, Y.J., 2017. Sugarsweetened beverage consumption and risks of obesity and hypertension in Chinese children and adolescents: a national cross-sectional analysis. Nutrients 9. Guxens, M., Aguilera, I., Ballester, F., Estarlich, M., Fernandez-Somoano, A., Lertxundi, A., Lertxundi, N., Mendez, M.A., Tardon, A., Vrijheid, M., Sunyer, J., 2012. Prenatal exposure to residential air pollution and infant mental development: modulation by antioxidants and detoxification factors. Environ. Health Perspect. 120, 144–149. Haberzettl, P., O'Toole, T.E., Bhatnagar, A., Conklin, D.J., 2016. Exposure to fine particulate air pollution causes vascular insulin resistance by inducing pulmonary oxidative stress. Environ. Health Perspect. 124, 1830–1839. Kelishadi, R., Mirghaffari, N., Poursafa, P., Gidding, S.S., 2009. Lifestyle and environmental factors associated with inflammation, oxidative stress and insulin resistance in children. Atherosclerosis 203, 311–319. Kim, J.H., Hong, Y.C., 2012. GSTM1, GSTT1, and GSTP1 polymorphisms and associations between air pollutants and markers of insulin resistance in elderly Koreans. Environ. Health Perspect. 120, 1378–1384. Kim, E., Park, H., Hong, Y.C., Ha, M., Kim, Y., Kim, B.N., Kim, Y., Roh, Y.M., Lee, B.E., Ryu, J.M., Kim, B.M., Ha, E.H., 2014. Prenatal exposure to PM(1)(0) and NO(2) and children's neurodevelopment from birth to 24 months of age: mothers and Children's Environmental Health (MOCEH) study. Sci. Total Environ. 481, 439–445. Koskinen, J., Magnussen, C.G., Sinaiko, A., Woo, J., Urbina, E., Jacobs, D.J., Steinberger, J., Prineas, R., Sabin, M.A., Burns, T., Berenson, G., Bazzano, L., Venn, A., Viikari, J., Hutri-Kahonen, N., Raitakari, O., Dwyer, T., Juonala, M., 2017. Childhood age and associations between childhood metabolic syndrome and adult risk for metabolic syndrome, type 2 diabetes mellitus and carotid intima media thickness: the international childhood cardiovascular cohort consortium. J Am Heart Assoc 6. Liang, Y., Hou, D., Zhao, X., Wang, L., Hu, Y., Liu, J., Cheng, H., Yang, P., Shan, X., Yan, Y., Cruickshank, J.K., Mi, J., 2015. Childhood obesity affects adult metabolic syndrome and diabetes. Endocrine 50, 87–92. Lin, H., Guo, Y., Di, Q., Zheng, Y., Xian, H., Li, X., Liu, T., Xiao, J., Zeng, W., Howard, S.W., Vaughn, M.G., Qian, Z.M., Ma, W., Wu, F., 2018. Consumption of fruit and vegetables might mitigate the adverse effects of ambient PM2.5 on lung function among adults. Environ. Res. 160, 77–82. Liu, C., Ying, Z., Harkema, J., Sun, Q., Rajagopalan, S., 2013. Epidemiological and experimental links between air pollution and type 2 diabetes. Toxicol. Pathol. 41, 361–373. Liu, C., Yang, C., Zhao, Y., Ma, Z., Bi, J., Liu, Y., Meng, X., Wang, Y., Cai, J., Chen, R., Kan, H., 2016. Associations between long-term exposure to ambient particulate air pollution and type 2 diabetes prevalence, blood glucose and glycosylated hemoglobin levels in China. Environ. Int. 92–93, 416–421. Lu, M.C., Wang, P., Cheng, T.J., Yang, C.P., Yan, Y.H., 2017. Association of temporal distribution of fine particulate matter with glucose homeostasis during pregnancy in women of Chiayi City, Taiwan. Environ. Res. 152, 81–87. Lucht, S.A., Hennig, F., Matthiessen, C., Ohlwein, S., Icks, A., Moebus, S., Jockel, K.H., Jakobs, H., Hoffmann, B., 2018. Air pollution and glucose metabolism: an analysis in non-diabetic participants of the Heinz Nixdorf Recall study. Environ. Health Perspect. 126, 47001. Malik, V.S., Popkin, B.M., Bray, G.A., Despres, J.P., Hu, F.B., 2010. Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. Circulation 121, 1356–1364. NCD Risk Factor Collaboration, 2017. Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies in 128.9 million children, adolescents, and adults. Lancet 390, 2627–2642. Peng, C., Bind, M.C., Colicino, E., Kloog, I., Byun, H.M., Cantone, L., Trevisi, L., Zhong, J., Brennan, K., Dereix, A.E., Vokonas, P.S., Coull, B.A., Schwartz, J.D., Baccarelli, A.A., 2016. Particulate air pollution and fasting blood glucose in nondiabetic individuals: associations and epigenetic mediation in the normative aging study, 2000-2011. Environ. Health Perspect. 124, 1715–1721. Poursafa, P., Mansourian, M., Motlagh, M.E., Ardalan, G., Kelishadi, R., 2014. Is air quality index associated with cardiometabolic risk factors in adolescents? The CASPIAN-III Study. Environ. Res. 134, 105–109. Rajagopalan, S., Brook, R.D., 2012. Air pollution and type 2 diabetes: mechanistic insights. Diabetes 61, 3037–3045. Rao, X., Montresor-Lopez, J., Puett, R., Rajagopalan, S., Brook, R.D., 2015a. Ambient air pollution: an emerging risk factor for diabetes mellitus. Curr. Diabetes Rep. 15, 603. Rao, X., Patel, P., Puett, R., Rajagopalan, S., 2015b. Air pollution as a risk factor for type 2 diabetes. Toxicol. Sci. 143, 231–241. Robledo, C.A., Mendola, P., Yeung, E., Mannisto, T., Sundaram, R., Liu, D., Ying, Q., Sherman, S., Grantz, K.L., 2015. Preconception and early pregnancy air pollution exposures and risk of gestational diabetes mellitus. Environ. Res. 137, 316–322. Romieu, I., Meneses, F., Ramirez, M., Ruiz, S., Perez, P.R., Sienra, J.J., Gerber, M., Grievink, L., Dekker, R., Walda, I., Brunekreef, B., 1998. Antioxidant supplementation and respiratory functions among workers exposed to high levels of ozone. Am. J. Respir. Crit. Care Med. 158, 226–232. Sade, M.Y., Kloog, I., Liberty, I.F., Katra, I., Novack, L., Novack, V., 2015. Air pollution and serum glucose levels: a population-based study. Medicine (Baltim.) 94, e1093. Schottker, B., Rathmann, W., Herder, C., Thorand, B., Wilsgaard, T., Njolstad, I., Siganos, G., Mathiesen, E.B., Saum, K.U., Peasey, A., Feskens, E., Boffetta, P., Trichopoulou, A., Kuulasmaa, K., Kee, F., Brenner, H., 2016. HbA1c levels in non-diabetic older adults - No J-shaped associations with primary cardiovascular events, cardiovascular and all-cause mortality after adjustment for confounders in a meta-analysis of individual participant data from six cohort studies. BMC Med. 14, 26. Schulz, A.J., Mentz, G.B., Sampson, N.R., Dvonch, J.T., Reyes, A.G., Izumi, B., 2015.

The authors have no conflict of interest to disclose. interests. Author contributions C.H., Y.C., and L.C. designed the study. L.C., X.Z., and Z.G. participated in the data collection. S.W. and L.C. performed the statistical analysis and drafted the first manuscript. P.G. and J.B. developed part of the algorithm. C.H., L.C., M.B. and Y.C. interpreting the data and critically revised the manuscript. B.J. and Q.W. contributed to discussion and edited the manuscript. All authors read and approved the final manuscript. Acknowledgements We thank all individuals who volunteered to participate in this study. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.envres.2019.108541. References Balti, E.V., Echouffo-Tcheugui, J.B., Yako, Y.Y., Kengne, A.P., 2014. Air pollution and risk of type 2 diabetes mellitus: a systematic review and meta-analysis. Diabetes Res. Clin. Pract. 106, 161–172. Bowe, B., Xie, Y., Li, T., Yan, Y., Xian, H., Al-Aly, Z., 2018. The 2016 global and national burden of diabetes mellitus attributable to PM2.5 air pollution. Lancet Planet Health 2, e301–e312. Bowler, R.P., Crapo, J.D., 2002. Oxidative stress in airways: is there a role for extracellular superoxide dismutase? Am. J. Respir. Crit. Care Med. 166, S38–S43. Brauer, M., Lencar, C., Tamburic, L., Koehoorn, M., Demers, P., Karr, C., 2008. A cohort study of traffic-related air pollution impacts on birth outcomes. Environ. Health Perspect. 116, 680–686. Brook, R.D., Xu, X., Bard, R.L., Dvonch, J.T., Morishita, M., Kaciroti, N., Sun, Q., Harkema, J., Rajagopalan, S., 2013. Reduced metabolic insulin sensitivity following sub-acute exposures to low levels of ambient fine particulate matter air pollution. Sci. Total Environ. 448, 66–71. Bureau, G.M., 2018. Guangzhou government. http://www.gz.gov.cn/, Accessed date: 16 January 2018. Cai, Y., Hansell, A.L., Blangiardo, M., Burton, P.R., de Hoogh, K., Doiron, D., Fortier, I., Gulliver, J., Hveem, K., Mbatchou, S., Morley, D.W., Stolk, R.P., Zijlema, W.L., Elliott, P., Hodgson, S., 2017. Long-term exposure to road traffic noise, ambient air pollution, and cardiovascular risk factors in the HUNT and lifelines cohorts. Eur. Heart J. 38, 2290–2296. Chen, Z., Salam, M.T., Toledo-Corral, C., Watanabe, R.M., Xiang, A.H., Buchanan, T.A., Habre, R., Bastain, T.M., Lurmann, F., Wilson, J.P., Trigo, E., Gilliland, F.D., 2016. Ambient air pollutants have adverse effects on insulin and glucose homeostasis in Mexican Americans. Diabetes Care 39, 547–554. Chuang, K.J., Yan, Y.H., Chiu, S.Y., Cheng, T.J., 2011. Long-term air pollution exposure and risk factors for cardiovascular diseases among the elderly in Taiwan. Occup. Environ. Med. 68, 64–68. Cohen, B.E., Barrett-Connor, E., Wassel, C.L., Kanaya, A.M., 2009. Association of glucose measures with total and coronary heart disease mortality: does the effect change with time? The Rancho Bernardo Study. Diabetes Res. Clin. Pract. 86, 67–73. Cui, H.Y., Chen, W.H., Dai, W., Liu, H., Wang, X.M., He, K.B., 2015. Source apportionment of PM2.5 in Guangzhou combining observation data analysis and chemical transport model simulation. Atmos. Environ. 116, 262–271 p. 116, 262-271. GBD 2016 Risk Factors Collaborators, 2017. Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 390, 1345–1422. Dadvand, P., Poursafa, P., Heshmat, R., Motlagh, M.E., Qorbani, M., Basagana, X., Kelishadi, R., 2018. Use of green spaces and blood glucose in children; a populationbased CASPIAN-V study. Environ. Pollut. 243, 1134–1140. Eze, I.C., Hemkens, L.G., Bucher, H.C., Hoffmann, B., Schindler, C., Kunzli, N., Schikowski, T., Probst-Hensch, N.M., 2015. Association between ambient air pollution and diabetes mellitus in Europe and North America: systematic review and metaanalysis. Environ. Health Perspect. 123, 381–389. Fleisch, A.F., Gold, D.R., Rifas-Shiman, S.L., Koutrakis, P., Schwartz, J.D., Kloog, I., Melly, S., Coull, B.A., Zanobetti, A., Gillman, M.W., Oken, E., 2014. Air pollution exposure and abnormal glucose tolerance during pregnancy: the project Viva cohort. Environ. Health Perspect. 122, 378–383. Franklin, B.A., Brook, R., Arden, P.C.R., 2015. Air pollution and cardiovascular disease. Curr. Probl. Cardiol. 40, 207–238.

8

Environmental Research 176 (2019) 108541

L. Cai, et al.

proximity with fasting plasma glucose and metabolic risk factors for cardiovascular disease in a cross-sectional study of cardiac catheterization patients. Environ. Health Perspect. 123, 1007–1014. Wolf, K., Popp, A., Schneider, A., Breitner, S., Hampel, R., Rathmann, W., Herder, C., Roden, M., Koenig, W., Meisinger, C., Peters, A., 2016. Association between longterm exposure to air pollution and biomarkers related to insulin resistance, subclinical inflammation, and adipokines. Diabetes 65, 3314–3326. World Health Orgnization, 2006. WHO Air quality guidelines for particulate matter, ozone, nitrogen dioxide and sulfur dioxide-Global update 2005. http://www.who. int/phe/health_topics/outdoorair/outdoorair_aqg/en/index.html, Accessed date: 29 September 2016. Wu, L., Jin, L., Shi, T., Zhang, B., Zhou, Y., Zhou, T., Bao, W., Xiang, H., Zuo, Y., Li, G., Wang, C., Duan, Y., Peng, Z., Huang, X., Zhang, H., Xu, T., Li, Y., Pan, X., Xia, Y., Gong, X., Chen, W., Liu, Y., 2017. Association between ambient particulate matter exposure and semen quality in Wuhan, China. Environ. Int. 98, 219–228. Yang, B.Y., Qian, Z.M., Li, S., Chen, G., Bloom, M.S., Elliott, M., Syberg, K.W., Heinrich, J., Markevych, I., Wang, S.Q., Chen, D., Ma, H., Chen, D.H., Liu, Y., Komppula, M., Leskinen, A., Liu, K.K., Zeng, X.W., Hu, L.W., Guo, Y., Dong, G.H., 2018a. Ambient air pollution in relation to diabetes and glucose-homoeostasis markers in China: a crosssectional study with findings from the 33 Communities Chinese Health Study. Lancet Planet Health 2, e64–e73. Yang, B.Y., Bloom, M.S., Markevych, I., Qian, Z.M., Vaughn, M.G., Cummings-Vaughn, L.A., Li, S., Chen, G., Bowatte, G., Perret, J.L., Dharmage, S.C., Heinrich, J., Yim, S.H., Lin, S., Tian, L., Yang, M., Liu, K.K., Zeng, X.W., Hu, L.W., Guo, Y., Dong, G.H., 2018b. Exposure to ambient air pollution and blood lipids in adults: the 33 Communities Chinese Health Study. Environ. Int. 119, 485–492. Zhang, X., Gregg, E.W., Williamson, D.F., Barker, L.E., Thomas, W., Bullard, K.M., Imperatore, G., Williams, D.E., Albright, A.L., 2010. A1C level and future risk of diabetes: a systematic review. Diabetes Care 33, 1665–1673. Zhong, J., Karlsson, O., Wang, G., Li, J., Guo, Y., Lin, X., Zemplenyi, M., Sanchez-Guerra, M., Trevisi, L., Urch, B., Speck, M., Liang, L., Coull, B.A., Koutrakis, P., Silverman, F., Gold, D.R., Wu, T., Baccarelli, A.A., 2017. B vitamins attenuate the epigenetic effects of ambient fine particles in a pilot human intervention trial. Proc. Natl. Acad. Sci. U. S. A. 114, 3503–3508.

Effects of particulate matter and antioxidant dietary intake on blood pressure. Am. J. Public Health 105, 1254–1261. Sly, P., Flack, F., 2008. Susceptibility of children to environmental pollutants. Ann. N. Y. Acad. Sci. 1, 163–183. Sun, Q., Wang, A., Jin, X., Natanzon, A., Duquaine, D., Brook, R.D., Aguinaldo, J.G., Fayad, Z.A., Fuster, V., Lippmann, M., Chen, L.C., Rajagopalan, S., 2005. Long-term air pollution exposure and acceleration of atherosclerosis and vascular inflammation in an animal model. J. Am. Med. Assoc. 294, 3003–3010. Sun, Q., Yue, P., Deiuliis, J.A., Lumeng, C.N., Kampfrath, T., Mikolaj, M.B., Cai, Y., Ostrowski, M.C., Lu, B., Parthasarathy, S., Brook, R.D., Moffatt-Bruce, S.D., Chen, L.C., Rajagopalan, S., 2009. Ambient air pollution exaggerates adipose inflammation and insulin resistance in a mouse model of diet-induced obesity. Circulation 119, 538–546. Thiering, E., Cyrys, J., Kratzsch, J., Meisinger, C., Hoffmann, B., Berdel, D., von Berg, A., Koletzko, S., Bauer, C.P., Heinrich, J., 2013. Long-term exposure to traffic-related air pollution and insulin resistance in children: results from the GINIplus and LISAplus birth cohorts. Diabetologia 56, 1696–1704. Thiering, E., Markevych, I., Bruske, I., Fuertes, E., Kratzsch, J., Sugiri, D., Hoffmann, B., von Berg, A., Bauer, C.P., Koletzko, S., Berdel, D., Heinrich, J., 2016. Associations of residential long-term air pollution exposures and satellite-derived greenness with insulin resistance in German adolescents. Environ. Health Perspect. 124, 1291–1298. Toledo-Corral, C.M., Alderete, T.L., Habre, R., Berhane, K., Lurmann, F.W., Weigensberg, M.J., Goran, M.I., Gilliland, F.D., 2018. Effects of air pollution exposure on glucose metabolism in Los Angeles minority children. Pediatr Obes 13, 54–62. Tong, H., Rappold, A.G., Diaz-Sanchez, D., Steck, S.E., Berntsen, J., Cascio, W.E., Devlin, R.B., Samet, J.M., 2012. Omega-3 fatty acid supplementation appears to attenuate particulate air pollution-induced cardiac effects and lipid changes in healthy middleaged adults. Environ. Health Perspect. 120, 952–957. Wang, S., Dong, Y.H., Wang, Z.H., Zou, Z.Y., Ma, J., 2017. [Trends in overweight and obesity among Chinese children of 7-18 years old during 1985-2014]. Zhonghua Yufang Yixue Zazhi 51, 300–305. Ward-Caviness, C.K., Kraus, W.E., Blach, C., Haynes, C.S., Dowdy, E., Miranda, M.L., Devlin, R.B., Diaz-Sanchez, D., Cascio, W.E., Mukerjee, S., Stallings, C., Smith, L.A., Gregory, S.G., Shah, S.H., Hauser, E.R., Neas, L.M., 2015. Association of roadway

9