- Email: [email protected]
Effect of residential exposure to green space on maternal blood glucose levels, impaired glucose tolerance, and gestational diabetes mellitus
Environmental Research 176 (2019) 108526
Contents lists available at ScienceDirect
Environmental Research journal homepage: www.elsevier.com/locate/envres
Effect of residential exposure to green space on maternal blood glucose levels, impaired glucose tolerance, and gestational diabetes mellitus
T
Jiaqiang Liaoa,1, Xinmei Chena,1, Shunqing Xua, Yuanyuan Lia, Bin Zhangb, Zhongqiang Caob, Yiming Zhangb, Shengwen Liangc, Ke Huc, Wei Xiaa,∗ a Key Laboratory of Environment and Health, Ministry of Education & Ministry of Environmental Protection, And State Key Laboratory of Environmental Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, People's Republic of China b Women and Children Medical and Healthcare Center of Wuhan, Wuhan, Hubei, People's Republic of China c Wuhan Environmental Monitoring Center, Wuhan, Hubei Province, 430000, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Green spaces Maternal blood glucose levels Impaired glucose tolerance Gestational diabetes mellitus
Background: Residential surrounding green spaces can affect human health. However, limited studies have examined their impacts on maternal blood glucose homeostasis outcomes. Objective: We examined the associations of residential exposure to green space with maternal blood glucose levels, gestational impaired glucose tolerance (IGT), and gestational diabetes mellitus (GDM). Methods: Pregnant women were recruited from a prospective birth cohort between October 2012 and September 2015. Exposure to green space was calculated as the mean value of the normalized difference vegetation index (NDVI) within a 300-m circular buffer area surrounding each residence. Maternal glucose was measured between 24 and 28 weeks of gestation, and gestational IGT and GDM were diagnosed using valid methods. We estimated the associations of residential NDVI with maternal glucose levels using multiple linear regression models with adjustment for age, education, BMI, passive smoking during pregnancy, parity, season of conception, income, and urbancity. We estimated the relative risks of residential NDVI with IGT and GDM using a generalized estimating equation model with modified Poisson regression. The mediation effects of residential exposure to air pollution and maternal physical activity were assessed using causal mediation analysis. Results: Of 6807 pregnant women, 751 (11.3%) and 604 (8.8%) were diagnosed with IGT and GDM, respectively. One SD increment of residential NDVI was associated with a decrease of 0.06 mmol/L (95% CI: −0.07, −0.05), 0.09 mmol/L (95% CI: −0.13, −0.05), and 0.06 mmol/L (95% CI: −0.09, −0.03) in maternal fasting glucose levels, 1-h glucose levels, and 2-h glucose levels, respectively, as well as reduced risks of incident IGT (RR: 0.92, 95% CI: 0.86, 0.99) and GDM (RR: 0.85, 95% CI: 0.79, 0.92). The association between residential NDVI and maternal fasting glucose levels was partly mediated by maternal exposure to PM2.5. Conclusion: Living with higher levels of green space was significantly associated with decreased maternal glucose levels and attenuated risks of incident maternal IGT and GDM. Our findings provide evidence linking green space to better maternal glucose outcomes. More studies are needed to further explore the maternal and child health benefits related to our findings.
1. Introduction Urban green spaces, such as parks and sidewalks, have been widely linked to improved health and well-being (WHO, 2016). Population studies have found that higher levels of surrounding green spaces were associated with reduced risks of cardiovascular disease mortality,
incident stroke, and incident depressive symptoms in younger and older populations (Banay et al., 2019; Orioli et al., 2019; Shen and Lung, 2016). Additionally, some evidence suggests that higher levels of residential surrounding green spaces may be protective against adverse pregnancy outcomes, such as preterm birth and low birth weight (Laurent et al., 2013; Markevych et al., 2014; Nichani et al., 2017).
Abbreviations: FGL, fast glucose level; GL1, 1-h glucose level; GL2, 2-h glucose level ∗ Corresponding author. School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, People's Republic of China. E-mail address: [email protected] (W. Xia). 1 These authors contributed equally. https://doi.org/10.1016/j.envres.2019.108526 Received 19 March 2019; Received in revised form 30 May 2019; Accepted 5 June 2019 Available online 06 June 2019 0013-9351/ © 2019 Elsevier Inc. All rights reserved.
Environmental Research 176 (2019) 108526
J. Liao, et al.
Fig. 1. Maps of NDVI (2013) and geospatial distributions of pregnant women in the studied area.aHigher values represent greater vegetative cover.
incident type 2 diabetes (Muller et al., 2018; B Yang et al., 2018). However, as far as we know, only one study examined the association of residential surrounding green space with GDM, and observed a nonsignificant association (Choe et al., 2018). No study has been conducted to examine the impact of residential surrounding green space on gestational impaired glucose tolerance (IGT), which has been widely reported to be associated with elevated risk of adverse maternal outcomes such as caesarean section rate, prematurity and infant diagnosis with macrosomia (Ostlund et al., 2003; Wong et al., 2007). Additionally, the potential mechanisms underlying these associations are not reported. Therefore, we hypothesized that exposure to higher levels of residential surrounding green spaces could benefit maternal glucose levels, decreasing the risks of incident maternal IGT and GDM. We examined these assumptions based on an ongoing prospective birth cohort. To explore the potential mechanisms underlying these
Limited evidence has been reported regarding the associations of exposure to green space with maternal blood glucose-homeostasis outcomes. Exposure to higher levels of air pollution and noise, and lower levels of physical activity were reported to be associated with elevated risks of incident diabetes and gestational diabetes mellitus (GDM) (Nasiri-Amiri et al., 2016; Sorensen et al., 2013; Wang et al., 2017; B Yang et al., 2018), while exposure to higher levels of residential surrounding green spaces is known to promote health by contributing to lower levels of air pollution and noise exposure, increased odds of physical activity, and reduced likelihood of stress and depression (Cohen-Cline et al., 2015; James et al., 2015; Margaritis and Kang, 2017; Nieuwenhuijsen et al., 2017; Roe et al. 2013; Sallis et al., 2012). A few studies have reported that exposure to higher levels of residential surrounding green spaces was associated with lower fasting glucose levels and 1-h glucose levels, and lower risks of insulin resistance and
2
Environmental Research 176 (2019) 108526
J. Liao, et al.
diabetes before conception. Plasma glucose levels were measured using a Roche Modular P800 automated biochemistry analyzer at the Wuhan Women and Children Medical and Healthcare Center. According to World Health Organization (WHO) 2006 recommendations, pregnant women were diagnosed with IGT if both of the following conditions were satisfied: 1) fasting blood glucose level < 7.0 mmol/L; and 2) 2-h blood glucose level ≥7.8 and < 11.10 mmol/L. According to the International Association of Diabetes and Pregnancy Study Group (Metzger et al., 2010), pregnant women were diagnosed with GDM if any of the following conditions were satisfied: 1) fasting blood glucose level 5.1 mmol/L; 2) 1-h blood glucose level 10 mmol/L; or 3) 2-h blood glucose level 8.5 mmol/L.
associations, we further examined the mediation effects of maternal exposure to air pollution and maternal physical activity. 2. Methods 2.1. Population and study design The studied pregnant women were from a prospective birth cohort conducted in Wuhan, China between October 2012 and September 2015. Pregnant women were invited to participate in the study when they came to the Women and Children Medical and Healthcare Center of Wuhan for their first prenatal care visit. The criteria for recruitment were residence in Wuhan City, with singleton pregnancy, and without diabetes before pregnancy. Then, the enrolled pregnant women were followed for the duration of the pregnancy. The research protocol was approved by the Institutional Review Board of Tongji Medical School, Huazhong University of Science and Technology. All participants provided written informed consent at enrolment. Based on the conditions described above, we initially included 6883 pregnant women in the study. We excluded pregnant women who had missing data on annual household income (n = 70), educational background (n = 1), and prepregnancy BMI (n = 5). Therefore, we had 6807 pregnant women in the final analysis.
3. Statistical analysis 3.1. Main analyses We explored associations of residential surrounding NDVI and maternal glucose levels by scatter plots. We then estimated the changes and 95% confidence intervals (CIs) of maternal blood glucose levels (fasting glucose levels, 1-h glucose levels, and 2-h glucose levels) associated with increases of residential surrounding green spaces (one SD increments and quartiles) using linear regression models. Since incidences of IGT and GDM in our study population were relatively common (nearly 10% or more), we estimated the relative risks (RRs) and 95% CIs of incident IGT and GDM associated with increases in residential surrounding green spaces (one SD increments and quartiles) using generalized estimating equation (GEE) models with modified Poisson regression, which corrected the standardized errors to better approximates of RRs for common outcomes (Zou, 2004). We modeled the median value within each stratum of residential NDVI quartiles as a continuous variable entering multiple regression models, and the GEE models and P values for trends were obtained by conducting Wald chisquared tests. We used directed acyclic graph (DAG) methods to determine the covariates in multiple adjusted models. DAG-defined covariates are listed in supplementary mateial Figure S1. The following covariates were included: residence area (rural, urban), maternal age (≤24, 25–35, ≥36 years), maternal education (≤9, 10–12, ≥13 years), maternal household income (< 30,000, 30,000–50000, 50,000–1000000, ≥100,000 yuan), maternal prepregnancy BMI (underweight: < 18.5, normal: 18.5–24, overweight or obesity: 24 kg/m2), passive smoking during pregnancy (yes, no), maternal parity (1: nulliparous women, ≥2: parous women), and season of conception (spring, summer, autumn, and winter). Since few pregnant women reported active smoking during pregnancy (n = 37, 0.54%), we did not adjust for active smoking in the multivariable analyses.
2.2. Data collection We collected the maternal-related characteristics such as residential address, maternal age, and maternal education, passive smoking during pregnancy, maternal weight and prepregnancy height using a questionnaire interview at enrolment. Information on the date of birth, infant's sex, and infant's birth weight was obtained from medical records. Gestational age was estimated by the 1st-trimester ultrasound and the correct date of the last menstrual period (LMP). The maternal prepregnancy BMI was evaluated using maternal prepregnancy weight in kg divided by the square of the maternal height in meters. 2.3. Evaluation of maternal exposure to green space We assessed the exposure to green space by calculating a satellitebased normalized difference vegetation index (NDVI) for pregnant women based on their residential addresses, which is a normalized score ranging from −1 to 1, with lower values representing water bodies or barren areas and higher values representing photosynthetically active vegetation. We generated the data from remotely sensed daily averaged NDVI values covering the study area using MODIS raster data with a spatial resolution of 500 m× 500 m, which was collected from the International Scientific & Technical Data Mirror Site, Computer Network Information Center, Chinese Academy of Sciences (http://www.gscloud.cn). Since NDVI was unlikely to change over the gestational period for each pregnant woman, we only calculated the NDVI for the conception year of each pregnant woman. We selected the daily NDVI maps with the lowest rate of cloud coverage during the spring season (most green season during each year) in Wuhan city, which were 1st May, 1st May, 1st Jun, and 1st April for pregnant women whose conception years were 2012, 2013, 2014, and 2015, respectively. The geospatial distribution of residential addresses of pregnant women and the map of NDVI in Wuhan city were listed in Fig. 1. Based on existing evidence, we evaluated the residential exposure to green space by calculating the averaged NDVI values within a 300-m buffer area surrounding the residential address for each pregnant woman.
3.2. Mediation analyses of air pollution and maternal physical outdoor activities To further explore the mechanisms underlying the associations of residential exposure to green space with maternal blood glucose outcomes, we examined the mediation effects of maternal exposure to air pollution and maternal physical outdoor activity on these associations. We measured air pollution by evaluating residential exposure to fine particulate matter (PM2.5). Maternal daily exposures to PM2.5 from the date of the LMP to the date of the OGTT examination were estimated using a spatial-temporal land use regression (LUR) model adjusted for short-term (weeks) and long-term (seasons) variations (within 12×4 knots of natural cubic splines in each year). The details of the evaluation were introduced in our previous study (Liao et al., 2018). In brief, we collected data from 20 ground-level surveillance stations recording daily average PM2.5 concentration across the studied area, which were used as outcome variables in the spatial-temporal LUR model. We included the following predictor variables in our spatial-temporal LUR
2.4. Measurements of maternal glucose levels, diagnosis of IGT and GDM A 75-g oral glucose tolerance test (OGTT) was conducted within gestational weeks 24–28 for the pregnant women who did not have 3
Environmental Research 176 (2019) 108526
J. Liao, et al.
resided in the lowest quartile, respectively.
model: residential areas of commercial and industrial land use with a circular buffer of 5 km, residential areas of farmland use with a circular buffer of 15 km, residential areas of lake with a circular buffer of 10 km, lengths of major highways with a circular buffer of 5 km around the residential address, numbers of pollutant-related enterprises with a circular buffer of 15 km, and daily average temperature. The 10-fold out-of-sample cross-validation (CV) indicated that our method had good ability to predict maternal daily PM2.5 exposures (model explained variance R2: 71.5%, average 10-fold CV R2: 72.6%). Since most of the pregnant women had physical activities equal to or above 3 times per week (75.9%), we reclassified the variable of maternal physical activity into a binary variable of physical inactivity (0: ≥3 times per week, 1: ≤3 times per week). We conducted the causal mediation analyses by examining a continuous mediator of average PM2.5 concentration from the date of the LMP to the date of the OGTT examination and a binary mediator of physical inactivity on the associations of residential exposure to green space with maternal blood glucose outcomes.
4.2. Results of mediation analyses Exposure to higher levels of residential surrounding green space was associated with lower levels of residential PM2.5 exposure and reduced risks of maternal physical inactivity (Supplementary Material Table S1). After adjusting for covariates, a one SD increment of residential NDVI was associated with a decrease of −14.87 μg/m3 (95% CI: −15.40, −14.33) in maternal PM2.5 exposure and a lower risk of 0.87 (95% CI: 0.80, 0.95) in maternal physical inactivity. Similar associations were observed in the associations of maternal PM2.5 exposure and maternal physical inactivity with residential NDVI by quartiles. Exposure to higher levels of maternal PM2.5 was associated with increased maternal fasting glucose levels and 1-h glucose levels and higher risk of incident GDM (Supplementary Material Table S2). After adjusting for covariates, one 10 μg/m3 increment of maternal PM2.5 exposure was associated with an increase of 0.02 mmol/L (95% CI: 0.02, 0.03) in maternal fasting blood glucose levels, an increase of 0.02 mmol/L (95% CI: 0.00, 0.03) in 1-h maternal glucose levels, and an elevated risk of 1.05 (95% CI: 1.02, 1.09) in incident GDM. No associations of maternal physical inactivity with maternal glucose homeostatic outcomes were observed. The results of mediation analyses further indicated that maternal exposure to PM2.5 mediated the association between residential exposure to green space and maternal fasting glucose levels (Table 3). After adjusting for covariates, a one SD increment of residential NDVI contributed to a decrease of 0.05 mmol/L (95% CI: −0.06, −0.03) in maternal fasting glucose level, and 22.1% (95% CI: 18.4%, 27.4%) of this association can be explained by increased maternal PM2.5 concentration. No significant mediation effects of maternal PM2.5 exposure on other maternal blood glucose outcomes were observed.
3.3. Sensitivity analyses To further evaluate the robustness of our conclusions, we conducted several sensitivity analyses to examine the associations of residential surrounding greenness with maternal glucose outcomes: (1) evaluating the residential surrounding greenness by a buffer area of 1000 m; and (2) evaluating the residential surrounding greenness by simply extracting cell values to the residence location. All statistical analyses were conducted by SAS 9.4 and STATA 13.0. The statistical significance level was set to 0.05. 4. Results Of 6807 pregnant women, 11.03% were diagnosed with IGT, and 8.87% were diagnosed with GDM (Table 1). Among them, those who came from urban areas, from families with lower levels of annual household income, had an older age, were less educated, were overweight or obese prepregnancy, and were multiparous were more likely to be diagnosed with IGT or GDM.
4.3. Results of sensitivity analyses The associations of residential exposure to green space with maternal blood glucose levels, incident IGT, and incident GDM were similar to the results of the main analyses when we measured residential surrounding NDVI by circular buffer areas of 1000 m, and simply extracted cell values to the residence location (Supplementary Material Table S3-S4).
4.1. Results of main analyses The scatter plots (Fig. 2(a)-(c)) indicated that residential surrounding NDVI was negatively associated with maternal fasting glucose level, 1-h glucose level, and 2-h glucose level. Table 2 depicts the associations of residential exposure to green space with maternal glucosehomeostasis outcomes. After adjusting for the covariates, a one SD increment of residential NDVI was associated with lower maternal fasting glucose levels (mean difference: −0.06 mmol/L, 95% CI: −0.08, 0.05), 1-h glucose levels (mean difference: −0.10 mmol/L, 95% CI: −0.14, −0.06), and 2-h glucose levels (mean difference: −0.07 mmol/L, 95% CI: −0.10, −0.04), and lower risks of incident IGT (RR: 0.92, 95% CI: 0.86, 0.99), and incident GDM (RR: 0.85, 95% CI: 0.78, 0.93). The doseresponse associations of quartile measurements of residential NDVI with maternal blood glucose outcomes were observed (all values of P for trend < 0.05). Pregnant women who resided in higher quartiles of residential NDVI had lower blood glucose levels. For example, compared with pregnant women who resided in the lowest quartile of residential NDVI, the adjusted mean differences in fasting glucose levels were −0.10 mmol/L (95% CI: −0.14, −0.07), −0.14 mmol/L (95% CI: −0.18, −0.11), and −0.17 mmol/L (95% CI: −0.21, −0.14) in those who resided in the second, third, and fourth quartiles of residential NDVI, respectively. Furthermore, pregnant women who resided in higher quartiles of residential NDVI had lower risks of incident IGT and GDM. The adjusted RRs for incident IGT and GMD were 0.77 (95% CI: 0.62, 0.96) and 0.66 (95% CI: 0.52, 0.84) for pregnant women who resided in the highest quartile of residential NDVI and those who
5. Discussion We examined the associations of residential exposure to green space with maternal blood glucose levels, incident gestational IGT, and incident GDM based on a prospective birth cohort. We have reported several key findings. First, we consistently observed that exposure to higher levels of residential surrounding green spaces was associated with lower levels of maternal fasting glucose, 1-h glucose, and 2-h glucose as well as reduced risks of incident IGT and GDM. Second, we demonstrated that reduced levels of maternal exposure to PM2.5 mediated 18.4%–27.4% of the association between residential exposure to green space and maternal fasting glucose levels. This study provides preliminary epidemiological evidence linking exposure to higher levels of residential surrounding green spaces to lower risks of gestational IGT. Given the extraordinary important roles of gestational IGT in elevating risks of maternal caesarean rate, prematurity, large for gestational age, and macrosomic infants (Ostlund et al., 2003; Wong et al., 2007), our findings were important to further determine the mechanisms of green space on maternal and child health. Our findings were consistent with the reported evidence linking green space to incident diabetes in the general population. A population study examined the association of neighborhood green space and the rate of type 2 diabetes mellitus (T2DM) in 267,072 Australians, which demonstrated that participants from greener neighborhoods had lower 4
Environmental Research 176 (2019) 108526
J. Liao, et al.
Table 1 Baseline characteristics of studied pregnant women stratified by IGT status and GDM status (n (%)). Characteristics
All
GDM
No GDM
IGT
No IGT
Total Residential areas Rural Urban Maternal age, years ≤24 25–29 30–34 ≥35 Education, years ≤9 10–12 ≥13 Annual household income, yuan < 30,000 30,000–50,000 50,000–100,000 > 100,000 Prepregnancy BMI, kg/m2 Underweight(< 18.5) Normal(18.5–24.0) Overweight or obesity(≥24) Parity 1: nulliparous women ≥2: parous women Passive smoking during pregnancy No Yes Season of LMP Spring Summer Autumn Winter
6807 (100.00)
604 (8.87)
6203 (91.13)
751 (11.03)
6056 (88.97)
1253 (18.41) 5554 (81.59)
101 (16.72) 503 (83.28)
1152 (18.57) 5051 (81.43)
128 (17.04) 623 (82.96)
1125 (18.58) 4931 (81.42)
565 (8.3) 4039 (59.34) 1827 (26.84) 376 (5.52)
27 (4.47) 278 (46.03) 217 (35.93) 82 (13.58)
538 (8.67) 3761 (60.63) 1610 (25.96) 294 (4.74)
23 (3.06) 379 (50.47) 260 (34.62) 89 (11.85)
542 (8.95) 3660 (60.44) 1567 (25.88) 287 (4.74)
455 (6.68) 1095 (16.09) 5257 (77.23)
61 (10.1) 115 (19.04) 428 (70.86)
394 (6.35) 980 (15.8) 4829 (77.85)
55 (7.32) 115 (15.31) 581 (77.36)
400 (6.61) 980 (16.18) 4676 (77.21)
726 (10.67) 1442 (21.18) 2889 (42.44) 1750 (25.71)
73 (12.09) 143 (23.68) 256 (42.38) 132 (21.85)
653 (10.53) 1299 (20.94) 2633 (42.45) 1618 (26.08)
81 (10.79) 168 (22.37) 311 (41.41) 191 (25.43)
645 (10.65) 1274 (21.04) 2578 (42.57) 1559 (25.74)
1401 (20.58) 4592 (67.46) 814 (11.96)
65 (10.76) 390 (64.57) 149 (24.67)
1336 (21.54) 4202 (67.74) 665 (10.72)
103 (13.72) 498 (66.31) 150 (19.97)
1298 (21.43) 4094 (67.6) 664 (10.96)
6106 (89.7) 701 (10.3)
503 (83.28) 101 (16.72)
5603 (90.33) 600 (9.67)
639 (85.09) 112 (14.91)
5467 (90.27) 589 (9.73)
5131 (75.38) 1676 (24.62)
458 (75.83) 146 (24.17)
4673 (75.33) 1530 (24.67)
562 (74.83) 189 (25.17)
4569 (75.45) 1487 (24.55)
955 (14.03) 1834 (26.94) 2403 (35.3) 1615 (23.73)
67 (11.09) 176 (29.14) 225 (37.25) 136 (22.52)
888 (14.32) 1658 (26.73) 2178 (35.11) 1479 (23.84)
116 168 265 202
839 (13.85) 1666 (27.51) 2138 (35.30) 1413 (23.33)
(15.45) (22.37) (35.29) (26.9)
Abbreviations: IGT: impaired glucose tolerance; GDM: gestational diabetes.
risks of T2DM (Astellburt et al., 2014). Similar studies have also reported that individuals living in the greener neighborhoods had lower risks of developing diabetes (Dalton et al. 2016; Muller et al., 2018). For children, those who spent more time in green spaces had lower fasting blood glucose levels (Dadvand et al., 2018). Another study that examined the associations of green space and markers related to blood glucose in adults reported that exposure to higher levels of residential surrounding greenness was associated with lower fasting and 2-h glucose levels, lower 2-h insulin levels, lower insulin resistance and higher
β -cell function (B Yang et al., 2018). As far as we know, only one study has examined the association of residential surrounding greenness with incident GDM, which found a nonsignificant association (Choe et al., 2018). In this study, we consistently demonstrated that pregnant women with higher levels of residential surrounding green spaces had lower fasting glucose levels, 1-h glucose levels, and 2-h glucose levels and lower risks of incident gestational IGT and GDM. Numerous studies have highlighted the consistent mediation role of reduced levels of air pollution in the associations of exposure to green
Fig. 2. The scatter plots of residential surrounding NDVI with maternal fasting glucose level, 1-h glucose level, and 2-h glucose levela. aRed lines represent the fitted lines by linear regression, and the green areas represent the 95% confidence intervals of predicted lines. 5
Environmental Research 176 (2019) 108526
J. Liao, et al.
Table 2 Adjusted changes of residential exposure to green space with maternal glucose levels, IGT, and GDM. Measurements of green space
Changes of maternal glucose levels (95% CI), mmol/La Fasting glucose
One SD (0.15) increment, Categorized into quartiles Q1: −0.09–0.13 Q2: 0.14–0.32 Q3: 0.33–0.40 Q4: ≥0.41 P for trend
−0.06 (−0.08, −0.05)
1-h glucose ∗
Reference −0.10 (−0.14, −0.07)∗ −0.14 (−0.18, −0.11)∗ −0.17 (−0.21, −0.14)∗ < .001
RR (95% CI) for IGTb
RR (95% CI) for GDMb
2-h glucose ∗
−0.10 (−0.14, −0.06)
−0.07 (−0.10, −0.04)∗
0.92 (0.86, 0.99)∗
0.85 (0.78, 0.93)∗
Reference −0.10 (−0.21, 0.02) −0.21 (−0.32, −0.10)∗ −0.27 (−0.39, −0.15)∗ < .001
Reference −0.05 (−0.14, 0.04) −0.09 (−0.18, 0.00) −0.19 (−0.28, −0.10)∗ < .001
Reference 0.95 (0.77, 1.17) 0.93 (0.75, 1.14) 0.77 (0.62, 0.96)∗ 0.022
Reference 0.98 (0.79, 1.22) 0.74 (0.59, 0.93)∗ 0.66 (0.52, 0.84)∗ < .001
Abbreviations: IGT: impaired glucose tolerance; GDM: gestational diabetes; RR: relative risk; CI: confidence interval. ∗ P < 0.05. a The changes in glucose levels and 95% CIs were estimated by multiple linear regression models, adjusting for age, education years, BMI, passive smoking during pregnancy, parity, season of conception, income, and urban areas.. b The RRs and 95% CIs of incident IGT were estimated by a generalized estimating equation model with modified Poisson regression, adjusting for age, education years, BMI, passive smoking during pregnancy, parity, season of conception, income, and urban areas. The RRs and 95% CIs of incident GDM were estimated by a generalized estimating equation model with modified Poisson regression, adjusting for age, education years, BMI, passive smoking during pregnancy, parity, season of conception, income, and urban areas..
≥3 times per week, 1: ≤3 times per week) might be not enough to classify the different levels of maternal physical activity in our study population. More mechanisms underlying the associations of residential exposure to green space with maternal blood glucose outcomes should be further explored. For example, reduced levels of noise exposure might be a potential mechanism linking green space to maternal blood glucose outcomes. Experimental evidence indicated that higher levels of urban green spaces were associated with lower levels of noise exposure (Margaritis and Kang, 2017; WHO, 2010), while higher levels of noise exposure were associated with elevated risk of incident diabetes and GDM (Eriksson et al., 2014; Margaritis and Kang, 2017; Pedersen et al., 2017). Second, disruptions of sleep duration and poorer sleep quality are potential mechanisms underlying the associations of exposure to green space with maternal blood glucose outcomes. Population evidence indicated that exposure to higher levels of green space was associated with better sleep duration (Astellburt et al., 2013; Johnson et al., 2018), while short sleep duration and poor levels of sleep quality were reported to be associated with higher risks of incident GDM (Cai et al., 2017; Facco et al., 2017; Wang et al., 2017). Finally, reduced odds of major stressful life events might mediate the associations of green space with maternal blood glucose outcomes. Exposure to higher levels of urban neighborhood green spaces was associated with lower levels of stress (Roe et al., 2013; Roe et al., 2017; Trigueromas et al., 2017). Exposure to higher levels of stress was associated with higher risk of incident diabetes (Ge et al., 2006; Mooy et al., 2000; Toshihiro et al., 2008). This study has several strengths. The main strength is that the studied population was from a prospective birth cohort and all pregnant women were suggested to participate in OGTT for screening GDM from 24 to 28 weeks of gestation, which offered us an opportunity to examine multiple maternal blood glucose outcomes with exposures. We
space with adverse health outcomes ( Dadvand et al., 2015a, 2015b; Lee et al., 2018; Nieuwenhuijsen et al., 2017; B Yang et al., 2018). In this study, we further found that reduced levels of maternal PM2.5 exposure partly mediated (from 19.4% to 27.7%) the association of residential exposure to green space with maternal fasting glucose levels. Mechanistic studies indicated that exposure to higher levels of ambient PM2.5 induced oxidative stress and contributed to insulin resistance (Haberzettl et al., 2016; Lodovici and Bigagli, 2011; Xu et al., 2011). Population evidence also reported consistent associations of PM2.5 exposure with glucose outcomes in the general population or in pregnant women. In the general population, exposure to higher levels of ambient PM2.5 was associated with higher fasting glucose levels and elevated risk of incident diabetes (Liang et al., 2019; Lucht et al., 2018; Park et al., 2015; BY Yang et al., 2018). In pregnant women, exposure to higher levels of ambient PM2.5 was associated with higher fasting OGTT glucose levels and elevated risks of incident IGT and GDM (Fleisch et al., 2014; Hu et al., 2015; Lu et al., 2017). We did not observe a significant mediation effect of maternal physical activity on the association of residential surrounding green space with maternal glucose outcomes. Several reasons can be used to explain our findings. Several studies have demonstrated that adjusting for physical activity did not significantly change the association of residential surrounding green space with incident diabetes in the general population (Astellburt et al., 2014; Thiering et al., 2016). Another population study further examined the mediation role of physical activity on the association of residential surrounding green space with incident diabetes and did not observe a significant mediation effect (Dalton et al. , 2016). This evidence indicates that maternal physical activity might be a less important pathway linking residential surrounding green space to maternal glucose outcomes. Additionally, the frequency of maternal physical activity was similar in our studied population, and a binary indicator (0:
Table 3 Results of mediation analyses of residential exposure to PM2.5 on the associations of residential exposure to green space with maternal glucose outcomes. Glucose outcomes
Fasting glucose 1-h glucose 2-h glucose IGT GDM
Direct effect
Indirect effect
Total effect
Proportion
(95% CI)
(95% CI)
(95% CI)
Mediated (%)
−0.014 (−0.021, −0.006)∗ 0.009 (−0.019, 0.036) 0.007 (−0.015, 0.028) 0.000 (−0.006, 0.006) −0.004 (−0.010, 0.001)
−0.062 −0.092 −0.063 −0.010 −0.018
−0.048 −0.101 −0.070 −0.011 −0.013
(−0.063, (−0.153, (−0.111, (−0.023, (−0.026,
−0.034)∗ −0.050)∗ −0.030)∗ 0.001) −0.002)∗
(−0.074, (−0.136, (−0.097, (−0.021, (−0.029,
Abbreviations: IGT: impaired glucose tolerance; GDM: gestational diabetes; RR: relative risk; CI: confidence interval. ∗ P < 0.05. 6
−0.050)∗ −0.049)∗ −0.029)∗ −0.001)∗ −0.008)∗
22.1 (18.4, 27.4)∗ −9.7 (−18.0, −6.5) −11.1 (−23.7, −7.2) −4.2 (−2.4, −1.9) 24.1 (14.6, 53.7)
Environmental Research 176 (2019) 108526
J. Liao, et al.
consent was provided from all participating mothers at enrollment.
prospectively evaluated the associations of residence surrounding green spaces with maternal blood glucose outcomes, which provided powerful epidemiological evidence linking green spaces to maternal blood glucose outcomes. As far as we know, we are the first to report an association of residential exposure to green spaces with gestational IGT. Additionally, based on the well-designed cohort, several known risk factors associated with maternal blood glucose outcomes, such as prepregnancy BMI and passive smoking during pregnancy, were adjusted, which enhanced the credibility of the studied conclusions. Some limitations still need to be acknowledged. We evaluated exposure to green spaces by residential surrounding greenness, which did not offer information to differentiate the types, quality, or use of green spaces. We collected residence information when pregnant women took their first prenatal visit at the studied hospital, which resulted in inaccurate evaluation of residential exposures to green space for pregnant women who changed their residence during pregnancy. The spatial resolution of MODIS data of 500, ×500 m used in this study was relatively coarse, which might cause small differences in analyzing the associations of residential surrounding green spaces with maternal glucose outcomes by different buffer areas. The measurement of physical activity was self-reported, which could not rule out recall bias, and more accurate measurements should be used in further studies.
Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests Acknowledgements We thank all of the pregnant women who participated in the study and all collaborators in the study hospital. Appendix A. Supplementary data Supplementary data related to this article can be found at https:// doi.org/10.1016/j.envres.2019.108526. References Astellburt, T., Feng, X., Kolt, G.S., 2013. Does access to neighbourhood green space promote a healthy duration of sleep? Novel findings from a cross- sectional study of 259 319 australians. BMJ Open 3, 1. Astellburt, T., Feng, X., Kolt, G.S., 2014. Is neighborhood green space associated with a lower risk of type 2 diabetes? Evidence from 267,072 australians. Diabetes Care 37, 197–201. Banay, R.F., James, P., Hart, J.E., Kubzansky, L.D., Spiegelman, D., Okereke, O.I., et al., 2019. Greenness and depression incidence among older women. Environ. Health Perspect. 127, 027001. Cai, S., Tan, S., Gluckman, P.D., Godfrey, K.M., Saw, S., Teoh, O.H., et al., 2017. Sleep quality and nocturnal sleep duration in pregnancy and risk of gestational diabetes mellitus. Sleep 40. Choe, S., Kauderer, S., Eliot, M.N., Glazer, K., Kingsley, S.L., Carlson, L., et al., 2018. Air pollution, land use, and complications of pregnancy. Sci. Total Environ. 645, 1057–1064. Cohen-Cline, H., Turkheimer, E., Duncan, G.E., 2015. Access to green space, physical activity and mental health: a twin study. J. Epidemiol. Community Health 69, 523–529. Dadvand, P., Nieuwenhuijsen, M.J., Esnaola, M., Forns, J., Basagana, X., Alvarezpedrerol, M., et al., 2015a. Green spaces and cognitive development in primary schoolchildren. Proc. Natl. Acad. Sci. U.S.A. 112, 7937–7942. Dadvand, P., Rivas, I., Basagana, X., Alvarez-Pedrerol, M., Su, J., De Castro Pascual, M., et al., 2015b. The association between greenness and traffic-related air pollution at schools. Sci. Total Environ. 523, 59–63. Dadvand, P., Poursafa, P., Heshmat, R., Motlagh, M.E., Qorbani, M., Basagana, X., et al., 2018. Use of green spaces and blood glucose in children; a population-based caspianv study. Environ. Pollut. 243, 1134–1140. Dalton, A.M., Jones, A.P., Sharp, S.J., Cooper, A.J., Griffin, S., Wareham, N.J., 2016. Residential neighbourhood greenspace is associated with reduced risk of incident diabetes in older people: a prospective cohort study. BMC Public Health 16, 1171. Eriksson, C., Hilding, A., Pyko, A., Bluhm, G., Pershagen, G., Ostenson, C., 2014. Longterm aircraft noise exposure and body mass index, waist circumference, and type 2 diabetes: a prospective study. Environ. Health Perspect. 122, 687–694. Facco, F., Grobman, W.A., Reid, K.J., Parker, C.B., Hunter, S., Silver, R.M., et al., 2017. Objectively measured short sleep duration and later sleep midpoint in pregnancy are associated with a higher risk of gestational diabetes. Am. J. Obstet. Gynecol. 217. Fleisch, A.F., Gold, D.R., Rifasshiman, S.L., Koutrakis, P., Schwartz, J., Kloog, I., et al., 2014. Air pollution exposure and abnormal glucose tolerance during pregnancy: the project viva cohort. Environ. Health Perspect. 122, 378–383. Ge, D., Dong, Y., Wang, X., Treiber, F.A., Snieder, H., 2006. The Georgia cardiovascular twin study: influence of genetic predisposition and chronic stress on risk for cardiovascular disease and type 2 diabetes. Twin Res. Hum. Genet. 9, 965–970. 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. Hu, H., Ha, S., Henderson, B.H., Warner, T.D., Roth, J., Kan, H., et al., 2015. Association of atmospheric particulate matter and ozone with gestational diabetes mellitus. Environ. Health Perspect. 123, 853–859. James, P., Banay, R.F., Hart, J.E., Laden, F., 2015. A review of the health benefits of greenness. Current Epidemiology Reports, vol. 2. pp. 131–142. Johnson, B.S., Malecki, K.M., Peppard, P.E., Beyer, K.M.M., 2018. Exposure to neighborhood green space and sleep: evidence from the survey of the health of Wisconsin. Sleep health 4, 413–419. Laurent, O., Wu, J., Li, L., Milesi, C., 2013. Green spaces and pregnancy outcomes in southern California. Health Place 24, 190–195. Lee, J.Y., Lamichhane, D.K., Lee, M., Ye, S., Kwon, J.H., Park, M.S., et al., 2018. Preventive effect of residential green space on infantile atopic dermatitis associated
6. Conclusions Living with higher levels of green space was significantly associated with decreased maternal glucose levels and attenuated risks of incident maternal IGT and GDM. Our findings provide evidence linking green space to better maternal glucose outcomes. Additional studies are needed to further explore the subsequent maternal and child health benefits related to our findings. Conflicts of interest The authors declare they have no actual or potential competing financial interests. Funding This work was supported by the National Natural Science Foundation of China [grant numbers: 91643207, 91743103, and 21437002], the National Key Research and Development Plan [grant numbers: 2016YFC0206700, 2016YFC0206203], and the Fundamental Research Funds for the Central Universities, HUST [grant numbers: 2016YXZD043, 2015ZDTD047, and 2018KFYXMPT00]. Availability of data and materials The datasets for the current study are available from the corresponding author on reasonable request. Authors' contributions Jiaqiang Liao and Xinmei Chen contributed to analyzing the data and drafting the paper. Zhongqiang Cao, Shengwen Liang, Ke Hu, and Yiming Zhang contributed to data collection and data analysis. Wei Xia, Shunqing Xu, Yuanyuan Li, and Bin Zhang contributed to the study concept and design, results interpretation, and revision of the manuscript and approved the final version of the manuscript. Ethics approval and consent to participate The study protocol was approved by the ethics committee of Tongji Medical College, Huazhong University of Science and Technology (No. S152) and Wuhan Children's Hospital (Wuhan Maternal and Child Healthcare Hospital) (No. 2,016,003). Individual written informed 7
Environmental Research 176 (2019) 108526
J. Liao, et al.
2017. Gestational diabetes mellitus and exposure to ambient air pollution and road traffic noise: a cohort study. Environ. Int. 108, 253–260. Roe, J., Thompson, C.W., Aspinall, P., Brewer, M.J., Duff, E.I., Miller, D., et al., 2013. Green space and stress: evidence from cortisol measures in deprived urban communities. Int. J. Environ. Res. Public Health 10, 4086–4103. Roe, J., Aspinall, P., Thompson, C.W., 2017. Coping with stress in deprived urban neighborhoods: what is the role of green space according to life stage? Front. Psychol. 8, 1760. Sallis, J.F., Floyd, M.F., Rodriguez, D.A., Saelens, B.E., 2012. Role of built environments in physical activity, obesity, and cardiovascular disease. Circulation 125, 729–737. Shen, Y., Lung, S.C., 2016. Can green structure reduce the mortality of cardiovascular diseases. Sci. Total Environ. 566, 1159–1167. Sorensen, M., Andersen, Z.J., Nordsborg, R.B., Becker, T., Tjonneland, A., Overvad, K., et al., 2013. Long-term exposure to road traffic noise and incident diabetes: a cohort study. Environ. Health Perspect. 121, 217–222. Thiering, E., Markevych, I., Bruske, I., Fuertes, E., Kratzsch, J., Sugiri, D., et al., 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. Toshihiro, M., Saito, K., Takikawa, S., Takebe, N., Onoda, T., Satoh, J., 2008. Psychosocial factors are independent risk factors for the development of type 2 diabetes in Japanese workers with impaired fasting glucose and/or impaired glucose tolerance. Diabet. Med. 25, 1211–1217. Trigueromas, M., Donairegonzalez, D., Seto, E., Valentin, A., Martinez, D., Smith, G., et al., 2017. Natural outdoor environments and mental health: stress as a possible mechanism. Environ. Res. 159, 629–638. Wang, H., Leng, J., Li, W., Wang, L., Zhang, C., Liu, H., et al., 2017. Sleep duration and quality, and risk of gestational diabetes mellitus in pregnant Chinese women. Diabet. Med. 34, 44–50. WHO, 2010. Urban Planning, Environment and Health: from Evidence to Policy Action. World Health Organization, Geneva. WHO, 2016. Urban Green Spaces and Health: A Review of Evidence. World Health Organization, Copenhagen, Denmark. Wong, M., Wong, W.H.S., Cheung, Y., 2007. Fetal myocardial performance in pregnancies complicated by gestational impaired glucose tolerance. Ultrasound Obstet. Gynecol. 29, 395–400. Xu, X., Liu, C., Xu, Z., Tzan, K., Zhong, M., Wang, A., et al., 2011. Long-term exposure to ambient fine particulate pollution induces insulin resistance and mitochondrial alteration in adipose tissue. Toxicol. Sci. 124, 88–98. Yang, B., Markevych, I., Heinrich, J., Bowatte, G., Bloom, M.S., Guo, Y., et al., 2018a. Associations of greenness with diabetes mellitus and glucose-homeostasis markers: the 33 communities Chinese health study. Int. J. Hyg Environ. Health. Yang, B.Y., Qian, Z., Li, S., Chen, G., Bloom, M.S., Elliott, M., et al., 2018b. 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. Zou, G., 2004. A modified Poisson regression approach to prospective studies with binary data. Am. J. Epidemiol. 159, 702–706.
with prenatal air pollution exposure. Int. J. Environ. Res. Public Health 15. Liang, F., Yang, X., Liu, F., Li, J., Xiao, Q., Chen, J., et al., 2019. Long-term exposure to ambient fine particulate matter and incidence of diabetes in China: a cohort study. Environ. Int. 126, 568–575. Liao, J., Yu, H., Xia, W., Zhang, B., Lu, B., Cao, Z., et al., 2018. Exposure to ambient fine particulate matter during pregnancy and gestational weight gain. Environ. Int. 119, 407–412. Lodovici, M., Bigagli, E., 2011. Oxidative stress and air pollution exposure. J. Toxicol. 2011 487074. Lu, M., Wang, P., Cheng, T., Yang, C., Yan, Y., 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., Hennig, F., Matthiessen, C., Ohlwein, S., Icks, A., Moebus, S., et al., 2018. Air pollution and glucose metabolism: an analysis in non-diabetic participants of the heinz nixdorf recall study. Environ. Health Perspect. 126, 047001. Margaritis, E., Kang, J., 2017. Relationship between green space-related morphology and noise pollution. Ecol. Indicat. 72, 921–933. Markevych, I., Fuertes, E., Tiesler, C.M., Birk, M., Bauer, C.P., Koletzko, S., et al., 2014. Surrounding greenness and birth weight: results from the giniplus and lisaplus birth cohorts in munich. Health Place 26, 39–46. Metzger, B.E., Gabbe, S.G., Persson, B., Buchanan, T.A., Catalano, P.A., Damm, P., et al., 2010. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care 33 e97; author reply e98. Mooy, J.M., De Vries, H., Grootenhuis, P.A., Bouter, L.M., Heine, R.J., 2000. Major stressful life events in relation to prevalence of undetected type 2 diabetes : the hoorn study. Diabetes Care 23, 197–201. Muller, G., Harhoff, R., Rahe, C., Berger, K., 2018. Inner-city green space and its association with body mass index and prevalent type 2 diabetes: a cross-sectional study in an urban German city. BMJ Open 8. Nasiri-Amiri, F., Bakhtiari, A., Faramarzi, M., Adib Rad, H., Pasha, H., 2016. The association between physical activity during pregnancy and gestational diabetes mellitus: a case-control study. Int. J. Endocrinol. Metab. 14, e37123. Nichani, V., Dirks, K., Burns, B., Bird, A., Morton, S., Grant, C., 2017. Green space and pregnancy outcomes: evidence from growing up in New Zealand. Health Place 46, 21–28. Nieuwenhuijsen, M.J., Khreis, H., Triguero-Mas, M., Gascon, M., Dadvand, P., 2017. Fifty shades of green: pathway to healthy urban living. Epidemiology 28, 63–71. Orioli, R., Antonucci, C., Scortichini, M., Cerza, F., Marando, F., Ancona, C., et al., 2019. Exposure to residential greenness as a predictor of cause-specific mortality and stroke incidence in the rome longitudinal study. Environ. Health Perspect. 127, 027002. Ostlund, I., Hanson, U., Bjorklund, A., Hjertberg, R., Eva, N., Nordlander, E., et al., 2003. Maternal and fetal outcomes if gestational impaired glucose tolerance is not treated. Diabetes Care 26, 2107–2111. Park, S.K., Adar, S.D., Oneill, M.S., Auchincloss, A.H., Szpiro, A.A., Bertoni, A.G., et al., 2015. Long-term exposure to air pollution and type 2 diabetes mellitus in a multiethnic cohort. Am. J. Epidemiol. 181, 327–336. Pedersen, M., Olsen, S.F., Halldorsson, T.I., Zhang, C., Hjortebjerg, D., Ketzel, M., et al.,
8
Contents lists available at ScienceDirect
Environmental Research journal homepage: www.elsevier.com/locate/envres
Effect of residential exposure to green space on maternal blood glucose levels, impaired glucose tolerance, and gestational diabetes mellitus
T
Jiaqiang Liaoa,1, Xinmei Chena,1, Shunqing Xua, Yuanyuan Lia, Bin Zhangb, Zhongqiang Caob, Yiming Zhangb, Shengwen Liangc, Ke Huc, Wei Xiaa,∗ a Key Laboratory of Environment and Health, Ministry of Education & Ministry of Environmental Protection, And State Key Laboratory of Environmental Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, People's Republic of China b Women and Children Medical and Healthcare Center of Wuhan, Wuhan, Hubei, People's Republic of China c Wuhan Environmental Monitoring Center, Wuhan, Hubei Province, 430000, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Green spaces Maternal blood glucose levels Impaired glucose tolerance Gestational diabetes mellitus
Background: Residential surrounding green spaces can affect human health. However, limited studies have examined their impacts on maternal blood glucose homeostasis outcomes. Objective: We examined the associations of residential exposure to green space with maternal blood glucose levels, gestational impaired glucose tolerance (IGT), and gestational diabetes mellitus (GDM). Methods: Pregnant women were recruited from a prospective birth cohort between October 2012 and September 2015. Exposure to green space was calculated as the mean value of the normalized difference vegetation index (NDVI) within a 300-m circular buffer area surrounding each residence. Maternal glucose was measured between 24 and 28 weeks of gestation, and gestational IGT and GDM were diagnosed using valid methods. We estimated the associations of residential NDVI with maternal glucose levels using multiple linear regression models with adjustment for age, education, BMI, passive smoking during pregnancy, parity, season of conception, income, and urbancity. We estimated the relative risks of residential NDVI with IGT and GDM using a generalized estimating equation model with modified Poisson regression. The mediation effects of residential exposure to air pollution and maternal physical activity were assessed using causal mediation analysis. Results: Of 6807 pregnant women, 751 (11.3%) and 604 (8.8%) were diagnosed with IGT and GDM, respectively. One SD increment of residential NDVI was associated with a decrease of 0.06 mmol/L (95% CI: −0.07, −0.05), 0.09 mmol/L (95% CI: −0.13, −0.05), and 0.06 mmol/L (95% CI: −0.09, −0.03) in maternal fasting glucose levels, 1-h glucose levels, and 2-h glucose levels, respectively, as well as reduced risks of incident IGT (RR: 0.92, 95% CI: 0.86, 0.99) and GDM (RR: 0.85, 95% CI: 0.79, 0.92). The association between residential NDVI and maternal fasting glucose levels was partly mediated by maternal exposure to PM2.5. Conclusion: Living with higher levels of green space was significantly associated with decreased maternal glucose levels and attenuated risks of incident maternal IGT and GDM. Our findings provide evidence linking green space to better maternal glucose outcomes. More studies are needed to further explore the maternal and child health benefits related to our findings.
1. Introduction Urban green spaces, such as parks and sidewalks, have been widely linked to improved health and well-being (WHO, 2016). Population studies have found that higher levels of surrounding green spaces were associated with reduced risks of cardiovascular disease mortality,
incident stroke, and incident depressive symptoms in younger and older populations (Banay et al., 2019; Orioli et al., 2019; Shen and Lung, 2016). Additionally, some evidence suggests that higher levels of residential surrounding green spaces may be protective against adverse pregnancy outcomes, such as preterm birth and low birth weight (Laurent et al., 2013; Markevych et al., 2014; Nichani et al., 2017).
Abbreviations: FGL, fast glucose level; GL1, 1-h glucose level; GL2, 2-h glucose level ∗ Corresponding author. School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, People's Republic of China. E-mail address: [email protected] (W. Xia). 1 These authors contributed equally. https://doi.org/10.1016/j.envres.2019.108526 Received 19 March 2019; Received in revised form 30 May 2019; Accepted 5 June 2019 Available online 06 June 2019 0013-9351/ © 2019 Elsevier Inc. All rights reserved.
Environmental Research 176 (2019) 108526
J. Liao, et al.
Fig. 1. Maps of NDVI (2013) and geospatial distributions of pregnant women in the studied area.aHigher values represent greater vegetative cover.
incident type 2 diabetes (Muller et al., 2018; B Yang et al., 2018). However, as far as we know, only one study examined the association of residential surrounding green space with GDM, and observed a nonsignificant association (Choe et al., 2018). No study has been conducted to examine the impact of residential surrounding green space on gestational impaired glucose tolerance (IGT), which has been widely reported to be associated with elevated risk of adverse maternal outcomes such as caesarean section rate, prematurity and infant diagnosis with macrosomia (Ostlund et al., 2003; Wong et al., 2007). Additionally, the potential mechanisms underlying these associations are not reported. Therefore, we hypothesized that exposure to higher levels of residential surrounding green spaces could benefit maternal glucose levels, decreasing the risks of incident maternal IGT and GDM. We examined these assumptions based on an ongoing prospective birth cohort. To explore the potential mechanisms underlying these
Limited evidence has been reported regarding the associations of exposure to green space with maternal blood glucose-homeostasis outcomes. Exposure to higher levels of air pollution and noise, and lower levels of physical activity were reported to be associated with elevated risks of incident diabetes and gestational diabetes mellitus (GDM) (Nasiri-Amiri et al., 2016; Sorensen et al., 2013; Wang et al., 2017; B Yang et al., 2018), while exposure to higher levels of residential surrounding green spaces is known to promote health by contributing to lower levels of air pollution and noise exposure, increased odds of physical activity, and reduced likelihood of stress and depression (Cohen-Cline et al., 2015; James et al., 2015; Margaritis and Kang, 2017; Nieuwenhuijsen et al., 2017; Roe et al. 2013; Sallis et al., 2012). A few studies have reported that exposure to higher levels of residential surrounding green spaces was associated with lower fasting glucose levels and 1-h glucose levels, and lower risks of insulin resistance and
2
Environmental Research 176 (2019) 108526
J. Liao, et al.
diabetes before conception. Plasma glucose levels were measured using a Roche Modular P800 automated biochemistry analyzer at the Wuhan Women and Children Medical and Healthcare Center. According to World Health Organization (WHO) 2006 recommendations, pregnant women were diagnosed with IGT if both of the following conditions were satisfied: 1) fasting blood glucose level < 7.0 mmol/L; and 2) 2-h blood glucose level ≥7.8 and < 11.10 mmol/L. According to the International Association of Diabetes and Pregnancy Study Group (Metzger et al., 2010), pregnant women were diagnosed with GDM if any of the following conditions were satisfied: 1) fasting blood glucose level 5.1 mmol/L; 2) 1-h blood glucose level 10 mmol/L; or 3) 2-h blood glucose level 8.5 mmol/L.
associations, we further examined the mediation effects of maternal exposure to air pollution and maternal physical activity. 2. Methods 2.1. Population and study design The studied pregnant women were from a prospective birth cohort conducted in Wuhan, China between October 2012 and September 2015. Pregnant women were invited to participate in the study when they came to the Women and Children Medical and Healthcare Center of Wuhan for their first prenatal care visit. The criteria for recruitment were residence in Wuhan City, with singleton pregnancy, and without diabetes before pregnancy. Then, the enrolled pregnant women were followed for the duration of the pregnancy. The research protocol was approved by the Institutional Review Board of Tongji Medical School, Huazhong University of Science and Technology. All participants provided written informed consent at enrolment. Based on the conditions described above, we initially included 6883 pregnant women in the study. We excluded pregnant women who had missing data on annual household income (n = 70), educational background (n = 1), and prepregnancy BMI (n = 5). Therefore, we had 6807 pregnant women in the final analysis.
3. Statistical analysis 3.1. Main analyses We explored associations of residential surrounding NDVI and maternal glucose levels by scatter plots. We then estimated the changes and 95% confidence intervals (CIs) of maternal blood glucose levels (fasting glucose levels, 1-h glucose levels, and 2-h glucose levels) associated with increases of residential surrounding green spaces (one SD increments and quartiles) using linear regression models. Since incidences of IGT and GDM in our study population were relatively common (nearly 10% or more), we estimated the relative risks (RRs) and 95% CIs of incident IGT and GDM associated with increases in residential surrounding green spaces (one SD increments and quartiles) using generalized estimating equation (GEE) models with modified Poisson regression, which corrected the standardized errors to better approximates of RRs for common outcomes (Zou, 2004). We modeled the median value within each stratum of residential NDVI quartiles as a continuous variable entering multiple regression models, and the GEE models and P values for trends were obtained by conducting Wald chisquared tests. We used directed acyclic graph (DAG) methods to determine the covariates in multiple adjusted models. DAG-defined covariates are listed in supplementary mateial Figure S1. The following covariates were included: residence area (rural, urban), maternal age (≤24, 25–35, ≥36 years), maternal education (≤9, 10–12, ≥13 years), maternal household income (< 30,000, 30,000–50000, 50,000–1000000, ≥100,000 yuan), maternal prepregnancy BMI (underweight: < 18.5, normal: 18.5–24, overweight or obesity: 24 kg/m2), passive smoking during pregnancy (yes, no), maternal parity (1: nulliparous women, ≥2: parous women), and season of conception (spring, summer, autumn, and winter). Since few pregnant women reported active smoking during pregnancy (n = 37, 0.54%), we did not adjust for active smoking in the multivariable analyses.
2.2. Data collection We collected the maternal-related characteristics such as residential address, maternal age, and maternal education, passive smoking during pregnancy, maternal weight and prepregnancy height using a questionnaire interview at enrolment. Information on the date of birth, infant's sex, and infant's birth weight was obtained from medical records. Gestational age was estimated by the 1st-trimester ultrasound and the correct date of the last menstrual period (LMP). The maternal prepregnancy BMI was evaluated using maternal prepregnancy weight in kg divided by the square of the maternal height in meters. 2.3. Evaluation of maternal exposure to green space We assessed the exposure to green space by calculating a satellitebased normalized difference vegetation index (NDVI) for pregnant women based on their residential addresses, which is a normalized score ranging from −1 to 1, with lower values representing water bodies or barren areas and higher values representing photosynthetically active vegetation. We generated the data from remotely sensed daily averaged NDVI values covering the study area using MODIS raster data with a spatial resolution of 500 m× 500 m, which was collected from the International Scientific & Technical Data Mirror Site, Computer Network Information Center, Chinese Academy of Sciences (http://www.gscloud.cn). Since NDVI was unlikely to change over the gestational period for each pregnant woman, we only calculated the NDVI for the conception year of each pregnant woman. We selected the daily NDVI maps with the lowest rate of cloud coverage during the spring season (most green season during each year) in Wuhan city, which were 1st May, 1st May, 1st Jun, and 1st April for pregnant women whose conception years were 2012, 2013, 2014, and 2015, respectively. The geospatial distribution of residential addresses of pregnant women and the map of NDVI in Wuhan city were listed in Fig. 1. Based on existing evidence, we evaluated the residential exposure to green space by calculating the averaged NDVI values within a 300-m buffer area surrounding the residential address for each pregnant woman.
3.2. Mediation analyses of air pollution and maternal physical outdoor activities To further explore the mechanisms underlying the associations of residential exposure to green space with maternal blood glucose outcomes, we examined the mediation effects of maternal exposure to air pollution and maternal physical outdoor activity on these associations. We measured air pollution by evaluating residential exposure to fine particulate matter (PM2.5). Maternal daily exposures to PM2.5 from the date of the LMP to the date of the OGTT examination were estimated using a spatial-temporal land use regression (LUR) model adjusted for short-term (weeks) and long-term (seasons) variations (within 12×4 knots of natural cubic splines in each year). The details of the evaluation were introduced in our previous study (Liao et al., 2018). In brief, we collected data from 20 ground-level surveillance stations recording daily average PM2.5 concentration across the studied area, which were used as outcome variables in the spatial-temporal LUR model. We included the following predictor variables in our spatial-temporal LUR
2.4. Measurements of maternal glucose levels, diagnosis of IGT and GDM A 75-g oral glucose tolerance test (OGTT) was conducted within gestational weeks 24–28 for the pregnant women who did not have 3
Environmental Research 176 (2019) 108526
J. Liao, et al.
resided in the lowest quartile, respectively.
model: residential areas of commercial and industrial land use with a circular buffer of 5 km, residential areas of farmland use with a circular buffer of 15 km, residential areas of lake with a circular buffer of 10 km, lengths of major highways with a circular buffer of 5 km around the residential address, numbers of pollutant-related enterprises with a circular buffer of 15 km, and daily average temperature. The 10-fold out-of-sample cross-validation (CV) indicated that our method had good ability to predict maternal daily PM2.5 exposures (model explained variance R2: 71.5%, average 10-fold CV R2: 72.6%). Since most of the pregnant women had physical activities equal to or above 3 times per week (75.9%), we reclassified the variable of maternal physical activity into a binary variable of physical inactivity (0: ≥3 times per week, 1: ≤3 times per week). We conducted the causal mediation analyses by examining a continuous mediator of average PM2.5 concentration from the date of the LMP to the date of the OGTT examination and a binary mediator of physical inactivity on the associations of residential exposure to green space with maternal blood glucose outcomes.
4.2. Results of mediation analyses Exposure to higher levels of residential surrounding green space was associated with lower levels of residential PM2.5 exposure and reduced risks of maternal physical inactivity (Supplementary Material Table S1). After adjusting for covariates, a one SD increment of residential NDVI was associated with a decrease of −14.87 μg/m3 (95% CI: −15.40, −14.33) in maternal PM2.5 exposure and a lower risk of 0.87 (95% CI: 0.80, 0.95) in maternal physical inactivity. Similar associations were observed in the associations of maternal PM2.5 exposure and maternal physical inactivity with residential NDVI by quartiles. Exposure to higher levels of maternal PM2.5 was associated with increased maternal fasting glucose levels and 1-h glucose levels and higher risk of incident GDM (Supplementary Material Table S2). After adjusting for covariates, one 10 μg/m3 increment of maternal PM2.5 exposure was associated with an increase of 0.02 mmol/L (95% CI: 0.02, 0.03) in maternal fasting blood glucose levels, an increase of 0.02 mmol/L (95% CI: 0.00, 0.03) in 1-h maternal glucose levels, and an elevated risk of 1.05 (95% CI: 1.02, 1.09) in incident GDM. No associations of maternal physical inactivity with maternal glucose homeostatic outcomes were observed. The results of mediation analyses further indicated that maternal exposure to PM2.5 mediated the association between residential exposure to green space and maternal fasting glucose levels (Table 3). After adjusting for covariates, a one SD increment of residential NDVI contributed to a decrease of 0.05 mmol/L (95% CI: −0.06, −0.03) in maternal fasting glucose level, and 22.1% (95% CI: 18.4%, 27.4%) of this association can be explained by increased maternal PM2.5 concentration. No significant mediation effects of maternal PM2.5 exposure on other maternal blood glucose outcomes were observed.
3.3. Sensitivity analyses To further evaluate the robustness of our conclusions, we conducted several sensitivity analyses to examine the associations of residential surrounding greenness with maternal glucose outcomes: (1) evaluating the residential surrounding greenness by a buffer area of 1000 m; and (2) evaluating the residential surrounding greenness by simply extracting cell values to the residence location. All statistical analyses were conducted by SAS 9.4 and STATA 13.0. The statistical significance level was set to 0.05. 4. Results Of 6807 pregnant women, 11.03% were diagnosed with IGT, and 8.87% were diagnosed with GDM (Table 1). Among them, those who came from urban areas, from families with lower levels of annual household income, had an older age, were less educated, were overweight or obese prepregnancy, and were multiparous were more likely to be diagnosed with IGT or GDM.
4.3. Results of sensitivity analyses The associations of residential exposure to green space with maternal blood glucose levels, incident IGT, and incident GDM were similar to the results of the main analyses when we measured residential surrounding NDVI by circular buffer areas of 1000 m, and simply extracted cell values to the residence location (Supplementary Material Table S3-S4).
4.1. Results of main analyses The scatter plots (Fig. 2(a)-(c)) indicated that residential surrounding NDVI was negatively associated with maternal fasting glucose level, 1-h glucose level, and 2-h glucose level. Table 2 depicts the associations of residential exposure to green space with maternal glucosehomeostasis outcomes. After adjusting for the covariates, a one SD increment of residential NDVI was associated with lower maternal fasting glucose levels (mean difference: −0.06 mmol/L, 95% CI: −0.08, 0.05), 1-h glucose levels (mean difference: −0.10 mmol/L, 95% CI: −0.14, −0.06), and 2-h glucose levels (mean difference: −0.07 mmol/L, 95% CI: −0.10, −0.04), and lower risks of incident IGT (RR: 0.92, 95% CI: 0.86, 0.99), and incident GDM (RR: 0.85, 95% CI: 0.78, 0.93). The doseresponse associations of quartile measurements of residential NDVI with maternal blood glucose outcomes were observed (all values of P for trend < 0.05). Pregnant women who resided in higher quartiles of residential NDVI had lower blood glucose levels. For example, compared with pregnant women who resided in the lowest quartile of residential NDVI, the adjusted mean differences in fasting glucose levels were −0.10 mmol/L (95% CI: −0.14, −0.07), −0.14 mmol/L (95% CI: −0.18, −0.11), and −0.17 mmol/L (95% CI: −0.21, −0.14) in those who resided in the second, third, and fourth quartiles of residential NDVI, respectively. Furthermore, pregnant women who resided in higher quartiles of residential NDVI had lower risks of incident IGT and GDM. The adjusted RRs for incident IGT and GMD were 0.77 (95% CI: 0.62, 0.96) and 0.66 (95% CI: 0.52, 0.84) for pregnant women who resided in the highest quartile of residential NDVI and those who
5. Discussion We examined the associations of residential exposure to green space with maternal blood glucose levels, incident gestational IGT, and incident GDM based on a prospective birth cohort. We have reported several key findings. First, we consistently observed that exposure to higher levels of residential surrounding green spaces was associated with lower levels of maternal fasting glucose, 1-h glucose, and 2-h glucose as well as reduced risks of incident IGT and GDM. Second, we demonstrated that reduced levels of maternal exposure to PM2.5 mediated 18.4%–27.4% of the association between residential exposure to green space and maternal fasting glucose levels. This study provides preliminary epidemiological evidence linking exposure to higher levels of residential surrounding green spaces to lower risks of gestational IGT. Given the extraordinary important roles of gestational IGT in elevating risks of maternal caesarean rate, prematurity, large for gestational age, and macrosomic infants (Ostlund et al., 2003; Wong et al., 2007), our findings were important to further determine the mechanisms of green space on maternal and child health. Our findings were consistent with the reported evidence linking green space to incident diabetes in the general population. A population study examined the association of neighborhood green space and the rate of type 2 diabetes mellitus (T2DM) in 267,072 Australians, which demonstrated that participants from greener neighborhoods had lower 4
Environmental Research 176 (2019) 108526
J. Liao, et al.
Table 1 Baseline characteristics of studied pregnant women stratified by IGT status and GDM status (n (%)). Characteristics
All
GDM
No GDM
IGT
No IGT
Total Residential areas Rural Urban Maternal age, years ≤24 25–29 30–34 ≥35 Education, years ≤9 10–12 ≥13 Annual household income, yuan < 30,000 30,000–50,000 50,000–100,000 > 100,000 Prepregnancy BMI, kg/m2 Underweight(< 18.5) Normal(18.5–24.0) Overweight or obesity(≥24) Parity 1: nulliparous women ≥2: parous women Passive smoking during pregnancy No Yes Season of LMP Spring Summer Autumn Winter
6807 (100.00)
604 (8.87)
6203 (91.13)
751 (11.03)
6056 (88.97)
1253 (18.41) 5554 (81.59)
101 (16.72) 503 (83.28)
1152 (18.57) 5051 (81.43)
128 (17.04) 623 (82.96)
1125 (18.58) 4931 (81.42)
565 (8.3) 4039 (59.34) 1827 (26.84) 376 (5.52)
27 (4.47) 278 (46.03) 217 (35.93) 82 (13.58)
538 (8.67) 3761 (60.63) 1610 (25.96) 294 (4.74)
23 (3.06) 379 (50.47) 260 (34.62) 89 (11.85)
542 (8.95) 3660 (60.44) 1567 (25.88) 287 (4.74)
455 (6.68) 1095 (16.09) 5257 (77.23)
61 (10.1) 115 (19.04) 428 (70.86)
394 (6.35) 980 (15.8) 4829 (77.85)
55 (7.32) 115 (15.31) 581 (77.36)
400 (6.61) 980 (16.18) 4676 (77.21)
726 (10.67) 1442 (21.18) 2889 (42.44) 1750 (25.71)
73 (12.09) 143 (23.68) 256 (42.38) 132 (21.85)
653 (10.53) 1299 (20.94) 2633 (42.45) 1618 (26.08)
81 (10.79) 168 (22.37) 311 (41.41) 191 (25.43)
645 (10.65) 1274 (21.04) 2578 (42.57) 1559 (25.74)
1401 (20.58) 4592 (67.46) 814 (11.96)
65 (10.76) 390 (64.57) 149 (24.67)
1336 (21.54) 4202 (67.74) 665 (10.72)
103 (13.72) 498 (66.31) 150 (19.97)
1298 (21.43) 4094 (67.6) 664 (10.96)
6106 (89.7) 701 (10.3)
503 (83.28) 101 (16.72)
5603 (90.33) 600 (9.67)
639 (85.09) 112 (14.91)
5467 (90.27) 589 (9.73)
5131 (75.38) 1676 (24.62)
458 (75.83) 146 (24.17)
4673 (75.33) 1530 (24.67)
562 (74.83) 189 (25.17)
4569 (75.45) 1487 (24.55)
955 (14.03) 1834 (26.94) 2403 (35.3) 1615 (23.73)
67 (11.09) 176 (29.14) 225 (37.25) 136 (22.52)
888 (14.32) 1658 (26.73) 2178 (35.11) 1479 (23.84)
116 168 265 202
839 (13.85) 1666 (27.51) 2138 (35.30) 1413 (23.33)
(15.45) (22.37) (35.29) (26.9)
Abbreviations: IGT: impaired glucose tolerance; GDM: gestational diabetes.
risks of T2DM (Astellburt et al., 2014). Similar studies have also reported that individuals living in the greener neighborhoods had lower risks of developing diabetes (Dalton et al. 2016; Muller et al., 2018). For children, those who spent more time in green spaces had lower fasting blood glucose levels (Dadvand et al., 2018). Another study that examined the associations of green space and markers related to blood glucose in adults reported that exposure to higher levels of residential surrounding greenness was associated with lower fasting and 2-h glucose levels, lower 2-h insulin levels, lower insulin resistance and higher
β -cell function (B Yang et al., 2018). As far as we know, only one study has examined the association of residential surrounding greenness with incident GDM, which found a nonsignificant association (Choe et al., 2018). In this study, we consistently demonstrated that pregnant women with higher levels of residential surrounding green spaces had lower fasting glucose levels, 1-h glucose levels, and 2-h glucose levels and lower risks of incident gestational IGT and GDM. Numerous studies have highlighted the consistent mediation role of reduced levels of air pollution in the associations of exposure to green
Fig. 2. The scatter plots of residential surrounding NDVI with maternal fasting glucose level, 1-h glucose level, and 2-h glucose levela. aRed lines represent the fitted lines by linear regression, and the green areas represent the 95% confidence intervals of predicted lines. 5
Environmental Research 176 (2019) 108526
J. Liao, et al.
Table 2 Adjusted changes of residential exposure to green space with maternal glucose levels, IGT, and GDM. Measurements of green space
Changes of maternal glucose levels (95% CI), mmol/La Fasting glucose
One SD (0.15) increment, Categorized into quartiles Q1: −0.09–0.13 Q2: 0.14–0.32 Q3: 0.33–0.40 Q4: ≥0.41 P for trend
−0.06 (−0.08, −0.05)
1-h glucose ∗
Reference −0.10 (−0.14, −0.07)∗ −0.14 (−0.18, −0.11)∗ −0.17 (−0.21, −0.14)∗ < .001
RR (95% CI) for IGTb
RR (95% CI) for GDMb
2-h glucose ∗
−0.10 (−0.14, −0.06)
−0.07 (−0.10, −0.04)∗
0.92 (0.86, 0.99)∗
0.85 (0.78, 0.93)∗
Reference −0.10 (−0.21, 0.02) −0.21 (−0.32, −0.10)∗ −0.27 (−0.39, −0.15)∗ < .001
Reference −0.05 (−0.14, 0.04) −0.09 (−0.18, 0.00) −0.19 (−0.28, −0.10)∗ < .001
Reference 0.95 (0.77, 1.17) 0.93 (0.75, 1.14) 0.77 (0.62, 0.96)∗ 0.022
Reference 0.98 (0.79, 1.22) 0.74 (0.59, 0.93)∗ 0.66 (0.52, 0.84)∗ < .001
Abbreviations: IGT: impaired glucose tolerance; GDM: gestational diabetes; RR: relative risk; CI: confidence interval. ∗ P < 0.05. a The changes in glucose levels and 95% CIs were estimated by multiple linear regression models, adjusting for age, education years, BMI, passive smoking during pregnancy, parity, season of conception, income, and urban areas.. b The RRs and 95% CIs of incident IGT were estimated by a generalized estimating equation model with modified Poisson regression, adjusting for age, education years, BMI, passive smoking during pregnancy, parity, season of conception, income, and urban areas. The RRs and 95% CIs of incident GDM were estimated by a generalized estimating equation model with modified Poisson regression, adjusting for age, education years, BMI, passive smoking during pregnancy, parity, season of conception, income, and urban areas..
≥3 times per week, 1: ≤3 times per week) might be not enough to classify the different levels of maternal physical activity in our study population. More mechanisms underlying the associations of residential exposure to green space with maternal blood glucose outcomes should be further explored. For example, reduced levels of noise exposure might be a potential mechanism linking green space to maternal blood glucose outcomes. Experimental evidence indicated that higher levels of urban green spaces were associated with lower levels of noise exposure (Margaritis and Kang, 2017; WHO, 2010), while higher levels of noise exposure were associated with elevated risk of incident diabetes and GDM (Eriksson et al., 2014; Margaritis and Kang, 2017; Pedersen et al., 2017). Second, disruptions of sleep duration and poorer sleep quality are potential mechanisms underlying the associations of exposure to green space with maternal blood glucose outcomes. Population evidence indicated that exposure to higher levels of green space was associated with better sleep duration (Astellburt et al., 2013; Johnson et al., 2018), while short sleep duration and poor levels of sleep quality were reported to be associated with higher risks of incident GDM (Cai et al., 2017; Facco et al., 2017; Wang et al., 2017). Finally, reduced odds of major stressful life events might mediate the associations of green space with maternal blood glucose outcomes. Exposure to higher levels of urban neighborhood green spaces was associated with lower levels of stress (Roe et al., 2013; Roe et al., 2017; Trigueromas et al., 2017). Exposure to higher levels of stress was associated with higher risk of incident diabetes (Ge et al., 2006; Mooy et al., 2000; Toshihiro et al., 2008). This study has several strengths. The main strength is that the studied population was from a prospective birth cohort and all pregnant women were suggested to participate in OGTT for screening GDM from 24 to 28 weeks of gestation, which offered us an opportunity to examine multiple maternal blood glucose outcomes with exposures. We
space with adverse health outcomes ( Dadvand et al., 2015a, 2015b; Lee et al., 2018; Nieuwenhuijsen et al., 2017; B Yang et al., 2018). In this study, we further found that reduced levels of maternal PM2.5 exposure partly mediated (from 19.4% to 27.7%) the association of residential exposure to green space with maternal fasting glucose levels. Mechanistic studies indicated that exposure to higher levels of ambient PM2.5 induced oxidative stress and contributed to insulin resistance (Haberzettl et al., 2016; Lodovici and Bigagli, 2011; Xu et al., 2011). Population evidence also reported consistent associations of PM2.5 exposure with glucose outcomes in the general population or in pregnant women. In the general population, exposure to higher levels of ambient PM2.5 was associated with higher fasting glucose levels and elevated risk of incident diabetes (Liang et al., 2019; Lucht et al., 2018; Park et al., 2015; BY Yang et al., 2018). In pregnant women, exposure to higher levels of ambient PM2.5 was associated with higher fasting OGTT glucose levels and elevated risks of incident IGT and GDM (Fleisch et al., 2014; Hu et al., 2015; Lu et al., 2017). We did not observe a significant mediation effect of maternal physical activity on the association of residential surrounding green space with maternal glucose outcomes. Several reasons can be used to explain our findings. Several studies have demonstrated that adjusting for physical activity did not significantly change the association of residential surrounding green space with incident diabetes in the general population (Astellburt et al., 2014; Thiering et al., 2016). Another population study further examined the mediation role of physical activity on the association of residential surrounding green space with incident diabetes and did not observe a significant mediation effect (Dalton et al. , 2016). This evidence indicates that maternal physical activity might be a less important pathway linking residential surrounding green space to maternal glucose outcomes. Additionally, the frequency of maternal physical activity was similar in our studied population, and a binary indicator (0:
Table 3 Results of mediation analyses of residential exposure to PM2.5 on the associations of residential exposure to green space with maternal glucose outcomes. Glucose outcomes
Fasting glucose 1-h glucose 2-h glucose IGT GDM
Direct effect
Indirect effect
Total effect
Proportion
(95% CI)
(95% CI)
(95% CI)
Mediated (%)
−0.014 (−0.021, −0.006)∗ 0.009 (−0.019, 0.036) 0.007 (−0.015, 0.028) 0.000 (−0.006, 0.006) −0.004 (−0.010, 0.001)
−0.062 −0.092 −0.063 −0.010 −0.018
−0.048 −0.101 −0.070 −0.011 −0.013
(−0.063, (−0.153, (−0.111, (−0.023, (−0.026,
−0.034)∗ −0.050)∗ −0.030)∗ 0.001) −0.002)∗
(−0.074, (−0.136, (−0.097, (−0.021, (−0.029,
Abbreviations: IGT: impaired glucose tolerance; GDM: gestational diabetes; RR: relative risk; CI: confidence interval. ∗ P < 0.05. 6
−0.050)∗ −0.049)∗ −0.029)∗ −0.001)∗ −0.008)∗
22.1 (18.4, 27.4)∗ −9.7 (−18.0, −6.5) −11.1 (−23.7, −7.2) −4.2 (−2.4, −1.9) 24.1 (14.6, 53.7)
Environmental Research 176 (2019) 108526
J. Liao, et al.
consent was provided from all participating mothers at enrollment.
prospectively evaluated the associations of residence surrounding green spaces with maternal blood glucose outcomes, which provided powerful epidemiological evidence linking green spaces to maternal blood glucose outcomes. As far as we know, we are the first to report an association of residential exposure to green spaces with gestational IGT. Additionally, based on the well-designed cohort, several known risk factors associated with maternal blood glucose outcomes, such as prepregnancy BMI and passive smoking during pregnancy, were adjusted, which enhanced the credibility of the studied conclusions. Some limitations still need to be acknowledged. We evaluated exposure to green spaces by residential surrounding greenness, which did not offer information to differentiate the types, quality, or use of green spaces. We collected residence information when pregnant women took their first prenatal visit at the studied hospital, which resulted in inaccurate evaluation of residential exposures to green space for pregnant women who changed their residence during pregnancy. The spatial resolution of MODIS data of 500, ×500 m used in this study was relatively coarse, which might cause small differences in analyzing the associations of residential surrounding green spaces with maternal glucose outcomes by different buffer areas. The measurement of physical activity was self-reported, which could not rule out recall bias, and more accurate measurements should be used in further studies.
Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests Acknowledgements We thank all of the pregnant women who participated in the study and all collaborators in the study hospital. Appendix A. Supplementary data Supplementary data related to this article can be found at https:// doi.org/10.1016/j.envres.2019.108526. References Astellburt, T., Feng, X., Kolt, G.S., 2013. Does access to neighbourhood green space promote a healthy duration of sleep? Novel findings from a cross- sectional study of 259 319 australians. BMJ Open 3, 1. Astellburt, T., Feng, X., Kolt, G.S., 2014. Is neighborhood green space associated with a lower risk of type 2 diabetes? Evidence from 267,072 australians. Diabetes Care 37, 197–201. Banay, R.F., James, P., Hart, J.E., Kubzansky, L.D., Spiegelman, D., Okereke, O.I., et al., 2019. Greenness and depression incidence among older women. Environ. Health Perspect. 127, 027001. Cai, S., Tan, S., Gluckman, P.D., Godfrey, K.M., Saw, S., Teoh, O.H., et al., 2017. Sleep quality and nocturnal sleep duration in pregnancy and risk of gestational diabetes mellitus. Sleep 40. Choe, S., Kauderer, S., Eliot, M.N., Glazer, K., Kingsley, S.L., Carlson, L., et al., 2018. Air pollution, land use, and complications of pregnancy. Sci. Total Environ. 645, 1057–1064. Cohen-Cline, H., Turkheimer, E., Duncan, G.E., 2015. Access to green space, physical activity and mental health: a twin study. J. Epidemiol. Community Health 69, 523–529. Dadvand, P., Nieuwenhuijsen, M.J., Esnaola, M., Forns, J., Basagana, X., Alvarezpedrerol, M., et al., 2015a. Green spaces and cognitive development in primary schoolchildren. Proc. Natl. Acad. Sci. U.S.A. 112, 7937–7942. Dadvand, P., Rivas, I., Basagana, X., Alvarez-Pedrerol, M., Su, J., De Castro Pascual, M., et al., 2015b. The association between greenness and traffic-related air pollution at schools. Sci. Total Environ. 523, 59–63. Dadvand, P., Poursafa, P., Heshmat, R., Motlagh, M.E., Qorbani, M., Basagana, X., et al., 2018. Use of green spaces and blood glucose in children; a population-based caspianv study. Environ. Pollut. 243, 1134–1140. Dalton, A.M., Jones, A.P., Sharp, S.J., Cooper, A.J., Griffin, S., Wareham, N.J., 2016. Residential neighbourhood greenspace is associated with reduced risk of incident diabetes in older people: a prospective cohort study. BMC Public Health 16, 1171. Eriksson, C., Hilding, A., Pyko, A., Bluhm, G., Pershagen, G., Ostenson, C., 2014. Longterm aircraft noise exposure and body mass index, waist circumference, and type 2 diabetes: a prospective study. Environ. Health Perspect. 122, 687–694. Facco, F., Grobman, W.A., Reid, K.J., Parker, C.B., Hunter, S., Silver, R.M., et al., 2017. Objectively measured short sleep duration and later sleep midpoint in pregnancy are associated with a higher risk of gestational diabetes. Am. J. Obstet. Gynecol. 217. Fleisch, A.F., Gold, D.R., Rifasshiman, S.L., Koutrakis, P., Schwartz, J., Kloog, I., et al., 2014. Air pollution exposure and abnormal glucose tolerance during pregnancy: the project viva cohort. Environ. Health Perspect. 122, 378–383. Ge, D., Dong, Y., Wang, X., Treiber, F.A., Snieder, H., 2006. The Georgia cardiovascular twin study: influence of genetic predisposition and chronic stress on risk for cardiovascular disease and type 2 diabetes. Twin Res. Hum. Genet. 9, 965–970. 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. Hu, H., Ha, S., Henderson, B.H., Warner, T.D., Roth, J., Kan, H., et al., 2015. Association of atmospheric particulate matter and ozone with gestational diabetes mellitus. Environ. Health Perspect. 123, 853–859. James, P., Banay, R.F., Hart, J.E., Laden, F., 2015. A review of the health benefits of greenness. Current Epidemiology Reports, vol. 2. pp. 131–142. Johnson, B.S., Malecki, K.M., Peppard, P.E., Beyer, K.M.M., 2018. Exposure to neighborhood green space and sleep: evidence from the survey of the health of Wisconsin. Sleep health 4, 413–419. Laurent, O., Wu, J., Li, L., Milesi, C., 2013. Green spaces and pregnancy outcomes in southern California. Health Place 24, 190–195. Lee, J.Y., Lamichhane, D.K., Lee, M., Ye, S., Kwon, J.H., Park, M.S., et al., 2018. Preventive effect of residential green space on infantile atopic dermatitis associated
6. Conclusions Living with higher levels of green space was significantly associated with decreased maternal glucose levels and attenuated risks of incident maternal IGT and GDM. Our findings provide evidence linking green space to better maternal glucose outcomes. Additional studies are needed to further explore the subsequent maternal and child health benefits related to our findings. Conflicts of interest The authors declare they have no actual or potential competing financial interests. Funding This work was supported by the National Natural Science Foundation of China [grant numbers: 91643207, 91743103, and 21437002], the National Key Research and Development Plan [grant numbers: 2016YFC0206700, 2016YFC0206203], and the Fundamental Research Funds for the Central Universities, HUST [grant numbers: 2016YXZD043, 2015ZDTD047, and 2018KFYXMPT00]. Availability of data and materials The datasets for the current study are available from the corresponding author on reasonable request. Authors' contributions Jiaqiang Liao and Xinmei Chen contributed to analyzing the data and drafting the paper. Zhongqiang Cao, Shengwen Liang, Ke Hu, and Yiming Zhang contributed to data collection and data analysis. Wei Xia, Shunqing Xu, Yuanyuan Li, and Bin Zhang contributed to the study concept and design, results interpretation, and revision of the manuscript and approved the final version of the manuscript. Ethics approval and consent to participate The study protocol was approved by the ethics committee of Tongji Medical College, Huazhong University of Science and Technology (No. S152) and Wuhan Children's Hospital (Wuhan Maternal and Child Healthcare Hospital) (No. 2,016,003). Individual written informed 7
Environmental Research 176 (2019) 108526
J. Liao, et al.
2017. Gestational diabetes mellitus and exposure to ambient air pollution and road traffic noise: a cohort study. Environ. Int. 108, 253–260. Roe, J., Thompson, C.W., Aspinall, P., Brewer, M.J., Duff, E.I., Miller, D., et al., 2013. Green space and stress: evidence from cortisol measures in deprived urban communities. Int. J. Environ. Res. Public Health 10, 4086–4103. Roe, J., Aspinall, P., Thompson, C.W., 2017. Coping with stress in deprived urban neighborhoods: what is the role of green space according to life stage? Front. Psychol. 8, 1760. Sallis, J.F., Floyd, M.F., Rodriguez, D.A., Saelens, B.E., 2012. Role of built environments in physical activity, obesity, and cardiovascular disease. Circulation 125, 729–737. Shen, Y., Lung, S.C., 2016. Can green structure reduce the mortality of cardiovascular diseases. Sci. Total Environ. 566, 1159–1167. Sorensen, M., Andersen, Z.J., Nordsborg, R.B., Becker, T., Tjonneland, A., Overvad, K., et al., 2013. Long-term exposure to road traffic noise and incident diabetes: a cohort study. Environ. Health Perspect. 121, 217–222. Thiering, E., Markevych, I., Bruske, I., Fuertes, E., Kratzsch, J., Sugiri, D., et al., 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. Toshihiro, M., Saito, K., Takikawa, S., Takebe, N., Onoda, T., Satoh, J., 2008. Psychosocial factors are independent risk factors for the development of type 2 diabetes in Japanese workers with impaired fasting glucose and/or impaired glucose tolerance. Diabet. Med. 25, 1211–1217. Trigueromas, M., Donairegonzalez, D., Seto, E., Valentin, A., Martinez, D., Smith, G., et al., 2017. Natural outdoor environments and mental health: stress as a possible mechanism. Environ. Res. 159, 629–638. Wang, H., Leng, J., Li, W., Wang, L., Zhang, C., Liu, H., et al., 2017. Sleep duration and quality, and risk of gestational diabetes mellitus in pregnant Chinese women. Diabet. Med. 34, 44–50. WHO, 2010. Urban Planning, Environment and Health: from Evidence to Policy Action. World Health Organization, Geneva. WHO, 2016. Urban Green Spaces and Health: A Review of Evidence. World Health Organization, Copenhagen, Denmark. Wong, M., Wong, W.H.S., Cheung, Y., 2007. Fetal myocardial performance in pregnancies complicated by gestational impaired glucose tolerance. Ultrasound Obstet. Gynecol. 29, 395–400. Xu, X., Liu, C., Xu, Z., Tzan, K., Zhong, M., Wang, A., et al., 2011. Long-term exposure to ambient fine particulate pollution induces insulin resistance and mitochondrial alteration in adipose tissue. Toxicol. Sci. 124, 88–98. Yang, B., Markevych, I., Heinrich, J., Bowatte, G., Bloom, M.S., Guo, Y., et al., 2018a. Associations of greenness with diabetes mellitus and glucose-homeostasis markers: the 33 communities Chinese health study. Int. J. Hyg Environ. Health. Yang, B.Y., Qian, Z., Li, S., Chen, G., Bloom, M.S., Elliott, M., et al., 2018b. 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. Zou, G., 2004. A modified Poisson regression approach to prospective studies with binary data. Am. J. Epidemiol. 159, 702–706.
with prenatal air pollution exposure. Int. J. Environ. Res. Public Health 15. Liang, F., Yang, X., Liu, F., Li, J., Xiao, Q., Chen, J., et al., 2019. Long-term exposure to ambient fine particulate matter and incidence of diabetes in China: a cohort study. Environ. Int. 126, 568–575. Liao, J., Yu, H., Xia, W., Zhang, B., Lu, B., Cao, Z., et al., 2018. Exposure to ambient fine particulate matter during pregnancy and gestational weight gain. Environ. Int. 119, 407–412. Lodovici, M., Bigagli, E., 2011. Oxidative stress and air pollution exposure. J. Toxicol. 2011 487074. Lu, M., Wang, P., Cheng, T., Yang, C., Yan, Y., 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., Hennig, F., Matthiessen, C., Ohlwein, S., Icks, A., Moebus, S., et al., 2018. Air pollution and glucose metabolism: an analysis in non-diabetic participants of the heinz nixdorf recall study. Environ. Health Perspect. 126, 047001. Margaritis, E., Kang, J., 2017. Relationship between green space-related morphology and noise pollution. Ecol. Indicat. 72, 921–933. Markevych, I., Fuertes, E., Tiesler, C.M., Birk, M., Bauer, C.P., Koletzko, S., et al., 2014. Surrounding greenness and birth weight: results from the giniplus and lisaplus birth cohorts in munich. Health Place 26, 39–46. Metzger, B.E., Gabbe, S.G., Persson, B., Buchanan, T.A., Catalano, P.A., Damm, P., et al., 2010. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care 33 e97; author reply e98. Mooy, J.M., De Vries, H., Grootenhuis, P.A., Bouter, L.M., Heine, R.J., 2000. Major stressful life events in relation to prevalence of undetected type 2 diabetes : the hoorn study. Diabetes Care 23, 197–201. Muller, G., Harhoff, R., Rahe, C., Berger, K., 2018. Inner-city green space and its association with body mass index and prevalent type 2 diabetes: a cross-sectional study in an urban German city. BMJ Open 8. Nasiri-Amiri, F., Bakhtiari, A., Faramarzi, M., Adib Rad, H., Pasha, H., 2016. The association between physical activity during pregnancy and gestational diabetes mellitus: a case-control study. Int. J. Endocrinol. Metab. 14, e37123. Nichani, V., Dirks, K., Burns, B., Bird, A., Morton, S., Grant, C., 2017. Green space and pregnancy outcomes: evidence from growing up in New Zealand. Health Place 46, 21–28. Nieuwenhuijsen, M.J., Khreis, H., Triguero-Mas, M., Gascon, M., Dadvand, P., 2017. Fifty shades of green: pathway to healthy urban living. Epidemiology 28, 63–71. Orioli, R., Antonucci, C., Scortichini, M., Cerza, F., Marando, F., Ancona, C., et al., 2019. Exposure to residential greenness as a predictor of cause-specific mortality and stroke incidence in the rome longitudinal study. Environ. Health Perspect. 127, 027002. Ostlund, I., Hanson, U., Bjorklund, A., Hjertberg, R., Eva, N., Nordlander, E., et al., 2003. Maternal and fetal outcomes if gestational impaired glucose tolerance is not treated. Diabetes Care 26, 2107–2111. Park, S.K., Adar, S.D., Oneill, M.S., Auchincloss, A.H., Szpiro, A.A., Bertoni, A.G., et al., 2015. Long-term exposure to air pollution and type 2 diabetes mellitus in a multiethnic cohort. Am. J. Epidemiol. 181, 327–336. Pedersen, M., Olsen, S.F., Halldorsson, T.I., Zhang, C., Hjortebjerg, D., Ketzel, M., et al.,
8