Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jun 10.
Published in final edited form as: Sci Total Environ. 2022 Feb 14;824:153848. doi: 10.1016/j.scitotenv.2022.153848

Mortality Risk Associated with Greenness, Air Pollution, and Physical Activity in a Representative U.S. Cohort

Carver J Coleman 1, Ray A Yeager 2, Zachari A Pond 3, Daniel W Riggs 4, Aruni Bhatnagar 4, C Arden Pope III 1
PMCID: PMC10243166  NIHMSID: NIHMS1782053  PMID: 35176374

Abstract

Several cohort studies suggest greenness is associated with decreased mortality risk. Potential confounding by or interactions between physical activity and air pollution remains unclear. This study evaluates associations of greenness, air pollution, and physical activity with mortality risk and investigates confounding and effect modification across these key risk factors. National Health Interview Survey (NHIS) data covering 1997–2014 were linked to the National Death Index to generate a cohort of 403,748 individuals with 39,528 deaths. Greenness, represented by census-tract Normalized Difference Vegetation Index (NDVI) for the seasonal period of May-October, was averaged over the years 2003–2016. Air pollution was estimated by census-tract level PM2.5 concentrations from 1999–2015. Cox Proportional Hazard Models were used to estimate hazard ratios (HR) for differences in greenness, air pollution, and physical activity. Alternative models that evaluated potential confounding and stratified models that evaluated effect modification were examined. Mortality risks were associated with PM2.5 (HR = 1.14, 95% CI: 1.09–1.19 per 10 μg/m3) and physical inactivity (1.49, 1.44–1.54 relative to sufficiently active), but not with greenness (1.01, 0.99–1.03 per IQR). The PM2.5-mortality association was mitigated at high levels of greenness (1.05, 0.91–1.22). There was no strong evidence of confounding between air pollution, physical activity, and greenness. However, stratified analysis suggested effect modification for PM2.5 and NDVI by physical activity. A significant protective greenness-mortality association was observed for only highly active individuals (0.91, 0.86–0.96). Also, relatively high PM2.5-mortality HRs were observed for more physically active individuals (1.25, 1.12–1.40). PM2.5 air pollution and physical inactivity are robustly associated with mortality risk. Greenness may be most beneficial and air pollution relatively harmful to highly active individuals. This analysis provides evidence that, in addition to not smoking, being physically active and living in a clean, green environment contributes to improved health and reduced risk of mortality.

Keywords: greenness, PM2.5, physical activity, mortality, air pollution

Graphical Abstract

graphic file with name nihms-1782053-f0004.jpg

1. Introduction

A growing body of literature has revealed associations between high levels of residential green space and reduced risk of all-cause mortality. Greenness is hypothesized to promote physical activity (James et al. 2015; James et al. 2017), improve mental health (Gascon et al. 2018; Pun et al. 2018; Hartley et al. 2021), heighten cognition (de Keijzer et al. 2016; Thygesen et al. 2020), reduce noise and heat (Szeremeta and Zannin 2009; Dzhambov and Dimitrova 2014; Maimaitiyiming et al. 2014), and reduce air pollution (Hwang et al. 2011; Nieuwenhuijsen et al. 2018; Crouse et al. 2019; Klompmaker et al. 2020; Kasdagli et al. 2020; Ji et al. 2020; Jaafari et al. 2020). Most cohort studies, primarily in the U.S. (James et al. 2016; Heo et al. 2019, Yitshak-Sade et al. 2019; Son et al. 2021; Coleman et al. 2021), Canada (Villeneuve et al. 2012; Crouse et al. 2017; Crouse et al. 2019; Chen et al. 2020), and China (Ji et al. 2019; Sun et al. 2020a; Ji et al. 2020), have used Normalized Difference Vegetation Index (NDVI) to study the greenness and mortality relationship. Even though most cohorts were restricted to the elderly (Ji et al. 2019; Zijlema et al. 2019; Heo et al. 2019; Sun et al. 2020a; Ji et al. 2020; Son et al. 2021), non-immigrants (Crouse et al. 2017; Crouse et al. 2019), a single sex (James et al. 2016; Zijlema et al. 2019), stroke victims (Wilker et al. 2014), or cancer patients and survivors (Coleman et al. 2021), several studies have identified greenness effects on all-cause mortality in more population representative cohorts, either in a single city (Villeneuve et al. 2012; Nieuwenhuijsen et al. 2018; Orioli et al. 2019; Chen et al. 2020), group of cities (Kim et al. 2019), U.S. state (Yitshak-Sade et al. 2019), or country (de Keijzer et al. 2017; Vienneau et al. 2017). No studies to date have analyzed greenness effects in a representative, U.S. cohort. Furthermore, only four of the nineteen identified cohort studies (James et al. 2016; Villeneuve et al. 2012; Ji et al. 2019; Ji et al. 2020) include information on physical activity. Physical activity is a potential confounder and effect modifier of the greenness-mortality association because it robustly affects mortality (Paffenbarger et al. 1986) and may be correlated with greenness (James et al. 2015; James et al. 2017).

Air pollution has been estimated to be the fourth largest contributor to global burden of disease (Murray et al. 2020). Many cohort studies have found associations between an increase in fine particulate matter air pollution (PM2.5) and increased risk of all-cause, cardiopulmonary, and lung cancer mortality (Dockery et al. 1993; Pope et al. 2002; Jerrett et al. 2017; Pinault et al. 2017; Di et al. 2017; Pope et al. 2019; Chen and Hoek 2020). However, several cohort studies have shown that air pollution effects are mitigated at high levels of greenness (Crouse et al. 2017; Kim et al. 2019; Heo et al. 2019; Yitshak-Sade et al. 2019; Sun et al. 2020a; Son et al. 2021; Coleman et al. 2021), or that greenness may confound estimates of air pollution-mortality associations (Crouse et al. 2019). Similarly, some studies have found that air pollution and physical activity also interact. The benefit of physical activity may be mitigated at high levels of air pollution (Tainio et al. 2016; Tainio et al. 2021) due to increased exposure to PM2.5, although other studies have found no interaction between air pollution and physical activity (Anderson et al. 2015; Cole-Hunter et al. 2018; Sun et al. 2020b; Elliott et al. 2020) or that physical activity reduces the harmful effect of air pollution (Matt et al. 2016; Laeremans et al. 2018).

The primary objectives of this study were to 1) assess risk of all-cause mortality associated with greenness, air pollution, and physical activity; 2) investigate potential confounding between greenness, air pollution, and physical activity; and 3) explore effect modification between greenness, air pollution, and physical activity. Secondary objectives were to estimate risk of cause-specific mortality and analyze all-cause associations stratified by several individual-level characteristics.

2. Methods

2.1. The Cohort

The National Health Interview Survey (NHIS), an annual survey carried out by the National Center for Health Statistics (NCHS), uses a geographically clustered probability sampling design to produce a representative sample of the civilian noninstitutionalized U.S. population (NCHS 1986, 2021a). Persons without a permanent, U.S. residence are excluded from the survey universe, such as persons in correctional facilities, active-duty military, persons in long-term care facilities, and U.S. citizens living in foreign countries (NCHS 2021b). Surveying of households by the NCHS began in 1958, though the current questionnaire was implemented in 1997. Publicly available NHIS data covering the years 1997–2014 were linked to restricted-use 2010 census-tract identifiers and restricted-use National Death Index data through 2015 (NCHS 2021c). U.S. census tracts include approximately 4,000 individuals. The final cohort only contained individuals aged 18–84 who were surveyed in the contiguous U.S. and for whom information was available on age, sex, race-ethnicity, leisure-time physical activity, smoking status, body mass index (BMI), income, marital status, educational attainment, census tract-level air pollution, census tract-level greenness, survey date, and date and cause of death (if deceased before December 31st, 2015).

Harmonization of several variables was required due to minor questionnaire changes between 1997 and 2014. In particular, household income was reported as inflation-adjusted household income using the Consumer Price Index. These changes, along with data linkage information, are documented elsewhere (Pope et al. 2019). The privacy of the individuals surveyed by the NCHS is protected by the Privacy Act of 1974. As such, analyses were conducted at an NCHS Research Data Center in Rockville, MD, and no efforts were made to identify individuals surveyed by the NCHS. Because research reported in this manuscript uses publicly accessible data that are de-identified, it is not subject to federal regulations on protection of human research subjects.

2.2. Physical Activity Data

Leisure-time aerobic physical activity information was collected by the NCHS beginning in 1997. Subjects were asked to describe the number of minutes per week they participate in vigorous and moderate leisure-time physical activity (NCHS 2021d). Vigorous physical activities are defined in the questionnaire as activities that cause heavy sweating or large increases in breathing or heart rate, while moderate activities cause only light sweating or a slight to moderate increase in breathing or heart rate. The physical activity of individuals was then classified following the recommendations of the 2008 Physical Activity Guidelines for Americans (HHS 2008) created by the U.S. Department of Health and Human Services (HHS). Mainly, weekly aerobic activity was calculated as the sum of a subject’s minutes spent per week participating in moderate physical activities and two times the minutes spent per week on vigorous aerobic activities. The four leisure-time physical activity classifications (as designated by the HHS) are inactive (0 min per week), insufficiently active (between 0 and 150 min per week), sufficiently active (between 150 and 300 min per week), and highly active (more than 300 min per week). More information on physical activity in the NHIS is available elsewhere (NCHS 2021d).

2.3. Greenness Estimates

Mosaiced Normalized Difference Vegetation Index (NDVI) data at the 2010 census tract-level were used as a proxy for residential green space. Collection of these estimates are described elsewhere (Coleman et al. 2021). In brief, NDVI estimates were obtained over the years 2003–2016 from the EarthExplorer website (U.S. Geological Survey 2019) and subsequently cleaned for water and clouds (See Supplemental Information 1). Because there was very little year-to-year variation in NDVI estimates (R > 0.97), a single average NDVI estimate for each census tract was calculated. Additionally, to avoid potential bias from snow during the winter, only warm months (May-October) NDVI estimates were used. Participants were classified into EPA level 1 ecoregions (12 locations where the ecosystems, type, quality, and quantity of environmental resources are generally similar) at the 2010 census tract-level to account for differing types of vegetation and climates (EPA 2022).

2.4. Air Pollution Estimates

Monitored fine particulate matter air pollution (PM2.5) data, provided by the EPA, were used in an integrated empirical geographic (IEG) modeling approach (Kim et al. 2020). Census tract-level estimates for PM2.5 were obtained using integrated ground-based monitored data with satellite-derived estimates of land use and air pollution, along with other land use and geographic variables. Information on the IEG modeling technique is available elsewhere (Pope et al. 2019; Kim et al. 2020; Coleman et al. 2020) and the data are publicly available (https://www.caces.us). Averaged 1999–2015 PM2.5 estimates were used for this study. A recent study using the same PM2.5 data and the NHIS cohort concluded results were not sensitive to the average time frame (Lefler et al. 2019; Pope et al. 2019).

2.5. Statistical Analyses

A priori Cox Proportional Hazard (CPH) Models were implemented in SAS software version 9.4 using the PHREG procedure (SAS Institute Inc., Cary, North Carolina). Hazard ratios (HRs) and 95% confidence intervals (95% CIs) were calculated for all-cause mortality using a full unstratified model and several derivations of this model. In the full model, all 532 combinations of age (1-year intervals), sex, and race-ethnicity (Black non-Hispanic, Hispanic, Other/unknown, White non-Hispanic) were allowed their own baseline hazard (using the STRATA statement). The STRATA statement allows the CHP model to flexible control for age, sex, and race. Warm-months NDVI (designated as NDVI) and PM2.5 were included as continuous variables, with NDVI scaled per IQR and PM2.5 scaled per 10 μg/m3. The following variables were included as categorical variables: physical activity (inactive, insufficiently active, sufficiently active, highly active), smoking status (current, former, never), BMI (< 20, 20–25, 2530, 30–35, > 35), inflation-adjusted household income ($0–35,000, $35–50,000, $50–75,000, > $75,000), educational attainment (< high school graduate, high school graduate, some college, college graduate, > college graduate), marital status (divorced, separated, never married, widowed, married), urban/rural, census region (Northeast, West, Midwest, South), and survey year.

Survival time was calculated in days and was the difference between the survey date and death. Censored survival time for survivors was calculated as the difference between survey date and end of follow-up (December 31, 2015) or loss to follow-up. For cause-specific analyses, censored survival time was similarly calculated as the difference between survey date and death of other causes. Cause-of-death coding used the tenth revision of the International Statistical Classification of Diseases, Injuries, and Causes of Death (ICD-10). Mainly, cardiopulmonary mortality was divided into cardiovascular (I00-I09, I11, I13, I20-I51), cerebrovascular (I60-I69), chronic lower respiratory (J40-J47), and influenza/pneumonia (J09-J18). Cancer mortality (C00-C97) was also divided into lung cancer (C33-C34). All HRs for all models are displayed in tables as supplementary materials for all-cause, cardiopulmonary, and cancer mortality.

Models with combinations of NDVI, PM2.5, and physical activity were used to assess confounding. Additionally, model sensitivity was assessed by progressively adding controls starting with age, sex, and race-ethnicity. HRs of the association between all-cause mortality and NDVI, PM2.5, and physical activity were calculated stratified across different quartiles of NDVI, quartiles of PM2.5, and categories of physical activity to assess effect modification. Additionally, HRs were calculated for NDVI, PM2.5, and physical activity stratified across the following individual characteristics: sex, race-ethnicity, age (18–64, 65–74, 75–84), smoking status, BMI, inflation-adjusted income, educational attainment, marital status, urban/rural, census region, survey year (1997–2002, 2003–2008, 2009–2014), and ecoregion (Eastern Temperate Forests, Great Plains, Mediterranean California, Other).

3. Results

Table 1 presents the summary statistics of individuals in the NHIS Cohort used in this analysis. The cohort consisted of 403,748 individuals with 39,528 deaths. More deaths were from cardiopulmonary causes (14,720 or 37.2%) than from any other cause of death. 54.7% of the individuals were female and 78.5% were surveyed in urban areas. Most of the individuals were either physically inactive (35.7%) or highly active (29.5%). The mean PM2.5 exposure was 10.59 μg/m3 with a standard deviation (SD) of 2.41 and an IQR of 3.18. Warm-months NDVI had a mean of 0.54 with a SD of 0.17 and an IQR of 0.27.

Table 1.

Unweighted, baseline summary statistics of the NHIS Cohort restricted to years 1997–2014 and ages 18–84 with mortality follow-up through 2015.

Summary Statistic Full Cohort
Total Individuals in Cohort 403,748 (100.0%)
 Total Number of Deaths 39,528 (9.8%)
  Cardiopulmonarya 14,720 (37.2%)
   Cardiovascular 9,491 (24.0%)
   Cerebrovascular 2,042 (5.2%)
   Chronic Lower Respiratory 2,450 (6.2%)
   Pneumonia and Influenza 737 (1.9%)
  Cancera 10,107 (25.6%)
   Lung Cancer 2,892 (7.3%)
  Other 14,701 (37.2%)
Age in years (mean ± SD) 45.88 ± 17.02
NDVI (mean ± SD [IQR]) 0.44 ± 0.14 [0.20]
May-Oct NDVI (mean ± SD [IQR]) 0.54 ± 0.17 [0.27]
PM2.5 (mean ± SD [IQR]) 10.59 ± 2.41 [3.18]
Physical Activity
 Inactive 144,265 (35.7%)
 Insufficiently Active 80,062 (19.8%)
 Sufficiently Active 60,301 (14.9%)
 Highly Active 119,120 (29.5%)
Smoking Status
 Current 88,481 (21.9%)
 Former 87,639 (21.7%)
 Never 227,628 (56.4%)
BMI
 < 20 23,182 (5.7%)
 20–25 131,000 (32.5%)
 25–30 141,085 (34.9%)
 30–35 67,127 (16.6%)
 > 35 41,354 (10.2%)
Sex
 Female 220,957 (54.7%)
 Male 182,791 (45.3%)
Race-ethnicity
 Black non-Hispanic 59,762 (14.8%)
 Hispanic 70,350 (17.4%)
 Other/Unknown 20,159 (5.0%)
 White non-Hispanic 253,477 (62.8%)
Household Income
 $0–35,000 156,565 (38.8%)
 $35–50,000 59,686 (14.8%)
 $50–75,000 73,400 (18.2%)
 > $75,000 114,097 (28.3%)
Marital Status
 Divorced 61,669 (15.3%)
 Separated 14,959 (3.7%)
 Never married 106,030 (26.23%)
 Widowed 31,879 (7.9%)
 Married 189,211 (46.9%)
Education
 < High-school Graduate 71,322 (17.7%)
 High-school Graduate 109,511 (27.1%)
 Some College 120,645 (29.9%)
 College Graduate 66,825 (16.6%)
 > College Graduate 35,445 (8.8%)
Urban/Rural
 Rural 86,825 (21.5%)
 Urban 316,923 (78.5%)
Census Region
 Northeast 66,070 (16.4%)
 Midwest 91,751 (22.7%)
 South 152,280 (37.7%)
 West 93,647 (23.2%)
Ecoregion
 Eastern Temperate Forests 237,283 (58.8%)
 Great Plains 64,640 (16.0%)
 Marine West Coast Forests 10,526 (2.6%)
 Mediterranean California 43,845 (10.9%)
 North American Deserts 24,866 (6.2%)
 Northern Forests 6,732 (1.7%)
 Northwestern Forested 5,794 (1.4%)
 Southern Semi-Arid Highlands 554 (0.1%)
 Temperate Sierras 532 (0.1%)
 Tropical Wet Forests 8,976 (2.2%)
Survey Year
 1997–2002 123,408 (30.6%)
 2003–2008 115,989 (28.7%)
 2009–2014 164,351 (40.7%)
Open in a new tab

Note: The data were complete for all variables. SD, standard deviation; IQR, interquartile range; NDVI, normalized difference vegetation index; PM2.5, particulate matter (< 2.5 μg/m3); BMI, body mass index.

a

Cause-of-death coding used the tenth revision of the International Statistical Classification of Diseases, Injuries, and Causes of Death (ICD-10). Mainly, cardiopulmonary mortality was divided into cardiovascular (I00-I09, I11, I13, I20-I51), cerebrovascular (I60-I69), chronic lower respiratory (J40-J47), and influenza/pneumonia (J09-J18). Cancer mortality (C00-C97) was also divided into lung cancer (C33-C34).

Table 2 shows the correlation coefficients between NDVI, PM2.5, and physical activity. NDVI and PM2.5 were weakly, inversely correlated (−0.12). NDVI was minimally correlated with physical activity and inversely correlated with inactivity (−0.04). PM2.5 was correlated with inactivity (0.06) and inversely correlated with high activity (−0.05).

Table 2.

Correlations of Exposures and Physical Activity

NDVI PM2.5
NDVI 1
PM2.5 −0.12 1
Inactive −0.04 0.06
Insufficiently Active 0.03 −0.00
Sufficiently Active 0.01 −0.01
Highly Active 0.01 −0.05
Open in a new tab

Note: NDVI, normalized difference vegetation index; PM2.5, particulate matter (< 2.5 μg/m3).

Table 3 shows HRs and 95% confidence intervals (95% CIs) for the full unstratified model (including NDVI, PM2.5, and physical activity together) across cause-specific mortality. No significant associations were observed for NDVI. A 10 μg/m3 increase in PM2.5 was significantly associated with an increased risk of all-cause, cardiopulmonary, cardiovascular, pneumonia and influenza, cancer, and lung cancer mortality. Physical inactivity relative to sufficient physical activity was significantly associated with an increased risk of all observed causes of death. The highest HR values were observed with Pneumonia/Influenza, followed by cardiopulmonary mortality. Table S1 shows HRs and 95% CIs for each risk factor in the full model for all-cause, cardiopulmonary, cancer, and lung cancer mortality and Table S2 displays the same for cardiovascular, cerebrovascular, chronic lower respiratory, and pneumonia and influenza mortality. The largest risk factor for all-cause mortality in the NHIS cohort was current smoking (2.09, 2.04–2.15).

Table 3.

Estimated HRs and 95% CIs for the full unstratified model across cause-specific mortality using the CPH model. Baseline physical activity is sufficiently active.

Cause of Deatha NDVI (Per IQR) PM2.5 (Per 10 μg/m3) Inactive Insufficiently Active Highly Active
All Cause 1.01 (0.99, 1.03) 1.01 (0.99, 1.03) 1.49 (1.44, 1.54) 1.14 (1.10, 1.19) 0.90 (0.87, 0.94)
 Cardiopulmonary 0.99 (0.96, 1.03) 1.16 (1.07, 1.25) 1.63 (1.54, 1.73) 1.20 (1.12, 1.28) 0.87 (0.81, 0.94)
  Cardiovascular 0.99 (0.95, 1.04) 1.25 (1.13, 1.37) 1.51 (1.41, 1.63) 1.17 (1.08, 1.27) 0.89 (0.81, 0.97)
  Cerebrovascular 1.03 (0.93, 1.13) 1.16 (0.95, 1.42) 1.58 (1.35, 1.85) 1.29 (1.08, 1.53) 0.93 (0.77, 1.12)
  Chronic Lower Respiratory 1.00 (0.91, 1.09) 0.82 (0.67, 0.99) 2.24 (1.91, 2.62) 1.23 (1.02, 1.47) 0.76 (0.63, 0.93)
  Pneumonia/Influenza 0.90 (0.77, 1.05) 1.39 (1.00, 1.94) 1.61 (1.24, 2.10) 1.25 (0.93, 1.67) 0.83 (0.61, 1.14)
 Cancer 1.00 (0.96, 1.05) 1.25 (1.14, 1.37) 1.30 (1.22, 1.39) 1.11 (1.03, 1.19) 0.91 (0.84, 0.98)
  Lung Cancer 1.01 (0.93, 1.10) 1.20 (1.01, 1.44) 1.38 (1.22, 1.57) 1.09 (0.94, 1.26) 0.90 (0.78, 1.04)
 Other 1.03 (0.99, 1.06) 1.05 (0.97, 1.13) 1.50 (1.42, 1.59) 1.12 (1.05, 1.19) 0.92 (0.86, 0.98)
Open in a new tab

Note: Adjusted for combinations of age, sex, and race-ethnicity, smoking status, body mass index, household income, educational attainment, marital status, urban/rural, census region, and survey year. HR, hazard ratio; CI, confidence interval; CPH, cox proportional hazard; IQR, interquartile range; NDVI, normalized difference vegetation index; PM2.5, particulate matter (< 2.5 μg/m3).

a

Cause-of-death coding used the tenth revision of the International Statistical Classification of Diseases, Injuries, and Causes of Death (ICD-10). Mainly, cardiopulmonary mortality was divided into cardiovascular (I00-I09, I11, I13, I20-I51), cerebrovascular (I60-I69), chronic lower respiratory (J40-J47), and influenza/pneumonia (J09-J18). Cancer mortality (C00-C97) was also divided into lung cancer (C33-C34).

Results from models with different combinations of NDVI, PM2.5, and physical activity are illustrated in Figure 1 and shown in Table S3. Each variable is first shown alone in the model and then controlling for the other two separately and together. Controlling for PM2.5 and physical activity had minimal impact on the NDVI-mortality HR. Also, PM2.5-mortality HRs were highly stable to controlling for NDVI and physical activity. The association between physical activity and mortality was essentially unaffected by controlling for PM2.5 and NDVI. NDVI-mortality and PM2.5-mortality estimates were not highly sensitive to modeling choices regarding covariates included in the model (Figure S1).

Figure 1.

Figure 1.

Open in a new tab

Estimated HRs and 95% CIs for all-cause mortality associated with an IQR increase in NDVI (square), a 10 μg/m3 increase in PM2.5 (circle), and physical activity relative to sufficiently active (diamond) with and without controlling for NDVI, PM2.5, and physical activity. HRs are estimated using the CPH model including the other two variables individually and together. HR, hazard ratio; CI, confidence interval; IQR, interquartile range; NDVI, normalized difference vegetation index; PM2.5, particulate matter (< 2.5 μg/m3); CPH, cox proportional hazard.

The HRs for NDVI, PM2.5, and physical activity stratified by quartiles of NDVI, quartiles of PM2.5, and categories of physical activity are illustrated in Figure 2 and shown in Tables S4–6. An IQR increase in NDVI was generally not associated with mortality risk. However, NDVI was associated with decreased mortality risk for highly active individuals (0.91, 0.86–0.96). Furthermore, PM2.5 was generally associated with increased mortality risk. However, PM2.5 was not significantly associated with increased mortality risk at high levels of greenness (1.05, 0.91–1.22), and the PM2.5-mortality association was strongest for highly active individuals (1.25, 1.12–1.40). Additionally, physical inactivity effects were mitigated at higher levels of PM2.5 and lower levels of NDVI, though these differences were insignificant.

Figure 2.

Figure 2.

Open in a new tab

Effect modification analysis of estimated HRs and 95% CIs from the CPH model for all-cause mortality associated with an IQR increase in NDVI (square), a 10 μg/m3 increase in PM2.5 (circle), and physically inactive relative to sufficiently active (diamond) stratified by quartiles of NDVI, quartiles of PM2.5, and categories of leisure-time physical activity. HR, hazard ratio; CI, confidence interval; CPH, cox proportional hazard; IQR, interquartile range; NDVI, normalized difference vegetation index; PM2.5, particulate matter (< 2.5 μg/m3).

Figure 3 and Tables S4–6 show HRs and 95% CIs for NDVI, PM2.5, and physical inactivity stratified by individual-level characteristics. No significant NDVI effects were observed for any strata. The relationship between PM2.5 and mortality risk was somewhat consistent across strata, though the associations were larger for Hispanics, individuals 18–64 years old, never smokers, those living in urban areas, and individuals living in the Great Plains. Physical inactivity was significantly associated with mortality risk across all strata.

Figure 3.

Figure 3.

Open in a new tab

Estimated HRs and 95% CIs for all-cause mortality associated with an IQR increase in NDVI (square), a 10 μg/m3 increase in PM2.5 (circle), and physically inactive relative to sufficiently active (diamond) stratified by sex, race-ethnicity, age, smoking status, body mass index, inflation-adjusted household income, educational attainment, marital status, urban/rural, census region, survey year, and EPA ecoregion. HRs are estimated using the CPH model adjusting for all the other covariates. HR, hazard ratio; CI, confidence interval; IQR, interquartile range; NDVI, normalized difference vegetation index; PM2.5, particulate matter (< 2.5 μg/m3); EPA, U.S. Environmental Protection Agency; CPH, cox proportional hazard.

4. Discussion

This study analyzed associations between mortality risk and greenness, air pollution, and leisure-time physical activity in a large, representative cohort of the U.S. population using a priori cox proportional hazard models. Normalized Difference Vegetation Index (NDVI) was not associated with all-cause mortality. Choices in modeling such as controlling for PM2.5 or physical activity as covariates in the models had minimal impact on NDVI estimates. Stratified analysis, however, suggest that NDVI-mortality associations were modified by levels of physical activity. An IQR increase in NDVI was significantly associated with decreased mortality risk only for highly active individuals. NDVI-mortality associations were also most significant for highly active individuals in the U.S. Nurses’ Health Study cohort (James et al. 2016). However, this study found little evidence of a correlation between NDVI and physical activity, suggesting that greenness-mortality associations for the highly active may not be due to facilitated physical activity. Highly active individuals may receive benefits from greenness because outdoor physical activity in green areas may provide psychological and related benefits (James et al. 2015). Additionally, physical activity may increase the likelihood of exposure to more vegetation-produced microbial antigens which may educate the immune system and prevent disease, a hypothesis commonly referred to as ‘Old Friends’ (Frew 2019; Murdaca et al. 2021). Heightened breathing during exercise may further facilitate intake of beneficial antigens (Tainio et al. 2021). Several studies have found interactions between greenness and socioeconomic status (de Keijzer et al. 2017; Crouse et al. 2017; Son et al. 2021), though no such interactions were observed in this study.

PM2.5 was significantly associated with increased risk of mortality. The PM2.5-mortality association was robust to almost all choices in modeling and across stratified analyses of individual-level characteristics. Interestingly, PM2.5-mortality associations tended to be larger at lower levels of greenness, consistent with several other studies (Crouse et al. 2017; Kim et al. 2019; Heo et al. 2019; Yitshak-Sade et al. 2019; Sun et al. 2020a; Son et al. 2021; Coleman et al. 2021), suggesting that low levels of greenness exacerbate the health effects of PM2.5 exposure. This may be due to the deposition of PM on tree leaves (Hwang et al. 2011), or higher levels of greenness may be associated with different sources or compositions of PM2.5 that have different toxicities. Additionally, PM2.5-mortality associations were somewhat larger for more active individuals, suggesting higher inhalation during exercise and heightened exposure to air particulates from outdoor exercise can exacerbate the adverse effects of PM2.5 (Tainio et al. 2021). This study also found that the individuals 18–64 years of age and never smokers were more sensitive to PM2.5, which is consistent with a previous observation that exposure to PM2.5 induces cardiovascular injury even in young, healthy adults (Pope et al. 2016) and that the effects of PM2.5 are more pronounced in the absence of high disease burden and advanced age.

The PM2.5 HR estimate for all-cause mortality observed in this study (1.14, 1.09–1.19) is similar to the Six Cities cohort (Dockery et al. 1993) (1.14, 1.07–1.22), the 2001 Canadian Census Health and Environment Cohort (Pinault et al. 2017) (1.15, 1.12–1.17), and a previous NHIS study with an extended cohort without physical activity data (Pope et al. 2019) (1.12, 1.08–1.15), but slightly larger than the ACS cohort (Pope et al. 2002; Jerrett et al. 2017) (1.07, 1.04–1.10), and the Medicare cohort (Di et al. 2017) (1.07, 1.07–1.08).

Physical inactivity was significantly and robustly associated with increased mortality risk. Tainio et al. reported that beneficial physical activity effects may be mitigated at extreme levels of PM2.5 (Tainio et al. 2016). In this analysis, there was suggestive, but not statistically significant, evidence that physical activity may be less beneficial to individuals living in areas with high PM2.5.

This study has several strengths: 1) The cohort is a well-documented, nationwide, representative sample of the adult population that is well-harmonized across years. 2) The cohort is reasonably large, with over 400,000 individuals surveyed over an 18-year period with up to 19 years of mortality follow-up with nearly 40,000 deaths. 3) Detailed information on leisure-time physical activity is available for each subject in the cohort and follows categorical conventions set forth by the HHS. 4) PM2.5-mortality HRs were not sensitive to including NDVI or physical activity, suggesting that air pollution studies lacking NDVI or physical activity data may still be informative. 5) Controls for individual characteristics such as age, sex, race-ethnicity, income, education, smoking status, and BMI were also available, though PM2.5 and NDVI estimates were not sensitive to including any single covariate. 6) Census tract-level data allowed for high spatial resolution (U.S. census tracts typically contain approximately 4,000 individuals). 7) Estimates of PM2.5 and NDVI are publicly available. 8) NHIS data files on the census tract-level with mortality follow-up are accessible to researchers with approved proposals and NHIS data with mortality follow-up for several metropolitan areas are publicly available.

This study, like all observational studies, may be limited by potential residual confounding due to inadequate control for variables that are both correlated with environmental exposures or physical activity and also mortality risk. The present study has attempted to minimize confounding by controlling for important and available characteristics (age, sex, race-ethnicity, marital status), behavioral factors (physical activity, smoking status, BMI), socioeconomic status (inflation-adjusted income, educational attainment), spatial identifiers (census region, urban/rural), and temporal identifiers (survey year). Furthermore, Figure S1 shows that PM2.5 and NDVI estimates remain reasonably stable as covariates are progressively added to the model, yet confounding by unknown variables remains possible. For physical activity there is also the concern of reverse causality. Physical inactivity can contribute to disease. Disease can contribute to lack of physical activity. For example, physical inactivity contributes to risk of diabetes and respiratory and cardiovascular disease; these diseases may affect physical activity, and both contribute to risk of mortality. Another limitation of this study is lack of follow-up on behavioral, socioeconomic, and spatial factors resulting in some measurement error of covariates. Individuals were surveyed upon entering the cohort, but only mortality was followed-up. Additionally, since no information is available on moving, PM2.5 and NDVI exposures represent exposures at survey date and some unknown window before and after survey date. Assuming classical measurement error from lack of follow-up, the HRs for PM2.5 and NDVI may be biased towards the null. However, more complicated measurement errors make it difficult to determine the likely direction of the potential bias (Sheppard et al. 2012; Haber et al. 2020).

In conclusion, this study identified significant associations between air pollution, physical activity, and mortality risk in the first representative, U.S. cohort study with information on greenness, air pollution, and physical activity. This analysis provides evidence that, in addition to not smoking, being physically active and living in a clean, green environment contributes to improved health and reduced risk of mortality.

Supplementary Material

1

Highlights.

  • Greenness, air pollution, and physical activity independently impact health.

  • Interactions of these factors were evaluated in a representative U.S. cohort.

  • Greenness was associated with reduced mortality for those more physically active.

  • Air pollution and physical inactivity were adversely associated with mortality risk.

  • Green, clean environments and physical activity jointly improve health.

Acknowledgments

This report is independent research funded in part by the National Institute of Environmental Health Sciences (ES029846, ES023716, and ES030283) and as part of the Center for Air, Climate, and Energy Solutions (CACES), which was supported under Assistance Agreement No. R835873 awarded by the U.S. Environmental Protection Agency. It has not been formally reviewed by the agency. EPA does not endorse any products or commercial services mentioned in this publication. The views expressed in this document are solely those of authors and do not necessarily reflect those of the U.S. EPA, the Centers for Disease Control and Prevention, the National Center for Health Statistics (NCHS), or the NCHS Research Data Center (RDC). We also acknowledge Brandon Ryan, who contributed substantial material support to developing the NDVI dataset utilized here.

Footnotes

CRediT Roles

Carver J. Coleman: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing - original draft, Writing - review & editing. Ray A. Yeager: Conceptualization, Data curation, Methodology, Writing - review & editing. Daniel W. Riggs: Conceptualization, Data curation, Methodology, Writing - review & editing. Zachari A. Pond: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing - review & editing. Aruni Bhatnagar: Conceptualization, Data curation, Funding acquisition, Project administration, Resources, Supervision, Writing - original draft, Writing - review & editing. C. Arden Pope III: Conceptualization, Formal analysis, Funding acquisition, Project administration, Resources, Supervision, Writing - original draft, Writing - review & editing.

Declaration of Interest

The authors declare they have no actual or potential competing financial interests.

Data Sharing

NHIS data are publicly available at the NCHS website (https://www.cdc.gov/nchs/nhis/), though proposals, signed author statement forms, and a visit to a NCHS Research Data Center are required to access census tract identifiers. Fully public NHIS data is available for several large, Metropolitan areas. Information on obtaining these data is available in this published article (https://doi.org/10.1007/s11869-017-0535-3). Air pollution estimates are available at the Center for Air, Climate, and Energy Solutions website (https://www.caces.us/). NDVI data are available, courtesy of the U.S. Geographical Survey, via the EarthExplorer website (https://doi.org/10.5066/F7H41PNT).

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Andersen ZJ, De Nazelle A, Mendez MA, Garcia-Aymerich J, Hertel O, Tjønneland A et al. 2015. A study of the combined effects of physical activity and air pollution on mortality in elderly urban residents: The danish diet, cancer, and health cohort. Environ Health Perspect 123(6):557–563. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Chen H, Burnett RT, Bai L, Kwong JC, Crouse DL, Lavigne E et al. 2020. Residential greenness and cardiovascular disease incidence, readmission, and mortality. Environ Health Perspect 128(8):087005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Chen J, Hoek G. 2020. Long-term exposure to PM and all-cause and cause-specific mortality: A systematic review and meta-analysis. Environ Int:105974. [DOI] [PubMed] [Google Scholar]
  4. Cole-Hunter T, de Nazelle A, Donaire-Gonzalez D, Kubesch N, Carrasco-Turigas G, Matt F et al. 2018. Estimated effects of air pollution and space-time-activity on cardiopulmonary outcomes in healthy adults: A repeated measures study. Environ Int 111:247–59. [DOI] [PubMed] [Google Scholar]
  5. Coleman CJ, Yeager RA, Riggs DW, Coleman NC, Garcia GR, Bhatnagar A et al. 2021. Greenness, air pollution, and mortality risk: A US cohort study of cancer patients and survivors. Environ Int 157:106797. [DOI] [PubMed] [Google Scholar]
  6. Coleman NC, Burnett RT, Ezzati M, Marshall JD, Robinson AL, Pope CA III. 2020. Fine particulate matter exposure and cancer incidence: Analysis of SEER cancer registry data from 1992–2016. Environ Health Perspect 128(10):107004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Crouse DL, Pinault L, Balram A, Brauer M, Burnett RT, Martin RV et al. 2019. Complex relationships between greenness, air pollution, and mortality in a population-based canadian cohort. Environ Int 128:292–300. [DOI] [PubMed] [Google Scholar]
  8. Crouse DL, Pinault L, Balram A, Hystad P, Peters PA, Chen H et al. 2017. Urban greenness and mortality in canada’s largest cities: A national cohort study. The Lancet Planetary Health 1(7):e289–e297. [DOI] [PubMed] [Google Scholar]
  9. de Keijzer C, Gascon M, Nieuwenhuijsen MJ, Dadvand P. 2016. Long-term green space exposure and cognition across the life course: A systematic review. Curr Environ Health Rep 3(4):468–477. [DOI] [PubMed] [Google Scholar]
  10. Di Q, Wang Y, Zanobetti A, Wang Y, Koutrakis P, Choirat C et al. 2017. Air pollution and mortality in the medicare population. N Engl J Med 376(26):2513–2522. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dockery DW, Pope CA, Xu X, Spengler JD, Ware JH, Fay ME et al. 1993. An association between air pollution and mortality in six US cities. N Engl J Med 329(24):1753–1759. [DOI] [PubMed] [Google Scholar]
  12. Dzhambov AM, Dimitrova DD. 2014. Urban green spaces’ effectiveness as a psychological buffer for the negative health impact of noise pollution: A systematic review. Noise and health 16(70):157. [DOI] [PubMed] [Google Scholar]
  13. Elliott EG, Laden F, James P, Rimm EB, Rexrode KM, Hart JE. 2020. Interaction between long term exposure to fine particulate matter and physical activity, and risk of cardiovascular disease and overall mortality in US women. Environ Health Perspect 128(12):127012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Frew J 2019. The hygiene hypothesis, old friends, and new genes. Front Immunol 10:388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gascon M, Sánchez-Benavides G, Dadvand P, Martínez D, Gramunt N, Gotsens X et al. 2018. Long-term exposure to residential green and blue spaces and anxiety and depression in adults: A cross-sectional study. Environ Res 162:231–239. [DOI] [PubMed] [Google Scholar]
  16. Haber G, Sampson J, Graubard B. 2020. Bias due to berkson error: Issues when using predicted values in place of observed covariates. Biostatistics kxaa002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hartley K, Perazzo J, Brokamp C, Gillespie GL, Cecil KM, LeMasters G et al. 2021. Residential surrounding greenness and self-reported symptoms of anxiety and depression in adolescents. Environ Res 194:110628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Heo S, Bell ML. 2019. The influence of green space on the short-term effects of particulate matter on hospitalization in the US for 2000–2013. Environ Res 174:61–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hwang H, Yook S, Ahn K. 2011. Experimental investigation of submicron and ultrafine soot particle removal by tree leaves. Atmos Environ 45(38):6987–6994. [Google Scholar]
  20. Jaafari S, Shabani AA, Moeinaddini M, Danehkar A, Sakieh Y. 2020. Applying landscape metrics and structural equation modeling to predict the effect of urban green space on air pollution and respiratory mortality in tehran. Environ Monit Assess 192:1–15. [DOI] [PubMed] [Google Scholar]
  21. James P, Banay RF, Hart JE, Laden F. 2015. A review of the health benefits of greenness. Curr Epidemiol Rep 2(2):131–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. James P, Hart JE, Banay RF, Laden F. 2016. Exposure to greenness and mortality in a nationwide prospective cohort study of women. Environ Health Perspect 124(9):1344–1352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. James P, Hart JE, Hipp JA, Mitchell JA, Kerr J, Hurvitz PM et al. 2017. GPS-based exposure to greenness and walkability and accelerometry-based physical activity. Cancer Epidemiol Prevent Biomark 26(4):525–532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jerrett M, Turner MC, Beckerman BS, Pope III CA, Van Donkelaar A, Martin RV et al. 2017. Comparing the health effects of ambient particulate matter estimated using ground-based versus remote sensing exposure estimates. Environ Health Perspect 125(4):552–559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Ji JS, Zhu A, Bai C, Wu C, Yan L, Tang S et al. 2019. Residential greenness and mortality in oldest-old women and men in china: A longitudinal cohort study. The Lancet Planetary Health 3(1):e17–e25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Ji JS, Zhu A, Lv Y, Shi X. 2020. Interaction between residential greenness and air pollution mortality: Analysis of the chinese longitudinal healthy longevity survey. The Lancet Planetary Health 4(3):e107–e115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kasdagli M, Katsouyanni K, de Hoogh K, Lagiou P, Samoli E. 2020. Associations of air pollution and greenness with mortality in greece: An ecological study. Environ Res:110348. [DOI] [PubMed] [Google Scholar]
  28. Kim S, Kim H, Lee J. 2019. Interactions between ambient air particles and greenness on cause specific mortality in seven korean metropolitan cities, 2008–2016. Int J Environ Res Public Health 16(10):1866. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kim S, Bechle M, Hankey S, Sheppard L, Szpiro AA, Marshall JD. 2020. Concentrations of criteria pollutants in the contiguous US, 1979–2015: Role of prediction model parsimony in integrated empirical geographic regression. PloS one 15(2):e0228535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Klompmaker JO, Hoek G, Bloemsma LD, Marra M, Wijga AH, van den Brink C et al. 2020. Surrounding green, air pollution, traffic noise exposure and non-accidental and cause specific mortality. Environ Int 134:105341. [DOI] [PubMed] [Google Scholar]
  31. Laeremans M, Dons E, Avila-Palencia I, Carrasco-Turigas G, Orjuela JP, Anaya E et al. 2018. Short-term effects of physical activity, air pollution and their interaction on the cardiovascular and respiratory system. Environ Int 117:82–90. [DOI] [PubMed] [Google Scholar]
  32. Lefler JS, Higbee JD, Burnett RT, Ezzati M, Coleman NC, Mann DD et al. 2019. Air pollution and mortality in a large, representative US cohort: Multiple-pollutant analyses, and spatial and temporal decompositions. Environ Health 18(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Maimaitiyiming M, Ghulam A, Tiyip T, Pla F, Latorre-carmona P, Halik Ü et al. 2014. ISPRS journal of photogrammetry and remote sensing effects of green space spatial pattern on land surface temperature: Implications for sustainable urban planning and climate change adaptation. ISPRS J Photogramm Remote Sens 89:59–66. [Google Scholar]
  34. Matt F, Cole-Hunter T, Donaire-Gonzalez D, Kubesch N, Martínez D, Carrasco-Turigas G et al. 2016. Acute respiratory response to traffic-related air pollution during physical activity performance. Environ Int 97:45–55. [DOI] [PubMed] [Google Scholar]
  35. Murdaca G, Greco M, Borro M, Gangemi S. 2021. Hygiene hypothesis and autoimmune diseases: A narrative review of clinical evidences and mechanisms. Autoimmun Rev 20(7): 102845. [DOI] [PubMed] [Google Scholar]
  36. Murray CJ, Aravkin AY, Zheng P, Abbafati C, Abbas KM, Abbasi-Kangevari M et al. 2020. Global burden of 87 risk factors in 204 countries and territories, 1990–2019: A systematic analysis for the global burden of disease study 2019. The Lancet 396(10258):1223–1249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. National Center for Health Statistics (NCHS). Division of Health Interview Statistics. 1986. National Health Interview Survey. US Public Health Service, National Center for Health Statistics. Accessed 1 May 2021. ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Dataset_Documentation/NHIS/2014/srvydesc.pdf [Google Scholar]
  38. NCHS. 2021a. National Health Interview Survey, 1986–2014. NHIS data, questionnaires, and related documentation. Accessed 1 May 2021. https://www.cdc.gov/nchs/nhis/dataquestionnaires-documentation.htm
  39. NCHS. 2021b. About the National Health Interview Survey. Accessed 1 May 2021. https://www.cdc.gov/nchs/nhis/about_nhis.htm
  40. NCHS. 2021c. NCHS Data Linked to NDI Mortality Files. Accessed 1 May 2021. https://www.cdc.gov/nchs/data-linkage/mortality.htm
  41. NCHS. 2021d. Adult Physical Activity Questions: List of Questionnaires. Accessed 1 May 2021. https://www.cdc.gov/nchs/nhis/physical_activity/pa_questions.htm
  42. Nieuwenhuijsen MJ, Gascon M, Martinez D, Ponjoan A, Blanch J, Garcia-Gil MdM et al. 2018. Air pollution, noise, blue space, and green space and premature mortality in barcelona: A mega cohort. Int J Environ Res Public Health 15(11):2405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Orellano P, Reynoso J, Quaranta N, Bardach A, Ciapponi A. 2020. Short-term exposure to particulate matter (PM10 and PM2.5), nitrogen dioxide (NO2), and ozone (O3) and all cause and cause-specific mortality: Systematic review and meta-analysis. Environ Int 142:105876. [DOI] [PubMed] [Google Scholar]
  44. 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(2):027002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Paffenbarger RS Jr, Hyde R, Wing AL, Hsieh C. 1986. Physical activity, all-cause mortality, and longevity of college alumni. N Engl J Med 314(10):605–613. [DOI] [PubMed] [Google Scholar]
  46. Pinault LL, Weichenthal S, Crouse DL, Brauer M, Erickson A, van Donkelaar A et al. 2017. Associations between fine particulate matter and mortality in the 2001 canadian census health and environment cohort. Environ Res 159:406–415. [DOI] [PubMed] [Google Scholar]
  47. Pope CA III, Bhatnagar A, McCracken JP, Abplanalp W, Conklin DJ, O’Toole T. 2016. Exposure to fine particulate air pollution is associated with endothelial injury and systemic inflammation. Circ Res 119(11):1204–1214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Pope CA III, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K et al. 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 287(9):1132–1141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Pope CA III, Lefler JS, Ezzati M, Higbee JD, Marshall JD, Kim S et al. 2019. Mortality risk and fine particulate air pollution in a large, representative cohort of US adults. Environ Health Perspect 127(7):077007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Pun VC, Manjourides J, Suh HH. 2018. Association of neighborhood greenness with self-perceived stress, depression and anxiety symptoms in older US adults. Environ Health 17(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Sheppard L, Burnett RT, Szpiro AA, Kim S, Jerrett M, Pope CA et al. 2012. Confounding and exposure measurement error in air pollution epidemiology. Air Qual Atmos Health 5(2):203–216. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Son J, Sabath MB, Lane KJ, Miranda ML, Dominici F, Di Q et al. 2021. Long-term exposure to PM2.5 and mortality for the older population: Effect modification by residential greenness. Epidemiology 32(4):477–486. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Sun S, Sarkar C, Kumari S, James P, Cao W, Lee RS et al. 2020a. Air pollution associated respiratory mortality risk alleviated by residential greenness in the chinese elderly health service cohort. Environ Res 183:109139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sun S, Cao W, Qiu H, Ran J, Lin H, Shen C et al. 2020b. Benefits of physical activity not affected by air pollution: A prospective cohort study. Int J Epidemiol 49(1):142–152. [DOI] [PubMed] [Google Scholar]
  55. Szeremeta B, Zannin PHT. 2009. Analysis and evaluation of soundscapes in public parks through interviews and measurement of noise. Sci Total Environ 407(24):6143–6149. [DOI] [PubMed] [Google Scholar]
  56. Tainio M, de Nazelle AJ, Götschi T, Kahlmeier S, Rojas-Rueda D, Nieuwenhuijsen MJ et al. 2016. Can air pollution negate the health benefits of cycling and walking? Prev Med 87:233–236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Tainio M, Jovanovic Andersen Z, Nieuwenhuijsen MJ, Hu L, de Nazelle A, An R et al. 2021. Air pollution, physical activity and health: A mapping review of the evidence. Environ Int 147:105954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Thygesen M, Engemann K, Holst GJ, Hansen B, Geels C, Brandt J et al. 2020. The association between residential green space in childhood and development of attention deficit hyperactivity disorder: A population-based cohort study. Environ Health Perspect 128(12):127011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. U.S. Department of Health and Human Services (HHS). 2008. Physical Activity Guidelines for Americans. Accessed 1 May 2021. https://health.gov/our-work/physicalactivity/previous-guidelines/2008-physical-activity-guidelines
  60. U.S. Environmental Protection Agency (EPA). Ecoregions. Accessed 4 January 2022. https://www.epa.gov/eco-research/ecoregions
  61. U.S. Geological Survey. 2021. eMODIS NDVI V6 10-day composite images, time period: 2003–2016, 250m resolution, accessed via USGS EarthExplorer website; US Geological Survey: Reston, VA, USA. Accessed 1 January 2019: doi: 10.5066/F7H41PNT. [DOI] [Google Scholar]
  62. Vienneau D, de Hoogh K, Faeh D, Kaufmann M, Wunderli JM, Röösli M. 2017. More than clean air and tranquillity: Residential green is independently associated with decreasing mortality. Environ Int 108:176–184. [DOI] [PubMed] [Google Scholar]
  63. Villeneuve PJ, Jerrett M, Su JG, Burnett RT, Chen H, Wheeler AJ et al. 2012. A cohort study relating urban green space with mortality in ontario, canada. Environ Res 115:51–58. [DOI] [PubMed] [Google Scholar]
  64. Wilker EH, Wu C, McNeely E, Mostofsky E, Spengler J, Wellenius GA et al. 2014. Green space and mortality following ischemic stroke. Environ Res 133:42–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Yitshak-Sade M, James P, Kloog I, Hart JE, Schwartz JD, Laden F et al. 2019. Neighborhood greenness attenuates the adverse effect of PM2. 5 on cardiovascular mortality in neighborhoods of lower socioeconomic status. Int J Environ Res Public Health 16(5):814. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Zijlema WL, Stasinska A, Blake D, Dirgawati M, Flicker L, Yeap BB et al. 2019. The longitudinal association between natural outdoor environments and mortality in 9218 older men from perth, western australia. Environ Int 125:430–436. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1782053-supplement-1.docx (330KB, docx)

RESOURCES