Secondary Logo

Journal Logo


Residential Greenspace in Childhood Reduces Risk of Pediatric Inflammatory Bowel Disease: A Population-Based Cohort Study

Elten, Michael MSc1,2; Benchimol, Eric I. MD, PhD1,3,4,5,6; Fell, Deshayne B. PhD1,3,4; Kuenzig, M. Ellen PhD3,4,5; Smith, Glenys MSc4; Kaplan, Gilaad G. MD, MPH7; Chen, Hong PhD4,8,9,10; Crouse, Dan PhD11,12; Lavigne, Eric PhD1,2,4

Author Information
The American Journal of Gastroenterology: February 2021 - Volume 116 - Issue 2 - p 347-353
doi: 10.14309/ajg.0000000000000990



Crohn's disease (CD) and ulcerative colitis (UC) are the 2 main subtypes of inflammatory bowel disease (IBD), a condition characterized by inflammation in the gastrointestinal tract because of an inappropriate immune response (1). The incidence of IBD increased in the Western nations during the 20th century, and the rates of pediatric-onset IBD continue to increase globally (2,3). Environmental factors related to Western culture and/or rapid socioeconomic development may be implicated in IBD pathogenesis (4).

The health effects of exposure to features of natural vegetation, often referred to as greenspace, is an emerging area of interest in environmental epidemiology. Urbanization has led to a decrease in the population's exposures to plants and the natural environment (5). Despite this, greenspace has not been considered in any published analysis of IBD incidence. Currently, residential greenspace has been shown to negatively correlate with other immune-mediated conditions such as atopic sensitization and asthma (6–8). Greenspace may improve health through many pathways, including the promotion of increased physical activity, psychosocial effects, decreased exposure to air pollution and urban noise, or direct exposure to a wider range of biodiversity (9,10).

The objective of this study was to investigate the potential independent effects of residential greenspace on the risk of developing pediatric-onset IBD.


Study setting and data sources

We conducted a retrospective cohort study using population-based health administrative data for the province of Ontario, Canada's most populous province (2014 population: 13.6 million) (11). We formed this cohort through deterministic linkage of data sets at Institute for Clinical Evaluative Sciences (ICES), an independent nonprofit research institute that collects information on legal residents with valid health cards (>99% of the population). Authors had complete access to all data used in this study. Ethics approval for this study was granted by the Research Ethics Boards of Health Canada and the Ottawa Health Science Network.

The study population was formed by the ICES-derived MOMBABY database, which links hospital admission records of mothers with their infants for all hospital births in Ontario using information from the Discharge Abstract Database, provided by the Canadian Institute for Health Information. This was linked to the Registered Persons Database (RPDB), a dataset provided by the Ontario Ministry of Health and Long-Term Care containing demographic information on those registered for universal provincial health insurance. The data were then linked to the CENSUS data set. Finally, the study population was linked to the Ontario Crohn's and Colitis Cohort (OCCC), which identifies persons with IBD according to algorithms that have been designed for both pediatric and adult populations, described below (12,13). All data sets were linked using unique encoded identifiers and analysed at ICES.

Study population

We included all births between April 1, 1991, and March 31, 2014. We excluded mother-infant pairs who had invalid identification numbers, multiple-birth deliveries, children with an invalid or missing recorded biological sex, or who were ineligible for provincial health insurance at birth. Finally, children who had invalid or missing covariates or greenspace estimates during both pregnancy and childhood were excluded from the study population.

Outcome and exposure assessment

Our study outcome was defined as diagnosis with new-onset IBD before the age of 18 years, as identified by the OCCC validated algorithms. The algorithms were validated in Ontario and use IBD diagnostic codes from the International Classification of Diseases-9 (555, 556) and International Classification of Diseases-10 (K50, K51). The optimal algorithm to capture pediatric-onset IBD cases required 4 physician contacts or 2 hospitalizations with IBD-related diagnostic codes within 3 years if a child had undergone a colonoscopy, and 7 physician contacts or 3 hospitalizations with IBD-related diagnostic codes within 3 years if the child had not undergone a colonoscopy. This algorithm had the following properties when evaluated in the Ontario pediatric population: a sensitivity of 89.6%–91.1%, a specificity of 99.5%–100%, a positive predictive value of 59.2%–76.0%, and a negative predictive value of 99.9%–100% (12). The diagnosis date for each child was the date with the first IBD diagnostic code. The diagnosis (CD or UC) associated with 5 of the last 7 most recent outpatient visits was used to determine the disease subtype. Cases where the subtype was unclassifiable (8.5% of cases) were included in the analyses for overall IBD but were excluded from subtype analyses.

Exposure to greenspace was estimated using the satellite-derived normalized difference vegetation index (NDVI). A validated measure of neighborhood greenspace, NDVI is a continuous measure between −1 and 1, with negative values indicating water and positive values representing the presence of green vegetation (14). For this study, we included only values greater than 0 (>99% of all values). We compiled NDVI readings from satellites Landsat 5 (1990–2011) and Landsat 8 (2013–2017) at a 30 m spatial resolution for each year of the study period, excluding 2012. These annual NDVI values were based on an average of readings taken approximately every 16 days based on the orbit pattern of the satellites (15). We assigned the maximum of the growing season (May 1 to August 31) NDVI values within a 250-m radius around the representative point (i.e., the corresponding geographic coordinates) of each individual's 6-digit residential postal code for each year. In urban areas, the representative point for 6-digit postal codes corresponds typically to one side of a street in a given block or the centre of an apartment building and has positional accuracy within approximately 100–160 m. Positional uncertainty is greater in rural areas. NDVI values calculated using different buffer sizes are often highly correlated, although there is no consensus on the most representative size to use (9). The use of a 250-m radius for NDVI assignment in this study was chosen to balance the need for a resolved exposure metric while limiting the introduction of potential misclassification caused by using 6-digit postal codes. The values were limited to measurements taken during the growing season to exclude months with possible snow coverage, which could provide inaccurate information on the underlying level of vegetation, and maximum values were used instead of averaged values because average values are based on a small amount of readings and may be overly influenced by artifacts in the satellite images caused by cloud coverage.

The mother's residence at delivery was used to assign greenspace exposures during pregnancy, whereas the child's residence as of July 1 in each year as recorded in the RPDB was used to account for possible intraprovincial changes in residence during the childhood period. The residence variable in the RPDB may be updated voluntarily when an individual changes residence, on an interaction with the healthcare system, or when a health card is renewed every 5 years. We then averaged these annual estimates over the pregnancy period to get the pregnancy average and any exposure after birth to get the childhood average exposures. Greenspace exposure was modeled both as a continuous variable (range 0–1) and transformed into quartiles.

Statistical analysis

We used standard and mixed-effects Cox proportional hazards models to evaluate the relationship between exposure to residential greenspace and the risk of developing pediatric-onset IBD, as well as CD and UC subtypes. Follow-up time was age in years, starting from birth and ending at the first of the diagnosis of IBD, 18 years, the end of follow-up (March 31, 2017), death, or ineligibility of provincial health insurance (primarily because of moving outside Ontario). We reported hazard ratios (HRs) and 95% confidence intervals (CIs) for the increase in the interquartile range of NDVI in the given exposure period.

We built models iteratively, starting with a univariate model (model 1). Next, we built a fully adjusted model (model 2) by adding potential confounders as covariates in the regression model. We selected the potential confounders a priori based on previous published literature and adjusted for the following variables: sex (male or female) (12), maternal IBD status (yes or no, obtained from the OCCC) (12,13,16), rural or urban residence at birth (based on the Statistics Canada 2006 Census “rural area” definition for postal codes) (17,18), and median neighborhood household income quintile (based on the 2006 Canadian census of the population and calculated at the dissemination area level, spatial units that contain approximately 400–700 people) (19), which is a validated proxy for individual-level income (20).

Because exposure to residential greenspace is dependent on geographical location, unmeasured risk factors of the population that tend to choose to live in these areas could result in confounding. To control for this, we fit a mixed-effects model (model 3) to the data, with 2 levels of random intercepts corresponding to geographical variables representing spatial clusters. We used this 2-level hierarchal model to fit the random effects of census divisions (large units, equivalent in size to counties) and then census tracts (smaller divisions of neighborhoods available in cities) or census subdivisions (roughly representing municipalities, used when there was no census tract assigned). This spatial survival model may reduce the amount of confounding from unmeasured characteristics (21).

We also explored further adjustments for exposure to ambient air pollution in models 4–7. Average nitrogen dioxide (NO2), fine particulate matter, and ozone exposures from birth to the end of follow-up were added separately (models 4, 5, and 6, respectively) and simultaneously (model 7) to the regression models. Detailed information on how ambient air pollutants were assessed and assigned to the study population are available (see Supplementary Material 1, Supplementary Digital Content 2,, description of air pollution exposure assessment).

Because the gut microbiome is believed to be established early in life, we conducted a preplanned sensitivity analysis to investigate effects of very early life exposures to residential greenspace by limiting the follow-up time to the first 10 years of life. Here, all children were censored after 10 years, and their childhood greenspace estimates were recalculated for this period.

We explored how the effects of greenspace might differ by maternal IBD status, rural/urban residence at birth, and high/low concentrations of each of the air pollution measures (split at the median concentration). We stratified the population by each potential effect measure modifier, ran the fully adjusted models, and then qualitatively compared the magnitude and direction of the HRs. We then tested the statistical significance of the effect measure modification using Wald test of the interaction term derived from the product of the effect modifier and the main exposure.

To investigate possible dose-response relationships, we split the study population into quartiles of greenspace. We reported HRs and 95% CIs for the adjusted model (model 3) at each increasing quartile of greenspace. To test the statistical significance of the dose response, we modeled the categorized exposure as a continuous variable (with values ranging from 1 to 4) and used the Wald test.

All analyses were conducted using R (version 3.1.4) packages survival (version: 2.42-3) and coxme (version: 2.2-5).


From 2,994,625 initial mother-infant pairs, there were 2,715,318 mother-infant pairs eligible for analysis (see Figure 1, Supplementary Digital Content 1,, study population flowchart). Among them, 3,444 developed pediatric-onset IBD (1,906 CD, 1,246 UC, and 292 IBD type unclassifiable). The characteristics of the study population at birth are shown in Table 1. Those that developed IBD were more likely to be men, lived in higher income neighborhoods, and were born to mothers with IBD in urban areas compared with those who did not develop IBD. The median exposure to greenspace for the study population during pregnancy and childhood were 0.69 and 0.72, respectively, whereas interquartile ranges were 0.10 and 0.08, respectively. A summary of quantile NDVI values is presented in Table 1 (see Supplementary Digital Content 2, Individuals were followed up for a total of 33,115,707 years, for an average of approximately 12.2 years per person. The average follow-up time for those born before April 1, 1999 (who would have been able to attain the age of 18 years), was 16.9 years.

Table 1.
Table 1.:
Demographic characteristics of the study population at birth stratified by disease status

The results of the iterative model building process are shown in Table 2. Residential greenspace during childhood was associated with a lower likelihood of developing IBD before the age of 18 based on the unadjusted model (model 1: HR 0.79, CI 0.75–0.82). This association did not change substantially after adding additional covariates and random effects (model 3: HR 0.77, 95% CI 0.74–0.81) or when also considering the effects of the 3 pollutants (model 7: HR 0.82 95% CI 0.77–0.87). These associations were consistent for CD (model 3: HR 0.81, 95% CI 0.76–0.87) and UC (model 3: HR 0.72, 95% CI 0.67–0.78). By contrast, there was no clear association between exposure to greenspace during pregnancy and risk of pediatric IBD, CD, or UC because the CIs of most models crossed the null (Table 2).

Table 2.
Table 2.:
Associations between exposures to greenspace and pediatric-onset IBD and Crohn's disease and ulcerative colitis subtypes

In the sensitivity analyses looking at only the first 10 years of follow-up (Table 3), the effect of greenspace was stronger for overall IBD (model 3: HR 0.70, 95% CI 0.69–0.71), with similar associations in CD (model 3: HR 0.70, 95% CI 0.68–0.72) and UC (model 3: HR 0.71, 95% CI 0.70–0.73). There were no statistically significant associations detected for exposures to greenspace during pregnancy. The results of the effect measure modification analyses are reported in Table 4. There were no major differences in the effect of childhood exposure to greenspace on IBD by rurality of residence at birth (P = 0.33), maternal IBD status (P = 0.72), or by any of the air pollutants (P range: 0.33–0.93). Effect measure modification was also explored across disease subtypes (see Tables 2 and 3, Supplementary Digital Content 2,, effect modification analysis for CD and UC, respectively). Rurality did not modify the association between greenspace and CD (P = 0.84), but the association between greenspace and UC seemed to be different for children in rural and urban areas (P = 0.05). There was no association between greenspace and UC in rural areas (HR 1.00, 95% CI 0.71–1.42), but there was a protective effect in urban areas (HR 0.71, 95% CI 0.66–0.77). NO2 did not modify the association of greenspace and IBD at all (P = 0.93) but seemed to have divergent effects in CD and UC. For CD, those residing in areas high in NO2 saw a slightly weaker protective effect than those living in areas lower in NO2 (HR 0.89, 95% CI 0.81–0.97 and HR 0.79, 95% CI 0.69–0.89, respectively, P = 0.05). For UC, it was the opposite, where those in areas high in NO2 had a somewhat stronger protective effect then those living in areas low in NO2 (HR 0.71, 95% CI 0.64–0.79 and HR 0.87, 95% CI 0.73–1.03, respectively, P = 0.05). There were no other differences across potential modifiers in the disease subtypes.

Table 3.
Table 3.:
Associations between pregnancy and childhood exposures to greenspace and IBD, Crohn's disease and ulcerative colitis during the first 10-year of follow-up
Table 4.
Table 4.:
Effect measure modification analysis for the association between childhood residential greenspace on overall pediatric-onset IBD

The results of the categorical analyses for childhood exposure to greenspace are reported in Table 5. Each increasing quartile of greenspace was associated with a decreasing risk of IBD, CD, and UC. The associations for overall IBD and being in quartile 2 (HR 0.77, 95% CI 0.71–0.84), quartile 3 (HR 0.72, 95% CI 0.66–0.79), and quartile 4 (HR 0.60, 95% CI 0.53–0.68) of greenspace are indicative of a linear dose-response pattern (P value for linear trend <0.0001).

Table 5.
Table 5.:
Association between increasing quartiles of greenspace during childhood and pediatric-onset IBD


In this retrospective population-based cohort study, we found that higher exposure to residential greenspace during childhood was associated with a reduced risk of pediatric-onset IBD. This relationship was observed for both disease subtypes and was stronger for very early onset IBD (diagnosed before 10 years of age). By contrast, greenspace during pregnancy did not show a consistent association with overall IBD or CD and UC subtypes. Future studies are necessary to confirm whether children playing in parks may serve as an approach to prevent some cases of pediatric-onset IBD.

This is the first study to demonstrate higher exposure to greenspace was associated childhood-onset IBD. The protective benefit of interactivity with greenspaces such as parks is consistent with other studies showing that healthy living lifestyle behaviours reduced the risk of IBD including low psychological stress (22), never smoking (23), and increased physical activity (24,25). In addition, greenspace may in part be an important mediator of the consistent finding that living in a rural area is protective of IBD (17,26,27). Similarly, exposures to greenspace may be a residual confounder that explains the inconsistent findings of an association between air pollution and IBD (28,29).

The observed protective association between residential greenspace and pediatric IBD may be explained by a variety of factors, including directly through increased exposure to biodiversity and indirectly through facilitating other healthy lifestyle factors such as physical activity and reduced stress. Exposure to microbes as a protective factor for IBD concurs with the widely supported “hygiene hypothesis,” which theorizes that the reduced microbial exposure and sterile environments that humans experience in industrialized nations may negatively affect the immune system. This reduced microbial load has been previously suggested to be responsible for the global increases in IBD incidence (30). Physical activity in children promoted by proximity and access to greenspace such as by parks and playgrounds may also be exerting influence on IBD risk (10,24). This could be because of changes in the composition of the gut microbiota caused by recreational exercise that has been reported in several animal studies (31).

This study is strengthened by its population-based design, which tend to be more robust to selection bias. In addition, we were able to consider the effects of other environmental exposures that may be spatially correlated with greenspace, such as urban/rural residence and exposures to air pollutants. Another strength is that we were able to assess exposures throughout the entire life-course starting from the estimated date of conception. This allowed us to characterize both prenatal and postnatal effects.

Several limitations to this study should be considered. First, exposure to greenspace may be spatially correlated with other unconsidered environmental risk factors or unmeasured characteristics of the population. For instance, those that reside in urban areas with high greenspace tend to be wealthier are less likely to be ethnic minorities and have better health literacy and behaviours (32). To help control for this possible confounding, competing environmental factors (rurality and air pollution), neighborhood income, and random effects representing geographical variables were added in the regression models. Another limitation of the study includes the use of secondary data that were not collected for research purposes. This may mean that there may be errors in diagnostic coding or changes in coding practice over the course of the study that are not accounted for, although as long-term estimates of greenspace were used, this would be unlikely to greatly affect our results. A small amount of misclassification of IBD diagnosis based on the accuracy of the algorithms is present as well, but this would not differ greatly across greenspace levels, limiting the impact on the observed associations. Residual confounding may be caused by confounders that were not available in the health administrative data sources, such as ethnicity or genetics. However, recent immigrants (that typically have lower rates of IBD) have lower average NDVI values in Canadian urban cities (33), meaning that the strength of our associations may have been underestimated. Finally, most of Ontario's population live in urban areas and the vegetation and other environmental exposures may be unique to this population, so caution should be exercised when interpreting the results of this study and extrapolating to other geographic regions.

Greenspace is an emerging area in environmental health research, and this study is the first to investigate and report its protective effect with IBD. Future studies should look at using alternative measures to NDVI to evaluate the specific features of greenspace responsible for this finding (such as type or access to vegetation) and look for effects on the gut microbiome, immune dysregulation, and genetic predisposition in IBD. Mediation analyses involving these factors are needed to identify possible causal mechanisms. Intervention studies designed to target specific environmental risk factors to reduce the risk of IBD should include greenspace in study design. It is possible that by changing our immediate environment, we could prevent childhood-onset IBD.


Guarantor of the article: Eric Lavigne, PhD.

Specific author contributions: E.L., E.I.B., and M.E.: concept and design. M.E., G.S., E.L., and E.I.B.: data collection. All authors: results interpretations. M.E., E.I.B., E.L., and G.G.K.: manuscript writing. All authors: manuscript proofing.

Financial support: This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). This study also received funding from: Health Canada. Parts of this material are based on data and information compiled and provided by: MOHLTC, Canadian Institute for Health Information. The analyses, conclusions, opinions and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.

Potential competing interests: M.E.K. was supported by a Post-Doctoral Fellowship Award from the Canadian Institutes of Health Research (CIHR), Canadian Association of Gastroenterology (CAG), Crohn's and Colitis Canada, and Mitacs Canada. E.I.B. was supported by a New Investigator Award from the CIHR, CAG and Crohn's and Colitis Canada; and by the Career Enhancement Program of the Canadian Child Health Clinician Scientist Program. The remaining authors have no financial, consultant, institutional or other relationships that might lead to bias or a conflict of interest to declare.

Study Highlights


  • ✓ Environmental exposures such as urbanization and air pollution have been associated with increased risk of inflammatory bowel disease (IBD) in children.
  • ✓ Greenspace is likely protective of other immune-mediated conditions, but the effects on IBD are unknown.


  • ✓ This is the first study to show that greenspace is protective of IBD diagnosis during childhood.
  • ✓ Lack of greenspace may help explain previous findings of higher incidence of IBD for those in urban areas.
  • ✓ Increased availability and interactivity with greenspace may represent a new avenue to prevent development of IBD in children.


NDVI metrics, indexed to DMTI Spatial Inc. postal codes, were provided by CANUE (Canadian Urban Environmental Health Research Consortium).


1. Podolsky DK. Inflammatory bowel disease. N Engl J Med 1991;325:928–37.
2. Molodecky NA, Soon IS, Rabi DM, et al. Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review. Gastroenterology 2012;142:46–54.e42.
3. Benchimol EI, Bernstein CN, Bitton A, et al. Trends in epidemiology of pediatric inflammatory bowel disease in Canada: Distributed network analysis of multiple population-based provincial health administrative databases. Am J Gastroenterol 2017;112:1120–34.
4. Ng SC, Yun Shi H, Hamidi N, et al. Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: A systematic review of population-based studies. Lancet 2017;390:2769–78.
5. Pauleit S, Ennos R, Golding Y. Modeling the environmental impacts of urban land use and land cover change: A study in Merseyside, UK. Landsc Urban Plan 2005;71:295–310.
6. Ruokolainen L, von Hertzen L, Fyhrquist N, et al. Green areas around homes reduce atopic sensitization in children. Allergy 2015;70:195–202.
7. Lovasi GS, Quinn JW, Neckerman KM, et al. Children living in areas with more street trees have lower prevalence of asthma. J Epidemiol Community Health 2008;62:647–9.
8. Sbihi H, Tamburic L, Koehoorn M, et al. Greenness and incident childhood asthma: A 10-year follow-up in a population-based birth cohort to. Am J Respir Crit Care Med 2015;192:1131–3.
9. Markevych I, Schoierer J, Hartig T, et al. Exploring pathways linking greenspace to health: Theoretical and methodological guidance. Environ Res 2017;158:301–17.
10. Almanza E, Jerrett M, Dunton G, et al. A study of community design, greenness, and physical activity in children using satellite, GPS and accelerometer data. Health Place 2012;18:46–54.
11. Statistics Canada. Table 17-10-0005-01: Population estimates on July 1st, by age and sex.
12. Benchimol EI, Guttmann A, Griffiths AM, et al. Increasing incidence of paediatric inflammatory bowel disease in Ontario, Canada: Evidence from health administrative data. Gut 2009;58:1490–7.
13. Benchimol EI, Guttmann A, Mack DR, et al. Validation of international algorithms to identify adults with inflammatory bowel disease in health administrative data from Ontario, Canada. J Clin Epidemiol 2014;67:887–96.
14. Rhew IC, Vander Stoep A, Kearney A, et al. Validation of the normalized difference vegetation index as a measure of neighborhood greenness. Ann Epidemiol 2011;21:946–52.
15. Robinson NP, Allred BW, Jones MO, et al. A dynamic landsat derived normalized difference vegetation index (NDVI) product for the conterminous United States. Remote Sens 2017;9:1–14.
16. Aujnarain A, Mack DR, Benchimol EI. The role of the environment in the development of pediatric inflammatory bowel disease. Curr Gastroenterol Rep 2013;15:326.
17. Benchimol EI, Kaplan GG, Otley AR, et al. Rural and urban residence during early life is associated with a lower risk of inflammatory bowel disease: A population-based inception and birth cohort study. Am J Gastroenterol 2017;112:1412–22.
18. du Plessis V, Beshiri R, Bollman RD, et al. Rural and small town Canada analysis bulletin. Rural Small T Can Anal Bull 2001;3:1–17.
19. Blanchard JF, Bernstein CN, Wajda A, et al. Small-area variations and sociodemographic correlates for the incidence of Crohn's disease and ulcerative colitis. Am J Epidemiol 2001;154:328–35.
20. Glazier RH, Creatore MI, Agha MM, et al. Socioeconomic misclassification in Ontario's Health Care Registry. Can J Public Heal 1996;94:140–3.
21. Sheppard L, Burnett RT, Szpiro AA, et al. Confounding and exposure measurement error in air pollution epidemiology. Air Qual Atmos Heal 2012;5:203–16.
22. Mawdsley JE, Rampton DS. Psychological stress in IBD: New insights into pathogenic and therapeutic implications. Gut 2005;54:1481–91.
23. Mahid SS, Minor KS, Soto RE, et al. Smoking and inflammatory bowel disease: A meta-analysis. Mayo Clin Proc 2006;81:1462–71.
24. Wang Q, Xu KQ, Qin XR, et al. Association between physical activity and inflammatory bowel disease risk: A meta-analysis. Dig Liver Dis 2016;48:1425–31.
25. Ananthakrishnan AN, Kaplan GG, Ng SC. Changing global epidemiology of inflammatory bowel diseases: Sustaining health care delivery into the 21st century. Clin Gastroenterol Hepatol 2020;18:1252–60.
26. Soon IS, Molodecky NA, Rabi DM, et al. The relationship between urban environment and the inflammatory bowel diseases: A systematic review and meta-analysis. BMC Gastroenterol 2012;12:51.
27. Verheij RA, Maas J, Groenewegen PP. Urban-rural health differences and the availability of green space. Eur Urban Reg Stud 2008;15:307–16.
28. Kaplan GG, Hubbard J, Korzenik J, et al. The inflammatory bowel diseases and ambient air pollution: A novel association. Am J Gastroenterol 2010;105:2412–9.
29. Elten M, Benchimol EI, Fell DB, et al. Ambient air pollution and the risk of pediatric-onset inflammatory bowel disease: A population-based cohort study. Environ Int 2020;138:105676.
30. Koloski NA, Bret L, Radford-Smith G. Hygiene hypothesis in inflammatory bowel disease: A critical review of the literature. World J Gastroenterol 2008;14:165–73.
31. Cook MD, Allen JM, Pence BD, et al. Exercise and gut immune function: Evidence of alterations in colon immune cell homeostasis and microbiome characteristics with exercise training. Immunol Cell Biol 2016;94:158–63.
32. Pham TTH, Apparicio P, Séguin AM, et al. Spatial distribution of vegetation in Montreal: An uneven distribution or environmental inequity?. Landsc Urban Plan 2012;107:214–24.
33. Villeneuve PJ, Jerrett M, Su JG, et al. A cohort study relating urban green space with mortality in Ontario, Canada. Environ Res 2012;115:51–8.

Supplemental Digital Content

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of The American College of Gastroenterology