Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Nov 1.
Published in final edited form as: Allergy. 2024 Oct 21;79(11):3036–3046. doi: 10.1111/all.16359

Early-life exposure to residential greenness and risk of asthma in a U.S. bronchiolitis cohort

Wojciech Feleszko 1,2,†, Heidi Makrinioti 1,3,†, Marta Nalej 4, Tadao Ooka 1,5, Zhaozhong Zhu 1, Ashley F Sullivan 1, Tuomas Jartti 6, Kohei Hasegawa 1, Carlos A Camargo Jr 1
PMCID: PMC11560528  NIHMSID: NIHMS2029375  PMID: 39429165

Abstract

Introduction:

Severe bronchiolitis (i.e., bronchiolitis requiring hospitalization) is linked to childhood asthma development. Despite a growing understanding of risk factors for developing post-bronchiolitis asthma, protective factors remain unclear. In this study, we aimed to investigate whether exposure to residential greenness between birth and bronchiolitis hospitalization is associated with asthma and atopic asthma development by age 6 years.

Methods:

We analyzed a U.S. severe bronchiolitis cohort from hospitalization to age 6 years, investigating how the normalized difference vegetation index (NDVI) and chlorophyll index green (CI green), measured in small (100m) and large (500m) radiuses around homes, relate to asthma and atopic asthma by age 6 years. We also explored whether maternal antibiotic use, daycare attendance, and respiratory virus type during hospitalization act as effect modifiers.

Results:

The study cohort included 861 infants, with 239 (28%) developing asthma by age 6 years—152 atopic, 17 non-atopic, and 70 unclassified. Early life residential exposure to high NDVI and CI green levels was associated with lower odds of asthma (ORAdj for NDVI within a 100m radius, 0.18; 95%CI, 0.05-0.78; and ORAdj for CI green levels within a 100m radius, 0.53; 95%CI, 0.31-0.90). Associations also were significant for the development of atopic asthma (ORAdj 0.16; 95%CI, 0.03-0.96; and ORAdj 0.46; 95%CI, 0.25-0.92; respectively). Results were similar for the 500m radius exposures. No effect modification was noted.

Conclusion:

In a U.S. bronchiolitis cohort, exposure to residential greenness between birth and bronchiolitis hospitalization is linked to lower asthma and atopic asthma risk by age 6 years.

Keywords: asthma, bronchiolitis, chlorophyll index, greenness, vegetation

Graphical Abstract

graphic file with name nihms-2029375-f0001.jpg

This study investigates whether exposure to residential greenness between birth and bronchiolitis hospitalization is associated with asthma and atopic asthma development by age 6 years.

We analyzed a severe bronchiolitis cohort from hospitalization to age 6 years, assessing how the NDVI and CI green, measured in small (100m) and large (500m) radiuses around homes, relate to asthma and atopic asthma by age 6 years.

Early-life exposure to residential greenness significantly reduces the risk of developing asthma and atopic asthma by age 6 in infants hospitalized with severe bronchiolitis, highlighting the potential protective role of natural environments in high-risk pediatric populations.

Abbreviations: CI green, chlorophyll index green; NDVI, normalized difference vegetation index; OD, odds ratio.

INTRODUCTION

Severe bronchiolitis (i.e., bronchiolitis requiring hospitalization) is the most frequent cause of hospitalization in the U.S. and Europe (1, 2). Among infants with severe bronchiolitis, approximately 30% will develop childhood asthma (3). Although several risk factors have been identified for post-bronchiolitis asthma development, such as specific respiratory viruses (4), exposure to increased levels of traffic-related air pollution (5), genetic (6), and immune-related factors (7), it remains unclear what factors protect against asthma development in these high-risk children.

Evidence from birth cohort studies suggests that exposure to environmental factors (e.g., residential surrounding greenness) may confer a protective effect against asthma development (8, 9). In these studies, the quantification of residential surrounding greenness has been mainly conducted through the calculation of the normalized difference vegetation index (NDVI) and, to a lesser extent, the leaf area index (LAI) (10). Although most findings support “protective” associations with childhood asthma (11, 12), a few studies highlight non-protective associations between exposures to residential surrounding greenness and the development of atopic asthma (13, 14). Specifically, these observations indicate that prolonged exposure to residential surrounding greenness may be linked to the development of allergic sensitization and atopic asthma (13, 14).

These inconclusive data leave a gap in our understanding of associations between residential surrounding greenness exposures and asthma/atopic asthma, particularly among children at high risk for asthma (i.e., infants with severe bronchiolitis). The investigation of possible underlying mechanisms in this population of children may yield novel interventions for asthma prevention (e.g., microbiome-related products or non-pharmacological interventions, such as “greenness prescriptions”) (15, 16).

Our study was designed to address this knowledge gap by investigating associations between exposure to residential surrounding greenness, from birth until bronchiolitis hospitalization, and asthma development. We accomplished this by analyzing data from the 35th Multicenter Airway Research Collaboration (MARC-35) cohort.

METHODS

A. Study design, setting, and participants

Details of the study design, settings, participants, and data collection have been previously reported (17-19). Briefly, from 2011 to 2014, MARC-35 recruited participants at 17 sites across 14 U.S. states (participating centers are listed in Table E1 in the online supplement, and the proportion of the cohort residing in main Census Regions is shown in Figure E1). Participants include infants (i.e., aged <1 year) hospitalized with a physician diagnosis of bronchiolitis, as defined following the American Academy of Pediatrics (AAP) guidelines (20). We excluded infants with a known heart-lung disease, immunodeficiency, immunosuppression, or gestational age <32 weeks. All patients were treated at the discretion of the treating physician. Of 1,016 infants initially enrolled into the MARC-35 cohort, 921 (91%) completed the run-in procedures (i.e., contact at both 1-week after hospital discharge and 3 weeks after hospitalization) and comprise the longitudinal cohort. Among these, 861 (94% of 921) completed the follow-up for asthma diagnosis by age 6 years and comprise the analytical cohort (21). The institutional review board at each participating hospital approved the study. Written informed consent was obtained from the parent or guardian (21).

B. Exposures: Normalized difference vegetation index (NDVI) and chlorophyll index green (CI green)

The primary exposures consisted of the NDVI and CI green levels within a small (100m [meters]) and a large (500m) radius around the residential address at index hospitalization. We georeferenced 861 participants according to their hospitalization residential address using ArcGIS Online World Geocoding Service (Esri Redlands, California, USA), thereby allowing matching participants to place and track environmental exposures. Average NDVI levels were calculated to represent the residential greenspace exposure during the participants’ baseline evaluation period (i.e., residence at the time of hospitalization). We used ArcMap 10.8.2 to process satellite images and extract the average NDVI levels within a small (100m) and a large (500m) radius around the residential address. We used a similar approach to calculate CI green levels (22, 23). For the calibration and testing of associations within different radiuses around the residential address, average NDVI and CI green levels were also calculated at 200m and 300m radiuses (Figures E2, E3, E4). Furthermore, by collecting residential address data at birth and tracking stability, we identified those infants who had not changed their residential address since birth. In a sensitivity analysis, we then tested associations in this group of infants to investigate whether associations were similar with a prolonged duration of exposure (i.e., between birth and bronchiolitis hospitalization)

C. Outcomes

The primary outcome was asthma by age 6 years. Asthma was defined using an epidemiologic definition, which included a physician’s diagnosis of asthma by age 6 years as reported during parent interviews and either the use of asthma medication (e.g., albuterol inhaler, inhaled corticosteroids) or the presence of asthma-related symptoms (e.g., wheezing, nocturnal cough) in the preceding year (24, 25).

The secondary outcomes were atopic and non-atopic asthma by age 6 years. Atopic asthma was defined as an asthma status (described above) and positive specific IgE (sIgE) to at least one food and/or aeroallergen, either from sIgE testing (Phadia) or ISAC chip at any point during the 6-year follow-up (Table E2) (26). Non-atopic asthma was defined as asthma status with negative specific IgE to all measured foods and/or aeroallergens.

D. Covariates and Potential Effect Modifiers

The main covariates were race/ethnicity, socioeconomic factors (e.g., annual household income, caregiver education), and indoor smoking. In addition, to maximally control for potential confounding by socioeconomic status, we fitted the logistic regression models by adding an additional socioeconomic factor (healthcare insurance type) and a composite socioeconomic status index based on annual household income, caregiver education, and healthcare insurance (Online Supplement methods) (27). Furthermore, to maximally control for traffic-related air pollution exposure, we utilized a proxy measure (i.e., distance from a major road). A major road was defined as a road with 10,000 vehicles or more per day. The distance from a major road was used both as a categorical variable (see Online Supplement methods) and as a dichotomous variable (i.e., less than 100m distance) based on previous evidence from our group (5). We also tested for effect modification by maternal antibiotics use during pregnancy, daycare attendance, and respiratory virus type (i.e., respiratory syncytial virus [RSV]-positive and rhinovirus-positive) of the participant’s index hospitalization.

E. Statistical analysis

We analyzed the NDVI and CI green levels both as continuous and as categorical variables; the radiuses for NDVI and CI green levels reported here are for 100m and 500m. We conducted additional analysis with NDVI and CI green levels at 200m and 300m to investigate the distribution of data at distances other than the closest and longest; we report these findings in the Online Supplement. Furthermore, we categorized the NDVI and CI green levels into three groups: lowest, intermediate, and highest groups. The definitions of these tertiles are described in the online supplement; the distribution of values for each variable at 100m, 200m, 300m, and 500m radiuses are described in Supplement Figures E2 and E3.

To compare the mean NDVI and CI green levels between participants who did and did not complete follow-up for asthma (overall, atopic, and non-atopic), we performed the Wilcoxon rank-sum test. To visualize the correlation between the NDVI and CI green levels at 100m, 200m, 300m, and 500m radiuses, we fitted hexagonal binning plots (Figure E4). To visualize the distribution of NDVI/CI green levels at the radiuses, as mentioned above, across different Census regions, we fitted violin plots (Figures E5 and E6). To visualize the relationship of the two primary exposures (i.e., NDVI and CI green levels within the distances described above) with the asthma outcomes (i.e., asthma and atopic asthma), we fitted locally estimated scatterplot smoothed (loess) plots (Figures E7 and E8). The causal hypothesis for the main analysis, including exposures, covariates, and outcomes, is depicted in the Online Supplement (Figure E9). To investigate the associations of NDVI and CI green levels with each outcome, we constructed unadjusted and adjusted logistic regression models (Table E3, E4). In the multivariable regression models, we adjusted for four patient-level potential confounders (race/ethnicity, annual household income, caregiver education, and indoor smoking) based on a priori knowledge (28). Since our hypothesis focused on the associations between greenness exposures between birth and bronchiolitis hospitalization, we performed a subgroup analysis focusing on the 845 infants (98% of the analytic sample) who had not changed residential addresses since birth (Table E3, E4). All statistical analysis was completed using the R statistical software version R 4.3.1. All p-values were two-tailed, with p<0.05 considered statistically significant.

RESULTS

A. Patient characteristics

The analytic cohort consisted of 861 infants, with a median age of 3.3 months (interquartile range, 2-6 months) at enrollment. Girls accounted for 40% (n=346) of the cohort; 44% were non-Hispanic White, 23% non-Hispanic Black, and 29% Hispanic. Baseline characteristics and clinical outcomes of infants hospitalized with bronchiolitis are summarized in Table 1. The geolocation and mapping of the cohort across Census regions revealed that 35% were residing in the US Northeast, 35% in the South, 21% in the West, and 9% in the Midwest (Figure E1).

Table 1.

Baseline characteristics and clinical outcomes of infants hospitalized with bronchiolitis

Overall
(n=861)
Clinical factors
 Age (month), median (IQR) 3.3 (2-6)
 Female sex 346 (40)
 Race/ethnicity
  Non-Hispanic White 380 (44)
  Non-Hispanic Black 194 (23)
  Hispanic 251 (29)
  Other or unknown 32 (4)
 Annual household income
  Less than $20,000 131 (15)
  $20,000 to $39,999 117 (14)
  $40,000 to $59,999 94 (11)
  $60,000 to $79,999 70 (8)
  $80,000 or $99,999 44 (5)
  $100,000 or more 154 (18)
  Prefer not to answer 244 (28)
  Unknown 2 (1)
  Caregiver education (i.e., completion of high school or equivalent) 693 (80)
  Indoor smoking 131 (15)
 Health insurance type for the child
  Public 519 (60)
  Private 352 (41)
  None 9 (1)
Maternal use of antibiotics during pregnancy 255 (30)
Daycare attendance 200 (23)
Laboratory findings
  RSV-positive 697 (81)
  Rhinovirus-positive 181 (21)
Exposures
 Census regions
  West 181 (21)
  Midwest 77 (9)
  Northeast 301 (35)
  South 301 (35)
 Surrounding greenness indexes, median (IQR)
  NDVI within 100 m 0.20 (0.10-0.28)
  NDVI within a 200 m 0.21 (0.10-0.27)
  NDVI within a 300 m 0.21 (0.10-0.28)
  NDVI within a 500 m 0.21 (0.10-0.29)
 CI green within a 100 m 1.50 (1.27-1.75)
 CI green within a 200 m 1.53 (1.28-1.77)
 CI green within a 300 m 1.54 (1.28-1.77)
 CI green within a 500 m 1.55 (1.28-1.77)
Clinical outcomes
 Asthma by age 6 years† 239 (28)
 Atopic asthma by age 6 years‡ 152 (18)
 Non-atopic asthma by age 6 years ⁂ 17 (2)
Open in a new tab

Abbreviations: CI green, chlorophyll index green; IgE, immunoglobulin E; IQR, interquartile range; NDVI, normalized difference vegetation index; RSV, respiratory syncytial virus

Note: Data are no. (%) of total infants unless otherwise indicated. Percentages may not equal 100, because of rounding and missingness.

†

Infants with a doctor diagnosis of asthma by age 6 years, as reported during parent interviews; AND either medication use (i.e., an inhaled bronchodilator, inhaled corticosteroid, oral corticosteroid/systemic corticosteroid, montelukast) OR asthma-related symptoms (i.e., report of “breathing problems”) between 5.0-5.9 years of age

‡

Infants positive for asthma by age 6 years AND positive specific IgE to at least one food and/or aeroallergen, either from sIgE testing or ISAC chip at any point during the 6-year follow-up

⁂

Infants positive for asthma by age 6 years AND negative specific IgE to all foods and/or aeroallergens, by either sIgE testing or ISAAC chip at any point during the 6-year follow-up. There is missing information regarding lgE levels at any point during the follow-up in 70 patients.

B. Exposures

The distribution of NDVI and CI green values with the visual representation of quantiles was depicted in quantile-quantile (QQ) plots in Figures E2 and E3. Higher NDVI levels correlated with higher CI green levels (Figure E4 and Figures E5-E6 for distributions as per Census region). To visually inspect differences between areas of robust greenness and ambient chlorophyll content and areas of limited greenness and reduced chlorophyll content, we downloaded world imagery maps for high NDVI and CI green levels and low NDVI and CI green levels at the 100m and 500m radiuses, respectively (Figure 1). We visually inspected differences in detailed data on the types of surfaces or covers in land cover images of the areas at the 100m and 500m radiuses (Figure 2).

Figure 1.

Figure 1.

Open in a new tab

World imagery maps for areas with normalized difference vegetation index (NDVI) and chlorophyll index green (CI green) values at the highest and lowest tertiles at 100m and 500m radiuses.

World Imagery maps at 0.5m resolution obtained from Esri, Maxar, Earthstar Geographics, and the GIS User Community. A. A buffer at 100m showing NDVI (0.52) and CI green (2.57) values of the highest tertiles; B. A buffer at 500m showing NDVI (0.51) and CI green (2.49) values of the highest tertiles; C. A buffer at 100m shows NDVI (−0.28) and CI green (0.62) values of the lowest tertiles; D. A buffer at 500m shows NDVI (−0.23) and CI green (0.66) values of the lowest tertiles.

Figure 2.

Figure 2.

Open in a new tab

Land cover images for areas with normalized difference vegetation index (NDVI) and chlorophyll index green (CI green) values at the highest and lowest tertiles at 100m and 500m radiuses

National Land Cover Database for the U.S. at 30m resolution obtained from the U.S. Geological Survey in: A. a buffer at 100m showing NDVI (0.52) and CI green (2.57) values at the highest tertiles; B. a buffer at 500m showing NDVI (0.51) and CI green (2.49) values at the highest tertiles; C. a buffer 100m showing NDVI (−0.28) and CI green (0.62) values at the lowest tertiles; D. a buffer at 500m showing NDVI (−0.23) and CI green (0.66) values at the lowest tertiles.

C. Associations of exposures with asthma by age 6 years

Subsection 1: Associations with asthma by age 6 years

By age 6 years, asthma was diagnosed in 28% of children (n=239). Our analysis revealed significant associations between residential exposure to surrounding greenness and asthma. Specifically, exposure to increased NDVI levels at 100m and 500m radiuses was associated with lower odds for asthma (ORUnadj 0.21; 95%CI, 0.06-0.76; ORUnadj 0.25; 95%CI, 0.07-0.95; respectively) (Figure 3, Table E3). In the multivariable model adjusting for potential confounders, associations between NDVI levels at 100m and 500m radiuses and asthma remained significant (ORAdj 0.18; 95%CI, 0.05-0.78; and ORAdj 0.18; 95%CI, 0.04-0.81; respectively) (Figure 3, Table E3).

Figure 3.

Figure 3.

Open in a new tab

Forest plot with odds ratios (ORs) and 95% confidence intervals (CIs) of unadjusted and adjusted associations between normalized difference vegetation index (NDVI) levels at 100m and 500m radiuses and asthma (left panel) or atopic asthma (right panel) by age 6 years

Unadjusted and adjusted associations of NDVI levels at 100m and 500m radiuses with asthma and atopic asthma outcomes for the analytical cohort (n=861 infants for the asthma outcome and n=774 infants for the atopic asthma outcome) and the cohort of infants who did not move between birth and hospitalization (“no move”; n=845 infants for the asthma outcome and n=774 infants for the atopic asthma outcome). The logistic regression models were adjusted for race/ethnicity, household income, caregiver education, and indoor smoking. P-values for all tested associations, in unadjusted and adjusted models, fall below 0.05. The red dotted line is placed just below the lower OR value (i.e., 0.11).

Abbreviations: AAS, atopic asthma; AS, asthma; CI, confidence interval; OR, odds ratio.

Subsection 2: Sensitivity analysis (infants who have not moved between birth and bronchiolitis hospitalization)

Testing these associations in the subgroup of infants whose families did not change residential address between birth and bronchiolitis hospitalization yielded similar point estimates for both the 100m and the 500m radiuses (Figure 3). In addition, exposure to CI green levels at 100m and 500m radiuses was associated with lower odds for asthma (ORUnadj 0.52; 95%CI, 0.32-0.83; ORUnadj 0.48; 95%CI, 0.29-0.79; respectively) (Figure 4, Table E4). In the multivariable model adjusting for confounders, associations between CI green levels at 100m and 500m radiuses and asthma remained significant (ORAdj 0.53; 95%CI, 0.31-0.90; and ORAdj 0.45; 95%CI, 0.26-0.79; respectively) (Figure 4, Table E4). Similar to the NDVI subgroup analysis, testing these CI green associations in the subgroup of infants whose families did not change residential address between birth and bronchiolitis hospitalization yielded similar point estimates (Figure 4). Moreover, adding an additional socioeconomic factor (healthcare insurance type) and a composite socioeconomic status index yielded similar results (Online Supplement).

Figure 4.

Figure 4.

Open in a new tab

Forest plot with odds ratios (ORs) and 95% confidence intervals (CIs) of unadjusted and adjusted associations between chlorophyll index green (CI green) levels at 100m and 500m radiuses and asthma (left panel) or atopic asthma (right panel) by age 6 years.

Unadjusted and adjusted associations of CI green levels at 100m and 500m radiuses with asthma and atopic asthma outcomes for the analytical cohort (n=861 infants for the asthma outcome and n=774 infants for the atopic asthma outcome) and the cohort of infants who did not move between birth and hospitalization (“no-move”; n=845 infants for the asthma outcome and n=774 infants for the atopic asthma outcome). The logistic regression models were adjusted for race/ethnicity, household income, caregiver education, and indoor smoking. P-values for all tested associations, in unadjusted and adjusted models, fall below 0.05. The red dotted line is placed just at the lower OR value (i.e., 0.40).

Abbreviations: AAS, atopic asthma; AS, asthma; CI, confidence interval; OR, odds ratio.

Subsection 3: NDVI levels as a categorical variable

The highest NDVI levels at 100m radius (i.e., highest tertile levels ≥ 0.270) were associated with lower asthma risk (ORUnadj 0.67; 95%CI, 0.46-0.98, and ORAdj 0.63; 95%CI, 0.42-0.96 ), while the associations were not significant for the 500m radius distance (Table 2). For CI green values, the highest CI green values (i.e., highest tertile levels ≥ 1.680) were significantly associated with asthma development within both 100m and 500m radiuses (Table 3).

Table 2.

Odds ratios for asthma, atopic, and non-atopic asthma under the highest tertile of normalized difference vegetation index (NDVI) values (i.e., above 66.67th centile) at 100m and 500 m radiuses.

Outcome Tertile 3 (range ≥ 0.270)
for NDVI
within 100 m		 Tertile 3 (range ≥ 0.271)
for NDVI
within 500 m
OR (95% CI) p-value OR (95% CI) p-value
Asthma by age 6 years (unadjusted model) (n=287) 0.67 (0.46 – 0.98) 0.03 0.73 (0.50 – 1.05) 0.09
Asthma by age 6 years (adjusted model)* (n=287) 0.63 (0.42 – 0.96) 0.03 0.74 (0.49 – 1.12) 0.16
Atopic asthma by age 6 years (unadjusted model) (n=258) 0.55 (0.35 – 0.87) 0.01 0.69 (0.44 – 1.09) 0.11
Atopic asthma by age 6 years (adjusted model)* (n=258) 0.50 (0.30 – 0.84) 0.009 0.73 (0.44 – 1.2) 0.21
Non-atopic asthma by age 6 years (unadjusted model) (n=258) 0.49 (0.14-1.67) 0.25 0.73 (0.21 – 2.52) 0.62
Open in a new tab

Abbreviations: CI, confidence interval; NDVI, normalized difference vegetation index; OR, odds ratio

*

The logistic regression model has been adjusted for race/ethnicity, household income, caregiver education, and smoking inside the household. The reference group is the lowest tertile, as this is described in the Online Supplement.

Table 3.

Odds ratios for asthma, atopic, and non-atopic asthma under the highest tertiles of chlorophyll index (CI) green values (i.e., above 66.67th centile) at 100m and 500 m radiuses.

Outcome Tertile 3 (range ≥ 1.68)
for CI green
within a 100 m		 Tertile 3 (range ≥ 1.68)
for CI green
within a 500 m
OR (95% CI) p-value OR (95% CI) p-value
Asthma by age 6 years (unadjusted model) (n=287) 0.58 (0.39 – 0.86) 0.006 0.60 (0.41 – 0.89) 0.01
Asthma by age 6 years (adjusted model) *(n=287) 0.47 (0.28 – 0.81) 0.006 0.57 (0.37 – 0.88) 0.01
Atopic asthma by age 6 years (unadjusted model) (n=258) 0.50 (0.31 – 0.81) 0.005 0.63 (0.40 – 1.00) 0.049
Atopic asthma by age 6 years (adjusted model) *(n=258) 0.55 (0.36 – 0.86) 0.008 0.61 (0.36 – 1.03) 0.06
Non-atopic asthma by age 6 years (unadjusted model) (n=258) 0.77 (0.22-2.89) 0.68 0.68 (0.21 – 2.17) 0.51
Open in a new tab

Abbreviations: CI green, Chlorophyll index green; OR, Odds ratio

*

The logistic regression model has been adjusted for race/ethnicity, household income, caregiver education, and smoking inside the household. The reference group is the lowest tertile, as this is described in the Online Supplement.

D. Effect modification by respiratory virus type, maternal antibiotic use, and daycare attendance

The interaction between NDVI and CI green levels and respiratory virus type was not statistically significant (all pinteraction>0.10). Similarly, the interaction between NDVI and CI green levels and maternal use of antibiotics during pregnancy or daycare attendance were not statistically significant (all pinteraction >0.19).

E. Associations of exposures with atopic and non-atopic asthma by age 6 years

Subsection 1: Associations with atopic and non-atopic asthma by age 6 years

We also investigated whether NDVI and CI green levels were significantly associated with atopic (n=152, 64%) or non-atopic (n=17, 7%) asthma. We identified significant associations of NDVI levels at 100m and 500m radiuses with atopic asthma (Figure 3). For example, in the multivariable model adjusting for potential confounders, associations between NDVI levels at 100 m and 500 m radiuses and atopic asthma remained significant (ORAdj 0.16; 95%CI, 0.03-0.96; and ORAdj 0.17; 95%CI, 0.03-0.90; respectively) and similarly lower odds of atopic asthma were observed in children exposed to higher CI green levels at 100 m and 500 m radiuses in children with (ORAdj 0.46; 95%CI, 0.25-0.92; and ORAdj 0.43; 95%CI, 0.22-0.84; respectively).

Subsection 2: Sensitivity analysis (infants who have not moved between birth and bronchiolitis hospitalization)

Similar to the NDVI subgroup analysis, testing the associations in the subgroup of infants whose families did not change residential address between birth and bronchiolitis hospitalization yielded similar results (Figure 4).

Statistical power for non-atopic asthma analyses was limited (n=17). Although results were not statistically significant, point estimates for the association between NDVI and green levels and non-atopic asthma were comparable to those for overall asthma and atopic asthma. Additional analyses, parallel to those done for overall asthma and atopic asthma, yielded consistent, non-significant results (Online Supplement). Associations between exposure to the highest levels of surrounding greenness and atopic asthma development by are depicted in Tables 2 and 3.

DISCUSSION

To our knowledge, this is the first severe bronchiolitis cohort linking exposure to residential greenness (i.e., NDVI and CI green levels) from birth until bronchiolitis hospitalization to lower the risk of developing asthma and atopic asthma by age 6 years. The main analysis was performed in the entire sample of 861 patients, and we also performed a subgroup analysis on 845 infants who did not change residential address between birth and bronchiolitis hospitalization, which confirmed our initial findings. The reported associations remained significant even after adjusting for common confounders, such as socioeconomic deprivation and traffic-related air pollution. Notably, no effect modification by maternal antibiotic use, daycare attendance, or respiratory virus type was noted.

Current research, mainly from birth cohort studies, indicates a growing body of evidence linking exposure to residential greenness from birth with asthma development (28, 29). Although such exposures are also associated with reduced odds of developing atopic asthma (8, 29) as well, this relationship remains unclear (13). There are, for example, studies describing associations between residential exposure to surrounding greenness and increased risk for atopic asthma development (13). One of these studies notably has conflicting results in different ages (i.e., age 5 years in comparison to age 6 years outcomes) (13). Therefore, it is still unclear whether exposure to residential surrounding greenness is associated with a reduced risk of developing atopic asthma. The variation in these findings might be linked to differences in study populations (e.g., children with a high risk of allergies) or specific types of vegetation exposures (e.g., exposure to grass versus trees) (13, 14).

Our study adds to the scientific literature in three ways. First, this is the first severe bronchiolitis cohort identifying protective associations between exposure to residential surrounding greenness from birth until bronchiolitis hospitalization and the development of childhood asthma (and atopic asthma). This is a significant contribution, as severe bronchiolitis and a parental history of asthma are two of the strongest early-life risk factors for the development of childhood asthma (3). Secondly, the focus of our exposure was during the first months of life, a crucial period for immune system adaptation (i.e., development of innate immune responses to various environmental triggers) (30). Thirdly, our study utilizes two greenness indexes (i.e., NDVI and CI green levels). NDVI levels consist of markers of vegetation health across different regions and time periods. In contrast, CI green levels provide information about the presence and concentration of chlorophyll. Therefore, any measurement errors not addressed using a single index (e.g., CI green levels not capturing other aspects of vegetation health as comprehensively as NDVI levels) were addressed.

With regard to the greenness index findings, our study utilized a novel approach. In most cases, NDVI levels were increased when CI green levels were increased too (Figure E4). Residing in areas with increased NDVI and CI green levels (Figures 1 and 2) was associated with reduced odds of asthma and atopic asthma development. The mechanisms underlying this observation could inform important public health policies. For example, the presence of healthy vegetation and increased chlorophyll content may indicate better air quality, which can be beneficial for individuals with asthma, as poor air quality can trigger asthma symptoms (31). In addition, joint healthy vegetation and increased chlorophyll content may indicate a reduced exposure to allergens, as healthier vegetation can act as a natural barrier, filtering and reducing the airborne concentration of allergens (e.g., pollen and mold spores), (29), main drivers of asthma exacerbations in children.

In addition, we included negative values in our NDVI raster. Negative values indicate possible distance from areas with water (32). More specifically, with an increasing interest in blue spaces (i.e., blue areas that represent nature albeit lacking greenness) (33), our NDVI raster, in contrast to other reported studies, included values that could indicate possible distance from areas with water (32). We, therefore, showed that the possible protective association does not lie only in the exposure to natural environments, as it has been described in past studies (16). There is a particular mechanism by which, out of all-natural environments, those with rich vegetation provide a more significant benefit for asthma prevention.

Understanding these mechanisms is crucial since it could shed more light on the associations noted with atopic asthma by age 6 years. We hypothesize that using both NDVI and CI green levels allowed a comprehensive assessment of plant biodiversity in our study (34, 35). Therefore, the biodiversity hypothesis can partially support our study’s findings in regard to atopic asthma.. Environmental biodiversity in residential areas may influence the skin microbiome diversity and composition and the subsequent susceptibility to allergens. (36) In addition, urbanization has been associated with an increase in abundance and diversity of pollen allergens, mainly driven by an increase in allergenic non-native low chlorophyll content plants.(37) ”Ecological analyses of the International Study of Asthma and Allergies in Childhood (ISAAC), a further testament to the hygiene hypothesis, also support these associations by suggesting that exposure to non-allergenic pollen may be protective against respiratory disease development (39, 40). Understanding underlying mechanisms, in particular, regarding the role of biodiversity, is important. Moreover, analysis of plant species around the residential address would have indeed been helpful in determining associations with allergenic pollens.

Our study has several limitations. First, there is a lack of information about the distribution of time spent close to areas of increased surrounding greenness (11). Although this would help with the duration of exposure, we did not follow up with children using daily diaries of time spent close to areas of increased surrounding greenness. Second, for defining atopic asthma, we included specific IgE level data measured at any follow-up point to maintain the cohort's statistical power due to missing sIgE information by age 6 years for some patients. High serum IgE levels during the first 6 years are linked to an increased risk of allergies, recurrent wheezing, and asthma (41).

Third, while our definition of asthma development by age 6 years aligns with many epidemiological studies, it may overestimate asthma cases. Linking our data to a broader set of lung function measurements could elucidate this. Yet, a single lung function measurement does not fully capture the pathophysiology, in contrast to the relatively stable nature of IgE concentrations. Consequently, we refrained from further analysis in this direction. Additionally, the absence of a healthy comparison group may limit the generalizability of our findings. Finally, we measured NDVI and CI green only during the first year, not over the entire 6-year observation period. Longitudinal measurements, accounting for seasonal variations and human interventions, would more accurately assess continuous exposure and its cumulative effects on asthma and allergic conditions in early childhood.

These novel results trigger several research questions that merit investigation. First, the findings require replication in a separate bronchiolitis cohort. Second, it would be helpful to test associations with the development of IgE sensitization to specific aeroallergens (9). Third, these findings call for research that examines the underlying mechanisms, with an emphasis on the role of the upper airway microbiome and its metabolism. Finally, we recognize that different study designs (e.g., randomized controlled trials) will be needed to identify whether exposure to residential surrounding greenness is truly protective among children with severe bronchiolitis or who are otherwise at higher risk of developing asthma (e.g., have a genetic susceptibility associated with increased risk for asthma (6)).

CONCLUSION

Our study presents compelling evidence of associations between exposure to residential greenness and reduced risk of asthma (and atopic asthma) development among infants with severe bronchiolitis. Notably, these associations were not modified by commonly reported risk factors for childhood asthma, such as respiratory virus type, maternal use of antibiotics, and daycare attendance. These findings are novel and may hold important implications for public health and could become instrumental in the primary prevention of childhood asthma.

Supplementary Material

Supinfo

Acknowledgments:

The MARC-35 cohort and investigators are supported by grants UG3/UH3 OD-023253 (Camargo), R01 AI-148338 (Liang & Hasegawa), and K01 AI-153558 (Zhu) from the U.S. National Institutes of Health (Bethesda, USA). WF was supported by the Respira Foundation Science and Travel Grant (Warsaw, Poland). TJ was supported by the Pediatric Research Foundation (Helsinki, Finland).

Footnotes

Conflict of interest: The authors declare no conflict of interest in relation to this work

References:

  • 1.Fujiogi M, Goto T, Yasunaga H, Fujishiro J, Mansbach JM, Camargo CA Jr., et al. Trends in Bronchiolitis Hospitalizations in the United States: 2000-2016. Pediatrics 2019;144(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Del Riccio M, Spreeuwenberg P, Osei-Yeboah R, Johannesen CK, Fernandez LV, Teirlinck AC, et al. Burden of Respiratory Syncytial Virus in the European Union: estimation of RSV-associated hospitalizations in children under 5 years. J Infect Dis 2023;228(11):1528–1538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Nanishi M, Chandran A, Li X, Stanford JB, Alshawabkeh AN, Aschner JL, et al. Association of Severe Bronchiolitis during Infancy with Childhood Asthma Development: An Analysis of the ECHO Consortium. Biomedicines 2022;11(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Makrinioti H, Hasegawa K, Lakoumentas J, Xepapadaki P, Tsolia M, Castro-Rodriguez JA, et al. The role of respiratory syncytial virus- and rhinovirus-induced bronchiolitis in recurrent wheeze and asthma-A systematic review and meta-analysis. Pediatr Allergy Immunol 2022;33(3):e13741. [DOI] [PubMed] [Google Scholar]
  • 5.Freid RD, Qi YS, Espinola JA, Cash RE, Aryan Z, Sullivan AF, et al. Proximity to Major Roads and Risks of Childhood Recurrent Wheeze and Asthma in a Severe Bronchiolitis Cohort. Int J Environ Res Public Health 2021;18(8). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Caliskan M, Bochkov YA, Kreiner-Moller E, Bonnelykke K, Stein MM, Du G, et al. Rhinovirus wheezing illness and genetic risk of childhood-onset asthma. N Engl J Med 2013;368(15):1398–1407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hasegawa K, Hoptay CE, Harmon B, Celedon JC, Mansbach JM, Piedra PA, et al. Association of type 2 cytokines in severe rhinovirus bronchiolitis during infancy with risk of developing asthma: A multicenter prospective study. Allergy 2019;74(7):1374–1377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Fuertes E, Markevych I, Thomas R, Boyd A, Granell R, Mahmoud O, et al. Residential greenspace and lung function up to 24 years of age: The ALSPAC birth cohort. Environ Int 2020;140:105749. [DOI] [PubMed] [Google Scholar]
  • 9.Cavaleiro Rufo J, Paciencia I, Hoffimann E, Moreira A, Barros H, Ribeiro AI. The neighbourhood natural environment is associated with asthma in children: A birth cohort study. Allergy 2021;76(1):348–358. [DOI] [PubMed] [Google Scholar]
  • 10.Liu W, Liu K, Cai J, Ma S, Yang M, Yang Q, et al. Associations between residential greenness and asthma and allergic rhinitis in children: A systematic review and meta-analysis. Sustainable Cities and Society 2023;94:104566. [Google Scholar]
  • 11.Zeng XW, Lowe AJ, Lodge CJ, Heinrich J, Roponen M, Jalava P, et al. Greenness surrounding schools is associated with lower risk of asthma in schoolchildren. Environ Int 2020;143:105967. [DOI] [PubMed] [Google Scholar]
  • 12.Donovan GH, Gatziolis D, Longley I, Douwes J. Vegetation diversity protects against childhood asthma: results from a large New Zealand birth cohort. Nat Plants 2018;4(6):358–364. [DOI] [PubMed] [Google Scholar]
  • 13.Lovasi GS, O'Neil-Dunne JP, Lu JW, Sheehan D, Perzanowski MS, Macfaden SW, et al. Urban tree canopy and asthma, wheeze, rhinitis, and allergic sensitization to tree pollen in a New York City birth cohort. Environ Health Perspect 2013;121(4):494–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Yu H, Zhou Y, Wang R, Qian Z, Knibbs LD, Jalaludin B, et al. Associations between trees and grass presence with childhood asthma prevalence using deep learning image segmentation and a novel green view index. Environ Pollut 2021;286:117582. [DOI] [PubMed] [Google Scholar]
  • 15.Lloyd CM, Saglani S. Early-life respiratory infections and developmental immunity determine lifelong lung health. Nat Immunol 2023;24(8):1234–1243. [DOI] [PubMed] [Google Scholar]
  • 16.Turunen AW, Halonen J, Korpela K, Ojala A, Pasanen T, Siponen T, et al. Cross-sectional associations of different types of nature exposure with psychotropic, antihypertensive and asthma medication. Occup Environ Med 2023;80(2):111–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Hasegawa K, Mansbach JM, Ajami NJ, Espinola JA, Henke DM, Petrosino JF, et al. Association of nasopharyngeal microbiota profiles with bronchiolitis severity in infants hospitalised for bronchiolitis. Eur Respir J 2016;48(5):1329–1339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Stewart CJ, Mansbach JM, Wong MC, Ajami NJ, Petrosino JF, Camargo CA, Jr., et al. Associations of Nasopharyngeal Metabolome and Microbiome with Severity among Infants with Bronchiolitis. A Multiomic Analysis. Am J Respir Crit Care Med 2017;196(7):882–891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Hasegawa K, Piedra PA, Bauer CS, Celedon JC, Mansbach JM, Spergel JM, et al. Nasopharyngeal CCL5 in infants with severe bronchiolitis and risk of recurrent wheezing: A multi-center prospective cohort study. Clin Exp Allergy 2018;48(8):1063–1067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Ralston SL, Lieberthal AS, Meissner HC, Alverson BK, Baley JE, Gadomski AM, et al. Clinical practice guideline: the diagnosis, management, and prevention of bronchiolitis. Pediatrics 2014;134(5):e1474–1502. [DOI] [PubMed] [Google Scholar]
  • 21.Dumas O, Erkkola R, Bergroth E, Hasegawa K, Mansbach JM, Piedra PA, et al. Severe bronchiolitis profiles and risk of asthma development in Finnish children. J Allergy Clin Immunol 2022;149(4):1281–1285 e1281. [DOI] [PubMed] [Google Scholar]
  • 22.Tayade R, Yoon J, Lay L, Khan AL, Yoon Y, Kim Y. Utilization of Spectral Indices for High-Throughput Phenotyping. Plants (Basel) 2022;11(13). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Lu S, Lu X, Zhao W, Liu Y, Wang Z, Omasa K. Comparing vegetation indices for remote chlorophyll measurement of white poplar and Chinese elm leaves with different adaxial and abaxial surfaces. J Exp Bot 2015;66(18):5625–5637. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Camargo CA Jr., Ingham T, Wickens K, Thadhani R, Silvers KM, Epton MJ, et al. Cord-blood 25-hydroxyvitamin D levels and risk of respiratory infection, wheezing, and asthma. Pediatrics 2011;127(1):e180–187. [DOI] [PubMed] [Google Scholar]
  • 25.Islam MS, Huq S, Ahmed S, Roy S, Schwarze J, Sheikh A, et al. Operational definitions of paediatric asthma used in epidemiological studies: A systematic review. J Glob Health 2021;11:04032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Onell A, Hjalle L, Borres MP. Exploring the temporal development of childhood IgE profiles to allergen components. Clin Transl Allergy 2012;2(1):24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Thakur N, Oh SS, Nguyen EA, Martin M, Roth LA, Galanter J, et al. Socioeconomic status and childhood asthma in urban minority youths. The GALA II and SAGE II studies. Am J Respir Crit Care Med 2013;188(10):1202–1209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Hartley K, Ryan PH, Gillespie GL, Perazzo J, Wright JM, Rice GE, et al. Residential greenness, asthma, and lung function among children at high risk of allergic sensitization: a prospective cohort study. Environ Health 2022;21(1):52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Lambert KA, Bowatte G, Tham R, Lodge C, Prendergast L, Heinrich J, et al. Residential greenness and allergic respiratory diseases in children and adolescents - A systematic review and meta-analysis. Environ Res 2017;159:212–221. [DOI] [PubMed] [Google Scholar]
  • 30.Georgountzou A, Papadopoulos NG. Postnatal Innate Immune Development: From Birth to Adulthood. Front Immunol 2017;8:957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Altman MC, Kattan M, O'Connor GT, Murphy RC, Whalen E, LeBeau P, et al. Associations between outdoor air pollutants and non-viral asthma exacerbations and airway inflammatory responses in children and adolescents living in urban areas in the USA: a retrospective secondary analysis. Lancet Planet Health 2023;7(1):e33–e44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Jimenez-Munoz JC, Sobrino JA, Plaza A, Guanter L, Moreno J, Martinez P. Comparison Between Fractional Vegetation Cover Retrievals from Vegetation Indices and Spectral Mixture Analysis: Case Study of PROBA/CHRIS Data Over an Agricultural Area. Sensors (Basel) 2009;9(2):768–793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Labib SM, Lindley S, Huck JJ. Spatial dimensions of the influence of urban green-blue spaces on human health: A systematic review. Environ Res 2020;180:108869. [DOI] [PubMed] [Google Scholar]
  • 34.Cilluffo G, Ferrante G, Fasola S, Malizia V, Montalbano L, Ranzi A, et al. Association between Asthma Control and Exposure to Greenness and Other Outdoor and Indoor Environmental Factors: A Longitudinal Study on a Cohort of Asthmatic Children. Int J Environ Res Public Health 2022;19(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Aerts R, Honnay O, Van Nieuwenhuyse A. Biodiversity and human health: mechanisms and evidence of the positive health effects of diversity in nature and green spaces. Br Med Bull 2018;127(1):5–22. [DOI] [PubMed] [Google Scholar]
  • 36.Bernard-Verdier M, Seitz B, Buchholz S, Kowarik I, Lasuncion Mejia S, Jeschke JM. Grassland allergenicity increases with urbanisation and plant invasions. Ambio 2022;51(11):2261–2277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Hanski I, von Hertzen L, Fyhrquist N, Koskinen K, Torppa K, Laatikainen T, et al. Environmental biodiversity, human microbiota, and allergy are interrelated. Proc Natl Acad Sci U S A 2012;109(21):8334–8339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Haahtela T. A biodiversity hypothesis. Allergy 2019;74(8):1445–1456. [DOI] [PubMed] [Google Scholar]
  • 39.Asher MI, Stewart AW, Mallol J, Montefort S, Lai CK, Ait-Khaled N, et al. Which population level environmental factors are associated with asthma, rhinoconjunctivitis and eczema? Review of the ecological analyses of ISAAC Phase One. Respir Res 2010;11(1):8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Beggs PJ, Clot B, Sofiev M, Johnston FH. Climate change, airborne allergens, and three translational mitigation approaches. EBioMedicine 2023;93:104478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Kulig M, Tacke U, Forster J, Edenharter G, Bergmann R, Lau S, et al. Serum IgE levels during the first 6 years of life. J Pediatr 1999;134(4):453–458. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supinfo
NIHMS2029375-supplement-Supinfo.docx (1.2MB, docx)

RESOURCES