Residential Green and Blue Space, the Natural Environment, Genetic Susceptibility, and Incident COPD

Abstract

Background

Although cigarette smoking is the predominant risk factor for chronic obstructive pulmonary disease (COPD), environmental exposures and genetic susceptibility also contribute to disease risk, yet their joint effects remain unclear. We therefore prospectively examined the associations of residential green and blue space and the natural environment (GBN) with incident COPD across levels of genetic susceptibility.

Methods

This population-based prospective cohort study included 385,502 UK Biobank participants. Residential GBN exposures within 300 m and 1000 m buffers were assessed. Genetic susceptibility was quantified using a polygenic risk score based on COPD-associated variants identified in a genome-wide association study. Cox proportional hazards models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations between GBN exposures, genetic risk, and incident COPD. Gene–environment interactions were evaluated on multiplicative and additive scales.

Results

Higher exposure to green space and the natural environment was associated with a lower risk of COPD (green space: HR = 0.87 [95% CI, 0.83–0.92], natural environment: HR = 0.83 [95% CI, 0.78–0.87] for the 1000 m buffer). Associations with blue space were weak and inconsistent. The natural environment exposure was inversely associated with COPD risk across genetic risk strata, with no significant multiplicative or additive interactions observed.

Conclusion

Greater residential exposure to green space and the natural environment was associated with lower COPD risk, with no evidence that genetic susceptibility modified these associations, highlighting the potential role of environmental interventions in COPD prevention.

Introduction

Chronic obstructive pulmonary disease (COPD) is a complex respiratory disorder characterized by persistent and progressive airflow limitation, affecting more than 400 million people globally and remaining the third leading cause of death worldwide, with its burden projected to continue rising.^1–3 The incidence of COPD continues to rise, driven not only by demographic aging but also by ongoing exposure to environmental risk factors and lifestyle determinants.⁴ Beyond the well-established role of cigarette smoking, growing evidence highlights the importance of broader environmental determinants, such as urbanization, industrial emissions, and traffic-related air pollution, in the development and progression of COPD.⁵

In parallel, increasing attention has been directed toward the potential respiratory benefits of contact with natural environments, including vegetation, parks, water bodies, and other natural features. Exposure to these environments has been linked to reduced mortality and improved cardiovascular and mental health outcomes.^6,7 However, epidemiological evidence on the role of green and blue space and the natural environment (GBN) in COPD is limited and inconsistent. Some studies have identified residential green space as a potential protective factor associated with better lung function.^8,9 Potential mechanisms include vegetation-mediated filtration of airborne pollutants, greater opportunities for physical activity, and reduced heat exposure and cumulative physiological stress.^10–13 However, findings have been mixed, with some studies reporting null or even adverse associations.^14,15

While environmental exposures are fundamental to COPD etiology, susceptibility to the disease cannot be fully explained by smoking or air pollution alone. In addition to alpha-1 antitrypsin deficiency, genetic loci such as hedgehog interacting protein (HHIP) and family with sequence similarity 13 member A (FAM13A) are known to be associated with increased COPD risk, reduced lung function, and disease severity.^16,17 Polygenic risk scores (PRSs), which aggregate multiple COPD-associated alleles, are predictive of incident all-cause COPD and enable quantification of genetic susceptibility.¹⁸ Therefore, examining whether genetic susceptibility, as quantified by a PRS, modifies the association between environmental exposures and COPD may clarify gene–environment interplay and inform precision prevention strategies.

Despite the increasing recognition of the gene–environment interplay in respiratory diseases, evidence examining the joint effects of GBN exposures and genetic susceptibility on COPD risk remains extremely limited. Most previous studies have examined environmental determinants or genetic predisposition separately and have often relied on cross-sectional designs or self-reported outcomes, limiting causal inference. Large prospective population cohorts with detailed environmental and genetic data therefore provide an important opportunity to overcome these limitations and clarify whether the benefits of GBN exposures extend to individuals with high genetic vulnerability.

Therefore, in this study, we investigated the associations between GBN exposures and incident COPD using data from 385,502 participants in the UK Biobank, a large population-based prospective cohort. We further examined whether genetic susceptibility, quantified through the COPD-PRS, modified these associations.

Methods

Study Population

This cohort study used data from the UK Biobank, a population-based prospective cohort of over 500,000 participants recruited between 2006 and 2010 from 22 research centers across the United Kingdom.¹⁹ The North West Multicenter Research Ethics Committee approved the UK Biobank study, and all participants provided written informed consent. Further details about the cohort are available at https://www.ukbiobank.ac.uk/.

Based on 502,301 individuals initially enrolled in UK Biobank, we excluded 1,297 who withdrew consent or were lost to follow-up, 61,386 with missing GBN exposure data, and 13,339 with missing covariate data (overall missingness <5%). We further excluded 9,148 participants with missing PRS data and 24,036 participants of non-European ancestry. Finally, 7,593 individuals with pre-existing COPD at baseline were excluded, resulting in 385,502 participants included in the final analysis (Figure S1).

Exposures

GBN exposures were defined as the proportion of green space, blue space (water), and the natural environment within 300 m and 1000 m buffers surrounding each participant’s residence, calculated as the percentage of the total area within each buffer classified as the corresponding environmental type, consistent with previous studies.^20,21

Green and blue space data were obtained from the 2005 Generalized Land Use Database (GLUD) for England (Department for Communities and Local Government, 2007), sourced from Neighborhood Statistics (http://www.neighbourhood.statistics.gov.uk/) and available for England only. Domestic gardens were classified separately from green space, potentially underestimating vegetation exposure in garden-rich areas, while blue space captured only mapped water coverage, not water-body type, quality and accessibility.

Natural environment data were derived from the Centre for Ecology and Hydrology’s (CEH) 2007 Land Cover Map (LCM).²² In the LCM classification, land was divided into natural and built environments, with gardens included in the built environment category, differing from their separate classification in GLUD. Besides, the minimum mapping unit for the LCM was 0.5 ha, meaning that smaller patches of the natural environment were not captured.

GLUD and CEH environmental data were spatially allocated to the defined buffers via the Geospatial Modelling Environment (GME; http://www.spatialecology.com/gme/). Although the natural environment measure partially overlapped with green and blue space, green and blue space represented specific GLUD land-use categories, whereas the LCM-based natural environment measure encompassed a broader range of natural land-cover types. Given the restricted availability of GLUD data, the analyses were limited to participants residing in England. Further details are available on the UK Biobank website: https://biobank.ndph.ox.ac.uk/showcase/refer.cgi?id=15374. In this study, GBN exposures were categorized into quartiles.²³

Outcomes

Participants were followed from the date of their first assessment center visit until the earliest of COPD diagnosis, death, or end of follow-up. COPD diagnoses were obtained from the UK Biobank “first occurrence” dataset (data category 2410), derived from multiple linked health records including death registries, primary care data, hospital admissions, and self-reports. In total, 16,480 participants with COPD were identified. Consistent with previous research, COPD was coded using the International Classification of Diseases, Tenth Revision (ICD-10): J43 and J44.²⁴

Polygenic Risk Score

In this study, COPD-PRS was constructed using 22 single nucleotide polymorphisms (SNPs) with minor allele frequencies > 0.05, identified from a previous genome-wide association study (GWAS).²⁵ To avoid sample overlap, we used summary statistics from this GWAS after excluding UK Biobank participants. The effect sizes and related information for each SNP are presented in Table S1.

PRSs were computed for all UK Biobank participants using PLINK 2.0 (https://www.cog-genomics.org/plink/2.0/). The weighted PRS was calculated as follows:

represents the number of risk alleles (0, 1, or 2) for the SNP, and denotes the corresponding effect size derived from the reference GWAS. The scaling factor was applied to standardize the total contribution of the 22 SNPs, ensuring comparability across variants. Missing genotypes were imputed using mean substitution.

A higher PRS indicates greater genetic susceptibility to COPD. Participants were categorized by COPD-PRS percentiles into low (lowest quintile), intermediate (quintiles 2–4), and high (highest quintile) genetic risk categories.

Covariates

Analyses were adjusted for potential confounders associated with GBN, COPD, and COPD-PRS, including age, sex, body mass index (BMI), household income before tax, educational attainment, smoking status and home area. The home areas were classified as urban or other areas based on participants’ home postcodes linked with 2001 census data from the Office for National Statistics. For covariates with more than 5% missingness, an “unknown” category was created. Participants with missing values for covariates with less than 5% missingness were excluded from the analysis.

Statistical Analysis

The general demographic characteristics of the study population were summarized using Student’s t-tests for normally distributed continuous variables, Wilcoxon rank-sum tests for non-normally distributed continuous variables, and Pearson’s χ²-tests for categorical variables. Continuous variables are expressed as the mean (standard deviation, SD) or median (interquartile range, IQR), and categorical variables as frequency (%).

Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations between GBN exposures and incident COPD, genetic risk categories and COPD risk. To minimize potential collinearity, the GBN indicators were modeled separately. Model 1 was adjusted for sex and age, and Model 2 was further adjusted for BMI, household income, educational attainment, smoking status, home area, and either genetic risk or GBN categories.

For environmental exposures associated with incident COPD in Cox models, mediation analyses were performed to evaluate whether particulate matter with an aerodynamic diameter ≤2.5 μm (PM_2.5) and nitrogen dioxide (NO₂) mediated these associations using the “mediation” package in R. Air pollution estimates for 2010 were derived from land-use regression (LUR) models developed within the European Study of Cohorts for Air Pollution Effects (ESCAPE, http://www.escapeproject.eu/). Environmental exposures were standardized using z-scores, and mediation effects were estimated for each SD increase in exposure. The average causal mediation effect (ACME), average direct effect (ADE), total effect, and proportion mediated were estimated using 1,000 non-parametric bootstrap simulations. All models were adjusted for the covariates included in Model 2.

The joint effects of GBN and genetic risk categories on COPD were examined using 12 combined groups, with the low genetic risk group and the highest quartile of exposure as the reference category. Additive interactions were evaluated using the relative excess risk due to interaction (RERI), the attributable proportion due to interaction (AP), and the synergy index (SI). Additive interaction was considered present when the 95% CI for RERI and AP excluded 0 and the 95% CI for SI excluded 1. For interaction analyses, GBN and PRS were dichotomized into low and high groups based on the median (50th percentile).

Restricted cubic spline (RCS) analyses were performed with Cox proportional hazards models to explore the dose–response relationships between GBN and the risk of incident COPD. The Cox models were adjusted for covariates in Model 2.

Several sensitivity analyses were conducted to evaluate the robustness of the main findings. We repeated analyses using a broader COPD definition (ICD-10 J41–J44), natural environment exposure measures in the full eligible population with available LCM data, and after excluding COPD cases diagnosed within the first two years of follow-up. Additional adjustments were made for alcohol consumption, physical activity, cardiovascular conditions, diabetes, and parental history of COPD. Alcohol consumption and physical activity were assessed using UK Biobank touchscreen questionnaire data at baseline.²³

All analyses were performed using R software, version 4.2.2, and statistical significance was defined as P < 0.05.

Results

Baseline Characteristics

As shown in Table 1, among the 385,502 participants included in the analysis, the mean age was 56.7 years (SD, 8.0 years), and 45.6% were male. During follow-up, 16,480 participants (4.3%) developed COPD. Compared with those who did not develop COPD, individuals with COPD were older (61.0 vs 56.6 years), more likely to be male (54.5% vs 45.2%), and had a higher BMI (27.6 vs 26.6 kg/m²).

Table 1.

Baseline Characteristics of Participants

Characteristic	Overall N = 385,502	Non-COPD N = 369,022	COPD N = 16,480	P value
Body-mass index, kg/m2	26.7 (24.1, 29.8)	26.6 (24.1, 29.7)	27.6 (24.6, 31.2)	<0.001
Age, years	56.7 (8.0)	56.6 (8.0)	61.0 (6.5)	<0.001
Sex				<0.001
Female	209,810 (54.4%)	202,309 (54.8%)	7,501 (45.5%)
Male	175,692 (45.6%)	166,713 (45.2%)	8,979 (54.5%)
Home area				<0.001
Urban	326,775 (84.8%)	312,012 (84.6%)	14,763 (89.6%)
Others	58,727 (15.2%)	57,010 (15.4%)	1,717 (10.4%)
Household income before tax per year				<0.001
Less than 18,000	72,194 (18.7%)	66,044 (17.9%)	6,150 (37.3%)
18,000–30,999	85,177 (22.1%)	81,319 (22.0%)	3,858 (23.4%)
31,000–51,999	88,221 (22.9%)	85,986 (23.3%)	2,235 (13.6%)
52,000–100,000	69,392 (18.0%)	68,366 (18.5%)	1,026 (6.2%)
Greater than 100,000	18,656 (4.8%)	18,460 (5.0%)	196 (1.2%)
Unknown a	51,862 (13.5%)	48,847 (13.2%)	3,015 (18.3%)
Educational group				<0.001
College or university degree	123,590 (32.1%)	121,090 (32.8%)	2,500 (15.2%)
Any school degree (A-level, AS-level, O-level, GCSE, CSE)	151,756 (39.4%)	146,632 (39.7%)	5,124 (31.1%)
Vocational qualification (NVQ, HND, or HNC) or other  professional qualifications	45,346 (11.8%)	42,838 (11.6%)	2,508 (15.2%)
None of the above	64,810 (16.8%)	58,462 (15.8%)	6,348 (38.5%)
Smoking status				<0.001
Never smokers	209,910 (54.5%)	206,760 (56.0%)	3,150 (19.1%)
Previous smokers	137,483 (35.7%)	130,020 (35.2%)	7,463 (45.3%)
Current smokers	38,109 (9.9%)	32,242 (8.7%)	5,867 (35.6%)
Genetic risk category b				<0.001
Low	77,102 (20.0%)	74,137 (20.1%)	2,965 (18.0%)
Intermediate	231,299 (60.0%)	221,456 (60.0%)	9,843 (59.7%)
High	77,101 (20.0%)	73,429 (19.9%)	3,672 (22.3%)

Notes: Data are presented as the mean (SD), n (%) or median (IQR). ^aThe “Unknown” category included “prefer not to answer,” “do not know,” and missing values in the UK Biobank. ^bGenetic risk category defined according to a polygenic risk score as low (lowest quintile), intermediate (quintiles 2 to 4), or high (highest quintile).

Abbreviations: COPD, Chronic obstructive pulmonary disease; GCSE, General certificate of secondary education; CSE, Certificate of secondary education; NVQ, National vocational qualification; HND, Higher national diploma; HNC, Higher national certificate.

Socioeconomic disparities were evident: participants who developed COPD were more likely to report lower household income (37.3% earning < £18,000 vs 17.9% among those without COPD) and lower educational attainment (15.2% with a college or university degree vs 32.8%). The distribution of genetic risk categories also differed significantly, with a higher proportion of participants who developed COPD in high genetic risk (22.3% vs 19.9%). Similarly, in the descriptive statistics by genetic risk category (Table S2), the proportion of participants who developed COPD increased from 3.8% in the low-risk group to 4.8% in the high-risk group.

Baseline characteristics according to GBN exposures are summarized in Tables S3–S5. Age, sex, and BMI were broadly similar across groups, while green space and natural environment exposures showed differences in urbanicity, socioeconomic status, and lifestyle factors. Compared with the lowest quartile, the highest quartile of green space and natural environment exposure within the 1000 m buffer had lower proportions of current smokers (7.5% vs 12.4% and 7.4% vs 13.0%, respectively), urban residents (44.0% vs 99.9% and 44.6% vs 99.9%), and participants with annual household income below £18,000 (14.9% vs 19.3% and 14.7% vs 20.7%). The proportion of participants who developed COPD was lower among participants with higher green space and natural environment exposure (1000 m buffer: 3.4% vs 4.6% and 3.3% vs 4.9%, respectively).

Association of GBN with Incident COPD

Higher exposure to the natural environment was consistently associated with a lower COPD risk across both buffer distances (Table 2). For green space, inverse associations were more evident for the 1000 m buffer and at the highest exposure quartile, whereas the association within the 300 m buffer was non-monotonic. Specifically, participants in the highest quartile of green space exposure had a significantly lower risk of COPD than those in the lowest quartile (HR 0.94 [95% CI, 0.89–0.98] for 300 m; HR 0.87 [95% CI, 0.83–0.92] for 1000 m; P for trend < 0.001). Greater exposure to natural environment was also inversely associated with COPD risk (HR 0.87 [95% CI, 0.83–0.91] for 300 m; HR 0.83 [95% CI, 0.78–0.87] for 1000 m; P for trend < 0.001). Further adjustment for genetic risk categories in Model 2 did not materially change the results, suggesting that the observed associations of green space and natural environment with COPD were independent of genetic susceptibility.

Table 2.

Independent Associations of Green Space, Blue Space and the Natural Environment with COPD

Exposure	No.of Cases/Person-Years	Model 1, HR (95% CI)	Model 2, HR (95% CI)
GS, 300 m buffer
First quartile (the lowest)	4,051/1,247,584	1[Reference]	1[Reference]
Second quartile	4,763/1,253,613	1.14 (1.10–1.19)***	1.06 (1.02–1.11)**
Third quartile	4,383/1,261,467	1.03 (0.99–1.08)	1.01 (0.97–1.06)
Fourth quartile (the highest)	3,283/1,270,989	0.75 (0.71–0.78)***	0.94 (0.89–0.98)*
P for trend		<0.001	<0.001
Per SD increase	16,480/5,033,653	0.87 (0.85–0.88)***	0.96 (0.95–0.98)***
GS, 1000 m buffer
First quartile (the lowest)	4,415/1,241,754	1[Reference]	1[Reference]
Second quartile	4,700/1,254,679	1.01 (0.97–1.05)	0.98 (0.94–1.02)
Third quartile	4,103/1,264,717	0.85 (0.81–0.89)***	0.90 (0.86–0.94)***
Fourth quartile (the highest)	3,262/1,272,503	0.66 (0.63–0.69)***	0.87 (0.83–0.92)***
P for trend		<0.001	<0.001
Per SD increase	16,480/5,033,653	0.84 (0.83–0.86)***	0.94 (0.92–0.96)***
BS, 300 m buffer
First quartile (the lowest)	4,573/1,273,087	1[Reference]	1[Reference]
Second quartile	4,019/1,236,130	0.88 (0.84–0.92)***	0.95 (0.91–1.00)*
Third quartile	3,963/1,260,195	0.85 (0.81–0.89)***	0.98 (0.94–1.03)
Fourth quartile (the highest)	3,925/1,264,242	0.85 (0.81–0.88)***	0.97 (0.93–1.01)
P for trend		<0.001	0.584
Per SD increase	16,480/5,033,653	0.98 (0.96–1.00)*	1.00 (0.98–1.01)
BS, 1000 m buffer
First quartile (the lowest)	4,200/1,253,470	1[Reference]	1[Reference]
Second quartile	3,966/1,255,461	0.93 (0.89–0.98)**	1.00 (0.96–1.04)
Third quartile	4,088/1,263,018	0.96 (0.92–1.00)*	1.02 (0.98–1.07)
Fourth quartile (the highest)	4,226/1,261,706	0.99 (0.95–1.03)	1.05 (1.01–1.10)*
P for trend		0.269	0.006
Per SD increase	16,480/5,033,653	0.99 (0.98–1.01)	1.01 (0.99–1.02)
NE, 300 m buffer
First quartile (the lowest)	4,592/1,262,595	1[Reference]	1[Reference]
Second quartile	4,354/1,256,489	0.93 (0.89–0.97)***	0.94 (0.90–0.98)**
Third quartile	4,230/1,253,150	0.89 (0.85–0.92)***	0.95 (0.91–0.99)*
Fourth quartile (the highest)	3,304/1,261,419	0.67 (0.64–0.70)***	0.87 (0.83–0.91)***
P for trend		<0.001	<0.001
Per SD increase	16,480/5,033,653	0.84 (0.83–0.86)***	0.95 (0.93–0.96)***
NE, 1000 m buffer
First quartile (the lowest)	4,706/1,241,155	1[Reference]	1[Reference]
Second quartile	4,523/1,256,914	0.90 (0.86–0.94)***	0.94 (0.90–0.98)**
Third quartile	4,077/1,264,504	0.78 (0.75–0.82)***	0.88 (0.85–0.92)***
Fourth quartile (the highest)	3,174/1,271,081	0.60 (0.57–0.63)***	0.83 (0.78–0.87)***
P for trend		<0.001	<0.001
Per SD increase	16,480/5,033,653	0.81 (0.80–0.83)***	0.91 (0.90–0.93)***

Notes: *P < 0.05, **P < 0.01, ***P < 0.001. Model 1 was adjusted for age and sex, whereas Model 2 was additionally adjusted for body mass index, household income, educational attainment, smoking status, home area, and categories of genetic risk.

Abbreviations: COPD, Chronic obstructive pulmonary disease; HR, Hazard ratio; CI, Confidence interval; SD, Standard deviation; GS, Green space; BS, Blue space; NE, Natural environment.

In contrast, associations with blue space were weaker and inconsistent. The inverse association observed in the minimally adjusted models was substantially attenuated after full adjustment (300 m: HR 0.97 [95% CI, 0.93–1.01]; P for trend = 0.584; 1000 m: HR 1.05 [95% CI, 1.01–1.10]; P for trend = 0.006).

The RCS further demonstrated generally inverse exposure–response associations for green space within the 1000 m buffer and natural environment exposure within both buffers. The association for green space within the 300 m buffer was nonlinear, with a lower COPD risk observed mainly at higher exposure levels, whereas no clear exposure–response pattern was observed for blue space (Figure S2).

Sensitivity analyses incorporating an alternative COPD definition (Table S6), analyses based on natural environment exposure in the full eligible population with available LCM data (Table S7), exclusion of participants diagnosed within the first two years of follow-up (Table S8) and additional covariate adjustment (Table S9) showed similar findings.

Mediation Analyses

Mediation analyses indicated significant indirect effects through PM_2.5 and NO₂ in separate single-mediator models for green space and natural environment exposures (Table S10). For natural environment exposure within the 1000 m buffer, the estimated proportions mediated by PM_2.5 and NO₂ were 80.8% and 94.5%, respectively.

Association of Genetic Predisposition with Incident COPD

The risk of COPD increased with higher genetic susceptibility (Table 3). Compared with participants in the lowest genetic risk group, those in the intermediate and high genetic risk categories had significantly higher risks of COPD (HR 1.15 [95% CI, 1.09–1.22] and HR 1.30 [95% CI, 1.21–1.40], respectively; P for trend < 0.001). When modeled as a continuous variable, each 1–SD increase in the COPD-PRS was associated with a 10% higher risk of COPD (HR 1.10 [95% CI, 1.08–1.13]).

Table 3.

The Risk of COPD According to Genetic Risk Category

	No.of Cases/Person-Years	Model 1, HR (95% CI)	Model 2, HR (95% CI)
Genetic risk
Low	2,965/1,008,731	1[Reference]	1[Reference]
Intermediate	9,843/3,019,902	1.11 (1.07–1.16)***	1.15 (1.09–1.22)***
High	3,672/1,005,021	1.25 (1.19–1.31)***	1.30 (1.21–1.40)***
P for trend		<0.001	<0.001
Per SD increase	16,480/5,033,653	1.08 (1.06–1.10)***	1.10 (1.08–1.13)***

Note: ***P < 0.001. Model 1 was adjusted for age and sex, whereas Model 2 was additionally adjusted for body mass index, household income, educational attainment, smoking status, home area, and categories of green space, blue space, and the natural environment.

Abbreviations: COPD, Chronic obstructive pulmonary disease; HR, Hazard ratio; CI, Confidence interval; SD, Standard deviation.

Interaction and Joint Analysis of GBN and Genetic Predisposition

Compared with participants with high GBN exposure and low genetic susceptibility, those with higher genetic susceptibility and lower GBN exposure generally exhibited higher COPD risk (Figure 1). Specifically, compared with the reference group, individuals with high genetic risk and the first quartile of green space and natural environment exposure within the 1000 m buffer had 52% and 65% higher risks, respectively (green space: HR = 1.52 [95% CI, 1.36–1.69], natural environment: HR = 1.65 [95% CI, 1.48–1.83]).

Figure 1.

The joint associations of GBN and genetic risk categories with the risk of incident COPD. (A and B) Joint effects of green space and genetic risk categories; (C and D) Joint effects of blue space and genetic risk categories; (E and F) Joint effects of natural environment and genetic risk categories. The combination of low genetic risk and the fourth exposure quartile served as the reference. The model was adjusted for age, sex, body mass index, household income, educational attainment, home area, and smoking status.

Abbreviations: GBN, Green space, blue space, and natural environment; COPD, Chronic obstructive pulmonary disease; GS, Green space; BS, Blue space; NE, Natural environment; HR, Hazard ratio; CI, Confidence interval.

Within each stratum of genetic susceptibility, greater exposure to green space and the natural environment was generally associated with a lower COPD risk compared with the first quartile (Figure S3). These protective associations of natural environment exposure (1000 m buffer) were consistent across low, intermediate, and high genetic risk categories, indicating benefits irrespective of genetic susceptibility. However, no significant multiplicative or additive interactions were observed (Table S11). In addition, no consistent associations were observed for blue space.

Discussion

In this large prospective cohort, participants with high genetic risk and lower GBN exposure had a significantly greater risk of incident COPD (Figure 1). Additionally, greater exposure to the natural environment within the 1000 m buffer was associated with lower COPD risk across all genetic risk strata (Figure S3). Genetic risk and exposures to green space and the natural environment were independently associated with incident COPD, and no significant interaction was observed between them, suggesting that genetic susceptibility did not modify these associations (Tables 2, 3 and S11).

A nationwide Chinese study reported that higher residential greenness was associated with higher forced expiratory volume in 1 second (FEV₁) (24.76 mL per IQR increase [95% CI, 13.32–36.20]) and lower COPD prevalence (odds ratio [OR] = 0.90 [95% CI, 0.83–0.97]).⁸ However, a cross-sectional study from China reported that neighborhood greenness was positively associated with COPD prevalence (OR = 1.08 [95% CI, 1.01–1.15]), rather than serving as a protective factor.²⁶ These inconsistent findings make it difficult to clarify the specific impact of the natural environment on COPD. More recently, a prospective study consistent with our results similarly reported that exposure to green space and natural environment was inversely associated with COPD incidence.²⁷ However, that study did not account for genetic factors or potential gene–environment interplay, thereby highlighting the novelty of our study, which provides greater insight on the joint associations of genetic and environmental factors with COPD onset.

To our knowledge, no previous study has investigated the joint associations of exposures to GBN and genetic susceptibility with the risk of COPD. Several studies have examined whether environmental exposures, such as smoking, modify COPD susceptibility by specific genetic variants. A population-based study from the Rotterdam cohort reported a significant interaction between an HHIP polymorphism and cumulative smoking exposure in relation to COPD risk.²⁸ Another study in a Korean population revealed that the FAM13A haplotype was associated with a lower ratio of FEV₁ to forced vital capacity and interacted with heavy smoking to modify pulmonary function decline, whereas a cross-cohort analysis found that neither HHIP nor FAM13A variants were significantly associated with ever smoking.^29,30

Although these studies suggest potential gene–environment interactions, they were limited by modest sample sizes and focused mainly on harmful exposures such as smoking rather than protective environmental factors. Unlike previous work emphasizing risk amplification, the present study shifted the perspective toward resilience. We examined whether favorable environmental exposures—specifically residential GBN—may mitigate COPD risk irrespective of genetic predisposition. Leveraging the large-scale UK Biobank cohort, this study integrated comprehensive and objectively measured indicators of GBN exposures and incorporated PRSs capturing key COPD susceptibility loci. This design allowed a systematic evaluation of whether beneficial environmental exposures can attenuate genetic vulnerability to COPD, offering new insights for precision prevention among individuals at high genetic risk.

Several biological and environmental mechanisms may explain our findings. Variants in HHIP and FAM13A are implicated in pathways regulating airway remodeling, epithelial repair, and oxidative stress responses.^31,32 Conversely, exposure to greener and more natural environments has been linked to lower concentrations of ambient air pollutants,³³ reductions in the oxidative burden,³⁴ enhanced physical activity,¹⁰ and improved mental well-being.³⁵ Together, these complementary mechanisms may promote pulmonary resilience across different levels of genetic susceptibility.

Our study has several strengths. The large prospective UK Biobank cohort with long-term follow-up enabled the assessment of incident COPD. Separate analyses of GBN exposures and COPD-PRS, together with joint and interaction analyses, provided a comprehensive evaluation of their independent associations, joint risk patterns, and potential gene–environment interactions. Mediation analyses explored potential pathways, while extensive sensitivity analyses supported the robustness of the findings.

Limitations

Several limitations should be acknowledged. First, GBN exposures in our study were assessed only within 300 m and 1000 m buffers, which may limit the ability to capture the full range of relevant spatial scales. Second, the absence of information on residential mobility may have led to attenuation of exposure estimates. Third, as the participants were restricted to those of European ancestry, the generalizability of our findings to other populations remains uncertain. Replication in other populations and studies incorporating residential histories are warranted. Finally, residual confounding from unmeasured factors, such as smoking intensity, occupational exposures, socioeconomic deprivation, and settlement quality, cannot be excluded.

Conclusion

The present study revealed that individuals with higher genetic susceptibility and lower exposures to green space and the natural environment had an increased risk of incident COPD. Moreover, the protective associations of the natural environment with COPD incidence persisted across all levels of genetic risk. No significant gene–environment interaction was detected. Collectively, our findings support further evaluation of residential natural environments as potentially modifiable factors in COPD prevention.

Funding Statement

This work was supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (grant number 2024ZD0528500; Lifeng Yan), the National Natural Science Foundation of China (grant numbers 82090015, 82400035 and 82570055; Weining Xiong, Huaqi Guo and Lifeng Yan, respectively), and the Henan Provincial Medical Science and Technology Program (grant number SBGJ202502089; Huaqi Guo).

Abbreviations

COPD, Chronic obstructive pulmonary disease; GBN, Green and blue space and the natural environment; HR, Hazard ratio; CI, Confidence interval; HHIP, Hedgehog interacting protein; FAM13A, Family with sequence similarity 13 member A; PRS, Polygenic risk score; GLUD, Generalized land use database; CEH, Centre for ecology and hydrology; LCM, Land cover map; GME, Geospatial modelling environment; ICD-10, International classification of diseases, tenth revision; SNP, Single nucleotide polymorphism; GWAS, Genome-wide association study; BMI, Body mass index; SD, Standard deviation; IQR, Interquartile range; PM_2.5, Particulate matter with an aerodynamic diameter ≤2.5 μm; NO₂, Nitrogen dioxide; LUR, Land use regression; ESCAPE, European study of cohorts for air pollution effects; ACME, Average causal mediation effect; ADE, Average direct effect; RERI, Relative excess risk due to interaction; AP, Attributable proportion due to interaction; SI, Synergy index; RCS, Restricted cubic spline; FEV1, Forced expiratory volume in 1 second; OR, Odds ratio.

Data Sharing Statement

The dataset used and analyzed during the current study is available from UK Biobank (www.ukbiobank.ac.uk).

Ethics Approval and Consent to Participate

This study is based on an approved UK Biobank application (Application Number: 91536). The UK Biobank study was approved by the North West Multicenter Research Ethics Committee (MREC reference: 21/NW/0157). All participants provided written informed consent. This study using UK Biobank data was approved by the Ethics Committee of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine (Approval No. SH9H-2026-T498-1).

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.