Skip to main content
NCBI home page
Springer logoLink to Springer
. 2026 Oct 5;145(1):83. doi: 10.1007/s00439-026-02875-w

Annual average exposure to ambient air pollution, polygenic risk, and incidence of major chronic diseases: a large-scale prospective cohort study

Jing Tian 1, Sicong Wang 1, Xinyi Chang 1, Yi Han 1,✉
PMCID: PMC13638686  PMID: 42831997

Abstract

Ambient air pollution is a major environmental risk factor for chronic diseases, yet its independent effects, as well as its joint associations and interactions with genetic susceptibility, on multiple chronic conditions remain understudied. This study aims to investigate the association between annual average exposure to multiple air pollutants and five major chronic diseases, while quantifying the role of polygenic risk in these relationships. We included 313,534 participants from the UK Biobank. Land use regression models were used to estimate annual average air pollution concentrations at residential addresses. Chronic diseases (cardiovascular, respiratory, neurodegenerative, hypertension, and diabetes) were identified using ICD-9 and ICD-10 codes. Weighted polygenic risk scores (PRS) were constructed for each disease. Cox proportional hazards models were used to assess associations, and additive interactions were evaluated using RERI and AP. Annual average exposure to PM2.5, PM10, NO2, and NOx was significantly associated with increased chronic disease risk. PM2.5 showed the strongest association (HR = 1.22 per 5 µg/m³, 95% CI: 1.19–1.26). High genetic risk was independently associated with disease incidence (e.g., diabetes: HR = 1.77 per SD increase in PRS). Significant additive interactions were observed between high genetic risk and high PM2.5 exposure for neurodegenerative disease (RERI = 0.14) and diabetes (RERI = 0.21). Participants with both high genetic risk and high pollution exposure had the highest disease burden. Annual average exposure to air pollutants is associated with increased risks of multiple chronic diseases, with effect modification by genetic susceptibility.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00439-026-02875-w.

Introduction

Chronic diseases, including cardiovascular diseases, diabetes, chronic respiratory diseases, and neurodegenerative disorders, represent a leading cause of global mortality and disability, posing a substantial public health and economic burden worldwide (Kimokoti and Millen 2016, Collaborators 2018, Hacker and Kaufmann 2024). Concurrently, ambient air pollution has been widely recognized as a major modifiable environmental risk factor, with its long-term health hazards being firmly established (Collaborators 2020, Rajagopalan et al. 2018). Upon inhalation, pollutants—particularly fine particulate matter (PM2.5) and nitrogen dioxide (NO2)—can penetrate deep into the alveoli, enter the systemic circulation, and induce systemic inflammation and oxidative stress, thereby driving the development and progression of a wide spectrum of chronic diseases (Glencross et al. 2020).

A growing body of epidemiological evidence has linked air pollution exposure to diverse health outcomes. For instance, recent studies have demonstrated that annual average exposure to air pollutants is positively associated with an increased risk of psoriasis incidence, the progression of chronic lung disease from onset to mortality, and the incidence of abdominal aortic aneurysm (Xiong et al. 2025, Ma et al. 2024, Wang et al. 2023). These findings not only broaden our understanding of the scope of health hazards posed by air pollution but also underscore the significant inter-individual variability in response to environmental insults. Growing evidence suggests that genetic background plays a pivotal role in modulating an individual’s sensitivity to environmental stressors (Miller and Newby 2019, Xiong and Xia 2025, Hill and Lim 2023). Specifically, polymorphisms in genes related to oxidative stress and inflammatory pathways may determine the magnitude of adverse health effects following air pollution exposure. Therefore, it is plausible that individuals with high genetic susceptibility may be particularly vulnerable to these environmental hazards.

Most current studies focus solely on a single disease or a limited number of conditions, lacking comprehensive assessments of the individual effects of several major air pollutants on a range of key chronic diseases within a single, unified large-scale prospective cohort. Furthermore, although individual genetic backgrounds may significantly alter health responses to air pollution, the compound effects and interactions between polygenic risk and long-term air pollution exposure in the development of various chronic diseases have yet to be systematically explored and quantitatively evaluated.

Therefore, this study leverages UK Biobank data to comprehensively assess the association between annual average exposure to four key air pollutants and the incidence of five major chronic diseases (cardiovascular disease, respiratory disease, neurodegenerative disease, hypertension, and diabetes). It further elucidates the independent and synergistic effects of polygenic risk scores and air pollution on disease risk, and finally quantifies the additive interaction between genetic susceptibility and environmental exposure.

Methods

Study design and participants

This study is based on the UK Biobank, a large-scale, population-based prospective cohort. This cohort study recruited over 500,000 participants through 22 assessment centers established in England, Scotland, and Wales between 2006 and 2010 (Sudlow et al. 2015, Halperin and Tirosh 2024). At baseline (the time of recruitment), all participants completed a detailed electronic questionnaire and underwent a series of physical measurements. The study protocol was approved by the National Health Service (NHS) North West Multi-Centre Research Ethics Committee. All participants provided written informed consent prior to enrollment. In this study, baseline was defined as the time of the participant’s first visit to the assessment center. To ensure a rigorous prospective design for assessing incident outcomes, we explicitly excluded participants who already had a pre-existing diagnosis of the respective chronic disease at baseline. Figure 1 illustrates the overall study design, including the detailed participant inclusion and exclusion processes.

Fig. 1.

Fig. 1

Open in a new tab

Inclusion and exclusion procedures for all chronic disease populations in this study

Air pollution measurement

The spatial variability of ambient air pollutants was estimated using Land Use Regression (LUR) models developed by the European Study of Cohorts for Air Pollution Effects (ESCAPE) project. These models were built upon standardized monitoring campaigns conducted between January 26, 2010, and January 18, 2011. To capture seasonal variations, the monitoring protocol involved three separate two-week measurement periods in different seasons within this timeframe at multiple monitoring sites across the study areas. To estimate individual exposure, the LUR models incorporated fine-scale Geographic Information System (GIS)-derived predictors, including traffic intensity on the nearest road, total traffic load within circular buffers, population density, and specific land-use categories. The detailed methodology regarding the development, performance, and rigorous cross-validation of these specific models has been extensively described in previous peer-reviewed publications (Eeftens et al. 2012, Beelen et al. 2013). By integrating these high-resolution spatial predictors, the models were capable of predicting annual average pollutant concentrations at the specific residential address level of each UK Biobank participant. This address-level approach effectively captures intra-urban spatial contrasts in air pollution that larger-scale resolution models might overlook. In this study, we assigned the 2010 annual average pollutant concentrations to the participants’ baseline residential addresses as a proxy for their environmental exposure. The models demonstrated robust performance upon cross-validation, with explained variances (R2) as follows: 88% for PM10, 87% for NO2, 88% for NOx, and 77% for PM 2.5.

Ascertainment of outcomes

Incident cases of chronic diseases were identified through the UK Biobank’s comprehensive linkage to electronic health records. Specifically, hospital inpatient admissions data were obtained from NHS Digital (for England), Public Health Scotland, and the SAIL Databank (for Wales). Mortality data were derived from national death registries provided by the NHS Information Centre (for England and Wales) and the NHS Central Register (for Scotland). Diagnoses were determined based on the first occurrence of the relevant International Classification of Diseases, 9th Edition (ICD-9) and 10th Edition (ICD-10) codes within these linked hospital and death datasets. Specifically, the broad disease categories evaluated in this study included the following specific diagnoses: cardiovascular disease (encompassing ischemic heart disease, heart failure, and stroke), respiratory disease (encompassing chronic obstructive pulmonary disease and asthma), neurodegenerative disease (encompassing Alzheimer’s disease, Parkinson’s disease, and multiple sclerosis), as well as hypertension and diabetes. The comprehensive list of exact ICD-9 and ICD-10 codes used for each specific condition is provided in Table S1. Participants were followed from baseline (2006–2010) until the occurrence of a chronic disease event or the time of review. Review was defined as the time of loss to follow-up, death, or completion of follow-up (up to December 31, 2022), whichever occurred first. While air pollution land use regression models were developed for the entire UK Biobank cohort, participants were excluded from the final analytical sample if they had missing exposure data or missing baseline covariates required for the adjusted models.

Covariates

Based on prior research, this analysis comprehensively adjusted for a range of potential confounding factors. Covariate data were primarily collected at baseline through standardized touch-screen questionnaires and physical measurements at the UK Biobank assessment centers. Demographic and socioeconomic variables included age (calculated continuously from date of birth to the baseline assessment date), sex (male or female), and ethnicity (categorized as White or others). Socioeconomic status was controlled for through educational attainment (categorized as high education level, other levels, or unknown) and average total before-tax household income (stratified into <£18,000, £18,000–£30,999, £31,000–£51,999, £52,000–£100,000, and >£100,000). Lifestyle confounders were defined as follows: Body mass index (BMI) was calculated from objectively measured weight and height and categorized according to the World Health Organization (WHO) guidelines (< 18.5, 18.5–24.9, 25.0–29.9, and ≥ 30.0 kg/m²). Smoking status and alcohol drinking status were self-reported and classified into ‘never’, ‘previous’, and ‘current’. Physical activity levels were self-reported via the baseline questionnaire and dichotomized simply as ‘yes’ (indicating engagement in regular physical activity) or ‘no’ (indicating a lack of regular physical activity), without further quantification of exercise intensity or duration. Furthermore, a healthy diet score (ranging from 0 to 7) was constructed based on seven dietary components (vegetables, fruit, fish, whole grains, refined grains, processed meat, and unprocessed meat), with detailed scoring components and intake goals provided in Table S2. Various dietary scoring systems have been applied to UK Biobank data to evaluate health outcomes (Navratilova et al. 2024). In this study, we constructed a 7-item healthy diet score based on the American Heart Association (AHA) recommendations (Lloyd-Jones et al. 2010).

Genotyping and weighted polygenic risk score

To construct polygenic risk scores (PRS) for chronic diseases, we utilized genome-wide association study (GWAS) data. For cardiovascular disease, hypertension, and diabetes, the weighted PRS were derived from previously reported and rigorously evaluated summary statistics within the UK Biobank database (Thompson et al. 2024). These scores have been validated for their predictive performance in large-scale UK Biobank analyses. For respiratory and neurodegenerative diseases, GWAS summary statistics were obtained from external sources as detailed in Table S3.

After applying linkage disequilibrium (LD) pruning and filtering for genome-wide significance (Inline graphic), we identified 32 independent SNPs for respiratory diseases and 75 for neurodegenerative diseases. The weighted PRS for each participant was calculated as the sum of allele counts multiplied by their respective effect sizes (beta coefficients) from the GWAS summary statistics.

We calculated a weighted PRS according to the following formula PRSiInline graphic, where PRSi is the weighted polygenic risk score of the ith individual, Xik is the dosage (i.e., the number of effect alleles, recoded as 0, 1, or 2) of the kth SNP for the ith individual, βk is the estimated effect of the kth SNP, and K is the total number of instrumental variables (Zhang et al. 2022). The PRS was then classified into low (the lowest tertile), intermediate (second tertile), and high (the highest tertile) genetic risk categories.

Statistical analysis

All statistical analyses were performed using R software (version 4.5.0). Baseline characteristics were presented as medians (interquartile ranges) for continuous variables and numbers (percentages) for categorical variables, with group comparisons tested by the Mann-Whitney U test or Chi-squared test, as appropriate. To investigate potential non-linear exposure-response relationships between ambient air pollutants and the incidence of chronic diseases, we incorporated restricted cubic splines (RCS) with three knots—placed at the 10th, 50th, and 90th percentiles of the pollutant concentrations—into the fully adjusted Cox proportional hazards models. The formal test for non-linearity was conducted using a likelihood ratio test. This test evaluated whether the model containing both the linear and the cubic spline terms provided a significantly better fit than a simpler model restricted to only the linear term. A P-value for non-linearity (P non−linear) < 0.05 indicated a statistically significant non-linear relationship. The associations between annual average exposure to air pollutants (PM2.5, PM10, NO2, NOx) and the incidence of chronic diseases were assessed using Cox proportional hazards models, calculating hazard ratios (HRs) and 95% confidence intervals (CIs). For continuous models, air pollutant concentrations were scaled per 5 µg/m³ increment (and 10 µg/m³ in sensitivity analyses) rather than per 1 µg/m³. This scaling was purposefully chosen to provide more clinically interpretable effect sizes and to facilitate direct comparability with previous large-scale epidemiological cohorts and meta-analyses. To maximize the use of available information, the Cox proportional hazards models for each specific pollutant utilized the maximum available sample size with non-missing data for that particular exposure and required covariates. Consequently, the sample sizes varied slightly across pollutants (385,605 for PMs and 415,178 for NOs). In contrast, Table 1 describes the ‘intersection cohort’ (n = 313,534) with complete data across all four pollutants and covariates to ensure a consistent baseline comparison. We constructed two models: Model a adjusted for age and sex, and Model b further adjusted for ethnicity, BMI, education, household income, smoking, alcohol drinking, physical activity, UK Biobank assessment centre and healthy diet score. PRS were constructed using published SNPs and analyzed both continuously (per standard deviation) and categorically (low, intermediate, high risk). The association of PRS with disease risk was evaluated using Cox models with sequential adjustment for demographic, lifestyle, and finally, air pollution factors. Furthermore, to identify potentially vulnerable subpopulations, we conducted stratified analyses by age, sex, and BMI to explore potential effect modifications. Regarding the joint associations between genetic susceptibility and environmental exposure, we purposefully evaluated interactions on the additive scale rather than the multiplicative scale. For the joint and interaction analyses, ambient air pollution exposures were dichotomized into “low”and “high” categories based on the 2005 World Health Organization (WHO) Air Quality Guidelines (which were the prevailing standards during the cohort’s baseline period): 10 µg/m³ for PM2.5, 20 µg/m³ for PM10, and 40 µg/m³ for NO2 and NOx. In epidemiological and public health contexts, additive interaction is considered more biologically meaningful and relevant for evaluating the public health impact of exposures. It directly quantifies the absolute excess risk caused by the co-exposure, making it highly valuable for identifying susceptible subgroups that would benefit the most from targeted environmental or clinical interventions. Additive interaction was mathematically quantified by calculating the relative excess risk due to interaction (RERI) and the attributable proportion (AP), where a RERI or AP significantly greater than 0 indicates a positive additive interaction.

Table 1.

Baseline Characteristics of 313,534 Participants in the UK Biobank Study

Characteristics Control group, N = 187,894 Chronic diseases, N = 125,640 p-value
Age, years 54.00 (47.00, 61.00) 60.00 (54.00, 65.00) < 0.001
Sex, N (%) < 0.001
Male 76,683.00 (40.81%) 62,067.00 (49.40%)
Ethnicity, N (%) < 0.001
White 177,816.00 (94.64%) 118,663.00 (94.45%)
BMI, N (%) < 0.001
BMI < 18.5 1,131.00 (0.60%) 537.00 (0.43%)
18.5 ≤ BMI < 25 76,335.00 (40.63%) 33,071.00 (26.32%)
25 ≤ BMI < 30 79,086.00 (42.09%) 55,375.00 (44.07%)
BMI ≥ 30 31,342.00 (16.68%) 36,657.00 (29.18%)
Education level, N (%) < 0.001
High education level 70,589.00 (37.57%) 35,498.00 (28.25%)
Others levels 94,948.00 (50.53%) 62,641.00 (49.86%)
Unknown 22,357.00 (11.90%) 27,501.00 (21.89%)
Smoking status, N (%) < 0.001
Never smoker 112,997.00 (60.14%) 63,587.00 (50.61%)
Previous smoker 58,081.00 (30.91%) 47,457.00 (37.77%)
Current smoker 16,816.00 (8.95%) 14,596.00 (11.62%)
Physical activity, N (%) < 0.001
No 67,913.00 (36.14%) 48,953.00 (38.96%)
Yes 119,981.00 (63.86%) 76,687.00 (61.04%)
Alcohol drinking status, N (%) < 0.001
Never drinking 6,809.00 (3.62%) 5,557.00 (4.42%)
Previous drinking 4,948.00 (2.63%) 4,588.00 (3.65%)
Current drinking 176,137.00 (93.74%) 115,495.00 (91.93%)
Healthy diet score (mean (SD)) 3.00 ± 1.23 3.00 ± 1.30 < 0.001
Household income, N (%) 0.622
<£18,000 42,789.00 (22.77%) 28,858.00 (22.97%)
£18,000−£30,999 47,910.00 (25.50%) 31,982.00 (25.46%)
£31,000−£51,999 49,109.00 (26.14%) 32,596.00 (25.94%)
£52,000−£100,000 38,055.00 (20.25%) 25,445.00 (20.25%)
>£100,000 10,031.00 (5.34%) 6,759.00 (5.38%)
PM2.5, µg/m3 9.90 (9.24, 10.53) 9.94 (9.30, 10.57) < 0.001
PM10, µg/m3 16.01 (15.20, 16.99) 16.03 (15.26, 17.01) < 0.001
NO2, µg/m³ 25.90 (21.06, 31.11) 26.10 (21.43, 31.07) < 0.001
NOx, µg/m³ 41.73 (33.63, 50.36) 42.25 (34.32, 50.64) < 0.001
Open in a new tab

Data are presented as median (interquartile range, IQR) for continuous variables, except for the healthy diet score, which is presented as mean ± standard deviation (SD). Categorical variables are presented as numbers (percentages). P-values were calculated using the Mann-Whitney U test for continuous variables (with the exception of the independent Student’s t-test used for the healthy diet score) and the Chi-squared test for categorical variables to compare differences between participants who developed any of the five specified chronic diseases during follow-up and those who did not

BMI body mass index, PM2.5 particulate matter with an aerodynamic diameter ≤ 2.5 μm, PM10 particulate matter with an aerodynamic diameter ≤ 10 μm, NO2 nitrogen dioxide, NOx nitrogen oxides

Results

Baseline characteristics of participants

As shown in the study flowchart in Fig. 1; Table 1 summarizes the baseline characteristics of 313,534 participants who did not have any of the five chronic diseases under study at baseline and for whom covariate data were complete. This cohort was further divided into two groups: 187,894 participants who remained free of all five diseases throughout the follow-up period (the control group) and 125,640 participants who developed at least one of these diseases during the follow-up period (the chronic disease group). Significant differences existed between the two groups for most variables (p < 0.001), except household income (p = 0.622). Demographically, the chronic disease group was older [median (IQR): 60.00 (54.00, 65.00) vs. 54.00 (47.00, 61.00) years] and had a higher male proportion (49.40% vs. 40.81%). For lifestyle and anthropometrics, the chronic disease group showed higher obesity (BMI ≥ 30: 29.18% vs. 16.68%), smoking (previous: 37.77% vs. 30.91%; current: 11.62% vs. 8.95%), and lower regular physical activity (61.04% vs. 63.86%) compared to the control group.

Notably, the chronic disease group had marginally higher median concentrations of all ambient air pollutants (all p < 0.001): PM2.5 [9.94 (9.30, 10.57) vs. 9.90 (9.24, 10.53) µg/m³], PM10 [16.03 (15.26, 17.01) vs. 16.01 (15.20, 16.99) µg/m³], NO2 [26.10 (21.43, 31.07) vs. 25.90 (21.06, 31.11) µg/m³], and NOx[42.25 (34.32, 50.64) vs. 41.73 (33.63, 50.36) µg/m³]. In stratified analyses, similar positive trends were observed for specific chronic diseases, including cardiovascular disease, respiratory diseases, neurodegenerative diseases, hypertension, and diabetes (Table S4–S8).

Ambient air pollution and chronic disease risk

As shown in Table 2, annual average exposure to PM2.5, PM10, NO2, and NOx was significantly associated with increased chronic disease risk (P for trend < 0.001). PM2.5 exhibited the strongest association, with each 5 µg/m³ increment corresponding to a 22% higher risk (adjusted HR = 1.22, 95% CI: 1.19–1.26). Compared with the lowest quartile (Q1), individuals in the highest quartile (Q4) had a 12% increased risk (adjusted HR = 1.14, 95% CI: 1.10–1.22). For PM10, each 5 µg/m³ increment was associated with an 13% increased risk (adjusted HR = 1.13, 95% CI: 1.04–1.18); comparing Q4 to Q1, the risk was elevated by 8% (adjusted HR = 1.08, 95% CI: 1.02–1.14). For both NO2 and NOx, each 5 µg/m³ increment was associated with a 5% increased risk (adjusted HR = 1.05, 95% CI: 1.03–1.11). Compared to their respective Q1, the Q4 of NO2 and Nox were associated with 10% (adjusted HR = 1.10) and 10% (adjusted HR = 1.10) higher risks, respectively. Sensitivity analyses using alternative exposure metrics (per 10 µg/m³ and per IQR increment) yielded consistent results, with PM2.5 remaining the most strongly associated pollutant (Table S9). Notably, PM2.5 consistently demonstrated the strongest associations across all disease subtypes (Table S10-S14). For instance, each 5 µg/m³ increase in PM2.5 was associated with a 24% higher risk of cardiovascular disease (adjusted HR = 1.24, 95% CI: 1.13–1.32), a 32% increased risk of respiratory disease (adjusted HR = 1.32, 95% CI: 1.26–1.38), a 43% elevated risk of neurodegenerative disease (adjusted HR = 1.43, 95% CI: 1.32–1.49), a 25% higher risk of hypertension (adjusted HR = 1.25, 95% CI: 1.21–1.28), and a 43% increased risk of diabetes (adjusted HR = 1.43, 95% CI: 1.37–1.54).

Table 2.

The association between annual average exposure to ambient air pollution and Chronic diseases risk

Pollution Increase Events Total/ person-years HR (95% CI) P for trend
Model a Model b
PM2.5
per 5 µg/m3 increment 156,195 385,605/4,143,358 1.43 (1.42–1.47) 1.22 (1.19–1.26)
Q1(< 9.27) 37,314 96,645/1,053,521 ref ref < 0.001
Q2 (9.27–9.92) 39,387 97,005/ 1,039,343 1.10 (1.08–1.11) 1.06 (1.02–1.10)
Q3 (9.92–10.55) 39,342 95,799/ 1,023,561 1.15 (1.13–1.16) 1.07 (1.06–1.09)
Q4 (> 10.55) 40,152 96,156/ 1,026,933 1.22 (1.21–1.24) 1.14 (1.10–1.22)
PM10
per 5 µg/m3 increment 156,195 385,605/4,143,358 1.23 (1.2–1.27) 1.13 (1.04–1.18)
Q1 (< 15.23) 38,068 96,426/ 1,047,306 ref ref < 0.001
Q2 (15.23–16.02) 39,604 96,909/ 1,042,663 1.07 (1.05–1.08) 1.06 (1.02–1.12)
Q3 (16.02–17.01) 39,289 96,348/ 1,028,284 1.10 (1.08–1.11) 1.06 (1.03–1.08)
Q4 (> 17.01) 39,234 95,922/ 1,025,105 1.11 (1.09–1.12) 1.08 (1.02–1.14)
NO2
per 5 µg/m3 increment 163,929 415,178/4,521,505 1.08 (1.08–1.09) 1.05 (1.03–1.11)
Q1 (< 21.37) 39,571 103,943/ 1,145,872 ref ref < 0.001
Q2 (21.37–26.16) 41,709 103,803/ 1,127,439 1.09 (1.08–1.11) 1.03 (1.02–1.12)
Q3 (26.16–31.20) 41,849 103,775/ 1,125,841 1.13 (1.11–1.14) 1.06 (1.02–1.14)
Q4 (> 31.20) 40,800 103,657/ 1,122,353 1.17 (1.16–1.19) 1.10 (1.02–1.09)
NOx
per 5 µg/m3 increment 163,928 415,178/4,521,505 1.08 (1.08–1.09) 1.03 (1.01–1.08)
Q1 (< 34.19) 39,517 103,847/ 1,144,219 ref ref < 0.001
Q2 (34.19–42.25) 41,138 103,756/ 1,128,726 1.08 (1.07–1.10) 1.04 (1.03–1.09)
Q3 (42.25–50.63) 41,548 103,791/ 1,124,317 1.13 (1.12–1.15) 1.04 (1.02–1.12)
Q4 (> 50.63) 41,725 103,784/ 1,124,242 1.20 (1.18–1.21) 1.10 (1.05–1.17)
Open in a new tab

Air pollutant concentrations were modeled both continuously (per 5 µg/m³ increment) and categorically (in quartiles, Q1-Q4) using Cox proportional hazards models. Cut-off values for the quartiles are as follows: PM2.5 (< 9.27, 9.27–9.92, 9.92–10.55, and > 10.55 µg/m³); PM10 (< 15.23, 15.23–16.02, 16.02–17.01, and > 17.01 µg/m³); NO2 (< 21.37, 21.37–26.16, 26.16–31.20, and > 31.20 µg/m³); NOx (< 34.19, 34.19–42.25, 42.25–50.63, and > 50.63 µg/m³)

Model a is adjusted for age (continuous) and sex

Model b is the fully adjusted model, accounting for age, sex, ethnicity, BMI category, educational attainment, household income, smoking status, alcohol drinking status, healthy diet score (continuous), physical activity, and UK Biobank assessment centre

P for trend was calculated by assigning the median value of each pollutant quartile to the respective categories and modeling this variable continuously

HR hazard ratio, CI confidence interval, Ref reference category, PM2.5 particulate matter with an aerodynamic diameter ≤ 2.5 μm, PM10 particulate matter with an aerodynamic diameter ≤ 10 μm, NO2 nitrogen dioxide, NOx nitrogen oxides

Exposure-response relationship results

Figure 2 depicts the exposure–response relationships between annual average exposure to four major air pollutants—PM2.5 (A), PM10 (B), NO2 (C), and NOx (D)—and the risk of chronic diseases. All HRs were adjusted for age, sex, ethnicity, BMI, education, household income, alcohol drinking status, smoking status, healthy diet score, physical activity and UK Biobank assessment centre.

Fig. 2.

Fig. 2

Open in a new tab

Exposure−response relationship between individual air pollutants and the risk of chronic disease. A PM2.5 and chronic disease; B PM10 and chronic disease; C NO2 and chronic disease; and D NOx and chronic disease. HRs adjusted for age, sex, ethnicity, BMI, education, household income, alcohol drinking status, smoking status, healthy diet score, physical activity and UK Biobank assessment centre. Abbreviations: PM2.5, particulate matter with aerodynamic diameters of ≤ 2.5 μm; PM10, particulate matter with aerodynamic diameters of ≤ 10 μm; NO2, nitrogen dioxide; NOx, nitrogen oxide. To prevent unstable estimations at the extremes of the data distributions, the exposure-response curves were truncated at the 1st and 99th percentiles of the observed concentrations for each air pollutant

A significant positive association was observed for each pollutant with chronic disease risk (p-overall < 0.001), and these relationships exhibited clear nonlinear characteristics (p-nonlinear < 0.001). In all cases, increasing pollutant concentrations were accompanied by a progressive rise in HRs above the reference level. The relationship for PM2.5 demonstrated the steepest increase in risk per unit concentration, indicating its strongest effect among the pollutants assessed, which aligns with the findings presented in Table 2. Similar, though less pronounced, increasing trends were evident for PM10, NO2, and NOx.

To assess potential variations in the associations, we performed stratified analyses by age, sex, and BMI (Figure S1). A significant modifying effect of sex was observed (all P for heterogeneity < 0.001), with consistently higher risk estimates for all four pollutants in women (HR range: 1.05–1.20) compared to men (HR range: 1.02–1.13). However, no significant effect modifications were found for age (< 60 vs. ≥60 years) or BMI status (overweight vs. obesity) (all P for heterogeneity > 0.05).

Joint associations and interactions between genetic risk and ambient air pollution on chronic disease risk

Tables S15–S19 evaluate the independent association of PRS with five chronic diseases. After adjusting for age, sex, ethnicity, BMI, education level, household income, smoking, alcohol consumption, UK Biobank assessment centre, physical activity, and healthy diet score (Model b), higher genetic risk was significantly associated with the risk of chronic diseases across all studies. For each standard deviation increase in PRS, the risk of chronic disease increased, specifically for cardiovascular disease (HR = 1.35, 95% CI: 1.28–1.40), respiratory disease (HR = 1.24, 1.18–1.30), neurodegenerative disease (HR = 1.38, 95% CI:1.34–1.40), hypertension (HR = 1.36, 95% CI:1.32–1.39), and diabetes (HR = 1.77, 95% CI:1.70–1.84). In Model c, after further adjusting for environmental air pollutants, individuals in the high genetic risk group still exhibited a significantly increased risk of chronic disease compared to those in the low genetic risk group.

We evaluated the synergistic associations of genetic risk and annual average exposure to individual ambient air pollutants with the risk of five chronic diseases (Fig. 3). The results demonstrated additive effects, where high pollution exposure consistently increased disease risk across all genetic risk strata (low, intermediate, and high). Specifically, for cardiovascular disease (Fig. 3A), high PM2.5 exposure incrementally increased the risk across genetic risk groups (HRs: 1.12, 1.47, and 1.96). A pronounced additive effect was observed for neurodegenerative disease (Fig. 3C), where high-risk individuals with high PM2.5 exposure had a significantly higher risk (HR = 2.01) than those with high genetic risk but low exposure (HR = 1.71). The strongest joint associations were observed for diabetes (Fig. 3E), with the highest risk for the combination of high genetic risk and high PM2.5 (HR = 3.84) or high NOx (HR = 3.94). Specific details are provided in Table S20-S24.

Fig. 3.

Fig. 3

Fig. 3

Fig. 3

Open in a new tab

Risk of different chronic diseases according to annual average exposure to ambient air pollutants based on different genetic risk categories. A Cardiovascular disease, B Respiratory disease, C Neurodegenerative disease, D Hypertension, E Diabetes. HRs adjusted for age, sex, ethnicity, BMI, education, household income, alcohol drinking status, smoking status, healthy diet score, physical activity and UK Biobank assessment centre

To further quantify the interaction between genetic susceptibility and environmental air pollution, we calculated the RERI and AP (Tables S25-S29). Analysis revealed significant positive additive interactions between genetic risk and air pollution (particularly PM2.5) for neurodegenerative disease and diabetes risks. For neurodegenerative disease, the interaction between high genetic risk and high PM2.5 exposure was most pronounced (RERI = 0.14, 95% CI: 0.03–0.25; AP = 1.89, 95% CI: 1.76–2.04). Diabetes also showed a significant interaction (high genetic risk & high PM2.5: RERI = 0.21, 95% CI: 0.11–0.30; AP = 2.90, 95% CI: 2.75–3.03). In contrast, no significant additive interactions were observed for cardiovascular and respiratory diseases.

Discussion

In this large-scale prospective study, our findings indicate that annual average exposure to environmental air pollutants is significantly associated with an increased risk of developing multiple chronic diseases. Each 5 µg/m³ increment in PM2.5 was associated with a 22% increased risk of chronic disease incidence in the fully adjusted model (HR = 1.22, 95% CI: 1.19–1.26), with stronger associations observed for diabetes (HR = 1.43, 95% CI: 1.37–1.54 per 5 µg/m³) and neurodegenerative disease (HR = 1.43, 95% CI: 1.32–1.49 per 5 µg/m³). Notably, a clear positive exposure-response relationship was evident, as individuals in the highest quartile of PM2.5 exposure consistently faced significantly greater risks compared to those in the lowest quartile across all disease outcomes. Furthermore, our study yielded two pivotal findings regarding individual susceptibility: first, a high polygenic risk was consistently associated with a substantially elevated disease risk, with the highest genetic risk group exhibiting markedly increased hazards (e.g., for CVD, HR = 1.83, 95% CI: 1.78–1.88); and second, we demonstrated a pronounced synergistic effect, whereby participants with concomitant high genetic risk and high air pollution exposure faced the greatest disease burden (e.g., for CVD, HR = 1.96, 95% CI: 1.90–2.02 for high PM2.5 + high genetic risk vs. low reference group).

Our findings align with and expand upon existing epidemiological literature demonstrating the deleterious health effects of ambient air pollution. Consistent with previous large-scale cohorts, including earlier analyses within the UK Biobank and the ESCAPE project, we observed significant positive associations between particulate matter and nitrogen oxides with cardiometabolic and respiratory outcomes (Rajagopalan et al. 2018, Carlsen et al. 2022). Notably, our study emphasizes the particularly strong effect sizes of PM2.5 on the incidence of diabetes and neurodegenerative diseases. This corroborates recent longitudinal studies suggesting that the central nervous system and metabolic pathways might be highly sensitive targets of fine particulate toxicity (Li et al. 2025, Fiffer et al. 2025). Furthermore, while previous research has largely established the independent main effects of air pollution and genetic risk, our findings regarding their synergistic interactions add a critical dimension to the literature. Similar to recent gene-environment interaction studies focusing on single outcomes (such as incident coronary artery disease or specific respiratory traits), our comprehensive multi-disease approach highlights that genetic susceptibility systematically amplifies the vulnerability to environmental insults across a broad spectrum of chronic conditions (Zhu et al. 2021, Cao et al. 2025).

This study observed a positive correlation between annual average exposure to multiple air pollutants such as PM2.5 and NO2 and increased risks of chronic diseases. This finding is biologically plausible and consistent with a substantial body of existing evidence (Carlsen et al. 2022, Cheng et al. 2024, Kshirsagar et al. 2022). The underlying mechanism may involve systemic inflammation and oxidative stress triggered by inhaled pollutants. The strong associations observed in our study between PM2.5 and the risks of diabetes and neurodegenerative diseases can be explained by a systemic cascade of events initiating in the respiratory system. First, inhaled PM2.5 primarily deposits in the lungs. Although there is a perspective that ultrafine particles may translocate into the systemic circulation, substantial evidence indicates that the majority of particles with diameters ≥ 100 nm are retained within the lungs and exert local toxicity (Miller et al. 2017, Palko et al. 2010, Wiebert et al. 2006). These particles, via organic chemicals and metals adhered to their surfaces, induce significant generation of reactive oxygen species (ROS) within the lungs, thereby initiating oxidative stress and local inflammatory responses (Haberzettl et al. 2016). This process is supported by temporal hierarchical pathway analyses in humans, showing that pulmonary inflammation and oxidative stress pathways are activated first following PM2.5 exposure, followed subsequently by systemic responses (Haberzettl et al. 2018). Second, the local inflammation and oxidative stress originating in the lungs can evolve into a systemic state. Studies have shown that PM2.5 exposure is significantly associated with elevated levels of lipid peroxidation products (e.g., malondialdehyde, F2-isoprostanes) and DNA oxidation damage markers (e.g., 8-oxo-deoxyguanosine) in human urine or blood (Kim et al. 2009, Sørensen et al. 2003, Riggs et al. 2020, Mu et al. 2021). This systemic oxidative stress constitutes a common soil for subsequent multi-organ damage. For the cardiovascular system, the consequences include vascular endothelial dysfunction, increased arterial stiffness, and impaired vascular repair capacity (Krishnan et al. 2012, Riggs et al. 2021, Pope et al. 2016). The significant association between PM2.5 and CVD risk observed in our study likely represents the clinical endpoint of the long-term accumulation of this series of vascular injuries. Particularly noteworthy is that systemic inflammation and oxidative stress are also central links connecting air pollution with metabolic and neurological disorders. Regarding metabolism, animal experiments have demonstrated that PM2.5-induced pulmonary oxidative stress can trigger both vascular and systemic insulin resistance (Sun et al. 2009, Liu et al. 2017). The strongest effect size in our study was observed for diabetes (HR = 1.45), providing strong population-based evidence for the hypothesis that “pulmonary oxidative stress acts as an initiator of metabolic disorders.” Air pollution likely promotes the onset and progression of diabetes by concurrently inducing visceral adipose tissue inflammation, hepatic steatosis, and hypothalamic inflammation (Zheng et al. 2013, Liu et al. 2014). In the nervous system, PM2.5 may invade the brain via multiple routes, including the circulatory system, the nasal-olfactory bulb pathway, and the gut-brain axis (Calderón-Garcidueñas et al. 2017, Calderón-Garcidueñas et al. 2019). Once inside, these particles can potentially induce neuroinflammation, activate microglia, and promote the accumulation of aberrant proteins such as amyloid-beta (Calderón-Garcidueñas et al. 2022, Calderón-Garcidueñas et al. 2020), providing a plausible mechanistic framework for explaining the association between PM2.5 and neurodegenerative disease risk (HR = 1.46) observed in our study. Therefore, inhaled air pollutants first trigger oxidative stress and inflammatory responses in the lungs; this localized reaction is subsequently amplified into a systemic state; ultimately, persistent systemic inflammation and oxidative damage lead to targeted injury in multiple organs, thereby increasing the risk of various chronic diseases observed in this study.

Our study further elucidates significant synergistic effects between genetic susceptibility and air pollution exposure. We found that individuals with both high genetic risk and high PM2.5 exposure faced the greatest disease burden. This finding indicates that an individual’s genetic background substantially amplifies their vulnerability to the adverse effects of ambient air pollution, potentially through shared pathways such as inflammatory responses and oxidative stress (Al-Kindi et al. 2020, Kaufman et al. 2024).This indicates that the disease burden attributable to air pollution is unevenly distributed across populations, with individuals possessing high genetic susceptibility bearing disproportionately higher risks. This phenomenon is well-supported by research, as consistent evidence demonstrates that populations living in highly polluted areas exhibit higher mortality rates and chronic disease incidence (Bhatnagar 2017, Bhatnagar 2022). When genetic susceptibility factors compound environmental exposure, this risk is further amplified.

Several limitations of our study should be considered. First, potential exposure misclassification exists because we utilized annual average air pollutant concentrations strictly from the year 2010 as a proxy for annual average exposure. Relying on a single-year spatial model does not account for temporal variations in absolute concentrations or the impact of participants’ residential mobility before or after cohort entry. Specifically, although the UK Biobank provides data on “time at current address”, we did not utilize this field to restrict our analyses to long-term residents. Consequently, assigning baseline address exposures to participants who relocated shortly before or after baseline may have introduced deceiving exposure estimates. Second, our analyses were restricted to single-pollutant exposure models. Because we did not implement multi-exposure models, our approach could not fully account for the confounding effects of co-exposures. Therefore, the observed associations for any single pollutant might partially reflect the combined effects of the broader, highly correlated ambient air mixture. Third, outcome misclassification may have occurred since our case ascertainment relied exclusively on hospital inpatient records and death registries. Consequently, individuals with milder forms of these chronic diseases who were managed solely in primary care settings might have been missed, potentially underestimating the true disease incidence and restricting the generalizability of our findings to more severe cases. Fourth, despite extensive covariate adjustments, residual confounding from unmeasured factors cannot be completely ruled out. Crucially, because our Land Use Regression (LUR) models implement geographical covariates to predict ambient exposures, our findings may be subject to spatial confounding. Unmeasured geographically clustered determinants of health—such as variations in water quality, neighborhood-level disparities, or localized cultural differences in sleep and diet—might share spatial structures with air pollution and partially contribute to the estimated disease risks. Additionally, our adjustment for alcohol consumption was based on drinking status (never, previous, or current) rather than quantitative volume (e.g., drinks per week). While this ensured consistency in covariate classification, it may not have fully accounted for the dose-dependent confounding effects of alcohol intake. Therefore, some degree of residual confounding related to alcohol consumption volume may still exist. Therefore, some degree of residual confounding related to alcohol consumption volume may still exist. Future studies would benefit from incorporating more granular data, such as the ‘alcohol intake frequency’ variable (Field ID 1558,https://biobank.ndph.ox.ac.uk/showcase/field.cgi? id=1558) available in the UK Biobank, to more comprehensively account for these dose-dependent effects. Finally, as discussed previously, our continuous Cox models implicitly assumed log-linearity, yielding an “average” risk estimate. However, we mitigated this by reporting categorical (quartile) models that successfully relaxed the strict linearity assumption. Future studies utilizing multi-pollutant statistical frameworks and time-varying exposures are warranted.

Conclusion

Annual average exposure to PM2.5, PM10, NO₂, and NOx is significantly associated with increased incidence rates of multiple chronic diseases, with particularly pronounced effects on diabetes and neurodegenerative disorders. Furthermore, a synergistic interaction exists between genetic susceptibility and air pollution: individuals with high polygenic risk face disproportionately higher disease burdens when exposed to high concentrations of air pollutants.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (274KB, docx)

Author contributions

Jing Tian conceived the study and analyzed the data. Sicong Wang and Xinyi Chang wrote the first and successive drafts of the manuscript. All authors revised the manuscript for important intellectual content. Yi Han had full access to the data and took responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.

Funding

Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0524006). Outstanding Youth Project of Heilongjiang Natural Science Foundation (No.JQ2024H004).

Data availability

Data were accessed from https://www.ukbiobank.ac.uk. Data and code for this study are available from the corresponding authors upon reasonable request.

Declarations

Conflict of interest

The authors declare no conflicts of interest.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Al-Kindi SG, Brook RD, Biswal S et al (2020) Environmental determinants of cardiovascular disease: lessons learned from air pollution. Nat Reviews Cardiol 17(10):656–672 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beelen R, Hoek G, Vienneau D et al (2013) Development of NO2 and NOx land use regression models for estimating air pollution exposure in 36 study areas in Europe – The ESCAPE project. Atmos Environ 72(Complete):10–23 [Google Scholar]
  3. Bhatnagar A (2017) environmental determinants of cardiovascular disease. Circul Res 121(2):162–180 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bhatnagar A (2022) cardiovascular effects of particulate air pollution. Annu Rev Med 73(1):393–406 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Calderón-Garcidueñas L, Reynoso-Robles R, Pérez-Guillé B et al (2017) Combustion-derived nanoparticles, the neuroenteric system, cervical vagus, hyperphosphorylated alpha synuclein and tau in young Mexico City residents[J]. Environ Res 159:186–201 [DOI] [PubMed] [Google Scholar]
  6. Calderón-Garcidueñas L, Reynoso-Robles R, González-Maciel A (2019) Combustion and friction-derived nanoparticles and industrial-sourced nanoparticles: The culprit of Alzheimer and Parkinson’s diseases. Environ Res 176:108574 [DOI] [PubMed] [Google Scholar]
  7. Calderón-Garcidueñas L, Torres-Jardón R, Kulesza RJ et al (2020) Alzheimer disease starts in childhood in polluted Metropolitan Mexico City. A major health crisis in progress. Environ Res 183:109137 [DOI] [PubMed] [Google Scholar]
  8. Calderón-Garcidueñas L, Stommel EW, Lachmann I et al (2022) TDP-43 CSF concentrations increase exponentially with age in metropolitan mexico city young urbanites highly exposed to pm2.5 and ultrafine particles and historically showing alzheimer and parkinson’s hallmarks. brain TDP-43 pathology in mmc residents is associated with high cisternal csf tdp-43 concentrations. Toxics 10(10):559 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cao Y, Hao J, Quan M et al (2025) Air pollution exposure and multimorbidity patterns: evidence from a national cohort study in China. BMC Public Health 25(1):3980 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Carlsen HK, Andersson EM, Molnár P et al (2022) Incident cardiovascular disease and long-term exposure to source-specific air pollutants in a Swedish cohort. Environ Res 209:112698 [DOI] [PubMed] [Google Scholar]
  11. Cheng B, Wei W, Pan C et al (2024) Air pollutants and the risk of incident hepatobiliary diseases: A large prospective cohort study in the UK Biobank. Sci Total Environ, : 175047 [DOI] [PubMed]
  12. Collaborators GDAI (2018) Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 392(10159):1789–1858 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Collaborators GRF (2020) Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet 10258(396):1223–1249 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Eeftens M, Beelen R, de Hoogh K et al (2012) Development of land use regression models for PM2.5, PM2.5 absorbance, PM10 and PMcoarsein 20 european study areas; results of the ESCAPE Project. Environ Sci Technol 46(20):11195–11205 [DOI] [PubMed] [Google Scholar]
  15. Fiffer MR, Chen J, Silva EL et al (2025) Long-term exposure to air pollution and incidence of Type 2 diabetes in the nurses’ health study and nurses’ health study II. Environ Health Perspect 133(6):67009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Glencross DA, Ho T, Camiña N et al (2020) Air pollution and its effects on the immune system. Free Radic Biol Med 151:56–68 [DOI] [PubMed] [Google Scholar]
  17. Haberzettl P, O Toole TE, Bhatnagar A et al (2016) Exposure to fine particulate air pollution causes vascular insulin resistance by inducing pulmonary oxidative stress. Environ Health Perspect 124(12):1830–1839 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Haberzettl P, Conklin DJ, Abplanalp WT et al (2018) Inhalation of fine particulate matter impairs endothelial progenitor cell function via pulmonary oxidative stress. Arterioscler Thromb Vasc Biol 38(1):131–142 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hacker K, Kaufmann R (2024) Chronic Disease mapping, an important strategy and tool for health promotion. Prev Chronic Dis 21:E28 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Halperin R, Tirosh A (2024) Increased cardiovascular risk in patients with germline VHL gene alterations - a UKBIOBANK-based genomic study. Endocr Abstracts
  21. Hill W, Lim EL (2023) Lung adenocarcinoma promotion by air pollutants. Nature 616(7955):159–167 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kaufman JD, Elkind MSV, Bhatnagar A et al Guidance to reduce the cardiovascular burden of ambient air pollutants: a policy statement from the american heart association. Circulation, 2020,142(23). [DOI] [PubMed]
  23. Kim K, Park E, Lee K et al (2009) Differential oxidative stress response in young children and the elderly following exposure to PM2.5. Environ Health Prev Med 14(1):60–66 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kimokoti RW, Millen BE (2016) nutrition for the prevention of chronic diseases. Med Clin North Am 100(6):1185–1198 [DOI] [PubMed] [Google Scholar]
  25. Krishnan RM, Adar SD, Szpiro AA et al (2012) vascular responses to long- and short-term exposure to fine particulate matter. J Am Coll Cardiol 60(21):2158–2166 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kshirsagar AV, Zeitler EM, Weaver A et al (2022) Environmental exposures and kidney disease. kidney360. 3(12):2174–2182 [DOI] [PMC free article] [PubMed]
  27. Li A, Zhou S, Chang J et al (2025) Associations of falls and genetic risk with dementia risk: a prospective cohort study from UK biobank. The Journal of the Alzheimer’s Association, Alzheimer’s & Dementia, p 21 [Google Scholar]
  28. Liu C, Fonken LK, Wang A et al (2014) Central IKKbeta inhibition prevents air pollution mediated peripheral inflammation and exaggeration of type II diabetes. Part Fibre Toxicol 11:53 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Liu C, Xu X, Bai Y et al (2017) Particulate Air pollution mediated effects on insulin resistance in mice are independent of CCR2. Part Fibre Toxicol 14(1):6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lloyd-Jones DM, Hong Y, LAbarthe D et al (2010) Defining and setting national goals for cardiovascular health promotion and disease reduction. Circulation 121(4):586–613 [DOI] [PubMed] [Google Scholar]
  31. Ma Y, Li D, Cui F et al (2024) Air pollutants, genetic susceptibility, and abdominal aortic aneurysm risk: a prospective study. Eur Heart J 45(12):1030–1039 [DOI] [PubMed] [Google Scholar]
  32. Miller MR, Newby DE (2019) Air pollution and cardiovascular disease: car sick. Cardiovascular Res, : cvz228 [DOI] [PubMed]
  33. Miller MR, Raftis JB, Langrish JP et al (2017) Inhaled nanoparticles accumulate at sites of vascular disease. ACS Nano 11(5):4542–4552 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mu G, Zhou M, Wang B et al (2021) Personal PM2.5 exposure and lung function: potential mediating role of systematic inflammation and oxidative damage in urban adults from the general population. Sci Total Environ 755:142522 [DOI] [PubMed] [Google Scholar]
  35. Navratilova HF, Lanham-New S, Whetton AD et al (2024) Associations of diet with health outcomes in the uk biobank: a systematic review. Nutrients 16(4):523 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Palko HA, Fung JY, Louie AY (2010) Positron emission tomography: A novel technique for investigating the biodistribution and transport of nanoparticles. Inhalation Toxicol 22(8):657–688 [DOI] [PubMed] [Google Scholar]
  37. Pope CA, Bhatnagar A, Mccracken JP et al (2016) Exposure to fine particulate air pollution is associated with endothelial injury and systemic inflammation. Circul Res 119(11):1204–1214 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rajagopalan S, Al-Kindi SG, Brook RD (2018) Air pollution and cardiovascular disease. J Am Coll Cardiol 72(17):2054–2070 [DOI] [PubMed] [Google Scholar]
  39. Riggs DW, Zafar N, Krishnasamy S et al (2020) Exposure to airborne fine particulate matter is associated with impaired endothelial function and biomarkers of oxidative stress and inflammation. Environ Res 180:108890 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Riggs DW, Yeager R, Conklin DJ et al (2021) Residential proximity to greenness mitigates the hemodynamic effects of ambient air pollution[J]. Am J Physiol Heart Circ Physiol 320(3):H1102–H1111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Sørensen M, Daneshvar B, Hansen M et al (2003) Personal PM2.5 exposure and markers of oxidative stress in blood. Environ Health Perspect 111(2):161–166 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sudlow C, Gallacher J, Allen N et al (2015) UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 12(3):e1001779 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sun Q, Yue P, Deiuliis JA et al (2009) Ambient air pollution exaggerates adipose inflammation and insulin resistance in a mouse model of diet-induced Obesity. Circulation 119(4):538–546 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Thompson DJ, Wells D, Selzam S et al (2024) A systematic evaluation of the performance and properties of the UK biobank polygenic risk score (PRS) release. PLoS ONE 19(9):e0307270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wang X, Chen L, Cai M et al (2023) Air pollution associated with incidence and progression trajectory of chronic lung diseases: a population-based cohort study. [J] Thorax 78(7):698–705 [DOI] [PubMed] [Google Scholar]
  46. Wiebert P, Sanchez-Crespo A, Seitz J et al (2006) Negligible clearance of ultrafine particles retained in healthy and affected human lungs. Eur Respir J 28(2):286–290 [DOI] [PubMed] [Google Scholar]
  47. Xiong Y, Xia Y (2025) Joint exposure to multiple air pollutants, genetic risk and incident psoriasis: a large-scale prospective cohort study. Br J Dermatol 192(3):420–429 [DOI] [PubMed] [Google Scholar]
  48. Xiong Y, Xia Y, Zhang X et al (2025) point exposure to multiple air pollutants, genetic risk and incident psoriasis: a large-scale prospective cohort study. Br J Dermatol 192(3):420–429 [DOI] [PubMed] [Google Scholar]
  49. Zhang Y, Yang R, Dove A et al (2022) Healthy lifestyle counteracts the risk effect of genetic factors on incident gout: a large population-based longitudinal study. BMC Med 20(1):138 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Zheng Z, Xu X, Zhang X et al (2013) Exposure to ambient particulate matter induces a NASH-like phenotype and impairs hepatic glucose metabolism in an animal model. J Hepatol 58(1):148–154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhu X, Ward J, Cullen B et al (2021) polygenic risk for schizophrenia, brain structure, and environmental risk in UK Biobank. Schizophrenia Bull Open 2(1):sgab042 [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Material 1 (274KB, docx)

Data Availability Statement

Data were accessed from https://www.ukbiobank.ac.uk. Data and code for this study are available from the corresponding authors upon reasonable request.

Articles from Human Genetics are provided here courtesy of Springer

RESOURCES