If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Ministry of Education Key Laboratory of Environment and Health, State Key Laboratory of Environmental Health (Incubating), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of ChinaDepartment of Maternal and Child Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of China
Ministry of Education Key Laboratory of Environment and Health, State Key Laboratory of Environmental Health (Incubating), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of ChinaDepartment of Maternal and Child Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of China
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, No.38 Xueyuan Road, Beijing, 100191, People's Republic of China
Corresponding author. School of Population Medicine and Public Health, Chinese Academy of Medical Sciences/Peking Union Medical College, No.31, Beijige-3, Dongcheng District, Beijing, People's Republic of China.
School of Population Medicine and Public Health, Chinese Academy of Medical Sciences/Peking Union Medical College, No.31, Beijige-3, Dongcheng District, Beijing, 100730, People's Republic of China
Corresponding author. School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of China.
Ministry of Education Key Laboratory of Environment and Health, State Key Laboratory of Environmental Health (Incubating), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of ChinaDepartment of Maternal and Child Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, No.13 Hangkong Road, Wuhan, 430030, People's Republic of China
Air pollutants are positively correlated with common cardiovascular diseases (CVDs) in the diabetic population.
•
Exposure to air pollutants increases the risk of some underappreciated CVDs.
•
Cardiovascular effects of air pollutants persist even at low concentrations.
Abstract
Background and aims
The adverse effects of air pollutants on the risk of most cardiovascular diseases (CVDs) are well-established, but the risk of CVDs such as deep vein thrombosis, pulmonary embolism, or aortic valve stenosis have been underappreciated, especially in the diabetic population. This study aimed to evaluate associations between long-term air pollutants exposure and the risk of incident CVDs among participants with diabetes.
Methods
This study included 27,827 participants with baseline diabetes from the UK Biobank. We then estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for CVDs associated with chronic air pollutant exposure in the diabetic population by fitting the Cox proportional hazards model. Moreover, we investigated the cardiovascular effects of air pollutants at concentrations below WHO guideline limits.
Results
After multivariable adjustment, long-term NO2 and NOx exposures were positively associated with the development of 8 and 6 types of CVDs in participants with diabetes, respectively. In term of particulate matters, the effect estimates ranged from 1.51 (1.13, 2.03) (coronary artery disease) to 4.65 (2.73, 7.92) (peripheral arterial disease) per 10 μg/m3 increase in PM2.5. Whereas, the effect estimates ranged from 1.15 (1.04, 1.27) (arterial hypertension) to 2.28 (1.40, 3.69) (pulmonary embolism) per 10 μg/m3 increase in PM10. In addition, our study discovered that for most of the cardiovascular events (8 of 9), the deleterious effects of air pollutants persisted even when participants were exposed to air pollutants concentrations below WHO guideline limits.
Conclusions
Long-term exposure to ambient NO2, NOx, PM2.5, and PM10, either at normal or low level, increased risk of various cardiovascular outcomes in the diabetic population.
Ambient air pollution is composed of a mixture of particles and gaseous pollutants (i.e., particulate matters with an aerodynamic diameter less than 2.5 μm (PM2.5) or 10 μm (PM10), nitrogen dioxide (NO2), nitrogen oxide (NOx), etc.). Long-term exposure to ambient air pollution is a major threaten to global life expectancy reduction [
Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a modeling study based on nationwide data.
]. For example, in regions with extreme air pollution levels (i.e., eastern Mediterranean region), particulate matter pollution contributed to approximately 25.67% of total cardiovascular disease (CVD) deaths in 2019 [
Burden of cardiovascular disease attributable to particulate matter pollution in the eastern mediterranean region: analysis of the 1990-2019 global burden of disease.
]. It is now well established from previous studies that exposure to ambient air pollutants is associated with an increased risk of most subtypes of CVD, but these results are usually summarized from the general population [
Taking a stand against air pollution-the impact on cardiovascular disease A joint opinion from the world heart federation, American college of cardiology, American heart association, and the European society of cardiology.
]. Furthermore, uncertainty remains as to whether air pollutants are associated with an increased risk of some underappreciated CVDs (i.e., deep vein thrombosis, pulmonary embolism, or aortic stenosis), particularly in the diabetic population.
As a metabolic disease, diabetes is characterized by chronically elevated blood glucose levels [
]. In recent years, as the prevalence of diabetes in the population has increased and the earlier age of diagnosis, the proportion of individuals with diabetes in the total population has increased considerably and the duration of the disease has been longer [
]. In this case, those people experience a greater risk of the occurrence of CVD and premature mortality over the long course of living with diabetes [
]. Additionally, the diabetic population was thought to be more sensitive to ambient air pollutants because both air pollution exposure and elevated blood glucose levels may contribute to cardiovascular damage through oxidative stress and inflammation [
]. Yet, further research is needed to investigate the specific details of the adverse cardiovascular effects induced by long-term exposure to air pollution in the diabetic population. Moreover, the pathophysiological evidence for the development of CVDs with long-term exposure to air pollutants is different, implying that air pollutants may pose different effects on subtypes of CVD [
Cardiovascular mortality and long-term exposure to particulate air pollution - epidemiological evidence of general pathophysiological pathways of disease.
Therefore, the purpose of this study was to prospectively investigate the associations between long-term exposure to air pollutants and the risk of developing various CVDs in participants with diabetes based on data from a large community cohort in the United Kingdom (UK). In addition to common CVDs, our study also focused on some cardiovascular events that have received less attention (i.e., peripheral arterial disease, atrial fibrillation, deep vein thrombosis, pulmonary embolism, and aortic valve stenosis). Furthermore, another purpose of this study was to provide an overview of the cardiovascular effects of long-term exposure to air pollutants at concentrations below the WHO guideline limits from a unique perspective.
2. Patients and methods
2.1 Study design and population
The UK Biobank, a large-scale population-based prospective cohort study, enrolled 502,480 participants aged 37–73 years since 2006 [
]. It assembled an unprecedented amount of biological and medical information from participants by the touchscreen questionnaire, verbal interview, physical measurements, and provided biological samples in different assessment centers distributed in England, Scotland, or Wales.
2.2 Air pollution estimates
According to the European Study of Cohorts for Air Pollution Effects project, the annual average concentrations of air pollutants (NO2, NOx, PM2.5, or PM10) in the UK Biobank was assessed using a Land Use Regression model (LUR) [
]. With the information of participants' residential addresses collected at baseline visit, the LUR model could estimate the spatial variations of annual average air pollutants concentrations around the participants’ home addresses via Geographic Information System-derived predictors (i.e., traffic, land use, and topography). The data of concentrations for NOx and PM2.5 were only available for 2010 while for NO2 (2005, 2006, 2007, and 2010) and PM10 (2007 and 2010) they were recorded in several years, and thus corresponding average values were included into the analyses.
2.3 Definitions of diabetes
The definition of diabetes was referenced by an existing study [
Association of serum 25-hydroxyvitamin D with cardiovascular outcomes and all-cause mortality in individuals with prediabetes and diabetes: results from the UK Biobank prospective cohort study.
], which included self-reported, physician diagnose, or the medication history of insulin and hypoglycemic drug using. Besides these, participants with glycated hemoglobin (HbA1c) level ≥48 mmol/mol (6.5%) were also identified as diabetes.
2.4 Assessment of outcomes
The outcomes for our study were a broad range of CVDs, including cerebrovascular diseases (stroke), thrombo-embolic diseases (deep vein thrombosis and pulmonary embolism), and other CVDs (coronary artery disease, aortic valve stenosis, atrial fibrillation, heart failure, and peripheral vascular disease) as well as arterial hypertension. Through the record linkage to Health Episode Statistics in UK (England, Wales, and Scotland), dates of diagnosis for hospital admissions were determined. Incident cases were confirmed in the national hospital registers as defined by the 9th and 10th Revision of International Classification of Diseases (ICD 9 and ICD 10) (Supplementary Table S1). Most types of CVDs are life-threatening once they occur and therefore require hospital admission for treatment. Therefore, the use of ICD codes from electronic health records to define outcomes of interest has the potential to reflect the incidence of most CVDs and is widely used in many studies. Previous studies have assessed the accuracy and validity of ICD codes in the UK Biobank for certain types of CVDs and reported a higher positive predictive value (>90%) [
Accuracy of electronic health record data for identifying stroke cases in large-scale epidemiological studies: a systematic review from the UK Biobank stroke outcomes group.
]. Thus, CVDs cases determined by ICD codes in this study were reasonable.
2.5 Ascertainment of covariates
The baseline questionnaire in the UK Biobank provided information on sociodemographic factors, physical measurements, and medical history. According to the literature [
], we included a series of sociodemographic factors of population, risk factors of CVD, or predictors of exposure as possible confounders, including age (years), sex (male or female), body mass index (BMI) (<25, 25–30, ≥30 kg/m2), household income (<£31,000 or ≥ £31,000, £31,000 is closest to the UK median household income), ethnicity (white European, mixed, Asian, black, Chinese, or other), educational background (degree level education or non-College level), alcohol status (never, former, or current drinkers), smoking status (never, former, or current smokers), healthy diet score (0, 1, 2, 3, 4, or 5), physical activity (never activity, low activity, medium activity, or high activity), hypertension (yes or no), hypercholesterolemia (yes or no), and hypertriglyceridemia (yes or no, hypertriglyceridemia was define as triglyceride levels of 1.7 mmol/L or greater). Physical activity was divided into four types according to responses to the questionnaire [
Frailty and pre-frailty in middle-aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493 737 UK Biobank participants.
]. A healthy diet score was derived from the self-reported diet information and ranged from 0 to 5. One point was added for each favorable dietary factor: vegetable intake ≥ four tablespoons/day, fruit intake ≥ three pieces/day; fish intake ≥ twice/week, unprocessed red meat intake ≤ twice/week, and processed meat intake ≤ twice/week [
Of the 502,480 participants in this study, we excluded participants without the diagnosed diabetes or abnormal glucose level at baseline (n = 472,319) and those with missing information on air pollutants (NO2, NOx, PM2.5, and PM10, n = 2334). A total of 27,827 participants with diabetes were included in the analysis (Supplementary Fig. S1).
2.7 Statistical analysis
Participants characteristics were illustrated as mean (± standard deviation (SD)) for continuous variables and number (proportion) for categorical variables. Missing data of variable were coded as a missing indicator category. Kolmogorov-Smirnov test was used to check normal distribution of continuous variables, and Levene's test was used to check variance heterogeneity in different groups. The continuous variables were then compared across groups by Student's t-test or Mann-Whitney U test, as appropriate. Categorical variables in different groups were compared using the Chi-Squared or Fisher exact tests, as appropriate.
Cox proportional hazards models were performed to assess the associations between long-term air pollutants exposure (continuous and categorized by quintile) and the risk of incidence of 9 cardiovascular events in the diabetic population. Survival time was estimated as the time from the response date of the baseline survey to the occurrence of any of the 9 cardiovascular events or date of censoring. The multivariable-adjusted Cox models were fitted by incorporating several potential covariates: age, sex, ethnicity, BMI, educational background, household income, alcohol status, smoking status, healthy diet score, physical activity, hypertension, hypercholesterolemia, and hypertriglyceridemia. Schoenfeld residuals were used to test the proportional hazards assumption, and no violation was observed. The results were expressed as hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs).
To examine concentration-response curves between air pollutants and the risk of 9 cardiovascular events in the diabetic population, the Cox proportional hazards models were fitted based on restricted cubic spline (RCS). Stratified analyses were also conducted according to age (≤60 years old or > 60 years old), sex (male or female), and household income (<£31,000 or ≥£31,000). Moreover, to explore whether these associations persisted at low-level of air pollutants exposure, analyses were performed among participants exposed to low-levels of NO2 ( 40 μg/m3), PM2.5 ( 10 μg/m3), and PM10 ( 20 μg/m3) according to the WHO air quality guideline limits, as well as among participants exposed to low-level of NOx (<the median level of NOx).
A series of sensitive analyses were conducted to examine the robustness of results. First, to examine whether the associations between chronic air pollutants exposure and the risk of incident CVDs differed by diabetes status, we stratified the analysis according to the diabetes status in all participants. Second, we excluded participants who were diagnosed with any of the mentioned cardiovascular conditions within the first 2 years of follow-up, to disentangle the temporal concerns regarding the relationship between air pollution and CVDs. Third, we excluded participants with missing covariates. Fourth, we adjusted baseline history of antihypertensive medication using (yes/no) and cholesterol lowering medication using (yes/no) in the Cox proportional hazards models. Fifth, we assumed that participants with unidentifiable lifestyle status as having optimal status and fitted the models. Finally, the sensitive analysis was only conducted in participants with length of living time at current address more than 5 years to assess the long-term effects of air pollutants on the risk of CVDs. All data analyses were performed by R program (version: 4.0.1). Two-sided p < 0.05 was considered as statistical significance.
3. Results
The baseline characteristics of the included participants are presented in Table 1. Of the total study population, the mean age was 59.39 (±7.27 years), males accounted for 60.6% and white European accounted for 85.6%. Median years of follow-up for the incident 9 cardiovascular outcomes in diabetic participants are documented in Supplementary Table S2. Compared to participants without cardiovascular events, those with cardiovascular events were older, predominantly males, had a high proportion of obesity, a poorer educational background, and lower household income. In addition, participants with cardiovascular events tended to be current smokers but less likely to be current drinkers, and had a higher prevalence of hypertension, hypercholesterolemia, and hypertriglyceridemia at baseline. The mean (±SD) concentrations of air pollutants (NO2, NOx, PM2.5, and PM10) among participants with CVDs were 30.74 (9.31), 46.52 (16.54), 10.16 (1.07), and 19.57 (1.94) μg/m3, respectively, and the corresponding concentrations were 30.43 (9.63), 45.62 (16.84), 10.08 (1.08), and 19.57 (2.00) μg/m3 for those without cardiovascular events.
Table 1Baseline characteristics of participants included in this study.
Total participants (n = 27,827)
Participants without CVDs (n = 7953)
Participants with CVDs (n = 19,874)
p values
Age (year)
59.39 (7.27)
56.40 (7.86)
60.59 (6.66)
< 0.001
Sex (Female, %)
10,971 (39.4)
3529 (44.4)
7442 (37.4)
< 0.001
Ethnicity (%)
< 0.001
White
23,824 (85.6)
6667 (83.8)
17,157 (86.3)
Mixed
195 (0.7)
71 (0.9)
124 (0.6)
Asian
1899 (6.8)
575 (7.2)
1324 (6.7)
Black
1057 (3.8)
332 (4.2)
725 (3.6)
Chinese
103 (0.4)
57 (0.7)
46 (0.2)
Other
518 (1.9)
177 (2.2)
341 (1.7)
Missing
231 (0.8)
74 (0.9)
157 (0.8)
BMI (%)
< 0.001
<25
3181 (11.4)
1404 (17.7)
1777 (8.9)
25–30
9471 (34.0)
3097 (38.9)
6374 (32.1)
≥30
15,175 (54.5)
3452 (43.4)
11,723 (59.0)
Educational background (%)
< 0.001
Degree level education
8017 (28.8)
2742 (34.5)
5275 (26.5)
Non-College education
11,629 (41.8)
3575 (45.0)
8054 (40.5)
Missing
8181 (29.4)
1636 (20.6)
6545 (32.9)
Alcohol status (%)
< 0.001
Never
2518 (9.0)
700 (8.8)
1818 (9.1)
Previous drinkers
1993 (7.2)
399 (5.0)
1594 (8.0)
Current drinkers
23,171 (83.3)
6808 (85.6)
16,363 (82.3)
Missing
145 (0.5)
46 (0.6)
99 (0.5)
Smoking status (%)
< 0.001
Never
12,713 (45.7)
4206 (52.9)
8507 (42.8)
Previous smokers
11,707 (42.1)
2792 (35.1)
8915 (44.9)
Current smokers
3140 (11.3)
870 (10.9)
2270 (11.4)
Missing
267 (1.0)
85 (1.1)
182 (0.9)
Healthy diet score (%)
0.150
0–1
4478 (16.1)
1262 (15.9)
3216 (16.2)
2–3
15,369 (55.2)
4344 (54.6)
11,025 (55.5)
4–5
7980 (28.7)
2347 (29.5)
5633 (28.3)
Physical activity (%)
< 0.001
Never
3858 (13.9)
843 (10.6)
3015 (15.2)
Low activity
2771 (10.0)
645 (8.1)
2126 (10.7)
Medium activity
17,285 (62.1)
5359 (67.4)
11,926 (60.0)
High activity
3200 (11.5)
981 (12.3)
2219 (11.2)
Missing
713 (2.6)
125 (1.6)
588 (3.0)
Household income (%)
< 0.001
<£31,000
14,667 (52.7)
3524 (44.3)
11,143 (56.1)
≥£31,000
7902 (28.4)
3069 (38.6)
4833 (24.3)
Missing
5258 (18.9)
1360 (17.1)
3898 (19.6)
Hypertension (%)
13,324 (47.9)
2422 (30.5)
10,902 (54.9)
< 0.001
Hypercholesterolemia (%)
6999 (25.2)
1667 (21.0)
5332 (26.8)
< 0.001
Hypertriglyceridemia (%)
16,125 (57.9)
4219 (53.0)
11,906 (59.9)
< 0.001
NO2 (μg/m3)
30.65 (9.40)
30.43 (9.63)
30.74 (9.31)
0.014
NOx (μg/m3)
46.26 (16.63)
45.62 (16.84)
46.52 (16.54)
< 0.001
PM2.5 (μg/m3)
10.14 (1.07)
10.08 (1.08)
10.16 (1.07)
< 0.001
PM10 (μg/m3)
19.57 (1.96)
19.57 (2.00)
19.57 (1.94)
0.719
BMI, body mass index; NO2, nitrogen dioxide; NOx, nitrogen oxides; PM2.5, fine particulate matter with diameter <2.5 μm; PM10, particulate matter with diameter <10 μm.
Fig. 1 shows the relationships between long-term air pollutants exposure and the risk of 9 cardiovascular outcomes in participants with diabetes. In the multivariable Cox models, exposure to NO2 was positively associated with 8 of the 9 outcomes (coronary artery disease, heart failure, arterial hypertension, peripheral arterial disease, stroke, atrial fibrillation, deep vein thrombosis, and pulmonary embolism). Per 10 μg/m3 increase in NO2 concentration, the HRs (95% CIs) for these 8 cardiovascular outcomes were 1.07 (1.03, 1.11), 1.09 (1.04, 1.14), 1.06 (1.04, 1.08), 1.16 (1.09, 1.25), 1.13 (1.07, 1.19), 1.05 (1.01, 1.10), 1.25 (1.11, 1.41), and 1.19 (1.08, 1.31), individually. In addition, each 10 μg/m3 increase in NOx was significantly associated with 2% (0%, 4%) increase in the incidence of coronary artery disease, 3% (1%, 6%) increase in the incidence of heart failure, 2% (1%, 4%) increase in the incidence of arterial hypertension, 7% (4%, 11%) increase in the incidence of peripheral arterial disease, 3% (0%, 6%) increase in the incidence of stroke, and 9% (4%, 14%) increase in the incidence of pulmonary embolism. In terms of particulate matters, exposures to ambient PM2.5 or PM10 were significantly associated with the initiation of 6 of 9 cardiovascular outcomes in the diabetic population. For each 10 μg/m3 increase in PM2.5, the effect estimates ranged from 1.51 (1.13, 2.03) (coronary artery disease) to 4.65 (2.73, 7.92) (peripheral arterial disease), while for each 10 μg/m3 increase in PM10, the effect estimates ranged from 1.15 (1.04, 1.27) (arterial hypertension) to 2.28 (1.40, 3.69) (pulmonary embolism).
Fig. 1Associations between long-term NO2 (A), NOx (B), PM2.5 (C), and PM10 (D) exposure and the risk of incident 9 cardiovascular outcomes in diabetic participants from the UK Biobank.
When fitting Cox models with categorical exposure, we found that long-term exposure to air pollutants was associated with development of 8 of 9 cardiovascular outcomes in participants with diabetes (p for trend <0.05, Supplementary Table S3). Fig. 2 presents the concentration-response relationships of PM2.5, PM10, NO2, and NOx concentrations with the risk of 9 cardiovascular outcomes in diabetic participants. We found monotonic increase in the dose-response relationships between particulate matters exposure and the risk of all cardiovascular outcomes in diabetic participants. As shown in Supplementary Table S4, when analyses were stratified by age, sex, and household income, the relationships of air pollutants and CVDs risk did not meaningfully differ by age, sex, or household income groups (all p for interaction >0.05).
Fig. 2Associations of air pollutants with the risk of incident 9 cardiovascular outcomes in diabetic participants from the UK Biobank.
A restricted cubic spline regression model with 3 knots (at the 10th, 50th, and 90th percentiles) was used to estimate the dose-response relations between air pollutants and incident coronary artery disease among participants with diabetes. Hazard ratios (solid lines) and 95% CIs (shaded areas) were adjusted for age, sex, ethnicity, educational background, alcohol status, smoking status, healthy diet score, physical activity, BMI, hypertension, hypercholesterolemia, and hypertriglyceridemia. HRs, hazard ratios; CIs, confidence intervals; NO2, nitrogen dioxide; NOx, nitrogen oxides; PM2.5, fine particulate matter with diameter <2.5 μm; PM10, particulate matter with diameter <10 μm; BMI, body mass index.
Table 2 illustrates the harmful effects of low-level air pollutants exposure on cardiovascular events in participants with diabetes. When exposure levels were limited below the WHO air quality guideline limits, exposure to NO2 was positively associated with almost all type of CVDs in participants with diabetes. Exposure to NOx was significantly associated only with arterial hypertension (1.00 (1.00, 1.01)) and peripheral arterial disease (1.03 (1.01, 1.04)). Low level PM2.5 exposure was related to the development of 4 cardiovascular outcomes, with HRs (CIs) of 1.10 (1.01, 1.21), 1.08 (1.02, 1.14), 1.39 (1.15, 1.69), and 1.16 (1.00, 1.34) for coronary artery disease, arterial hypertension, peripheral arterial disease, and stroke, respectively. However, no statistically significant association was found between low-level of PM10 exposure and the occurrence of any cardiovascular events.
Table 2Associations of various air pollutants (continual) in low-level concentrations with the risk of incident 9 cardiovascular outcomes among diabetic participants from the UK Biobank.
Outcomes
NO2
NOx
PM2.5
PM10
Cases/N
HRs (95% CIs)
Cases/N
HRs (95% CIs)
Cases/N
HRs (95% CIs)
Cases/N
HRs (95% CIs)
Coronary artery disease
1844/11,660
1.01 (1.01, 1.02)*
2523/15,560
1.00 (0.99, 1.01)
3429/20,944
1.10 (1.01, 1.21)*
1919/12,266
1.01 (0.97, 1.05)
Heart failure
999/12,955
1.02 (1.01, 1.02)*
1485/17,367
1.00 (0.99, 1.01)
2017/23,351
1.09 (0.96, 1.23)
1082/13,636
1.03 (0.98, 1.08)
Arterial hypertension
5189/9830
1.01 (1.00, 1.01)*
7036/12,945
1.00 (1.00, 1.01)*
9443/17,367
1.08 (1.02, 1.14)*
5384/10,163
1.01 (0.99, 1.03)
Peripheral arterial disease
434/13,087
1.02 (1.01, 1.03)*
644/17,511
1.03 (1.01, 1.04)*
887/23,541
1.39 (1.15, 1.69)*
472/13,743
1.05 (0.98, 1.13)
Stroke
751/13,037
1.01 (1.00, 1.02)*
1016/17,470
1.00 (0.99, 1.01)
1386/23,469
1.16 (1.00, 1.34)*
766/13,699
1.04 (0.99, 1.11)
Atrial fibrillation
1442/12,826
1.01 (1.00, 1.01)*
1979/17,169
1.00 (0.99, 1.01)
2608/23,084
1.09 (0.99, 1.21)+
1537/13,491
1.02 (0.98, 1.06)
Deep vein thrombosis
129/13,132
1.03 (1.01, 1.05)*
185/17,604
1.01 (0.99, 1.04)
251/23,670
1.23 (0.88, 1.74)
138/13,814
1.02 (0.90, 1.17)
Pulmonary embolism
214/13,154
1.02 (1.00, 1.03)+
304/17,625
0.99 (0.97, 1.01)
404/23,703
1.09 (0.83, 1.42)
220/13,834
1.09 (0.98, 1.21)
Aortic valve stenosis
358/13,179
1.00 (0.99, 1.02)
500/17,657
0.99 (0.98, 1.01)
656/23,744
0.97 (0.80, 1.19)
377/13,864
0.97 (0.90, 1.05)
HRs and 95% CIs in bold represent significance at p < 0.05 (*). p slightly higher than 0.05 are recognized as suggested significant (+).
In sensitivity analysis, the risk of most cardiovascular outcomes (heart failure, peripheral arterial disease, stroke, atrial fibrillation, deep vein thrombosis, and aortic valve stenosis) did not differ between diabetic and non-diabetic populations after long-term air pollutants exposure (p for interaction >0.05), except for coronary artery disease, arterial hypertension, and pulmonary embolism (p for interaction <0.05, Supplementary Table S5). Additionally, the associations between air pollutants and the risk of CVDs in participants with diabetes were not changed when participants diagnosed with any of the mentioned CVDs within the first 2 years of follow-up were excluded (Supplementary Table S6), when Cox models were fitted with the imputed dataset (Supplementary Table S7), and when the mean level of 24-h noise pollution was further adjusted in the Cox models (Supplementary Table S8). In addition, the results of the current study were generally robust when the Cox model was fitted by assuming that participants with unidentifiable lifestyle status had optimal status (Supplementary Table S9) and when the participants were restricted to those who had lived at their current address for more than 5 years (Supplementary Table S10).
4. Discussion
This nationwide prospective study has comprehensively evaluated the risk of chronic air pollutant exposure on the development of various cardiovascular events in the diabetic population. Besides the common types of CVDs, this study also discovered significant positive associations between long-term air pollutants exposure and the risk of some of underappreciated cardiovascular events (i.e., peripheral arterial disease, atrial fibrillation, deep vein thrombosis, and pulmonary embolism) in the diabetic population. Additionally, the cardiovascular effects persisted even when participants were exposed to air pollutants at concentrations below the WHO guideline limits. Moreover, the linear dose-response functions suggested that no evidence of a threshold for the association of air pollutants with the risk of cardiovascular events in diabetic population.
In the general population, the cardiovascular effects of acute or chronic air pollutants exposure have been well explored in the available publications [
]. However, most existing literatures focused on the detrimental effects of long-term air pollutants exposure on only one or a few cardiovascular events in the diabetic population. A study (n = 2072) from Taiwan investigated the associations of PM2.5 exposure with the incident risk of CVD among participants with diabetes and reported a positive association for increased risk of CVD (HR:1.040 (1.004, 1.073) per 1 μg/m3 increase in PM2.5 concentration) [
]. Another study using the Nurses’ Health Study found that women with diabetes were more vulnerable to the adverse cardiovascular effects of particulate matters (1.44 (1.23, 1.68) and 1.19 (1.10,1.28) per 10 μg/m3 increase in PM2.5 and PM10 concentrations, respectively) [
]. Similarly, an Iranian study observed stronger associations between PM2.5 exposure and the occurrence of a set of CVD events (myocardial infarctions, ischemic heart disease, and stroke) among subjects with diabetes [
Long-term exposure to PM2.5 and cardiovascular disease incidence and mortality in an Eastern Mediterranean country: findings based on a 15-year cohort study.
]. Findings of the above-mentioned research are consistent with our results. For the uncommon types of CVDs, such as peripheral arterial disease, deep vein thrombosis, pulmonary embolism, and aortic valve stenosis, evidence on the cardiovascular effects of air pollutants exposure mainly concentrated in the general population [
Exposure to long-term air pollution and incidence of peripheral arterial disease in the general population: a Korean national population-based retrospective cohort study.
]. However, this study estimated relatively larger HRs and broader CIs for deep vein thrombosis and pulmonary embolism associated with long-term air pollutants exposure. The underlying reason may be related to the small number of incident cases of deep vein thrombosis (n = 306) and pulmonary embolism (n = 478) during follow-up. The small sample size may have resulted in a wide range of CIs, which will further yield less definitive conclusions [
]. Future cohort studies should be carefully consider this issue.
When it comes to whether the extent of cardiovascular effects of air pollutants differs between diabetic and non-diabetic populations, our study found that the risk of certain types of cardiovascular disease (i.e., coronary artery disease, arterial hypertension, and pulmonary embolism) differed between diabetic and non-diabetic populations. In line with our study, previous studies have shown that adults with diabetes may be more susceptible to the cardiovascular effects induced by long-term air pollutants exposure [
]. In term of biological mechanisms, our findings are plausible. Available evidence to date suggests that diabetic population is often affected by cardiovascular risk factors, such as chronic hyperglycemia, diabetic nephropathy, cardiac autonomic neuropathy, hypoglycemia, or glucose variability [
]. These factors, in turn, induce the incident risk of CVD by promoting oxidative stress, vascular inflammation, monocyte adhesion, arterial wall thickening and endothelial dysfunction, even in well-controlled diabetes patients [
]. Thus, pathophysiological progressions may make the diabetic population more likely to develop CVDs.
Demographic characters analysis has played an increasing role in the epidemiology of air pollution, but conclusions regarding the effect modifications by sociodemographic characters remain uncertain in the available studies focusing on long-term air pollution exposure [
Exposure to ambient air pollution and the incidence of congestive heart failure and acute myocardial infarction: a population-based study of 5.1 million Canadian adults living in Ontario.
]. Of note, through stratified analyses, our study identified no significant effect modification of age, sex, and household income on the cardiovascular effects of air pollutants exposure in diabetic population, suggesting that there are no significant differences in the risk of incident CVDs after long-term exposure to air pollutants among diabetics of different ages, genders, and household income.
When exposure levels were below air quality threshold limits, some previous studies have reported no or only weak associations between air pollutants exposure and the development of cardiovascular events in general population [
Long-term exposure to low-level ambient air pollution and incidence of stroke and coronary heart disease: a pooled analysis of six European cohorts within the ELAPSE project.
High-resolution mapping of traffic related air pollution with Google street view cars and incidence of cardiovascular events within neighborhoods in Oakland, CA.
]. However, this study highlighted that the cardiovascular effects of long-term exposure to air pollutants persisted even when exposure levels are below the WHO air quality threshold limits. More importantly, in addition to common CVDs (coronary artery disease, heart failure, arterial hypertension, and stroke), our findings added novel cohort evidence that the risk of developing peripheral arterial disease, atrial fibrillation, and deep vein thrombosis is increased even at low levels of air pollutant exposure. Concentration-response relationship is of great research interest in the field of air pollution epidemiology. In our study, the concentration-response function curves for almost all cardiovascular events were monotonically increasing in the low range of air pollutants (NO2 < 30 μg/m³, NOx < 40 μg/m³, PM2.5<10 μg/m³, PM10 < 20 μg/m³), which is consistent with studies conducted in general population of UK and Canada [
Exposure to ambient air pollution and the incidence of congestive heart failure and acute myocardial infarction: a population-based study of 5.1 million Canadian adults living in Ontario.
]. Our findings suggested that exploring more reasonable pollution control strategies (such as addressing the problem that excessive use of fossil fuels [
]) has the potential to reduce the incidence of air-pollution-induced CVDs.
There were several major strengths in this study. First, this study provided new information related to a number of underappreciated cardiovascular events (i.e., peripheral arterial disease, atrial fibrillation, deep vein thrombosis, pulmonary embolism, and aortic valve stenosis) that have not been specifically evaluated in previous studies on the relationship between air pollutants and incident CVDs in the diabetic population. Second, based on a nationwide population study, our analysis has a large sample size and multiple types of cardiovascular outcomes, so the statistical power and robustness of results were higher than most analyses at present. Third, the cardiovascular effects persists even when participants were exposed to air pollutants at concentrations below the WHO guideline limits, which has a potential guiding significance for the primary prevention of cardiovascular events in the susceptible populations.
However, there were still some limitations in our study. First, the exposure data of air pollutants might be over- or underestimated in UK Biobank, because environmental exposure data were not available for locations other than home address. Second, the LUR model allows the reconstruction of historical exposure to air pollution among the UK Biobank participants, but the uncertainties in the air pollution exposure assessment may lead to bias to some extent. Third, only four air pollutants were available from UK Biobank, but other pollutants (i.e., ozone, ultrafine particles, sulfur dioxide, etc.) may also be associated with CVDs [
Associations of long-term exposure to ultrafine particles and nitrogen dioxide with increased incidence of congestive heart failure and acute myocardial infarction.
]. Fourth, despite WHO has now released updated air quality guideline, this study did not follow the latest version of the low-level definition of air pollutants exposure, as participants in UK Biobank were exposed to air pollutants at levels higher than the latest cutoff values. Fifth, our study is limited in terms of ethnic diversity (>85% White Europeans), so the results may not be directly generalizable to other ethnic groups. Sixth, cardiovascular events were identified using ICD coding systems in the current study. However, certain types of CVDs, such as atrial fibrillation, deep vein thrombosis, pulmonary embolism, or aortic valve stenosis, have considerable variation in diagnostic criteria and disease subtypes [
]. Moreover, ICD diagnosis codes mainly reflect inpatient diagnoses. For some less severe diseases, if we rely solely on ICD codes to define outcomes, this will lead to variance in cases and underreporting of cases. Future studies should consider combining ICD coding systems and other outcome ascertainment methods (i.e., electrocardiographic records) to define these CVDs. Seventh, due to the lack of biomarkers information related to pathogenesis in UK Biobank, we cannot confirm the study results mechanistically. Finally, although many covariates potentially affecting CVDs were considered and adjusted for in our statistical models, there may still be unmeasured or residual confounders affecting the analysis.
4.1 Conclusion
In summary, we found that long-term exposure to a series of air pollutants were pronouncedly associated with increased risk of a broad range of common or uncommon cardiovascular events in diabetic population. Additionally, in diabetic individuals, the cardiovascular effects of air pollutants persisted even at concentrations below the WHO air quality guideline limits. These findings have important implications for the recommendation and implementation of more stringent air quality guideline to protect the susceptible population.
Author contributions
YHT and YDYM contributed to the conception and design of the study. YHT and YDYM advised on all statistical aspects and interpreted the data. YDYM, DKL, JQX, and YHH performed the literature search and the analyses. All authors critically reviewed this and previous drafts. All authors approved the final draft for submission, with final responsibility for publication. All authors approved the final version of the manuscript. The corresponding author attests that all the listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Declaration of competing interest
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.
Acknowledgments
We thank all the staff and participants of the UK Biobank for their hard work and dedication in collecting the underlying data. The UK Biobank received ethics approval from the North West Multicenter Research Ethics Committee (reference no. 16/NW/0274). The study involved all participants providing written informed consent.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a modeling study based on nationwide data.
Burden of cardiovascular disease attributable to particulate matter pollution in the eastern mediterranean region: analysis of the 1990-2019 global burden of disease.
Taking a stand against air pollution-the impact on cardiovascular disease A joint opinion from the world heart federation, American college of cardiology, American heart association, and the European society of cardiology.
Cardiovascular mortality and long-term exposure to particulate air pollution - epidemiological evidence of general pathophysiological pathways of disease.
Association of serum 25-hydroxyvitamin D with cardiovascular outcomes and all-cause mortality in individuals with prediabetes and diabetes: results from the UK Biobank prospective cohort study.
Accuracy of electronic health record data for identifying stroke cases in large-scale epidemiological studies: a systematic review from the UK Biobank stroke outcomes group.
Frailty and pre-frailty in middle-aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493 737 UK Biobank participants.
Long-term exposure to PM2.5 and cardiovascular disease incidence and mortality in an Eastern Mediterranean country: findings based on a 15-year cohort study.
Exposure to long-term air pollution and incidence of peripheral arterial disease in the general population: a Korean national population-based retrospective cohort study.
Exposure to ambient air pollution and the incidence of congestive heart failure and acute myocardial infarction: a population-based study of 5.1 million Canadian adults living in Ontario.
Long-term exposure to low-level ambient air pollution and incidence of stroke and coronary heart disease: a pooled analysis of six European cohorts within the ELAPSE project.
High-resolution mapping of traffic related air pollution with Google street view cars and incidence of cardiovascular events within neighborhoods in Oakland, CA.
Exposure to ambient air pollution and the incidence of congestive heart failure and acute myocardial infarction: a population-based study of 5.1 million Canadian adults living in Ontario.
Associations of long-term exposure to ultrafine particles and nitrogen dioxide with increased incidence of congestive heart failure and acute myocardial infarction.
00065-5&id=ga1.jpg){kind=link}
00065-5&id=gr1.jpg){kind=link}
00065-5&id=gr2.jpg){kind=link}