Table 1.
Studies of the association between air pollution exposure and atherosclerosis
| Location of study, author |
Study design | Exposure metric(s) | Major findings |
|---|---|---|---|
| Los Angeles, USA Künzli 200536 |
798 healthy subjects >40 years old without diabetes or CVD participating in unrelated clinical trials. Cross-sectional association of CIMT with exposures |
Mean ambient PM2.5 exposures estimated using 2000 levels, home address, and GIS modelling Outcome: change in IMT per 10 μg/m3 ↑ in PM2.5 (range 5.2–26.9 μg/m3) |
Unadjusted 5.9 (95% CI 1.0–10.9, P = 0.018) ↑ in IMT Non-significant trend for 4.4 (95% CI 0.0–9.0, P = 0.056) ↑ in IMT (adjusted for age, sex, income) Larger and more significant effects seen in subgroups of women, those older than 60 years, and subjects taking lipid lowering medications. |
| Ruhr Area, Germany Hoffmann 200737 |
4494 subjects in a population-based cohort (age 45–75 years) from the German Heinz Nixdorf Recall Study Cross-sectional association of CAC with exposures |
Mean ambient PM2.5 exposures using 2002 levels, home addresses, and GIS modelling Distance between residence and major road used as a second exposure metric to traffic-related pollution Mean PM2.5 22.8 ± 1.5 μg/m3 |
Trend for 17.2% ↑ in CAC (95% CI −5.6–45.5) per interdecile range (3.91 μg/m3) of PM2.5 in fully adjusted models controlling for other risk factors Significant increase in CAC of 7.0 (95% CI 0.1–14.4) per reduction in distance to major road by 50% Adjusted OR 1.45 (95% CI 1.15–1.82) for CAC >75th % for age for those living within 100 m of major roadway Stronger associations in men, younger, and less-educated subjects. |
| USA (multiple locations) Diez Roux 200838 |
5172 adults without CV disease in the MESA study. Cross-sectional associations of exposure to IMT, ABI, and CAC |
Imputed prior 20-year mean ambient PM2.5 exposure levels by residential history and spatio-temporal models PM2.5 range: 12.8–24.1 μg/m3 |
CIMT increased by 1 (95% CI 0–2%) for a 10–90th percentile increase in PM2.5 (12.5 μg/m3) after controlling for demographics and risk factors. CIMT increased 3 (95% CI 0–5%) per 12.5 μg/m3 increase in PM2.5 using year 2001 mean levels ABI and CAC were not related to metrics of PM2.5 exposures. |
| Ruhr Area, Germany Hoffmann 200934 |
4348 subjects in a population-based cohort (age 45–75 years) from the German Heinz Nixdorf Recall Study Cross-sectional associations of exposures to ABI and prevalence of peripheral arterial disease |
Mean PM2.5 exposures estimated using year 2002 levels, home addresses, and GIS PM2.5 range: 19.8–26.9 μg/m3 Distance between residence and major road used as a second exposure metric to traffic-related pollution Mean PM2.5 levels 22.8 ± 1.5 μg/m3 |
PM2.5 exposures were not related to ABI Living within 50 m of roadway after adjusting for other risk factors was associated with a 0.024 (95% CI 0.047–0.001) reduction in ABI. The OR for having an ABI <0.9 or treatment for peripheral vascular disease was 1.77 (95% CI 1.01–2.1) for individuals living within 50 m of a roadway compared with those living >200 m away. Greater effects were seen in women with no consistent associations observed in men |
| USA, multiple locations Allen 200939 |
1147 adults (age 45–84 years) without CV disease in the MESA study. Cross-sectional associations of exposures to AAC |
PM2.5 exposures based upon average levels over 2 years (2000–2002). Subjects considered traffic exposed if they resided within 100 m of highway or 50 m of major roadway. PM2.5 range: 10.9 ± 0.1 to 22.8 ± 0.9 μg/m3 |
Non-significant 6% (95% CI 0.96–1.16) increase in risk for presence of AAC per 10 μg/m3 increase in PM2.5. Risks for presence of AAC were stronger in those living near a monitor (RR 1.11; 95% CI 1.00–1.24) and those not employed outside home (RR 1.10; 95% CI 1.00–1.22) per 10 μg/m3 increase in PM2.5. Non-significant association with traffic exposures |
| Ruhr Area, Germany Bauer 201040 |
4814 subjects in a population-based cohort (age 45–75 years) from the German Heinz Nixdorf Recall Study Cross-sectional associations of exposure to CIMT |
PM2.5 and PM10 using chemistry transport model (European Air Pollution Dispersion) with input data from emission inventories, meteorology, and regional topography. Values assigned to address. Range of PM2.5: 13.4–22.4 mg/m3 |
An interdecile range increase in PM2.5 (4.2 μg/3), PM10 (6.7 μg/m3, and distance to high traffic (1939 m) associated with a 4.3 [95% confidence interval (CI): 1.9–6.7%], 1.7 (95% CI: −0.7–4.1%), and 1.2 (95% CI: −0.2–2.6%) increase in CIMT, respectively. The association was stronger in younger, obese subjects and in statin users |
| Los Angeles, USA Künzli 201041 | 1483 subjects in five treatment trials related to IMT in the Los Angeles area. Age range: 52.6 ± 8.9–63.7 ± 6.5 years. Annual rate of change of CIMT measured over 1.8 (0.4–2.4)– 3.3 (0.5–5.1) years. Longitudinal association of CIMT progression with PM2.5 and proximity to roadway |
Mean ambient PM2.5 exposures in year 2000 using zip codes and GIS modelling Living <100 m to highway (n = 1.6% of all subjects) PM2.5 range: 20.12 ± 2.82–21.87 ± 1.1. |
Non-significant 2.53 μm (95% CI −0.31–5.38, P = 0.081) increase in IMT progression rate per year per 10 μg/m3 elevation PM2.5 exposure in main model accounting for multiple risk factors. Significant 5.46 μm (95% CI 0.13–10.79, P = 0.044) increase in IMT progression per year for subjects living within 100 m of a highway. This is twice the average annual rate of IMT progression. Larger and significant effects observed in those with low socioeconomic status among those living <100 m from a highway. |
| Greater London, UK Tonne 201242 |
2348 participants of the Whitehall II cohort of British civil servants who had CIMT measured between 2003 and 2005. Cross-sectional association of CIMT with PM10 and PM10 oxidative potential (OP) |
Weekly PM10 and PM10 OP in year prior to scan Mean ambient PM2.5 exposures in year 2000 using zip codes and GIS modelling Weekly OP predicted using measurements of antioxidant-reduced glutathione |
Significant 5% increase in CIMT with 5.2 μgm/m3(interquartile range) in PM10 (95% confidence interval = 1.9–8.3%) after covariate adjustment. The association for an interquartile range change in PM10*OP (1.5 m(−3)) was weaker: 1.2 (0.2–2.2%). |
| Greater Boston, USA Wilker 201343 |
CIMT 350 subjects of the Normative Aging Study between 2004 and 2008 |
1-year average black carbon (BC) using spatiotemporal models Distance to a major roadway and traffic density within a 100-m buffer of residence during the year before the first CIMT measurement. Median predicted BC at baseline for each subject was 0.29 µg/m3, IQR equivalent to 0.26 µg/m3 (25th to 75th quartile |
A 0.26 µg/m3 (one IQR) increase in BC associated with a 1.1% higher CIMT (95% CI: 0.4, 1.7%) based on a fully adjusted model. Living <200 m associated with –1.4 (95% CI: –7.11, 4.6%) lower CIMT (95% CI: –5.7, 1.5%). Higher traffic density associated with higher CIMT |
| Multiple sites, USA Adar 201344 |
5660 subjects of MESA between 2000 and 2005 Cross-sectional and longitudinal associations of CIMT and its progression with PM2.5, respectively |
PM2.5 estimated in year preceding and between CIMT measurements using spatio-temporal models. | A 2.5 µg/m3 higher level of residential PM2.5 during the follow-up period associated with 5.0 µm/year (95% CI 2.6–7.4 µm/year) greater IMT progression among persons in the same metropolitan area. Reduction in PM2.5 over follow-up associated with slowed IMT progression (−2.8 µm/year [95% CI −1.6−3.9 µm/year] per 1 µg/m3 reduction). |
| Ruhr Area Kälsch 201345 |
Thoracic Aortic Calcification (TAC) 4814 patients from the Heinz-Nixdorf Study Cross-sectional association between TAC and PM2.5 and road noise |
Exposure to PM2.5 in the year prior to TAC using chemistry transport model (EURAD-CTM), Road traffic noise using façade levels from noise models as weighted 24 h mean noise (Lden) and night-time noise (Lnight). |
PM2.5 and Lnight associated with TAC-burden of 18.1 (95% CI: 6.6–30.9%)/2.4 µg/m3 PM2.5 and 3.9 (95% CI 0.0; 8.0%)/5 dB(A) Lnight, respectively, in the full model and after mutual adjustment. No effect modification of PM2.5 association by Lnight or vice versa. Low correlation between PM2.5 and noise (r = 0.07–0.10) Increase of 20% in TAC/IQR of PM2.5 corresponded to ∼1 year of older vascular age in this cohort |
| Girona, Spain Rivera 201346 |
2780 participants in the REGICOR (Registre Gironí del Cor: the Gerona Heart Register) study Cross-sectional association between ABI and CIMT and exposures |
Long-term residential NO2 exposure (10 years' time-weighted average based on land-use regression and. traffic intensity). Associations with IMT and ABI using linear regression and multinomial logistic regression, |
Increased NO2, 5th–95th percentiles (25 µg/m3), traffic intensity (15 000 vehicles/day), and traffic load within 100 m (7 200 000 vehicle-m/day) associated with 0.56 (95% CI: −1.5, 2.6), 2.32 (95% CI: 0.48, 4.17), and 1.91 (95% CI: −0.24, 4.06) increase in IMT, respectively. No association with CIMT in adjusted models Exposures were positively associated with an ABI of >1.3, but not <0.9. Stronger association observed among those with a high level of education and in men ≥60 years. |
PM, particulate matter; CV, cardiovascular; CIMT, carotid intima–media thickness; CAC, coronary artery calcium; CI, confidence interval; IMT, intima–media thickness; OR, odds ratio; ABI, ankle–brachial index; AAC, abdominal aortic calcification; GIS, geographic information systems.