Air pollution, vascular disease and thrombosis: linking clinical data and pathogenic mechanisms


Antonio Coppola, Regional Reference Centre for Coagulation Disorders, Federico II University Hospital, Via S. Pansini, 5 – 80131 Naples, Italy.
Tel.: +39 0817462317; fax: +39 0815466152.


Summary.  The public health burden of air pollution has been increasingly recognized over the last decades. Following the first assessed adverse effects on respiratory diseases and lung cancer, a large body of epidemiologic and clinical studies definitely documented an even stronger association of air pollution exposure with cardiovascular mortality and morbidity, particularly related to atherothrombotic (coronary and cerebrovascular) disease. Particulate matter (PM), mainly that with lower aerodynamic diameter (fine and ultrafine PM), is responsible for the most severe effects, due to its capacity to transport toxic substances deep into the lower airways. These effects have been shown to occur not only after short-term exposure to elevated concentrations of pollutants, but even after long-term relatively low levels of exposure. Vulnerable subjects (elderly persons and those with preexisting cardiopulmonary diseases) show the highest impact. Fewer and conflicting data also suggest an association with venous thromboembolism. Although not completely elucidated, a series of mechanisms have been hypothesized and tested in experimental settings. These phenomena, including vasomotor and cardiac autonomic dysfunction, hemostatic unbalance, oxidative stress and inflammatory response, have been shown to change over time and differently contribute to the short-term and long-term adverse effects of pollution exposure. Beyond environmental health policies, crucial for improving air quality and reducing the impact of such an elusive threat to public health, the recognition and assessment of the individual risk, together with specific advice, should be routinely implemented in the strategies of primary and secondary cardiovascular prevention.


The public health burden of environmental air pollution has been increasingly recognized [1]. Since the evidence of the dramatic effects of unusual acutely increased exposure to air pollution (i.e. Meuse Valley, Belgium, 1930, or the ‘great smog’ in London, 1952) [2,3], a large body of literature has provided epidemiological and pathophysiological data supporting the association between air pollution and general mortality and morbidity from respiratory and cardiovascular diseases [1,4–7]. These effects have been shown to occur not only after acute exposure to markedly elevated concentrations of air pollutants, but even after short-term and long-term relatively low levels of exposure [8,9]. Beyond the exacerbation of individual risk in vulnerable subjects (elderly persons or those with preexisting cardio-respiratory disease), the unavoidable chronic exposure of large populations to these risks makes air pollution a major health problem [10]. The World Health Organization (WHO) estimates that urban air pollution was responsible for 1.3 million annual deaths in the year 2008 (i.e. 2.4% of the total deaths worldwide) and caused about 9% of the lung cancer deaths, 5% of cardiopulmonary deaths and about 1% of respiratory infection deaths [11].

Sources of air pollution include natural phenomena (biogenic sources, for example volcanoes, wildfires or land dust) and human activities (anthropogenic sources). Natural pollutants may be problematic in some circumstances, but most harmful health effects are related to human-generated sources of air pollution, categorized as mobile and stationary sources. The former include motor vehicle exhausts, the latter consist of non-moving sources, such as household combustion devices, industrial facilities and power plants [12]. In addition to primary pollutants, directly emitted from a source (for example, carbon monoxide from vehicle exhausts or sulphur dioxide from industrial processes), secondary pollutants are produced by chemical reactions of primary emissions in the atmosphere (for example, ozone generated by the reaction of volatile hydrocarbons with the sunlight). The result of these multiple, concurrent sources and processes is a complex mixture (Table 1) of gaseous and particulate matter (PM). Depending on its sources and geographical location, the composition of air pollution may be extremely heterogeneous, and further variations in a single location are related to meteorological conditions and differences in human activities over time, including time of the day, day of the week, or trends (for example the increase of diesel engines, which produce less carbon dioxide but about 100-fold higher particulate emissions than gasoline engines) [13].

Table 1.   Main components of air pollution
  1. *Particulate matter is distinguished according to its aerodynamic diameters in coarse, fine and ultrafine particles; numbers after the abbreviation indicate the maximum value in μm.

Gaseous substancesCarbon monoxide (CO), carbon dioxide (CO2), nitrogen dioxide (NO2), nitric oxide (NO3), sulphur dioxide (SO2), ozone (O3), volatile organic compounds (hydrocarbons, quinones)
Particulate matter (PM)*Metals, coarse particles (PM10), fine particles (PM2.5), ultrafine particles (PM0.1), nanoparticles

All components of air pollution are harmful to health, but the most numerous and severe effects have been attributed to PM, because particles may contain and transport in the respiratory tract a broad range of toxic substances [14]. In this respect, the toxic potential of PM is inversely related to its aerodynamic diameter (AD) [14–16]. On inhalation, coarse PM (PM10, i.e. particles with AD between 2.5 and 10 μm) is arrested in the nasal cavities and upper airways, whereas fine and, particularly, ultrafine PM (PM2.5 and PM0.1, i.e. particles with < 2.5 and 0.1 μm in AD, respectively) may penetrate deeper up to the lung alveoli and even pass to the bloodstream, where some components may exert direct adverse effects [15–17]. However, more convincing evidence of such translocations (directly as naked particles, or via endocytosis by alveolar macrophages or endothelial cells) has been achieved in animal models than in human studies [16–19]. The heterogeneous composition of PM includes combustion products and suspended crustal materials, as well as biological materials such as pollen, endotoxins, bacteria, spores and viruses. In urban areas, traffic of motor vehicles is a major source of PM, with diesel exhaust accounting for up to 40% of airborne PM [13]. On the whole, ambient PM is considered the most important indicator of the adverse effects on health of global air pollution. The monitoring policies and legislations to improve air quality led to significant benefits in developed countries, whereas many middle-income countries are disproportionately experiencing this burden, due to the lack of strategies for environmental health protection. Indeed, WHO statistics presently report the highest PM10 concentrations in the air of cities in China, India, Pakistan and other rapidly developing countries [20].

If the consequences of air pollutants on respiratory diseases, such as chronic obstructive pulmonary disease and asthma, are easily conceivable, the even stronger impact on cardiovascular disease, more recently assessed, is probably unexpected and pathogenic mechanisms are more difficult to explain [21,22]. The heterogeneity and complexity of air pollution composition is a further challenge to gain insight into these mechanisms [23]. Interestingly, beyond the effects on atherothrombotic (coronary and cerebrovascular) disease, some recent findings suggest an association of air pollution with venous thromboembolism [24]. In the following paragraphs literature data reporting the relationships between air pollution and cardiovascular and thrombotic diseases will be reviewed, together with the experimental studies that provided possible pathophysiological support for these robust and consistent clinical evidence.

Epidemiologic and clinical data

Large and rigorous epidemiologic studies published starting from the 1990s definitely documented the relationships between air pollution and cardiovascular mortality and morbidity. Among these clinical data, studies addressing short-term and long-term effects and outcomes have been usually distinguished [4,6–9,21,22,25,26]. Indeed, a series of researches evaluated the short-term consequences of exposure to increased concentrations of pollutants in metropolitan areas, particularly in terms of mortality or hospitalizations for cardiovascular causes [24–47], whereas other studies highlighted the long-term effects of chronic, persistent exposures through survival analysis and cardiovascular risk assessment of large populations over many years [4,6,48–57]. This categorization is not only consistent with the study design, but also reflects pathophysiological findings, clearly showing that adverse effects of air pollution on the cardiovascular system change over time, differently triggered by short-term, acute or long-term chronic exposure to air pollution [22]. As regards atherothrombotic manifestations, short-term studies (Table 2) investigated the occurrence of fatal and non-fatal events, particularly with respect to coronary artery disease. Fewer data are available concerning cerebrovascular disease and venous thromboembolism (VTE). Long-term studies (Table 3) also addressed the effects on markers of asymptomatic atherosclerosis [58–64]. In some studies, all cardiovascular or heart diseases are pooled, thus specific data concerning atherothrombotic manifestations are not available. In the frame of long-term assessments, conflicting studies recently addressed the issue of the association with VTE (Table 4) [65–67].

Table 2.   Main studies reporting associations of short-term exposure to air pollution with cardiovascular and atherothrombotic disease
Author, years [ref.], Study acronymStudy designCountry, yearsPrimary endpoint(s)Study populationPollutant(s)/assessment*Findings
  1. ACS, acute coronary syndrome (unstable angina and acute myocardial infarction); AMI, acute myocardial infarction; APHEA, Air Pollution and Health European Approach; CHF, congestive heart failure; CI, confidence interval; CO, carbon monoxide; HEAPSS, Health Effects of Air Pollution on Susceptible Subpopulation; IHCS, Intermountain Heart Collaborative Study; IHD, ischemic heart disease; MISA, Metanalysis of Italian Studies on Air pollution; NA, not applicable; NMMAPS, National Morbidity, Mortality and Air Pollution Study; NO2, nitrogen dioxide; O3, ozone; OR, odds ratio; PM0.1, particulate matter with aerodynamic diameter below 0.1 μm; PM2.5, particulate matter with aerodynamic diameter between 2.5 and 0.1 μm; PM10, particulate matter with aerodynamic diameter between 10 and 2.5 μm; PNC, particle number concentration; Pts, patients; RR, rate ratio; SO2, sulphur dioxide. *Only significant associations are reported. Significant associations were also found for black carbon exposure. Weaker associations with ozone levels during the summer, no significant association with levels of CO, NO2 and SO2. §Including myocardial infarction, angina pectoris, dysrhythmia and heart failure.More unstable data were found for NO2 and SO2. **CO and NO2 also showed significant effects regarding fatal cases aged > 75 years. ††No significant association with ozone was found.

(A) Cardiovascular mortality
Samet, 2000 [25], NMMAPSTime seriesUS, 1987–1994All-cause, cardio-respiratory∼50 million of residents in 20 citiesDaily PM10 (10 μg m−3 increase)Relative rate (95% CI) 0.51 (0.07 to 0.93) and 0.68 (0.20 to 1.16) from all causes and cardiovascular or respiratory causes, respectively
Le Tertre, 2002 [28]Time seriesFrance, 1990–1995All cause, cardiovascular, respiratoryData from 9 citiesDaily NO2, SO2, O3 (50 μg m−3 increase)Cardiovascular causes, increase, % (95% CI): NO2 3.8 (2.0 to 5.5), SO2 3.6 (2.1 to 5.2), O3 2.7 (1.3 to 4.1)
Biggeri, 2004 [27], MISA-2Meta-analysisItaly, 1996–2002All-cause, cardiovascular, respiratory∼9.1 million of residents in 9 citiesDaily NO2, SO2, PM10 (10 μg m−3 increase), CO (1 mg m−3 increase)Cardiovascular causes, increase, % (95% CI): NO2 0.4 (0.06 to 1.05), PM10 0.54 (0.02 to 1.02), CO 0.93 (0.10 to 1.77), SO2 1.11 (0.64 to 3.12)
Analitis, 2006 [26], APHEA-2Time seriesEurope, 3 years after 1990All cause, cardiovascular, respiratory∼4.3 million of residents in 29 citiesDaily PM10 (10 μg m−3 increase)Increase, % (95% CI) 0.76 (0.47 to 1.05) and 0.58 (0.21 to 0.95) from all causes and cardiovascular causes, respectively
Kettunen, 2007 [42]Time seriesFinland, 1998–2004StrokeResidents in Helsinki > 65 yearsSame and previous day CO, PM2.5 and PM0.1 (interquartile increase)Warm season, increase, % (95% CI): 6.9 (0.8 to 13.8) and 7.4 (1.3 to 13.8), same- and previous-day PM2.5; 8.3 (0.6 to 16.6) and 8.5 (−1.2 to 19.1), previous day CO and PM0.1
(B) Cardiovascular morbidity
Peters, 2001 [38]Case cross-overUS, 1995–1996AMI771 AMI pts in BostonPM2.5 (25 μg m−3 increase)OR (95% CI) 1.48 (1.09 to 2.02) and 1.69 (1.13 to 2.34) for exposure 2-h and 24-h before the onset
Morris, 2001 [29]Meta-analysisNAHospital admissionsNADaily PM10 (10 μg m−3 increase)Increase, % (95% CI) 0.68 (0.41 to 0.96) for IHD, 0.83 (0.50 to 1.15) for CHF and 0.18 (−0.21 to 0.57) for stroke
Peters, 2004 [37]Case cross-overGermany, 1999–2001AMI691 AMI ptsTraffic exposure within 1 hOR (95% CI) 2.92 (2.22 to 3.83); higher in cyclists (OR 3.94) than in automobile users (OR 2.60)
Von Klot, 2005 [40], HEAPSSCase cross-overEurope, 1992–2001Hospital re-admissions for cardiac causes§22 006 AMI survivors in 5 citiesDaily PM10 (10 μg m−3 increase), CO (200 μg m−3), NO2 (8 μg m−3), O3 (10 μg m−3)Increase, RR (95% CI): PM10 1.021 (1.004 to 1.039), CO 1.014 (1.001 to 1.026), NO2 1.032 (1.013 to 1.051), O3 1.026 (1.001 to 1.051)
Ballester, 2006 [30]Time seriesSpain, 1995–1999Hospital admissionsData from 14 citiesDaily O3, PM10 (10 μg m−3 increase), CO (1 mg m−3 increase)Cardiovascular disease, increase, % (95% CI): O30.7 (0.3 to 1.0), PM10 0.9 (0.4 to 1.5), CO 2.1 (0.7 to 3.5)
Dominici, 2006 [32]Time seriesUS, 1999–2002Hospital admissions∼11.5 million of persons > 65 years in 204 urban countiesDaily PM2.5 (10 μg m−3 increase)Increase, %, (95% CI): 1.28 (0.78 to 1.78) for CHF, 0.86 (−0.6 to 1.79) for peripheral vascular disease, 0.81 (0.30 to 1.32) for stroke, 0.57 (−0.01 to 1.15) for arrhythmia, 0.44 (0.02 to 0.86) for IHD
Zanobetti, 2006 [33]Case cross-overUS, 1985–1999AMI, hospital admissions300 000 AMI pts in 21 citiesDaily PM10 (10 μg m−3 increase)Increase, % (95% CI) 0.65 (0.3 to 1.0), higher in men and in subjects with respiratory disease
Wellenius, 2006 [34]Time seriesUS, 1986–1999CHF, hospital admissions292 918 pts (> 65 years), 7 citiesDaily PM10 (10 μg m−3 increase)Increase, % (95% CI) 0.72 (0.35 to 1.10). No significant effect of age and gender
Lanki, 2006 [39], HEAPSSCase cross-overEurope, 1992–2001AMI, hospital admissions26 958 first AMI in 5 citiesDaily CO (200 μg m−3 increase) and PNC (104 cm−3)**Increase, RR (95% CI): CO 1.005 (1.000 to 1.010), PNC 1.005 (0.996 to 1.015). Effects more pronounced in the warm season
Pope, 2006 [36], IHCSCase cross-overUS, 1994–2004ACS12 865 ACS pts undergoing angiographyDaily PM2.5 (10 μg m−3 increase)Increase, % (95% CI) 4.5 (1.1 to 8.0), larger for those with angiographically demonstrated coronary artery disease
Mustafic, 2012 [41]Meta-analysisNAAMINADaily NO2, SO2, PM2.5PM10 (10 μg m−3 increase), CO (1 mg m−3 increase)††Relative risk (95% CI): CO 1.048 (1.026 to 1.070); NO2 1.011 (1.006 to 1.016); SO2 1.010 (1.003 to 1.017); PM10 1.006 (1.002 to 1.009); PM2.5 1.025 (1.015 to 1.036)
Wellenius, 2005 [43]Case cross-overUS, 1986–1999Stroke, hospital admissions174 817 stroke pts (155 503 ischemic) in 9 citiesSame day PM10, CO, NO2, SO2 (interquartile increase)Ischemic stroke, increase, % (95% CI): PM10 1.03 (0.04 to 2.04); similar increase for the other pollutants. No association for hemorrhagic stroke
Anderson, 2010 [44]Case cross-overDenmark, 2003–2006Stroke, hospital admissions7485 stroke pts (6798 ischemic) in Copenhagen5-day average PM0.1, CO, NO2 (interquartile increase)Ischemic stroke without atrial fibrillation, % (95% CI): PM0.111 (0 to 23), higher for mild strokes (21%, 4 to 41); CO10 (−2 to 24); lower for NO2
Table 3.   Main studies reporting associations of long-term exposure to air pollution with cardiovascular and atherothrombotic disease
Author, years [ref.], Study acronymStudy designCountry, yearsPrimary endpoint(s)Study populationPollutant(s)/assessment*Findings
  1. ACS, American Cancer Society Cancer Prevention Study; AMI, acute myocardial infarction; CAC, coronary artery calcification; CI, confidence interval; CIMT, carotid intima-media thickness; CT, computed tomography; CTS, California Teachers Study; CV, cardiovascular; HNR, Heinz Nixdorf Recall Study; HR, hazard ratio; HSCS, Harvard Six Cities Study; IHD, ischemic heart disease; MESA, Multi-Ethnic Study of Atherosclerosis; NO2, nitrogen dioxide; OR, odds ratio; PM2.5, particulate matter with aerodynamic diameter between 2.5 and 0.1 μm; PM10, particulate matter with aerodynamic diameter between 10 and 2.5 μm; RR, relative risk. *Only significant associations are reported. Less clear association was found with urban background concentration. The original report of this study [6] provided cumulative data on mortality from cardiopulmonary causes. In this extended follow-up, all-cause mortality showed a 6% increase (95% CI 2–11). Mortality attributable to respiratory disease had relatively weak associations. §No increased risk of mortality for cerebrovascular disease was reported (RR, 0.97; 95% CI, 0.88–1.07). Extended follow-up of the original study [4], which reported all-cause and pooled cardiopulmonary mortality. In this study all-cause mortality showed a 14% increase (95% CI, 7–22). A previous follow-up [44] reported an improved overall mortality in parallel with a reduction of PM2.5 (RR, 0.73, 0.57–0.95), not definitely confirmed in this more recent analysis. ††Interdecile range increases were: PM2.5, 4.2 μg m−3; PM10, 6.73 μg m−3; and residential proximity to major roads, 1939 m.

(A) Cardiovascular mortality and morbidity
Hoek, 2002 [52]ProspectiveThe Netherlands, 1986–1994Mortality4492 Dutch subjectsResidential proximity to major roadsRR (95% CI), cardiopulmonary deaths: 1.95 (1.09 to 3.52) and, less consistently, with the estimated ambient background concentration (1.34, 0.68 to 2.64)
Pope, 2004 [50], ACS-CPS IIProspectiveUS, 1984–1998Mortality∼500 000 adults in all USPM2.5 (10 μg m−3 increase)RR (95% CI): all CV causes 1.12 (1.06 to 1.15), IHD 1.18 (1.14 to 1.23). Higher in current than former smokers
Jerrett, 2005 [51]ProspectiveUS, 1982–2000Mortality22 905 residents in the Los Angeles areaPM2.5 (10 μg m−3 increase)RR: cardiopulmonary causes 1.20, IHD 1.49
Toren, 2007 [54]ProspectiveSweden, 1971–2002IHD and cerebrovascular disease mortality176 309 PM-exposed and 71 778 unexposed construction workersAny PM exposure, inorganic dust, fumes, diesel exhaustRR (95% CI), IHD mortality: any PM 1.13 (1.07 to 1.19), inorganic dust 1.07 (1.03 to 1.12), fumes 1.05 (1.00 to 1.10), diesel exhaust 1.18 (1.13 to 1.24)§
Miller, 2007 [53], WHIProspectiveUS, 1994–1998Mortality, vascular events65 893 post-menopausal women in 36 citiesPM2.5 (10 μg m−3 increase)HR (95%CI): Mortality: any CV 1.76 (1.25 to 2.47), definite IHD 2.21 (1.17 to 4.16), cerebrovascular disease 1.83 (1.11 to 3.00); CV events: any 1.24 (1.09 to 1.41), IHD 1.21 (1.04 to 1.42), cerebrovascular 1.35 (1.08 to 1.68)
Maheswaran, 2010 [57]RetrospectiveUK, 1995–2006Mortality after stroke3320 survivors after stroke in south LondonNO2, PM10 (10 μg m−3 increase)HR (95% CI): PM10 1.52 (1.06 to 1.18), NO21.28 (1.11 to 1.48)
Lipsett, 2011 [55], CTSProspectiveUS, 1996–2005Mortality, IHD, stroke∼100 000 female school professionalsPM2.5, PM10 (10 μg m−3 increase)RR (95% CI): PM2.5, IHD mortality 1.20 (1.02 to 1.41), incident stroke 1.19 (1.02 to 1.38); PM10, IHD mortality 1.06 (0.99 to 1.14), incident stroke 1.06 (1.00 to 1.13)
Lepeule, 2012 [49], HSCSProspectiveUS, 1974–2009Mortality∼8000 white residents in 6 citiesPM2.5 (10 μg m−3 increase)RR (95% CI): all CV causes 1.26 (1.14 to 1.40)
Rosenbloom, 2012 [56], Onset StudyCase cross-overUS, 1989–1996Mortality after AMI3886 AMI survivors in 64 centersResidential proximity to major roadwayHR (95% CI): living ≤ 100 m 1.27 (1.01 to 1.60), 100 to ≤ 200 m 1.19 (0.93 to 1.60), 200 to ≤ 1000 m 1.13 (0.99 to 1.30) vs. living > 1000 m (P trend = 0.016)
(B) Subclinical atherosclerosis
Iannuzzi, 2010 [64]CohortItalyArterial stiffness, ultrasonography52 childrenResidential proximity to the major roadSignificantly higher in children living < 330 m and 330 to 730 m vs. > 730 m. No differences in CIMT
Kunzli, 2005 [58]CohortUSCIMT, ultrasonography798 adults living in the Los Angeles areaPM2.5 (10 μg m−3 increase)Increase, % (95% CI): 5.9 (1 to 11), larger in women, > 60 years, never smokers and those on lipid lowering drugs
Kunzli, 2010 [59]Meta-analysisUSCIMT, ultrasonographySubjects living in the Los Angeles areaResidential proximity to major roadsAccelerated progression in subjects living < 100 m from a highway (5.5 μm year−1, 95% CI 0.13 to 10.79), more than twice the population mean
Bauer, 2010 [60], HNRCohortGermanyCIMT, ultrasonography3380 subjects (45–75 years)PM2.5, PM10, residential proximity to major roads, interdecile increase††Increase, % (95% CI): PM2.5 4.3 (1.9 to 6.7), PM10 1.7 (−0.7 to 4.1), residential proximity 1.2 (−0.2 to 2.6)
Hoffman, 2007 [61], HNRCohortGermanyCAC, electron-beam CT4494 subjects (45–75 years)Residential proximity to major roadsOR (95% CI), CAC > 75th centile: 1.63 (1.14 to 2.33), 1.34 (1.00 1.79), living < 50 m, 51 to 100 m vs. > 200 m. 7% (0.1 to 14.4) increase for reduction of distance by half
Lambrechtsen, 2012 [62], DanRiskCohortDenmarkCAC, electron-beam CT1225 adults (20% living in city centers)Urban exposureOR (95% CI), CAC presence: 1.8 (1.3 to 2.4), living in city centers vs. outside
Allen, 2009 [63], MESACohortUSAbdominal aorta calcification, CT1147 adults in 5 metropolitan areasPM2.5 (10 μg m−3 increase)RR (95% CI) 1.06 (0.96 to 1.16); no difference in the quantity of calcification (Agatston score)
Table 4.   Studies addressing the association of air pollution with venous thromboembolism
Author, years [ref.], Study acronymStudy designCountry, yearsPrimary endpoint(s)Study populationPollutant(s)/assessmentFindings
  1. ARIC, Atherosclerosis Risk in Communities Study; DVT, deep-vein thrombosis; NO2, nitrogen dioxide; O3, ozone; OR, odds ratio; PM2.5, particulate matter with aerodynamic diameter between 2.5 and 0.1 μm; PM10, particulate matter with aerodynamic diameter between 10 and 2.5 μm; RR, relative risk; SO2, sulphur dioxide; VTE, venous thromboembolism; WHI, Women’s Health Initiative. *Specific concentration increase considered: 58.4 p.p.b. increase in ozone; 5.85 p.p.b. increase in SO2; 29.25 μg m−3 increase in NO2; 20.02 μg m−3 increase in PM2.5. No association with PM2.5 was reported.

(A) Short-term exposure
Dales, 2010 [46]Time seriesChile, 2001–2005VTE, hospital admissions∼5.4 million residents in SantiagoDaily O3, NO2, SO2, PM2.5 (concentration increase)*RR (95% CI): O3 1.07 (1.05 to 1.09); SO2 1.06 (1.02 to 1.09); NO2 1.08 (1.03 to 1.12); PM2.5 1.05 (1.03 to 1.06). Similar association with pulmonary embolism
Martinelli, 2012 [47]Case cross-overItaly, 2007–2009VTE, hospital admissions302 VTE patients in VeronaDaily PM10 (inter-quartile increase)OR (95% CI) 1.69 (1.13 to 2.53)
(B) Long-term exposure
Baccarelli, 2008 [24]Case–controlItaly, 1995–2005Deep-vein thrombosis870 DVT patients and 1210 controlsPM10 (10 μg m−3 increase)OR (95% CI) 1.70 (1.30 to 2.23). Weaker risk in women, particularly those on hormonal treatments
Baccarelli, 2009 [65]Case–controlItaly, 1995–2005Deep-vein thrombosis663 DVT patients and 859 controls from cities > 15 000 inhabitantsResidential proximity to major roadwayOR (95% CI) 1.47 (1.10 to 1.96), living near (10th centile, 3 m) vs. farther (90th centile, 245 m), not modified after adjusting for PM10 area levels
Shih, 2011 [67], WHIProspectiveUS, 1993–2004Deep-vein thrombosis26 450 post-menopausal women in 40 citiesPM2.5, PM10 (10 μg m−3 increase)HR (95% CI): PM10 0.93 (0.54 to 1.60), PM2.5 1.05 (0.72 to 1.53)
Kan, 2011 [66], ARICProspectiveUS, 1987–2005VTE13 123 subjects (45–64 years) in 4 areasResidential proximity to major roads, traffic densityHR (95% CI): 1.16 (95% CI 0.95 to 1.42), living < 150 m vs. > 150 m; 1.18 (0.88 to 1.57), 0.99 (0.74 to 1.34), 1.14 (0.86 to 1.51), quartiles of traffic density vs. lowest quartile

Short-term studies

Cardiovascular mortality  Two large population studies carried out in the United States (US) and in Europe consistently reported a strong association between the exposure to acute peaks of air pollution and cardiovascular deaths, by observing the daily concentrations of pollutants and mortality. Data from 50 million people living in 20 of the largest cities and metropolitan areas in the US (the National Morbidity, Mortality and Air Pollution Study, NMMAPS), revealed that total mortality rates were independently associated with fine PM concentrations and, particularly, that each 10 μg m−3 rise in PM10 on the day before death resulted in a 0.68% increase of cardiopulmonary mortality [25]. The Air Pollution and Health European Approach (APHEA-2) study, evaluating 43 million people living in 29 major European cities, showed an even stronger association, with an estimated increase in cardiovascular deaths of 0.76% for each 10 μg m−3 rise in PM10, higher than that calculated (0.58%) for mortality from respiratory disease [26]. Consistent data were reported by a recent update of the Metanalysis of Italian studies (MISA), also focusing on the role of other pollutants (NO2, CO and SO2), associated with increases of cardiovascular mortality ranging from 0.4% to 1.1% [27], and by the 6-year observation in nine French cities [28].

Coronary artery disease and cardiac morbidity  Other studies reported the short-term effects of air pollution in terms of increase of hospitalizations for cardiovascular causes or specific cardiovascular events, including atherothrombotic manifestations (Table 2B). A meta-analysis of studies providing data on hospital admissions for cardiovascular causes, documented that each 10 μg m−3 rise in PM10 was associated with an increase of hospitalizations for ischemic heart disease (0.7%) and for congestive heart failure (0.8%) [29]. Higher figures (increase up to 1.6% due to heart disease) were more recently reported from hospitalization data in 14 Spanish cities between 1995 and 1999 [30]. Interestingly, a sub-project of the European APHEA-2 study evaluating data from eight cities, showed that the increase in admission rates for cardiac causes was more relevant in the elderly population (0.7% in subjects > 65 years vs. 0.5% in the whole study population) [31]. In this respect, in a large US population aged > 65 years, admission rates due to congestive heart failure showed the highest impact, with a 1.28% increase of risk per each 10 μg m−3 elevation of PM2.5 concentrations in the same day [32]. The relevant role of PM in exacerbation of congestive heart failure was also shown in a study collecting hospital admission for this disease in seven US cities [33].

Specific data concerning acute myocardial infarction (AMI) were reported by a study on emergency admissions in Boston, showing a significant association of NO2 or PM2.5 with the risk of hospitalization [34]. The same authors previously reported an association with PM10 in a large study from 21 US cities [35]. The increased risk of acute coronary syndromes (unstable angina and AMI) was also shown by the Intermountain Heart Collaborative Study, in which elevation by 10 μg m−3 of PM2.5 was associated with a 4.5% increase of acute events, a larger effect being detectable in subjects with angiographically documented coronary artery disease [36].

The role of traffic-related air pollutants as a trigger for AMI emerged in a German study evaluating the relationship of such an exposure (time spent in cars, or on motorcycles, bicycles or public transport) with the onset of symptoms [37]. The exposure within 1 h was associated with an approximately 3-fold increase of risk (odds ratio, OR, 2.92). Interestingly, cyclists showed a greater risk (OR 3.94) than car drivers (OR 2.60), possibly indicating synergic effects of air pollution with physical activity [37]. Similarly, elevation by 25 μg m−3 of PM2.5 within 2 h from the onset of symptoms resulted in a 48% increased risk of AMI in a US study [38]. Another European study also confirmed this relationship, which was most consistent for fatal events in patients aged > 75 years and in the warm season [39]. Moreover, this study revealed an association of traffic-borne pollutants with the risk of hospital re-admission in individuals with previous AMI [40]. A recent meta-analysis of 34 studies quantified the association of short-term exposure to different air pollutants with AMI, highlighting a significant relationship for CO, NO2, SO2, PM10 and PM2.5, whereas no statistically significant results were obtained for ozone [41].

Cerebrovascular disease  Several studies reported the association of air pollution with an increase of mortality or hospital admissions for stroke. A Finnish study reported an association of mortality due to stroke in the Helsinki area, usually characterized by low levels of air pollution, particularly with PM2.5 concentrations, but also with ultrafine PM and CO. This effect was detectable only in the warm season, suggesting a role for seasonal differences of exposure to the pollution mixture [42]. Increased hospital admissions for stroke were associated with elevation of coarse and fine PM in different US studies [29,32,43], and more recently also with ultrafine PM in a study from Denmark [44]. Interestingly, Low et al. [45] reported an association of asthma exacerbators (i.e. warmer and drier air, grass pollen, upper respiratory infections, SO2 and PM10) with stroke admissions in the US, supporting the role of pulmonary inflammation as a link between atherothrombosis and air pollution.

Venous thromboembolism  The association between increased exposure to air pollution and VTE has been more recently studied and, therefore, is presently less established than the recognized relationship with atherothrombotic manifestations (Table 4A). A study conducted in Santiago, Chile, between 2001 and 2005 first reported a positive association of hospital admissions for venous thrombosis and/or pulmonary embolism with elevations in the concentrations of gaseous pollutants (ozone, SO2 and NO2) and fine PM [46]. Recent data from an Italian study confirmed that an increase of admissions to the emergency department for VTE was associated with elevations of coarse but not fine PM, after adjustment for other atmospheric parameters [47].

Long-term studies

Cardiovascular mortality and morbidity  The Harvard Six Cities study was the first study documenting the association of increased general and cardiopulmonary mortality with air pollution from the 14 to 16 year-survival analysis of a population of about 8000 US citizens (1.26-fold and 1.37% higher in the most polluted than in the least polluted area) [4]. Two analyses from the extended follow-up of this study clarified the specific increase of cardiovascular mortality, reporting relative risks (RRs) of 1.28 and 1.26 for every 10 μg m−3 elevation of PM2.5 concentrations, respectively [48,49]. Interestingly, comparing data from the period 1990 to 1998 with a previous phase of the study (1980–1985), the decrease of mean concentration of fine PM was associated with a reduction of all-cause (RR 0.73, 95% CI 0.57–0.95) and cardiovascular (0.69, 0.46–1.01) mortality [48]. However, the most recent analysis extending up to 2009 did not show clear reduction trends, in spite of a substantial drop in the pollutant concentrations [49].

Positive associations between long-term exposure to air pollutants and all-cause and cardiovascular mortality were consistently confirmed in other prospective studies in large cohorts of the general population [50–52] or in specific groups of individuals and/or different conditions of exposure [53–55]. On the whole, increases of cardiovascular mortality ranging between 12% and 76% per elevation by 10 μg m−3 of PM2.5 were documented (Table 3A). The extended 16-year follow-up of the American Cancer Society study revealed that ischemic heart disease was the single largest cause of mortality (23.7%) [50]. Recently, Rosenbloom et al. [56] evaluated the relationship between mortality in US survivors of AMI and major roadway distance. Lower distance (< 100 m) was correlated with a higher 10-year all-cause and cardiovascular mortality than living > 1000 m. In a model fully adjusting for patients’ demographic, social and clinical (risk factors, medication and cardiovascular morbidity) characteristics, this increased risk was estimated as high as 27% and 19% for all-cause and cardiovascular mortality, respectively. Regarding the latter, however, although statistically significant, the exposure groups showed a substantially similar magnitude of mortality rates [56]. A positive association between exposure to ambient air pollution and mortality was also previously reported in a retrospective evaluation of survivors of a first stroke living in south London [57].

Subclinical atherosclerosis  Several reports documented the association of long-term pollution exposure with markers/progression of subclinical atherosclerosis (Table 3B). Carotid intima-media thickness (CIMT) was found to increase by 4% per every 10 μg m−3 rise in PM2.5 in 798 residents of Los Angeles according to a geostatistical model recording local rise of pollution measurements and adjusting for potential confounders [58]. Moreover, pooling data from five randomized studies conducted in the same area, these authors found an accelerated progression of CIMT (5.5 μm year−1) in residents within 100 m of a highway, more than twice the population mean progression [59]. A similar relationship between CIMT and fine PM exposure was reported by a German prospective study [60], which also documented a strong inverse association between distance of residence from main roads and coronary artery calcification (CAC) measured by electron-beam computed tomography [61]. A recent Danish study confirmed increased CAC scores in subjects living in urban areas compared with rural residents [62]. The risk of aortic calcifications was also positively correlated with PM2.5 exposure [63]. Interestingly, some data suggest that the impact of traffic-related pollution on early atherosclerotic markers is detectable since childhood, as shown by the significantly higher carotid arterial stiffness documented in children living closer to major roads [64].

Venous thromboembolism  Two retrospective case–control studies first reported an association of deep vein thrombosis with air pollution in northern Italy, estimating an overall increase of 70% (OR, 1.70; 95% CI, 1.30–2.23) in risk per each 10 μg m−3 elevation in PM10 level during the year before diagnosis [24] and a 47% increase of risk in subjects living closer to major roads [65]. Conflicting results were reported from recent prospective studies (Table 4B), in which no significant association between the risk of VTE and traffic exposure, main road distance or PM concentrations was found in the general population [66] or in a large cohort of post-menopausal women on randomized exposure to hormone therapy [67].

Experimental studies and pathophysiological mechanisms

Although consistently and rigorously achieved by prospective studies, particularly with respect to cardiovascular mortality and morbidity, the available clinical evidence may only describe associations and, by definition, cannot prove the causative role of air pollution [21,22]. However, these studies provided a strong background for formulating pathophysiological hypotheses and exploring potential mechanisms. In this respect, growing data are being collected by controlled exposure studies, which helped to assess the effects of single pollutant or controlled composition mixtures and corroborated observations from ‘real world’, unpredictable exposures. As mentioned above, the complex mechanisms underlying cardiovascular and thrombotic effects of air pollution show a recognized time course, with acute and chronic responses contributing to accelerate vascular disease and trigger thrombotic complications in the short term and long term. These mechanisms are summarized in Fig. 1 and briefly described as follows.

Figure 1.

 Postulated pathophysiologic mechanisms of cardiovascular and thrombotic adverse effects of air pollution. Smaller particles are able to penetrate deeper into the respiratory tract, up to the lung alveoli, whereas PM10 only enter the upper airways. Key events of short-term exposure are the autonomic dysregulation and local lung inflammation, mediating vascular and cardiac dysfunctions. Moreover, a hypercoagulable state and the systemic propagation of inflammation are strictly related to the local lung inflammation. A role for direct effects of ultrafine particulate matter (or its components), able to pass into the bloodstream, has been also proposed. Systemic inflammation and hypercoagulability are the main mechanisms of long-term effects, associated with the acceleration of atherogenesis and triggering of thrombotic complications. While consistent and robust evidence supports the association with atherothrombotic and general cardiovascular disease, fewer and conflicting data are presently available concerning venous thromboembolism.

Short-term effects

The acute increase of cardiovascular morbidity and mortality after exposure to air pollution has been correlated with abnormal vascular and myocardial responses, mainly mediated by the autonomic nervous system. Moreover, a role for the acute lung inflammatory response and hemostatic unbalance has been suggested. These immediate effects result in facilitation of ischemia, impairment of circulatory compensation mechanisms and increased vulnerability to plaque rupture, which, together with a prothrombotic state, may trigger acute atherothrombotic events. In this context, a pro-arrythmogenic state has been also hypothesized as a contributor to the short-term complications of air pollution exposure [8,21,22].

Vascular and endothelial functions  Vasoconstriction of the brachial artery has been documented immediately after exposure to concentrated PM and ozone [68] or to diesel exhaust [69]. The latter also consistently determined an increase of central arterial stiffness evaluated by applanation tonometry [70]. In controlled exposure studies, this vascular response was also associated with increased diastolic and mean arterial pressure, which was related to the concentration of PM (but not to that of ozone), and particularly to the organic carbon fraction of the exposure [68,71,72]. An impaired vasomotor vascular function has been also revealed by abnormalities in flow-mediated dilatation (FMD) of the brachial artery. FMD was impaired 24 h after exposure to concentrated PM [72] and negatively correlated with their ambient concentrations in ‘real-world’ studies [72,73]. Even in this case, the composition of PM may exert different effects, FMD being impaired only on exposure in some urban areas [72], richer in combustion-derived matter. Along this line, an impaired response to vasodilators has been observed after 2-h exposure to diesel exhaust, but not to its gaseous phase, supporting the role of PM in these vascular effects [74,75]. In parallel with an autonomic dysregulation, the role of endothelial mediators of these impaired functions is still debated. Conflicting data are reported regarding the increase of endothelin-1 [76], whereas some findings suggest changes after pollutant exposure in the bioavailability of nitric oxide (NO), a key mediator in vasodilatation, inhibition of platelet activation and regulation of inflammatory response. These changes have been related to another crucial mechanism in vascular disease, the generation of reactive oxygen species (ROS), which react with NO forming peroxynitrites, thus reducing NO availability and beneficial functions [77].

Cardiac functions  A German study reported an association between increased heart rate and concentrations of CO, SO2 and PM, irrespective of other cardiovascular risk factors [78]. Other studies documented that increased exposure to air pollution resulted in an impairment of heart rate variability [79–81]. The latter is recognized as a risk factor for ventricular arrhythmias [82] and such complications have been reported to increase in parallel with concentrations of ozone and PM2.5 in the previous 24 h [83]. These findings are consistent with an autonomic dysfunction and the effects on heart rate, blood pressure and vascular reactivity are thought to play an important role in the short-term increase of cardiovascular mortality related to air pollution [22]. Although an activation of pulmonary reflex arcs by inhaled PM is hypothesized, a direct effect of pollutants on cardiac ion channels has been also suggested [72].

Other studies highlight an increased risk of electrocardiographic ischemic abnormalities associated with PM exposure, particularly in elderly subjects or in those with coronary artery disease [84–86]. Interestingly, these as well as the other effects on heart rate and blood pressure were reduced in subjects in a highly polluted area when protected from inhaling PM by the use of efficient facemasks [86].

Hemostatic unbalance  Ethical concerns hamper the conduction of human studies on thrombogenic effects of air pollution, therefore most information comes from animal models and ‘real life’ evaluations of components of the coagulation system [21,22]. Animal studies showed the role of PM in activating platelets and coagulation factors. The intratracheal instillation of PM from diesel exhaust in hamsters enhanced thrombus formation after endothelial injury through mechanisms involving platelet activation [87]. This effect, lasting up to 24 h, was reduced by pretreatment with sodium cromoglycate, a stabilizer of basophil and mast cell degranulation [88]. These findings suggest mechanisms linking PM-related pulmonary inflammation and thrombosis, through the release of histamine and other cytokines, the propagation of systemic inflammation and platelet activation. However, as the increase of histamine and the attenuation of thrombogenic effects by its antagonists are detectable only 6 and 24 h after exposure, the earliest prothrombotic mechanisms are likely to be triggered by direct effects of particles (or their components) reaching the circulation [88,89]. The prothrombotic effects of particle-induced lung inflammation were further explored in the same animal model after exposure to engineered nanoparticles, carbon nanotubes. Platelet-leukocyte aggregates and microvescicular tissue factor (TF) activity were detected 6 and 24 h after pulmonary instillation [90]. The abrogation of these effects by P-selectin neutralization supports a crucial role for this adhesion molecule in the particle-mediated rapid activation of platelets and their adhesion to circulating leukocytes [90]. An increase of soluble P-selectin levels and of platelet-monocyte aggregates was also shown in studies on healthy volunteers exposed to diesel exhaust. Enhanced thrombus formation was detected in both low- and high-shear conditions using whole blood in the Badimon chamber [91].

Several studies in humans evaluated the systemic detection of hemostatic abnormalities, suggesting a hypercoagulable state. A relevant increase of plasma viscosity was reported during an acute peak of air pollution in Germany [92]. Consistent with this, plasma levels of fibrinogen, an acute-phase inflammatory protein and a major determinant of plasma viscosity, were increased in healthy volunteers exposed to concentrated PM [93] and correlated with the concentrations of ambient pollutants on the previous day in London, particularly in the warm season [94]. This finding was not confirmed in an Italian study, in which, moreover, an inverse association with the shortening of prothrombin time was reported as a possible marker of hypercoagulability [95]. These data were more recently extended in a study on professional exposure to high PM concentrations (steel production workers), showing that higher exposure levels were associated with shorter prothrombin time, increased thrombin generation measured by endogenous thrombin potential (ETP) and higher tissue-type plasminogen activator (t-PA) [96]. The concomitant higher C-reactive protein (CRP) highlighted the link between inflammation and hypercoagulability. However, no significant change was detected comparing measurements at the baseline and after the 4-day work exposure, possibly indicating that PM-related inflammation and activation of coagulation play a role in long-term effects more than in the short-term exposure.

An impairment of endogenous fibrinolysis, measured by the endothelial release of t-PA, has been shown in studies of exposure to diesel exhaust of healthy volunteers or patients with coronary artery disease [74,85]. This abnormality, detectable at 6 h but not in earlier or later assessments, might contribute to the prothrombotic unbalance on acute pollution exposure. Another potential prothrombotic mechanism is represented by increased circulating microvescicles. These vescicles (mean diameter < 1 μm), released from stimulated or apoptotic cells in the vascular bed, carry negatively charged phospholipids and TF on their membranes, thus creating a procoagulant surface on which coagulation factors can bind and be activated. Elevated numbers of circulating microvesicles have been demonstrated in patients with VTE [97]. Recently, an association of high levels of circulating microvescicles with current or recent past PM exposure has been reported in non-smoker diabetic patients [98]. Interestingly, 1-week exposure was also associated with elevated CRP, leukocytes, fibrinogen and TF-dependent thrombin generation. Moreover, the mean PM10 exposure over 1 year showed the highest effect on microvescicle number and procoagulant potential. These data further support the link between inflammation and hypercoagulability in this setting, and that such effects may contribute particularly to the long-term prothrombotic mechanisms of air pollution, even regarding the debated association with VTE [98].

Long-term effects

As mentioned above, a major role in late vascular effects of air pollution is likely to be attributable to systemic inflammation and its links with atherogenesis and activation of the coagulation system.

Systemic inflammation  A series of studies showing an association between markers of systemic inflammation and pollutant exposure have been mentioned above [72,93–96,98]. This inflammatory response is likely to be the consequence of prolonged stimulation by mediators released on local inflammation within the lungs [99]. In this respect, ambient or concentrated PM exposure was associated with increases in cytokines such as interleukin (IL)-1β, IL-6, tumor necrosis factor-α and granulocyte-macrophage colony-stimulating factor [100,101] and with ROS generation via NADPH oxidase and toll-like receptor 4 mechanisms [102,103], which in turn may enhance the inflammatory response. Moreover, the secretion of adhesion molecules (like P-selectin) by endothelial cells in the inflamed alveoli results in binding and activation of leukocyte and platelets, generation of platelet-leukocyte aggregates and TF-bearing microvescicles, all potential prothrombotic mechanisms [90,91,98]. As an alternative/concurrent mechanism, inhaled ultrafine particles can cross the lung–blood barrier through the interstitium or cell (macrophage, endothelial) endocytosis, directly interact with vascular endothelium and cause oxidative and proinflammatory effects [19,90].

Atherogenesis  Oxidative stress and vascular inflammation are the key mechanisms by which air pollution may promote atherogenesis. This concept is supported by several animal studies. In Apolipoprotein E knock-out (ApoE−/−) mice on atherogenic diet, 6 months exposure to concentrated PM2.5 resulted in a 1.5-fold increase of aortic atheroma and evidence of macrophage infiltration and oxidative stress, as revealed by the increased expression of inducible NO-synthase and ROS generation [104]. The exposure to diesel exhaust for 7 weeks of the same animals was associated with significant changes in the plaque characteristics, with a relevant increase of lipid content, cellularity, numbers of foam cells and smooth muscle cells, and markers of oxidative stress. Interestingly, the latter were also correlated with the magnitude of exposure [105]. These changes, typically observed in unstable vulnerable plaques, have been also detected in coronary lesions of PM-exposed Watanabe hyperlipidemic rabbits, another animal model of atherosclerosis [106]. These data are consistent with the recognized implication of oxidative stress in the generation of oxidized low-density lipoproteins (oxLDL), which stimulate atherogenesis through lipid accumulation by macrophages in the vascular wall. Increased concentrations of oxLDL have been documented in ApoE−/− mice exposed to vehicle emissions [107]. Little evidence from human studies of oxidative stress is available, because of methodological problems in measuring ROS or other markers. However, a meta-analysis of studies reporting associations between environmental exposure to PM and oxidative damage to DNA and lipids provided reliable evidence of such relationships by measuring markers in urine, blood and breath, ruling out the potential for errors in the exposure assessment and in the analysis of each biomarker [108].


Air pollution increasingly emerged as an insidious, universal cardiovascular and thrombotic risk factor. The increase of cardiovascular mortality and morbidity due to air pollution exposure is unequivocally supported by a large body of clinical data, which also highlight the most harmful effects of fine and ultrafine PM. Consistent evidence is available for atherothrombtic disease, including coronary and cerebrovascular events, which are influenced in the short term by acute cardiovascular effects and prothrombotic mechanisms, whereas in the long term the acceleration of atherogenesis is involved. The individual risk conferred by air pollution is poorly predictable; however, a greater impact is documented in vulnerable subjects, such as elderly persons and patients with previous cardiovascular disease. More recent data also suggest a possible association with venous thromboembolism, although conflicting results have been recently reported for long-term effects of air pollution as a risk factor in this setting.

Although yet to be completely elucidated, a series of cardiovascular and prothrombotic effects of air pollution have been hypothesized and tested in experimental studies. However, the effects on the coagulation system and thrombogenic mechanisms have been poorly investigated in human studies. Moreover, the complexity and variability of the composition of air pollution further complicate the comprehension of its full impact on human health. As an example, studies concerning ultrafine PM are still limited.

Recent data show that among triggers of AMI, exposure to traffic exhaust-derived PM had overall the greatest population effect [10]. The recognition of this elusive threat to public health should lead to further improvement in monitoring policies and environmental health legislation, air pollution being almost completely anthropogenic and, therefore, a modifiable risk factor. The WHO suggested a target for PM concentrations below 20 μg mm−3; however, in spite of relevant efforts, this objective is still not reached in the air of many cities in Western countries. Even higher is the threat in the metropolitan areas of the largest cities in many middle-income countries in Asia, Africa, Central and Southern America and Eastern Europe, where presently the highest PM concentrations are reported and environmental health policies are poorly developed [20]. The decrease of PM exposure is expected to be associated with an increase in life expectancy and reduction of cardiopulmonary morbidity, although available data are still not sufficient to determine a definite trend and magnitude of effects [48,49].

Beyond strategies for public health, the recognition of PM exposure in the frame of the individual profile of cardiovascular risk and providing specific advice for subjects at risk, as stated by the recent recommendations from the American Heart Association [7], should be more definitely implemented in our clinical practice in standardized approaches for primary and secondary cardiovascular prevention.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.