Table 1.
Cardiovascular effectsa associated with personal and ambient air pollution exposure: selected studies.
| Studies | Design and population | Outcomes | Findings for PM mass and components | Findings for gases |
|---|---|---|---|---|
| Cohort and cross-sectional studies | ||||
| Dockery et al. 1993 | Cohort study examining ambient air pollution exposure and mortality in 8,111 adults in six U.S. cities with 14–16 years of follow-up | Cardiopulmonary mortality | Compared with the least polluted city, the most polluted city had an adjusted RR for cardiopulmonary mortality of 1.37 (95% CI, 1.11–1.68) | No association with O3, but SO2 and NO2 tracked between-city trends in PM concentrations |
| Pope et al. 2004a | Cohort study examining ambient PM exposure and cardiovascular mortality in 319,000–500,000 persons in the Cancer Prevention Study II, with 16 years of follow-up across U.S. urban areas | Cardiovascular mortality: ischemic heart disease, dysrhythmias, heart failure, and cardiac arrest | A 10-μg/m3 increase in PM2.5 was associated with 8–18% increases in mortality due to ischemic heart disease, dysrhythmias, heart failure, and cardiac arrest | Not assessed |
| Abbey et al. 1999 | Cohort study examining ambient PM10 exposure, total suspended sulfates, SO2, O3, and NO2 in relation to mortality in 6,338 non-smoking California Seventh-Day Adventists with 19 years of follow-up | Cardiopulmonary mortality | No associations | No associations |
| Hoek et al. 2002 | Cohort study examining ambient traffic-related air pollutant exposure (black smoke, NO2) and cause-specific mortality in 5,000 persons with 8 years of follow-up in the Netherlands Cohort Study on Diet and Cancer | Cardiopulmonary mortality | Cardiopulmonary mortality was associated with living near high traffic density (100 m to freeway or 50 m to major urban road) adjusted RR = 1.95 (95% CI, 1.09–3.52) and was associated with an increase of 10 μg/m3 black smoke from background (central sites) plus local sources (street proximity), RR = 1.71 (95% CI, 1.10–2.67) | Cardiopulmonary mortality was associated with an increase of 30 μg/m3 background plus local NO2, RR 1.81 (95% CI, 0.98–3.34) |
| Künzli et al. 2004 | Cross-sectional study on the relationship between ambient PM2.5 and CIMT, using baseline data from two clinical trials in Los Angeles; annual mean PM2.5 exposure was estimated using data from 23 monitoring stations linked to home addresses with geostatistical models | CIMT | For each increase of annual mean 10 μg/m3 PM2.5, CIMT increased by 5.9% (95% CI, 1–11%); adjustment for age reduced the coefficients, but further adjustment for covariates indicated robust estimates in the range of 3.9–4.3% | Estimates for O3 linked to ZIP code centroids were positive in relation to CIMT but not significant and smaller than PM2.5 |
| Cardiac ischemia and related outcomes | ||||
| Pekkanen et al. 2002 | Panel study examining ambient PM, NO2, and CO exposure and ischemia during 342 submaximal exercise tests in 45 subjects with CHD in Helsinki, Finland | ECG ST segment depression > 0.1 mV | Increased risk for ST depression (72 events) was associated with a change of lag-2 1,000 particles/cm3 NC0.1–1, OR = 3.29 (95% CI, 1.57–6.92), 10 μg/m3 PM2.5, OR = 2.84 (95% CI, 1.42–5.66), and 10,000 UFP/cm3 NC0.01–0.1, OR = 3.14 (95% CI, 1.56–6.32); UFPs were independent of PM2.5 | NO2 and CO were also associated with an increased risk for ST depression. |
| de Hartog et al. 2003 | Panel study examining ambient exposure to PM and NO2, SO2, and CO in relation to HRV and BP in 131 subjects with CHD in Helsinki, Finland; Amsterdam, the Netherlands; and Erfurt, Germany | Cardiorespiratory symptoms: chest pain, shortness of breath, avoidance of activities | A 10-μg/m3 increase in PM2.5 associated with shortness of breath, OR = 1.12 (95% CI, 1.02–1.24) and avoidance of activities, OR = 1.10 (95% CI, 1.01–1.19) | Not assessed |
| Peters et al. 2004 | Case-crossover study examining ambient traffic-related air pollution exposure and MI in 691 subjects from the Augsburg Myocardial Infarction Registry who had survived 24 hr postinfarct; time–activity diary data on activities during the 4 days before symptom onset were used to assess traffic exposures | MI | Exposure to traffic was associated with onset of MI 1 hr afterward, OR = 2.92 (95% CI, 2.22–3.83); a significant association was also seen for exposure to traffic 2 hr before onset, and there was evidence for effects up to 6 hr; key exposures influencing overall associations with traffic included times spent in cars and in public transportation; associations changed minimally, adjusting for exercise, and there was no confounding by reports of extreme anger or joy | As with PM, gases were not directly assessed, but traffic exposures involve pollutant gases as well as particles |
| Blood pressure (BP) | ||||
| Linn et al. 1999 | Panel study in Los Angeles, California, examining BP and lung function in 30 subjects with COPD, with only 4 consecutive days of air sampling: personal exposure to PM2.5, indoor and outdoor home PM2.5 and PM10, and ambient PM10, O3, NO2, and CO | BP | Systolic BP increased 0.172 mm Hg for every 1-μg/m3 increase in ambient lag-1 PM10 (p = 0.006); diastolic BP increased 0.095 mm Hg for every 1-μg/m3 increase in PM10 (p = 0.03); outdoor home PM10 was similarly associated with BP, but no significant associations were reported for PM2.5 or any indoor or personal PM measurement | No association of BP with exposure to central site O3, NO2, or CO |
| Brauer et al. 2001 | Panel study examining personal exposure over 7 non-consecutive days to PM2.5 and sulfate, and ambient exposure to PM2.5, PM10 , sulfate, and gaseous pollutants, in relation to BP, HRV, and lung function in 16 COPD patients in Vancouver, Canada | BP, HRV, SVE | Weak associations were observed between particle concentrations and increased SVE and with decreased systolic BP; ambient PM10 had the largest effect on cardiovascular end points and the only statistically significant association (SVE); use of personal exposure measurements did not show a larger or more consistent effect | CO was inversely associated with systolic BP and reduced estimates for ambient PM |
| Ibald-Mulli et al. 2001 | Retrospective analysis examining the relationship between ambient air pollution exposure (TSP, SO2, and CO) and BP in 2,607 men and women 25–64 years of age from a general population survey in Augsburg, Germany | Systolic BP | A 90-μg/m3 increase in TSP was associated with an increase in systolic BP of 1.79 mm Hg (95% CI, 0.63–2.95); in subgroups with high plasma viscosity levels or increased HR, systolic BP increased by 6.93 mm Hg (95% CI, 4.31–9.75) and 7.76 mm Hg (95% CI, 5.70–9.82) in association with TSP, respectively | An 80-μg/m3 increase in SO2 was associated with an increase in systolic BP of 0.74 mm Hg (95% CI, 0.08–1.40) |
| Ibald-Mulli et al. 2004 | Panel study examining ambient exposure to PM and NO2, SO2, and CO in relation to HRV and BP in 131 subjects with CHD in Helsinki, Finland; Amsterdam, the Netherlands and Erfurt, Germany | BP and HR | A small decrease in systolic BP (−0.72 mm Hg; 95% CI, −1.92 to 0.49) and diastolic BP (−0.70 mm Hg; 95% CI, −0.02 to −1.38) was found to be associated with a 5-day average increase of 10,000 UFPs/cm3 (NC0.01–0.1); slightly stronger and more significant associations were found for accumulation mode particle number concentration (NC0.1–1.0), but smaller associations were found for a 10 μg/m3 increase in PM2.5 mass; small decreases in HR were also found for PM exposures | The magnitude and significance of inverse BP associations with CO were similar to those of PM0.1–1.0; a small decrease in HR (−0.40 beats/min; 95% CI, −0.82 to 0.01) was found for an increase of lag-1, 5 μg/m3 SO2 |
| Zanobetti et al. 2004 | Panel study examining ambient PM2.5, O3, NO2, SO2, and CO in relation to BP among 62 patients with preexisting heart disease using data from 631 repeated visits for cardiac rehabilitation in Boston | BP | Increasing from the 10th to the 90th percentile in 5-day mean PM2.5 (10.5 μg/m3) resulted in increases of 2.8 mm Hg (95% CI, 0.1–5.5) in systolic, 2.7 mm Hg (95% CI, 1.2–4.3) in diastolic, and 2.7 mm Hg (95% CI, 1.0–4.5) in mean arterial BP; black carbon was associated with diastolic BP | Diastolic BO was associated with 120-hr average SO2 (3.9% increase; 95% CI, 0.3–76), O3 (2.7% increase; 95% CI, 0.02–5.4) |
| Autonomic control of cardiac rhythm | ||||
| Holguin et al. 2003 | Panel study in Mexico City examining indoor and outdoor nursing home measurements of PM2.5 and ambient exposure to O3, NO2, CO, and SO2 in relation to HRV in 34 elderly residents followed every other day for 3 months; personal PM2.5 was predicted using indoor and outdoor home PM2.5 plus time–activity data | HRV, frequency Domain | A 10-μg/m3 increase in predicted personal PM2.5 was associated with a 5.0% decrease in high-frequency HRV (β = −0.049; 95% CI, −0.090 to −0.007); associations with indoor PM2.5 were stronger than outdoor home PM2.5; among 13 subjects with hypertension, the association with predicted personal PM2.5 was stronger (−7.1%) | O3 was inversely associated with high-and low-frequency HRV among 13 subjects with hypertension (2% decrease per 10 ppb O3), but this association was confounded by PM2.5 |
| Pope et al. 2004b | Panel study of ambient exposure to PM and HRV and blood markers in 88 elderly subjects living in Salt Lake City and Provo/Orem, Utah | HRV | A 100-μg/m3 increase in PM2.5 was associated with a 35 (SE = 8) msec decrease in SDNN and a 42 (SE = 11) msec decrease in r-MSSD | Not assessed |
| Autonomic control of cardiac rhythm | ||||
| Magari et al. 2001, 2002a, 2002b | Panel study examining personal exposure to PM in relation to HRV in 20 (Magari et al. 2002a), 40 (Magari et al. 2001), and 39 (Magari et al. 2002b) healthy boilermakers exposed to welding fumes and residual oil fly ash | HRV | Each 100-μg/m3 increase in 3-hr average PM2.5 (laser photometer light scatter) was associated with a 1.4% (95% CI, −2.1 to −0.6%) decrease in 5-min SDNN in the 20 subjects (Magari et al. 2002a); in the 40 subjects, each 1-mg/m3 increase in 4-hr average PM2.5 was associated with a 2.66% (95% CI, −3.75 to −1.58%) decrease in 5-min SDNN SDNN (Magari et al. 2001); however, in 39 of these 40 subjects, PM2.5 metals on filters, lead and vanadium, were associated with an increase in workday average of the 5-min SDNN (Magari et al. 2002b) | Not assessed |
| Riediker et al. 2004 | Panel study of in-vehicle exposure to PM and HRV and blood markers of inflammation in 9 healthy male North Carolina Highway Patrol troopers | HRV | In-vehicle 10-μg/m3 PM2.5 increase was associated with increased ectopic beats throughout exposure (20%, p = 0.005); PM2.5 was positively associated with heart beat cycle length (6%, p = 0.01) as well as HF HRV and SDNN the next morning after exposure | NO2 and CO were not significant |
| Chan et al. 2004 | Panel study in Taipei, Taiwan, examining personal exposure to sub-micrometer particles and HRV over one 16-hr daytime period in 9 young healthy adults 19–29 years of age (2 females) and 10 older male subjects 42–97 years of age with lung function impairments (FEV1/FVC < 85%) | HRV | Personal exposure to NC0.02–1 was associated with decreased in both time-domain and frequency-domain HRV indices; in young subjects, a 10,000 particles/cm3 increase in the last 1–4 hr average NC0.02–1 was associated with 0.68–1.35% decrease in SDNN, 1.85–2.58% decrease in r-MSSD; in the older panel they found 10,000-particles/cm3 increase in the last 1- to 3-hr average NC0.02–1 was associated 1.72–3.00% decreases in SDNN and 2.72–4.65% decreases in r-MSSD; there were similar associations for high- and low-frequency domain indices | Not assessed |
| Tarkiainen et al. 2003 | Panel study in Kuopio, Finland, examining personal exposure to carbon monoxide and HRV in 6 subjects with CHD followed for three separate 24-hr ambulatory monitoring periods | HRV | Not assessed | r-MSSD increased by 2.4 msec (p = 0.03) with exposure to CO (> 2.7 ppm) |
| Peters et al. 2000 | Panel study of arrhythmias in 100 subjects in eastern Massachusetts with implanted defibrillators (63,628 person-days of follow-up) with ambient measurements of PM mass, black carbon, NO2, CO, O3, and SO2 | Defibrillator discharge interventions for ventricular tachycardias or fibrillation (33 subjects with at least one) | Only 6 subjects with ≥ 10 defibrillator discharges had increased arrhythmias associated with black carbon and PM2.5, which showed a weaker association; both PM metrics were confounded by NO2, but the effect estimate of NO2 was unchanged | 26-ppb increase in NO2 lagged 1 day was associated with increased defibrillator interventions in the full panel (OR = 1.8; 95% CI, 1.1–2.9). Subjects with ≥ 10 defibrillator discharges had increased arrhythmias associated with CO and NO2 across several lags |
| Systemic inflammation and thrombosis | ||||
| Seaton et al. 1999 | Panel study examining 3-day personal exposure estimated (from a one 24-hr personal exposure measurement) and city center ambient exposure to PM10 in relation to hematologic factors in 112 elderly subjects in Belfast and Edinburgh, UK | Hematologic factors: hemoglobin, packed red cells, red blood cell count, platelets, white blood cell count, CRP, fibrinogen, factor VII, IL-6 | An increase of 100 μ/m3 in personal PM10 and ambient PM10 exposure resulted in significant decreased mean percentage changes of ≤ 1% in hemoglobin concentration, packed cell volume, and red blood cell count; only personal PM10 was associated with an 11% decrease in platelets and a 7% decrease in factor VII; CRP increased with ambient PM10 (+147%; 95% CI, 20–477), but not with personal PM (p = 0.73); fibrinogen decreased with ambient PM10 (−9%; 95% CI, −19 to 0) | Not assessed |
| Systemic inflammation and thrombosis | ||||
| Schwartz 2001 | Cross-sectional study examining the relationship between ambient PM10, NO2, SO2, and blood biomarkers using data from a cohort study (NHANES III) | Fibrinogen, and platelet and white blood cell counts | For an interquartile range change in PM10 (26 μg/m3), the relative odds for being above the 90th percentile of fibrinogen was 1.77 (95% CI, 1.26–2.49); platelets, 1.27 ( 95% CI, 0.97–1.67); and white blood cells, 1.64 (95% CI, 1.17–2.30) | SO2 was positively associated with white cell counts, and NO2 with platelet counts and fibrinogen, but both gases were confounded by PM10 |
| Pekkanen et al. 2000 | Cross-sectional study examining the association between ambient PM10, NO2, CO, SO2, O3, and fibrinogen among 7,205 subjects in London at baseline enrollment in a cohort study | Fibrinogen | No association between PM10 and fibrinogen was seen after adjustment for confounders | NO2 increase from the 10th to the 90th percentile was associated with a 1.5% higher fibrinogen concentration (95% CI, 0.4–2.5%); similar increase for CO resulted in 1.5% higher fibrinogen concentration (95% CI, 0.5–2.5%); no association with SO2 or O3 |
| Peters et al. 1997a, 2001b | Cohort study in Augsburg, Germany, examining relationships of ambient TSP, SO2, and CO exposure to CRP in 631 men 45–64 years of age with no history of MI at their baseline assessment; two CRP measurements were 3 years apart | CRP | An increase of 26 μg/m3 (5-day mean) in TSP increased the odds of observing a CRP level above the 80th percentile, OR = 1.31 (95% CI, 1.09–1.56); CRP and plasma viscosity (Peters et al. 1997a) were increased during an air pollution episode in 1985 | An increase of 30 μg/m3 (5-day mean) in SO2 increased the odds of observing a CRP level above the 90th percentile, OR = 1.24 (95% CI, 1.03–1.49) |
| Pope et al. 2004b | Panel study of ambient exposure to PM and HRV and blood markers in 88 elderly subjects living in Salt Lake City and Provo/Orem, Utah | CRP, white blood cell count, whole blood viscosity, granulocytes, lymphocytes, monocytes, basophils, eosinophils, red blood cells, platelets | A 100-μg/m3 increase in PM2.5 was associated with a 0.81 (SE = 0.17) mg/dL increase in CRP; one subject’s data had a strong influence on estimates; there was no association with other outcomes | Not assessed |
| Riediker et al. 2004 | Panel study of in-vehicle exposure to PM and HRV and blood markers of inflammation in 9 healthy male North Carolina Highway Patrol troopers | CRP, plasminogen, von Willebrand factor, lymphocyte count, lymphocytes, neutrophils, hematocrit, red blood cell indices, uric acid | In-vehicle 10-μg/m3 PM2.5 increase was associated with decreased lymphocytes (−11%, p = 0.03), increased red blood cell indices (1%, p = 0.03), neutrophils (6%, p = 0.04), CRP (32%, p = 0.02), and von Willebrand factor (12%, p = 0.02) | NO2 and CO were not significant |
| Systemic inflammation and thrombosis | ||||
| Schwartz 2001 | Cross-sectional study examining the relationship between ambient PM10, NO2, SO2, and blood biomarkers using data from a cohort study (NHANES III) | Fibrinogen, and platelet and white blood cell counts | For an interquartile range change in PM10 (26 μg/m3), the relative odds for being above the 90th percentile of fibrinogen was 1.77 (95% CI, 1.26–2.49); platelets, 1.27 ( 95% CI, 0.97–1.67); and white blood cells, 1.64 (95% CI, 1.17–2.30) | SO2 was positively associated with white cell counts, and NO2 with platelet counts and fibrinogen, but both gases were confounded by PM10 |
| Pekkanen et al. 2000 | Cross-sectional study examining the association between ambient PM10, NO2, CO, SO2, O3, and fibrinogen among 7,205 subjects in London at baseline enrollment in a cohort study | Fibrinogen | No association between PM10 and fibrinogen was seen after adjustment for confounders | NO2 increase from the 10th to the 90th percentile was associated with a 1.5% higher fibrinogen concentration (95% CI, 0.4–2.5%); similar increase for CO resulted in 1.5% higher fibrinogen concentration (95% CI, 0.5–2.5%); no association with SO2 or O3 |
| Peters et al. 1997a, 2001b | Cohort study in Augsburg, Germany, examining relationships of ambient TSP, SO2, and CO exposure to CRP in 631 men 45–64 years of age with no history of MI at their baseline assessment; two CRP measurements were 3 years apart | CRP | An increase of 26 μg/m3 (5-day mean) in TSP increased the odds of observing a CRP level above the 80th percentile, OR = 1.31 (95% CI, 1.09–1.56); CRP and plasma viscosity (Peters et al. 1997a) were increased during an air pollution episode in 1985 | An increase of 30 μg/m3 (5-day mean) in SO2 increased the odds of observing a CRP level above the 90th percentile, OR = 1.24 (95% CI, 1.03–1.49) |
| Pope et al. 2004b | Panel study of ambient exposure to PM and HRV and blood markers in 88 elderly subjects living in Salt Lake City and Provo/Orem, Utah | CRP, white blood cell count, whole blood viscosity, granulocytes, lymphocytes, monocytes, basophils, eosinophils, red blood cells, platelets | A 100-μg/m3 increase in PM2.5 was associated with a 0.81 (SE = 0.17) mg/dL increase in CRP; one subject’s data had a strong influence on estimates; there was no association with other outcomes | Not assessed |
| Riediker et al. 2004 | Panel study of in-vehicle exposure to PM and HRV and blood markers of inflammation in 9 healthy male North Carolina Highway Patrol troopers | CRP, plasminogen, von Willebrand factor, lymphocyte count, lymphocytes, neutrophils, hematocrit, red blood cell indices, uric acid | In-vehicle 10-μg/m3 PM2.5 increase was associated with decreased lymphocytes (−11%, p = 0.03), increased red blood cell indices (1%, p = 0.03), neutrophils (6%, p = 0.04), CRP (32%, p = 0.02), and von Willebrand factor (12%, p = 0.02) | NO2 and CO were not significant |
Abbreviations: FEV1/FVC, forced expiratory volume in 1 sec/forced vital capacity; HF, high frequency; RR, relative risk; SVE, supraventricular ectopic heartbeat.
a
The focus is on cardiovascular outcomes. Although some studies may have examined other outcomes, they are not reported.