Secondary Logo

Journal Logo

Long-term Exposure to Air Pollution and Cardiovascular Mortality: An Analysis of 22 European Cohorts

Beelen, Roba; Stafoggia, Massimob; Raaschou-Nielsen, Olec; Andersen, Zorana Jovanovicc,d; Xun, Wei W.e,f; Katsouyanni, Kleag; Dimakopoulou, Konstantinag; Brunekreef, Berta,h; Weinmayr, Gudruni,j; Hoffmann, Barbaraj; Wolf, Kathrink; Samoli, Evangeliag; Houthuijs, Dannyl; Nieuwenhuijsen, Markm,n; Oudin, Annao; Forsberg, Bertilo; Olsson, Davido; Salomaa, Veikkop; Lanki, Timop; Yli-Tuomi, Tarjap; Oftedal, Benteq; Aamodt, Geirq; Nafstad, Perq,r; De Faire, Ulfs; Pedersen, Nancy L.t; Östenson, Claes-Göranu; Fratiglioni, Laurav; Penell, Johannas; Korek, Michals; Pyko, Andreis; Eriksen, Kirsten Thorupc; Tjønneland, Annec; Becker, Thomasw; Eeftens, Marloesa; Bots, Michielh; Meliefste, Keesa; Wang, Menga; Bueno-de-Mesquita, Basl; Sugiri, Dorotheaj; Krämer, Ursulaj; Heinrich, Joachimx; de Hoogh, Keese; Key, Timothyy; Peters, Annettek; Cyrys, Josefk,z; Concin, Hansaa; Nagel, Gabrielei,aa; Ineichen, Alexbb,cc; Schaffner, Emmanuelbb,cc; Probst-Hensch, Nicolebb,cc; Dratva, Juliabb,cc; Ducret-Stich, Reginabb,cc; Vilier, Alicedd,ee,ff; Clavel-Chapelon, Françoisedd,ee,ff; Stempfelet, Morganegg; Grioni, Sarahh; Krogh, Vittoriohh; Tsai, Ming-Yibb,cc,ii; Marcon, Alessandrojj; Ricceri, Fulviokk; Sacerdote, Carlottall; Galassi, Claudiall; Migliore, Enricall; Ranzi, Andreamm; Cesaroni, Giuliab; Badaloni, Chiarab; Forastiere, Francescob; Tamayo, Ibonn,nn; Amiano, Pilarn,nn; Dorronsoro, Mirenn,nn; Katsoulis, Michailoo; Trichopoulou, Antoniaoo; Vineis, Paoloe; Hoek, Gerarda

doi: 10.1097/EDE.0000000000000076
Air Pollution
Free
SDC

Background: Air pollution has been associated with cardiovascular mortality, but it remains unclear as to whether specific pollutants are related to specific cardiovascular causes of death. Within the multicenter European Study of Cohorts for Air Pollution Effects (ESCAPE), we investigated the associations of long-term exposure to several air pollutants with all cardiovascular disease (CVD) mortality, as well as with specific cardiovascular causes of death.

Methods: Data from 22 European cohort studies were used. Using a standardized protocol, study area–specific air pollution exposure at the residential address was characterized as annual average concentrations of the following: nitrogen oxides (NO2 and NOx); particles with diameters of less than 2.5 μm (PM2.5), less than 10 μm (PM10), and 10 μm to 2.5 μm (PMcoarse); PM2.5 absorbance estimated by land-use regression models; and traffic indicators. We applied cohort-specific Cox proportional hazards models using a standardized protocol. Random-effects meta-analysis was used to obtain pooled effect estimates.

Results: The total study population consisted of 367,383 participants, with 9994 deaths from CVD (including 4,992 from ischemic heart disease, 2264 from myocardial infarction, and 2484 from cerebrovascular disease). All hazard ratios were approximately 1.0, except for particle mass and cerebrovascular disease mortality; for PM2.5, the hazard ratio was 1.21 (95% confidence interval = 0.87–1.69) per 5 μg/m3 and for PM10, 1.22 (0.91–1.63) per 10 μg/m3.

Conclusion: In a joint analysis of data from 22 European cohorts, most hazard ratios for the association of air pollutants with mortality from overall CVD and with specific CVDs were approximately 1.0, with the exception of particulate mass and cerebrovascular disease mortality for which there was suggestive evidence for an association.

Supplemental Digital Content is available in the text.

From the aInstitute for Risk Assessment Sciences, Utrecht University, Utrecht, The Netherlands; bDepartment of Epidemiology, Lazio Regional Health Service, Rome, Italy; cDanish Cancer Society Research Center, Copenhagen, Denmark; dCenter for Epidemiology and Screening, Department of Public Health, University of Copenhagen, CSS, København K, Denmark; eMRC-HPA Centre for Environment and Health, Department of Epidemiology and Biostatistics, Imperial College London, St Mary’s Campus, London, United Kingdom; fUniversity College London, CeLSIUS, London, United Kingdom; gDepartment of Hygiene, Epidemiology, and Medical Statistics, Medical School, University of Athens, Athens, Greece; hJulius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands; iInstitute of Epidemiology and Medical Biometry, Ulm University, Ulm, Germany; jIUF–Leibniz Research Institute for Environmental Medicine, Düsseldorf, Germany, and Medical Faculty, University of Düsseldorf, Düsseldorf, Germany; kInstitute of Epidemiology II, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Germany; lNational Institute for Public Health and the Environment, Bilthoven, The Netherlands; mCentre for Research in Environmental Epidemiology (CREAL), Barcelona, and Parc de Recerca Biomèdica de Barcelona–PRBB (office 183.05) C. Doctor Aiguader, Barcelona, Spain; nConsortium for Biomedical Research in Epidemiology and Public Health (CIBER en Epidemiología y Salud Pública-CIBERESP), Melchor Fernández Almagro 3-5, Madrid, Spain; oDivision of Occupational and Environmental Medicine, Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden; pNational Institute for Health and Welfare, Kuopio, Finland; qNorwegian Institute of Public Health, Oslo, Norway; rInstitute of Health and Society, University of Oslo, Oslo, Norway; sInstitute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden; tDepartment of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden; uDepartment of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden; vAging Research Center, Department of Neurobiology, Care Sciences, and Society, Karolinska Institutet, Stockholm, Sweden; wDepartment of Environmental Science, Aarhus University, Roskilde, Denmark; xInstitute of Epidemiology I, Helmholtz Zentrum München, German Research Center of Environmental Health, Neuherberg, Germany; yCancer Epidemiology Unit, Nuffield Department of Clinical Medicine, University of Oxford, Oxford, United Kingdom; zEnvironmental Science Center, University of Augsburg, Augsburg, Germany; aaAgency for Preventive and Social Medicine, Bregenz, Austria; bbSwiss Tropical and Public Health Institute, Basel, Switzerland; ccUniversity of Basel, Basel, Switzerland; ddInserm, Centre for Research in Epidemiology and Population Health (CESP), U1018, Nutrition, Hormones and Women’s Health Team, Villejuif, France; eeUniversity Paris Sud, UMRS 1018, Villejuif, France; ff IGR, Villejuif, France; ggFrench Institute for Public Health Surveillance (InVS) 12, Saint-Maurice, France; hhEpidemiology and Prevention Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; iiDepartment of Environmental and Occupational Health Sciences, University of Washington, Seattle, WA; jjUnit of Epidemiology and Medical Statistics, Department of Public Health and Community Medicine, University of Verona, Verona, Italy; kkHuman Genetics Foundation–HuGeF, Turin, Italy; llUnit of Cancer Epidemiology, AO Citta’ della Salute e della Scienza—University of Turin and Center for Cancer Prevention, Turin, Italy; mmEnvironmental Health Reference Centre, Regional Agency for Environmental Prevention of Emilia-Romagna, Modena, Italy; nnPublic Health Division of Gipuzkoa, Basque Government, Gipuzkoa, Spain; and ooHellenic Health Foundation, Athens, Greece.

Supported by the European Community’s Seventh Framework Program (FP7/2007–2011 [grant agreement number: 211250]). For the Finnish part, additional funding received from the Academy of Finland (project number: 129317). For HUBRO, the data collection was conducted as part of the Oslo Health Study 2000–2001 and financed by the Norwegian Institute of Public Health, the University of Oslo, and the Municipality of Oslo. Financial support for the combined work with the Stockholm studies was received from the Swedish Environmental Protection Agency, the Swedish Heart-Lung Foundation, and the Swedish Council for Working Life and Social Research. The Swedish Ministry for Higher Education financially supports the Swedish Twin Register. SALT was supported by the Swedish Council for Working Life and Social Research and a grant from the NIH (grant number: AG-08724). TwinGene was supported by the Swedish Research Council (grant number: M-2005-1112), GenomEUtwin (grant numbers: EU/QLRT-2001-01254, QLG2-CT-2002-01254, NIH DK U01-066134). The Swedish Foundation for Strategic Research (SSF), and the Heart and Lung Foundation (grant number: 20070481). The EPIC-MORGEN and EPIC-PROSPECT cohorts were supported by the Dutch Ministry of Public Health, Welfare and Sports (VWS), Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), and Statistics Netherlands. The baseline study and the mortality follow-up of SALIA were funded by the Ministry of the Environment of North-Rhine-Westfalia (Germany). The KORA research platform and the MONICA Augsburg studies were initiated and financed by the Helmholtz Zentrum München, German Research Center for Environmental Health, which is funded by the German Federal Ministry of Education and Research and by the State of Bavaria. The VHM&PP is supported by the State of Vorarlberg, Austria. SAPALDIA received funds from the The Swiss National Science Foundation (grants numbers: 33CSCO-134276/1, 33CSCO-108796, 3247BO-104283, 3247BO-104288, 3247BO-104284, 3247–065896, 3100–059302, 3200–052720, 3200–042532, 4026–028099), the Federal Office for Forest, Environment, and Landscape and several Federal and Cantonal authorities. SAPALDIA received support in mortality record linkage from the Swiss National Cohort Study (grant numbers: 108806 and 134273).

The authors report no conflicts of interest.

Supplemental digital content is available through direct URL citations in the HTML and PDF versions of this article (www.epidem.com). This content is not peer-reviewed or copy-edited; it is the sole responsibility of the author.

Correspondence: Rob Beelen, Institute for Risk Assessment Sciences, Utrecht University PO Box 80178, 3508 TD Utrecht, The Netherlands. E-mail: r.m.j.beelen@uu.nl.

Cohort studies assessing effects of long-term exposure to air pollution on cardiovascular mortality have generally found increased risks.1–4 Studies in the United States have mostly reported associations for particles with diameters of less than 2.5 μm (PM2.5),5 whereas studies in Europe (including studies in The Netherlands,6 France,7 Norway,2 Denmark,8 United Kingdom,9 and Italy)10 have also reported associations with long-term exposure to nitrogen oxides (NO2 and NOx), which are more related to traffic pollutants than particle mass. Effect estimates differ across studies, with some studies showing little or no association of various air pollutants with all cardiovascular mortality.6,9,11–13 There is therefore an interest in investigating the mortality effects of a range of air pollutants.

The category of cardiovascular diseases (CVDs) is broad, and it is unlikely that the risk associated with air pollution exposure is uniform for the specific cardiovascular mortality causes. However, only a few studies have investigated specific causes of cardiovascular mortality, including ischemic heart disease and cerebrovascular disease.1,2,8,14–16

The aim of the current study was to investigate the effects of long-term exposure to air pollution on all cardiovascular mortality, as well as the more specific causes of ischemic heart disease mortality, myocardial infarction (MI) mortality, and cerebrovascular disease mortality, for a range of air pollution measures. In the framework of the collaborative European Study of Cohorts for Air Pollution Effects (ESCAPE), data from 22 ongoing cohort studies were used, with a standardized exposure assessment of particle mass and nitrogen oxides.

Back to Top | Article Outline

METHODS

The association between air pollution and cardiovascular mortality was analyzed in each cohort separately, following the standardized analysis protocol of the ESCAPE study.17 A standardized statistical protocol and STATA script were used, as explained in a training workshop for all local analysts. Cohort-specific results were sent to the coordinating institute (IRAS, Utrecht University) for central evaluation. We combined cohort-specific effect estimates by random-effects meta-analysis. Pooling of the cohort data was not possible due to data transfer and privacy issues. Random-effects meta-analysis has the advantage of taking into account interarea and intercohort differences not entirely addressed by the available confounders.

Back to Top | Article Outline

Study Populations

Twenty-two ongoing cohorts from 13 countries across Europe were included (Table 1, and eAppendix 1; http://links.lww.com/EDE/A767). All cohorts were included samples from the general population. The study areas of most cohorts consisted of a large city with surrounding smaller rural communities. Some (multicenter) cohorts included large regions of the country, such as EPIC-MORGEN in The Netherlands, SALIA in the Ruhr area in Germany, EPIC-Oxford covering much of the United Kingdom, the VHM&PP cohort in Austria, and SAPALDIA in three cities in Switzerland. The use of cohort data in ESCAPE was approved by the local ethical and data protection authorities. Each cohort study followed the rules for ethics and data protection set up in the country in which it was based.

TABLE 1

TABLE 1

Back to Top | Article Outline

Mortality Outcome Definition

In all cohorts, follow-up was based on linkage to mortality registries. Outcomes were defined on the basis of the underlying cause of death recorded on death certificates: all CVD mortality (International Classification of Diseases [ICD]-9: 400–440; ICD-10: I10-I70), ischemic heart disease mortality (ICD-9: 410–414; ICD-10: I20-I25), MI mortality (ICD-9: 410; ICD-10: I21, I22), and cerebrovascular disease mortality (ICD-9: 430–438; ICD-10: I60-I69) (eAppendix 2; http://links.lww.com/EDE/A767).

Back to Top | Article Outline

Exposure Assessment

Air pollution concentrations at the baseline residential addresses of study participants were estimated by land-use regression models following a standardized procedure described elsewhere.18,19 In brief, air pollution was monitored for 1 year between October 2008 and May 2011 in all study areas to obtain the following annual average concentrations: NO2 and NOx; particles with aerodynamic diameters of less than 2.5 μm (PM2.5) and less than 10 μm (PM10) plus PM2.5 absorbance (determined as the reflectance of PM2.5 filters).20,21 PMcoarse was calculated as PM10 minus PM2.5. PM and NOx were both measured in 19 of the 22 study areas; NOx alone was measured in the remaining three areas. Study area–specific land-use regression models were developed to explain the spatial variation of measured annual average air pollution concentrations within each area using traffic and land-use predictor variables from a Geographic Information System. The results of the land-use regression models were then used to estimate ambient air pollution concentration at the participants’ baseline addresses. In addition to air pollution concentrations, traffic intensity on the nearest road (vehicles per day) and total traffic load (intensity × length) on all major roads within a 100-m buffer were used as indicators of exposure. A detailed description of exposure-assessment procedures, including back-extrapolation of concentrations to the baseline year and fit of land-use regression models, is presented in eAppendix 3 (http://links.lww.com/EDE/A767).

Back to Top | Article Outline

Statistical Analyses

Cohort-specific Analyses

Cox proportional hazards models were used for the cohort-specific analyses. We used age as the time scale because of evidence that this better adjusts for potential confounding by age.22 Censoring occurred at the time of death for non-CVD causes, emigration, loss to follow-up for other reasons, or at end of follow-up, whichever came first. Air pollution exposure was analyzed as a linear variable. Information on potential confounders was available from questionnaires at baseline.

A priori, we specified three confounder models with increasing level of adjustment. Confounder models were decided based on previous cohort studies of air pollution and mortality, as well as availability of data in a majority of the cohorts. Model 1 included only age (time axis), sex, and calendar time (year(s) of enrollment). Model 2 added individual-level variables: smoking status (never/former/current), smoking intensity, smoking duration, environmental tobacco smoke, fruits intake, vegetables intake, alcohol consumption (linear and squared term), body mass index (BMI) (linear and squared term), educational level (low, medium, or high), occupational class (white/blue collar classification), employment status, and marital status. Model 3 added to Model 2 area-level socioeconomic status (SES) variables (mostly mean income of neighborhood or municipality). Model 3 was selected as the main confounder model. Only subjects with complete information for Model 3 variables were included in the analyses.

In sensitivity analyses, we added prevalent hypertension, physical activity, diabetes mellitus, and cholesterol level to Model 3. Extended confounder models were used in sensitivity analyses because some potential effect of air pollution might be mediated by hypertension, diabetes mellitus, and cholesterol level.

We further evaluated the impact of the addition of modeled road traffic noise to Model 3 because noise and air pollution have been shown to be correlated and may both affect CVD mortality. Road traffic noise was modeled at the highest exposed facade at the baseline address (eAppendix 3; http://links.lww.com/EDE/A767). Noise was used as continuous variable and as categorical variable (5 dB categories).1

Effect modification by a priori-specified variables was investigated by stratified analyses for age during follow-up (<60, 60–75, ≥ 75 year), sex, smoking status, educational level, fruits intake (<150, 150–300, ≥300 g/day), and BMI (<25, 25–30, ≥ 30 kg/m2). These variables were selected based on previous studies.5,23

We tested whether back-extrapolation of the air pollution concentrations to the baseline year affected the results (details in eAppendix 3; http://links.lww.com/EDE/A767). Sensitivity analyses restricted to subjects who did not move during follow-up were conducted. We conducted analyses without the large Austrian cohort.We checked for spatial clustering of residuals of the models using random effects of the spatial area units (often neighborhood or municipality) in each cohort. The two traffic indicator variables were analyzed in combination with background NO2 concentration.

All cohort-specific analyses were done in STATA versions 10 to 12 (StataCorp, College Station, TX) except models with random effects, which were done using R software (R Foundation for Statistical Computing, 2004 [ISBN 3-900051-00-3], Vienna, Austria).

Back to Top | Article Outline

Meta-analysis

Meta-analyses of cohort-specific effect estimates were conducted using the DerSimonian–Laird24 method with random effects. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated for fixed increments. Heterogeneity between cohorts was quantified by the I2 statistic and tested by the χ2 test from Cochran’s Q statistic.25

Effect modification was tested by meta-analysis of the pooled estimates from the various strata, and by computing the χ2 test of heterogeneity.

We tested whether effect estimates differed for cohorts for which the variance explained by the land-use regression model cross-validation was smaller or larger than 60% for PM2.5. In addition, we tested whether effect estimates differed by region of Europe (“North”: Sweden, Norway, Finland, Denmark; “West and Middle”: United Kingdom, The Netherlands, Germany, France, Austria, and Switzerland; and “South”: Italy and Greece). All meta-analyses were conducted in STATA, version 12 (StataCorp).

Back to Top | Article Outline

RESULTS

Characteristics of the Study Population

The total study population consisted of 367,383 participants contributing 5,119,317 person-years at risk (average time of follow-up = 13.9 years). Of the participants, 29,076 died from a natural cause during follow-up, with 9994 deaths due to CVD (4992 to ischemic heart disease, 2264 to MI, and 2484 cerebrovascular disease) (Table 1). The ratio of specific causes to all CVD deaths varied substantially among cohorts. Cohorts were recruited mostly in the 1990s. Cohorts differed in the number of participants, the mean baseline age, and the availability of information on confounders (Table 2, and eAppendix 4; http://links.lww.com/EDE/A767). Age, sex, smoking status, and area-level SES were available for all cohorts. Smoking intensity and duration were available as continuous variables for all cohorts except for the Austrian and French. The Austrian cohort had data on occupation and employment status but not on education.

TABLE 2

TABLE 2

Back to Top | Article Outline

Air Pollution Exposure

Air pollution concentrations varied among and within study areas (eAppendix 5; http://links.lww.com/EDE/A767). Concentrations showed an increase from northern to southern Europe cohorts. The average NO2 concentration ranged from 5.2 μg/m3 (EPIC-Umeå) to 59.8 μg/m3 (SIDRIA-Turin) and the average PM2.5 concentration from 6.6 μg/m3 (SDPP) to 31.0 μg/m3 (SIDRIA-Turin). Median differences between area-specific 5th and 95th percentiles in NO2 and PM2.5 concentrations were 21.1 and 4.3 μg/m3, respectively. Contrasts were within-city contrasts for most study areas but could also include between-city contrasts for the cohorts covering larger geographical areas. Correlations between exposure measures were moderate to high in each cohort (eAppendix 6; http://links.lww.com/EDE/A767).

Back to Top | Article Outline

Main Results

Most HRs for the association between mortality outcomes and the air pollutants were approximately 1.0 (Table 3, Figure, and eAppendix 7; http://links.lww.com/EDE/A767), with the exception of the association of particulate mass and cerebrovascular disease mortality (Table 3 and Figure). The HR for PM2.5 was 1.21 (95% CI = 0.87–1.69) per 5 μg/m3 and for PM10 the HR was 1.22 (0.91–1.63) per 10 μg/m3. In addition, associations between all outcomes and traffic intensity on the nearest road were slightly increased; the HR for total CVD mortality was 1.02 (0.99–1.05) per 5000 motor vehicles per day, an association that remained after adjustment for noise (eAppendix 8; http://links.lww.com/EDE/A767).

TABLE 3

TABLE 3

FIGURE. Ad

FIGURE. Ad

HRs for confounder Model 1 (adjusted only for calendar year and sex) were highest; after adjustment for individual-level confounders, HRs decreased. Sensitivity analyses showed that smoking variables were primarily responsible for this decrease. Inclusion of area-level SES variables led to a small further decrease in HRs.

No heterogeneity among cohorts was found for total CVD mortality, except for the HRs for traffic intensity on major roads in a 100-m buffer. HRs for random-effects (default method) and fixed-effects CVD meta-analyses were similar (data not shown). For the more specific causes, low to moderate heterogeneity was found (I2 < 50%).

Back to Top | Article Outline

Sensitivity Analyses

Additional adjustment for hypertension and physical activity, diabetes mellitus and cholesterol, and noise did not change the pooled HRs compared with the main model HRs (eAppendix 8; http://links.lww.com/EDE/A767). Back-extrapolation for NO2 was possible in 18 cohorts, whereas back-extrapolation for PM10 was possible for seven cohorts spread over Europe. HRs were not different between the back-extrapolated concentrations at baseline in the year of recruitment and the concentrations based on 2008–2011 measurements. Pooled HRs for CVD mortality for back-extrapolated NO2 concentrations based on the difference and the ratio method were 1.02 (95% CI = 0.97–1.07) and 1.01 (0.98–1.05)—essentially the same as the 1.02 (0.97–1.07) for the main ESCAPE exposure for the 18 cohorts with back-extrapolated NO2 concentrations. For cerebrovascular disease mortality, pooled HRs for back-extrapolated PM10 concentrations based on the difference and the ratio method were 1.36 (0.92–2.02) and 1.22 (0.93–1.59) also the same as the 1.21 (0.79–1.85) for the main ESCAPE exposure for the seven cohorts with back-extrapolated PM10 concentrations. Analyses restricted to subjects who did not move during follow-up resulted in similar HRs as for the main analyses (data not shown). HRs for total CVD mortality without the influential Austrian cohort were similar to those from the main analyses. HRs for cerebrovascular disease mortality for PM2.5 and PM10 increased to 1.32 (0.93–1.89) and 1.30 (0.96–1.77), respectively. HRs with and without accounting for spatial clustering of residuals were similar. PM2.5 effect estimates for cerebrovascular disease mortality were similar for the cohorts for which the variance explained by land-use regression model cross-validation was smaller or larger than 60%: 1.24 (0.71–2.15) (N = 6) and 1.22 (0.78–1.92) (N = 12), respectively. PM2.5 effect estimates for cerebrovascular disease mortality were also not statistically different among the cohorts in different regions. For the other pollutants also, effect estimates did not differ based on validation R2 or region.

Back to Top | Article Outline

Effect Modification

For none of the evaluated variables, was there an indication of effect modification, except for PM2.5, for which men had higher HRs than women (eAppendix 9; http://links.lww.com/EDE/A767). For NO2, there was no difference in HRs between men and women.

Back to Top | Article Outline

DISCUSSION

We found that most HRs for the association between air pollutants and mortality from overall CVD and specific CVDs were approximately 1.0, with the exception of particulate mass and cerebrovascular disease mortality for which there was suggestive evidence for an association.

An association between PM2.5 exposure and overall cardiovascular mortality was first identified in the Harvard Six Cities study (risk ratio [RR] = 1.28 [95% CI = 1.13–1.44] per 10 μg/m3),4 and the American Cancer Society study (RR = 1.09 [1.03–1.16] per 10 μg/m3),3 both from the United States. Overall, subsequent cohort studies have confirmed these associations.5,23 A recent review documented large heterogeneity of effect estimates across studies, with several studies showing no association.23 A recent study of a national United Kingdom cohort found associations between ambient air pollution and all-cause mortality, but associations were approximately 1.0 for cardiovascular mortality.9

Associations between exposure to NO2 or elemental carbon (which correlates highly with PM2.5 absorbance) and overall cardiovascular mortality were observed in studies conducted in Europe, Japan, Canada, and the United States.6–8,26–31 The number of studies on the association between long-term coarse particle exposure and cardiovascular mortality is still small, and there is no clear evidence that coarse particles are associated with overall cardiovascular mortality.3,11

The diversity of diseases included in the broad category of CVDs makes it unlikely that the risk associated with air pollution exposure is the same for all diseases.5 Studies have therefore evaluated specific cardiovascular causes, with focus on ischemic heart disease, and cerebrovascular disease mortality. Most studies investigating ischemic heart disease mortality found increased risks for PM2.5 or PM10 exposure,14,15,32–34 and for NO2 or NOx exposure.2,8,29–31 A few studies, however, found no association between air pollution exposure and ischemic heart disease mortality.1,11 Fewer studies have investigated cerebrovascular disease mortality. In the Dutch cohort study and in the Women’s Health Initiative Study, a strong association was found between cerebrovascular disease mortality and black smoke and PM2.5, respectively.1,15 In contrast, no such association was found in the American Cancer Society study,14 a Norwegian cohort study,2 a recent Danish cohort study,8 and recent large Canadian population-based studies.30,34

In previous cohort studies assessing effects of long-term exposure to air pollution on mortality, it has generally been found that effect estimates were larger for cardiovascular mortality than for all-cause or natural-cause mortality,23 with a few exceptions.6,9,35,36 However, the heterogeneity in effect estimates across studies was much larger for CVD mortality than for natural-cause mortality.23 In our study, we found the opposite; in the same cohorts, using the same exposure and statistical methods, an association of natural-cause mortality with PM2.5 (HR = 1.07 [95% CI = 1.02–1.13] per 5 μg/m3) was found.37 We do not have a clear explanation for the lack of an association with cardiovascular mortality within the ESCAPE project, whereas we did find an association with all natural-cause mortality. This discrepancy could be due to the fact that cardiovascular causes of death contributed no more than 30% of all deaths in 15 of the 22 cohorts. In the Harvard Six Cities and in the American Cancer Society studies, participants were born on average several decades earlier than in most of our cohorts, and the relative contribution of cardiovascular mortality to total mortality in these two cohorts was much larger.

We speculate that changes in cardiovascular risk profiles over time (eg, reduced smoking and increased medication and medical treatment)38 have altered the relation between air pollution and CVD mortality. There is some support for this speculation in a recent German study showing stronger association for early-period versus later-period exposures.31 These changes over time result in a lower fatality rate for cardiovascular events, suggesting that risk factors may increasingly have a different association with incident cardiovascular events and cardiovascular mortality.38 We found an association between PM10 and incident (fatal plus not-fatal) coronary events (HR = 1.12 [95% CI = 1.01–1.25] for an increase of 10 μg/m3 in PM10) in a subset of 11 cohorts within the ESCAPE project, using the same exposure and statistical methods.39 Furthermore, in the same subset of 11 cohorts, PM2.5 was associated with stroke incidence (HR = 1.19 [95% CI = 0.88–1.62]) (M. Stafoggia, unpublished data, 2014). For the incidence of both coronary events and stroke, the number of cases was approximately six times higher than the number of ischemic heart disease and CVD deaths, respectively, in our analysis. Incident events may be less affected by medication use than cardiovascular deaths, which are likely often preceded by nonfatal events. This suggests that a large number of incident coronary and stroke events do not lead to cardiovascular death soon after the event. Compared with our natural-cause mortality analysis, we observed more substantial heterogeneity of effect estimates across cohorts, especially for the more specific causes.

Another limitation is that we have to rely on data from mortality registries. There may be coding differences in death certificates among countries and among ESCAPE study areas. Such differences might have contributed to the heterogeneous results among ESCAPE cohorts. Heterogeneity in effect estimates due to coding differences might also have affected the overall pooled analyses, possibly leading to the lack of an association with cardiovascular mortality. We did observe increased effect estimates in models adjusted only for age and sex; these estimates were reduced after adjustment for confounders. It is unlikely that we over-adjusted our models, because our confounder models were similar to previous studies and because smoking (a well-established risk factor) was mainly responsible for the drop in effect estimates. We further found no evidence for CVD mortality associations in never-smokers.

CIs were small for overall CVD mortality and wider for specific causes. The increased HRs for cerebrovascular disease mortality could be a chance finding among the many associations studied. However, the consistency across models and the coherence with the stroke incidence analysis support these associations.

One source of variability of effect estimates among previous studies is likely related to varying degrees of exposure misclassification. A strength of ESCAPE is that air pollution exposure assessment within the ESCAPE study areas was conducted in a standardized way.

A limitation of our study is that the land-use regression models used for exposure assessment were based on air pollution measurements in the period 2008–2011, whereas the cohort studies included in ESCAPE started in the past (1985–2007, with most studies starting in the mid-1990s). Analyses using exposures back-extrapolated to the recruitment date showed similar HRs. Four recent studies in The Netherlands,40 Rome,41 United Kingdom,42 and Vancouver43 have shown that, for periods up to 10 years and longer, spatial air pollution contrasts often remained the same, even with a decrease in concentrations over time. Thus, land-use regression models based on current air pollution data may be valid predictors of historic spatial contrasts. Measurement error is an unlikely explanation of the lack of associations, because we did observe an association with natural-cause mortality using the same exposure variables and the same set of cohorts.

In conclusion, most HRs for the association of air pollutants with mortality from overall cardiovascular and from specific CVDs were approximately 1.0 in 22 European cohort studies, with the exception of particulate mass and cerebrovascular disease mortality for which there was suggestive evidence for an association.

Back to Top | Article Outline

ACKNOWLEDGMENTS

Mortality data for SAPALDIA were provided by the Swiss National Cohort Study which performed probabilistic record linkage between anonymized SAPALDIA data, including available information on vital status and date of death, on the one hand, and census and mortality data of the Federal Office of Statistics on the other hand. The SIDRIA cohort studies were co-founded by the Italian Ministry of Health. EPIC-San Sebastian was co-founded by the CIBERESP, Consortium for Biomedical Research in Epidemiology and Public Health, and Health Department of Basque Region Government. We thank Markus Wallner, Christian Bernhard, Elmar Bechter, Andrea Kaufmann, and Gabriela Dür from the Vorarlberg State Government. We thank Marjan Tewis, Marieke Oldenwening, Marta Cirach, Audrey de Nazelle, Bernhard Anwander, Paolo Crosignani, Jon Wickmann, Daniela Raffaele, Marco Gilardetti, Thomas Kuhlbusch, Ulrich Quass, Mohammad Vossoughi, Simone Bucci, Patrizio Pasquinelli, Giuseppe Costa, L.-J. Sally Liu, Andrea Kleiner, Pekka Taimisto, Arto Pennanen (from the Department of Environmental Health at National Institute for Health and Welfare), Jacopo Fogola and Daniele Grasso (Regional Agency for the Protection of the Environment of Piedmont), and Roberta Onorato, for their help with exposure assessment and data management within ESCAPE.

Back to Top | Article Outline

REFERENCES

1. Beelen R, Hoek G, Houthuijs D, et al. The joint association of air pollution and noise from road traffic with cardiovascular mortality in a cohort study. Occup Environ Med. 2009;66:243–250
2. Nafstad P, Håheim LL, Wisløff T, et al. Urban air pollution and mortality in a cohort of Norwegian men. Environ Health Perspect. 2004;112:610–615
3. Pope CA 3rd, Burnett RT, Thun MJ, et al. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA. 2002;287:1132–1141
4. Laden F, Schwartz J, Speizer FE, Dockery DW. Reduction in fine particulate air pollution and mortality: extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med. 2006;173:667–672
5. Brook RD, Rajagopalan S, Pope CA, et al. Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the american heart association. Circulation. 2010;121:2331–2378
6. Beelen R, Hoek G, van den Brandt PA, et al. Long-term effects of traffic-related air pollution on mortality in a Dutch cohort. Environ Health Perspect. 2008;116:196–202
7. Filleul L, Rondeau V, Vandentorren S, et al. Twenty five year mortality and air pollution: results from the French PAARC survey. Occup Environ Med. 2005;62:453–460
8. Raaschou-Nielsen O, Andersen ZJ, Jensen SS, et al. Traffic air pollution and mortality from cardiovascular disease and all causes: a Danish cohort study. Environ Health. 2012;11:60
9. Carey IM, Atkinson RW, Kent AJ, van Staa T, Cook DG, Anderson HR. Mortality associations with long-term exposure to outdoor air pollution in a national English cohort. Am J Respir Crit Care Med. 2013;187:1226–1233
10. Cesaroni G, Badaloni C, Gariazzo C, et al. Long-term exposure to urban air pollution and mortality in a cohort of more than a million adults in Rome. Environ Health Perspect. 2013;121:324–331
11. Puett RC, Hart JE, Suh H, Mittleman M, Laden F. Particulate matter exposures, mortality, and cardiovascular disease in the health professionals follow-up study. Environ Health Perspect. 2011;119:1130–1135
12. Ueda K, Nagasawa SY, Nitta H, Miura K, Ueshima HNIPPON DATA80 Research Group. . Exposure to particulate matter and long-term risk of cardiovascular mortality in Japan: NIPPON DATA80. J Atheroscler Thromb. 2012;19:246–254
    13. Ostro B, Reynolds P, Goldberg D, et al. Assessing long-term exposure in the California teachers study. Environ Health Perspect. 2011;119:A242–A243
      14. Pope CA 3rd, Burnett RT, Thurston GD, et al. Cardiovascular mortality and long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of disease. Circulation. 2004;109:71–77
      15. Miller KA, Siscovick DS, Sheppard L, et al. Long-term exposure to air pollution and incidence of cardiovascular events in women. N Engl J Med. 2007;356:447–458
      16. Puett RC, Hart JE, Yanosky JD, et al. Chronic fine and coarse particulate exposure, mortality, and coronary heart disease in the Nurses’ Health Study. Environ Health Perspect. 2009;117:1697–1701
      17. Raaschou-Nielsen O, Andersen ZJ, Beelen R, et al. Air pollution and lung cancer incidence in 17 European cohorts within the ESCAPE study. Lancet Oncol. 2013;14:813–822
      18. Eeftens M, Beelen R, de Hoogh K, et al. Development of Land Use Regression models for PM(2.5), PM(2.5) absorbance, PM(10) and PM(coarse) in 20 European study areas; results of the ESCAPE project. Environ Sci Technol. 2012;46:11195–11205
      19. Beelen R, Hoek G, Vienneau D, et al. Development of NO2 and NOx land use regression models for estimating air pollution exposure in 36 study areas in Europe—the ESCAPE project. Atmos Environ. 2013;72:10–23
      20. Eeftens M, Tsai M, Ampe C, et al. Spatial variation of PM2.5, PM10, PM2.5 absorbance and PM coarse concentrations between and within 20 European study areas and the relationship with NO2—results of the ESCAPE project. Atmos Environ. 2012;62:303–317
      21. Cyrys J, Eeftens M, Heinrich J, et al. Variation of NO2 and NOx concentrations between and within 36 European study areas: results from the ESCAPE study. Atmos Environ. 2012;62:374–390
      22. Thiébaut AC, Bénichou J. Choice of time-scale in Cox’s model analysis of epidemiologic cohort data: a simulation study. Stat Med. 2004;23:3803–3820
      23. Hoek G, Krishnan MK, Beelen R, Peters A, Ostro B, Kaufman J. Long-term air pollution exposure and cardio- respiratory mortality. Environ Health. 2013;12:43
      24. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188
      25. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–1558
      26. Gehring U, Heinrich J, Krämer U, et al. Long-term exposure to ambient air pollution and cardiopulmonary mortality in women. Epidemiology. 2006;17:545–551
      27. Smith KR, Jerrett M, Anderson HR, et al. Public health benefits of strategies to reduce greenhouse-gas emissions: health implications of short-lived greenhouse pollutants. Lancet. 2009;374:2091–2103
      28. Gan WQ, Koehoorn M, Davies HW, Demers PA, Tamburic L, Brauer M. Long-term exposure to traffic-related air pollution and the risk of coronary heart disease hospitalization and mortality. Environ Health Perspect. 2011;119:501–507
      29. Yorifuji T, Kashima S, Tsuda T, et al. Long-term exposure to traffic-related air pollution and mortality in Shizuoka, Japan. Occup Environ Med. 2010;67:111–117
      30. Chen H, Goldberg MS, Burnett RT, Jerrett M, Wheeler AJ, Villeneuve PJ. Long-term exposure to traffic-related air pollution and cardiovascular mortality. Epidemiology. 2013;24:35–43
      31. Heinrich J, Thiering E, Rzehak P, et al. Long-term exposure to NO2 and PM10 and all-cause and cause-specific mortality in a prospective cohort of women. Occup Environ Med. 2013;70:179–186
      32. Puett RC, Schwartz J, Hart JE, et al. Chronic particulate exposure, mortality, and coronary heart disease in the nurses’ health study. Am J Epidemiol. 2008;168:1161–1168
      33. Lipsett MJ, Ostro BD, Reynolds P, et al. Long-term exposure to air pollution and cardiorespiratory disease in the California teachers study cohort. Am J Respir Crit Care Med. 2011;184:828–835
      34. Crouse DL, Peters PA, van Donkelaar A, et al. Risk of nonaccidental and cardiovascular mortality in relation to long-term exposure to low concentrations of fine particulate matter: a Canadian national-level cohort study. Environ Health Perspect. 2012;120:708–714
      35. Hales S, Blakely T, Woodward A. Air pollution and mortality in New Zealand: cohort study. J Epidemiol Community Health. 2012;66:468–473
      36. Elliott P, Shaddick G, Wakefield JC, de Hoogh C, Briggs DJ. Long-term associations of outdoor air pollution with mortality in Great Britain. Thorax. 2007;62:1088–1094
      37. Beelen R, Raaschou-Nielsen O, Stafoggia M, et al. Effects of long-term exposure to air pollution on natural-cause mortality: an analysis of 22 European cohorts within the multicentre ESCAPE project. Lancet. 2013;13:62158–3 DOI: 10.1016/S0140-6736
      38. Nichols M, Townsend N, Scarborough P, Rayner M. Trends in age-specific coronary heart disease mortality in the European Union over three decades: 1980–2009. Eur Heart J. 2013;34:3017–3027
      39. Cesaroni G, Forastiere F, Stafoggia M, et al. Long term exposure to ambient air pollution and incidence of acute coronary events: prospective cohort study and meta-analysis in 11 European cohorts from the ESCAPE Project. BMJ. 2014 DOI: 10.1136/bmj.f7412
      40. Eeftens M, Beelen R, Fischer P, Brunekreef B, Meliefste K, Hoek G. Stability of measured and modelled spatial contrasts in NO(2) over time. Occup Environ Med. 2011;68:765–770
      41. Cesaroni G, Porta D, Badaloni C, et al. Nitrogen dioxide levels estimated from land use regression models several years apart and association with mortality in a large cohort study. Environ Health. 2012;11:48
      42. Gulliver J, de Hoogh K, Hansell A, Vienneau D. Development and back-extrapolation of NO2 land use regression models for historic exposure assessment in Great Britain. Environ Sci Technol. 2013;47:7804–7811
      43. Wang R, Henderson SB, Sbihi H, Allen RW, Brauer M. Temporal stability of land use regression models for traffic-related air pollution. Atmos Environ. 2013;64:312–319

      Supplemental Digital Content

      Back to Top | Article Outline
      © 2014 by Lippincott Williams & Wilkins, Inc