The fraction of lung cancer incidence attributable to fine particulate air pollution in France: Impact of spatial resolution of air pollution models

(1)

HAL Id: hal-02089354

https://hal.archives-ouvertes.fr/hal-02089354

Submitted on 3 Apr 2019

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

The fraction of lung cancer incidence attributable to fine particulate air pollution in France: Impact of spatial

resolution of air pollution models

Ivana Kulhanova, Xavier Morelli, Alain Le Tertre, Dana Loomis, Barbara Charbotel, Sylvia Medina, Jean-Nicolas Ormsby, Johanna Lepeule, Rémy

Slama, Isabelle Soejormataram

To cite this version:

Ivana Kulhanova, Xavier Morelli, Alain Le Tertre, Dana Loomis, Barbara Charbotel, et al.. The

fraction of lung cancer incidence attributable to fine particulate air pollution in France: Impact of

spatial resolution of air pollution models. Environment International, Elsevier, 2018, 121 (Part 2),

pp.1079-1086. �10.1016/j.envint.2018.09.055�. �hal-02089354�

(2)

Contents lists available atScienceDirect

Environment International

journal homepage:www.elsevier.com/locate/envint

The fraction of lung cancer incidence attributable to ﬁ ne particulate air pollution in France: Impact of spatial resolution of air pollution models

Ivana Kulhánová

^a,⁎,1

, Xavier Morelli

^a,b,1

, Alain Le Tertre

^c

, Dana Loomis

^d

, Barbara Charbotel

^e,f

, Sylvia Medina

^c

, Jean-Nicolas Ormsby

^g

, Johanna Lepeule

^h

, Rémy Slama

^h

,

Isabelle Soerjomataram

^a

aSection of Cancer Surveillance, International Agency for Research on Cancer, Lyon, France

bDepartment of Cancer and Environment, Centre Léon Bérard, Lyon, France

cSanté publique France, Paris, France

dMonographs Section, International Agency for Research on Cancer, Lyon, France

eUniversité de Lyon, Université Claude Bernard Lyon 1, IFSTTAR, UMRESTTE, UMR_T9405, Lyon, France

fService des maladies professionnelles, Hospices Civils de Lyon, Lyon, France

gFrench Agency for Food, Environmental and Occupational Health & Safety, Maisons-Alfort, France

hInserm, CNRS, University Grenoble-Alpes, IAB (Institute for Advanced Biosciences), Team of Environmental Epidemiology, Grenoble, France

A R T I C L E I N F O

Handling Editor: X Querol Keywords:

Air pollution Lung cancer

Population attributable fraction France

A B S T R A C T

Outdoor air pollution is a leading environmental cause of death and cancer incidence in humans. We aimed to estimate the fraction of lung cancer incidence attributable tofine particulate matter (PM2.5) exposure in France, and secondarily to illustrate the influence of the input data and the spatial resolution of information on air pollution levels on this estimate. The population attributable fraction (PAF) was estimated using a nationwide spatially refined chemistry-transport model with a 2-km spatial resolution, neighbourhood-scale population density data, and a relative risk from a published meta-analysis. We used the WHO guideline value for PM2.5

exposure (10μg/m³) as reference. Sensitivity analyses consisted in attributing the nation-wide median exposure to all areas and using alternative input data such as reference of PM2.5 exposure level and relative risk.

Population-weighted median PM2.5level in 2005 was 13.8μg/m³; 87% of the population was exposed above the guideline value. The burden of lung cancer attributable to PM_2.5exposure corresponded to 1466 cases, or 3.6%

of all cases diagnosed in 2015. Sensitivity analyses showed that the use of a national median of PM2.5exposure would have led to an underestimation of the PAF by 11% (population-weighted median) and by 72% (median of raw concentration), suggesting that our estimates would have been higher with even moreﬁnely spatially- resolved models. When the PM2.5reference level was replaced by the 5th percentile of country-scale exposure (4.9μg/m³), PAF increased to 7.6%. Other sensitivity analyses resulted in even higher PAFs. Improvements in air pollution are crucial for quantitative health impacts assessment studies. Actions to reduce PM_2.5levels could substantially reduce the burden of lung cancer in France.

1. Introduction

Outdoor air pollution is a leading environmental cause of death and is related to increased deaths from cardiovascular and respiratory diseases (WHO Regional Oﬃce for Europe, 2013) as well as to the incidence of numerous diseases. In 2010, Lelieveld and colleagues reported that about 3.15 million deaths worldwide were attributable to atmospheric ﬁne particulate matter (PM2.5) (Lelieveld et al., 2015), which represented about 6% of total mortality worldwide. A recent

Global Burden of Disease (GBD) study estimated that 4.2 million deaths (or 7.5% of total global mortality) in 2015 were attributable to ambient air pollution (GBD 2015 Risk Factors Collaborators, 2016). These results took into account only outdoor PM2.5levels, household PM2.5le- vels from biomass burning were excluded. In France, PM2.5were estimated to be responsible for 9% of the total mortality (Pascal et al., 2016). Regarding cancer, a group of experts convened by the Interna- tional Agency for Research on Cancer (IARC) concluded that PM2.5in outdoor air is carcinogenic to humans (IARC Group 1), in particular for

https://doi.org/10.1016/j.envint.2018.09.055

Received 11 June 2018; Received in revised form 21 September 2018; Accepted 27 September 2018

⁎Corresponding author at: Section of Cancer Surveillance, International Agency for Research on Cancer, 150 Cours Albert Thomas, 69372 Lyon, France.

1These authors contributed equally to this study.

E-mail address:kulhanovai@fellows.iarc.fr(I. Kulhánová).

Available online 30 October 2018

T

(3)

lung cancer (Loomis et al., 2013). In France, only estimates at the urban area level are available, restricted to the Grenoble area (about 0.7% of the total population in France) assessing that 6.8% of lung cancer cases in this area were attributable to PM2.5(Morelli et al., 2016).

Technological development has markedly advanced exposure assessment to ambient air pollution. The majority of early etiologic studies on PM2.5health eﬀects have utilised data from air quality monitoring networks (Ballester et al., 1996;Carugno et al., 2016;Fattore et al., 2011;Filleul et al., 2005;Künzli et al., 2000;Medina et al., 2004), which generally have a poor spatial resolution, with an increasing number of studies relying on satellite imagery (Evans et al., 2013;Lim et al., 2012) or chemical transport models (Anenberg et al., 2010).

When it comes to quantitative health impact assessment (HIA) studies, background monitoring stations are generally used to assess exposure (Pascal et al., 2013). However, recent HIA studies increasingly consider air pollution models, some of which with high spatial resolution (Orru et al., 2009; Pascal et al., 2016). A study in two French urban areas suggested that assessing exposures from background monitoring stations may underestimate attributable risk of death by 10 to 20%, compared toﬁne-scale models with spatial resolutions taking into account street-scale variations in the pollutants concentration (Morelli et al., 2016). In France, suchﬁne-scale models are available for small geographical units such as the urban area (Bentayeb et al., 2014).

Here, we aimed to estimate the proportion and the number of new lung cancer cases in France in 2015 attributable to exposure to PM2.5. A secondary aim was to study the impact of the spatial resolution of PM2.5

exposure model and of information on population density. We also tested the impact of changing the reference level and the relative risk estimates (dose-response function) used. While outdoor air pollution is a complex mixture containing various known carcinogens, PM2.5was chosen as an indicator of exposure as it is a prominent component of the mixture, with proven long-term eﬀects (WHO Regional Oﬃce for Europe, 2013).

2. Materials and methods 2.1. Study design and area

The burden of lung cancer attributable to PM2.5 was calculated

using the population attributable fraction (PAF) based on Comparative Risk Assessment method (Cohen et al., 2005;Ezzati et al., 2002). This methodology is based on combining exposure to air pollution and its distribution in the population with exposure-risk estimates at each level of exposure (World Health Organization, 2016). This study design is similar to HIA (1999). The study area was metropolitan France. We assumed a latency time between exposure to PM2.5and lung cancer incidence of 10 years; as such the number of lung cancer cases attributable to exposure to PM2.5in the year 2015 was based on PM2.5

concentrations in 2005 (Yanagi et al., 2012). Information on population density by neighbourhood was obtained for the year 2006 (the closest to 2005 with data available by neighbourhood, or IRIS, a geographical census unit generally consisting of 2000 persons) from the National Institute of Statistics and Economic Studies (INSEE) (INSEE, n.d.). The study population and the lung cancer incidence considered in this study were restricted to adults at age of 30 years and above.

2.2. Air pollution exposure

Air pollution exposure (annual mean concentration of PM2.5) was estimated from Gazel-Air model. This model is a nationwide air quality model which provides annual mean concentration of PM2.5 on a 2 × 2 km grid in metropolitan France from 1989 to 2008. It was de- veloped based on the CHIMERE chemistry-transport model, mesh re- ﬁnement and data assimilation with geostatistical analyses (Bentayeb et al., 2014). Emission data were derived from the European Monitoring and Evaluation Program emission cadastre consisting of gridded annual national emissions at 50 km resolution on the European scale. The CHIMERE model was used to map PM2.5concentrations at 10 km resolution in France. Finally, mesh reﬁnement was performed to improve the spatial distribution of PM2.5 concentrations at 2 km resolution (Bentayeb et al., 2014). The modelling strategy has been described in more details elsewhere (Bentayeb et al., 2014;Bentayeb et al., 2015). A geographic information system (GIS) was used to combine the Gazel-Air model with the population distribution by neighbourhood in order to obtain a population-weighted PM2.5exposure. The spatial distribution of PM2.5concentrations and the population density are illustrated in Fig. 1.

Fig. 1.PM2.5concentrations estimated by the Gazel-Air model (yearly averages inμg/m³; 2 km spatial resolution) in 2005 in metropolitan France (A) and population density by neighbourhood (number of inhabitants/km²) in 2006 (B).

I. Kulhánová et al. Environment International 121 (2018) 1079–1086

1080

(4)

2.3. Health outcome

Lung cancer incidence (ICD-10: C33–C34) was obtained from the French network of cancer registries (FRANCIM) for 2013. To estimate the number of lung cancer cases in 2015, we applied the 2013 age- and sex-speciﬁc lung cancer incidence rates to the 2015 population. In France, population-based cancer registries for adult population provide incidence data for only 20% of the population (Santé publique France 2010), while lung cancer mortality is available for the entire country at the municipality level. Assuming similarity between lung cancer mortality and incidence due to high case-fatality for lung cancer (Alberg et al., 2013), we used mortality data to estimate the fraction of lung cancer attributable to PM2.5excess exposure. Lung cancer deaths (ICD- 10: C34) by municipality were obtained from the Epidemiology Centre on Medical Causes of Death of the National Institute of Health and Medical Research (CépiDc, INSERM) for the years 2011–2013. How- ever, lung cancer deaths were not disclosed for municipalities where a recorded total number of deaths was lower than 5 between 2011 and 2013 due to conﬁdentiality. These missing deaths accounted for approximately 3% (or 2769 deaths) of the total lung cancer deaths recorded in France.

We redistributed these missing lung cancer deaths using the fol- lowing imputation method. First, we calculated the expected number of lung cancer deaths in each municipality with missing information by multiplying the population size of the municipality with the national lung cancer mortality rate (48.1 per 100,000) reported for the year 2012 (Inserm/CépiDc, n.d.). We chose the year 2012 as the mid-year of the period 2011–2013 of the mortality data. Then, we computed a correction factor to restrict the number of the imputed lung cancer deaths to 2769. This correction factor consisted of a ratio calculated in the municipalities with missing information by dividing the total number of missing lung cancer deaths by the total number of expected lung cancer deaths. Finally, the expected number of lung cancer deaths in municipalities with < 5 deaths was weighted by the correction factor to account for the missing 3% of lung cancer deaths (formula 1).

=

∑

=

∑ ×

=

∑ ×

= = =

( )

LCdeaths D D

D

N MR N

( )

2769

i n

i ected i

n

i LC

i n

i 1

exp

1 1

48.1

100000 (1)

whereDis the total number of missing lung cancer deaths (2769),Niis the population size in each municipality i, MRLC is the lung cancer mortality rate (48.1 per 100,000) reported in France for the year 2012, andnis the number of municipalities with missing information. About 11,400 (or 31% of total) municipalities were imputed due to missing information on lung cancer deaths. Attributing a homogeneous rate is expected to reduce the variance of ourfinal estimate and thus artifi- cially reduce the confidence intervals.

2.4. Risk characterization

To estimate the proportion of lung cancer incidence attributable to PM2.5in France, weﬁrst calculated the number (AN) of lung cancer deaths attributable to past excess exposure to PM2.5in each neigh- bourhoodiaccording to theformula (2).

= ⎛

⎝ ⎞

⎠

⎛

⎝

⎜

× −

×

⎞

⎠

⎟

−

( )

AN n D N

RR RR

exp 1

exp

i i

X X

10

i ref

(2) whereni is the population size in neighbourhood i,DandNare the number of lung cancer deaths and population in the smallest geographical unit available (IRIS),RR is the relative risk of lung cancer associated with a 10μg/m³increases in exposure to PM2.5,Xiis the PM2.5concentration observed in the neighbourhoodi,Xref is the reference level chosen for PM2.5. When a neighbourhood was spatially located across two (or more) grid cells of the Gazel-Air model, its

population was proportionally distributed to the corresponding grid cells by applying a ratio based on the areas of the diﬀerent neighbourhoods sections.

The current World Health Organization (WHO) air quality guideline –an annual mean concentration of 10μg/m³–was chosen as the reference level of exposure to PM2.5(World Health Organization, 2005).

This guideline represents the lowest level at which total, cardiopulmonary and lung cancer mortality have been demonstrated to increase with more than 95% conﬁdence in response to long-term exposure to PM2.5according to a study of the American Cancer Society (Pope et al., 2002;World Health Organization, 2005). However, there is no evidence of a threshold in the relation between PM2.5exposure and lung cancer incidence, which is why we also used an alternative reference value in a sensitivity analysis (Pascal et al., 2016). Regarding the relative risk (RR) of lung cancer associated with PM2.5exposure, we used the one from a meta-analysis byHamra et al. (2014), which is equal to 1.09 (95% conﬁdence interval, CI: 1.04–1.14) for each 10μg/

m³increase in PM2.5exposure. It is assumed to be the same for both sexes. The meta-analysis from which the RR was derived included 14 studies of lung cancer mortality and incidence from Europe, North America, China and Japan. All studies were adjusted for sex and age.

More than two thirds of studies were additionally adjusted for smoking and socioeconomic status (Hamra et al., 2014).

Then, the PAF of lung cancer deaths related to excess exposure to PM2.5 (i.e., above the reference level) was computed by summing neighbourhood-speciﬁc PM2.5attributable deaths and dividing them by the total number of lung cancer deaths in the country for the same

period = ∑

=

PAF AN D/

i n

i 1

. Finally, the PAF was multiplied by the estimated lung cancer incidence in 2015 to derive the number of new lung cancer cases in 2015 attributable to PM2.5exposure in France. The results based on this analysis refer to the“main model”. The 95% con- ﬁdence intervals of the PAF were based on the variability of the RR from the meta-analyses. All analytical steps necessary for the calculations are illustrated in Fig. S1 (Online supplementary material).

2.5. Sensitivity analyses

In addition to the main model, the impact of PM2.5on lung cancer incidence was assessed usingﬁve alternative hypotheses (H1–H5). In theﬁrst hypothesis (H1), we estimated the number of lung cancer cases attributable to PM2.5 assuming homogeneous population-weighted PM2.5exposure at the department (H1a) and country levels (H1b), i.e.

the same value of the PM2.5exposure for each geographical unit observed. The homogeneous values corresponded to the median PM2.5

exposure in each department or in the country. This allowed testing the sensitivity of our estimate to the spatial resolution of the PM2.5con- centrations. In a second sensitivity analysis (H2), we assumed homogeneous raw value of the PM2.5concentration at department (H2a) or country level (H2b) testing the impact of population density.

The PM2.5reference level and the concentration-response function are central parameters of the estimation of the number of attributable lung cancer cases. The third (H3) and fourth (H4) hypotheses assumed an alternative reference level of PM2.5and an alternative RR, respectively. The reference level of PM2.5in H3 was set to 4.9μg/m³ and corresponded to the 5th percentile of the PM2.5annual concentrations in rural areas in France observed in a recent study based on the same exposure model (Pascal et al., 2016).

The RR applied in H4 was equal to 1.18 (95% CI: 0.96–1.46) per 5μg/m³increase in PM2.5, corresponding to 1.40 (95% CI: 0.92–2.13) for a 10μg/m³increase in PM2.5. It was obtained from a meta-analysis from ESCAPE project including 14 cohorts from eight European countries (Sweden, Norway, Denmark, the Netherlands, UK, Austria, Italy, and Greece) (Raaschou-Nielsen et al., 2013). The estimate of the meta- analysis was adjusted for age, sex, calendar time, smoking status, smoking intensity, square of smoking intensity, smoking duration, time

(5)

since quitting smoking, environmental tobacco smoke, occupation, fruit intake, marital status, educational level, employment status, and area- level socioeconomic status.

Theﬁfth hypothesis (H5) was a combination of H3 and H4. The hypotheses H3, H4, and H5 were applied to the main model, and to the models at the department and country level. Data management and analyses were performed in Stata version 13.0. QGIS software version 2.8 was used for analysing spatial data and cartography.

3. Results

3.1. Fine particulate matter exposure

The exposure to PM2.5was assessed for 38.3 million inhabitants aged 30 years and above and living in a total of 36,250 municipalities, or 49,800 neighbourhoods.Fig. 2describes the distribution of the population-weighted exposure to PM2.5in 2005. The 5th, 50th and 95th percentiles of the population-weighted PM2.5levels in France were 8.2, 13.8, 21.8μg/m³, respectively. The population-weighted annual average exposure was 14.4μg/m³. Approximately 87% of the study population was exposed above the yearly average WHO guideline value of 10μg/m³, while 99% was above the 4.9μg/m³value.

3.2. Lung cancer cases attributable to PM2.5under the main hypotheses The total number of lung cancer cases in subjects aged 30 or more was estimated to be 40,451 (29,098 in men; 72% of total) in France in 2015. The estimated fraction of lung cancer cases attributable to PM2.5

in France in 2015 based on the main model was 3.6% of lung cancer incidence, corresponding to 1466 cases (Table 1).

3.3. Sensitivity analyses

The sensitivity analyses are depicted inTable 1. Under hypothesis H1, which assumed spatially homogeneous PM2.5exposure at various administrative scales, compared to the main model, the fraction and the number of attributable lung cancer cases diﬀered in relative value by about 0.4% (overestimation under H1, when PM2.5level was assumed homogeneous within each department) and -11% (underestimation, assuming homogeneity of PM2.5level at the country level). Population density information (H2) had a larger impact on estimates, with an underestimation of the number of attributable cases by 34% at the department level and 72% at the country level if raw PM2.5 concentration was applied for the calculations.

The PAF and the number of attributable lung cancer cases were estimated to be 7.6% and 3065 cases, respectively, if the reference level for PM2.5was changed to 4.9μg/m³(H3). If the alternative RR of 1.40 per 10μg/m³increase in PM2.5was used (H4), then the PAF and the number of attributable lung cancer cases were estimated to be 12.9%

and 5232 cases, respectively. Under the assumption of combining the alternative reference level of PM2.5with the alternative RR (H5), the PAF and the number of attributable lung cancer cases were even higher, reaching 26.0% (or 10,510 cases).

4. Discussion

4.1. Summary of mainﬁndings

This nation-wide study indicates that the proportion of lung cancer cases in France in 2015 attributable to PM2.5exposure could be between 1% (416 cases) and 26.5% (10,733 cases), depending on the spatial resolution PM2.5 exposure and on the assumptions made regarding the dose-response function and the reference level of PM2.5. Based on the main model estimates, about 1500 lung cancer cases could be attributable to PM2.5exposure, representing approximately 4% of all lung cancer cases in France in 2015. Whereas decreasing the spatial distribution of PM2.5exposure underestimated the number of attributable lung cancer cases by 11%, the spatial resolution of information on population density proved to be of high importance to obtain reliable estimates. Indeed, the PAF and the number of attributable lung cancer cases were underestimated by about 34% (department level) and 72%

(country level) when population density distribution was neglected in terms of PM2.5levels. Additionally, we conﬁrmed that the PAF estimate was highly sensitive to the choice of the concentration-response function and the reference level of PM2.5.

4.2. Underlying assumptions and entry parameters

This study has several limitations. The PM2.5concentration was assessed based on the Gazel-Air model (Bentayeb et al., 2014). This model is derived using the CHIMERE chemistry-transport model, mesh refinement and data assimilation with geostatistical analyses and provides information on exposure to PM2.5at a 2 × 2 km grid, which is a much finer spatial scale than that provided by monitoring stations (Bentayeb et al., 2014). This model was the best available model of air pollution in France for the year 2005. Although the Gazel-Air model coupled to neighbourhood-scale population density provides a fine spatial resolution for a nationwide exposure assessment, it has been suggested that models which do not take into account spatial contrasts in exposure at a local (street) scale could underestimate the health impact of air pollution (Morelli et al., 2016). Therefore, in this regard, and also considering that our results confirm that further degrading the spatial resolution of the air pollution model leads to lower PAF, our estimates may be biased towards zero. New modelling techniques exist at regional scales and allow a better characterization of the spatial contrasts in the exposure to air pollution (Atmo Auvergne-Rhône-Alpes, 2015). Future studies may possibly rely on such improved models when they become available at a national scale. The dose-response function is based on findings from epidemiological studies. Most of them were usingfixed monitoring stations to assess PM2.5levels. Finally,Bentayeb et al. (2014) have shown that Gazel-Air model may underestimate PM2.5levels compared to monitoring stations, leading again to under- estimating the PAF.

The PAF estimate strongly depends on other input data used for the analysis. In particular, the choice of the PM2.5concentration-response function markedly inﬂuences the results (see the sensitivity analysis H4). For the main analysis, we applied a RR from a meta-analysis by Hamra et al. that was based on international studies (Hamra et al., 2014). We used a robust RR with a low heterogeneity of the included studies (I²= 53% with p= 0.01). The meta-estimate combined Fig. 2.Population-weighted exposures to PM2.5estimated in France in 2005,

among adults aged 30 years or more. Estimates combine Gazel-Air model with ﬁne-scale population density data.

1082

(6)

incidence and mortality studies due to the high fatality rate among incidence lung cancer cases. Most of the information was obtained from cohort studies. The RR was based on 14 studies from mainly high income countries in Europe, North America and Japan, but also one in China. Some of these studies were adjusted for sex, smoking status and socioeconomic position measured by income, education or occupation (Hamra et al., 2014). Yet when comparing to other studies, our sensitivity analyses call for careful examination of the assumptions on the RR. The RR used for the sensitivity analysis was based on a meta-analysis of 14 cohorts from eight European countries adjusted for the fol- lowing confounders: age, sex, calendar time, smoking status, smoking intensity, square of smoking intensity, smoking duration, time since quitting smoking, environmental tobacco smoke, occupation, fruit intake, marital status, education level, employment status and area-level socioeconomic status (Raaschou-Nielsen et al., 2013). Advantage of this RR is that it is based on standardized exposure assessment (land use regression), includes only lung cancer incidence and was adjusted for a large number of potential confounders. This study was included in the meta-analysis byHamra et al. (2014)contributing only by 1.2% to the overall meta-estimate due to small population size.

Prevalence of tobacco smoking is still high in France, and corresponded to 30% of the adult population in 2005, ten years before the assessment of cancer incidence (Beck et al., 2007). Even though smoking is the main risk factor for lung cancer,Turner et al. (2011) have reported positive associations between mean long-term PM2.5

concentrations and lung cancer mortality in lifelong never-smokers, suggesting that outdoor air pollution plays an important role in development of lung cancer. However, given the strong relationship between cigarette smoking and lung cancer risk, a possible residual confounding by smoking could be of a concern if the RR applied in the PAF estimation was not adjusted for smoking behaviour. The RR from the meta-analysis used in the main model was based on 14 studies out of which 11 adjusted for smoking status. By restricting the meta-analysis only to the studies that adjusted for smoking as a confounder, the overall RR was 1.10 (95%-CI: 1.04–1.17) instead of 1.09 (95%-CI:

1.04–1.14). The RR among never smokers was even higher (1.18; 95%- CI: 1.00–1.39) (Hamra et al., 2014). This shows that if there is any smoking-related confounding bias in the RR of lung cancer associated with PM2.5 that we used, it leads to underestimation, and not

overestimation, of PM2.5attributable fraction.

No information on national lung cancer incidence was available withfine spatial distribution. We instead used lung cancer mortality data by municipality to derive the PAF. As the prognosis of lung cancer remains poor in France, with an overall 5-year survival of only about 15% (Cowppli-Bony et al., 2016), mortality level from lung cancer is assumed to be similar to incidence. Finally, for data confidentiality purposes, it was impossible to obtain mortality data at afiner scale than the municipality. Even at this level, confidentiality reason prevents dissemination of data in area where death counts were below 5 deaths.

Therefore, we had to impute the missing lung cancer deaths, which however represented only 3% of the total number of lung cancer deaths in France.

4.3. Interpretation

This study was unique in design since it utilised national data of PM2.5concentrations at aﬁne spatial scale in combination with lung cancer data at the ﬁnest scale available, namely the municipality, coupled to population distribution data at the neighbourhood level in France. We estimated that about 1500 lung cancer cases were attributable to PM2.5exposure in France under the assumption that PM2.5

reference level is 10μg/m³. This number of lung cancer cases represents about 4% of all new lung cancer cases in France in 2015. A recent GBD study provided PAF estimates for lung cancer deaths due to PM2.5ap- plying uniform distribution between 2.4μg/m³and 5.9μg/m³ as reference level of PM2.5(GBD 2015 Risk Factors Collaborators, 2016). For France, the GBD study estimated that about 2500 lung cancer deaths (or 7.4% of total) were due to ambient particulate matter pollution. The results from the H3 sensitivity analysis may be compared to the GBD study as the reference level of PM2.5under this hypothesis was 4.9μg/

m³similar to the reference level applied in the GBD. Under this H3, we estimated that 7.6% of lung cancer cases could be attributable to PM2.5

in France in 2015, a value very close to the GBD estimate for deaths. For neighbouring countries, the fraction of lung cancer deaths attributable to PM2.5was estimated to be 9.4% in Belgium, 8.4% in Germany, 7.7%

in Switzerland, 11.6% in Italy, 5.6% in Spain, and 7.3% in the United Kingdom (GBD 2015 Risk Factors Collaborators, 2016). A recent study in Alberta, Canada, where air pollution levels are much lower than in Table 1

Population attributable fraction (PAF) and number of lung cancer cases attributable toﬁne particulate matter (PM_2.5) with the corresponding 95% conﬁdence intervals (95% CI), metropolitan France, ages 30 years and more, for the year 2015.

Hypothesis PM2.5exposure:

5th–50th–95th percentiles (μg/m³)

PAF (%) (95% CI) Number of attributable lung cancer cases (95% CI) Relative diﬀerence compared to main model (%)

Main model 8.2–13.8–21.8 3.6 (1.7–5.4) 1466 (679–2193) –

Sensitivity analyses

H1: median of population-weighted PM2.5exposure

(a) Department 9.7–13.8–19.1 3.6 (1.7–5.4) 1471 (680–2203) 0.4

(b) Country 13.8–13.8–13.8 3.2 (1.5–4.9) 1303 (598–1965) −11.1

H2: median of raw PM2.5concentration

(a) Department 6.0–11.1–16.4 2.4 (1.1–3.6) 964 (445–1446) −34.2

(b) Country 11.2–11.2–11.2 1.0 (0.5–1.6) 416 (190–631) −71.6

H3: alternative reference level for PM2.5(4.9μg/m³)

(a) Neighbourhood 8.2–13.8–21.8 7.6 (3.5–11.2) 3065 (1431–4550) 109.1

(b) Department 9.7–13.8–19.1 7.7 (3.6–11.5) 3120 (1455–4634) 112.8

(c) Country 13.8–13.8–13.8 7.4 (3.4–11.0) 2987 (1388–4453) 103.7

H4: alternative RR (1.40 per 10μg/m³increase in PM2.5)

(a) Neighbourhood 8.2–13.8–21.8 12.9 (0.2–25.3) 5232 (78–10,221) 256.8

(b) Department 9.7–13.8–19.1 13.1 (0.1–25.8) 5287 (27–10,429) 260.6

(c) Country 13.8–13.8–13.8 12.0 (0.0–25.0) 4855 (0–10,102) 231.1

H5: combination H3 and H4

(a) Neighbourhood 8.2–13.8–21.8 26.0 (0.0–47.9) 10,510 (2–19,360) 616.8

(b) Department 9.7–13.8–19.1 26.5 (0.0–49.1) 10,733 (0–19,845) 632.1

(c) Country 13.8–13.8–13.8 25.9 (0.0–49.0) 10,468 (0–19,813) 614.0

(7)

France, estimated that about 2%–6% of incident lung cancer cases in 2012 may be attributable to PM2.5 exposure applying diﬀerent reference levels of PM2.5 (7.5μg/m³ and 3.18μg/m³, respectively) (Poirier et al., 2017). Using the same reference level of PM2.5(7.5μg/

m³) as the Canadian study, 5.1% of lung cancer mortality was estimated in South Africa in 2000 (Norman et al., 2007). The results of our sensitivity analyses also show that, as expected, the PAF is larger when a lower level of PM2.5is set as the reference.

Commercial, institutional and household energy use and supply, road transport, and industrial processes are the main contributors to PM2.5 emissions in France accounting for about 90% of all PM2.5

emissions. To improve air quality in France, public policies should focus on interventions reducing the emission from these particular sources (European Environment Agency, 2014). International literature tends to show that public policies and interventions can improve air quality with consequently wide-ranging health beneﬁts. Historically, introdu- cing a ban on coal sales in Dublin in 1990 was reported to lead to an immediate drop in black smoke (Clancy et al., 2002). The collapse of industrial and agricultural structure in East Germany after German re- uniﬁcation in 1990 and the subsequent conversion from brown coal to natural gas were associated with remarkable decreases in pollutant emissions between 1989 and 1991 (Ebelt et al., 2001). A ban on re- sidential wood burning in two urban areas of California in 2003 contributed to a decline in PM2.5concentrations in the range of 13%–14%

(Lighthall et al., 2009). Traﬃc related initiatives such as improvements to bus networks and walking and cycling schemes in London in 2003 and in Stockholm in 2006 led to a decrease in PM10concentrations of 0.8% (Tonne et al., 2008) and 7.6% (Johansson et al., 2009), respectively.

The implementation of low emission zones, the improvement in engine design and the installation of exhaust gas treatment systems were shown to reduce PM2.5concentrations in several cities worldwide.

For instance, after the introduction of low emission zones, PM2.5con- centrations were reduced by 5% in Copenhagen (Jensen et al., 2011) and PM10 concentrations were reduced by about 10% in Munich (Fensterer et al., 2014). PM2.5concentrations were also positively cor- related with traffic volumes and in particular with the volume of diesel trucks. An ambitious traffic-related regulation policy in Tokyo led to a decrease in PM2.5 concentration by 49.8% between 2001 and 2010 (Hara et al., 2013). Improvements in air quality were also reported during the 1996 Summer Olympic Games in Atlanta (Friedman et al., 2001) and in 2008 Summer Olympic Games in Beijing (Rich et al., 2015). During the Olympic period, Atlanta experienced a 16% reduction in PM10 concentrations, attributed to traffic decline (Friedman et al., 2001). As a condition for hosting the 2008 Olympic Games, the Chinese government agreed to temporarily and substantially improve air quality in Beijing for the Olympics by implementing heightened vehicular emissions standards, by restricting use of vehicles by licence plate number, by relocating and closing industrial facilities, and by stopping construction activities (Zhang et al., 2013). These govern- mental regulations resulted in a reduction by 27% in PM2.5 concentrations (Rich et al., 2012;Zhang et al., 2013).

Individual measurement of exposure to ambient PM2.5is challen- ging in large-scale studies. The projection of future health impacts of air pollution may play an important role in policy discussion. However, consistent methods for implementing such projections have not yet emerged (Likhvar et al., 2015). In addition, studies based on cumulative air pollution exposure over several years taking into account migration and therefore accounting for different exposure to air pollution level over life course will further improve the accuracy of health effects estimates related to the pollutants in the air. The chemical composition of PM2.5is diverse. Although factors and sources of PM more strongly related to specific health outcomes has not been isolated yet (Adams et al., 2015), some components and sources of PM may be more toxic than others. The evaluation of the degree of toxicity for specific chemical components of PM2.5is therefore essential for future assessments

of the impact of outdoor air pollution on the lung cancer incidence.

PM2.5accounts only for a small part of air pollution mixture. Estimating the eﬀect of other pollutants than PM2.5and/or the whole air pollution mixture on lung cancer incidence should be priority in future etiologic studies, so that these estimates can be incorporated in estimations of the attributable fraction such as ours. Challenge for future HIAs is to better understand and describe the uncertainties and public-health implica- tions of the diﬀerent choices made for the exposure modelling and scenarios (Likhvar et al., 2015).

5. Conclusion

This study is theﬁrst nation-wide lung cancer risk assessment associated with outdoor PM2.5exposure in France considering diﬀerent dose-response functions. Based on the main model, about 1500 lung cancer cases (4% of total) could be attributed to PM2.5yearly in France.

Depending on the risk models used, the total number of lung cancer cases associated with PM2.5exposure for the year 2015 ranged from 416 to 10,733. The calculations suggest that of the 40,451 lung cancer cases in France that year, from 1.0% to 26.5% may be attributed to PM2.5

exposure. These results suggest that PM2.5 is one of the main con- trollable causes of lung cancer in non-smokers. Policy action to reduce PM2.5 concentrations in France has a large potential to reduce lung cancer cases. Furthermore, improvements in the quality and spatial coverage of the air pollution data and a deeper understanding of the air pollution mixture are crucial for the investigation of future health impacts. The PAF estimates and the number of attributable lung cancer cases were highly sensitive to the choice of the concentration-response function and the reference level of PM2.5. Taking into account population distribution in future studies estimating the PAF is warranted due to its high impact on the estimates.

Supplementary data to this article can be found online athttps://

doi.org/10.1016/j.envint.2018.09.055.

Funding

This study wasﬁnancially supported by the French National Cancer Institute (INCa), grant number 2015-002.

Conﬂict of interest

No conﬂict of interest declared.

Acknowledgement

We would like to thank the FRANCIM Network for providing cancer incidence data and Mireille Eb from the CépiDc, INSERM for providing lung cancer mortality data by municipality.

References

Adams, K., Greenbaum, D.S., Shaikh, R., van Erp, A.M., Russell, A.G., 2015. Particulate matter components, sources, and health: systematic approaches to testing eﬀects. J.

Air Waste Manag. Assoc. 65, 544–558.

Alberg, A.J., Brock, M.V., Ford, J.G., Samet, J.M., Spivack, S.D., 2013. Epidemiology of Lung Cancer: Diagnosis and Management of Lung Cancer, 3rd ed. Chest 143.

American College of Chest Physicians Evidence-Based Clinical Practice Guidelines, pp. e1S–e29S.

Anenberg, S.C., Horowitz, L.W., Tong, D.Q., West, J.J., 2010. An estimate of the global burden of anthropogenic ozone andﬁne particulate matter on premature human mortality using atmospheric modeling. Environ. Health Perspect. 118, 1189–1195.

Atmo Auvergne-Rhône-Alpes, 2015. Regional exposure to air pollution in the Rhône- Alpes region in 2015. Available from.http://www.air-rhonealpes.fr/ﬁche-carte/

exposition-la-pollution-atmospherique-en-2015, Accessed date: 24 February 2017.

Ballester, F., Corella, D., Pérez-Hoyos, S., Hervás, A., 1996. Air pollution and mortality in Valencia, Spain: a study using the APHEA methodology. J. Epidemiol. Community Health 50, 527–533.

Beck, F., Guilbert, P., Gautier, A., 2007. Baromètre santé 2005: attitudes et comporte- ments de santé. Editions Inpes.

Bentayeb, M., Stempfelet, M., Wagner, V., Zins, M., Bonenfant, S., Songeur, C., Sanchez,

1084

(8)

O., Rosso, A., Brulfert, G., Rios, I., Chaxel, E., Virga, J., Armengaud, A., Rossello, P., Rivière, E., Bernard, M., Vasbien, F., Deprost, R., 2014. Retrospective modeling outdoor air pollution at aﬁne spatial scale in France, 1989–2008. Atmos. Environ.

92, 267–279.

Bentayeb, M., Wagner, V., Stempfelet, M., Zins, M., Goldberg, M., Pascal, M., Larrieu, S., Beaudeau, P., Cassadou, S., Eilstein, D., Filleul, L., Le Tertre, A., Medina, S., Pascal, L., Prouvost, H., Quenel, P., Zeghnoun, A., Lefranc, A., 2015. Association between long- term exposure to air pollution and mortality in France: a 25-year follow-up study.

Environ. Int. 85, 5–14.

Carugno, M., Consonni, D., Randi, G., Catelan, D., Grisotto, L., Bertazzi, P.A., Biggeri, A., Baccini, M., 2016. Air pollution exposure, cause-speciﬁc deaths and hospitalizations in a highly polluted Italian region. Environ. Res. 147, 415–424.

Clancy, L., Goodman, P., Sinclair, H., Dockery, D.W., 2002. Eﬀect of air-pollution control on death rates in Dublin, Ireland: an intervention study. Lancet 360, 1210–1214.

Cohen, A.J., Ross Anderson, H., Ostro, B., Pandey, K.D., Krzyzanowski, M., Kunzli, N., Gutschmidt, K., Pope, A., Romieu, I., Samet, J.M., Smith, K., 2005. The global burden of disease due to outdoor air pollution. J. Toxicol. Environ. Health A 68, 1301–1307.

Cowppli-Bony, A., Uhry, Z., Remontet, L., Guizard, A., Voirin, N., Monnereau, A., Bouvier, A., Colonna, M., Bossard, N., Woronoﬀ, A., Grosclaude, P., 2016. Survie des personnes atteintes de cancer en France métropolitaine, 1989–2013. Partie 1– Tumeurs solides. Institut de veille sanitaire, Saint-Maurice, France.

Ebelt, S., Brauer, M., Cyrys, J., Tuch, T., Kreyling, W.G., Wichmann, H.E., Heinrich, J., 2001. Air quality in postuniﬁcation Erfurt, East Germany: associating changes in pollutant concentrations with changes in emissions. Environ. Health Perspect. 109, 325–333.

European Environment Agency, 2014. Air Pollution Fact Sheet 2014–France. European Environment Agency, Copenhagen, Denmark.

Evans, J., van Donkelaar, A., Martin, R.V., Burnett, R., Rainham, D.G., Birkett, N.J., Krewski, D., 2013. Estimates of global mortality attributable to particulate air pollution using satellite imagery. Environ. Res. 120, 33–42.

Ezzati, M., Lopez, A.D., Rodgers, A., Vander Hoorn, S., Murray, C.J., 2002. Comparative Risk Assessment Collaborating Group, Selected major risk factors and global and regional burden of disease. Lancet 360, 1347–1360.

Fattore, E., Paiano, V., Borgini, A., Tittarelli, A., Bertoldi, M., Crosignani, P., Fanelli, R., 2011. Human health risk in relation to air quality in two municipalities in an in- dustrialized area of Northern Italy. Environ. Res. 111, 1321–1327.

Fensterer, V., Kuchenhoﬀ, H., Maier, V., Wichmann, H.E., Breitner, S., Peters, A., Gu, J., Cyrys, J., 2014. Evaluation of the impact of low emission zone and heavy traﬃc ban in Munich (Germany) on the reduction of PM10in ambient air. Int. J. Environ. Res.

Public Health 11, 5094–5112.

Filleul, L., Rondeau, V., Vandentorren, S., Le Moual, N., Cantagrel, A., Annesi-Maesano, I., Charpin, D., Declercq, C., Neukirch, F., Paris, C., Vervloet, D., Brochard, P., Tessier, J.F., Kauﬀmann, F., Baldi, I., 2005. Twentyﬁve year mortality and air pollution:

results from the French PAARC survey. Occup. Environ. Med. 62, 453–460.

Friedman, M.S., Powell, K.E., Hutwagner, L., Graham, L.M., Teague, W.G., 2001. Impact of changes in transportation and commuting behaviors during the 1996 Summer Olympic Games in Atlanta on air quality and childhood asthma. JAMA 285, 897–905.

GBD 2015 Risk Factors Collaborators, 2016. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 388, 1659–1724.

Hamra, G.B., Guha, N., Cohen, A., Laden, F., Raaschou-Nielsen, O., Samet, J.M., Vineis, P., Forastiere, F., Saldiva, P., Yorifuji, T., Loomis, D., 2014. Outdoor particulate matter exposure and lung cancer: a systematic review and meta-analysis. Environ.

Health Perspect. 122, 906–911.

Hara, K., Homma, J., Tamura, K., Inoue, M., Karita, K., Yano, E., 2013. Decreasing trends of suspended particulate matter and PM2.5 concentrations in Tokyo, 1990–2010. J.

Air Waste Manage. Assoc. 63, 737–748.

HIA, 1999. Health Impact Assessment: Main Concepts and Suggested Approach.

Gothenburg Consensus Paper, December 1999. European Centre for Health Policy, WHO Regional Oﬃce for Europe, Brussels.

INSEE, 2016INSEE, n.d. IRIS - Deﬁnition. Available from:https://www.insee.fr/en/

metadonnees/deﬁnition/c1523, Accessed date: 22 February 2017.

Inserm/CépiDc, 2016Inserm/CépiDc, n.d. Interrogation des données sur les causes de décès de 1979 à 2013. Available at:http://www.cepidc.inserm.fr/inserm/html/

index2.htm, Accessed date: 27 June 2016.

Jensen, S.S., Ketzel, M., Nøjgaard, J.K., Becker, T., 2011. What are the Impacts on Air Quality of Low Emission Zones in Denmark? Proceedings from the Annual Transport Conference at Aalborg University. Available at:http://www.traﬁkdage.dk/papers_

2011/31_SteenSolvangJensen.pdf, Accessed date: 7 March 2017.

Johansson, C., Burman, L., Forsberg, B., 2009. The eﬀects of congestions tax on air quality and health. Atmos. Environ. 43, 4843–4854.

Künzli, N., Kaiser, R., Medina, S., Studnicka, M., Chanel, O., Filliger, P., Herry, M., Horak Jr., F., Puybonnieux-Texier, V., Quénel, P., Schneider, J., Seethaler, R., Vergnaud, J.C., Sommer, H., 2000. Public-health impact of outdoor and traﬃc-related air pollution: a European assessment. Lancet 356, 795–801.

Lelieveld, J., Evans, J.S., Fnais, M., Giannadaki, D., Pozzer, A., 2015. The contribution of outdoor air pollution sources to premature mortality on a global scale. Nature 525, 367–371.

Lighthall, D., Nunes, D., Tyner, T., 2009. Environmental Health Evaluation of Rule 4901:

Domestic Wood Burning. A Case Study of the Fresno/Clovis and Bakersﬁeld Metropolitan Areas. Available at:https://www.fresnostate.edu/chhs/cvhpi/

documents/wood-burning-report.pdf, Accessed date: 7 March 2017.

Likhvar, V.N., Pascal, M., Markakis, K., Colette, A., Hauglustaine, D., Valari, M., Klimont, Z., Medina, S., Kinney, P., 2015. A multi-scale health impact assessment of air pollution over the 21st century. Sci. Total Environ. 514, 439–449.

Lim, S.S., Vos, T., Flaxman, A.D., Danaei, G., Shibuya, K., Adair-Rohani, H., Amann, M., Anderson, H.R., Andrews, K.G., Aryee, M., Atkinson, C., Bacchus, L.J., Bahalim, A.N., Balakrishnan, K., Balmes, J., Barker-Collo, S., Baxter, A., Bell, M.L., Blore, J.D., Blyth, F., Bonner, C., Borges, G., Bourne, R., Boussinesq, M., Brauer, M., Brooks, P., Bruce, N.G., Brunekreef, B., Bryan-Hancock, C., Bucello, C., Buchbinder, R., Bull, F., Burnett, R.T., Byers, T.E., Calabria, B., Carapetis, J., Carnahan, E., Chafe, Z., Charlson, F., Chen, H., Chen, J.S., Cheng, A.T., Child, J.C., Cohen, A., Colson, K.E., Cowie, B.C., Darby, S., Darling, S., Davis, A., Degenhardt, L., Dentener, F., Des Jarlais, D.C., Devries, K., Dherani, M., Ding, E.L., Dorsey, E.R., Driscoll, T., Edmond, K., Ali, S.E., Engell, R.E., Erwin, P.J., Fahimi, S., Falder, G., Farzadfar, F., Ferrari, A., Finucane, M.M., Flaxman, S., Fowkes, F.G., Freedman, G., Freeman, M.K., Gakidou, E., Ghosh, S., Giovannucci, E., Gmel, G., Graham, K., Grainger, R., Grant, B., Gunnell, D., Gutierrez, H.R., Hall, W., Hoek, H.W., Hogan, A., Hosgood 3rd, H.D., Hoy, D., Hu, H., Hubbell, B.J., Hutchings, S.J., Ibeanusi, S.E., Jacklyn, G.L., Jasrasaria, R., Jonas, J.B., Kan, H., Kanis, J.A., Kassebaum, N., Kawakami, N., Khang, Y.H., Khatibzadeh, S., Khoo, J.P., Kok, C., Laden, F., Lalloo, R., Lan, Q., Lathlean, T., Leasher, J.L., Leigh, J., Li, Y., Lin, J.K., Lipshultz, S.E., London, S., Lozano, R., Lu, Y., Mak, J., Malekzadeh, R., Mallinger, L., Marcenes, W., March, L., Marks, R., Martin, R., McGale, P., McGrath, J., Mehta, S., Mensah, G.A., Merriman, T.R., Micha, R., Michaud, C., Mishra, V., Mohd Hanaﬁah, K., Mokdad, A.A., Morawska, L., Mozaﬀarian, D., Murphy, T., Naghavi, M., Neal, B., Nelson, P.K., Nolla, J.M., Norman, R., Olives, C., Omer, S.B., Orchard, J., Osborne, R., Ostro, B., Page, A., Pandey, K.D., Parry, C.D., Passmore, E., Patra, J., Pearce, N., Pelizzari, P.M., Petzold, M., Phillips, M.R., Pope, D., Pope 3rd, C.A., Powles, J., Rao, M., Razavi, H., Rehfuess, E.A., Rehm, J.T., Ritz, B., Rivara, F.P., Roberts, T., Robinson, C., Rodriguez-Portales, J.A., Romieu, I., Room, R., Rosenfeld, L.C., Roy, A., Rushton, L., Salomon, J.A., Sampson, U., Sanchez-Riera, L., Sanman, E., Sapkota, A., Seedat, S., Shi, P., Shield, K., Shivakoti, R., Singh, G.M., Sleet, D.A., Smith, E., Smith, K.R., Stapelberg, N.J., Steenland, K., Stockl, H., Stovner, L.J., Straif, K., Straney, L., Thurston, G.D., Tran, J.H., Van Dingenen, R., van Donkelaar, A., Veerman, J.L., Vijayakumar, L., Weintraub, R., Weissman, M.M., White, R.A., Whiteford, H., Wiersma, S.T., Wilkinson, J.D., Williams, H.C., Williams, W., Wilson, N., Woolf, A.D., Yip, P., Zielinski, J.M., Lopez, A.D., Murray, C.J., Ezzati, M., AlMazroa, M.A., Memish, Z.A., 2012. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010.

Lancet 380, 2224–2260.

Loomis, D., Grosse, Y., Lauby-Secretan, B., El Ghissassi, F., Bouvard, V., Benbrahim- Tallaa, L., Guha, N., Baan, R., Mattock, H., Straif, K., 2013. International Agency for Research on Cancer Monograph Working Group, The carcinogenicity of outdoor air pollution. Lancet Oncol. 14, 1262–1263.

Medina, S., Plasencia, A., Ballester, F., Mucke, H.G., Schwartz, J., Apheis, G., 2004.

Apheis: public health impact of PM10 in 19 European cities. J. Epidemiol.

Community Health 58, 831–836.

Morelli, X., Rieux, C., Cyrys, J., Forsberg, B., Slama, R., 2016. Air pollution, health and social deprivation: aﬁne-scale risk assessment. Environ. Res. 147, 59–70.

Norman, R., Cairncross, E., Witi, J., Bradshaw, D., South African Comparative Risk Assessment Collaborating Group, 2007. Estimating the burden of disease attributable to urban outdoor air pollution in South Africa in 2000. S. Afr. Med. J. 97, 782–790.

Orru, H., Teinemaa, E., Lai, T., Tamm, T., Kaasik, M., Kimmel, V., Kangur, K., Merisalu, E., Forsberg, B., 2009. Health impact assessment of particulate pollution in Tallinn usingﬁne spatial resolution and modeling techniques. Environ. Health 8, 7.

Pascal, M., Corso, M., Chanel, O., Declercq, C., Badaloni, C., Cesaroni, G., Henschel, S., Meister, K., Haluza, D., Martin-Olmedo, P., Medina, S., 2013. Assessing the public health impacts of urban air pollution in 25 European cities: results of the Aphekom project. Sci. Total Environ. 449, 390–400.

Pascal, M., de Crouy Chanel, P., Wagner, V., Corso, M., Tillier, C., Bentayeb, M., Blanchard, M., Cochet, A., Pascal, L., Host, S., Goria, S., Le Tertre, A., Chatignoux, E., Ung, A., Beaudeau, P., Medina, S., 2016. The mortality impacts ofﬁne particles in France. Sci. Total Environ. 571, 416–425.

Poirier, A.E., Grundy, A., Khandwala, F., Friedenreich, C.M., Brenner, D.R., 2017. Cancer incidence attributable to air pollution in Alberta in 2012. CMAJ Open 5, E524–E528.

Pope, I.C., Burnett, R.T., Thun, M.J., et al., 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure toﬁne particulate air pollution. JAMA 287, 1132–1141.

Raaschou-Nielsen, O., Andersen, Z.J., Beelen, R., Samoli, E., Stafoggia, M., Weinmayr, G., Hoﬀmann, B., Fischer, P., Nieuwenhuijsen, M.J., Brunekreef, B., Xun, W.W., Katsouyanni, K., Dimakopoulou, K., Sommar, J., Forsberg, B., Modig, L., Oudin, A., Oftedal, B., Schwarze, P.E., Nafstad, P., De Faire, U., Pedersen, N.L., Ostenson, C.G., Fratiglioni, L., Penell, J., Korek, M., Pershagen, G., Eriksen, K.T., Sorensen, M., Tjonneland, A., Ellermann, T., Eeftens, M., Peeters, P.H., Meliefste, K., Wang, M., Bueno-de-Mesquita, B., Key, T.J., de Hoogh, K., Concin, H., Nagel, G., Vilier, A., Grioni, S., Krogh, V., Tsai, M.Y., Ricceri, F., Sacerdote, C., Galassi, C., Migliore, E., Ranzi, A., Cesaroni, G., Badaloni, C., Forastiere, F., Tamayo, I., Amiano, P., Dorronsoro, M., Trichopoulou, A., Bamia, C., Vineis, P., Hoek, G., 2013. Air pollution and lung cancer incidence in 17 European cohorts: prospective analyses from the European Study of Cohorts for Air Pollution Eﬀects (ESCAPE). Lancet Oncol. 14, 813–822.

Rich, D.Q., Kipen, H.M., Huang, W., Wang, G., Wang, Y., Zhu, P., Ohman-Strickland, P., Hu, M., Philipp, C., Diehl, S.R., Lu, S.E., Tong, J., Gong, J., Thomas, D., Zhu, T., Zhang, J.J., 2012. Association between changes in air pollution levels during the Beijing Olympics and biomarkers of inﬂammation and thrombosis in healthy young adults. JAMA 307, 2068–2078.

Rich, D.Q., Liu, K., Zhang, J., Thurston, S.W., Stevens, T.P., Pan, Y., Kane, C., Weinberger, B., Ohman-Strickland, P., Woodruﬀ, T.J., Duan, X., Assibey-Mensah, V., Zhang, J., 2015. Diﬀerences in birth weight associated with the 2008 Beijing Olympics air pollution reduction: results from a natural experiment. Environ. Health Perspect. 123,

(9)

880–887.

Santé publique France, 2010. Surveillance épidémiologique des cancers. Available from:

http://invs.santepubliquefrance.fr/Dossiers-thematiques/Maladies-chroniques-et- traumatismes/Cancers/Surveillance-epidemiologique-des-cancers/Introduction, Accessed date: 25 January 2017.

Tonne, C., Beevers, S., Armstrong, B., Kelly, F., Wilkinson, P., 2008. Air pollution and mortality beneﬁts of the London Congestion Charge: spatial and socioeconomic in- equalities. Occup. Environ. Med. 65, 620–627.

Turner, M.C., Krewski, D., Pope 3rd, C.A., Chen, Y., Gapstur, S.M., Thun, M.J., 2011.

Long-term ambientﬁne particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am. J. Respir. Crit. Care Med. 184, 1374–1381.

WHO Regional Oﬃce for Europe, 2013. Review of Evidence on Health Aspects of Air Pollution - REVIHAAP Project. Technical Report. WHO Regional Oﬃce for Europe, Copenhagen.

World Health Organization, 2005. WHO Air Quality Guidelines for Particulate Matter, Ozone, Nitrogen Dioxide and Sulfur Dioxide. Available at:http://apps.who.int/iris/

bitstream/10665/69477/1/WHO_SDE_PHE_OEH_06.02_eng.pdf, Accessed date: 23 March 2016.

World Health Organization, 2016. Ambient Air Pollution: A Global Assessment of Exposure and Burden of Disease. World Health Organization, Geneva, Switzerland.

Yanagi, Y., Assuncao, J.V., Barrozo, L.V., 2012. The Impact of Atmospheric Particulate Matter on Cancer Incidence and Mortality in the City of Sao Paulo, Brazil, Cad Saude Publica. 28. pp. 1737–1748.

Zhang, J., Zhu, T., Kipen, H., Wang, G., Huang, W., Rich, D., Zhu, P., Wang, Y., Lu, S.E., Ohman-Strickland, P., Diehl, S., Hu, M., Tong, J., Gong, J., Thomas, D., HEI Health Review Committee, 2013. Cardiorespiratory biomarker responses in healthy young adults to drastic air quality changes surrounding the 2008 Beijing Olympics. Res. Rep.

Health Eﬀ. Inst. 5–174.

1086