Article Text

Original article
Occupational exposure to lead and lung cancer: results from two case-control studies in Montreal, Canada
  1. Willy Wynant1,2,
  2. Jack Siemiatycki1,3,
  3. Marie-Élise Parent1,3,4,
  4. Marie-Claude Rousseau1,3,4
  1. 1Research Centre of the CHUM, Montreal, Quebec, Canada
  2. 2Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Quebec, Canada
  3. 3Department of Social and Preventive Medicine, Faculty of Medicine, University of Montreal, Montreal, Quebec, Canada
  4. 4INRS-Institut Armand-Frappier, University of Quebec, Laval, Quebec, Canada
  1. Correspondence to Dr Marie-Claude Rousseau, INRS-Institut Armand-Frappier, Université du Québec, 531 boul. des Prairies, Laval, QC H7V 1B7, Canada; marie-claude.rousseau{at}iaf.inrs.ca

Abstract

Objectives We investigated the association between workplace lead exposure and lung cancer risk, separately for organic lead and for inorganic lead, from either engine emissions or from other sources.

Methods Two population-based case-control studies were carried out in Montreal (1979–1986 and 1996–2002) to investigate occupational factors in relation to lung cancer among 1593 men with histologically confirmed incident lung cancer, and 1426 controls from the general population. Interviews elicited information on sociodemographic characteristics, lifetime smoking and occupational history. Chemists translated each job into potential chemical exposures. Cumulative indices of exposure were derived and classified into non-substantial and substantial exposure. ORs adjusted for several potential confounders including smoking, and 95% CIs were estimated by logistic regression.

Results Lifetime prevalences of exposure in Study I were 3% for organic lead, 40% for inorganic lead from engine emissions and 17% for inorganic lead from other sources; corresponding prevalences in Study II were 4%, 19% and 16%, respectively. No associations were observed when comparing ever to never exposed subjects in pooled analyses (organic lead, OR=1.39, 95% CI 0.77 to 2.52; inorganic lead from engine emissions: OR=0.89, 95% CI 0.72 to 1.09; inorganic lead from other sources: OR=0.99, 95% CI 0.76 to 1.29). Nor were these exposures associated with lung cancer in subjects with substantial cumulative exposure.

Conclusions In this large study, using a blinded expert-based assessment of lifetime occupational exposure and adjustment for several potential confounders, we observed no increased risk of lung cancer with exposure to lead compounds.

  • inorganic lead
  • organic lead
  • lung cancer
  • case-control study
  • occupational exposure

Statistics from Altmetric.com

What this paper adds

  • There is uncertainty concerning the carcinogenic potential of lead compounds.

  • In 2006, the International Agency for Research on Cancer classified inorganic lead as a probable carcinogen and judged organic lead as unclassifiable based on limited and inadequate evidence on humans.

  • Among men exposed through various occupations, we found no excess risk of lung cancer associated with occupational exposure to organic lead or to inorganic lead either from engine emissions or other sources.

Introduction

Lead was one of the first metals exploited by man; it has been used for at least 6000 years.1 The properties which have encouraged its historic and ongoing usefulness include: softness, malleability, ductility, poor conductivity, resistance to corrosion and resistance to various chemicals.1 One of the most common metals found in the general environment, environmental lead contamination mainly results from human activity.2 In adults, lead toxicity has been demonstrated among highly exposed workers in a variety of occupations and industries such as refining industries, battery manufacturing plants, construction, painting and printing.2 ,3 Lead compounds have been suspected to be carcinogenic at least since the 1950s;4 the suspicion has focussed mainly on cancers of the lung, stomach, kidney, brain and nervous system as the target organs. In 2006, the International Agency for Research on Cancer classified inorganic lead as a probable carcinogen with limited evidence for humans and sufficient evidence for animals, and organic lead as unclassifiable with inadequate evidence for both humans and animals.2 The most convincing evidence on inorganic lead was for the lung and stomach cancers. However, evidence among humans was considered to be limited even in these two cases.5 The Environmental Protection Agency through its Integrated Risk Information System, also classified inorganic lead as a probable human carcinogen.6 Nevertheless, the relationship between lead exposure and lung cancer risk remains controversial. While an association has been observed in several studies,2 it has been hypothesised that this could be due to residual confounding by arsenic exposure or smoking habits.7–10 Because of the widespread usage and environmental persistence of lead compounds, it is particularly important to establish more definitively whether such exposures entail cancer risk, with a particular concern about lung cancer risk.

Lung cancer is the most common cause of cancer mortality, responsible for over one million deaths worldwide each year.11 In industrialised nations, the contribution of occupational exposures to the lung cancer burden may be small compared with that of cigarette smoking, but it is large when compared with the contributions of most other risk factors.12 In particular, a number of occupational exposures are known to be associated with this tumour, including asbestos, radon, inorganic arsenic, chromium, nickel and polycyclic aromatic hydrocarbons.13 The present report, carried out in the context of two large population-based case-control studies in Montreal, Canada, aimed to provide new evidence regarding the possible association between workplace exposure to lead compounds and lung cancer risk. A unique feature, in addition to the ability to exercise good control over potential confounders, is our ability to distinguish the effects of three important classes of lead compounds: organic lead, inorganic lead resulting from engine emissions, and inorganic lead from other sources.

Methods

Population and design

This study incorporates two population-based case-control studies carried out in Montreal, Quebec, Canada. The methodology of these studies has already been described.14 ,15 The first one, Study I, was conducted from 1979 to 1986 and included only men aged between 35 and 70 years and diagnosed with any one of about 20 cancer types considered. The second one, Study II, was conducted from 1996 to 2002 and included both men and women aged from 35 to 75 years old, and diagnosed with lung cancer. The present report is restricted to men because the total number of women in Study II was relatively small (465 cases), and their lifetime prevalence of occupational exposure to each category of lead compounds ranged from less than 1% to 2.5%.

In both studies, the cases were primary, incident and histologically confirmed lung cancers, identified from all the major Montreal area hospitals providing almost complete coverage (estimated at 97%) of cancer cases diagnosed. Among the 1082 eligible patients with histologically confirmed primary incident lung cancer in Study I, 857 (79%) completed the interview. Among the 857 eligible cases of men from Study II, 738 (86%) completed the interview.

The population controls were randomly selected from the electoral list for the Province of Quebec, which is maintained by an active ongoing registration process, and is considered to be an almost complete list of citizens of voting age. In Study I, controls were frequency-matched by age and electoral district—comprised of about 40 000 electors—to cancer cases, and 533 (72%) of the 740 eligible population controls completed the interview. In Study II, controls were frequency—matched to cases according to the distribution of age, sex and electoral district, and 899 (70%) out of the 1294 eligible male population controls completed the interview. Given that the age distribution of cases in Study II was roughly similar to that of cases in Study I, and that both studies shared the same procedures, we pooled the data, after verifying that the results did not differ.

Ethics committees at all participating institutions approved the study protocols, and all participants provided informed consent.

Data collection

Face-to-face interviews were conducted with each participating subject. If the subject could not respond, a surrogate respondent provided information (18% in Study I, and 24% in Study II, for men). Interviews were divided in two parts. First, a structured interview allowed collecting sociodemographic characteristics, lifestyle information (dietary habits, smoking, etc) and the medical history of the subject. Then, a semistructured interview was conducted in order to obtain a detailed description of each job held by the subject: the company where the subject worked, the products used, the nature of the worksite, the subject's tasks and the equipment used.

A team of chemists and industrial hygienists translated each job into potential exposures from a list of almost 300 substances, without knowledge of the subject's case or control status. The exposure codes attributed to each participant were based on consensus among the coders. For each exposure thought to be present in any job, they rated three distinct dimensions of information, each of them on a three-point scale. First, they defined the degree of confidence that the exposure really occurred with the following levels: ‘possible’, ‘probable’ and ‘definite’. Then, they estimated the frequency of exposure in a normal workweek as a proportion of the working time. In Study I, this information was rated into pre-established categories (<5%, 5–30%, >30%), whereas in Study II, it was expressed in percentage, on a continuous scale. The team of experts also quantified the relative level of concentration of the substances as ‘low’, ‘medium’ or ‘high’. Finally, the duration of exposure to a given substance during a specific job was considered equivalent to the duration of the job. Non-exposure was interpreted as exposure at a level similar to the level found in the general population environment. Since the interviews elicited detailed and individual information on each job held, different subjects with the same job title could be attributed different exposures and for a given exposure, different levels of confidence, frequency and concentration of exposure. Conversely, subjects with the same exposure profile could have held different occupations. An example of such occupational exposure coding is provided in the appendix given by Parent et al.16

Occupational exposure to lead compounds

Three families of lead compounds were considered: organic lead, inorganic lead from gasoline engine emissions and inorganic lead from other sources. Organic lead includes lead from uncombusted leaded gasoline and aviation gasoline. Tetraethyl lead can be added to gasoline to increase the fuel's octane number, which improves the antiknock characteristics of the fuel in spark ignition engines. About 70–75% of this lead is transformed into inorganic lead in vehicles’ engines upon combustion and emitted to the atmosphere through the exhaust pipe along with 1% of the organic lead that passes through the engine unchanged. The rest of the lead remains trapped within the exhaust system. Although lead in gasoline accounts for less than 10% of all refined lead production, about 80–90% of lead in global ambient air comes from combustion of leaded gasoline.17 In Canada, leaded gasoline was phased out from the mid-1970s to the early 1990s.18 Inorganic lead from other sources includes lead fumes, chromates, oxides, dusts and lead from lubricating oils and greases as additives containing lead naphthenate (inorganic lead) is present in lubricating formulations. The two categories of inorganic lead were kept separate because of the distinctive character of lead compounds from gasoline emissions in contrast with the wider variety of other inorganic lead compounds, the very high exposure prevalence to the former compared with the latter, and the fact that the coexposures and potential confounders would be very different between the two categories.

For lead-containing mixtures, such as gasoline, gasoline emissions and lubricating oils and greases, the experts assigned a concentration of exposure for the entire mixture and separately, at a lower level, for the lead compound component. The algorithms they used to assign the concentration level to the lead compound fraction thus had the effect of restricting the distribution to a maximum of: (1) low concentration level for lead in gasoline, aviation gasoline and lubricating oils and greases and (2) medium concentration level for lead in gasoline emissions.

We excluded exposures which began within the 5-year period prior to diagnosis/recruitment as we expected a lag period between lead exposure and lung cancer. Additionally, exposures coded with the lowest confidence levels were excluded from the analyses. These exclusions dropped eight subjects (two cases and six controls in Study II) for whom the experts assessed all lead exposures with a low confidence level that these exposures actually occurred. Finally, for each lead category, we estimated a time-weighted average level over all jobs to divide exposure into a substantial level (concentration and frequency must be at level ‘medium’ or ‘high’, and the total duration of exposure higher than 5 years) and non-substantial level. Non-exposed subjects were those who were not occupationally exposed to lead (any form). For example, participants unexposed to organic lead but exposed to inorganic lead, were excluded from the analyses for the association between organic lead and lung cancer.

Statistical analyses

ORs and 95% CIs were estimated using unconditional multivariate logistic regression. We adjusted for age, educational level, respondent type (proxy/self), ethnocultural origin and smoking habits. We also adjusted for known occupational exposure to lung carcinogens, and for which the prevalence of exposure was at least 3% in the studies: asbestos, silica, arsenic, chromium VI and cadmium. We adjusted for smoking using the Comprehensive Smoking Index, a variable which summarises the information on smoking duration, smoking intensity and time since cessation.19 This allowed us to use only one degree of freedom for modelling smoking, and has been shown to provide superior adjustment for smoking.19 Pooled analyses were conducted adding a binary variable to adjust for study. To avoid residual confounding from age, we used age as a continuous variable using a non-parametric modelling with generalised additive models.20 Analyses were carried out by level of exposure (non-exposed/non-substantial/substantial), and by duration (less or more than 10 years).

For some carcinogens such as tobacco smoke,21 ionising radiation22 and arsenic,23 it is well established that risk to develop lung cancer diminishes with time since last exposure. Therefore, an adjustment for time since last exposure was also considered (eg, years since last exposure included as a continuous variable in the regression model). However, this variable was not associated with lung cancer risk, and it was not included in the final analyses.

Finally, analyses were also conducted separately for the most prevalent histological subtypes of lung cancer in our study: squamous cell carcinoma (38.9% of the cases in the pooled studies), adenocarcinoma (25.6% in the pooled studies) and small cell lung cancer (17.8% in the pooled studies).

Results

Selected characteristics of the study groups are presented in table 1. Cases and controls were of similar ages, reflecting the frequency matching, although participants were slightly older in Study II. The majority of participants were of French Canadian ethnocultural origin. In both studies, controls had attained a higher educational level than cases. As expected, cases were more likely to be smokers than controls in both studies. Reflecting population trends in smoking behaviour, there were fewer current and more former smokers among controls in Study II than in Study I. The proportion of proxy respondents among cases was higher in Study II. Finally, exposure to inorganic lead from engine emissions was half as common in Study II as compared with Study I, most likely reflecting the increased use of unleaded gas in more recent time periods.

Table 1

Distribution of selected characteristics among participants from two case-control studies conducted in Montreal, QC, Canada

The main occupations to which exposure to lead was attributed, and the most typical exposure coding per occupation for Study I have been published previously,24 and those for Study II are shown in table 2. Only occupational groups which contributed ≥5% of exposed subjects in any category of lead exposure are presented. Subjects exposed to organic lead were principally exposed if they were mechanics or repairmen. Those exposed through occupations in sales and services either worked in garages with a gas station and reported contact with gasoline, or mentioned using gasoline to clean mechanical parts. Subjects exposed in metal shaping and forming occupations reported using gasoline as a grease remover in order to clean hands or metal parts. Exposure to organic lead was attributed in such situations solely when the occupation was held before the phasing out of leaded gasoline. Those exposed to inorganic lead resulting from engine emissions were exposed through employment as mechanics or in forestry and logging occupations. In general, the usual concentration of exposure was defined as ‘low’. Subjects exposed to inorganic lead from other sources were more likely to be exposed via protective service occupations (policemen, firefighters, security guards) and electrical or electronics occupations. Painting was not a predominant occupation resulting in exposure to inorganic lead; when a painter was considered exposed, it was usually at ‘high’ frequency, whereas frequency was ‘low’ for the protective service occupations.

Table 2

Main occupations for which lead compounds exposures were coded (examples from Study II)

ORs for the associations between ever exposure to lead compounds and lung cancer are presented in table 3. Despite some variations, ORs did not differ between the two studies, as judged by interaction terms between the variables ‘study indicator’ and ‘ever exposure to organic lead’ (p=0.34), ‘ever exposure to inorganic lead resulting from engine emissions’ (p=0.63), and ‘ever exposure to inorganic lead from other sources’ (p=0.51). Neither in the individual studies nor in the pooled analysis did any OR depart from the null in any remarkable way; the overall pattern is compatible with an absence of any excess risks. The only possible exception was a point estimate OR of 1.4 (95% CI 0.8 to 2.5) for ever exposure to organic lead, but exposure level could not be explored as all exposures to organic lead were at the non-substantial level.

Table 3

ORs for association of occupational exposure to lead with lung cancer, according to lifetime level of exposure

ORs by duration of exposure, less than 10 years or 10 years or more, are shown in table 4. In the pooled analysis, none of these ORs was statistically significant, though there was an elevated point estimate of 1.9 for ≥10 years exposure to organic lead compounds. Once again, ORs did not differ between the two studies, as judged by the interaction terms in the models (data not shown).

Table 4

ORs for association of occupational exposure to lead with lung cancer, according to lifetime duration of exposure

Overall, there were no statistically significant associations between lead exposure and any of the histological subtypes of lung cancer when comparing ever exposed to non-exposed (table 5). We observed a non-significant slight increased risk of lung cancer for those exposed to organic lead compared with those not exposed to organic lead for squamous cell carcinoma.

Table 5

ORs for association of occupational exposure to lead compounds with lung cancer, according to histological group

Sensitivity analyses were conducted without adjusting for educational level. The unadjusted ORs did not differ from those adjusted for educational level (see data on web supplemental material). More particularly, although generally slightly further from the null value, none of the unadjusted ORs was statistically significant.

Discussion

We found no excess risk of lung cancer associated with occupational exposure to organic lead and to inorganic lead either from engine emissions or other sources. These results are consistent with some previous analyses.24–27 By contrast, excesses of lung cancer have been reported at plants producing lead chromate pigments in the USA,28 Germany and The Netherlands.29 In 2000, Steenland and Boffetta summarised results from eight studies of workers highly exposed to lead.30 Two of the eight studies showed an increased risk of lung cancer (at least 50%). Combining the eight studies, the risk was about 30% higher among exposed individuals as compared with non-exposed. However, that combined result depended heavily on one study where a threefold excess risk of lung cancer was found, and in which workers may have been exposed to both lead and arsenic. Since arsenic is known to cause lung cancer, it is not clear whether the increase in lung cancer was due to lead, arsenic, or a combination of the two. Excluding this study, all other studies combined result in a risk estimate at around 14% higher among those exposed to lead as compared with those unexposed. Moreover, these studies could not determine whether this increase was because of lead exposure or if it happened because the lead-exposed workers tended to smoke more than the comparison groups, as these studies did not control for smoking.

Evidence regarding the mutagenic, clastogenic and carcinogenic properties of lead remain inconclusive.31 Lead chromate has been consistently found to increase the frequency of chromosome aberrations in in vitro studies, whereas results are almost all negative for lead nitrate or lead glutamate. In a recent updated review on human studies, 15 studies out of 21 reported an increase in the frequency of chromosome aberrations in humans exposed to lead. The in vitro studies reviewed also reported contradictory results about the alterations in sister chromatid exchanges. However, in human studies, most of the results showed an increase in the frequency of sister chromatid exchanges in populations exposed to lead. Overall, several studies have considered populations of workers occupationally exposed to lead, and found evidence of genotoxic activity in the workers’ lymphocytes, a DNA repair deficiency and more DNA breaks.31 More particularly, Wise and colleagues, indicated that lead chromate was cytotoxic and genotoxic to human lung cells.32 Finally, it seems that lead compounds do not appear to cause genetic damage directly, but may do so through several indirect mechanisms.31

As with any epidemiological study, there were several possible sources of error or bias in our results. The numbers of study subjects and the estimated prevalence of exposure were limited and gave rise to imprecise OR estimates. The exposure information was estimated retrospectively by experts, and there was undoubtedly some degree of error in assessing exposures; such error would have been non-differential since the raters were blinded to case/control status. Concentration of exposure could not be estimated in absolute terms; it was only done on an ordinal scale. Only a minority of exposed workers were considered to have been exposed at the highest concentration levels, and it may well be that the exposure levels in such a general population sample are insufficient to detect a risk that would manifest at higher levels. There was not detailed enough exposure data to be able to assess risks in relation to different temporal patterns of exposure. There may have been selection biases if workers who were particularly susceptible to lung cancer and other respiratory diseases could have selected themselves out of exposure after relatively brief employment because they suffered adverse short-term respiratory effects from the dusty conditions.7 Actually, one of lead's primary targets is the nervous system,33 and adults whose work exposes them to lead have been shown to develop nervous system problems.34 While the most important potential confounders were available, some potential risk factors, such as diet, were not available for the whole sample. However, it is not likely that diet was an important confounder. Indeed, it is unlikely to be strongly associated with lead exposure, given the wide range of occupations in which lead exposure was coded. It is hypothetically possible that diet could be an effect modifier, given that lead absorption increases in diets with low levels of calcium;35 however, we were unable to assess such a phenomenon. We also have to note that there were quite high proportions of proxy response, but it was reassuring that the prevalence of exposure was similar between proxy and self-respondents, and that the overall results were similar when we excluded subjects with proxy respondents from the analyses (data not shown). Lastly, statistical power was limited for our analyses by histological subtype of lung cancer, as evident from the wide CIs.

Our study has several strengths. The numbers of subjects, while limited, were still large for a population-based case-control study with detailed exposure assessment. While conducting the study in a general population rather than in a selected high-exposure cohort has the disadvantage of including a large fraction of exposed workers whose exposure levels were relatively low, it has the concomitant advantage of providing representative and generalisable risk estimates for the entire population of workers exposed to lead, rather than a small subset of workers. While exposure information was imperfect, a great deal of time and effort was devoted by technical experts to estimating exposure. Occupational exposure was attributed to subjects on the basis of their detailed lifetime job history reported at the interview. We have previously shown that the subjects’ reports of occupational history were valid.36 Our team of chemists and industrial hygienists attributed exposure to the subjects using a method for which reasonable reliability37 ,38 and validity39 have been shown in our own study population and using our own study personnel. However, as our exposure assessment protocol was based on expert opinion rather than direct measurement or biomarkers, it inevitably entailed some degree of measurement error. Because this work was done blindly with respect to disease status, we assume that any misclassification of the exposure variables under consideration would have occurred non-differentially with respect to the outcome. Our assessment of exposure frequency and concentration was semiquantitative, based on descriptions provided by the subjects and established by expert chemists and industrial hygienists. Although errors in exposure assessment for lead would likely result in an attenuation of risk estimates, errors in exposure assessment for occupational confounders or failure to include true occupational confounders could lead to bias in any direction. Adjustment for potential confounders was more extensive than in previous studies. We collected detailed information on potential confounders, covering sociodemographic and lifestyle factors, including detailed smoking history, as well as other occupational exposures pertinent to lung cancer, such as asbestos, silica, chromium VI, arsenic and cadmium. Whereas it is notoriously difficult to control for smoking in cohorts of workers due to a lack of detailed information on smoking, a case-control study allows for ascertainment of a complete history of cigarette smoking, a potentially crucial confounder of the relation between occupational lead exposure and lung cancer. For the parameterisation of smoking history variables, we used an approach based on a risk model derived from our study subjects.19 Although this is an optimal approach, it is conceivable that some degree of residual confounding remains. Whereas previous studies typically focused on populations with particularly high exposure levels of lead, our population-based case-control studies covered the entire range of workplace exposure circumstances. Our studies also allowed for the integration of lifetime job histories, rather than only focusing on the worker's history with only one employer. The case definition of incident histologically confirmed lung cancer, allowed for the collection of more detailed diagnostic information from medical records than that typically found on death certificates. We also assessed risks in relation to relative levels of exposure. Response rates were quite high, over 80% for case groups and 70% for population control groups. Although the participation proportions were lower among controls than among cases, these proportions were as high as or higher than those in most recent studies,40 thus diminishing the likelihood of selection bias. For all these reasons, results from our case-control studies constitute an important complement to previous studies.

The interpretation of epidemiological studies of cancer risks in lead workers and the carcinogenicity of lead exposures is complicated by the variability and complexity of the various forms of lead present in the work environment, and by the changes in work conditions that have occurred over the last decades. In our study, with the exception of a hint of a possible excess in relation to prolonged exposure to organic lead compounds, we observed no increased risk of lung cancer when comparing exposed to unexposed subjects, or when considering relatively high and prolonged lead exposure. Although inorganic lead has been classified as a probable carcinogen, evidence for carcinogenicity among humans remains limited, and our study does not support a possible association between occupational exposure to lead and lung cancer. Nevertheless, there is some unresolved evidence that lead compounds may cause stomach cancer.2 ,24 Our findings are of course important for workplace health, but they also inform us on environmental health in the general population, as lead compounds find their way into the general environment through various routes.2 ,24

Acknowledgments

The fieldwork was supervised by Lesley Richardson, and the chemical coding was carried out by Dr Michel Gerin, Dr Louise Nadon, Denis Begin, Ramzan Lakhani and Benoît Latreille. This study was supported by research and personnel support grants from the Canadian Cancer Society, the Institut de recherche en santé et securité au travail du Quebec, the Fonds de la recherche en santé du Quebec (FRSQ), the Canadian Institutes of Health Research (CIHR), and the Guzzo-Cancer Research Society Chair in Environment and Cancer. MEP is the recipient of a FRSQ Investigator Award. MCR is the recipient of a CIHR New Investigator Award.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Contributors JS designed data collection tools, monitored data collection, participated to the analysis plan and interpretation of results and revised the paper. M-EP contributed to the analysis plan, participated in the interpretation of the results and revised the paper. M-CR developed the analysis plan with the first author, participated in the interpretation of the results and revised the paper. She is guarantor. WW developed the analysis plan with MCR, analysed the data and drafted and revised the paper.

  • Funding This study was supported by The National Cancer Institute of Canada (grant # 010736 Oc 86565), The National Health Research and Development Program (grant # 6605-4730-800), The Medical Research Council of Canada (grant # 37673 MT 14704), the Canadian Institutes of Health Research (grant # MOP 14704), and the Guzzo-Cancer Research Society Chair in Environment and Cancer.

  • Competing interests None.

  • Ethics approval This study was approved by ethics committees at the Institut Armand-Frappier (University of Quebec), McGill University and each of the 18 hospitals in which cases were ascertained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.