Objectives Established risk factors for leukaemia do not explain the majority of leukaemia cases. Previous studies have suggested the importance of occupation and related exposures in leukaemogenesis. We evaluated possible associations between job title and selected hazardous agents and leukaemia in the European Prospective Investigation into Cancer and Nutrition.
Methods The mean follow-up time for 241 465 subjects was 11.20 years (SD 2.42 years). During the follow-up period, 477 incident cases of myeloid and lymphoid leukaemia occurred. Data on 52 occupations considered a priori to be at high risk of developing cancer were collected through standardised questionnaires. Occupational exposures were estimated by linking the reported occupations to a job exposure matrix. Cox proportional hazard models were used to explore the association between occupation and related exposures and risk of leukaemia.
Results The risk of lymphoid leukaemia significantly increased for working in chemical laboratories (HR 8.35, 95% CI 1.58 to 44.24), while the risk of myeloid leukaemia increased for working in the shoe or other leather goods industry (HR 2.54, 95% CI 1.28 to 5.06). Exposure-specific analyses showed a non-significant increased risk of myeloid leukaemias for exposure to benzene (HR 1.15, 95% CI 0.75 to 1.40; HR=1.60, 95% CI 0.95 to 2.69 for the low and high exposure categories, respectively). This association was present both for acute and chronic myeloid leukaemia at high exposure levels. However, numbers were too small to reach statistical significance.
Conclusions Our findings suggest a possible role of occupational exposures in the development of both lymphoid and myeloid leukaemia. Exposure to benzene seemed to be associated with both acute and chronic myeloid leukaemia.
Statistics from Altmetric.com
What this paper adds
Established risk factors for leukaemias are viral infections by human T-cell leukaemia/lymphotropic virus type 1, exposure to ionising radiation, benzene and alkylating agents, which do not explain most leukaemia cases. Evidence suggests that a broad spectrum of occupational and environmental exposures may increase the risk of leukaemia.
This study showed that working in chemical laboratories and in the shoe or other leather goods industry increased the risk of lymphoid and myeloid leukaemia, respectively. Exposure to benzene seemed to be associated with both acute and chronic myeloid leukaemia.
Our findings suggest a possible role of occupational exposures (solvents) in the development of both lymphoid and myeloid leukaemia.
Leukaemias are a malignancy of the haematopoietic system classified into two major components corresponding to the myeloid and lymphoid cell linage of leukaemias. There were an estimated 350 434 new cases and 257 161 deaths from leukaemia in the year 2008 in the world.1 The most established risk factors for leukaemia are viral infections by human T-cell leukaemia/lymphotropic virus type 1, exposure to ionising radiation and benzene and alkylating agents.2 However, these risk factors do not explain most leukaemia cases, necessitating further research in lifestyle and occupational/environmental risk factors of leukaemia.2
Previous studies have suggested the importance of occupational and environmental risk factors for leukaemias.2–4 It has been estimated that known occupational factors contribute to approximately 10% of leukaemia cases in the USA and Europe.4 For example, leukaemia has been found in excess among workers involved in the manufacturing of petroleum products,5 ,6 rubber,6 ,7 plastic,6 ,8 nuclear energy,6 ,9 munitions,6 electronic manufacturing6 ,10 and agriculture.3 ,8 An increased risk of leukaemias has also been reported among mechanics, iron and metalware workers, launderers, dry cleaners,11 nursing, healthcare workers, janitors, light truck drivers, employers in plumbing, heating and air conditioning,3 welders/flame cutters and telephone line workers10 and workers in abattoirs and meat processing plants.12 ,13 In addition to risks associated with jobs and industries several specific occupational exposures such as benzene,14 ionising radiation,2 pesticides,3 ,15 ,16 formaldehyde17 and hair dye18 have been linked to an excess risk of leukaemia. However, it should be noted that the degree of carcinogenicity evidence (sufficient or limited) differs for the mentioned jobs and exposures.19
The classification of leukaemias has changed in recent years in which leukaemias of the lymphoid line are currently grouped under lymphomas.20 In the present study we evaluated possible associations between job title and selected hazardous agents (ie, pesticides, solvents, metals, contact with animals, ionising) and leukaemia in a large prospective cohort (European Prospective Investigation into Cancer and Nutrition (EPIC)) using both the old and new definition of leukaemia so as to retain comparison with previously published reports. In addition, we analysed risks per major subtype (acute myeloid leukaemia (AML), chronic myeloid leukaemia (CML) and chronic lymphoid leukaemia (CLL)) as some studies have suggested that occupational exposures appear to be related to a specific histological type of leukaemia.6
Materials and methods
The EPIC project is a European network of prospective cohorts that was set up to examine relationships of cancer risk with nutrition and metabolic risk factors.21 ,22 Briefly, recruitment of subjects took place between 1992 and 2000 in 23 centres located in 10 European countries (Denmark, France, Greece, Germany, Italy, The Netherlands, Norway, Spain, Sweden and the UK). The cohort includes participants of both genders, mostly in the age range 35–70 years at recruitment. After providing informed consent, standardised lifestyle and personal history questionnaires were collected from most participants. The EPIC study was approved by the review board of the International Agency for Research on Cancer (IARC) and by all local institutes recruiting participants. Further details can be found elsewhere.22 After the exclusion of subjects without information on lifestyle (n=1272), prevalent cancer cases (10 527), and centres without occupational histories: France (n=68 741), Oxford (n=53 712), The Netherlands (n=37 356), Sweden (n=49 686), Norway (n=35 890) and Naples (n=5 057), 241 465 subjects were included in the present analyses. Out of the whole study population, 91.55% was alive; 7.53% was dead; 0.43% did not want to continue to participate; 0.26% emigrated to another region; 0.18% emigrated to another country and for 0.06% status was unknown.
Outcomes and exposure assessment
Incident primary leukaemia cancers were identified by either population cancer registries (Denmark, Italy, Spain and the UK) or other methods, such as health insurance records, pathology registries, and active contact of study subjects or next of kin (Germany and Greece). The diagnosis, tumour site classification and morphology of each case were based on the International Classification of Disease (ICD-O-2), which was reclassified according to the recently published WHO classification of tumours of the haematopoietic and lymphoid tissues (ICD-O-3) (see data supplement 1, available online only). The follow-up period for the present study was for data reports received at IARC to the end of 2010. Follow-up time was accrued up to the date of last known contact, the diagnosis date or the date of death, whichever came first. The present analysis included 477 leukaemia cases (282 men and 195 women) and 240 988 non-cases (101 807 men and 139 181 women).
Occupational status was obtained using a questionnaire almost exclusively at the time of study recruitment in which participants were asked if they had ever worked in any of 52 occupations considered a priori to be at high risk of developing cancer (see data supplement 2, available online only). Information on educational level, smoking status, alcohol intake, body mass index (BMI), physical activity and family history of cancer were also obtained among other information.22
Job exposure matrix
The International Standard Classification of Occupations (ISCO-88) was used for coding of reported high-risk occupations. Occupational exposures were estimated by linking the reported occupations to a general population job exposure matrix (ALOHA-JEM).23 The occupational exposure categories included were: pesticides; herbicides (subset of pesticides); insecticides (subset of pesticides); aromatic solvents; benzene (subset of aromatic solvents); chlorinated solvents; trichloroethylene (subset of chlorinated solvents); formaldehyde; metals (organic and inorganic metals); contact with animal products; and ionising radiation. Three categories of exposure (formaldehyde, contact with animal products and ionising radiation) were not in the original ALOHA-JEM and were newly generated by two industrial hygienists (RV and YC). The JEM classified subjects based on job code into high, low and no exposure categories for the aforementioned exposures.
All statistical analyses were conducted using SAS, V.9.2. Cox proportional hazard regression models were used to examine the association between job title and associated occupational exposures and leukaemia. Age was used as the underlying time variable, with entry time defined as the subject's age at recruitment, and exit time defined as the subject's age at diagnosis of leukaemia, death, loss to follow-up or censoring date at the end of the follow-up period, whichever came first. Tests of linear trend across categories were conducted by fitting a model treating the tertile categories as an ordinal variable. Analyses were stratified by lymphoid versus myeloid leukaemia and subtypes of AML, CML and CLL. The effect of physical activity (sex-specific quartiles of combined recreational, household and occupational physical activity; categorical) and educational level (indicator of socioeconomic status; categorical), BMI (in kg/m2; continuous) and family history of cancer (ie, colon and breast cancer) as potential confounding variables were examined, but they did not appreciably (more than 10% change) change the risk estimates and therefore were not included in the final models. Multivariate analyses were adjusted for sex, smoking status (categorical: never, former and current smoker), alcohol intake (categorical: never, former, current and past and current drinker), age at recruitment and country.
Table 1 shows the general characteristics of the cohort and lymphoid and myeloid leukaemia cases. The mean follow-up time for 241 465 subjects was 11.20 years (SD 2.42 years). The 102 089 men and 139 376 women contributed 1 130 513 and 1 572 863 person-years of follow-up, respectively.
The crude incidence rate (IR) was 17.64 (95% CI 16.10 to 19.30) (IR 7.44, 95% CI 6.44 to 8.54 for myeloid and IR 8.77, 95% CI 7.69 to 9.96 for lymphoid) cases per 100 000 person-years based on 477 newly diagnosed leukaemia cases. Lymphoid and myeloid leukaemia cases were slightly older than the whole study population at recruitment. Myeloid leukaemia cases reported more family history of cancer compared to lymphoid leukaemia and the whole cohort. Smoking habits, educational levels, alcohol consumption, physical activity and BMI did not differ considerably among both leukaemia cases and the whole study population. Of all leukaemia cases 49.7% (n=237) were diagnosed with lymphoid leukaemia, 42.1% (n=201) with myeloid, 4.0% (n=19) with leukaemia Not Otherwise Specified (NOS) and 4.2% (n=20) with other leukaemias.
Job title-based analyses
Due to sample size limitations, we examined only job title categories with more than five exposed cases (see data supplement 3, available online only). Multivariate analysis restricted to myeloid leukaemia cases showed a significant increased risk for subjects working in shoe or other leather goods (HR 2.54, 95% CI 1.28 to 5.06) and a non-significant increased risk for painters (see data supplement 3, available online only). We found a non-significant risk reduction of myeloid leukaemias for workers in medical and health services and turning/metal engineering. The risk of lymphoid leukaemia (HR 8.35, 95% CI 1.58 to 44.24) and in particular CLL (HR 11.32, 95% CI 1.32 to 97.2) (data not shown) was increased significantly for working in a chemical laboratory. A non-significant increased risk of lymphoid leukaemia and CLL (data not shown) was observed for roofers (HR 2.19, 95% CI 0.79 to 6.09 and HR 2.55, 95% CI 0.93 to 7.02, respectively) and workers in turning/metal engineering (HR 2.31, 95% CI 0.91 to 5.82 and HR 2.98, 95% CI 0.93 to 9.54, respectively).
Exposure-specific analyses showed a non-significant increased risk of myeloid leukaemias for benzene (all categories) (table 2). Lymphoid leukaemia showed no significant association. Subsequent subtype-specific analyses indicated that benzene was associated, albeit non-significantly, with both acute and chronic myeloid leukaemia at high exposure levels (HR 1.52, 95% CI 0.78 to 2.98; HR 1.98, 95% CI 0.75 to 5.19, respectively) while exposure to aromatic solvent increased the risk of AML but not CML (table 3). In addition, we observed a non-significant increased risk of CLL for exposure to formaldehyde and ionising radiation.
In the present study, we investigated the associations between occupations considered a priori to be at high risk of developing cancer and the risk of developing leukaemia in a large multicentre prospective cohort study. Subsequently, we studied the possible association between leukaemia and selected occupational exposures by linking the reported occupations to a JEM. This study is one of the few population-based studies on occupational exposures that has several important strengths including its large size, ability to control for potential confounders and to investigate subtypes of leukaemia as well as the prospective design of the study. Nevertheless, some of the analyses were based on small numbers of exposed cases, and should be interpreted with caution. This particularly affects the analyses based on job titles.
We found a significant increased risk of lymphoid leukaemia and in particular CLL (data not shown) for working in chemical laboratories. An increased risk of mortality from malignant lymphoma was reported among members of the American Chemical Society who died between 1948 and 1967.24 Another study among professional chemists in England and Wales showed an excess mortality from CLL.25 Moreover, a study among Finnish laboratory workers showed a slight excess in leukaemia and non-Hodgkin lymphoma risk.26 A recent follow-up of 15 million people in five Nordic countries showed an excess in the standardised incidence ratio of non-Hodgkin lymphoma but not leukaemia among male technical workers including chemists.27 Together with our findings this would suggest that chemical laboratory workers are potentially exposed to leukaemogens and/or lymphomagens.
There has been some indication of an increased risk of total leukaemia and CLL among individuals involved in metal working occupations in previous studies.3 ,11 ,28 ,29 Consistently, in our study an increased risk of lymphoid leukaemias and CLL (data not shown) was found for turning/metal engineering. However, no increased risk was observed for metalworking or for metal exposure based on the JEM. It should, however, be noted that metal working covers a wide range of work situations and therefore includes a correspondingly wide range of potential exposures, such as mineral and metal dust, aromatic and chlorinated solvents and metalworking fluids.
We found an increased risk of lymphoid leukaemia among roofers who might be exposed to bitumen fumes/asphalt and/or coal tar-based materials, which are known to be human carcinogens. This finding is consistent with a meta-analysis of epidemiological studies that showed an excess risk of leukaemia among roofers,30 and a recent case–control study among male construction workers in California (1988–2007) that showed a significant increased risk of total leukaemia among roofers.31
In our study, a significant increased risk of myeloid leukaemia, in particular AML, among subjects working in the shoe/leather industries was found. Some previous studies showed that people working in the shoe/leather goods industry are at an increased risk of leukaemia32–34 while others did not show such an association.6 ,35 ,36 Glues in the shoe industry historically contained benzene. The content of benzene in glues, however, varied across time and might have varied from country to country.33 These differences may explain some of the observed heterogeneity in results.
The IARC has classified painting as an occupationally related cause of cancer. Consistent with our finding of an increased risk of myeloid leukaemia among painters, a study among Swedish workers employed in 1960 and/or 1970 in various painting trades and the paint and varnish manufacturing industry reported an increased risk of non-lymphocytic leukaemia among paint and varnish workers.37 However, a recent large study among the general population in five Nordic countries did not show a higher risk of leukaemia among painters.27
Our results showed that occupational exposure to high levels of aromatic solvents, in particular benzene, increased the risk of myeloid leukaemia and its subtypes AML and CML. Benzene exposure is a well-known cause of AML,4 ,38 but for CML, there is no consistent evidence to date.39 ,40 Most cohort studies have not found CML in benzene-exposed workers, while the more suggestive evidence comes from population-based case–control studies.39 Given the suggestive evidence from this study and others it seems prudent that future studies on benzene and/or CML would focus on this putative association.
Despite the large number of studies dealing with occupation and leukaemias in general, only a few studies focused on specific subtypes of the disease. The subjects included in our study constituted a large group with many years of follow-up and reasonably complete vital status ascertainment, which enables us to examine the associations by histological types of AML, CML, acute lymphoblastic leukaemia and CLL, although the numbers of acute lymphoblastic leukaemia cases for analyses were prohibitively low. Moreover, we had detailed information on several possible confounders, which has been frequently lacking in other studies.
Our study has some limitations. First, a large number of cohort participants were excluded due to the lack of occupational histories. Second, as the cohort was designed to address questions other than occupational exposures, the information on jobs was limited. The lack of information on the duration of employment in a particular job, on detailed job tasks and on exposure to particular carcinogens in/outside of the workplace will have caused a degree of exposure misclassification. In addition, we did not have full occupational job histories of the study subjects, which would have caused an underestimation of the exposure prevalence. However, given that most high-risk job groups were ascertained in our study, we have probably captured the most significant occupational exposures within the population. Moreover, the resulting exposure misclassification of the mentioned study limitations is likely to have been non-differential, and as such would have resulted in an underestimation of risk but would not have led to false positive findings. The latter might explain the failure to show a clear effect for known risk factors of leukaemia, in particular myeloid leukaemia, including exposure to ionising radiation and formaldehyde especially in combination with the very low prevalence of high exposed subjects to these agents (<1%).
In summary, our study showed some evidence supporting the role of occupational exposures in the development of both lymphoid and myeloid leukaemia. Exposure to benzene seemed to be associated with both acute and chronic myeloid leukaemia at high exposure levels. However, given the relative low numbers of exposed cases, the results have to be interpreted with some caution.
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:
Contributors All authors other than FSH and YC were responsible for the conception, design and conduct of the EPIC study. FSH, RV and YC were responsible for the execution of the research reported in the paper, the statistical analysis and the interpretation of data. The paper was drafted by FSH and was revised with contributions from co-authors. All authors reviewed and approved the paper.
Funding The work described in this paper was carried out with financial support of the Europe Against Cancer Program of the European Commission (SANCO), Deutsche Krebshilfe, Deutsches Krebsforschungszentrum, German Federal Ministry of Education and Research, Danish Cancer Society, Health Research Fund (FIS) of the Spanish Ministry of Health, Spanish regional governments of Andalucia, Asturias, Basque Country, Murcia, Navarra and the Catalan Institute of Oncology, La Caixa (BM 06-130), RTICC-RD06/0020, Spain, Cancer Research UK, Medical Research Council, UK, Stroke Association, UK, British Heart Foundation, Department of Health, UK, Food Standards Agency, UK, Wellcome Trust, UK, Greek Ministry of Health, Greek Ministry of Education, Greek Ministry of Health and Social Solidarity, the Hellenic Health Foundation, and the Stavros Niarchos Foundation, Italian Association for Research on Cancer (AIRC) and the Italian National Research Council.
Competing interests None.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the IARC ethics committee and all the local institutional review boards.
Provenance and peer review Not commissioned; externally peer reviewed.
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.