Objectives We used data from a large, population-based case–control study in Maine, New Hampshire, and Vermont to examine relationships between occupation, industry and bladder cancer risk.
Methods Lifetime occupational histories were obtained by personal interview from 1158 patients newly diagnosed with urothelial carcinoma of the bladder in 2001–2004, and from 1402 population controls. Unconditional logistic regression was used to calculate ORs and 95% CIs, adjusted for demographic factors, smoking and employment in other high-risk occupations.
Results Male precision metalworkers and metalworking/plasticworking machine operators had significantly elevated risks and significant trends in risk with duration of employment (precision metalworkers: OR 2.2, 95% CI 1.4 to 3.4, ptrend=0.0065; metalworking/plasticworking machine operators: OR 1.6, 95% CI 1.01 to 2.6, ptrend=0.047). Other occupations/industries for which risk increased significantly with duration of employment included: for men, textile machine operators, mechanics/repairers, automobile mechanics, plumbers, computer systems analysts, information clerks, and landscape industry workers; for women, service occupations, health services, cleaning and building services, management-related occupations, electronic components manufacturing and transportation equipment manufacturing. Men reporting use of metalworking fluids (MWF) had a significantly elevated bladder cancer risk (OR 1.7, 95% CI 1.1 to 2.5).
Conclusions Our findings support the hypothesis that some component(s) of MWF may be carcinogenic to the bladder. Our results also corroborate many other previously reported associations between bladder cancer risk and various occupations. More detailed analyses using information from the study's job-specific questionnaires may help to identify MWF components that may be carcinogenic, and other bladder carcinogens associated with a variety of occupations.
- Bladder cancer
- case-control study
Statistics from Altmetric.com
What this paper adds
Over 40 occupations have been associated with an elevated risk of bladder cancer in epidemiological studies, but the evidence is compelling for only a few.
A large, population-based case–control study with lifetime occupational histories and detailed exposure assessment techniques provided the opportunity to further examine occupational risk factors for bladder cancer.
Men reporting use of metalworking fluids (MWF) had a significantly elevated bladder cancer risk, supporting the hypothesis that these fluids contain bladder carcinogens and contribute to the often observed association between bladder cancer and metalworking.
Significant findings were observed for many other occupations/industries, including: for men, textile machine operators, automobile mechanics, plumbers and landscape workers; and for women, health services, cleaning and building services, electronic components manufacturing and transportation equipment manufacturing.
More detailed analyses using information collected in job-specific questionnaires administered in this study may help to identify components of MWF that may be carcinogenic, and other bladder carcinogens to which people are exposed in a variety of occupations.
Occupational risk factors for bladder cancer have been examined in many studies. Although over 40 occupations have been associated with an elevated risk, the evidence is compelling for only a small number of occupations: dyestuffs workers and dye users, aromatic amine manufacturing workers, rubber workers, leather workers, painters, truck drivers and aluminium workers. There is strong evidence that certain aromatic amines are carcinogenic to the human bladder. Associations have been observed for several other occupational exposures, including polycyclic aromatic hydrocarbons (PAHs), diesel engine exhaust, leather dust, mineral oils, combustion and pyrolysis products, chlorinated solvents, creosote, herbicides/pesticides and asbestos.1 We used interview data from a large, population-based case–control study in New England, where bladder cancer mortality rates have been elevated for decades,2 to examine the relationships between occupation, industry and bladder cancer risk. This study builds upon earlier studies of bladder cancer and occupation in this region3–5 by studying disease incidence rather than mortality and by using detailed exposure assessment techniques. The extent to which occupational exposures might explain the New England bladder cancer excess will be examined as part of a broader effort integrating findings from this analysis with those for several other risk factors.
The study was conducted in Maine, Vermont and New Hampshire. All residents newly diagnosed with a histologically-confirmed carcinoma of the urinary bladder (including carcinoma in situ) at ages 30–79 years between 1 September 2001 and 31 October 2004 (Maine and Vermont) or between 1 January 2002 and 31 July 2004 (New Hampshire) were eligible for study. Rapid patient ascertainment in each state was conducted using data from hospital pathology departments, hospital cancer registries and the state cancer registries. We interviewed 1213 cases (65% of 1878 eligible cases). Of non-participating cases, 50% refused, 22% were deceased, 12% were too ill, 5% did not speak English, 5% had a physician refusal and 5% were not locatable. Based on a diagnostic slide review by the study's expert pathologist (AS), we excluded 20 patients determined not to have bladder cancer and 23 patients with non-urothelial carcinoma, leaving 1170 cases.
Controls aged 30–64 years were selected randomly from Department of Motor Vehicle (DMV) records in each state, and controls aged 65–79 years were selected from beneficiary records of the Centers for Medicare and Medicaid Services (CMS). Controls were frequency matched to cases by state, gender and age at diagnosis or control selection (within 5 years). We interviewed 1418 (594 DMV and 824 CMS) controls (65% of eligible DMV and 65% of eligible CMS controls). Of non-participating controls, 70% and 65% of DMV and CMS controls, respectively, refused, 24% of DMV and 11% of CMS controls were not locatable, 3% of DMV and 10% of CMS controls did not speak English, 1% of DMV and 7% of CMS controls were too ill, and 1% of DMV and 7% of CMS controls were deceased.
All respondents gave written informed consent to participate in this study. The study protocol was approved by the National Cancer Institute Special Studies Institutional Review Board, as well as the human subjects review boards of each participating institution.
Individuals who agreed to participate were asked to complete a mailed residence and work history calendar before the home visit. During the home visit, a trained interviewer reviewed the calendar and administered a computer-assisted personal interview. The interviewer recorded all jobs held for at least 6 months since age 16, excluding unpaid jobs and absentee business ownership. For each job, participants were queried about the employer name and location, years job started and ended, type of business, products made or services provided, and the participant's main duties, tools used and chemicals handled. For certain occupations, job-specific questionnaires were administered to solicit detailed information about exposures of interest6 (see online appendix A). The computer-assisted personal interview also covered demographics, tobacco use and other exposures.
A total of 1158 of 1170 cases (99%) and 1402 of 1418 controls (99%) were included in the analysis; the remainder had no qualifying jobs or were missing information on potential confounders.
Occupational coding and statistical analysis
We coded each job using the 1980 Standard Occupational Classification (SOC)7 and the 1987 Standard Industrial Classification (SIC)8 schemes. Using unconditional logistic regression models, we computed sex-specific ORs and 95% CIs for bladder cancer for each two-, three- and four-digit SOC and SIC code using SAS statistical software, version 9.1. For each OR, the reference group comprised those never employed in the occupation or industry. ORs were adjusted for age (<55, 55–64, 65–74, 75+ years), race (white only, Native American/white, other races), Hispanic ethnicity (yes/no), state, smoking status (never, occasional (<100 cigarettes over the lifetime), former, current) and employment in a high-risk occupation other than the one being analysed (ever/never). Occupations were identified as ‘high-risk’ (see online appendix B) for men and women separately if ORs in this study were 1.5 or higher prior to the high-risk occupation adjustment and 10 or more individuals were employed; however, if a three-digit occupational code met these criteria but had substantial variation in risk among the four-digit codes it comprised, we selected only those four-digit codes with elevated risk, regardless of the number of people employed. We further evaluated smoking effects by replacing smoking status with smoking duration (never smoked, smoked occasionally, smoked for <10 years, 10–19 years, 20–29 years, 30–39 years, 40–49 years, 50+ years) and found minimal changes in the ORs; therefore, the final models were adjusted for smoking status only. ORs and 95% CIs for ever/never employment in each SOC and SIC code are presented in the tables if the occupation or industry was held by at least 15 men or 15 women in our study and (1) had an OR that was either statistically significant or ≥2.0 or ≤0.5 in our study, or (2) is considered to be an a priori high-risk occupation (ie, an occupation with multiple reports of elevated risk in the literature and listed in Schottenfeld and Fraumeni's textbook).1 Online appendices B and C provide complete results for ever/never employment in all occupations and industries held by our study participants.
We assessed risk by duration of employment for occupations and industries with a significant association for ever/never employed. In addition to the reference category (never held the job), two duration categories (with 5 or 10 years of employment as cut-points) were used if fewer than 50 people held the job, and three (with 5 and either 10 or 15 years of employment as cut-points) were used if there were 50+ people. Selection of cut-points was job specific, based on the distribution of years of employment. Tests of linear trend were performed by treating the median duration of employment among controls for each duration category as a continuous variable; participants who never held the job of interest were assigned a duration of zero.
For occupations with a significant positive trend in risk with increasing duration of employment, we examined the relationship between initial year of employment and bladder cancer risk if numbers permitted. We also tested for interactions between smoking and occupation by adding a cross-product term to the logistic model. Finally, because controls under age 65 were limited to those with driver's licences, we also restricted the cases under age 65 to those with a valid driver's licence at the time of diagnosis; the changes in the risk estimates were minimal and the results are not presented.
We focus mainly on occupations and industries with significant trends in risk with duration of employment. All statistically significant associations, both positive and negative, are indicated by asterisks in tables 1 and 2.
Risk by SOC code
Among men, bladder cancer risk was significantly elevated (table 1) and increased significantly with duration of employment (table 3) in several a priori high-risk occupations: information clerks (SOC 464, OR 2.6, 95% CI 1.1 to 6.5, ptrend=0.011), mechanics and repairers (SOC 61, OR 1.3, 95% CI 1.04 to 1.6, ptrend=0.0030), vehicle and mobile equipment mechanics (SOC 611, OR 1.5, 95% CI 1.1 to 2.0, ptrend=0.0023), automobile mechanics (SOC 6111, OR 1.6, 95% CI 1.05 to 2.4, ptrend=0.030), precision metalworkers (SOC 681–2, OR 2.2, 95% CI 1.4 to 3.4, ptrend=0.0065), machine operators and tenders (metal and plastic) (SOC 75, OR 1.6, 95% CI 1.05 to 2.4, ptrend=0.023), metalworking/plasticworking machine operators (SOC 751–2, OR 1.6, 95% CI 1.01 to 2.6, ptrend=0.047) and textile/apparel/furnishings machine operators (SOC 765, OR 2.0, 95% CI 1.2 to 3.3, ptrend=0.0013). Plumbers/pipefitters/steamfitters (SOC 645), another a priori high-risk occupation, had a significant positive trend in risk (ptrend=0.032) and a non-significantly increased risk for ever/never employed (OR 1.5, 95% CI 0.8 to 2.5). The only a posteriori high-risk occupation with a significant duration trend was computer systems analysts (SOC 1712, OR 4.4, 95% CI 1.3 to 14, ptrend=0.021). Bladder cancer risk was significantly reduced among male accountants/auditors/financial specialists (SOC 141, OR 0.5, 95% CI 0.3 to 0.9), with a significant negative trend with duration of employment (ptrend=0.045).
Among women, bladder cancer risk was significantly elevated (table 1) and increased significantly with duration of employment (table 3) in management-related occupations (SOC 14, OR 2.7, 95% CI 1.3 to 5.5, ptrend=0.0018) and service occupations, except private household and protective (SOC 52, OR 1.6, 95% CI 1.1 to 2.3, ptrend=0.031). Among the service occupations, risks were elevated in health services (SOC 523, OR 2.5, 95% CI 1.4 to 4.4, ptrend=0.0013) and in cleaning and building service occupations (SOC 524, OR 3.7, 95% CI 1.8 to 7.3, ptrend=0.0010) (both a priori high-risk occupations). Women in miscellaneous administrative support occupations (SOC 479) had a significantly reduced risk (OR 0.4, 95% CI 0.2 to 0.8) with a significant duration effect (ptrend=0.043).
Risk by SIC code
Men in the landscape and horticultural services industry (SIC 078) had a significantly elevated risk of bladder cancer (OR 2.4, 95% CI 1.2 to 4.8) (table 2) and a significant positive trend in risk with duration of employment (ptrend=0.023) (table 3), as did women in the electronic components and accessories industry (SIC 367, OR 2.2, 95% CI 1.1 to 4.7, ptrend=0.028) and the transportation equipment industry (SIC 37, OR 8.7, 95% CI 2.0 to 37, ptrend=0.023). Significant reductions in risk with significant duration effects were observed for men working as real estate agents/managers (SIC 6531, OR 0.5, 95% CI 0.3 to 0.9, ptrend=0.011) and in nursing and personal care facilities (SIC 805, OR 0.3, 95% CI 0.1 to 0.97, ptrend=0.037) and for women working in colleges and universities (SIC 8221, OR 0.05, 95% CI 0.2 to 0.9, ptrend=0.012).
Risk by decade first employed
The overall increased risk among vehicle and mobile equipment mechanics, particularly automobile mechanics, was limited to men who began working before 1970 and was highest for those first employed before 1950 (table 4). Only those plumbers/pipefitters/steamfitters first employed before 1960 were at elevated risk. The elevated risk for metalworking/plasticworking machine operators was evident only among men who first held this job before 1970. Men who first operated textile/apparel/furnishing machines in the 1950s had an eightfold risk of bladder cancer that was statistically significant, but there was no discernable pattern in risk by year of first employment.
Interactions with cigarette smoking
The only occupation with a statistically significant smoking interaction was precision metalworking (table 5). Although the interaction was not significant for metalworking/plastic working machine operators, the joint effect was stronger than an additive model would yield, with a sixfold risk among smokers compared to non-smoking men never employed in that occupation. The elevated risk observed among male precision metalworkers and metalworking/plasticworking machine operators was evident only among smokers.
Exposure to metalworking fluids
The excess risk of bladder cancer among male precision metalworkers and metalworking/plasticworking machine operators prompted us to explore the association between metalworking fluids (MWF) and bladder cancer risk. These complex chemical mixtures are used as coolants and lubricants in the machining process and are characterised as: straight oils (typically mineral oil), soluble oils, semi-synthetic fluids (mineral oil emulsified in water) and synthetic fluids (water with organics and additives; no oil content). Using information from the occupational histories and job-specific questionnaires, we grouped study participants into four mutually exclusive categories: (1) people reporting use of MWF in a job-specific questionnaire; (2) people assessed as having possible exposure to MWF by the study's industrial hygienist (PS); (3) people assessed as having no exposure to MWF, but possible exposure to mineral oil; and (4) people assessed as having no exposure to either MWF or mineral oil (referent group). We observed statistically significant elevations in risk among men exposed to MWF (OR 1.7, 95% CI 1.1 to 2.5) and those with possible exposure to mineral oil (OR 1.3, 95% CI 1.1 to 1.7) (table 6).
Several occupations and industries were associated with a significantly elevated risk of bladder cancer in this study. Of particular note are the findings for precision metalworkers and metalworking/plasticworking machine operators (most of whom worked with metals). There is strong a priori evidence of an excess bladder cancer risk among metalworkers,4 9–27 and this has been widely hypothesised as being attributable to MWF exposure. In our study, reported use of MWF carried a 70% excess risk of bladder cancer that was statistically significant, adding to the evidence that some of these fluids contain bladder carcinogens. However, changes over time in the composition and prevalence of use of the different MWF make it difficult to identify the components potentially associated with bladder cancer.
The early, ‘straight’ MWF were typically mineral oils that contained PAHs, one class of possible bladder carcinogens.17 28 29 PAH removal began in the 1950s30 and levels were drastically lower by the mid-1980s31; however, small amounts may be created during machining.32 Synthetic MWF, first marketed in the 1950s,31 contained nitrites and amine additives for corrosion inhibition, often leading to the formation of N-nitrosamines,32 33 which are bladder carcinogens in animal models.34 Although nitrite was eliminated as an additive in the 1980s,31 nitrosamines are still found at low levels.35 Synthetic MWF may also contain N-phenyl-2-naphthylamine32 and other additives and contaminants. In our study, only those metalworking/plasticworking machine operators first employed before 1970 had an elevated risk, but it is not possible to distinguish between secular changes in exposure and latency effects in this study. A recent cohort study of automotive workers in Michigan found that exposure to straight MWF increased bladder cancer risk but was unable to determine whether this was attributable to PAH exposure; synthetic MWF were not associated with an elevated risk of bladder cancer in that study.36 Also noteworthy in our study were the significant interaction between smoking and precision metalworking, and the observation that the increased bladder cancer risk was observed largely among smokers.
The trend in risk with duration of employment among male operators of textile/apparel/furnishing machines was highly statistically significant (ptrend=0.0013), with a 10-fold risk for long-term employment. Most of these jobs were in the textile industry or the leather industry, with higher risk in the latter (OR 3.3, 95% CI 1.4 to 8.0) than the former (OR 1.6, 95% CI 0.8 to 3.3) (not shown). Most of the textile machine operators in the leather industry worked in shoe manufacturing, where there are many potential exposures; leather dust and solvents are the most common. Many investigators have reported elevated bladder cancer risk among shoemakers/repairers,12 13 29 37–39 but it is unclear whether these studies included shoe manufacturers. Machine operators in the textile industry worked mostly in wool or cotton mills operating winding/twisting machines or knitting/weaving machines. The excess risk observed among operators of these types of machines is consistent with several previous studies.17 40–42 Chemicals used in spinning and weaving areas have included mineral oils, polymers and sizing agents.42 43
The bladder cancer excess risk observed among male mechanics and repairers, particularly automobile mechanics, has been reported previously.9 12 17 23 27 29 44 In our study, only those who started this work before 1970 had elevated risk. This could indicate that exposure to a carcinogen has diminished over time, or it could be a latency effect. The work environment of mechanics may involve exposure to many substances, including asbestos, oils and greases, metal dust, welding and soldering fumes, solvents, machining fluids and paints.10 12 27 44–46
Men working as plumbers for 15 years or more had a significantly elevated risk of bladder cancer in our study. Previous studies of plumbers have been inconsistent, with some reporting significant bladder cancer excesses18 27 and others reporting non-significant excesses4 10 37 47 or null findings.12 17 41 44 46 Plumbers have been reported to be exposed to many hazardous materials (eg, lead and other welding fumes, various solvents, tar, greases and asbestos), making it difficult to identify putative agents. Only those men who worked as plumbers before 1960 had an excess bladder cancer risk.
Risk was significantly elevated among men in the landscape and horticultural services industry, consistent with previous reports of excess bladder cancer risk among nursery workers, gardeners and lawn care service employees.9 12 14 17 25 48 49 Although it has been postulated that exposure to pesticides and/or fertilisers might be responsible, the evidence is conflicting; some studies of pesticides and fertilisers reported an elevated risk,14 17 50–54 while others did not.11 12 55 56
We observed a significant, twofold risk of bladder cancer among women in the electronic components and accessories industry that was attributable mainly to the manufacture of semiconductors and related devices. Employment in computer manufacturing, a related industry, was identified as a high-risk industry in Detroit women,57 but a study of cancer incidence among male and female workers in a semiconductor manufacturing plant and a plant manufacturing computer hard drives and other electronic storage devices did not observe a bladder cancer excess.58 Workers in the semiconductor and computer manufacturing industries may have been exposed to a variety of known or suspected carcinogens, including metals (arsenic, nickel, chromium), electromagnetic fields, asbestos, acids and various solvents.58 59
Risk was not significantly elevated among male construction/maintenance painters in our study, contrary to several previous studies.23 27 60 61 However, when we identified men who had worked as painters (construction and maintenance painters; coating/painting/spraying machine operators; painting supervisors; or hand painting, coating and decorating occupations) and whose job titles or reported duties, tools, or exposures indicated that they had performed spray painting, there were many more cases (n=14) than controls (n=5). Several other investigators have described elevated risks for spray painters,11 12 14 52 who have been reported to be exposed to many known or suspected carcinogens, including solvents and metals.
The strengths of our study include the population-based design and ascertainment of complete occupational histories from direct interviews with study participants, and the ability to adjust for smoking, employment in high-risk occupations and other risk factors. Limitations include the small number of people in many occupations and industries, and the large number of comparisons performed, allowing for chance associations. We reduced the number of comparisons and therefore the likelihood of chance findings to some extent by restricting the analysis to occupations held by at least 15 participants of a single gender, but it is likely that some of the significant associations reported here arose by chance. The 65% participation rate for cases and controls can be viewed as a limitation; however, because study participation is unlikely to have differed between cases and controls in an exposure-dependent manner, we do not believe that this led to bias in the risk estimates. Assigning SOC and SIC codes was challenging for some jobs because of limitations in the information available from the interview; coders were blinded to case–control status and any misclassification was likely to be non-differential. Finally, occupation and industry titles are only a crude surrogate for exposures. An individual job or industry title may be associated with a wide range of exposures, and grouping people with a given job title who may be highly exposed with those potentially unexposed attenuates the strength of an association.
In conclusion, our results lend support to the hypothesis that some as yet unidentified components of MWF are carcinogenic to the bladder in humans and corroborate many other previously-reported associations between bladder cancer risk and various occupations. More detailed analyses using information collected in job-specific questionnaires administered in this study may help to identify components of MWF that may be carcinogenic, and other bladder carcinogens to which people are exposed in a variety of occupations.
Funding The NIH funded this study.
Competing interests None.
Ethics approval The study protocol was approved by the National Cancer Institute Special Studies Institutional Review Board, as well as the human subjects review boards of each participating institution.
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.