|Home | About | Journals | Submit | Contact Us | Français|
Unlike cancer of the bladder, cancer of the renal pelvis is not considered an occupational cancer and little is known about risks among women.
Using the Swedish national census and cancer registry-linked data (1971-1989), we identified transitional cell cancers of the renal pelvis (N=1374) and bladder (N=21,591). Correlation between cancer sites for the Standardized Incidence Ratios (SIR) were determined using Pearson's coefficient of the log SIR. Relative risks of job exposure matrix variables were calculated using Poisson regression.
Both cancer sites were significantly elevated among women and men employed in the machine/electronics industry, sedentary work, and indoor work, as well as among men employed in the shop and construction metal industry, contributing 10-14% of cases among men. Risks by industry were more highly correlated among women (r=0.49, p=0.002) than men (r=0.24, p=0.04).
Cancers of the renal pelvis and bladder share common occupational risk factors that may be more frequent among women. In addition, there may be several jobs that pose an increased risk specifically for cancer of the renal pelvis but not bladder.
In Sweden, the incidence of bladder cancer has doubled since the 1960s and reasons for this are not entirely clear [National Board of Health and Welfare of Sweden 2000, Thorn, et al. 1997, Mattsson and Wallgren 1984]. In addition, although bladder cancer accounts for a higher proportion of all cancers among men (7.3%) than women (2.9%), women have experienced a comparatively lower survival, and a greater increase in incidence [Ries, et al. 2003, McLaughlin, et al. 1996, Madeb and Messing 2004]. A significant proportion of bladder cancers among men (10-25%) have been attributed to occupational exposures, and changes in the workforce suggest that these exposures are also becoming more relevant for women [Kogevinas, et al. 2003, Vineis and Magnani 1985, Mannetje, et al. 1999].
At the same time, very little is known about occupational risk factors for cancers of the renal pelvis, particularly among women. Cancers of the bladder and renal pelvis are both predominantly of transitional cell origin (90-95%). Secondary cancer studies suggest that these two cancer sites are etiologically related because approximately 1-4% of bladder cancer patients will develop cancers of the renal pelvis and ureter, and 20-50% of renal pelvis cancer patients will also have bladder cancer [Rabbani, et al. 2001, Sharir and Jewett 1999, Kirkali and Tuzel 2003, Huguet-Perez, et al. 2001]. Two known carcinogens that are associated with both cancers include tobacco smoking and ionizing radiation.[Silverman, et al. 2006]
In this study, our purpose is to explore similar occupational risk factors between these two cancer sites among employed Swedish men and women.
The cohort was defined as all male and female Swedish residents reporting gainful employment at either the 1960-or-1970 censuses, and alive on January 1, 1971. Person-years were accrued for each member of the cohort beginning January 1, 1971 until a diagnosis of cancer, emigration, death or end of follow-up on December 31, 1989, whichever occurred first. We used the Swedish Cancer-Environment Registry (CER), Version III, to determine the incidence of transitional cell cancers of the renal pelvis (International Classification of Diseases, 7th Revision (ICD-7) code 180.1), and bladder (181.0) between 1971-1989. We also investigated mortality from these two cancers, but do not report these results in the tables. Cancers were limited by histology according to the Swedish Cancer Registry PAD (codes 114 and 116). Microscopic confirmation occurred for 97% of cancers in this study. CER III is the linked dataset between the 1960-and-1970 National Population and Housing Censuses and the Swedish cancer registry. The proportion of unlinked cancer cases was less than 1%, and the Census response rate was 99% [Wiklund and Eklund 1986]. Job titles reported to the Census were converted to the three-digit National Swedish Classification of Occupations and Industries, as described elsewhere [Shields, et al. 2002]. Urban residence was defined according to the place of residence at the 1970 Census (i.e. residence in one of the 3 largest Swedish cities of Stockholm, Göteborg, and Malmö).
Job exposure matrices (JEM), developed by American and Swedish experts in occupational medicine, were used to assess risks associated with the following occupational exposures: asbestos, ionizing radiation, low physical activity and indoor work. JEMs for other known or suspected bladder carcinogens have not yet been developed for this dataset. All scoring was blinded to the health status of subjects. Ionizing radiation and asbestos exposures were scored in the categories of probability (none, low, medium, high), intensity (none, low, medium, high), and confidence (low, high) of exposure, and have been used in previous research [Shields, et al. 2002]. Occupational physical activity was categorized as heavy, medium, light, and sedentary. This JEM has been shown to correlate reasonably well with self-reported occupational physical activity (spearman rank coefficient=0.62) [Moradi, et al. 1998] . Work location was classified into 3 categories, (outdoor only, both indoor and outdoor, and indoor only). Individuals were classified into a JEM category according to the occupation held in 1970.
We calculated the standardized incidence ratio (SIR), the ratio of observed to expected number of cancers, for each two- and three-digit occupation and industry category. Individuals were assigned to an industry or occupation category in a specified category if they reported employment in either 1960 or 1970. Thus, an individual reporting a different occupation in 1960 than in 1970 would be considered separately for each occupational exposure group. According to our calculations, 34% and 37% of men and 24% and 24% of women with cancer of the bladder and renal pelvis cancer, respectively, held a different occupation in 1960 than in 1970. The expected number of cases was based on the incidence rates in attained age (by 5-year-age groups), sex, site and calendar-year (by 4-year calendar periods from 1971-1989) specific cancer incidence rates. The expected rate was based on cancer incidence rates in the total employed population. The total employed population was defined as those individuals reporting employment at either the 1960 or 1970 census. Autopsy-only reported cases were excluded from both observed and expected rate calculations.
We present SIR results for occupations and industries with at least 3 exposed cases and meeting at least one the following criteria: 1) statistically significant (p<0.05), 2) SIR=2.0 or greater, and 3) a priori risk based on published literature [Kogevinas, et al. 2003]. For completeness, if the SIR met these criteria for renal pelvis cancer, we also reported the SIR value for bladder cancer, even if it did not meet the above criteria.
In order to test the hypothesis that bladder and renal pelvis cancers were similarly elevated in each occupation and industry category, the Pearson's correlation coefficient and two-tailed p-values were calculated from the log of the SIR values, stratified by sex. In addition, because occupational correlations could be related to smoking, we also estimated the correlation for SIR values using a conservative approach to adjust the SIR values for smoking. Using this approach, we identified occupations that were significantly elevated for lung cancer in this dataset as previously reported [Pollan and Gustavsson, 1999]. The list of male occupations elevated for bladder and/or renal pelvis cancer was then reviewed by the authors and placed into the following categories based on knowledge of the existing literature: definitely not tobacco-related only (occupation codes 633, 754, 757, 781, 794 and 826); probably not tobacco-related only (occupation codes 662, 751, 752, 795 and 932); and possibly tobacco-related only (295, 333, 603, 793, 941). The list of female occupations with both bladder and/or renal pelvis cancer elevations occurred in only three occupations (085, 758 and 921) all three of which were categorized as possibly tobacco-related only. Based on the existing literature that reports attenuation of occupation-related risks for cancers of the bladder and lung to range between 5 to 30 percent [Blair, et al. 1985, Siemiatycki, et al., 1988, Haldorsen, et al., 2004, Mannetje, et al., 1999, Richiardi, et al., 2005], we conservatively adjusted the number of observed cancers in each of the occupation groups above by reducing them by 30 percent and then calculated the tobacco-adjusted correlation between bladder and renal pelvis SIR values.
We used multivariate Poisson regression to estimate the relative risk of bladder and renal pelvis cancer with respect to the job-exposure matrix (JEM) estimated exposure to asbestos, ionizing radiation, physical activity, and indoor/outdoor work. The JEM variables were adjusted for urban residence, as well as attained age and calendar year.
Among the 4,197,684 employed Swedish residents there were 70,083,912 person-years of follow-up, with a mean follow-up of 16.70 years. There were 1,014 and 360 renal pelvis cancers and 18,244 and 3,347 bladder cancers among men and women, respectively.
Among women, there were no occupations with significantly elevated risks for both cancer sites (Table 1). Risk for both cancer sites was significantly reduced among women employed as agriculture and animal management workers. The correlation between bladder and renal pelvis risks was statistically significant, and this correlation was not altered once we adjusted the SIR values for smoking (r=0.47, p=0.02). Occupations which demonstrated excess risk for cancer of the bladder among women included: personnel work, business administration, bank cashier, clerk/secretary/stenographer, insurance clerk, switchboard operator, graphic work, bookbinder, photographic lab work, and waitress. Cancers of the renal pelvis were significantly elevated among women employed as pharmacists, journalists/editors, business executives, food-related workers, food process workers, and chemical/cellulose workers, unspecified. The largest proportion of bladder and renal pelvis cancers occurred among clerk/secretary/stenographers and waitresses combined (16% of bladder cancers and 18% of renal pelvis cancers), data not shown.
Among men, both cancer sites were elevated for purchasers/office sales people, shop and construction metal work, and, civil security and enforcement. Risks for both cancer sites were significantly reduced among men employed in agriculture and forestry work. The highest proportion of cases occurred in shop and construction metal work, accounting for 12% of bladder and 14% of renal pelvis cancers. There was not a statistically significant correlation between the non-smoking adjusted SIR values for occupation among men (r=0.17, p=0.12). However, once we adjusted SIR values for smoking the correlation increased slightly and became statistically significant (r=0.25, p=0.02).
Among women, risks for both cancer sites were elevated in the machine/electronics industry, and both cancer sites were significantly reduced in the agriculture and stock raising industry (Table 3). The correlation coefficient for the bladder and renal pelvis SIR industry values was statistically significant among women (r=0.49, p=0.002). Bladder cancer was significantly elevated among women employed in the following industries: graphics and publishing, book binderies, machine/electronics, radio/TV, wholesale durable goods (fuels, chemical, hardware and machines), banking, insurance, state administration (general), universities and higher learning, public functions, and hotel and restaurant work. Renal pelvis cancer was elevated among women employed in the food, chocolate and candy, machine/electronics, transportation equipment construction, hardware and machine business, and drugstore industries.
Among men, both cancer sites were significantly elevated among men employed in the machine/electronics industry, and both cancer sites were reduced among men employed in agriculture. The correlation coefficient between bladder and renal pelvis cancer risks was statistically significant (r=0.25, p=0.04). Employees in the machine/electronics industry accounted for 10% of bladder and 12% of renal pelvis cancers.
Urban residence was associated with a 26 to 36% elevated risk for cancers of the bladder among men and women, respectively, but was not statistically significant for renal pelvis cancer (Table 5). Among both men and women, sedentary (versus physically active) and indoor (versus outdoor) work were associated with a significantly increased risk of both cancer sites and trends for both sexes and cancer sites were statistically significant (p-trends ranging from <0.001 to <0.03).
Among men, moderate occupational ionizing radiation exposure was associated with an elevated risk of bladder cancer, however, there was no evidence of a trend. There was also no association with bladder cancer mortality (data not shown).
Both cancer sites were significantly elevated among workers in the following industries: machine/electronics (men and women), machine (men); and occupations: civil security (men), purchasing and office sales (men), and shop and construction metal (men); as well as indoor and sedentary workers, demonstrating evidence of a trend. Among all industry and occupational categories, shop and construction metal work contributed the highest proportion of cancers among men.
Cancer of the renal pelvis occurs less frequently than cancers of the bladder, and unlike bladder cancer, is not generally referred to as an occupational cancer. Our data suggest that among the employed population, the occupation-related correlations between these two cancer sites is moderate (25 percent among men and nearly 50 percent among women) and persists after proportionate adjustment for smoking-related cancers. We also found fewer commonly associated occupations between these two cancer sites than might be expected, given that both cancer sites are associated with tobacco smoking [Silverman, et al. 2006], the fact that cancers of the bladder and renal pelvis frequently occur simultaneously or as secondary cancers in the same individual [Rabbani, et al. 2001, Sharir and Jewett 1999], as well as data from molecular studies of genetic and epigenetic changes in tumors of the bladder and renal pelvis [Vriesema, et al. 2001, Paterson, et al. 2003, Hafner, et al. 2001].
In particular, we identified several jobs for which there was an elevated risk of cancer of the renal pelvis but not bladder. Several of these occupations and industrial groups also had a small number of cases (i.e. 5 or below), suggesting that these results could be due to chance. Those jobs with a significantly elevated risk for renal pelvis cancer, but not bladder cancer, and having more than 5 cases, included: women employed as food processors and employed in the food, transportation and drug store industries; and men employed as insurance clerks, butchers, cleaners and employed in the pulp grinding, metal manufacturing, scientific/surgical instruments, insurance and legal service industries. Cancers of the renal pelvis have been previously associated with work in the dry cleaning and iron and steel industry[McCredie and Stewart 1993]. Other occupational studies have reported an elevated risk of breast cancer in female pharmacists and an elevated risk of bladder cancer in the pharmaceutical industry.[Pelucchi, et al. 2002, Notani, et al. 1993, Baker, et al. 1986] Pharmacologic agents, such as phenacetin-containing analgesics, have been associated with papillary scarring and cancer of the renal pelvis while others are known to significantly alter intra-pelvic pressure as well as the flow rate of urine from the renal pelvis to the ureter[Stewart, et al. 1999, Jung, et al. 2006]. However, the extent of exposure to such agents in our study population is not known. Mycotoxins produced by fungal contamination of grains and cereals have been associated with Balkan endemic nephropathy, and this condition is in turn associated with an elevated risk of transitional cell cancers of the renal pelvis.[Grollman, et al. 2007] Certain mycotoxins, like ochratoxin A, are nephrotoxic and cause renal cell cancer in animal models.[Pfohl-Leszkowicz and Manderville 2007] Exposure to mycotoxins is known to occur in the food processing, meat, and pulp grinding industries[Prazmo, et al. 2003, Thuvander, et al. 2001]. One might hypothesize that that mycotoxin exposure is related to the commonly elevated risk of renal pelvis cancer in these occupational groups, although this is not known.
Conversely, it is notable that in our study all occupations and industrial groups which were protective for renal pelvis cancer (i.e. had a SIR value significantly below 1.0, p<0.05), were also protective for bladder cancer. In addition, consistent with an earlier study, we identified a reduced risk for cancer of the renal pelvis for employment in agriculture, which was lower among women than among men[McLaughlin, et al. 1987]. A large proportion of women in our study sample were employed in agriculture, which may explain why trends for outdoor work and occupational physical activity were steeper among women than men, although the confidence intervals among men and women were overlapping.
Differences in risk factors for cancers of the bladder and renal pelvis for certain occupation and industry groups may be explained because certain carcinogens can act selectively or more intensively upon either the bladder or the renal pelvis, as is evidenced by animal studies [Murai, et al. 1993, Bach 1991], and as suggested by differences in genetic instability between bladder and renal pelvis tumors occurring within the same patient [Takahashi, et al. 2001]. In summary, however, our results suggest that while there are a few occupations with shared increased risk, these occupations account for a moderate proportion of the total number of cases in the employed population.
Due to the sparse amount of information on risks for cancers of the renal pelvis, by necessity, the following discussion will focus primarily on the consistency of our findings with other studies of bladder cancer.
In other studies, greater physical activity has been related to a reduced risk of bladder cancer, although these reports were not-statistically significant [Dosemeci, et al. 1993, Hardman 2001, Hu, et al. 2003, McTiernan, et al. 1998, Mellemgaard, et al. 1995, Severson, et al. 1989, Sommer, et al. 2004, Thune and Furberg 2001, Tripathi, et al. 2002, Brownson, et al. 1991]. While the biologic mechanisms of physical activity in cancer prevention are not clear, it has been hypothesized that moderate physical activity may improve mucosal immunity and alter cytokine expression [Sommer, et al. 2004, Tripathi, et al. 2002, Gleeson, et al. 2004, Bevers, et al. 2004, Griffiths and Mellon 2004, Drela, et al. 2004]. Notably, intravesical administration of Bacillus Calmette-Guerin (BCG) used to treat superficial bladder cancer, results in altered cytokine production (as measured in urine), and in vitro, BCG has been shown to induce T-cell and NK-cell activity [Moldoveanu, et al. 2001, Saint, et al. 2003, Ma, et al. 1987]. In addition, other epidemiologic studies and animal models suggest that other physical-activity related disease pathways, including IGF-1 and diabetes, may also be relevant [Tripathi, et al. 2002, Hursting, et al. 2004, Coughlin, et al. 2004].
An association with indoor work has not been previously reported. Sunlight is the major source of vitamin D [Holick 2004]. The active metabolite of Vitamin (1α,25 (OH)2D3) is reported to promote apoptosis in human bladder cancer cell lines, and intravesical administration of vitamin D3 in the bladder has been shown to reduce the incidence of bladder tumors in animal models [Konety, et al. 2001, Yazawa, et al. 2000]. However, it is important to note that most physically active occupations are also outdoors and our study is unable to disentangle the differential relation of outdoor exposure and physical activity. Follow-up in future studies is necessary.
Reduced bladder cancer risk has been frequently found among farmers [McLaughlin, et al. 1987, Lee, et al. 2004, Blair, et al. 1992, Acquavella, et al. 1998, Porru, et al. 1996, Wiklund and Dich 1994]. However, elevated risks have been reported for professional gardeners [Silverman, et al. 1989, Teschke, et al. 1997, Colt, et al. 2004], orchard and greenhouse workers [Kristensen, et al. 1996], nursery workers [Kogevinas, et al. 2003], herbicide applicators [Lee, et al. 2004, Rusiecki, et al. 2004, la Vecchia, et al. 1990], and field crop workers [Mannetje, et al. 1999]. Several other cancers are reduced among farmers, possibly related to the generally lower smoking rate among farmers [Wiklund and Dich 1994, Statistics 1965, Wiklund and Dich 1995], although studies adjusting for smoking still find a reduced risk of bladder cancer among farmers [Barbone, et al. 1994, Folsom, et al. 1996]. Similarities in the risk for bladder cancer among men and women employed in farming and the agricultural sector suggest that non-occupational factors may be important since job-related exposures may be quite different between men and women [Alavanja, et al. 1999].
Urban residence has been previously associated with higher risk of bladder cancer, possibly related to environmental agents more frequently encountered in urban environments, including chlorination by-products in drinking water, tobacco smoke, and decreased fluid intake [Villanueva, et al. 2004].
Ionizing radiation is an established cause of bladder cancer, primarily through studies of high-dose radiotherapy, radioactive iodine therapy, and atomic bomb survivors [Silverman, et al. 2006]. Only a few studies have reported excess bladder cancer risk due to ‘low-dose’ ionizing radiation exposures which are generally found in the occupational setting, and most of these studies focus on cancer mortality rather than incidence [McGeoghegan and Binks 2000, Sont, et al. 2001, Cardis, et al. 1995, Ugnat, et al. 2004]. Our study found evidence of an association between ionizing radiation and bladder cancer incidence. However, we did not find a trend with higher exposure, possibly due to the small number of individuals in highly exposed jobs or to exposure misclassification. The plausibility of lower-dose exposure in bladder cancer risk has been suggested by the positive relation between Cs 137 levels in urine, chronic proliferative atypical cystitis (Chernobyl cystitis), and the development of dysplasia and carcinoma in situ of the bladder among individuals living in the Chernobyl region [Romanenko, et al. 2003].
Several studies report elevated risk of bladder cancer among machinists [Kogevinas, et al. 2003, Colt, et al. 2004, Tolbert 1997] and a previous follow-up of this cohort found elevated risk for cancer of the renal pelvis among this occupational group [McLaughlin, et al. 1987]. Metal-workers and machinists account for approximately 15% of occupationally-related bladder cancers in Europe [Kogevinas, et al. 2003]. Other electronics-related work associated with increased bladder cancer risk in this study include: radio and TV, electrical installation, electrical power/gas/waterworks, and electrical work. Electric power plant workers may be at excess risk for kidney and bladder cancer [Mattos, et al. 2002, Mattos and Koifman 1996], although this has not been consistently found [Mattos and Koifman 1996, Baris, et al. 1996, Keller and Howe 1993]. Electricians and electrical workers are exposed to polychlorinated biphenyls, asbestos, and low frequency EMF, none of which have been conclusively linked to occupational bladder cancer. Workers in the semiconductor industry exposed to elemental arsenic and/or gaseous arsenic hydride in the in the manufacture of silicon chips have been shown to have higher urinary metabolites of arsenic [Hwang, et al. 2002]. While arsenic is classified as a known bladder carcinogen, the carcinogenic potential of occupational exposures is not clear [Cancer 2004].
Our results are consistent with previous studies among women finding excess bladder cancer risks among, secretarial and clerical [Zheng, et al. 2002], cashiers [Colt, et al. 2004] [Colt, et al. 2004], education work [Zheng, et al. 2002], sales [Mannetje, et al. 1999, Colt, et al. 2004, Swanson and Burns 1995, Silverman, et al. 1990], graphics and publishing [Bulbulyan, et al. 1999], machine [Barbone, et al. 1994], electronics [Mannetje, et al. 1999, Silverman, et al. 1990, Simpson, et al. 1999], telephone industry [Dosemeci and Blair 1994], and waitresses [Simpson, et al. 1999]. In addition, among women, Gridley et al. (1999) report a 20% elevated risk of bladder cancers among the employed population, compared with the non-employed female population in Sweden [Gridley, et al. 1999]. Because our study was limited to the employed population, this suggests that the occupational bladder cancer risks among women reported in this study may be higher when compared with the general population. We did not confirm observations of significantly elevated bladder cancer risk among women employed in pulp and paper processing [Langseth and Andersen 1999], health services [Colt, et al. 2004, Simpson, et al. 1999, Carpenter and Roman 1999], leather [Garabrant and Wegman 1984], metal work [Mannetje, et al. 1999, Silverman, et al. 1990], chemical [Pelucchi, et al. 2002, Silverman, et al. 1990, Dolin and Cook-Mozaffari 1992], rubber [Swanson and Burns 1995, Silverman, et al. 1990, Simpson, et al. 1999, Carpenter and Roman 1999], and textile [Mannetje, et al. 1999, Simpson, et al. 1999, Carpenter and Roman 1999] industries. A large pooled analysis of case-control studies in Europe among leather and rubber work also did not report elevated risks [Mannetje, et al. 1999]. Elevations in bladder cancer risk among women employed as photographic laboratory workers, book binders, and business administrators, and the radio and TV industry have not been previously reported. Finally, although the relative risks we found in this study among women employed as clerks/secretaries and waitresses were modest, these occupations accounted for a large proportion of the total number of cases.
Our results are similar to research reporting elevated bladder cancer risk among men employed in: administrative and clerical work [Porru, et al. 1996, Schumacher, et al. 1989], artistic work [Brown, et al. 2002], automotive repair [Zheng, et al. 2002], chemical industry [Silverman, et al. 1989, Teschke, et al. 1997, la Vecchia, et al. 1990], electronics and electrical work [Mattos, et al. 2002, Dolin and Cook-Mozaffari 1992, Tynes, et al. 1992], engineering [Dolin and Cook-Mozaffari 1992], electrical power/gas and waterworks [Dolin and Cook-Mozaffari 1992], executives [Dolin and Cook-Mozaffari 1992], food processors [Dolin and Cook-Mozaffari 1992], gas stations [Schoenberg, et al. 1984], glass processing [Dolin and Cook-Mozaffari 1992], health care (including physicians) [Wynder and Goldsmith 1977, Howe, et al. 1980], machine industry [Kogevinas, et al. 2003, Colt, et al. 2004, Tolbert 1997, Dolin and Cook-Mozaffari 1992, Howe, et al. 1980, Claude, et al. 1988], metal work [Silverman, et al. 1989, Zheng, et al. 2002], photographic lab work, printing/graphics work [Lynge, et al. 1995, Pesch, et al. 2000], railroad work [Dolin and Cook-Mozaffari 1992], rubber industry [Vineis and Magnani 1985, Zheng, et al. 2002, Straif, et al. 1998], ship's machine command [Dolin and Cook-Mozaffari 1992], shop and metal industry [Kogevinas, et al. 2003, Teschke, et al. 1997, Colt, et al. 2004, Pesch, et al. 2000], transportation workers [Colt, et al. 2004, Dolin and Cook-Mozaffari 1992, Pesch, et al. 2000, Hoar and Hoover 1985, Silverman, et al. 1983, Guo, et al. 2004], and waiters [Porru, et al. 1996]. This study did not confirm elevations in bladder cancer among men employed as asbestos workers [Ugnat, et al. 2004], dry cleaners [Teschke, et al. 1997, Zheng, et al. 2002, Pesch, et al. 2000, Ruder, et al. 2001, Brown and Kaplan 1987], dyestuffs workers [Vineis and Magnani 1985, Ugnat, et al. 2004, Vineis and Pirastu 1997], hairdressers [Teschke, et al. 1997, Lynge 1990, Miller and Bartsch 2001], leather workers [Dolin and Cook-Mozaffari 1992, Schumacher, et al. 1989, Pesch, et al. 2000, Montanaro, et al. 1997], miners [Teschke, et al. 1997, Dolin and Cook-Mozaffari 1992], painters [Silverman, et al. 1989, Teschke, et al. 1997, Zheng, et al. 2002, Dolin and Cook-Mozaffari 1992], petroleum workers [Teschke, et al. 1997], textile workers [Kogevinas, et al. 2003, Teschke, et al. 1997, Frumin, et al. 1990], or welders [Silverman, et al. 1989]. A large pooled case-control study also reported no association with hairdressing [Mannetje, et al. 1999], and the lack of association during the time period of this study may be related to changes in the constituents of the hair dyes used [Czene et al. 2003]. There may also be unique aspects of the Swedish work environment which might be important in comparing our results with other studies [Kleinman 1984]. Similar to our study, Axelson et al. (1994) did not observe an increased risk of bladder cancer in Swedish dry cleaners [Axelson, et al. 1994]. In addition, while previous studies have found quite high risks associated with dyestuffs work (ranging from 4 to 64-fold), benzidene and 2-naphthylamine are two known carcinogenic compounds which have not been widely used in the dye industry in Nordic countries [Vineis and Magnani 1985, la Vecchia, et al. 1990, Naito, et al. 1995], [Dreyer, et al. 1997].
The strengths of this study include: a large number of cases in a population with excellent cancer surveillance, the investigation of risk factors among the working population in order to control for the ‘healthy-worker bias,’ and the ascertainment of occupational exposure prior to cancer diagnosis. Important limitations of this study include: 1) no adjustment for the effects of other potential confounding factors in the estimation SIR values (e.g. diet, tobacco use, use of protective equipment on the job), 2) exposure misclassification due to classification of occupational exposures based on job title occurring at one point in time, which may be greater for women than for men [Stewart and Blair 1994]; and 3) the large number of comparisons in this analysis, which increases the possibility that a statistically significant finding is due to chance alone [Law, et al. 2001].
Methodologic studies report that adjustment for tobacco use in occupational studies may alter occupational risk estimates [Shavers, et al. 2005, Ji, et al. 2005, Richiardi, et al. 2005]. Studies using diverse methods including the assessment of confounding in case-control studies using self-reported smoking status, and those employing an a priori correction factor for smoking-related occupations have found adjustment for smoking to result in attenuation of approximately 10 to 25% [Blair, et al. 1985, Siemiatycki, et al., 1988, Haldorsen, et al., 2004, Mannetje, et al., 1999, Richiardi, et al., 2005]. A recent study reports that, among Swedish men employed in 1960 and 1970, the risk for bladder cancer among men was attenuated downwards by a range 5 to 29%, depending upon the occupation [Ji et al., 2005]. Richiardi et al. (2005) adjusted for self-reported smoking status and found lung cancer risk among men employed in metal and machine industry to be attenuated downwards by 6% and for blue collar workers by 11% [Richiardi, et al. 2005]. We used the upper limit of 30% to adjust for the possibility that smoking could be confounding our SIR results. The smoking-adjusted SIR resulted in an increased correlation between bladder and renal pelvis risks among men. This suggests that the correlations we observed between bladder and renal pelvis cancers were not due to tobacco alone. It is likely, however, that our correction may be too stringent as not all occupations with an elevated risk for lung cancer are solely due to smoking, but may be attributed to recognized workplace carcinogens [Siemiatycki et al., 2004, Rousseau et al., 2005].
Our work suggests that there are common occupational risks for cancers of the bladder and renal pelvis, particularly among women. Further research is needed to confirm commonly elevated risks found for the machine/electronics industry, physical inactivity and indoor work. In addition, there may be several jobs that pose an increased risk specifically for cancer of the renal pelvis but not bladder.
The authors would like to thank Drs. Debra Silverman, Wong-Ho Chow, Sheila Zahm, and Mitchell Gail of the National Cancer Institute, United States, for their thoughtful review of the manuscript; and Drs. Nils Plato and Tahereh Moradi of Karolinska Institutet, Stockholm, Sweden for the development of Job Exposure Matrices. The authors also appreciate the clerical assistance of Ms. Aimee Seisay, Ms. Barbara Hynum and Ms. Diane Pague in the preparation of the tables and manuscript and the computer programming assistance of Ms. Heather Morris, Information Management Services, Inc. This research was supported by the Intramural Research Program of the National Cancer Institute.
Grant Sponsor: This work was conducted as part of a Post-Doctoral Fellowship at the National Cancer Institute, Division of Cancer Epidemiology and Genetics
Institutions where statistical analysis performed: National Cancer Institute and Pennsylvania State University
Conflict of Interest
The authors do not presently have affiliations with organizations that have a direct financial interest in the subject matter or materials discussed.