![]() |
![]() ![]() |
||||||||||||||||
![]() ![]() ![]() |
|
||||||||||||||||
|
![]() Occupational exposure to chemical and petrochemical industries and bladder cancer risk in four western Canadian provincesAnne-Marie Ugnat, Wei Luo, Robert Semenciw, Yang Mao and The Canadian Cancer Registries Epidemiology Research Group AbstractOccupational factors have been proposed to play a critical role in bladder cancer. This population-based case-control study was conducted to confirm the association between selected occupational and non-occupational risk factors and risk of bladder cancer using data collected from the four western Canadian provinces. Unconditional logistic regression analyses were based on 549 histologically confirmed bladder cancer cases and 1,099 controls. Bladder cancer risk was found to increase with increasing pack-years of cigarette smoking with an odds ratio (OR) in the highest quartile of 3.32 (95% confidence interval [CI], 2.28-4.82). A dose-response relationship was demonstrated between bladder cancer and pack-years of smoking (p<0.0001). A positive trend was observed with coffee consumption in men (p<0.0001), with the highest risk in the highest category of exposure: drinkers of four cups or more per day had an OR of 1.77 (95% CI 1.11-2.82). Increased bladder cancer risk was associated with self-reported exposure at work to several chemicals: asbestos (OR 1.69 [95% CI 1.07-2.65]); mineral, cutting or lubricating oil (1.64 [95% CI 1.06-2.55]); benzidine (2.20 [95% CI 1.00-4.87]). The population attributable fraction (PAF) estimates were 51% for cigarette smoking, 17% for heavy coffee consumption, 10% for mineral, cutting or lubricating oil exposure, 6% for asbestos exposure, and 1% for benzidine exposure. Although self-reported chemical exposures have important limitations, the findings are suggestive of increased risk for several associations previously reported between chemical agents or industries and risk of bladder cancer. Key words: bladder cancer; occupation BackgroundBladder cancer is the fourth most frequently diagnosed cancer in Canadian men.1 It has been estimated that 3,700 new cases will be diagnosed in 2004, corresponding to an annual rate of 23 per 100,000 Canadian men. Environmental exposure to certain chemicals has been linked to increased risk of bladder cancer.2-4 In industrialized countries, cigarette smoking has been identified as the most important risk factor, accounting for about 50% of bladder cancer cases in men.5 Occupational exposure ranks second as a risk factor.6 Many chemicals evaluated by the International Agency for Research on Cancer (IARC)7 because of their potential carcinogenic risk to humans originate in the chemical and petrochemical industries. It is hypothesized that most chemical carcinogens affect the urothelial cells through their presence in the urine. The National Enhanced Cancer Surveillance System (NECSS) coordinated by Health Canada provided an excellent opportunity to confirm the association between selected occupational and non-occupational risk factors and risk of bladder cancer in a large population-based Canadian sample. Because other studies have shown that, in industrialized countries, the population-attributable risk for bladder cancer associated with occupational exposure is less than 10% among women8 compared with up to 25% among men,9 we chose to focus our study on Canadian men. Materials and methodsSelection of cases and controlsDeveloped in the mid-1990s, the NECSS used questionnaires to collect detailed risk factor information on 19 types of cancer from a nationwide sample of 20,755 recently diagnosed cancer patients and 5,039 population control subjects. It was designed to explore the relation between environmental factors and risk of cancer.10 From the NECSS data, we assembled all histologically confirmed incident cases of bladder cancer (International Classification of Diseases, Ninth Revision, code C67) in men aged 20 to 75 that were reported between 1994 and 1997 in the four western Canadian provinces (British Columbia, Alberta, Saskatchewan and Manitoba). We based our study on these provinces because the NECSS did not collect bladder cancer data in Ontario, Quebec or New Brunswick, and the remaining major Canadian chemical and petrochemical industries are located in the west. Applying province-specific methods, representative samples of each province were accumulated to form a population control pool in the NECSS database. From this pool, control groups were selected with an age distribution (by five-year age groups) similar to that of the bladder cancer cases. In British Columbia, Saskatchewan and Manitoba, a random sample was obtained through the provincial health insurance plans, which generally cover more than 95% of the provincial populations. Random-digit dialing was used to derive the control sample in Alberta. Data collection Mailed self-administered questionnaires, with telephone follow-up when necessary for clarification, were used to collect information on all subjects' residential and occupational histories and on other risk factors for bladder cancer. Questions were included on demographic data, socioeconomic status, lifestyle, behaviour, diet pattern, physical activity and occupational history. Overall, response rates in the western provinces were 60% for male bladder cancer cases and 59% for controls. Occupational exposure The data collection questionnaire included questions on employment history. For each job held for at least 12 months, details were requested about the type of industry, main job duties, job location, title, status (full, part-time or seasonal) and duration. Subjects were also asked whether they had ever worked with any of 17 particular chemical substances for more than one year, and the total years of exposure. Occupational exposure was obtained in two ways.
Years of occupation in these industries were categorized into long-term (20 years or more), medium long-term (10 to 19 years), short-term (1 to 9 years) and no exposure. Statistical analysisDue to the availability of information on a large number of variables, we implemented a screening process based on biological and statistical associations supported by the literature and the univariate analysis. All the variables on which the differences between cases and controls in the univariate analysis were statistically significant were included in the final multivariate analysis model. Unconditional multiple logistic regression models (SAS software)12 were used to estimate the odds ratios (ORs) while adjusting for potential confounders. Since different sampling methods were used in each province, the province of residence was controlled for in the univariate analysis. The multivariate analysis model included age (20-54, 55-64, 65-74), years of exposure, province of residence, education, pack-years of smoking, coffee consumption and tea consumption. The PAF was calculated using a method presented by Bruzzi et al.,13 which allows estimation on the basis of data from case-control studies. The calculations apply the knowledge of the OR and distribution of exposure only among cases, assuming that they are representative of all cases in the population. In addition to the total PAF, the summary attributable risk for each of the multiple factors was estimated with adjustment for the other risk factors. ResultsA total of 549 male bladder cancer cases and 1,099 male controls were included in the study. Table 1 compares cases and controls with respect to characteristics reported in the literature to be risk factors for bladder cancer. The controls were more likely to have more years of education than the cases. The risk increased with pack-years of smoking and with greater coffee consumption. A dose-response relationship was observed in men, with the highest risk in the highest category of exposure: smokers with smoking pack years of 30 or more had an OR of 3.32 (95% CI 2.28-4.82) and drinkers of four cups or more per day had an OR of 1.77 (95% CI 1.11-2.82). Self-reported chemical exposure at work Table 2 shows the ORs for bladder cancer based on occupational exposure to the chemical substances specifically asked about in the questionnaire. After adjusting for province of residence, pack-years of smoking, age, education, years of exposure to chemical substance, coffee consumption and tea consumption, we found significantly elevated risks of bladder cancer among subjects who reported occupational exposure to these chemical agents: asbestos (OR = 1.69, 95% CI 1.07-2.65); mineral, cutting or lubricating oil (OR = 1.64, 95% CI 1.06-2.55); and benzidine (OR = 2.20, 95% CI 1.00-4.87). A non-significantly elevated risk was observed for men who reported exposure to the following chemical substances: dyestuffs (OR = 3.05, 95% CI 0.87-10.68); pesticides (OR = 1.50, 95% CI 0.89-2.53); herbicides (OR = 1.18, 95% CI 0.66-2.12); radiation sources (OR=1.39, 95% CI 0.75-2.56); isopropyl oil (OR = 1.31, 95% CI 0.43-3.98); asphalt and creosote (OR = 1.30, 95% CI 0.89-1.92); and welding materials (OR = 1.11, 95% CI 0.74-1.66).
Table 3 presents adjusted ORs for bladder cancer in men in relation to their length of exposure to the chemicals studied. The risk of bladder cancer increased with increasing years of exposure to dyestuffs. The ORs were 1.1 (95% CI 0.4-2.8) for 1 to 9 years; 3.4 (95% CI 0.6-20.9) for 10 to19 years; and 4.7 (95% CI 0.9-23.8) for 20 years or more. The tests for trend were significant (p = 0.03). SOC-grouped chemical-related industries The results of examining occupational exposure based on SOC groups (Table 4) indicated that, compared with men who worked only in occupations unrelated to chemicals, men who had ever worked in chemical-related occupations (but not in major chemical-related industries) had a non- significantly higher risk of bladder cancer (OR = 1.20, 95% CI 0.90-1.60). Men who had ever worked in metal processing or pulp and paper making industries had a non-significantly higher risk of bladder cancer: the ORs were 1.30 (95% CI 0.55-3.30) and 2.33 (95% CI 0.75-7.25) respectively. Men who had ever worked in any of the major chemical related industries (metal processing, pulp and paper making, or petroleum-, rubber- or plastic-related industries) also showed a non-significantly higher risk of bladder cancer (OR = 1.27, 95% CI 0.65-2.47). Similar results were observed when adjusting for province of residence and age only.
Population attributable fraction (PAF) Estimation of PAF is presented in Table 5. Overall, PAF was 51% for smoking, 33% for coffee consumption, and 16% for occupational exposure to asbestos, mineral, cutting or lubricating oil, and benzidine. These three factors together explained more than 70% of bladder cancer cases in this population. The largest PAF was observed in those with 30 or more pack-years of somking (26%). Coffee consumption of four cups or more accounted for 17% of bladder cancer. The risk attributable to occupation ranged from 1.4 to 10.1%. DiscussionIn this case-control study, we found a significant positive association between risk of bladder cancer and occupational exposure of Canadian men to specific chemicals: mineral, cutting or lubricating oil, asbestos, and benzidine. It was estimated that 16% of bladder cancer could have been prevented by elimination of occupational exposure to these chemicals. Compared with men who never worked in chemical-related occupations, men who were ever employed in metal processing or pulp and paper industries, or any of the major chemical-related industries (metal processing, pulp and paper making, or petroleum processing) had a non-significantly higher risk of bladder cancer. We also confirmed that smoking and coffee consumption were significantly related to increased bladder cancer risk. The proportion of bladder cancer attributable to smoking and heavy coffee consumption were 51% and 17% respectively. An IARC review concluded that "there is sufficient evidence from studies in humans that mineral oils (containing various additives and impurities) that have been used in occupations such as mulespinning, metal machining and jute processing are carcinogenic to humans."14 Several bladder cancer case-control studies have noted an association between bladder cancer risk and work as a machinist, with ORs ranging from 1.5 to 5.0.15 In our study, we also found an increased risk of bladder cancer among men with occupational exposure to mineral, cutting or lubricating oil. Prevention strategies to reduce the impact of occupational exposure to such oils should be considered. The US Department of Labour - Occupational Safety and Health Administration has established two exposure limits that may apply to cutting fluids. Employees should be exposed to no more than 5mg/m3 of mineral oil mist for an eight hour time weighted average (TWA), and no more than 15 mg/m3 for any particulate, as an eight hour TWA.16 Bladder cancer has long been known to be related to occupational exposure in the dye-producing industries.17 We observed a dose-response relationship between number of years exposed to dyestuffs and bladder cancer risk in men in the present study as observed elsewhere.17-19 The carcinogenic risk to the bladder of benzidine is well known,20,21 and our study provides further support. The results of epidemiological studies of bladder cancer and asbestos are inconsistent. Silverman and colleagues22 suggested that asbestos and insulation workers may be at increased risk of bladder cancer, whereas other researchers found no association.23,24 In our study, the higher bladder cancer risk seen among men exposed to asbestos could have occurred because the men exposed to asbestos were highly likely to have been exposed to mineral, cutting or lubricating oil (more than half of the men exposed to asbestos [118/206] also had been exposed to lubricating oil). The PAFs of occupational exposure were estimated to be 10% for mineral, cutting or lubricating oil, 6% for asbestos and 1% for benzidine. The PAF for mineral, cutting or lubricating oil is considerably greater than asbestos and benzidine. Although the OR for mineral, cutting or lubricating oil is moderate (1.64), the frequency of occupational exposure to this substance is greater than asbestos and benzidine. We examined the SOCs of people who self-reported with occupational exposure to mineral, cutting or lubricating oil. The results indicated that 26% of men had occupations related to farming, fishing, forestry, mining and quarrying; 63% of men had occupations related to processing, machining, product fabricating, construction trades, transport equipment operating, material handling, and equipment operating. In our study, statistically significant elevations of two- to three-fold were observed in the ORs at all smoking levels above 10 pack-years. A dose-response relationship was demonstrated between bladder cancer and pack-years of smoking. Such associations are well-documented in the literature.25-27 Our data support an association between coffee consumption and risk of bladder cancer which is consistent with other studies.28,29 To rule out a possibility of residual effect of smoking, coffee consumption and the risk of bladder cancer was also examined among non-smokers. A statistically significant excess risk was consistently observed for subjects having drunk more than 4 cups per day (OR=6.17, [95% CI 1.73-21.96]). However, a recent systematic literature review has suggested that coffee and tea consumption are probably not associated with bladder cancer.30 Coffee drinking has been studied extensively as a potential risk factor, but the inconsistency of the observed associations suggests that the relationship is either quite weak, noncausal, or dependent in a complex way on unmeasured factors.31 The attributable risk estimates indicated 33% of bladder cancers among men could have been prevented by elimination of coffee consumption. A previous study by D'Avanzo et al.32 found that coffee consumption was potentially responsible for 23% of bladder cancer. The possibility of recall bias cannot be discarded in our study. The main advantage of this study is that it is population based. We were able to examine bladder cancer risk according to self-reported exposure to certain chemical substances at work and by occupational group based on SOCs. We also controlled for major bladder cancer risk factors. The average response rates of both bladder cancer cases and controls were low (60% and 59% respectively). Since exposure information for non-respondents was not available for analysis, we could not examine whether or not systematic differences existed between respondents and non-respondents with respect to occupational exposure. Potential non-response bias might be introduced due to the low response rate. This study only considered jobs held for more than one year, but this might be a substantial issue for people who were seasonally employed (and exposed). However, the underestimation of exposure would likely be the same for cases and controls, resulting in an OR biased towards 1.
We developed our exposure assessment methods without peer review. Because these methods were never validated, the potential for misclassification could not be quantified. Information on occupational chemical exposure was based solely on subjects' recall of contact with chemical agents selected a priori and listed on the questionnaire. We made no attempt to obtain all the job titles held by subjects to link with local exposure patterns. However, we examined the SOCs of respondents' self-reported occupational exposure to asbestos, mineral, cutting or lubricating oil, or benzidine. The results show that more than 90% of these people had ever been employed in chemical-related industries. Therefore, it is unlikely that this method introduced a strong information bias, it is very likely that it introduced exposure misclassification and caused true associations to be missed or underestimated. Self-reported exposure to chemicals at work could also introduce recall bias since people who have adverse health outcomes tend to remember and report past exposures differently than do those who did not develop such health outcomes. We used only the first three digits of the SOC codes to determine subjects' occupations. Grouping jobs meant that people with different work duties and exposures were included in the same occupational group; this is usually expected to result in non-differential misclassification and bias of risk estimates to the null.33 Multiple comparison is also a concern for this study. There is a potential of generating false-positive results due to random variation. The results of this study should be interpreted with caution. ConclusionIn conclusion, this study suggests that occupational exposure to certain chemical substances is associated with increased risk of bladder cancer. However, further investigations are needed in order to provide sufficient evidence of causal relations and to instigate preventive measures. We confirmed that chemical-related occupations pose a risk of bladder cancer in Canadian men. The findings of this investigation support an association between lifestyle factors (cigarette smoking and coffee consumption) and bladder cancer in men. AcknowledgementsThe authors gratefully acknowledge Dr. K Johnson for his comments on the manuscript. The authors wish to thank L Anderson for assistance with manuscript editing, and L On and SY Pan for technical assistance.
Author References Anne-Marie Ugnat, Wei Luo, Robert Semenciw, Yang Mao, Surveillance and Risk Assessment Division, Centre for Chronic Disease Prevention and Control, Population and Public Health Branch, Health Canada, Ottawa, Ontario, Canada The Canadian Cancer Registries Epidemiology Research Group Correspondence: Wei Luo, Surveillance and Risk Assessment Division, Centre for Chronic Disease Prevention and Control, Population and Public Health Branch, Health Canada, PL 6702A, 120 Colonnade Road, Ottawa, Ontario, Canada K1A 0K9; Fax: (613) 941-2057; E-mail: Wei_Luo@hc-sc.gc.ca [Previous] [Table of Contents] [Next]
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
![]() |
|||
![]() |
Last Updated: 2004-08-19 | ![]() |