AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 57:896–905 (2014)

Prostate Cancer and Occupational Exposure to Whole-Body Vibration in a National Population-Based Cohort Study Marcella K. Jones, MPH,1,2 M. Anne Harris, PhD,1,3 Paul A. Peters, Michael Tjepkema, MPH,4 and Paul A. Demers, PhD1,2

PhD,

4

Background Following preliminary evidence from observational studies, we test the potential relationship between whole-body vibration (WBV) and prostate cancer in a cohort study. Methods WBV exposure was assigned based on occupation in 1991 and 1,107,700 participants were followed for incident prostate cancer until the end of 2003. Adjusted hazard rate ratios (HRs) were calculated using Cox proportional hazards modeling. Results 17,922 incident prostate cancer cases were observed. WBV-exposed men in Natural and Applied Sciences Occupations had a 37% elevated risk of prostate cancer (95% CI 1.09–1.72) and WBV-exposed men in Trades, Transport, and Equipment Operators Occupations had a 9% reduced risk (95% CI 0.86–0.97). Independent of WBV exposure, small but significant differences in risk were seen for several occupational categories. Conclusions We found no consistent relationship between WBV and prostate cancer. Further research could focus on other exposures or specific occupations in the studied categories to determine what may be contributing to the observed differences in prostate cancer risk. Am. J. Ind. Med. 57:896–905, 2014. ß 2014 Wiley Periodicals, Inc. KEY WORDS: occupational epidemiology; prostate cancer; whole-body vibration; cohort study; risk factors

INTRODUCTION Prostate cancer is the most common type of cancer diagnosed in Canadian men, with an estimated incidence of 121 cases per 100,000 men in 2012 [Canadian Cancer Society’s Steering Committee on Cancer Statistics, 2012].

1

Occupational Cancer Research Centre,Toronto, Canada Dalla Lana School of Public Health, University of Toronto,Toronto, Canada 3 School of Occupational and Public Health, Ryerson University,Toronto, Canada 4 Health Analysis Division, Statistics Canada, Ottawa, Canada  Correspondence to: Marcella K. Jones, MPH, Division of Epidemiology, Dalla Lana School of Public Health,c/o Graduate Office, 6th Floor,155 College St.,Toronto ON,Canada M5T 3M7. E-mail: [email protected] 2

Accepted 5 May 2014 DOI10.1002/ajim.22354. Published online 25 June 2014 in Wiley Online Library (wileyonlinelibrary.com).

ß 2014 Wiley Periodicals, Inc.

Age, family history of prostate cancer, and ethnicity are established risk factors for prostate cancer [Brawley, 2012; Leitzmann and Rohrmann, 2012]. Other potential risk factors studied include diet, genetics, physical activity, obesity, and smoking, though the results of these studies have been mixed [Brawley, 2012; Leitzmann and Rohrmann, 2012]. Occupational exposure to arsenic, cadmium, nightshift work, pesticides, and employment in rubber industry and agriculture have also been investigated with varied results [Parent and Siemiatycki, 2001; Cogliano et al., 2011; Parent et al., 2012; Koutros et al., 2013]. Recently, whole-body vibration (WBV) has been considered as a possible occupational risk factor for prostate cancer [Young et al., 2009; Nadalin et al., 2012]. WBV occurs when mechanical energy from a vibrating surface is transmitted to the body through the feet or trunk. It is commonly experienced by workers who operate forklift

Prostate Cancer and Occupational Exposure to Whole-Body Vibration

trucks, transport trucks, tractors, buses, loaders, cars, and vans [Palmer et al., 2000]. Occupational exposure to WBV is a logical starting point for evaluating health effects as variation in vibration doses are more attributable to workplace than leisure activities at a population level [Palmer et al., 2000]. Carcinogenic mechanisms by which WBV could increase prostate cancer risk have not yet been proposed. However, WBV exposure has been associated with prostatitis [Milby and Spear, 1974; Helmkamp et al., 1984; Wasserman et al., 1984] and elevated testosterone levels [Bosco et al., 2000], both of which may increase the risk of prostate cancer [Gann et al., 1996; Dennis et al., 2002; Roberts et al., 2004]. Preliminary epidemiological evidence suggests that occupational exposure to WBV may be associated with an increase in prostate cancer risk [Sass-Kortsak et al., 2007; Young et al., 2009; Nadalin et al., 2012]. A case-control study in Northeastern Ontario (Canada) that looked at occupational risk factors for prostate cancer found a 20% significant increase in prostate cancer risk among workers employed in trades, transport, and equipment operator occupations [Sass-Kortsak et al., 2007]. The authors proposed that WBV contributed to this relationship, since numerous known confounders were accounted for. A second case-control study in Montreal that assigned WBV exposure through a job-exposure matrix (JEM) of intensity and daily duration of exposure found a 44% increased risk of prostate cancer in those who ever had occupational exposure to WBV, although the result was not statistically significant [Nadalin et al., 2012]. A 90% significantly increased risk of prostate cancer was also seen in workers that were employed in transport equipment operations. Lastly, a meta-analysis looked at driving occupations and prostate cancer with the intent of determining whether WBV might play a role in this relationship [Young et al., 2009]. The pooled relative risk estimate found a slightly elevated non-significant risk [Young et al., 2009]. While there have been longitudinal studies that have looked at the risk of prostate cancer among all occupations [Sharma-Wagner et al., 2000; Zeegers et al., 2004; Pukkala et al., 2009], no cohort study has been conducted that specifically looks at WBV as a risk factor for prostate cancer. Following the recent evidence of a possible WBV and prostate cancer association [Sass-Kortsak et al., 2007; Young et al., 2009; Nadalin et al., 2012], the main objective of this study was to test the relationship between WBV and prostate cancer using a large population-based cohort of the Canadian population and a JEM similar to that used in a previous study [Nadalin et al., 2012]. A secondary objective of the study was to examine the risk of prostate cancer in different occupational categories to help contextualize findings from analyses of exposure to WBV.

897

METHODS Study Population The 1991 census cohort used in this study was originally created for the “Canadian census mortality follow-up study, 1991 through 2001” [Wilkins et al., 2008]. Individuals were eligible for the census cohort if they were 25 years of age and older, were a usual resident of Canada on the census reference day (June 4, 1991), were not a long-term resident of an institution such as a prison, hospital, or nursing home, and had been amongst the 20% of Canadian households selected to complete the long-form questionnaire. The cohort was established via probabilistic matching of 1991 in-scope census records to non-financial tax-filer data from 1990 and 1991, using dates of birth and postal codes of the individual and his or her spouse or common-law partner (if any). About three quarters of the in-scope persons were successfully linked to tax-filer data, creating a cohort of 2.7 million people (representing 15% of the 1991 non-institutionalized Canadian population 25 years and older) for follow-up in the mortality and cancer incidence data. Subjects in the census cohort were linked to the Canadian Mortality Database using deterministic and probabilistic linkage methods [Wilkins et al., 2008]. The 1991 Canadian census follow-up study received approval by the Statistics Canada Policy Committee after consultation with Statistics Canada Confidentiality and Legislation Committee, Data Access and Control Services Division, and the Federal Privacy Commissioner. The 1991 Canadian census follow-up study was amended to extend the years of mortality follow-up, link to cancer incidence from the Canadian Cancer Database, and link to annual place of residence from Historical Tax Summary Files (HTSF) [Peters and Tjepkema, 2010; Peters et al., 2013]. The Canadian Cancer Database contains national reporting of cancer incidence from 1969 to present, with 316,003 individuals and 338,085 incidences (individuals may have multiple incidences of cancer) linked to the end of 2003 in the present cohort. Age (at entry to cohort), sex, province of residence, income adequacy quintile, highest level of education, and occupation were obtained from the 1991 census data. Prostate cancer cases, year of diagnosis, and age of diagnosis were provided by the Canadian Cancer Database. Year of death was retrieved from the Canadian Mortality Database. Lastly, the HTSF was used to identify individuals who had either emigrated or were lost to follow-up during the study period. For the purposes of this study, the cohort was further restricted to working males to minimize the healthy worker effect [Li and Sung, 1999] and to males under the age of 75 to minimize survival bias by excluding those men who did not report an occupation or were older than 74 on census day. Study methods were approved by the University of Toronto’s Health Sciences Research Ethics Board (protocol

898

Jones et al.

reference # 26517). Due to the use of anonymized secondary linked data sources, no recruitment or contact with participants occurred for this study and therefore study specific consent was not possible.

Exposure Assessment Census respondents were asked to describe the kind of work they were doing and the most important work activities in their job over the week prior to the census. If no job was held in the week prior, respondents were asked to answer for the job of longest duration since January 1, 1990. If more than one job was held in the week prior to the census, respondents answered for the job at which the most hours were worked [Statistics Canada, 1992]. This reported occupation information for each individual was then coded to different occupational classification systems by a Statistics Canada employee. This study used the Standard Occupational Classification (SOC) 1991 system [Statistics Canada, 1991]. In this system, occupations are grouped into 10 general categories that are denoted by letters A–J; occupational groups are further specified using 1–3 numbers. A total of 512 occupation groups are defined by the SOC 1991. Information about occupational exposure to WBV was obtained from a WBV exposure database provided by authors of a previous case-control study, which included information on WBV intensity (in m/s2) and daily duration of exposure (in hours/ day) [Nadalin et al., 2012]. Using this WBV database, one of three levels of WBV exposure was assigned to every SOC 1991 category: no exposure, low exposure, or high exposure.

 Low WBV exposure: o



1–8 hr/day duration and low intensity (0.45 m/s2) or; o 1–4 hr/day duration and medium intensity (>0.45 and 0.90 m/s2). High WBV exposure: o o

5–8 hr/day duration and medium intensity (>0.45 and 0.90 m/s2) or; 1–8 hr/day duration and high intensity (>0.90 m/s2).

The WBV exposure database was supplemented with additional occupational exposure information from the scientific literature in order to create a simple job-exposure matrix (JEM) by SOC 1991 code. Each occupational category was reviewed for WBV exposure by an occupational hygiene expert (PD). A tabulation was then created to compile the WBV exposure for each SOC code, as indicated by three types of data sources (the WBV exposure database from Nadalin et al. [2012], additional exposure assessment literature, and expert assessment). From these, two exposure

variables were created. The main exposure variable assigned an exposure (i.e., none, low, or high) only when all sources agreed on the level of exposure. If the sources disagreed, the lower exposure was assigned (i.e., none or low). A second “sensitive” exposure accounted for this disagreement, and assigned a low or high exposure even if only a single source indicated that level of exposure. Refer to the supplementary Table S1 provided online for a copy of the JEM and notes on the supporting sources for the assigned exposure estimates.

Prostate Cancer Case Attainment Prostate cancer was defined based on the 3rd revision of the International Classification of Diseases for Oncology (ICD-O-3) (code C619) or the 9th revision of the International Classification of Diseases (ICD-9) (code 1850). Both codes were used to account for provincial differences in cancer registration and to ensure that all incident cases of prostate cancer in the Canadian Cancer Database were captured. In this study, only the first prostate cancer since cohort inception in 1991 in an individual was considered a case.

Statistical Methods Hazard rate ratios (HRs) and their 95% confidence intervals (CI) were calculated using Cox proportional hazard regression models. Person-years were calculated from an individual’s entry into the cohort in 1991 until prostate cancer diagnosis, death, emigration, lost to follow-up, or the end of 2003, whichever occurred first. Three main analyses were completed with distinct constructions of WBV exposure as the independent variable of interest: (1) level of WBV exposure; (2) binary WBV exposure; (3) binary WBV exposure within a SOC 1991 letter category. Second, a sub-analysis was completed in a cohort subset of men under the age of 50. Prostate cancer screening increases in prevalence after age 50 (Beaulac et al., 2006), and so we hypothesize that prostate cancer in this younger age group is less sensitive to trends in prostate cancer screening and may represent more aggressive disease. Third, a sensitivity analysis was also conducted using the sensitive WBV exposure assessment to explore the effect of the exposure assignment on the results. Finally, a secondary analysis was conducted to estimate the risk of prostate cancer between different SOC 1991 letter categories, independent of exposure to WBV. In this analysis a rolling reference group was used so that prostate cancer risk in each occupational category was compared to the prostate cancer risk in all other occupational categories together. All HR estimates were adjusted for age at entry to the cohort (in 10-year age categories) and province of residence

Prostate Cancer and Occupational Exposure to Whole-Body Vibration

at the time of the census. The census-assessed socioeconomic indicators of income and education were also included in the model to control for potential confounding by socio-economic status (SES), screening rates, and lifestyle factors, such as diet and level of physical activity; factors which may be more similar within groups of similar SES [Giles-Corti and Donovan, 2002; Beaulac et al., 2006; Darmon and Drewnowski, 2008]. Income adequacy quintiles were constructed as follows: First, for each economic family or unattached individual, total pre-tax, post-transfer income from all sources was combined across all family members. The ratio of total income of the economic family to the Statistics Canada low income cut-off (pre-tax, post-transfer) for the applicable family size and community size group was then calculated based on the low income cut-offs shown in the 1991 Census Dictionary [Statistics Canada, 1992]. Quintiles of the population based on this ratio were then constructed within each census metropolitan area, census agglomeration, or rural and small town area (provincial residual) [Wilkins et al., 2008]. Education level was determined by assigning an individual to one of four categories based on the highest level of schooling attained by the time of the 1991 census. In accordance with Statistics Canada disclosure rules, case counts of less than six were suppressed in the reported tables, all frequencies (with the exception of prostate cancer cases) were rounded to the nearest 100, and person-years were rounded to the nearest 10. All statistical analyses were performed using SAS 9.2 (SAS Institute, Cary, NC).

RESULTS The derivation of the cohort and the flow of participants through the cohort study are illustrated in Figure 1. There were 1,107,700 males in the working cohort at the beginning of follow-up in 1991. At the end of follow-up in 2003, 959,400 males remained in the cohort and a total of 12,675,590 person-years were accumulated. Deaths and loses to follow-up were similar between exposure groups, and approximately 86% of each exposure group remained in the cohort after 12 years of follow-up. The crude incidence density rate of prostate cancer was 1.41 cases per 1,000 person-years, based on 17,922 diagnoses during the observation period. Baseline characteristics of the cohort population by WBV exposure status are presented in Table I. 173,700 males were classified as exposed to WBV; 82,400 exposed to low levels of WBV and 91,300 to high levels. Age distribution was similar between exposure groups. The income adequacy quintile distributions between exposure groups were similar, though the proportion of exposed males in the richest quintile (22.9%) was smaller than for unexposed males in the richest quintile (26.2%). Those exposed to WBV were less educated

899

than those unexposed to WBV, with only 3.9% of exposed males having received a university degree by 1991, compared to 19.3% of unexposed males. Most individuals exposed to WBV worked in SOC category H (Trades, Transport, and Equipment Operators and Related Occupations) (45.4%) or category I (Occupations Related to Primary Industry) (34.1%). Exposed individuals were also far more likely to be employed in a blue collar profession (85.6%) than a white collar profession (14.4%). A total of 3,142 prostate cancer cases were diagnosed in men exposed to WBV, representing a crude incidence density rate of 1.58 cases per 1,000 person-years. In the unexposed group, a total of 14,780 prostate cancer cases were diagnosed; a crude incidence density rate of 1.38 cases per 1,000 personyears. Table II shows Cox proportional hazard ratios for the associations between WBV and prostate cancer incidence, adjusted for age and province or age, province, and SES. No relationship was found between level of WBV exposure or binary WBV exposure and prostate cancer. Exposed males who were employed in SOC category C (Natural and Applied Sciences and Related Occupations) had a significantly elevated risk of prostate cancer (age, province and SES adjusted HR ¼ 1.37, 95% CI 1.09–1.72) compared to all unexposed working males. Exposed males employed in SOC category H (Trades, Transport, and Equipment Operators and Related Occupations) had significantly reduced risk of prostate cancer (HR ¼ 0.91, 95% CI 0.86–0.97). A sub-analysis was conducted in men under 50 years of age. A total of 271 prostate cancer cases were diagnosed in men under 50 during the 7,748,160 years of follow-up accumulated in the under 50 cohort; a crude incidence density rate of 0.035 cases per 1,000 person-years. No significant associations were detected, though the trends were similar to the results in the main cohort. Exposed males under 50 in SOC category C had an elevated risk of prostate cancer (HR ¼ 2.04, 95% CI 0.5–8.24), greater than what was observed in this category in the main cohort. Exposed males under 50 in SOC category G (Sales and Service Occupations) also had an increased risk of prostate cancer (HR ¼ 1.37, 95% CI 0.72–2.58), which was not observed in the main cohort. In the sensitivity analyses, using a definition of WBV exposure designed to be sensitive rather than specific, similar results were seen; no relationship was found between binary WBV exposure or level of WBV exposure and prostate cancer, and the relationships between WBV exposure and prostate cancer in specific SOC categories were similar to the results in the main analysis. In general, the confidence intervals around the sensitive exposure estimates were wider than for the main (specific exposure) analysis (results not shown). Table III shows Cox proportional hazard ratios for the associations between employment in SOC categories and prostate cancer incidence, independent of exposure to WBV.

900

Jones et al.

FIGURE1. Derivation of the male working cohort and status of individuals throughout the follow-up period. Follow-up frequencies are rounded to the nearest 100 in accordance with Statistics Canada disclosure rules. WBV ¼ whole-body vibration.

These estimates are also adjusted for age and province or age, province, and SES. A small significantly elevated risk was present in three SOC categories: A (Management Occupations) (HR ¼ 1.05, 95% 1.01–1.09), B (Business, Finance,

and Administrative Occupations) (HR ¼ 1.07, 95% CI 1.01– 1.12), and I (Primary Industry Occupations) (HR ¼ 1.06, 95% CI 1.01–1.12), relative to all other SOC categories. Men employed in SOC category H occupations were at

Prostate Cancer and Occupational Exposure to Whole-Body Vibration

901

TABLE I. Baseline Characteristics of the Cohort (N ¼ 1,107,700) by Whole-Body Vibration Exposure Status Characteristic Age, at entry to cohort Mean (SD) 25^34 35^44 45^54 55^64 65^74 Income adequacy quintile Poorest Quintile 2 Quintile 3 Quintile 4 Richest Highest level of education No high school High school with or without trade certificate Post-secondary non-university University degree SOC 1991category (A) Management occupations (B) Business, finance, and administrative occupations (C) Natural and applied sciences and related occupations (D) Health Occupations (E) Occupations in social science, education, government service, and religion (F) Occupations in art, culture, recreation, and sport (G) Sales and Service Occupations (H) Trades, transport, and equipment operators and related occupations (I) Occupations unique to primary industry (J) Occupations unique to processing, manufacturing, and utilities Job type Blue collar White collar Prostate cancer Cases

WBV exposed (n ¼1,73,700)

WBV unexposed (n ¼ 9,34,000)

42.6 (11.8) 53,500 (30.8) 51,400 (29.6) 35,800 (20.6) 25,000 (14.4) 8,000 (4.6)

41.5 (11.2) 305,500 (32.7) 290,100 (31.1) 193,700 (20.7) 118,900 (12.7) 25,800 (2.8)

22,600 (13.0) 31,200 (18.0) 38,700 (22.3) 41,400 (23.8) 39,800 (22.9)

97,400 (10.4) 158,200 (16.9) 204,900 (21.9) 229,100 (24.5) 244,400 (26.2)

83,900 (48.3) 67,000 (38.6) 16,000 (9.2) 6,700 (3.9)

237,900 (25.5) 377,300 (40.4) 138,100 (14.8) 180,700 (19.3)

10,500 (6.0) 0 (0.0) 3,100 (1.8) 0 (0.0) 0 (0.0) 0 (0.0) 22,100 (12.7) 78,800 (45.4) 59,200 (34.1) 0 (0.0)

147,500 (15.8) 98,200 (10.5) 82,300 (8.8) 22,600 (2.4) 64,500 (6.9) 19,500 (2.1) 152,500 (16.3) 221,800 (23.7) 18,600 (2.0) 106,500 (11.4)

148,700 (85.6) 25,000 (14.4)

424,900 (45.5) 509,100 (54.4)

3,142 (1.8)

14,780 (1.6)

All figures are frequency (%) unless otherwise stated. Frequencies are rounded to the nearest100 in accordance with Statistics Canada disclosure rules. WBV,whole-body vibration; SOC, Standard Occupational Classification.

significantly reduced risk of prostate cancer (HR ¼ 0.93, 95% CI 0.90–0.97), compared to all other SOC categories.

DISCUSSION These results do not support an association between occupational WBV exposure and prostate cancer. We found no relationship between our WBV variables and prostate cancer in this population-based cohort. A previous casecontrol study found a borderline significant odds ratio between their WBV exposure variable and prostate cancer

(OR ¼ 1.44; 95% CI 0.99–2.09) using WBV exposure assigned by interview data and an exposure database [Nadalin et al., 2012]. Though we used a JEM that was based partly on the same exposure assessment, we did not replicate the findings of an association. It is possible that the case-control design allowed more precise assignment of exposure, due to the availability of a complete occupational history in that study compared with a one-time assessment in the current study. We did find some associations between WBV and prostate cancer in specific SOC categories. Exposed workers in SOC category C had a 37% significantly elevated risk of

902

Jones et al.

TABLE II. Associations (HR and 95% CI) Between WBVand Prostate Cancer Incidence, Canada, 1991^2003 Age and Age, province and province adjusted SES adjustedy Exposure Category WBV unexposed WBV low WBV high WBV unexposed WBV exposed SOC 1991category unexposed (A) Management occupations exposed (C) Natural and applied sciences and related occupations exposed (G) Sales and Service Occupations exposed (H) Trades, transport, and equipment operators and related occupations exposed (I) Occupations unique to primary industry exposed

Person-years at risk§ Cases 10,686,660 947,110 1,041,820 10,686,660 1,988,930 10,686,660 120,320 34,880 259,470 903,110 671,150

14,780 1,223 1,919 14,780 3,142 14,780 195 75 211 1,105 1,556

HR (95% CI)

HR (95% CI)

1.00 () 0.97 (0.92^1.03) 0.96 (0.91^1.00) 1.00 () 0.96 (0.93^1.00) 1.00 () 1.07 (0.93^1.23) 1.44 (1.14^1.80) 1.02 (0.89^1.17) 0.87 (0.82^0.93) 1.01 (0.95^1.06)

1.00 () 1.01 (0.95^1.07) 1.00 (0.95^1.05) 1.00 () 1.00 (0.96^1.04) 1.00 () 1.03 (0.90^1.19) 1.37 (1.09^1.72) 1.00 (0.88^1.15) 0.91 (0.86^0.97) 1.06 (1.00^1.12)

Boldface font indicates significance at the 95% confidence level. § Rounded to nearest10 in accordance with Statistics Canada disclosure rules.  Adjusted for age at entry (in10-year age categories) and province at baseline. y Adjusted for age at entry (in10-year age categories), province, income adequacy quintile, and highest level of education at baseline.

prostate cancer compared to all unexposed workers. This increased risk in exposed category C workers was not seen in category C as a whole. There are three occupation codes which were considered exposed in category C: Air Pilots Flight Engineers and Flying Instructors (C171), Deck Officers Water Transport (C173), and Engineer Officers Water Transport (C174). It is unlikely that WBV exposure would have a different effect in each SOC category, since

these categories are largely arbitrary, and therefore it is more plausible that other exposure(s) in least one of these occupations is driving the observed elevation in prostate cancer risk. This observation is consistent with other studies of pilots and navigators. Pilots were found to have an elevated risk of prostate cancer in two Canadian studies [Band et al., 1996, 1999], and a death certificate study also found that airplane pilots and navigators had a significantly

TABLE III. Associations (HR and 95% CI) Between SOC1991Category and Prostate Cancer Incidence, Canada, 1991^2003

SOC1991 category (A) Management occupations (B) Business, finance, and administrative occupations (C) Natural and applied sciences, and related occupations (D) Health occupations (E) Occupations in social science, education, government service, and religion (F) Occupations in art, culture, recreation, and sport (G) Sales and service occupations (H) Trades, transport, and equipment operators and related occupations (I) Occupations unique to primary industry (J) Occupations unique to processing, manufacturing, and utilities

Age and province adjusted

Age, province and SES adjustedy

Person-years at risk§

Cases

HR (95% CI)

HR (95% CI)

1,806,620 1,122,690 978,640 257,060 738,150 222,150 1,984,970 3,452,730 884,270 1,228,310

3,030 1,642 1,067 386 1,267 271 2,935 4,161 1,854 1,309

1.11 (1.07^1.15) 1.08 (1.03^1.14) 1.02 (0.96^1.08) 1.04 (0.94^1.16) 1.16 (1.10^1.23) 1.01 (0.89^1.13) 0.96 (0.92^1.00) 0.90 (0.87^0.93) 1.01 (0.96^1.06) 0.91 (0.86^0.96)

1.05 (1.01^1.09) 1.07 (1.01^1.12) 0.95 (0.89^1.01) 0.94 (0.85^1.04) 1.07 (1.00^1.14) 0.98 (0.87^1.11) 1.00 (0.96^1.04) 0.93 (0.90^0.97) 1.06 (1.01^1.12) 0.94 (0.89^1.00)

Note: A rolling reference group was used (i.e., risk in each occupational category was compared to the risk in all other occupational categories together). Boldface font indicates significance at the 95% confidence level. § Rounded to nearest10 in accordance with Statistics Canada disclosure rules.  Adjusted for age at entry (in10-year age categories) and province at baseline. y Adjusted for age at entry (in10-year age categories), province, income adequacy quintile, and highest level of education at baseline.

Prostate Cancer and Occupational Exposure to Whole-Body Vibration

increased risk of mortality from prostate cancer [Krstev et al., 1998a]. A Nordic cohort study of pilots also found an excess risk of prostate cancer in this occupation group, though the risk was not significant [Pukkala et al., 2003]. It is possible that this risk in pilots could be attributable to shared exposures other than WBV, such as night shift work or cosmic radiation. In water transport workers, the relationship is less consistent. A few studies have found that this group experiences an elevated risk of prostate cancer [Aronson et al., 1996; Sharma-Wagner et al., 2000; Pukkala et al., 2009], whereas another paper reports decreased risks in deck officers and water transport operators [Band et al., 1999]. We also found a small 9% significant reduction of prostate cancer risk in WBV exposed workers in SOC category H compared to all unexposed workers. Again, as it is unlikely that WBV increases the risk of prostate cancer in one category but decreases it in another, another exposure or random variation is likely responsible for this observed reduction in risk. Additionally, since the majority of workers who are exposed to WBV are employed in a category H occupation (45.4%), this observed negative association has a large influence on our null findings between WBV and prostate cancer. There are 17 exposed occupations in category H, including heavy equipment operators, transportation equipment operators, transportation equipment laborers, and some electrical trades. Our findings are unexpected, given that several other studies have found elevated risks of prostate cancer in heavy equipment and transportation equipment operators [Aronson et al., 1996; Krstev et al., 1998a,b; Brown and Delzell, 2000; Sharma-Wagner et al., 2000; Järvholm and Silverman, 2003; Nadalin et al., 2012]. However, a few studies have also found either a reduction in prostate cancer risk [Krstev et al., 1998a,b; Zeegers et al., 2004] or no association [Pukkala et al., 2009] in these occupations. A meta-analysis of driving occupations that included many of these aforementioned studies found a slightly increased non-significant relative risk of prostate cancer, though when they restricted their pooled analysis to studies from Canada only, they found a non-statistically significant 10% decreased risk of prostate cancer [Young et al., 2009]. It was a secondary objective of this study to look at the risk of prostate cancer in different occupational groups, independent of exposure to WBV. We found slightly elevated risks in three SOC categories. Workers employed in Management Occupations (A) and Business, Finance, and Administrative Occupations (B) have a small (5–7%) elevated risk of prostate cancer compared to workers in all other occupations. This observation is consistent with previous studies, which show that workers in these occupations have an elevated risk of prostate cancer [Band et al., 1999; Sharma-Wagner et al., 2000; Pukkala et al., 2009], possibly attributable to their sedentary working environment [Lynch, 2010]. We also found a slight elevated

903

risk of prostate cancer in Occupations Unique to Primary Industry (I), which is in agreement with other studies [Sharma-Wagner et al., 2000; Sass-Kortsak et al., 2007]. The reduced risk of prostate cancer that we saw in WBV-exposed workers in category H was apparent in this occupational group as a whole: those employed in Trades, Transport, and Equipment Operators and Related Occupations (category H) have a small 7% reduction in risk compared to all other workers. Again, this finding is contradictory to what other studies have found in this occupational category. SassKortsak et al. [2007] used the same occupational classification system and found that category H had a 20% increase in prostate cancer risk, though they were able to classify their participants based on longest occupation held. Another study, which recoded the SOC categories, found that transport equipment operators had a 90% significantly elevated risk [Nadalin et al., 2012]. Our study has several strengths. It is the first populationbased longitudinal study that considers the association between WBV and prostate cancer, in a large sample of the non-institutionalized Canadian population of working men. Our study therefore has strong generalizability to the Canadian population of working men age 25–74. To our knowledge, it is also the first large cohort study that looks at the risk of prostate cancer by occupation in Canada. Through the use of this study design we are able to capture all incident cases of prostate cancer in our cohort during 1991–2003 that are reported in the Canadian Cancer Database. While our exposure assessment was retrospective, we assigned exposure based on occupation data that was collected prospectively, rather than relying on recall. While we based our study and exposure assignment on previous studies [Sass-Kortsak et al., 2007; Nadalin et al., 2012], our exposure assessment presented several limitations. We relied on previous assessments [Nadalin et al., 2012] and expert review (PD) to assign exposure, as we are not aware of other existing, validated JEMs for WBV exposure. Unlike other exposure assessments undertaken for WBV [Harris et al., 2012], we had no specific information on individual exposures, and so we assigned the same WBV exposure to all individuals who were categorized in the same SOC occupation code. It is likely that individuals within the same occupational code experience heterogeneous levels of WBVexposure, resulting in non-differential misclassification between levels of exposure. There is also a chance of nondifferential misclassification occurring between the exposed and unexposed groups. However, this non-differential misclassification would attenuate associations and would not explain our findings of increased risk, but may have contributed to the null association we observed between WBV exposure and prostate cancer. The relative arbitrariness of SOC letter categories and heterogeneity of occupational exposures within these categories is another limitation, although we sought to stay consistent with the occupational

904

Jones et al.

categories previously found to increase risk [Sass-Kortsak et al., 2007; Nadalin et al., 2012]. We only had data on occupation at the time of cohort entry (the 1991 census), introducing a further source of error into our occupational exposure assessments as workers classified as exposed represent different durations of exposure and some of those classified as unexposed may have previously or subsequently been exposed to WBV. Assuming this dynamic is non-differential with respect to disease risk, this is a further conservative bias and would bias our results to the null effect. To account for job turnover and minimize non-differential misclassification, future longitudinal studies of occupational cancer could consider incorporating a measure of annual occupational stability derived from labor survey data (e.g., see Pukkala et al., 2005). This study only includes cancer diagnoses following cohort inception in 1991. Therefore it is possible that some participants had a cancer diagnosis prior to follow-up and this may affect the reported occupation at the time of the census. However, we would expect such misclassification to attenuate effects, since some of those who were cancer cases and truly exposed would be categorized as controls or unexposed. The cancer data available for the period of follow-up does not necessarily include all potentially clinically relevant variables (e.g., grade variables such as Gleason score). Although it is not fully understood how these variables could be related to etiology, it would be of interest for future studies with later follow-up to test their effects, as the Canadian Cancer Registry began collecting relevant grading variables in 2004. The data in the 1991 census did not allow us to control for all potential confounding factors such as family history of prostate cancer. However, though we know that family history is associated with the outcome, it is not likely associated with the exposure, and thus not a major source of concern. Lifestyle factors, which may be important in the relationship, such as physical activity and diet, could be confounding our estimates, however, the effect these factors have on the risk of prostate cancer is small, if any [Leitzmann and Rohrmann, 2012], and we controlled for SES as a proxy for these factors. Finally, this study is limited in the available follow-up period. Prostate cancer has a long latency and appears late in life [Brawley, 2012]. The 12 years of follow-up in this study limits the number of observable prostate cancer cases. However, this limitation is partly addressed by the large cohort size and the fact that a diversity of ages are represented at cohort entry. Future cohort studies of prostate cancer will be improved by the use of a longer follow-up period in order to better observe this relationship.

CONCLUSIONS While this study did not find an association between WBV and prostate cancer, it did identify some occupational

groups, such as air pilots and water transport occupations (both SOC category C occupations), which may have an elevated prostate cancer risk. Further research should be conducted on these occupations in order to identify any occupational exposures that are contributing to this risk. It will also be informative to conduct a more specific analysis of relevant occupations (i.e., category H: Trades, Transport, and Equipment Operators and Related Occupations) to determine, which particular jobs are driving the apparent reduction of prostate cancer risk in this category. Where possible, WBV exposure should be measured directly, or as precisely as possible, in order to more accurately determine the level and duration of WBV exposure each individual experiences.

ACKNOWLEDGMENTS We thank the authors of Nadalin et al. [2012] and particularly Alan Salmoni for providing details of the exposure assessment used in their analysis. Grant sponsor: Workers Safety and Insurance Board of Ontario; Grant number 11024.

REFERENCES Aronson KJ, Siemiatycki J, Dewar R, Gérin M. 1996. Occupational risk factors for prostate cancer: Results from a case-control study in Montréal, Québec, Canada. Am J Epidemiol 143:363–373. Band PR, Le ND, Fang R, Deschamps M, Coldman AJ, Gallagher RP, Moody J. 1996. Cohort study of Air Canada pilots: Mortality, cancer incidence, and leukemia risk. Am J Epidemiol 143:137–143. Band PR, Le ND, Fang R, Threlfall WJ, Gallagher RP. 1999. Identification of occupational cancer risks in British Columbia: Part II: A population-based case-control study of 1516 prostatic cancer cases. J Occup Environ Med 41:233–247. Beaulac JA, Fry RN, Onysko J. 2006. Lifetime and recent prostate specific antigen (PSA) screening of men for prostate cancer in Canada. Can J Public Heal Rev 97:171–176. Bosco C, Iacovelli M, Tsarpela O, Cardinale M, Bonifazi M, Tihanyi J, Viru M, De Lorenzo A, Viru A. 2000. Hormonal responses to wholebody vibration in men. Eur J Appl Physiol 81:449–454. Brawley OW. 2012. Prostate cancer epidemiology in the United States. World J Urol 30:195–200. Brown DA, Delzell E. 2000. Motor vehicle manufacturing and prostate cancer. Am J Ind Med 38:59–70. Canadian Cancer Society’s Steering Committee on Cancer Statistics. 2012. Toronto, ON: Canadian Cancer Statistics 2012. Cogliano VJ, Baan R, Straif K, Grosse Y, Lauby-Secretan B, El Ghissassi F, Bouvard V, Benbrahim-Tallaa L, Guha N, Freeman C, et al. 2011. Preventable exposures associated with human cancers. J Natl Cancer Inst 103:1827–1839. Darmon N, Drewnowski A. 2008. Does social class predict diet quality? Am J Clin Nutr 87:1107–1117. Dennis L, Lynch C, Torner J. 2002. Epidemiologic association between prostatitis and prostate cancer. Urology 60:78–83.

Prostate Cancer and Occupational Exposure to Whole-Body Vibration

905

Gann P, Hennekens C, Ma J, Longcope C, Stampfer M. 1996. Prospective study of sex hormone levels and risk of prostate cancer. J Natl Cancer Inst 88:1118–1126.

Peters PA, Tjepkema M, Wilkins R, Fines P, Crouse DL, Chan PCW, Burnett RT. 2013. Data resource profile: 1991 Canadian Census Cohort. Int J Epidemiol 42:1319–1326.

Giles-Corti B, Donovan R. 2002. Socioeconomic status differences in recreational physical activity levels and real and perceived access to a supportive physical environment. Prev Med 35:601–611.

Pukkala E, Aspholm R, Auvinen A, Eliasch H, Gundestrup M, Haldorsen T, Hammar N, Hrafnkelsson J, Kyyrönen P, Linnersjö A, et al. 2003. Cancer incidence among 10,211 airline pilots: A Nordic study. Aviat Space Environ Med 74:699–706.

Harris MA, Cripton PA, Teschke K. 2012. Retrospective assessment of occupational exposure to whole-body vibration for a case-control study. J Occup Environ Hyg 9:371–380. Helmkamp JC, Talbott EO, Marsh GM. 1984. Whole body vibration—A critical review. Am Ind Hyg Assoc J 45:162–167.

Pukkala E, Guo J, Kyyrönen P, Lindbohm M-L, Sallmén M, Kauppinen T. 2005. National job-exposure matrix in analyses of census-based estimates of occupational cancer risk. Scand J Work Environ Health 31(2):97–107.

Järvholm B, Silverman D. 2003. Lung cancer in heavy equipment operators and truck drivers with diesel exhaust exposure in the construction industry. Occup Environ Med 60:516–520.

Pukkala E, Martinsen JI, Lynge E, Gunnarsdottir HK, Sparén P, Tryggvadottir L, Weiderpass E, Kjaerheim K. 2009. Occupation and cancer—Follow-up of 15 million people in five Nordic countries. Acta Oncol 48:646–790.

Koutros S, Beane Freeman LE, Lubin JH, Heltshe SL, Andreotti G, Barry KH, DellaValle CT, Hoppin AJ, Sandler DP, Lynch CF, et al. 2013. Risk of total and aggressive prostate cancer and pesticide use in the Agricultural Health Study. Am J Epidemiol 177:59–74.

Roberts RO, Bergstralh EJ, Bass SE, Lieber MM, Jacobsen SJ. 2004. Prostatitis as a risk factor for prostate cancer. Epidemiology 15:93–99.

Krstev S, Baris D, Stewart P, Dosemeci M, Swanson GM, Greenberg RS, Schoenberg JB, Schwartz AG, Liff JM, Hayes RB. 1998. Occupational risk factors and prostate cancer in U.S. blacks and whites. Am J Ind Med 34:421–430. Krstev S, Baris D, Stewart P, Hayes RB, Blair A, Dosemeci M. 1998. Risk for prostate cancer by occupation and industry: A 24-state death certificate study. Am J Ind Med 34:413–420. Leitzmann MF, Rohrmann S. 2012. Risk factors for the onset of prostatic cancer: Age, location, and behavioral correlates. Clin Epidemiol 4:1–11. Li C-Y, Sung F-C. 1999. A review of the healthy worker effect in occupational epidemiology. Occup Med 49:225–229. Lynch BM. 2010. Sedentary behavior and cancer: A systematic review of the literature and proposed biological mechanisms. Cancer Epidemiol Biomarkers Prev 19:2691–2709. Milby TH, Spear RC. 1974. Relationship between whole body vibration and morbidity patterns among heavy equipment operators. Cincinnati, Ohio: National Institute for Occupational Safety and Health. Nadalin V, Kreiger N, Parent M-E, Salmoni A, Sass-Kortsak A, Siemiatycki J, Sloan M, Purdham J. 2012. Prostate cancer and occupational whole-body vibration exposure. Ann Occup Hyg 56:968–974. Palmer KT, Griffin MJ, Bendall H, Pannett B, Coggon D. 2000. Prevalence and pattern of occupational exposure to whole body vibration in Great Britain: Findings from a national survey. Occup Environ Med 57:229–236. Parent M-É, El-Zein M, Rousseau M-C, Pintos J, Siemiatycki J. 2012. Night work and the risk of cancer among men. Am J Epidemiol 176:751–759. Parent M-É, Siemiatycki J. 2001. Occupation and prostate cancer. Epidemiol Rev 23:138–143. Peters PA, Tjepkema M. 2010. 1991–2011 Census health outcomes follow-up. In: Social statistics: The interplay among censuses, surveys, and administrative data—Proceedings of the 2010 International Methodology Symposium (Catalogue 11-522-XCB). Ottawa, ON: Statistics Canada. pp. 150–156.

Sass-Kortsak AM, Purdham JT, Kreiger N, Darlington G, Lightfoot NE. 2007. Occupational risk factors for prostate cancer. Am J Ind Med 50:568–576. Sharma-Wagner S, Chokkalingam P, Malker HS, Stone BJ, McLaughlin JK, Hsing AW. 2000. Occupation and prostate cancer risk in Sweden. J Occup Environ Med 42:517–525. Statistics Canada. 1991. Standard Occupational Classification 1991. http://www.statcan.gc.ca/subjects-sujets/standard-norme/soc-cnp/1991/ soc-ctp91_menu-eng.htm (accessed 21 May 2012). Statistics Canada. 1992. 1991 Census Dictionary. Ottawa, ON. Wasserman DE, Erdreich J, Wilcox T, Doyle T, Spaeth S. 1984. Vibration exposure of heavy equipment operators in paperboard manufacturing. J Acoust Soc Am 76:6–7. Wilkins R, Tjepkema M, Mustard C, Choinière R. 2008. The Canadian census mortality follow-up study, 1991 through 2001. Health Rep 19:25–43. Young E, Kreiger N, Purdham J, Sass-Kortsak A. 2009. Prostate cancer and driving occupations: Could whole body vibration play a role? Int Arch Occup Environ Health 82:551–556. Zeegers MPA, Friesema IHM, Goldbohm RA, van den Brandt PA. 2004. A prospective study of occupation and prostate cancer risk. J Occup Environ Med 46:271–279.

Supporting Information Additional Supporting Information may be found in the online version of this article.

Institution at which the work was performed: Occupational Cancer Research Centre, CancerCare Ontario, 505 UniversityAvenue,17th Floor,Toronto,Ontario,Canada M5G1X3. Disclosure Statement: The authors report no conflicts of interests.

Prostate cancer and occupational exposure to whole-body vibration in a national population-based cohort study.

Following preliminary evidence from observational studies, we test the potential relationship between whole-body vibration (WBV) and prostate cancer i...
414KB Sizes 0 Downloads 3 Views