Volume 25
Number 3/4
2004

[Table of Contents]

Do work-related breast cancer risks in pre-menopausal women depend on family history?

Chris D Bajdik, Raymond Fang, Pierre R Band, Nhu Le and Richard P Gallagher

Abstract

Our objective was to determine work-related pre-menopausal breast cancer risks that depend on a woman's family history of the disease. In a large case-control study, 318 women with breast cancer and 340 healthy women completed a mailed questionnaire. All of the women were pre-menopausal and controls were matched to cases by age. All risk estimates were adjusted for women's smoking history and whether they reported a prior breast biopsy. There was an odds ratio (OR) of 6.9 (95% confidence interval: 1.5-31.9) for breast cancer among pre-menopausal women with no family history if they ever worked in material processing occupations. Among women with a family history of breast cancer, there was an OR of 6.4 (0.7-55.9) if they ever worked as miscellaneous salesclerks and salespersons of commodities, and an OR of 5.7 (0.6-50.9) if they ever worked in department stores. Despite changes in the OR, none of the estimates were significantly different in women with and without a family history.

Key words: breast cancer; disease susceptibility; family history; industry; occupation

Introduction

Breast cancer is the most commonly diagnosed malignancy among women in British Columbia (BC) Canada. About 2,700 women in BC were diagnosed with breast cancer in 2003 and about 640 died from the disease.¹ Established breast cancer risk factors include age, family history of breast cancer, previous benign breast disease, and hormonal factors such as age at menarche and parity.² A family history of breast cancer can occur because of shared genetic factors among relatives, shared environment among relatives, or by chance. Germline mutations in BRCA1 and BRCA2 account for 15%-20% of families in which there is more than one woman with breast cancer,³ and other genes may explain additional familial clustering.⁴ There is substantial variation in breast cancer incidence rates around the world⁵ and the probability of a family history also depends on local disease rates.⁶ (If disease rates are increased in a population, there is a greater probability that a family member will be affected.) Finally, the probability of familial clustering depends on the number, type, and age of relatives in a family.⁶

A large population-based case-control study of breast cancer was conducted in BC during 1988 and 1989.^7-9 That study reported a significantly-increased pre-menopausal breast cancer risk for electronic data processing operators, barbers and hairdressers, women in sales and material processing occupations, and in the food, clothing, chemical and transportation industries. Many studies have investigated the association between occupations and breast cancer, but the findings are inconsistent. Goldberg and Labreche reviewed 115 studies of occupation and breast cancer published between 1971 and 1994 and concluded that there was some, albeit limited, evidence of increased breast cancer risk among women working in the pharmaceutical industry, and among cosmetologists and beauticians.¹⁰ They also found some evidence of an increased risk for chemists, for workers with electromagnetic field exposure, and for workers exposed to solvents. An updated review in 2001 drew similar conclusions.¹¹

A family history of breast cancer might indicate that a woman has increased susceptibility to the disease. This seems particularly likely for pre-menopausal breast disease because an early age of diagnosis itself suggests predisposition. Several studies have found that women with a family history of breast or ovarian cancer have different breast cancer risks associated with reproductive factors than do women without a family history.^12-15 A recent US study found that women under age 50 with a family history of breast or ovarian cancer appeared to have increased breast cancer risks associated with medical radiation.¹⁶ We hypothesize that pre-menopausal women with a family history might have increased breast cancer risks as the result of work-related factors. In this analysis, we wanted to identify occupations and industries that differentially affect risk in women with and without a family history of breast cancer.

Methods

We analyzed data from a large case-control study of breast cancer.^7-9 Cases were all women under age 75 who were diagnosed with breast cancer in BC between June 1, 1988 and June 30, 1989. Controls were women who were randomly selected from the 1989 provincial voters list and age-matched to the cases. One thousand, four hundred and eighty-nine (1,489) cases were identified, and 1,502 controls were selected. All of these women were mailed a questionnaire that asked, among several things, about their occupational history and their family history of breast cancer. Upon its return, each questionnaire was reviewed for completeness and women were telephoned about missing information. A total of 1,018 cases and 1,025 controls completed the questionnaire, yielding a response rate of 68% in each group. Women were considered pre-menopausal if they were either menstruating or had undergone a hysterectomy without oophorectomy and were under 50 years of age. Women were considered post-menopausal if they had undergone natural menopause or a bilateral oophorectomy, or if they had undergone a hysterectomy with ovarian conservation but were 50 years of age or older. There were 318 cases and 340 controls who were pre-menopausal, and women participating in the study gave written, informed consent. Approval for the study was granted by the Behavioural Research Ethics Committee at the University of British Columbia.

The study questionnaire asked women whether they had ever had a breast biopsy, and the answer was interpreted as evidence of prior benign breast disease. It also asked women whether they had smoked more than 100 cigarettes in their lifetime, the age they began smoking, whether they had ever quit (and for how long), and the average number of cigarettes that they smoked each day.

With regards to their occupational history, the questionnaire asked women to describe every job they had ever held for a year or more. Details of each job included its title, the dates that it began and ended, whether the work was part-time or seasonal, the duties involved, the name of the employer and where the job was located. The questions were presented in a tabular format and an example was provided. This simplified both answering the questions and interpreting the responses. Occupations were coded according to the Canadian Standard Occupational Classification¹⁷ (SOC) and the Canadian Standard Industrial Classification¹⁸ (SIC). In each system, the most specific groups are defined by four-digit codes. Groups based on the first two or three digits of those codes define broader occupational and industrial classes. A positive family history of breast cancer was defined when women reported having a mother or sister with breast cancer. The women in this analysis were mostly Caucasian with some post-secondary education, and had been pregnant at least once.⁹ The main differences between the case and control groups were 1) 17% of cases and only 8% of controls had a prior breast biopsy (suggesting prior benign breast disease; p < 0.05 for difference) and 2) 23% cases reported more than 20 pack- years of cigarette smoking as compared to 16% as reported by controls (p < 0.1).9 The data are the same as those described in the original analysis of occupational breast cancer risk (i.e., ref 9).

We restricted the analysis to occupational and industrial groups in which there was at least one case and one control with and without a family history of breast cancer, and at least five women with a family history and five women without a family history in total. These restrictions were adopted to avoid conclusions based on the absence of observations (although this itself might be considered evidence) and provide a minimum level of statistical stability in the results. Following these restrictions, we analyzed risks for four-digit occupational and industrial codes. When there were insufficient data in a four-digit group, broader groups defined by three-digit and two-digit codes were used.

Logistic regression was used to calculate the odds ratio (OR) as an estimate of the relative risk for breast cancer. Separate analyses were performed for women with and without a family history of breast cancer. We estimated relative risks adjusting for age (in years) and factors that were significant in a previous analysis:⁹ pack-years of cigarettes smoked (in three categories: 0, <20 and >= 20) and history of breast biopsy (any or none). Logistic regression was also used to test whether there was a significant interaction between family history and occupation/industry. All tests were two-sided with an alpha of 5%.

Results

Sixteen percent (16%) of cases and 7% of controls had a family history of breast cancer. The OR estimates associated with occupation and industry groups are given in Table 1. The mean OR for all occupational and industrial groups was 1.2 (range: 0.4-4.2), adjusting for family history. Restricting analyses to women with a family history, the mean OR for the same groups was 1.6 (range: 0.3-6.4), and 1.3 (range: 0.3-6.9) when the analysis was restricted to women without a family history. There were two occupational groups and one industrial group in which there was a roughly five-fold or greater difference in breast cancer risk for women with and without a family history of the disease. Among women with a family history of breast cancer, there was an OR of 6.4 if they ever worked for a year or more as miscellaneous salesclerks and salespersons of commodities (SOC 5135). The corresponding OR was 1.3 for women without a family history. This job group excludes supervisors, technical sales occupations and commercial travellers. Among women with a family history of breast cancer, there was an OR of 5.7 if they ever worked for a year or more in department stores (SIC 6411). The corresponding OR was 0.9 for women without a family history. Women without a family history of breast cancer had an OR of 6.9 if they ever worked for a year or more in material processing occupations (SOC 81/82). The corresponding OR was 0.6 for women with a family history. This job group includes occupations involving the processing of mineral ores, metals, stone, chemicals, rubber, plastic, textiles, food and wood. All models were adjusted for women's smoking history and whether they reported a prior breast biopsy. Models fitted using all women yielded no significant interactions involving family history and occupational group, industrial group, or other model variable.

TABLE 1
Relative risk of pre-menopausal breast cancer associated with ever being employed in an occupation or industry for one year or more: number of cases, number of controls (Cntls), odds ratio (OR) and 95% confidence interval (95% CI). A family history is defined as having a mother or sister who ever had breast cancer. Occupation and industry groups are ordered according to the first 2 digits of the SOC or SIC code, then the 3rd digit and 4th digit where they are present

Discussion

After adjustment for known risk factors, we identified occupational and industrial groups for which there was a substantial increase in pre-menopausal breast cancer risk depending on whether women had a family history of the disease. There was a seven-fold increase in the relative risk of breast cancer among pre-menopausal women with no family history who were ever employed in material processing occupations. There was a six-fold increase in the relative risk of breast cancer among pre-menopausal women with a family history if they had ever worked in department stores, and a five-fold increase if they had ever worked as miscellaneous salesclerks and salespersons of commodities. None of the differences between women with and without a family history were statistically significant. Earlier analysis of this data found significantly increased OR estimates of 1.7 and 4.2 for breast cancer among pre-menopausal women who ever worked as miscellaneous sales clerks and salespersons, and in materials processing occupations - regardless of whether they had a family history of breast cancer.⁹ Our analysis indicates these risks are highly dependent upon whether the woman has a family history.

A family history of breast cancer can result from genetic or environmental factors, but individual cases within a family might not be caused by the same exposures. All of the women in a family might have inherited susceptibility, but the co-factor that exploits this susceptibility could come from several sources. The genes responsible for a family history might not affect cancer directly, but rather affect a process such as DNA repair, chemical metabolism or cell cycling. Where members of a family share a genetic deficiency in DNA repair capability, DNA damage might occur as the result of exposure to sunlight or pesticides. A family history might also be the result of a non-genetic factor such as shared environment or shared lifestyle amongst family members.

The hypothesis that motivated our analysis is that family history of breast cancer is a marker of increased disease susceptibility and, as a consequence, women with a family history might experience greater risks as the result of work-related exposures. Our results are consistent with a recent study of disease risk in the relatives of women with breast cancer.¹⁹ This study concluded that a high proportion of breast cancers arises in a susceptible minority of women. Other studies suggest that the susceptibility might be due to multiple alleles that affect processes related to hormones or tumor suppression.^4,20,21 Work in department stores might entail elevated exposure to artificial lighting. This is consistent with the hypothesis that sunlight exposure decreases breast cancer risk. If the work is performed in the evening, the result is also consistent with the hypothesis that light-at-night (LAN) increases a woman's breast cancer risk. Unfortunately, we had no data as about LAN or sunlight exposure to test these hypotheses directly.

For either the LAN or sunlight hypothesis, the increased risk might only be applicable to women with a susceptibility to the exposure, and where the susceptibility is familial. Alternatively, the results could be due to a confounding factor that is familial, such as a tendency towards indoor or sedentary work. Many people find work through a family member's employer and factors that affect familial clustering might also appear to affect their occupation. Furthermore, a woman's child-rearing experience has a direct effect on her work experience outside the home. It has been suggested that working women might have fewer children, have their first child at a later age, and breast-feed their children less.²² We attempted to adjust for this phenomenon by estimating breast cancer risks after correcting for known confounders identified in earlier analyses. This also allowed us to compare the relative risk estimates with those of earlier results. Recent studies in Canada²³ and Finland²⁴ have shown that breast cancer risk is reduced for women who are more physically active. The role of physical activity is especially interesting because of the apparent complexity of body-weight's effect in determining breast cancer risk. Unfortunately, information about physical activity was not available in our study and we could not test for this factor's effect directly.

We adjusted our analyses for factors that were significant in earlier analyses of the same data, thereby allowing us to compare the results. Our analysis considered only pre-menopausal breast cancer risk, but not whether a woman's family history of breast cancer was a post-menopausal or pre-menopausal one. A family history of breast cancer is dependent on a woman's age and family size.¹² Someone with no family history of cancer at age 20 might have several affected relatives by age 40. Likewise, someone from a large family is more likely to have an affected relative than is someone from a small family, unless increased family size is somehow associated with a reduced disease risk. The BC study matched case and control groups by age. The study did not match the groups based on family size, but this is not expected to influence results unless family size is somehow associated with a woman's job history.

The information used in the study was obtained by a self-administered questionnaire, and reporting bias can occur if women with breast cancer respond to questions differently than healthy women. The questions were worded to minimize this phenomenon, and the questionnaire was designed with the same intention. Occupational and industrial risks were based on job titles, and these titles might encompass jobs with very different duties and exposures. Finally, caution is necessary when interpreting these results because of the number of occupations and industries we considered. In a case-control study, even a reasonably large one, the number of women reporting any particular occupation or industry will be small. The numbers will be smaller still when split according to whether or not a woman has a family history of breast cancer. Our main analysis estimated relative risk estimates separately for women with and without a family history of breast cancer. We also fitted models for each of the 26 occupational and industrial groups to test for an interaction between women with and without a family history, and none of the differences was statistically significant. However, three of the differences were roughly five-fold or more and this should not be ignored.

In studies of breast cancer and occupation, few associations have been found consistently. The lack of consistency might be explained by a difference in the underlying susceptibility of women. The allowable exposures in a work-place are typically based on risks that affect all employees. An alternative strategy would be to set exposure limits based on the most susceptible group.

Acknowledgements

This research was partially supported by a grant from the Workers' Compensation Board of British Columbia. Chris Bajdik is supported by a Scholar Award from the Michael Smith Foundation for Health Research.

References

1. NCIC. National Cancer Institute of Canada: Canadian Cancer Statistics 2003. Toronto, 2003.
2. McPherson K, Steele CM, Dixon JM. Breast cancer - epidemiology, risk factors, and genetics. Br J Med 2000;321:624-8.
3. Nathanson KN, Wooster R, Weber WL. Breast cancer genetics: What we know and what we need. Nat Medicine 2001;7:552-6.
4. Pharoah PDP, Antoniou A, Bobrow M, Zimmern RL, Easton DF, Ponder BAJ. Polygenic susceptibility to breast cancer and implications for prevention. Nat Genet 2002;31:33-6.
5. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J. Cancer incidence in five continents. volume VII. IARC Scientific Publications, Lyon, 1997.
6. Bajdik CD. Family history of breast/ovarian cancer and referral criteria for a BRCA1 testing program. PhD thesis, Department of Health Care and Epidemiology, Faculty of Medicine, University of British Columbia, 2001.
7. Yang CP, Daling JR, Band PR, Gallagher RP, White E, Weiss NS. Noncontraceptive hormone use and breast cancer. Cancer Causes Control 1992;3:475-9.
8. Yang CP, Weiss NS, Band PR, Gallagher RP, White E, Daling JR. History of lactation and breast cancer risk. Am J Epidemiol 1993;138: 1050-6.
9. Band PR, Le ND, Fang R, Deschamps M, Gallagher RP, Yang P. Identification of occupational cancer risks in British Columbia. A population-based case-control study of 995 incident breast cancer cases by menopausal status, controlling for confounding factors. J Occup Environ Med 2000;42:284-310.
10. Goldberg MS, Labreche F. Occupational risk factors for female breast cancer: a review. Occup Environ Med 1996;53:145-56.
11. Labreche F. Chapter 6. Literature review of associations between breast cancer and occupational exposures. In: Review of lifestyle and environmental risk factors for breast cancer: Report of the Working Group on primary prevention of breast cancer. Health Canada, 2001.
12. Hill DA, Preston-Martin S, Ross RK, Bernstein L. Medical radiation, family history of cancer, and benign breast disease in relation to breast cancer risk in young women, USA. Cancer Causes Control 2002;13:711-8.
13. Colditz GA, Rosner BA, Speizer FE. Risk factors for breast cancer according to family history of breast cancer. J Natl Cancer Inst 1996;88:365-71.
14. Andrieu N, Prevost T, Rohan TE, Luporsi E, Le MG, Gerber M et al. Variation in the interaction between familial and reproductive factors on the risk of breast cancer according to age, menopausal status, and degree of familiality. Int J Epidemiol 2000;29:214-23.
15. Olsson H, Bladstrom A. A cohort study of reproductive factors and family history of breast cancer in southern Sweden. Breast Cancer Res Treat 2002;76:203-9.
16. Becher H, Schmidt S, Chang-Claude J. Reproductive factors and familial predisposition for breast cancer by age 50 years. A case-control-family study for assessing main effects and possible gene-environment interaction. Int J Epidemiol 2003;32:38-48.
17. Statistics Canada. Standard Occupational Classification 1980. Catalogue No 12-565-E.
18. Statistics Canada. Standard Industrial Classification 1980. Catalogue No 12-501-E 528.
19. Peto J, Mack TM. High constant incidence in twins and other relatives of women with breast cancer. Nature Genetics 2002;26:411-4.
20. Pharoah PD, Lipscombe JM, Redman KL, Day NE, Easton DF, Ponder BA. Familial pre-disposition to breast cancer in a British population: implications for prevention. Eur J Cancer 2000;36:773-9.
21. Mack TM, Hamilton AS, Press MF, Diep A, Rappaport EB. Heritable breast cancer in twins. Br J Cancer 2002;87:294-300.
22. Threlfall WJ, Gallagher RP, Spinelli JJ, Band PR. Reproductive variables as possible confounders in occupational studies of breast and ovarian cancer in females. J Occ Med 1985;27:448-50.
23. Friedenreich CM, Courneya KS, Bryant HE. Influence of physical activity in different age and life periods on the risk of breast cancer. Epidemiology 2001;12:604-12.
24. Rintala PE, Pukkala E, Paakkulainen HT, Vihko VJ. Self-experienced physical workload and risk of breast cancer. Scand J Work Envivon Health 2002;28:158-62.

Author References

Chris D Bajdik, Raymond Fang, Nhu Le, Richard P Gallagher, Cancer Control Research Program and Occupational Oncology Research Program, BC Cancer Agency, Vancouver, BC, Canada

Pierre R Band, Occupational Oncology Research Program, BC Cancer Agency, Vancouver, BC, Canada

Correspondence: CD Bajdik, Cancer Control Research Program, BC Cancer Agency, 600 West 10th Avenue, Vancouver, BC, Canada V5Z 4E6; Fax: (604) 877-1868; E-mail: cbajdik@bccancer.bc.ca

[Previous] [Table of Contents] [Next]