- Email: [email protected]
Environmental features influencing the epidemiology of breast cancer
The Breast (2001) 10 (Supplement 3), 1··-8 r 2001 Harcourt Publishers Ltd
1. BREAST CANCER EPIDEMIOLOGY AND THE MEDIA
Environmental features influencing the epidemiology of breast cancer P. Boyle l and T. Zheng 2 JDivision of Epidemiology and Biostatistics, European Institute of Oncology, Milan, Italy,· 2 Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven CT, USA
SUM MAR Y. A recent study claimed that a larger proportion of breast cancer was due to environmental exposures than to hereditary factors. This controversial publication clearly indicates the essential importance of investigating the gene-environment interaction and having a clear idea of what are the essential environmental factors to investigate. Environmental exposures are a subject of current interest in breast cancer aetiology and two common issues are exposure to electromagnetic fields (EMFs) and certain pesticides, fungicides and their residues. Overall, there does not appear to be an identifiable increased risk associated with domestic exposure to EMF, although there is a need for good studies with direct estimates of exposure. Oestrogens from a variety of sources, including chemical exposures, are determinant in breast cancer risk in women. Attention has recently increasingly focused on environmental exposures and environmental chemical pollutants as risk factors for breast cancer. The common nature of breast cancer in women, and such exposures, make such an association plausible, although far from proven as causal. Given also the resistant nature of concentrations in adipose tissue, it is important to attempt to identify any risk resulting from this exposure. Based on current evidence, the risk of breast cancer associated with exposure to organochlorine pesticides and their residues appears to be small, if it exists at all. However, one caveat is that the effect of individual or groups of PCBs as risk factors for breast cancer merits some further investigation. Indeed, a clear and powerful message which comes from recent work is that great care and attention must be given to any statistical analysis of such chemical exposures given the potential for collinearity which exists between the various chemical compounds and their congeners. © 2001 Harcourt Publishers Ltd
INTRODUCTION
The debate about the portion of cancer that can be ascribed to nature or nurture has been ongoing for many decades. Doll and Peto l concluded that there was compelling evidence that upwards of 80%, or even 90% of human cancers had environmental causes and could be avoidable. Different populations around the world theoretically experience different levels of different forms of cancer and these levels change with time in an orderly, and frequently predictable, manner. Population groups whose lifestyle differentiates themselves from members of the same population frequently experience different levels of cancer (e.g. Mormons and Seventh Address correspondence to: Prof. Peter Boyle, Divisione di Epidemiology Biostatistics, European Institute of Oncology, Via Ripamonti 435,20141 Milano, Italy. Fax: + 39 02 5748 9922; E-mail: [email protected]
Day Adventists in the USA). Migrant groups from one country to another quickly adopt the cancer profile of their new home: thus the Japanese who left Japan with its low levels of breast cancer quickly attained the levels of breast cancer in the population of California within two generations/. Apart from certain types of familial cancer, such as those arising in adenomatous polyposis coli, the contribution of hereditary factors to the development of cancer is generally held to be small. 3- 5 This, of course, refers to dominant genes which have been investigated in family studies which go back two or three generations while no determination has yet been made of the contribution of low-penetrance single genes, recessive genes and mechanisms that involve several genes, which are much more difficult to investigate. Studies of twins open one possibility of estimating the contribution of inherited genes to the development of
2
Adjuvant Therapy of Primary Breast Cancer
cancer. A combined study of the Twin Cohorts from Sweden, Denmark and Finland 6 assembled data on nearly 45 000 pairs of twins. Controversy has followed this publication which attempted to gauge the relative impact of genetic constitution and environmental exposure in the causation of cancer." The authors used 45000 pairs of twins in Scandinavia to investigate the contribution of inherited and environmental factors by comparing the risk of cancer arising in identical (monozygotic) twins (who have the same set of genes) with that of fraternal (dizygotic) twins (who share about 50% of their genes). Among the major forms of cancer, estimates of the genetic component ranged from 270/0 for breast cancer to 42% for prostate cancer (Table 1). Overall, the twin of a woman affected by breast cancer had an increased risk of having the same cancer. The twin of a monozygotic twin who had breast cancer had a risk of breast cancer which was 5.2 times that of a monozygotic twin of a person without breast cancer. The concordance for breast cancer in female monozygotic twins was 0.14, which means that there is a 14% probability that the identical twin of a woman with breast cancer will have the same cancer.' The authors then attempted to calculate the proportion of phenotypic variance accounted for by inherited genetic differences among persons (heritability), by shared environmental risk factors (the proportion of phenotypic variance accounted for by environmental factors shared by both twins, thus contributing to similarity between them) and non-shared environmental factors (the proportion of phenotypic variance accounted for by environmental factors causing differences between twins) (Table 1). These estimates of heritability in the twin study were greater than those previously derived from family-based surveys of cancer. However, the publication was widely interpreted as demonstrating that inherited genes play only a limited role in the causation of cancer, in contrast to lifestyle and other environmental exposures. The timing of this publication, coming soon after the announcement that scientists were rapidly nearing complete deciphering of the human genome, raised concerns in the media that the attention devoted to the Human Genome project and to the study of cancerassociated genes might be misplaced. Twin studies have been useful for a long time for distinguishing the relative importance of inherited and environmental factors in disease. Lichtenstein et a1. 6 can be criticized since the statistical methods employed in this Twin Study assumed that genes and environment were acting in isolation and that 100% of cancer risk needs to be explained. There is increasing evidence that genetic makeup influences susceptibility or even resis-
Table 1 Effects of heritable and environmental factors in cancer in twins in Denmark, Sweden and Finland Environmental Cancer Stomach Colo rectum Lung Breast Ovary Prostate
Heritable
Shared
28% 35% 26% 27% 22% 42%
10% 5% 12% 6% 0% 0%
NonShared 62% 60% 62% 67% 78% 58%
Data abstracted from Lichtenstien et a1.6
tance to cancer-causing exposures and this study emphasizes the necessity for epidemiological studies to focus on the gene-environment interaction. BREAST CANCER: ENVIRONMENTAL AND GENETIC FACTORS Conventional risk factors for breast cancer include hormonally-related factors (including nulliparity, age at menarche, age at first and last births, age at menopause and oophorectomy), use of exogenous oestrogens (oral contraceptives and hormone replacement therapy) and dietary factors (including alcohol consumption and intake of vegetables and fruits). More recently, attention to a series of environmental exposures has increased. Notable among these have been electro magnetic fields (EMF) and certain pesticides, particularly organochlorine pesticides. These factors have been carefully studied in a case-control study conducted in New Haven, Connecticut which will be used below to illustrate the recent findings in this area of research. Cases were histologically confirmed, incident cases of breast cancer who either had breast-related surgery at the Yale-New Haven Hospital, or who were residents of Tolland Country, between 1 January 1994 and 31 December 1997. A total of 561 incident breast cancers were identified in Yale-New Haven Hospital with 432 (77%) completing in-person interviews. From the hospital database, 569 potential control patients who had breast-related surgery and who were diagnosed histologically with normal tissue or benign breast diseases (excluding atypical hyperplasia), were selected. Of these, 404 patients (71 %) participated in the study. In addition, 238 cases was identified among residents of Tolland County, of which 176 (74%) completed interviews. A total of 205 controls were also recruited. A total of 490 women (304 cases and 186 controls) aged 40-79 years had 0.4 g of breast adipose tissue
Environmental features in epidemiology available for chemical analysis. This was frozen and sent to Colorado State University for chemical analysis.
3
with more direct measures of EMF exposure, would be welcomed. However, it can be concluded that within our current knowledge, domestic exposures to EMF do not present a major source of breast cancer risk.
ELECTROMAGNETIC FIELDS Chronic exposure to EMFs has been hypothesized to increase female breast cancer risk." Although no studies with direct measures have been reported, household electrical appliances frequently provide the highest magnetic field exposure levels in any residence, with electric blankets among the strongest sources. Limited epidemiological studies investigating electric blanket use and breast cancer risk have produced inconclusive results. A case-eontrol study from Upstate New York found a borderline significantly increased risk of breast cancer for postmenopausal women (OR = 1.5, 95% CI 1.0-2.2) and premenopausal women (OR = 1.4, 95% CI 0.9-2.2) among those who reported using electric blankets continuously throughout the night compared to never-users. When the pre- and postmenopausal groups were combined, the odds ratio was 1.5 (1.1-1.9).9 Subsequently, Gammon et al.!" found an odds ratio of 1.0 (0.9-1.2) for breast cancer in women age 45 and an odds ratio of 1.1 (0.9-1.4) for those 45 years and over, among women who reported using electric blankets, mattress pads or heated water beds when compared to never users. Coogan and Aschengrau II reported an odds ratio of 1.0 (0.7-1.4) among women who reported using an electric blanket compared with women who did not. In Connecticut, 40% of breast cancer cases, and 430/0 of controls, reported regular use of electric blankets which produced an odds ratio of 0.9 (0.7-1.1).12 For those who used an electric blanket continuously throughout the night, the adjusted odds ratio was 0.9 (0.7-1.2) when compared to never users. Risk did not vary according to age at first use, duration of use, or menopausal and oestrogen receptor status. Besides electric blankets, there are a number of other household appliances that could be significant sources for domestic exposure to EMF. There was no association found in the Connecticut study between breast cancer risk and use of domestic appliances including a hair dryer, curling iron, electric razor, electric toothbrush, vacuum cleaner, popcorn maker, computer or electric iron. 12 Increasing concerns that exposure to EMFs from domestic appliances may increase breast cancer risk appear to be unfounded from results obtained from recent epidemiological studies. This does not rule out the presence of any risk and further studies, somehow
PESTICIDE EXPOSURE, PESTICIDE RESIDUES AND BREAST CANCER RISK Organochlorine compounds including organochlorine pesticides, have been suggested to be associated with the risk of breast cancer in women. However these studies have generally been small, with low statistical power, and have unsurprisingly produced inconsistent findings. It is, however, biologically plausible that exposure to these environmental contaminents may increase breast cancer risk, since some organochlorine compounds are animal carcinogens, oestrogenically active and induce P450 mixed-function oxidase enzymes, which are intimately involved in steroid hormone metabolism. While epidemiological studies have chiefly focused on polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT), there are a number of related compounds which could potentially be involved. These include trans-nonachlor (TNC) and Oxychlordane (OCD), a major metabolite of the insecticide chlordane and Hexachlorobenzene (RCB), organochlorine fungicide, and one of its isomers B-benzene hexachloride (BHC). Oxychlordane (OCD) and trans-nonachlor (TNC) Oxychlordane (OCD) is a major oxidative metabolite of the insecticide chlordane. TNC is a component of technical chlordane and technical heptachlor. Chlordane and heptachlor were primarily used for the protection of buildings, lawns and gardens from soil insects and termites. Chlordane was first produced commercially in the USA in 1947 as an insecticide. Heptachlor was isolated from technical chlordane and, aside from the above uses, has also been used to control mosquitos. Although the use of chlordane has been suspended since 1988, and heptachlor restricted in its use, significant exposure to OCD, TNC and their metabolites has occurred, since they are highly persistent in the environment. Residues of these compounds have been detected throughout the food chain and in human tissues. Their oestrogenic activity and the persistent and widespread nature of their exposure, indicate these as potential breast cancer risk factors. Two small studies have examined the association between OeD and TNC~
4
Adjuvant Therapy of Primary Breast Cancer
Table 2 Breast cancer risk associated with breast adipose-tissue levels of oxychlordane and trans-nonachlor Odds ratio
95%
CI
Oxychlordane <26.0 26.0-33.6 33.7-47.5 2:47.6 p for trend
1.0 0.7 0.7 0.7 0.29
0.4-1.2 0.4-1.2 0.4--1.2
Trans-nonachlor <36.4 36.4--53.1 53.2-71.0 2:71.1 P for trend
1.0 1.2 0.7 1.1 0.44
0.7-2.1 0.4-1.3 0.6-1.9
Measurements in parts per billion (ppb). Odds ratio adjusted for age, lifetime months of lactation, age and menarche, age at first fullterm pregnancy and race. Data abstracted from Zheng et a1.15
and breast cancer risk, Dewailly et al.':' found a nonsignificant higher mean adipose-tissue level of TNC in nine ER + ve breast cancer cases (50.3 ppb) then in 17 (benign breast disease) (BBD) controls (31.1 ppb) (P = 0.07). Falck et al. 14 found non-significantly higher breast adipose-tissue levels of OCD and Heptachlorepoxide in 20 breast cancer cases (136 ppb) than in 20 BBD controls (121 ppb). However, TNC was found to be nonsignificantly lower among the cases (103 ppb) than in the control (118 ppb). Zheng et al. 15 reported an analysis of breast adipose tissue of 304 cases of breast cancer and 186 BBD controls. The age and lipid-adjusted geometric mean adipose-tissue levels of OCD were similar between cases (36.4ppb) and controls (38.0ppb). The age and lipidadjusted geometric mean adipose tissue levels of TNC between cases (55.5ppb) and control (58.1 ppb) were also similar. There was no association between breast cancer risk and mean adipose tissue levels of OCD and TNC (Table 2). The covariate adjusted odds ratio (OR) was 0.7 (0.4-1.3) for OCD and 1.1 (0.6-1.9) for TNC, when the highest quartile was compared to the lowest (Table 2). The association did not differ by menopausal status, by histological type of breast cancer, disease stage at diagnosis or by history of lactation nor were tissue levels of OCD and TNC different by ER and PR status.i ' These findings do not support an overall association between breast adipose-tissue levels of OCD and TNC and risk of breast cancer in women, Such findings are consistent with those of other recent studies that environmental exposures to various environmental oestrogens does not seen to have an overall significant impact on the risk of breast cancer.
Hexachlorobenzene and B-benzene hexachloride Hexachlorobenzene (HCB) is an organochlorine fungicide which has been widely used for several decades mainly as a seed protectant on grain, wheat and field crops.!? Although it is no longer produced commercially, due to its resistance to chemical and biological breakdown HCD has contaminated air, water, soil and plants and has been detected in mammals and humans. Current exposure to humans is primarily through the food chain. Three small studies have found no significant difference in adipose tissue levels of HCB between breast cancer cases and controls 13,14, 17 although a recent case-control study'" reported a non-significantly elevated odds ratio of 1.8 (95% CI 0.6-5.4) for those with the highest serum levels of HCB, when compared to those with the lowest levels among parous women who never lactated. Zheng et al. 16 reported no significant difference in mean adipose-tissue levels of HCB between breast cancer patients (21.0 ppb) and BBD controls (19.1 ppb) in a large study involving 304 breast cancer cases and 186 controls. Risk did not vary according to menopausal status, oestrogen or progesterone receptor status, breast cancer histology, stage of diagnosis or type of benign breast disease. 16 Residues of B-benzene hexachloride (B-BHC), one of the isomers of hexachlorobenzene, has been reported to be higher in breast fat tissue from breast cancer patients than from controls.l ' After adjustment for age and parity, breast cancer tissue containing more than 0.1 mg/ kg fat of B-BHC was associated with an odds ratio of 10.5 (95% CI 2.0-55.3) for breast cancer compared with tissue containing 0.1 or less.l ' Zheng et al. 19 found no significant difference in breast adipose tissue levels of PBHC between 304 cases and 186 controls either overall or when broken down by menopausal status or oestrogen or progesterone receptor levels. There seems little epidemiological support at present for an association between either hexachlorobenzene (HCB) or ~-benzene hexachloride (P-BHC) and risk of breast cancer.
DDT and DDE Dichlorodiphenyltrichloroethane (DDT) and its most stable metabolite dichrlorodiphenyldichloroethane (DDE) have both been suggested to be associated with an increased risk of breast cancer. Again, this is biologically plausible given their oestrogenic activity and ability to induce P450 enzymes, which are intimately associated with steroid hormone metabolism.r"
Environmental features in epidemiology Falck et al.!" reported from study of 20 breast cancer cases and 20 BBD controls, that a 10 ppb increase in adipose-tissue levels of DDE corresponds to a 10/ 0 increase in breast cancer risk. Dewailly et al. 13 estimated a 8.9-fold increase in breast cancer risk among women whose DDE levels were above 1292 ppb in adipose tissue compared to those with a level of 238 ppb or lower, based on a study of 18 breast cancer cases and 17 BBD controls. Wolff et a1. 2 1 conducted a nested case-eontrol study with 58 cases and reported a 4-fold increase in risk of breast cancer for an elevation of serum DDE concentrations from 2.0ng/ml (10th percentile) to 19.1 ng/ml (90th percentile). In another nested case-eontrol study, involving 50 Black and 50 Asian women with breast cancer, a 2- to almost 4-fold increased risk of breast cancer was reported for Black women and White women when comparing women in the third tertile with those in the lowest tertile of DDE in serum. There was no increased risk with higher serum levels observed in Asian women in this trial. 22 More recent studies with larger sample sizes, ranging from 141-265 cases, do not support the association 18,23,24 with one study reporting a significant inverse between adipose tissue levels of DDE and breast cancer risk.25 Breast adipose tissue was collected from 304 breast cancer cases and 180 benign breast disease controls from Yale-New Haven Hospital. DDE and DDT in adipose tissue were measured using gas chromatography.i" Statistical significance for multiple means of tissue levels of DDE and DDT was calculated using analysis of variance and rank sum tests. A logistic regression model was used to control confounding and estimate the exposure/disease association. The age-adjusted geometric mean adipose tissue levels of DDE for cases (736.5ppb) was not significantly different (P = 0.40) than that for the controls (784.1 ppb). Mean tissue level of DDT was nearly identical for cases (51.8 ppb) and controls (55.6ppb). DDE and DDT levels did not vary significantly based on Oestrogen/progesterone receptor status of the cases, menopausal status, histology, stage, or type of benign breast disease. The adjusted OR is 0.9 (0.5-1.5) for DDE and 0.8 (0.5-1.5) for DDT when the highest quarter was compared with the lowest quarter (Table 3). There are two major contradictory hypothesis around at the present time. Firstly, women and men now basically live in an environment that is a virtual sea of oestrogens and one long-term effect of low-level exposure to these environmental oestrogens could be to increases the risk of breast cancer in women.r" Alternatively, and completely contradictory, it is argued that the exposure to such relatively low levels of DDE
5
Table 3 Breast cancer risk associated with breast adipose-tissue levels of dichlorodiphenyldichloroethane (DDE) and dichlorodiphenyltrichloroethane (DDT) Odds ratio Dichlorodiphenyldichloroethane (D DE) <412.6 1.0 412.6-779.2 1.3 779.3-1355.9 0.9 ~1356.0 0.9 p for trend 0.46 Dichlorodiphenyltrichloroethane (DDT) <28.7 1.0 28.7-49.4 0.8 49.5-78.0 0.6 ~78.1 0.8 P for trend 0.38
0.7-2.2 0.5-1.6 0.5-1.5
0.5-1.4 0.4-1.1 0.5-1.5
Measurements in parts per billion (ppb). Odds ratio adjusted for age, body mass index, lifetime months of location, age at menarche, age at first full-term pregnancy, menopausal status, income and race. Data abstracted from Zheng et a1. 26
and other weakly oestrogenic organochlorine compounds may not have a significant effect on a woman's risk of breast cancer. 28 Experimental studies have provided a substantial body of evidence that DDE and DDT are not mammary carcinogens.i" Laboratory studies also show that the relative oestrogenic potencies for most of the organochlorine pesticides are approximately 10- 4_ 10 --6 lower than that for the natural oestrogen, 17~-oestradiol.30 Furthermore, p,p'-DDE, the predominant environmental DDT pollutant, has not been consistently found to possess oestrogen activity, while its anti-androgenic activity is well recognized.r'' Epidemiological studies of women occupationally exposed to high levels of DDE and DDT have not shown an increased risk of breast cancer including the recent study from Mexico, an area where DDT has been widely used for agricultural purposes and malaria control, did not report an increased risk of breast cancer associated with serum levels of DDT and DDE. 23 One interesting, and relevant, observation from descriptive epidemiology is that while the risk of breast cancer is higher in White women than in Black women in the USA, the whole-body burden of DDE and DDT is much higher in Black women." Taken all together, it would appear that the evidence available at the present time is not compatible with an increased risk of breast cancer being associated with exposures to DDE and DDT. Polychorinated biphenyls (PCBs) Polychlorinated biphenyls (PCB) have been a major environmental health concern because of their. wide .
6
Adjuvant Therapy of Primary Breast Cancer
distribution and persistence in the environment. Experimental studies show that hormonal and non-hormonal activities of polychlorinated biphenyls (PCBs) are structure dependent, suggesting that the breast cancer risk associated with PCBs may vary according to specific PCB congeners. Four small studies have reported higher levels of PCBs in adipose tissue among breast cancer cases than among controls. 14 , 17,32,33 Moysich et al. 18 reported an increased risk of breast cancer associated with serum levels of PCBs among postmenopausal parous women who had never breast fed an infant (OR = 2.9, 95% CI 1.0-7.3). Zheng et a1. 34 reported a total of 475 cases and 502 controls who had their serum samples analyzed for PCBs in 1995-1997. The age- and lipid-adjusted geometric mean serum level of PCBs was comparable between cases (733.1 ppb) and controls (747.6ppb). After adjustment for confounding factors, the odds ratio of 0.95 (0.68-1.32) for PCBs were observed when the third tertile was compared with the lowest. Further stratification by parity, lactation, and menopausal and oestrogen receptor status also showed no significant association with serum levels or PCBs. Material was available on a total of 304 breast cancer cases and 186 controls aged 40-79 years who were recruited to measure congeners of PCBs. Fresh breast adipose tissue was analysed for PCBs. The age- and lipid-adjusted geometric mean tissue levels of total PCBs were not significantly different (P = 0.46) for the cases (478.6ppb) and controls (494.1 ppb). The covariateadjusted odds ratio was 0.7 (0.4-1.1) for all study participants when the third tertile was compared with the lowest tertile. No individual congeners or groups of congeners were associated with a significantly increased risk of breast cancer. Further stratification by type of breast disease, menopausal, parity, and lactation status, and body size also showed no significant association with body levels of PCBs. These results suggest that environmental exposure to PCBs may not substantially affect breast cancer risk. 35 Estimating joint effects of all congeners in a single analysis is complicated by correlation among exposure levels, and the resulting collinearity makes the results difficult to interpret. Patients with breast-related surgery at Yale-New Haven Hospital were interviewed using a standardized questionnaire, and breast adipose tissue samples were analysed for nine PCB congeners (74, 118, 138, 153, 156, 170, 180, 183, 187). The study recruited 490 women (304 cases and 186 controls) between 1994 and 1997.36 Logistic ridge regression was used to analyse the instability caused by collinearity. Although total PCB did not appear to be associated with breast cancer
risk, significant differences in effect were observed among the nine congeners. Logistic ridge regression demonstrated a protective effect on breast cancer risk for a potentially anti-oestrogenic and dioxin-like congener, 156, while two phenobarbital, CYPIA- and CYP2B-inducers had an adverse effect, 180 and 183. This analysis also suggested that a protective effect for another phenobarbital congener, 153, was largely explained by instability caused by collinearity.:" A hospital-based case-eontrol study was conducted in Ontario, Canada to evaluate the association between breast cancer risk and breast adipose tissue concentrations of several organochlorines.V Women scheduled for excision biopsy of the breast were enrolled and completed a questionnaire. The biopsy tissue of 217 cases and 213 benign controls frequency matched by study site and age in 5-year groups was analysed for 14 polychlorinated biphenyl (PCB) congeners, total PCBs, and 10 other organochlorines, including p,p' -1, I-dichloro-2,2-bis(p-chlorophenyl)ethylene. Multiple logistic regression was used to assess the magnitude of risk. While adjusting for age, menopausal status, and other factors, odds ratios (ORs) were above 1.0 for almost all organochlorines except five pesticide residues. The ORs were above two in the highest concentration categories of PCB congeners 105 and 118, and the ORs for these PCBs increased linearly across categories (P for trend ~0.01). Differences by menopausal status are noted especially for PCBs 105 and 118, with risks higher among premenopausal women, and for PCBs 170 and 180, with risks higher among postmenopausal women.:" The organochlorine compounds were examined in a prospective study environment recently. 38 A nested case--eontrol study was conducted to examine the association between serum concentrations DDE, DDT and PCBs and the development of breast cancer up to 20 years later. Cases (n = 346) and controls (n = 346) were selected from cohorts of women who donated blood in 1974, 1989, or both, and were matched on age, race, menopausal status, and month and year of blood donation. Median concentrations of DDE were lower among cases than controls in both time-periods (11.7% lower in 1974 (P = 0.06) and 8.6% lower in 1989 (P=0.41». Median concentrations of PCBs were similar among cases and controls (P=0.21 for 1974 and P=0.37 for 1989). The risk of developing breast cancer among women with the highest concentrations of ODE was roughly half that among women with the lowest concentrations, whether based on concentrations in 1974 (OR = 0.50, 95% CI 0.27-0.89, P (trend) =0.02) or in 1989 (OR=0.53, 95% CI 0.24-1.17, P (trend) = 0.08).
Environmental features in epidemiology The associations between circulating concentrations of PCBs and breast cancer were less pronounced but still in the same direction (1974: OR=0.68, 95% CI 0.3612.9, P (trend)=0.2; 1989: OR=0.73, 95% CI 0.371.46, P (trend) = 0.6). Adjustment for family history of breast cancer, body mass index, age at menarche or first birth, and months of lactation, did not materially alter these associations. These associations remained consistent regardless of lactation history and length of the follow-up interval, with the strongest inverse association observed among women diagnosed 16-20 years after blood-drawing. The authors concluded that the results from this prospective, community-based nested casecontrol study are reassuring. Even after 20 years of follow-up, exposure to relatively high concentrations of DDE or PCBs showed no evidence of contributing to an increased risk of breast cancer. 38
CONCLUSION Oestrogens, whether endogenous or exogenous, whether direct exposure to oestrogens or to chemicals with oestrogenic properties, hold the key to determining breast cancer risk in women. The subject of breast cancer risk factors such as hormonal factors (age at menarche, age at first birth, parity), exogenous hormonal sources (oral contraceptives and hormone replacement therapy) or dietary intakes (fat intake) has been well studied and some risk factors have been clearly identified, some of which are theoreticaly subject to alteration.:" Attention has recently increasingly focused on environmental exposures and environmental chemical pollutants as risk factors for breast cancer in women. The common nature of breast cancer in women, and the common nature of such exposures make such an association plausible, although far from proven as causal. Given also the resistant nature of concentrations in adipose tissue, it is important to attempt to identify any risk resulting from this exposure. Furthermore, when the prospective study" was analysed, median DDE and PCB concentrations among the controls were 59% and 147% higher in 1974 than 1989, respectively. Based on current evidence, the risk of breast cancer associated with exposure to organochlorine pesticides and their residues appears to be small, if it exists at all. However, one caveat is that the effect of individual or groups of PCBs as risk factors for breast cancer merits some further investigation. Indeed, a clear and powerful message which comes from recent work in Connecticut is that great care and attention must be taken to any statistical analysis of such chemical exposures, given the
7
potential for collinearity which exists between the various congeners. 12,15,16,19,26,34-36 These results indicate that studies of PCB congeners and health require an indepth statistical analysis in order to better understand the complex issues related to their collinearity.
Acknowledgements
It is a pleasure to acknowledge that this was conducted within the framework of support from the Associazione Italiana per la Ricerca sui Cancro (Italian Association for Research on Cancer).
References 1. Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 1981; 66: 1191-1308 2. Haenszel W, Kurihara M. Studies of Japanese migrants I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst. 1968; 40:43-68 3. Li F P. Phenotypes, genotypes, and interventions for hereditary cancers. Cancer Epidemiol Biomarkers Prev 1995; 4: 579-582 4. Easton DF. The inherited component of cancer. Br Med Bull 1994; 50: 527-535 5. Fearon ER. Human cancer syndromes: clues to the origin and nature of cancer. Science 1997; 278: 1043-1050 6. Lichtenstein P, Holm N V, Verkasalo P K, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A, Hemminki K. Environmental and heritable factors in the causation of cancer - analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 2000; 343(2): 78-85 7. Various correspondence. N Engl J Med 2000; 343 8. Stevens R G. Electric Power use and breast cancer: an hypothesis. Am J Epidemiol 1987; 125: 556-561 9. Vena J E, Marshall J R, Freudenheim J L et al. Risk of premenopausal breast cancer and use of electric blankets. Am J Epidemiol 1995; 142: 446-447 10. Gammon M 0, Schoenberg J B, Britton J A et al. Electric blanket use and breast cancer risk among younger women. Am J Epidemiol 1998; 148: 556-563 11. Coogan P F, Ashenau A. Exposure to power frequency magnetic fields and risk of breast cancer in Upper Cape Cod Cancer Incidence Study. Arch Environ Health 1998; 53: 359-367 12. Zheng T, Holford T R, Mayne STet al. Exposure to electromagnetic fields from use of electric blankets and other inhome electrical appliances and breast cancer risk. Am J Epidemiol 2000; 151(11): IIOl-1111 13. Dewailly E, Dodin S, Verreault Ret al. High organochlorine body burden in women with estrogen receptor-positive breast cancer. J Nat! Cancer Inst 1994; 86: 232-234. 14. Falck F, Ricci A, Wolff M S et al. Pesticides and polychlorinated biphenyl residues in human breast lipids and their relation to breast cancer. Arch Environ Health 1992; 47: 143-146. 15. Zheng T, Holford T. Tessari J et al. Oxychlordane and transnonachlor in breast adipose tissue and risk of female breast cancer. J Epidemiol Biostat 2000; 5: 153-160. 16. Zheng T, Holford T, Mayne STet al. Environmental exposure to hexachlorobenzene and risk of female breast cancer in Connecticut. Cancer Epidemiol Biomark Prevent 1999; 8: 407-411 17. Mussalo-Rauhamma H, Hasanen E, Pyysalo H, Antervo K, Kauppila R, Pantzar P. Occurrence of ~-hexachlorocyclohexanein breast cancer patients. Cancer 1990; 66: 2134-2128
8
Adjuvant Therapy of Primary Breast Cancer
18. Moysich K B, Ambrosone C B, Vena J E et al. Environmental organochlorine exposure and postmenopausal breast cancer risk. Cancer Epidemiol Biomarker Prev 1998; 7: 181-188 19. Zheng T, Holford T, Mayne STet al. Beta-benzene hexachloride in breast adipose tissue and risk of breast carcinoma. Cancer 1999; 85: 2212-2218 20. Houghton D L, Ritter L. Organochlorine residues and risk of breast cancer. J Am CoIl Toxico1 1995; 14-71-89. 21. Wolff M S, Toniolo P G, Lee E W et al. Blood levels of organochlorine residues and risk of breast cancer. J Natl Cancer Inst 1993; 85: 648-652. 22. Krieger N, Wolff M S, Hiatt R A et al. Breast cancer and serum organochlorines: a prospective study among White, Black, and Asian women. J Natl Cancer Inst 1994; 86: 589-599. 23. Lopez-Carrillo L, Blair A, Lopez-Cervantes M et al. Dichlorodiphenyltrichloro-ethane serum levels and breast cancer risk: a case-control study from Mexico. Cancer Res 1997; 57: 3728-3732. 24. Hunter D J, Hankinson S E, Laden F et al. Plasma organochlorine levels and the risk of breast cancer. New Engl J Med 1997; 337: 1253-1258. 25. Van 't Veer P, Lobbezoo I E, Martin-Moreno J M et al. DDT (dicophane) and postmenopausal breast cancer in Europe: casecontrol study. BMJ 1997; 315: 81-85 26. Zheng T, Holford T R, Mayne STet al. DDE and DDT in breast adipose tissue and risk of female breast cancer. Am J Epidemiol 1999; 150: 453-458. 27. Wo1ffM S, Toniolo P G. Environmental organochlorine exposure as a potential etiologic factor in breast cancer. Environ Health Perspect 1995; 103: 141-145 28. Safe S H, Zacharewski T. Organochlorine exposure and risk for breast cancer. Prog Clin BioI Res 1997; 396; 133-145 29. Shekhar P V M, Werdell J, Basrur V S. Environmental estrogen stimulation of growth and estrogen receptor function in preneoplastic and cancerous human breast cell lines. J Natl Cancer Inst 1997; 89: 1774-1782.
30. Safe S. Toxicology, structure-function relationship, and human and environmental health impacts of polychlorinated biphenyls: progress and problems. Environ Health Persp 1992; 100: 259-268 31. Kutz F W, Ybs A R, Strassman S C. Racial stratification of organochlorine insecticide residues in human adipose tissue. 10M 1977; 19: 619-622 32. Wasserman M, Nogueira D P, Tomatis L et al. Organochlorine compounds in neoplastic and adjacent apparently normal breast tissue. Bull Environ Contam Toxicol 1976; 15: 478-484 33. Unger M, Kiaer H, Blichert-Toft M, Olsen J, Clausen J. Organochlorine compounds in human breast fat from deceased with and without breast cancer and in a biopsy material from newly diagnosed patients undergoing breast surgery. Environ Res 1984; 34(1): 24-28 34. Zheng T, Holford T R, Tessari J et al. Risk of female breast cancer associated with serum of polychlorinated biphen 1,I-dichloro-2,2'bis (p-chlorophenyl) ethylene. Cancer Epidemiol Biomarkers Prev 2000; 9: 167-174 35. Zheng T, Holford T R, Tessari J et al. Breast Cancer risk associated with congeners of polychlorinated biphenyls. Am. J. Epidemiol 2000; 152: 50-58 36. Holford T R, Zheng T, Mayne S T, Zahm S H, Tessari J D, Boyle P. Joint effects of Nine polychlorinated biphenyl (PCB) congeners on breast cancer risk. Int J Epidemiol 2000; 29: 975-982 37. Aronson K J, Miller A B, Woolcott C G et al. Breast adipose tissue concentrations of polychlorinated biphenyls and other organochlorines and breast cancer risk. Cancer Epidemiol Biomarkers Prev 2000; 9(1): 55-63 38. Helzlsouer K J, Alberg A J, Huang H Y et al. Serum concentrations of organochlorine compounds and the subsequent development of breast cancer. Cancer Epidemiol Biomarkers Prev 1999; 8(6): 525-532 39. Mezzetti M, La Vecchia C, Decarli A, Boyle P, Talamini R, Franceschi S. Population Attributable risk for breast cancer: diet, nutrition and physical exercise. J Natl Cancer Inst 1998; 90: 389-394
1. BREAST CANCER EPIDEMIOLOGY AND THE MEDIA
Environmental features influencing the epidemiology of breast cancer P. Boyle l and T. Zheng 2 JDivision of Epidemiology and Biostatistics, European Institute of Oncology, Milan, Italy,· 2 Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven CT, USA
SUM MAR Y. A recent study claimed that a larger proportion of breast cancer was due to environmental exposures than to hereditary factors. This controversial publication clearly indicates the essential importance of investigating the gene-environment interaction and having a clear idea of what are the essential environmental factors to investigate. Environmental exposures are a subject of current interest in breast cancer aetiology and two common issues are exposure to electromagnetic fields (EMFs) and certain pesticides, fungicides and their residues. Overall, there does not appear to be an identifiable increased risk associated with domestic exposure to EMF, although there is a need for good studies with direct estimates of exposure. Oestrogens from a variety of sources, including chemical exposures, are determinant in breast cancer risk in women. Attention has recently increasingly focused on environmental exposures and environmental chemical pollutants as risk factors for breast cancer. The common nature of breast cancer in women, and such exposures, make such an association plausible, although far from proven as causal. Given also the resistant nature of concentrations in adipose tissue, it is important to attempt to identify any risk resulting from this exposure. Based on current evidence, the risk of breast cancer associated with exposure to organochlorine pesticides and their residues appears to be small, if it exists at all. However, one caveat is that the effect of individual or groups of PCBs as risk factors for breast cancer merits some further investigation. Indeed, a clear and powerful message which comes from recent work is that great care and attention must be given to any statistical analysis of such chemical exposures given the potential for collinearity which exists between the various chemical compounds and their congeners. © 2001 Harcourt Publishers Ltd
INTRODUCTION
The debate about the portion of cancer that can be ascribed to nature or nurture has been ongoing for many decades. Doll and Peto l concluded that there was compelling evidence that upwards of 80%, or even 90% of human cancers had environmental causes and could be avoidable. Different populations around the world theoretically experience different levels of different forms of cancer and these levels change with time in an orderly, and frequently predictable, manner. Population groups whose lifestyle differentiates themselves from members of the same population frequently experience different levels of cancer (e.g. Mormons and Seventh Address correspondence to: Prof. Peter Boyle, Divisione di Epidemiology Biostatistics, European Institute of Oncology, Via Ripamonti 435,20141 Milano, Italy. Fax: + 39 02 5748 9922; E-mail: [email protected]
Day Adventists in the USA). Migrant groups from one country to another quickly adopt the cancer profile of their new home: thus the Japanese who left Japan with its low levels of breast cancer quickly attained the levels of breast cancer in the population of California within two generations/. Apart from certain types of familial cancer, such as those arising in adenomatous polyposis coli, the contribution of hereditary factors to the development of cancer is generally held to be small. 3- 5 This, of course, refers to dominant genes which have been investigated in family studies which go back two or three generations while no determination has yet been made of the contribution of low-penetrance single genes, recessive genes and mechanisms that involve several genes, which are much more difficult to investigate. Studies of twins open one possibility of estimating the contribution of inherited genes to the development of
2
Adjuvant Therapy of Primary Breast Cancer
cancer. A combined study of the Twin Cohorts from Sweden, Denmark and Finland 6 assembled data on nearly 45 000 pairs of twins. Controversy has followed this publication which attempted to gauge the relative impact of genetic constitution and environmental exposure in the causation of cancer." The authors used 45000 pairs of twins in Scandinavia to investigate the contribution of inherited and environmental factors by comparing the risk of cancer arising in identical (monozygotic) twins (who have the same set of genes) with that of fraternal (dizygotic) twins (who share about 50% of their genes). Among the major forms of cancer, estimates of the genetic component ranged from 270/0 for breast cancer to 42% for prostate cancer (Table 1). Overall, the twin of a woman affected by breast cancer had an increased risk of having the same cancer. The twin of a monozygotic twin who had breast cancer had a risk of breast cancer which was 5.2 times that of a monozygotic twin of a person without breast cancer. The concordance for breast cancer in female monozygotic twins was 0.14, which means that there is a 14% probability that the identical twin of a woman with breast cancer will have the same cancer.' The authors then attempted to calculate the proportion of phenotypic variance accounted for by inherited genetic differences among persons (heritability), by shared environmental risk factors (the proportion of phenotypic variance accounted for by environmental factors shared by both twins, thus contributing to similarity between them) and non-shared environmental factors (the proportion of phenotypic variance accounted for by environmental factors causing differences between twins) (Table 1). These estimates of heritability in the twin study were greater than those previously derived from family-based surveys of cancer. However, the publication was widely interpreted as demonstrating that inherited genes play only a limited role in the causation of cancer, in contrast to lifestyle and other environmental exposures. The timing of this publication, coming soon after the announcement that scientists were rapidly nearing complete deciphering of the human genome, raised concerns in the media that the attention devoted to the Human Genome project and to the study of cancerassociated genes might be misplaced. Twin studies have been useful for a long time for distinguishing the relative importance of inherited and environmental factors in disease. Lichtenstein et a1. 6 can be criticized since the statistical methods employed in this Twin Study assumed that genes and environment were acting in isolation and that 100% of cancer risk needs to be explained. There is increasing evidence that genetic makeup influences susceptibility or even resis-
Table 1 Effects of heritable and environmental factors in cancer in twins in Denmark, Sweden and Finland Environmental Cancer Stomach Colo rectum Lung Breast Ovary Prostate
Heritable
Shared
28% 35% 26% 27% 22% 42%
10% 5% 12% 6% 0% 0%
NonShared 62% 60% 62% 67% 78% 58%
Data abstracted from Lichtenstien et a1.6
tance to cancer-causing exposures and this study emphasizes the necessity for epidemiological studies to focus on the gene-environment interaction. BREAST CANCER: ENVIRONMENTAL AND GENETIC FACTORS Conventional risk factors for breast cancer include hormonally-related factors (including nulliparity, age at menarche, age at first and last births, age at menopause and oophorectomy), use of exogenous oestrogens (oral contraceptives and hormone replacement therapy) and dietary factors (including alcohol consumption and intake of vegetables and fruits). More recently, attention to a series of environmental exposures has increased. Notable among these have been electro magnetic fields (EMF) and certain pesticides, particularly organochlorine pesticides. These factors have been carefully studied in a case-control study conducted in New Haven, Connecticut which will be used below to illustrate the recent findings in this area of research. Cases were histologically confirmed, incident cases of breast cancer who either had breast-related surgery at the Yale-New Haven Hospital, or who were residents of Tolland Country, between 1 January 1994 and 31 December 1997. A total of 561 incident breast cancers were identified in Yale-New Haven Hospital with 432 (77%) completing in-person interviews. From the hospital database, 569 potential control patients who had breast-related surgery and who were diagnosed histologically with normal tissue or benign breast diseases (excluding atypical hyperplasia), were selected. Of these, 404 patients (71 %) participated in the study. In addition, 238 cases was identified among residents of Tolland County, of which 176 (74%) completed interviews. A total of 205 controls were also recruited. A total of 490 women (304 cases and 186 controls) aged 40-79 years had 0.4 g of breast adipose tissue
Environmental features in epidemiology available for chemical analysis. This was frozen and sent to Colorado State University for chemical analysis.
3
with more direct measures of EMF exposure, would be welcomed. However, it can be concluded that within our current knowledge, domestic exposures to EMF do not present a major source of breast cancer risk.
ELECTROMAGNETIC FIELDS Chronic exposure to EMFs has been hypothesized to increase female breast cancer risk." Although no studies with direct measures have been reported, household electrical appliances frequently provide the highest magnetic field exposure levels in any residence, with electric blankets among the strongest sources. Limited epidemiological studies investigating electric blanket use and breast cancer risk have produced inconclusive results. A case-eontrol study from Upstate New York found a borderline significantly increased risk of breast cancer for postmenopausal women (OR = 1.5, 95% CI 1.0-2.2) and premenopausal women (OR = 1.4, 95% CI 0.9-2.2) among those who reported using electric blankets continuously throughout the night compared to never-users. When the pre- and postmenopausal groups were combined, the odds ratio was 1.5 (1.1-1.9).9 Subsequently, Gammon et al.!" found an odds ratio of 1.0 (0.9-1.2) for breast cancer in women age 45 and an odds ratio of 1.1 (0.9-1.4) for those 45 years and over, among women who reported using electric blankets, mattress pads or heated water beds when compared to never users. Coogan and Aschengrau II reported an odds ratio of 1.0 (0.7-1.4) among women who reported using an electric blanket compared with women who did not. In Connecticut, 40% of breast cancer cases, and 430/0 of controls, reported regular use of electric blankets which produced an odds ratio of 0.9 (0.7-1.1).12 For those who used an electric blanket continuously throughout the night, the adjusted odds ratio was 0.9 (0.7-1.2) when compared to never users. Risk did not vary according to age at first use, duration of use, or menopausal and oestrogen receptor status. Besides electric blankets, there are a number of other household appliances that could be significant sources for domestic exposure to EMF. There was no association found in the Connecticut study between breast cancer risk and use of domestic appliances including a hair dryer, curling iron, electric razor, electric toothbrush, vacuum cleaner, popcorn maker, computer or electric iron. 12 Increasing concerns that exposure to EMFs from domestic appliances may increase breast cancer risk appear to be unfounded from results obtained from recent epidemiological studies. This does not rule out the presence of any risk and further studies, somehow
PESTICIDE EXPOSURE, PESTICIDE RESIDUES AND BREAST CANCER RISK Organochlorine compounds including organochlorine pesticides, have been suggested to be associated with the risk of breast cancer in women. However these studies have generally been small, with low statistical power, and have unsurprisingly produced inconsistent findings. It is, however, biologically plausible that exposure to these environmental contaminents may increase breast cancer risk, since some organochlorine compounds are animal carcinogens, oestrogenically active and induce P450 mixed-function oxidase enzymes, which are intimately involved in steroid hormone metabolism. While epidemiological studies have chiefly focused on polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT), there are a number of related compounds which could potentially be involved. These include trans-nonachlor (TNC) and Oxychlordane (OCD), a major metabolite of the insecticide chlordane and Hexachlorobenzene (RCB), organochlorine fungicide, and one of its isomers B-benzene hexachloride (BHC). Oxychlordane (OCD) and trans-nonachlor (TNC) Oxychlordane (OCD) is a major oxidative metabolite of the insecticide chlordane. TNC is a component of technical chlordane and technical heptachlor. Chlordane and heptachlor were primarily used for the protection of buildings, lawns and gardens from soil insects and termites. Chlordane was first produced commercially in the USA in 1947 as an insecticide. Heptachlor was isolated from technical chlordane and, aside from the above uses, has also been used to control mosquitos. Although the use of chlordane has been suspended since 1988, and heptachlor restricted in its use, significant exposure to OCD, TNC and their metabolites has occurred, since they are highly persistent in the environment. Residues of these compounds have been detected throughout the food chain and in human tissues. Their oestrogenic activity and the persistent and widespread nature of their exposure, indicate these as potential breast cancer risk factors. Two small studies have examined the association between OeD and TNC~
4
Adjuvant Therapy of Primary Breast Cancer
Table 2 Breast cancer risk associated with breast adipose-tissue levels of oxychlordane and trans-nonachlor Odds ratio
95%
CI
Oxychlordane <26.0 26.0-33.6 33.7-47.5 2:47.6 p for trend
1.0 0.7 0.7 0.7 0.29
0.4-1.2 0.4-1.2 0.4--1.2
Trans-nonachlor <36.4 36.4--53.1 53.2-71.0 2:71.1 P for trend
1.0 1.2 0.7 1.1 0.44
0.7-2.1 0.4-1.3 0.6-1.9
Measurements in parts per billion (ppb). Odds ratio adjusted for age, lifetime months of lactation, age and menarche, age at first fullterm pregnancy and race. Data abstracted from Zheng et a1.15
and breast cancer risk, Dewailly et al.':' found a nonsignificant higher mean adipose-tissue level of TNC in nine ER + ve breast cancer cases (50.3 ppb) then in 17 (benign breast disease) (BBD) controls (31.1 ppb) (P = 0.07). Falck et al. 14 found non-significantly higher breast adipose-tissue levels of OCD and Heptachlorepoxide in 20 breast cancer cases (136 ppb) than in 20 BBD controls (121 ppb). However, TNC was found to be nonsignificantly lower among the cases (103 ppb) than in the control (118 ppb). Zheng et al. 15 reported an analysis of breast adipose tissue of 304 cases of breast cancer and 186 BBD controls. The age and lipid-adjusted geometric mean adipose-tissue levels of OCD were similar between cases (36.4ppb) and controls (38.0ppb). The age and lipidadjusted geometric mean adipose tissue levels of TNC between cases (55.5ppb) and control (58.1 ppb) were also similar. There was no association between breast cancer risk and mean adipose tissue levels of OCD and TNC (Table 2). The covariate adjusted odds ratio (OR) was 0.7 (0.4-1.3) for OCD and 1.1 (0.6-1.9) for TNC, when the highest quartile was compared to the lowest (Table 2). The association did not differ by menopausal status, by histological type of breast cancer, disease stage at diagnosis or by history of lactation nor were tissue levels of OCD and TNC different by ER and PR status.i ' These findings do not support an overall association between breast adipose-tissue levels of OCD and TNC and risk of breast cancer in women, Such findings are consistent with those of other recent studies that environmental exposures to various environmental oestrogens does not seen to have an overall significant impact on the risk of breast cancer.
Hexachlorobenzene and B-benzene hexachloride Hexachlorobenzene (HCB) is an organochlorine fungicide which has been widely used for several decades mainly as a seed protectant on grain, wheat and field crops.!? Although it is no longer produced commercially, due to its resistance to chemical and biological breakdown HCD has contaminated air, water, soil and plants and has been detected in mammals and humans. Current exposure to humans is primarily through the food chain. Three small studies have found no significant difference in adipose tissue levels of HCB between breast cancer cases and controls 13,14, 17 although a recent case-control study'" reported a non-significantly elevated odds ratio of 1.8 (95% CI 0.6-5.4) for those with the highest serum levels of HCB, when compared to those with the lowest levels among parous women who never lactated. Zheng et al. 16 reported no significant difference in mean adipose-tissue levels of HCB between breast cancer patients (21.0 ppb) and BBD controls (19.1 ppb) in a large study involving 304 breast cancer cases and 186 controls. Risk did not vary according to menopausal status, oestrogen or progesterone receptor status, breast cancer histology, stage of diagnosis or type of benign breast disease. 16 Residues of B-benzene hexachloride (B-BHC), one of the isomers of hexachlorobenzene, has been reported to be higher in breast fat tissue from breast cancer patients than from controls.l ' After adjustment for age and parity, breast cancer tissue containing more than 0.1 mg/ kg fat of B-BHC was associated with an odds ratio of 10.5 (95% CI 2.0-55.3) for breast cancer compared with tissue containing 0.1 or less.l ' Zheng et al. 19 found no significant difference in breast adipose tissue levels of PBHC between 304 cases and 186 controls either overall or when broken down by menopausal status or oestrogen or progesterone receptor levels. There seems little epidemiological support at present for an association between either hexachlorobenzene (HCB) or ~-benzene hexachloride (P-BHC) and risk of breast cancer.
DDT and DDE Dichlorodiphenyltrichloroethane (DDT) and its most stable metabolite dichrlorodiphenyldichloroethane (DDE) have both been suggested to be associated with an increased risk of breast cancer. Again, this is biologically plausible given their oestrogenic activity and ability to induce P450 enzymes, which are intimately associated with steroid hormone metabolism.r"
Environmental features in epidemiology Falck et al.!" reported from study of 20 breast cancer cases and 20 BBD controls, that a 10 ppb increase in adipose-tissue levels of DDE corresponds to a 10/ 0 increase in breast cancer risk. Dewailly et al. 13 estimated a 8.9-fold increase in breast cancer risk among women whose DDE levels were above 1292 ppb in adipose tissue compared to those with a level of 238 ppb or lower, based on a study of 18 breast cancer cases and 17 BBD controls. Wolff et a1. 2 1 conducted a nested case-eontrol study with 58 cases and reported a 4-fold increase in risk of breast cancer for an elevation of serum DDE concentrations from 2.0ng/ml (10th percentile) to 19.1 ng/ml (90th percentile). In another nested case-eontrol study, involving 50 Black and 50 Asian women with breast cancer, a 2- to almost 4-fold increased risk of breast cancer was reported for Black women and White women when comparing women in the third tertile with those in the lowest tertile of DDE in serum. There was no increased risk with higher serum levels observed in Asian women in this trial. 22 More recent studies with larger sample sizes, ranging from 141-265 cases, do not support the association 18,23,24 with one study reporting a significant inverse between adipose tissue levels of DDE and breast cancer risk.25 Breast adipose tissue was collected from 304 breast cancer cases and 180 benign breast disease controls from Yale-New Haven Hospital. DDE and DDT in adipose tissue were measured using gas chromatography.i" Statistical significance for multiple means of tissue levels of DDE and DDT was calculated using analysis of variance and rank sum tests. A logistic regression model was used to control confounding and estimate the exposure/disease association. The age-adjusted geometric mean adipose tissue levels of DDE for cases (736.5ppb) was not significantly different (P = 0.40) than that for the controls (784.1 ppb). Mean tissue level of DDT was nearly identical for cases (51.8 ppb) and controls (55.6ppb). DDE and DDT levels did not vary significantly based on Oestrogen/progesterone receptor status of the cases, menopausal status, histology, stage, or type of benign breast disease. The adjusted OR is 0.9 (0.5-1.5) for DDE and 0.8 (0.5-1.5) for DDT when the highest quarter was compared with the lowest quarter (Table 3). There are two major contradictory hypothesis around at the present time. Firstly, women and men now basically live in an environment that is a virtual sea of oestrogens and one long-term effect of low-level exposure to these environmental oestrogens could be to increases the risk of breast cancer in women.r" Alternatively, and completely contradictory, it is argued that the exposure to such relatively low levels of DDE
5
Table 3 Breast cancer risk associated with breast adipose-tissue levels of dichlorodiphenyldichloroethane (DDE) and dichlorodiphenyltrichloroethane (DDT) Odds ratio Dichlorodiphenyldichloroethane (D DE) <412.6 1.0 412.6-779.2 1.3 779.3-1355.9 0.9 ~1356.0 0.9 p for trend 0.46 Dichlorodiphenyltrichloroethane (DDT) <28.7 1.0 28.7-49.4 0.8 49.5-78.0 0.6 ~78.1 0.8 P for trend 0.38
0.7-2.2 0.5-1.6 0.5-1.5
0.5-1.4 0.4-1.1 0.5-1.5
Measurements in parts per billion (ppb). Odds ratio adjusted for age, body mass index, lifetime months of location, age at menarche, age at first full-term pregnancy, menopausal status, income and race. Data abstracted from Zheng et a1. 26
and other weakly oestrogenic organochlorine compounds may not have a significant effect on a woman's risk of breast cancer. 28 Experimental studies have provided a substantial body of evidence that DDE and DDT are not mammary carcinogens.i" Laboratory studies also show that the relative oestrogenic potencies for most of the organochlorine pesticides are approximately 10- 4_ 10 --6 lower than that for the natural oestrogen, 17~-oestradiol.30 Furthermore, p,p'-DDE, the predominant environmental DDT pollutant, has not been consistently found to possess oestrogen activity, while its anti-androgenic activity is well recognized.r'' Epidemiological studies of women occupationally exposed to high levels of DDE and DDT have not shown an increased risk of breast cancer including the recent study from Mexico, an area where DDT has been widely used for agricultural purposes and malaria control, did not report an increased risk of breast cancer associated with serum levels of DDT and DDE. 23 One interesting, and relevant, observation from descriptive epidemiology is that while the risk of breast cancer is higher in White women than in Black women in the USA, the whole-body burden of DDE and DDT is much higher in Black women." Taken all together, it would appear that the evidence available at the present time is not compatible with an increased risk of breast cancer being associated with exposures to DDE and DDT. Polychorinated biphenyls (PCBs) Polychlorinated biphenyls (PCB) have been a major environmental health concern because of their. wide .
6
Adjuvant Therapy of Primary Breast Cancer
distribution and persistence in the environment. Experimental studies show that hormonal and non-hormonal activities of polychlorinated biphenyls (PCBs) are structure dependent, suggesting that the breast cancer risk associated with PCBs may vary according to specific PCB congeners. Four small studies have reported higher levels of PCBs in adipose tissue among breast cancer cases than among controls. 14 , 17,32,33 Moysich et al. 18 reported an increased risk of breast cancer associated with serum levels of PCBs among postmenopausal parous women who had never breast fed an infant (OR = 2.9, 95% CI 1.0-7.3). Zheng et a1. 34 reported a total of 475 cases and 502 controls who had their serum samples analyzed for PCBs in 1995-1997. The age- and lipid-adjusted geometric mean serum level of PCBs was comparable between cases (733.1 ppb) and controls (747.6ppb). After adjustment for confounding factors, the odds ratio of 0.95 (0.68-1.32) for PCBs were observed when the third tertile was compared with the lowest. Further stratification by parity, lactation, and menopausal and oestrogen receptor status also showed no significant association with serum levels or PCBs. Material was available on a total of 304 breast cancer cases and 186 controls aged 40-79 years who were recruited to measure congeners of PCBs. Fresh breast adipose tissue was analysed for PCBs. The age- and lipid-adjusted geometric mean tissue levels of total PCBs were not significantly different (P = 0.46) for the cases (478.6ppb) and controls (494.1 ppb). The covariateadjusted odds ratio was 0.7 (0.4-1.1) for all study participants when the third tertile was compared with the lowest tertile. No individual congeners or groups of congeners were associated with a significantly increased risk of breast cancer. Further stratification by type of breast disease, menopausal, parity, and lactation status, and body size also showed no significant association with body levels of PCBs. These results suggest that environmental exposure to PCBs may not substantially affect breast cancer risk. 35 Estimating joint effects of all congeners in a single analysis is complicated by correlation among exposure levels, and the resulting collinearity makes the results difficult to interpret. Patients with breast-related surgery at Yale-New Haven Hospital were interviewed using a standardized questionnaire, and breast adipose tissue samples were analysed for nine PCB congeners (74, 118, 138, 153, 156, 170, 180, 183, 187). The study recruited 490 women (304 cases and 186 controls) between 1994 and 1997.36 Logistic ridge regression was used to analyse the instability caused by collinearity. Although total PCB did not appear to be associated with breast cancer
risk, significant differences in effect were observed among the nine congeners. Logistic ridge regression demonstrated a protective effect on breast cancer risk for a potentially anti-oestrogenic and dioxin-like congener, 156, while two phenobarbital, CYPIA- and CYP2B-inducers had an adverse effect, 180 and 183. This analysis also suggested that a protective effect for another phenobarbital congener, 153, was largely explained by instability caused by collinearity.:" A hospital-based case-eontrol study was conducted in Ontario, Canada to evaluate the association between breast cancer risk and breast adipose tissue concentrations of several organochlorines.V Women scheduled for excision biopsy of the breast were enrolled and completed a questionnaire. The biopsy tissue of 217 cases and 213 benign controls frequency matched by study site and age in 5-year groups was analysed for 14 polychlorinated biphenyl (PCB) congeners, total PCBs, and 10 other organochlorines, including p,p' -1, I-dichloro-2,2-bis(p-chlorophenyl)ethylene. Multiple logistic regression was used to assess the magnitude of risk. While adjusting for age, menopausal status, and other factors, odds ratios (ORs) were above 1.0 for almost all organochlorines except five pesticide residues. The ORs were above two in the highest concentration categories of PCB congeners 105 and 118, and the ORs for these PCBs increased linearly across categories (P for trend ~0.01). Differences by menopausal status are noted especially for PCBs 105 and 118, with risks higher among premenopausal women, and for PCBs 170 and 180, with risks higher among postmenopausal women.:" The organochlorine compounds were examined in a prospective study environment recently. 38 A nested case--eontrol study was conducted to examine the association between serum concentrations DDE, DDT and PCBs and the development of breast cancer up to 20 years later. Cases (n = 346) and controls (n = 346) were selected from cohorts of women who donated blood in 1974, 1989, or both, and were matched on age, race, menopausal status, and month and year of blood donation. Median concentrations of DDE were lower among cases than controls in both time-periods (11.7% lower in 1974 (P = 0.06) and 8.6% lower in 1989 (P=0.41». Median concentrations of PCBs were similar among cases and controls (P=0.21 for 1974 and P=0.37 for 1989). The risk of developing breast cancer among women with the highest concentrations of ODE was roughly half that among women with the lowest concentrations, whether based on concentrations in 1974 (OR = 0.50, 95% CI 0.27-0.89, P (trend) =0.02) or in 1989 (OR=0.53, 95% CI 0.24-1.17, P (trend) = 0.08).
Environmental features in epidemiology The associations between circulating concentrations of PCBs and breast cancer were less pronounced but still in the same direction (1974: OR=0.68, 95% CI 0.3612.9, P (trend)=0.2; 1989: OR=0.73, 95% CI 0.371.46, P (trend) = 0.6). Adjustment for family history of breast cancer, body mass index, age at menarche or first birth, and months of lactation, did not materially alter these associations. These associations remained consistent regardless of lactation history and length of the follow-up interval, with the strongest inverse association observed among women diagnosed 16-20 years after blood-drawing. The authors concluded that the results from this prospective, community-based nested casecontrol study are reassuring. Even after 20 years of follow-up, exposure to relatively high concentrations of DDE or PCBs showed no evidence of contributing to an increased risk of breast cancer. 38
CONCLUSION Oestrogens, whether endogenous or exogenous, whether direct exposure to oestrogens or to chemicals with oestrogenic properties, hold the key to determining breast cancer risk in women. The subject of breast cancer risk factors such as hormonal factors (age at menarche, age at first birth, parity), exogenous hormonal sources (oral contraceptives and hormone replacement therapy) or dietary intakes (fat intake) has been well studied and some risk factors have been clearly identified, some of which are theoreticaly subject to alteration.:" Attention has recently increasingly focused on environmental exposures and environmental chemical pollutants as risk factors for breast cancer in women. The common nature of breast cancer in women, and the common nature of such exposures make such an association plausible, although far from proven as causal. Given also the resistant nature of concentrations in adipose tissue, it is important to attempt to identify any risk resulting from this exposure. Furthermore, when the prospective study" was analysed, median DDE and PCB concentrations among the controls were 59% and 147% higher in 1974 than 1989, respectively. Based on current evidence, the risk of breast cancer associated with exposure to organochlorine pesticides and their residues appears to be small, if it exists at all. However, one caveat is that the effect of individual or groups of PCBs as risk factors for breast cancer merits some further investigation. Indeed, a clear and powerful message which comes from recent work in Connecticut is that great care and attention must be taken to any statistical analysis of such chemical exposures, given the
7
potential for collinearity which exists between the various congeners. 12,15,16,19,26,34-36 These results indicate that studies of PCB congeners and health require an indepth statistical analysis in order to better understand the complex issues related to their collinearity.
Acknowledgements
It is a pleasure to acknowledge that this was conducted within the framework of support from the Associazione Italiana per la Ricerca sui Cancro (Italian Association for Research on Cancer).
References 1. Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 1981; 66: 1191-1308 2. Haenszel W, Kurihara M. Studies of Japanese migrants I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst. 1968; 40:43-68 3. Li F P. Phenotypes, genotypes, and interventions for hereditary cancers. Cancer Epidemiol Biomarkers Prev 1995; 4: 579-582 4. Easton DF. The inherited component of cancer. Br Med Bull 1994; 50: 527-535 5. Fearon ER. Human cancer syndromes: clues to the origin and nature of cancer. Science 1997; 278: 1043-1050 6. Lichtenstein P, Holm N V, Verkasalo P K, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A, Hemminki K. Environmental and heritable factors in the causation of cancer - analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 2000; 343(2): 78-85 7. Various correspondence. N Engl J Med 2000; 343 8. Stevens R G. Electric Power use and breast cancer: an hypothesis. Am J Epidemiol 1987; 125: 556-561 9. Vena J E, Marshall J R, Freudenheim J L et al. Risk of premenopausal breast cancer and use of electric blankets. Am J Epidemiol 1995; 142: 446-447 10. Gammon M 0, Schoenberg J B, Britton J A et al. Electric blanket use and breast cancer risk among younger women. Am J Epidemiol 1998; 148: 556-563 11. Coogan P F, Ashenau A. Exposure to power frequency magnetic fields and risk of breast cancer in Upper Cape Cod Cancer Incidence Study. Arch Environ Health 1998; 53: 359-367 12. Zheng T, Holford T R, Mayne STet al. Exposure to electromagnetic fields from use of electric blankets and other inhome electrical appliances and breast cancer risk. Am J Epidemiol 2000; 151(11): IIOl-1111 13. Dewailly E, Dodin S, Verreault Ret al. High organochlorine body burden in women with estrogen receptor-positive breast cancer. J Nat! Cancer Inst 1994; 86: 232-234. 14. Falck F, Ricci A, Wolff M S et al. Pesticides and polychlorinated biphenyl residues in human breast lipids and their relation to breast cancer. Arch Environ Health 1992; 47: 143-146. 15. Zheng T, Holford T. Tessari J et al. Oxychlordane and transnonachlor in breast adipose tissue and risk of female breast cancer. J Epidemiol Biostat 2000; 5: 153-160. 16. Zheng T, Holford T, Mayne STet al. Environmental exposure to hexachlorobenzene and risk of female breast cancer in Connecticut. Cancer Epidemiol Biomark Prevent 1999; 8: 407-411 17. Mussalo-Rauhamma H, Hasanen E, Pyysalo H, Antervo K, Kauppila R, Pantzar P. Occurrence of ~-hexachlorocyclohexanein breast cancer patients. Cancer 1990; 66: 2134-2128
8
Adjuvant Therapy of Primary Breast Cancer
18. Moysich K B, Ambrosone C B, Vena J E et al. Environmental organochlorine exposure and postmenopausal breast cancer risk. Cancer Epidemiol Biomarker Prev 1998; 7: 181-188 19. Zheng T, Holford T, Mayne STet al. Beta-benzene hexachloride in breast adipose tissue and risk of breast carcinoma. Cancer 1999; 85: 2212-2218 20. Houghton D L, Ritter L. Organochlorine residues and risk of breast cancer. J Am CoIl Toxico1 1995; 14-71-89. 21. Wolff M S, Toniolo P G, Lee E W et al. Blood levels of organochlorine residues and risk of breast cancer. J Natl Cancer Inst 1993; 85: 648-652. 22. Krieger N, Wolff M S, Hiatt R A et al. Breast cancer and serum organochlorines: a prospective study among White, Black, and Asian women. J Natl Cancer Inst 1994; 86: 589-599. 23. Lopez-Carrillo L, Blair A, Lopez-Cervantes M et al. Dichlorodiphenyltrichloro-ethane serum levels and breast cancer risk: a case-control study from Mexico. Cancer Res 1997; 57: 3728-3732. 24. Hunter D J, Hankinson S E, Laden F et al. Plasma organochlorine levels and the risk of breast cancer. New Engl J Med 1997; 337: 1253-1258. 25. Van 't Veer P, Lobbezoo I E, Martin-Moreno J M et al. DDT (dicophane) and postmenopausal breast cancer in Europe: casecontrol study. BMJ 1997; 315: 81-85 26. Zheng T, Holford T R, Mayne STet al. DDE and DDT in breast adipose tissue and risk of female breast cancer. Am J Epidemiol 1999; 150: 453-458. 27. Wo1ffM S, Toniolo P G. Environmental organochlorine exposure as a potential etiologic factor in breast cancer. Environ Health Perspect 1995; 103: 141-145 28. Safe S H, Zacharewski T. Organochlorine exposure and risk for breast cancer. Prog Clin BioI Res 1997; 396; 133-145 29. Shekhar P V M, Werdell J, Basrur V S. Environmental estrogen stimulation of growth and estrogen receptor function in preneoplastic and cancerous human breast cell lines. J Natl Cancer Inst 1997; 89: 1774-1782.
30. Safe S. Toxicology, structure-function relationship, and human and environmental health impacts of polychlorinated biphenyls: progress and problems. Environ Health Persp 1992; 100: 259-268 31. Kutz F W, Ybs A R, Strassman S C. Racial stratification of organochlorine insecticide residues in human adipose tissue. 10M 1977; 19: 619-622 32. Wasserman M, Nogueira D P, Tomatis L et al. Organochlorine compounds in neoplastic and adjacent apparently normal breast tissue. Bull Environ Contam Toxicol 1976; 15: 478-484 33. Unger M, Kiaer H, Blichert-Toft M, Olsen J, Clausen J. Organochlorine compounds in human breast fat from deceased with and without breast cancer and in a biopsy material from newly diagnosed patients undergoing breast surgery. Environ Res 1984; 34(1): 24-28 34. Zheng T, Holford T R, Tessari J et al. Risk of female breast cancer associated with serum of polychlorinated biphen 1,I-dichloro-2,2'bis (p-chlorophenyl) ethylene. Cancer Epidemiol Biomarkers Prev 2000; 9: 167-174 35. Zheng T, Holford T R, Tessari J et al. Breast Cancer risk associated with congeners of polychlorinated biphenyls. Am. J. Epidemiol 2000; 152: 50-58 36. Holford T R, Zheng T, Mayne S T, Zahm S H, Tessari J D, Boyle P. Joint effects of Nine polychlorinated biphenyl (PCB) congeners on breast cancer risk. Int J Epidemiol 2000; 29: 975-982 37. Aronson K J, Miller A B, Woolcott C G et al. Breast adipose tissue concentrations of polychlorinated biphenyls and other organochlorines and breast cancer risk. Cancer Epidemiol Biomarkers Prev 2000; 9(1): 55-63 38. Helzlsouer K J, Alberg A J, Huang H Y et al. Serum concentrations of organochlorine compounds and the subsequent development of breast cancer. Cancer Epidemiol Biomarkers Prev 1999; 8(6): 525-532 39. Mezzetti M, La Vecchia C, Decarli A, Boyle P, Talamini R, Franceschi S. Population Attributable risk for breast cancer: diet, nutrition and physical exercise. J Natl Cancer Inst 1998; 90: 389-394