2212 b-Benzene Hexachloride in Breast Adipose Tissue and Risk of Breast Carcinoma Tongzhang Zheng, Sc.D.1 Theodore R. Holford, Ph.D.1 Susan T. Mayne, Ph.D.1 Patricia H. Owens, M.Sc.1 Barbara Ward, M.D.1 Darryl Carter, M.D.1 Robert Dubrow, M.D.1 Shelia H. Zahm, Sc.D.2 Peter Boyle, Ph.D.3 John Tessari, Ph.D.4 1 Yale University School of Medicine and Cancer Center, New Haven, Connecticut. 2 National Cancer Institute, National Institutes of Health, Bethesda, Maryland. 3 European Institute of Oncology, Milan, Italy. 4 Colorado State University, Fort Collins, Colorado. BACKGROUND. Epidemiologic studies have recently related benzene hexachloride (BHC) to breast carcinoma risk. Experimental studies have also shown that b-BHC is weakly estrogenic, hence supporting the alleged association. By directly comparing b-BHC levels in breast adipose tissue from incident breast carcinoma cases and controls, this study examined the hypothesis that exposure to b-BHC increases the risk of breast carcinoma in females. METHODS. A total of 490 Connecticut women (304 cases and 186 controls) were enrolled in the study during the period 1994 –1997. Cases were patients ages 40 –79 years with histologically confirmed incident primary breast carcinoma. Controls were patients with histologically confirmed incident benign breast disease. Breast adipose tissue was collected and analyzed for BHC isomers. A linear logistic regression model was used to adjust for potential confounders in estimating the association of exposure with disease. RESULTS. No significant differences in breast adipose tissue levels of b-BHC were observed between the cases and their controls overall, nor by menopausal status or estrogen and progesterone receptor status of the breast carcinoma cases. A nonsignificant reduced risk was observed among all subjects and among pre- and postmenopausal women when the highest quartile was compared with the lowest. Parous women with higher b-BHC levels, regardless of lactation status, had a nonsignificantly reduced breast carcinoma risk, whereas a nonsignificantly increased risk was observed among nulliparous women with higher b-BHC levels, based on very few study subjects. CONCLUSIONS. The results of this study do not support the hypothesis that increasing adipose tissue levels of b-BHC are associated with an increased risk of breast carcinoma in females. Cancer 1999;85:2212– 8. © 1999 American Cancer Society. KEYWORDS: b-benzene hexachloride, breast carcinoma, case– control, environment, epidemiology. E The authors thank Dr. Leticia DeDios and Dr. Christine Howe for their assistance in the collection of study materials. The authors also thank Donna Carrano, Melita Bosnyak, Heather Hutson, and Sylvia Ullman for their high quality interviewing, and the surgeons and nurses at the Yale–New Haven Hospital and the Temple Medical Center for their support of the study. Address for reprints: Dr. Tongzhang Zheng, 129 Church Street, Suite 700, New Haven, CT 06510. Received September 4, 1998; revision received December 23, 1998; accepted January 15, 1999. © 1999 American Cancer Society nvironmental estrogens have recently been associated with the risk of breast carcinoma. Most of the studies, however, have centered on polychlorinated biphenyls and 1,1-dichloro-2,29bis(p-chlorophenyl)ethylene as risk factors.1– 8 Benzene hexachloride (BHC), also known as hexachlorocyclohexane (HCH), an organochlorine insecticide, has also recently been related to breast carcinoma risk. Mussalo-Rauhamaa et al.9 reported that residues of b-BHC, one of the isomers of BHC, were higher in breast fat tissue from breast carcinoma patients than that from controls. After adjustment for age and parity, b-BHC remained significant as a risk factor for breast carcinoma. For residue levels of b-BHC in breast adipose tissue containing more than 0.1 mg/kg fat, the odds ratio was 10.5 (95% CI, 2.0 –55.3) compared with tissue containing 0.1 or less mg/kg fat. Experimental studies show that b-BHC is weakly estrogenic, sup- b-BHC and Breast Carcinoma Risk/Zheng et al. porting a potential association with breast carcinoma risk. In a breast carcinoma cell line, b-BHC was able to increase progesterone receptor (PR) content10 and stimulate proliferation in a dose-dependent manner in the estrogen receptor positive (ER1) cell line MLF-7, but not in an ER2 cell line.11 b-BHC was also found to augment rat uterine weights in a dose-dependent fashion,12 and a daily b-BHC dose of 0.5 mg/kg body weight given orally has also produced disturbances of the estrous cycle in female rats.13 It should be noted that DDT and PCBs also possess estrogenicity and have the ability to induce cytochrome p450 mixed-function oxidase enzymes, which are intimately involved in steroid hormone metabolism. Although some epidemiologic studies have found an association between DDT or PCBs and breast carcinoma,1–5 recent epidemiologic studies with large sample sizes have not shown an increased risk of breast carcinoma.6 – 8 In fact, an inverse association was reported in one study.8 Results relating more potent exogenous estrogens, such as postmenopausal hormone replacement therapy (HRT) and oral contraceptive use, to the risk of breast carcinoma have also been inconsistent.14 –16 Since 1978, technical grade BHC, which consists of a mixture of five configurational isomers of a-, b-, g-, d-, and e-BHC, has not been in use in the U.S.17 Therefore, in recent years, isomers of BHC other than b-BHC have rarely been found in human serum or fat tissue in the U.S. population because of their rapid biotransformation and subsequent conjugation and excretion. b-BHC, however, due to its resistance to chemical and biologic breakdown, has contaminated the environment and has been stored in fat tissue as a result of its extensive use during the 1950s and 1960s.18 Because of the ubiquitous nature of exposure to b-BHC through the food chain and the inconclusive data regarding the health effects of b-BHC, we decided to examine its correlation with the risk of breast carcinoma in females. This effort was part of a case– control study of organochlorine compounds and breast carcinoma risk in Connecticut. This study compared b-BHC levels in breast adipose tissue from incident breast carcinoma cases and noncancer controls. Breast adipose tissue was chosen because organochlorine compounds are lipophilic and highly persistent; thus, they tend to accumulate in fat tissue. Adipose tissue levels of organochlorine compounds, including BHC, are generally considered the best indicators of human lifetime environmental exposure. We measured three isomers of BHC (a-, b-, and g-BHC), but only 2 of the 490 subjects had quantifiable a-BHC (quantitation limit, 5.0 pg/mg) and none of the subjects had quantifiable g-BHC (quanti- 2213 tation limit, 10.0 pg/mg). In this article we report the results relating b-BHC and breast carcinoma risk. MATERIALS AND METHODS Study subjects were women ages 40 –79 years who had breast-related biopsy or surgery at Yale–New Haven Hospital (YNHH) in New Haven, Connecticut, and from whose breast pathology specimens we could collect at least 0.4 g of residual breast adipose tissue for chemical analyses. Breast adipose tissue not needed for diagnostic purposes was collected and placed into a glass vial on ice by personnel from the Tissue Retrieval Facility at the Yale Cancer Center. The samples were then coded and frozen within 30 minutes of being excised and stored at 284°C. Study subjects were enrolled consecutively between January 1, 1994, and December 30, 1997. Potential cases were patients with histologically confirmed, incident primary breast carcinoma (International Classification of Diseases [ICD-O] 174.0–174.9). The study pathologist (D.C.), who is responsible for almost all of the breast tissue diagnoses made at YNHH, classified the potential participants as either potential cases or controls and staged carcinomas according to the TNM system.19 Potential controls were patients with histologically confirmed incident benign breast disease (excluding atypical hyperplasia). Cases of benign breast disease were classified by the study pathologist and grouped according to the 1985 Pathologists’ Consensus Statement as proliferative benign breast disease without atypia (n 5 91) or nonproliferative disease (n 5 95; 21 patients with normal breast tissue, 25 with fibroadenoma, and 49 with other nonproliferative disease). Each participant, after approval by his or her physician, was approached by letter and then by phone. Those who consented were interviewed in person, generally in the woman’s home or in another convenient location. A standardized, structured questionnaire was used to obtain information about major known or suspected risk factors for breast carcinoma, including reproductive history, lactation history, past medical history, occupation, and demographic factors. The dietary information was collected through a scannable semiquantitative food frequency questionnaire, developed by the Fred Hutchinson Cancer Research Center, that was designed to optimize estimation of fat intake. Each subject was asked to characterize her usual diet in the year before she had the biopsy. Potential cases and controls were excluded if they had had a previous diagnosis of cancer, with the exception of nonmelanoma skin cancer. Interviews were completed for 79% of the potentially eligible cases and 74% of the controls. Frozen breast adipose tissue samples were sent in 2214 CANCER May 15, 1999 / Volume 85 / Number 10 batches to the study laboratory at Colorado State University, where they remained frozen until analysis. Tissue samples were analyzed in batches of 12, with each batch containing approximately 6 cases, 4 controls, and 2 quality-control samples. Samples were batched and coded at Yale; therefore, laboratory personnel in Colorado were blinded to the case– control status of the samples being analyzed. The laboratory method for analyzing BHC isomers in breast adipose tissue has been described elsewhere.20 Briefly, the method involved extraction in hexane, separation of the organochlorine pesticides from the PCBs and purification of the sample using Florisil (U.S. Silica Co., Berkeley Springs, WV) chromatography, and identification and quantification of the compounds using gas chromatography. The quantitation limit of this method was 15 ppb for b-BHC. All analyses were conducted under an established quality-control/quality-assessment program that included method spikes, reagent blanks, and quality-control windows. Quality-control spike mean recovery for b-BHC was 95% during the sample analyses, with a coefficient of variation of 15%. Adipose tissue levels of b-BHC were reported as parts per billion (ppb), equivalent to pg/mg lipid. The amount of lipid in the sample was quantified gravimetrically. Lipid adjustment was necessary because BHC isomers are lipid soluble and tissue sample lipid contents vary between subjects. Therefore, lipid adjustment facilitates comparability of the results across studies.21 Breast adipose tissue levels of b-BHC were compared between cases and controls, among pre- and postmenopausal women, and according to parity and lactation history. Quartiles of adipose tissue levels of b-BHC were formed based on the frequency distribution in the controls. Because earlier studies suggest that environmental estrogens may only affect the incidence of hormone-responsive breast carcinoma, b-BHC levels were also compared based on the cases’ ER and PR status. Both ER and PR levels were considered positive if the H-score was greater than 75, as described by McCarty et al.23 b-BHC levels were also compared based on breast carcinoma histology (lobular vs. ductal carcinoma), stage of diagnosis (Stage 0, I, II vs. III and IV), and type of benign breast disease (proliferative vs. nonproliferative disease). The statistical significance of the difference among multiple means of adipose tissue levels of b-BHC was calculated using analysis of variance and rank sum tests, and analysis of covariance was used to adjust for potential confounders. Because the distribution of b-BHC was skewed, we present the median as a measure of location and the first and third quartile cutpoints (25% and 75%) as summaries of the degree of variability. The log transformation was used to better approximate the normality assumption, and thus the antilog of the resulting adjusted mean values (i.e., the adjusted geometric mean) was used as a summary statistic. A linear logistic regression model was used to adjust for confounders when the association between exposure and disease was estimated. The variables included in the final model were age, body mass index (kg/m2), lifetime months of lactation, age at menarche, age at first full-term pregnancy (nulliparous, ,25 or $25 years), race (white, black, and other), family breast carcinoma history, fat intake, and lipidadjusted breast adipose tissue levels of total DDE and PCBs (ppb). Odds ratios (OR) and 95% confidence intervals were calculated using SAS statistical software.24 RESULTS As shown in Table 1, the cases were older than the controls. Because age was associated with body burden of b-BHC (r 5 0.30, P , 0.01), it was a potential confounder and hence was controlled in all subsequent analyses. Also shown in Table 1 is that women who experienced menarche at an early age had a nonsignificantly higher risk for breast carcinoma. Compared with those who had a first full-term pregnancy before age 20 years, women who were older at the time of their first full-term pregnancy showed a higher risk. Those with a lifetime lactation of more than 12 months experienced a reduced risk (OR 5 0.7, 95% CI 0.4 –1.1) compared with those who had never lactated. Dietary fat intake at the second tertile, but not the third tertile, was associated with a 60% increased risk. No other baseline factors showed a clear association with breast carcinoma risk in this study. The age-adjusted geometric means for adipose tissue levels of b-BHC were similar for the 304 breast carcinoma cases (27.1 ppb) and their 186 controls (26.3 ppb), as shown in Table 2. There were also no significant differences in age-adjusted geometric mean adipose tissue levels of b-BHC between pre- or postmenopausal cases and controls. The age-adjusted geometric mean adipose tissue levels of b-BHC for breast carcinoma cases and controls, based on the cases’ ER and PR status, are presented in Table 3. There were no significant differences in mean adipose tissue levels of b-BHC between controls and various ER and PR status subgroups. The cases’ mean adipose tissue levels of b-BHC also did not differ by ER and PR status. The age-adjusted geometric mean adipose tissue level of b-BHC for 304 cases (27.1 ppb) was almost identical to that of 91 controls with proliferative be- b-BHC and Breast Carcinoma Risk/Zheng et al. TABLE 1 Selected Characteristics of Breast Carcinoma Cases and Controls with Benign Breast Disease Characteristics Age (yrs) ,50 $50 Age at menarche (yrs) $15 13–14 ,13 Unknown Age at first full-term pregnancy (yrs) ,20 20–25 $26 Nulliparous Lifetime lactation (mos) 0 1–6 6–12 $13 Family breast carcinoma history Yes No BMI (kg/m2) ,21 21–24 $25 Fat intake (g/day) ,46 46–71 $72 Unknown Race White Black Other Cases Controls OR 81 105 1.0 1.7 25 130 147 2 21 72 92 1 1.0 1.5 1.3 — 31 129 107 37 34 68 56 28 1.0 2.1 2.2 1.5 1.2–3.8 1.2–3.9 0.7–3.0 197 46 22 39 107 33 14 32 1.0 0.8 0.9 0.7 0.5–1.3 0.4–1.7 0.4–1.1 229 75 142 44 1.0 1.1 0.7–1.6 37 116 151 27 62 97 1.0 1.4 1.1 0.8–2.4 0.7–2.0 78 128 88 10 62 62 61 1 1.0 1.6 1.1 — 266 32 6 157 21 8 1.0 0.9 0.4 TABLE 2 Lipid-Adjusted Adipose Tissue Levels of b-BHC (ppb) among Breast Carcinoma Cases and Benign Breast Disease Controls 95% CI 95 209 1.2–2.5 0.8–2.9 0.7–2.5 All subjects Cases Controls Premenopausal Cases Controls Postmenopausal Cases Controls 0.5–1.6 0.2–1.3 CI: confidence interval; BMI: body mass index. nign breast disease (27.2 ppb) and similar to that of 95 controls with nonproliferative benign breast disease (25.0 ppb). The two control groups were also not significantly different from each other (P 5 0.39). The geometric mean adipose tissue levels of b-BHC were also similar for 186 controls (26.3 ppb), for 262 women with ductal carcinoma (27.1 ppb), and for 33 women with lobular carcinoma (25.5 ppb). The geometric mean tissue level of b-BHC for 273 women with Stage 0 –II disease was 27.2 ppb, which was similar to that for 19 women with Stage III–IV disease (28.1 ppb) and not significantly different from the controls (26.3 ppb). Age and covariate-adjusted ORs for the association between b-BHC and breast carcinoma risk are presented in Table 4. There was an inverse association between b-BHC in breast adipose tissue and breast carcinoma risk when the highest quartile was com- No. of subjects Median (25%, 75%)a Age-adjusted geometric mean 304 186 27.1 (16.1, 41.6) 25.2 (16.3, 41.2) 27.1 26.3 0.61 87 75 17.3 (13.3, 28.3) 19.7 (14.4, 30.4) 18.8 19.7 0.62 217 111 30.1 (20.7, 46.0) 31.3 (19.3, 47.6) 31.3 31.9 0.82 P valueb b-BHC: b-benzene hydrochloride; ppb: parts per billion, equivalent to pg/mg lipid. a The cutpoints at the first quartile (25%) and the third quartile (75%) for the lipid adjusted breast adipose tissue levels of b-BHC. b P values for test of geometric means between cases and controls after adjustment for age using analysis of covariance. TABLE 3 Lipid-Adjusted Adipose Tissue of b-BHC (ppb) in Breast Carcinoma Cases and Controls, by Estrogen and Progesterone Receptor Status Hormone status 1.0–2.6 0.7–1.8 2215 Controls Cases ER1 ER2 Unknown Cases PR1 PR2 Unknown No. of subjects Median (25%, 75%)a Age-adjusted geometric mean 186 25.2 (16.3, 41.2) 26.3 157 126 21 27.5 (16.4, 44.9) 26.3 (15.5, 40.3) 23.6 (14.4, 37.5) 28.1 26.5 23.4 0.34 0.90 0.46 116 128 60 26.7 (15.7, 42.5) 27.9 (16.8, 43.9) 26.4 (15.1, 38.1) 27.6 27.7 24.8 0.52 0.45 0.55 P valueb b-BHC: b-benzene hexachloride; ER: estrogen receptor; PR: progesterone receptor; ppb: parts per billion, equivalent to pg/mg lipid. a The cutpoints at the first quartile (25%) and the third quartile (75%) for the lipid-adjusted breast adipose tissue levels of b-BHC. b P value for the geometric mean difference between controls and each type of case adjusted for age using analysis of covariance. pared with the lowest for the covariate-adjusted odds ratio (OR) (OR 5 0.6, 95% CI 0.3–1.1). An inverse association was also observed among pre- and postmenopausal women, but none of these achieved statistical significance at the nominal 5% level. Further analysis by parity and history of lactation showed an inverse association of breast carcinoma risk with increasing tissue levels of b-BHC among parous women (Table 5). For those who reported breastfeeding, the covariate adjusted OR was 0.4 (95% 2216 CANCER May 15, 1999 / Volume 85 / Number 10 TABLE 4 Breast Carcinoma Risk Associated with Breast Adipose Tissue Levels of b-BHC (ppb), by Menopausal Status b-BHC All subjects ,16.5 16.5–25.2 25.3–41.4 41.5– P value for trend Premenopausal ,14.7 14.7–20.0 20.1–30.3 30.4– P value for trend Postmenopausal ,20.0 20.0–31.2 31.3–47.6 47.7– P value for trend Cases Controls OR1 95% CI OR2 95% CI 81 59 87 77 48 46 46 46 1.0 0.7 0.8 0.5 0.30 0.4–1.1 0.4–1.3 0.3–1.0 1.0 0.7 0.8 0.6 0.26 0.4–1.2 0.5–1.5 0.3–1.1 33 16 21 17 19 19 18 19 1.0 0.5 0.6 0.4 0.98 0.2–1.1 0.3–1.5 0.2–1.1 1.0 0.5 0.6 0.4 0.90 0.2–1.2 0.2–1.6 0.1–1.1 54 56 62 45 28 27 29 27 1.0 0.9 0.7 0.5 0.27 0.5–1.8 0.4–1.5 0.2–1.1 1.0 1.0 0.8 0.6 0.27 0.5–1.9 0.4–1.6 0.3–1.3 b-BHC: b-benzene hexachloride; CI: confidence interval; ppb: parts per billion, equivalent to pg/mg lipid. OR1: odds ratio adjusted only for age; OR2: odds ratio adjusted for age, body mass index (kg/m2), lifetime mos of lactation, age at menarche, age at first full-term pregnancy (nulliparous, ,25, $25 yrs), race (white, black, and other), family breast carcinoma history, fat intakes, and lipid-adjusted breast adipose tissue levels of total DDE: 1,1-dichloro-2,29bis(p-chlorophenyl)ethylene and PCB: polychlorinated biphenyls1 (ppb). TABLE 5 Odds Ratios Associated with Adipose Tissue Levels of b-BHC among Parous and Nulliparous Women, by Lactation History b-BHC Parous women with history of lactation ,16.2 16.2–22.6 22.7–40.4 40.5– P value for trend Parous women with no history of lactationa ,16.9 16.9–27.7 27.8–42.9 43.0– P value for trend Nulliparous womenb ,18.6 18.6–31.8 31.9– P value for trend Cases Controls OR1 95% CI OR2 95% CI 35 19 34 19 20 20 19 20 1.0 0.5 0.7 0.3 0.10 0.2–1.2 0.3–1.6 0.1–0.8 1.0 0.6 0.9 0.4 0.16 0.2–1.4 0.4–2.3 0.1–1.1 37 34 44 45 19 19 20 19 1.0 0.7 0.7 0.6 0.96 0.3–1.7 0.3–1.6 0.2–1.4 1.0 0.7 0.7 0.6 0.91 0.3–1.7 0.3–1.6 0.2–1.7 9 12 16 10 10 10 1.0 1.3 1.5 0.63 0.4–4.4 0.4–5.5 1.0 1.4 3.1 0.88 0.4–5.5 0.6–15.1 b-BHC: b-benzene hexachloride; CI: confidence interval; ppb: parts per billion, equivalent to pg/mg lipid; OR1: odds ratio adjusted only for age; OR2: odds ratio adjusted for age, body mass index (kg/m2), lifetime mos of lactation, age at menarche, age at first full-term pregnancy (nulliparous, ,25, $25 yrs), race (white, black, and other), family breast carcinoma history, fat intakes, and lipid-adjusted breast adipose tissue levels of total DDE and PCBs (ppb). a No adjustment for lactation status. b No adjustment for lactation and age at first full-term pregnancy. b-BHC and Breast Carcinoma Risk/Zheng et al. CI, 0.1–1.1) when the highest quartile was compared with the lowest. Among nulliparous women, however, an adjusted OR of 3.1 (95% CI 0.6 –15.1) was observed when the third tertile was compared with the lowest based on only a very few subjects in this group. DISCUSSION No significant difference in breast adipose tissue levels of b-BHC was observed between breast carcinoma patients and their controls in this study. The mean adipose tissue levels of b-BHC also did not vary by case– control status when the subjects were stratified by menopausal status or by ER or PR status for the breast carcinoma cases. A nonsignificant inverse association with breast carcinoma risk was observed among all study subjects and among pre- and postmenopausal women when the highest quartile was compared with the lowest. Parous women, regardless of lactation history, experienced a nonsignificantly reduced risk of breast carcinoma. Therefore, our results do not support the hypothesis that increasing adipose tissue levels of b-BHC are associated with an increased risk of breast carcinoma in females. The lack of a positive association in this study contradicts the results of a case– control study of Finnish women reported by Mussalo-Rauhamaa et al.9 They found significantly higher (P 5 0.026) breast adipose tissue levels of b-BHC for breast carcinoma cases (130 ppb) than for controls (80 ppb). The risk was found to be almost 11-fold higher for women whose breast adipose tissue contained more than 100 ppb b-BHC compared with women whose tissue contained less. Chance offers one possible explanation for the difference between the results of our study and those of Mussalo-Rauhamaa et al.,9 especially because of the relatively small sample size of the latter (44 breast carcinoma cases and 33 controls). It could be hypothesized that the discrepancy between the study of Mussalo-Rauhamaa and this study is due to a threshold effect of b-BHC on breast carcinoma risk; that is, the risk may only exist at much higher exposure levels. The mean adipose tissue level of b-BHC in the cases of Mussalo-Rauhamaa et al. was 3.6 times that of cases in our study. Data published in 1985 from the U.S. National Human Adipose Tissue Survey show that the median residue level of b-BHC was 140 ppb for the entire nation for the years 1970 –1983. The mean adipose tissue levels of b-BHC decreased from approximately 370 ppb in 1970 to approximately 100 ppb in 1983.18 Because the tissue samples studied by Mussalo-Rauhamaa were actually collected between 1985 and 1986,9 the levels of b-BHC appear to be quite comparable to those found in the U.S. population and 2217 the Finnish population in the 1980s. The levels found in our study are consistent with a continuing decline in the level of exposure to this chemical. Inclusion of BBD controls may also be partially responsible for the weakness of or lack of association between breast carcinoma and some of the established risk factors shown in Table 1. The risk factors for BBD are currently poorly understood as reviewed by Ernster,25 and it is possible that some of the reproductive and demographic variables shown in Table 1 are risk factors for both benign breast disease and breast carcinoma. The weakness of or lack of association could also be a function of sample size, however; a similar lack of association for established risk factors has been seen in three recent case– control studies of similar size4,5,7 that investigated the association between DDE and breast carcinoma risk. Similarly, it could be argued that the overall lack of association between b-BHC and breast carcinoma risk in this study is due to the participation of patients with benign breast disease as controls. If there were an association between b-BHC and benign breast disease (excluding atypical hyperplasia), this could lead to an underestimation of the true relative risk for breast carcinoma. Although this possibility cannot be ruled out entirely, it is unlikely that the lack of association of b-BHC with breast carcinoma risk is only attributable to the participation of benign breast disease patients as controls. Two previously observed positive associations between organochlorine compounds and female breast carcinoma risk came from studies in which patients with benign breast disease served as controls.2,3 In addition, the age-adjusted geometric mean of adipose tissue levels of b-BHC for 91 women diagnosed with proliferative benign breast disease (27.2 ppb) was not significantly (P 5 0.39) higher than that for 95 women with nonproliferative disease or normal tissue (25.0 ppb). It is a concern that breast adipose tissue levels of b-BHC may be affected by case status. In particular, the tissue levels of b-BHC for patients with advanced disease may be affected by mobilization of energy from fat stores. In our study, only 20 breast carcinoma patients were diagnosed with Stage III/IV disease, and exclusion of these patients from the study did not result in any material change to the conclusion. A recent follow-up study7 also does not support that disease stage at diagnosis significantly impacts serum levels of DDE and PCBs. Another concern is the possibility that some patients in this study may have had chemotherapy or radiation therapy prior to surgery to reduce tumor bulk, and that such treatment could affect lipids and lipophilic substances in the breast. In the current 2218 CANCER May 15, 1999 / Volume 85 / Number 10 study, breast adipose tissue not needed for diagnostic purposes was collected by personnel from the Tissue Retrieval Facility at the Yale Cancer Center for chemical analyses. The majority of the breast adipose tissue samples were collected at the time of biopsy, before the final diagnosis was made. Although a small number of tissue samples may have been collected at the time of surgical treatment, according to the information provided by YNHH surgeons, only in rare instances would a patient at YNHH receive chemotherapy or radiation therapy prior to surgical treatment. The age difference between the cases and the controls in this study also cannot be used to explain the observed inverse association. Age was found to be positively associated with the body burden of b-BHC in our study (r 5 0.30, P , 0.01). Therefore, if the residual confounding from the age difference has any impact on the observed effect, it should cause a falsepositive association rather than a negative one. In conclusion, our overall results do not support a positive association between environmental exposure to b-BHC and risk of breast carcinoma in females. We did not find an increased risk among either pre- or postmenopausal women, among different types of breast carcinoma histologies, or among different types of benign breast disease. The risk also did not vary based on the ER or PR status of the cases. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. REFERENCES 1. 2. 3. 4. 5. 6. 7. Wolff MS, Toniolo PG, Lee EW, Rivera M, Dubin N. Blood levels of organochlorine residues and risk of breast carcinoma. J Natl Cancer Inst 1993;85:648 –52. Falck F, Ricci A, Wolff MS, Godbold J, Deckers P. Pesticides and polychlorinated biphenyl residues in human breast lipids and their relation to breast cancer. Arch Environ Health 1992;47:143– 6. Dewailly E, Dodin S, Verreault R, Ayotte P, Sauve L, Morin J, et al. High organochlorine body burden in women with estrogen receptor-positive breast cancer. J Natl Cancer Inst 1994;86:232– 4. Krieger N, Wolff MS, Hiatt RA, Rivera M, Vogelman J, Oretreich N. Breast cancer and serum organochlorines: a prospective study among white, black, and Asian women. J Natl Cancer Inst 1994;86:589 –99. Moysich KB, Ambrosone CB, Vena JE, Shields PG, Mendola P, Kostyniak P, et al. Environmental organochlorine exposure and postmenopausal breast cancer risk. Cancer Epidemiol Biomarkers Prev 1998;7:181– 8. Lopez-Carrillo L, Blair A, Lopez-Cervantes M, Cebrian M, Rueda C, Reyes R, et al. Dichlorodiphenyltrichloro-ethane serum levels and breast cancer risk: a case– control study from Mexico. Cancer Res 1997;57:3728 –32. Hunter DJ, Hankinson SE, Laden F, Colditz G, Manson JE, Willett WC, et al. Plasma organochlorine levels and the risk of breast cancer. N Engl J Med 1997;337:1253– 8. 18. 19. 20. 21. 22. 23. 24. 25. Van’t Veer P, Lobbezoo IE, Martin-Moreno JM, Gualler E, Gomez-Aracena J, Kardinaal AFM, et al. DDT (dicophane) and postmenopausal breast cancer in Europe: case– control study. BMJ 1997;315:81–5. Mussalo-Rauhamaa H, Hasanen E, Pyysalo H, Antervo K, Kauppila R, Pantzar P. Occurrence of beta-hexachlorocyclohexane in breast cancer patients. Cancer 1990;66: 2124 – 8. Coosen R, van Velsen FL. Effects of the beta-isomer of hexachlorocyclohexane on estrogen-sensitive human mammary tumor cells. Toxicol Appl Pharmacol 1989;101:310 – 8. Steinmetz R, Young PC, Caperell-Grant A, Gize EA, Madhukar BV, Ben-Jonathan N, et al. Novel estrogenic action of the pesticide residue beta-hexachlorocyclohexane in human breast cancer cells. Cancer Res 1996;56: 5403–9. Loeber JG, van Velsen FL. Uterotropic effects of the betaBHC, a food chain contaminant. Food Addit Contam 1984; 1:63– 6. Naishtein SY, Leibovich DL. Effect of small doses of DDT and lindane and their mixture on sexual function and embryogenesis in rats. Hyg Sanit 1971;36:190 –5. Hulka BS, Liu ET, Lininger RA. Steroid hormones and risk of breast cancer. Cancer 1994;74:1111–24. Kelsey J. Breast cancer epidemiology: summary and future directions. Epidemiol Rev 1993;15:256 – 63. Feigelson HS, Henderson BE. Estrogens and breast cancer. Carcinogenesis 1996;17:2279 – 84. Ahlborg UG, Lipworth L, Titus-Ernstoff L, Hsieh CC, Hanberg A, Baron J, et al. Organochlorine compounds in relation to breast cancer, endometrial cancer, and endometriosis: an assessment of the biological and epidemiological evidence. Crit Rev Toxicol 1995;25:463–531. Kutz FW, Wood PH, Bottimore DP. Organochlorine pesticides and polychlorinated biphenyls in human adipose tissue. Rev Environ Contam Toxicol 1991;120:1– 83. Beahrs OH, Myers MH, editors. Manual for staging cancer. 2nd edition. American Joint Committee on Cancer. Philadelphia, PA: J. B. Lippincott, 1983:127–34. Archibeque-Engle AL, Tessari JD, Winn DT, Nett TM, Zheng T. Comparison of organochlorine pesticide and polychlorinated biphenyl residues in human breast adipose tissue and serum. J Toxicol Environ Health 1997;52:285–93. Phillips DL, Pirkle JL, Burse VW, Bernert JT, Henderson O, Needham LL. Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding. Arch Environ Contam Toxicol 1989;18:495–500. Shekhar PVM, Werdell J, Basrur VS. Environmental estrogen stimulation of growth and estrogen receptor function in preneoplastic and cancerous human breast cell lines. J Natl Cancer Inst 1997;89:1774 – 82. McCarty KS, Miller LS, Cox EB, Konrath J, McCarty KS Sr. Estrogen receptor analyses. Arch Pathol Lab Med 1985;109: 716 –21. SAS Institute. SAS/STAT user’s guide. Version 6. Cary, NC: SAS Institute, 1990. Ernster VL. The epidemiology of benign breast disease. Epidemiol Rev 1981;3:184 –204.