Environmental Organochiorine Exposure and Postmenopausal ... · Preventive Medicine. State University of New York at Buffalo. 270 Farber Hall, Buffalo, NY 14214. Phone: (716) 829-2975;

Vol. 7, 181-1 88, March 1998 Cancer Epidemiology, Biomarkers & Prevention 181

Environmental Organochiorine Exposure and Postmenopausal

Breast Cancer Risk’

Kirsten B. Moysich,2 Christine B. Ambrosone,John E. Vena, Peter G. Shields, Pauline Mendola,Paul Kostyniak, Hebe Greizerstein, Saxon Graham,James R. Marshall, Enrique F. Schisterman, andJo L. Freudenheim

Department of Social and Preventive Medicine, School of Medicine andBiomedical Sciences [K. B. M., J. E. V., P. M., S. G., E. F. S., J. L. F.] and

Toxicology Research Center [P. K., H. 0.], State University of New York atBuffalo, Buffalo, New York 14214; Arizona Cancer Center, Tucson, Arizona

85724 [J. R. M.]; National Cancer Institute, Bethesda, Maryland 20892[P. G. S.]; and National Center for Toxicological Research, Jefferson, Arizona72079 [C. B. A.]

Abstract

Environmental exposure to organochlorine compoundshas been associated with a potential role in breast canceretiology, but results from previous investigations yieldedinconsistent results.

In this case-control study, we examined the effect of1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDE),hexachlorobenzene (HCB), mirex, and several measuresof polychlorinated biphenyls (PCBs) on postmenopausalbreast cancer risk.

The study sample included 154 primary, incident,histologically confirmed, postmenopausal breast cancer

cases and 192 postmenopausal community controls. Usualdiet, reproductive and medical histories, and otherlifestyle information was obtained by an extensive inperson interview. Serum levels (ng/g) of DDE, HCB,mirex, and 73 PCB congeners were determined by gaschromatography with electron capture. PCB exposurewas examined as total measured PCB levels, total number

of detected PCB peaks, and three PCB congener groups.In the total sample, there was no evidence of an

adverse effect of serum levels of DDE [odds ratio (OR),1.34; 95% confidence interval (CI) 0.71-2.55], HCB (OR,0.81; 95% CI, 0.43-1.53), or mirex (OR, 1.37; 95% CI,0.78-2.39). Further, higher serum levels of total PCBs

(OR, 1.14; 95% CI, 0.61-2.15), moderately chlorinatedPCBs (OR, 1.37; 95% CI, 0.73-2.59), more highlychlorinated PCBs (OR, 1.19; 95% CI, 0.60-2.36), orgreater number of detected peaks (OR, 1.34; 95% CI,0.72-2.47) were not associated with increased risk. There

Received 8/12/97; revised I l!24/97; accepted 12/1 1/97.

The costs of publication of this article were defrayed in part by the payment of

page charges. This article must therefore be hereby marked advertisement in

accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

I Supported in part by Grants DAMD1 7-94-J-4108, CAl 1535, and CA/ES 62995from the National Cancer Institute.

2 To whom requests for reprints should be addressed, at Department of Social andPreventive Medicine. State University of New York at Buffalo. 270 Farber Hall,

Buffalo, NY 14214. Phone: (716) 829-2975; Fax: (716) 829-2979; [email protected].

was some indication of a modest increase in risk forwomen with detectable levels of less chlorinated PCBs

(OR, 1.66; 95% CI, 1.07-2.88). Among parous womenwho had never lactated, there was some evidence forincreased risk, associated with having detectable levels ofmirex (OR, 2.42; 95% CI, 0.98-4.32), higher serumconcentrations of total PCBs (OR, 2.87; 95% CI, 1.01-7.29), moderately chlorinated PCBs (OR, 3.57; 95% CI,1.10-8.60), and greater numbers of detected PCBcongeners (OR, 3.31; 95% CI, 1.04-11.3).

These results suggest that an increase in risk ofpostmenopausal breast cancer associated withenvironmental exposure to PCBs and mirex, if at allpresent, is restricted to parous women who had neverbreast-fed an infant.

Future studies should consider lactation history ofparticipants, as well as use similar epidemiological andlaboratory methodologies, to ensure comparability ofresults across studies.

Introduction

Environmental factors have been implicated in breast cancer

etiology, due to the steady increase in incidence over the last

decades ( 1), regional and international differences in incidence,and observed changes in incidence rates in migrant populations

(2). One group of environmental exposures that has been ex-amined in relation to breast cancer are organochlorine corn-

pounds, such as DDE,3 the major metabolite of DDT, PCBs,

HCB, and mirex. From the early l940s to 1972, DDT was one

of the most widely used chemicals for controlling insect pests

on agricultural crops and for controlling insects that carry

diseases such as malaria and typhus. PCBs have been manu-

factured commercially since 1929 for a variety of applications,

including use as dielectrics in transformers and capacitors and

for cooling fluids in hydraulic systems. HCB is a widespreadchlorinated hydrocarbon originating from agricultural and in-

dustrial sources. It was originally used as a fungicide, but

currently, the major source of HCB is industrial emission as a

side product related to the manufacture of organochlorinated

products. Mirex was extensively used in the southeastern

United States for the control of the red fire ant and as a fire

retardant coating for various materials. Although commercialproduction of all of these compounds was banned in the early

19705, measurable levels of organochlorine residues are stillfound in human tissue and blood samples. This persistence hasbeen attributed to their slow metabolism and high lipid solu-

bility, leading to storage in adipose stores, including breast

3 The abbreviations used are: DDE. I.l-dichloro-2.2-bis(p-chlorophenyl)ethyl-

ene: DDT, 2.2-bis(p-chlorophenyl)- I , 1 . I -trichloroethane: PCB. polychlorinated

biphenyl; HCB, hexachlorobenzene: LOD. limit of detection: OR. odds ratio: CI.

confidence interval.

on June 6, 2021. © 1998 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

http://cebp.aacrjournals.org/

182 Organochlorines and Breast Cancer Risk

tissue. The primary route of excretion for these compounds for

women is lactation (3).Evidence from laboratory studies has demonstrated a com-

plex diversity of biological effects associated with these com-pounds. DDE, HCB, mirex, and some PCB congeners have

been associated with induction of cytochrome P450 enzymes(4-9), which may or may not be associated with estrogenic(10-14) and antiestrogenic effects (12) shown in some inves-tigations. Studies have also noted changes in immune responses(15-17) and tumor-promoting effects (4, 7, 18-20).

The body of evidence on the effect of organochlorines on

breast cancer risk is limited (21-30). Several studies comparedorganochlorine concentrations in neoplastic tissue with tissuelevels of women with benign breast disease or women who diedin accidents (21-25). Results from these mammary tissue stud-ies are inconsistent and were hampered by several methodolog-ical limitations, including small sample size and absent or

inadequate control for known breast cancer risk factors.Only recently, several epidemiological studies examined

this risk relationship with more precision (26-30). Wolff et a!.(26) examined DDE and total PCB concentrations in sera from

58 breast cancer cases and 1 7 1 controls. Although risk wasassociated with higher levels of serum DDE, there was no

evidence of a linear relationship with greater body burden ofPCBs. Increased risk, either statistically significant or of bor-

derline significance, was observed for all PCB exposure levelsabove the lowest. A nested case-control study was also con-

ducted by Krieger et a!. (27), consisting of 150 breast cancercases from three ethnic groups (Caucasian, African-American,

and Asian) and 150 matched controls, all of whom had bloodsamples stored in the late 1960s. No excess risks of breastcancer in association with serum levels of DDE and total PCBswere observed for the total sample, although there was some

evidence for nonsignificant increases in risk for Caucasian andAfrican-American women with higher DDE levels. More re-cently, a European study compared DDE adipose tissue levelsof 265 postmenopausal women with breast cancer to DDElevels of 341 controls (28). Participants were recruited in study

centers in Germany, the Netherlands, Northern Ireland, Swit-zerland, and Spain. Results indicated that mean adipose tissuelevels were lower among women with breast cancer than incontrols, and no increase in risk was observed in associationwith higher DDE body burden. Similarly, in a Mexican study

(29), no difference in mean serum concentrations of DDE andp’p’-DDT were found between 141 breast cancer patients and141 hospital controls, nor were higher serum levels of these

compounds related to increased risk. Finally, a case-controlstudy nested within the Nurses’ Health Study cohort was con-

ducted among 236 women with breast cancer and 236 controls(30). Blood samples were obtained before breast cancer wasdiagnosed in the case group. There was no evidence for greaterrisk of breast cancer among women with higher plasma levels

of DDE or PCBs.In this case-control study, we examined the overall asso-

ciation between serum levels of DDE, HCB, mirex, total PCBs,

and PCB congener groups and risk of postmenopausal breastcancer, and possible effect modification by history of lactation.This research adds to the existing body of evidence in severalways: first, the effect of organochlorines on breast cancer risk

was examined in a group homogenous with respect to meno-

pausal status, postmenopausal women; second, because of alarger sample size, we were able to determine the effects ofthese compounds in subgroups of women, defined by history of

lactation; third, because of the availability of the extensive

questionnaire data, we were able to adjust risk estimates for all

known and suspected confounders; fourth, all risk estimateswere adjusted for serum lipids; and fifth, we were able to

examine the effect of PCBs in much greater detail, due to the

availability of congener specific data.

Materials and Methods

The effect of environmental exposure to organochlonnes onbreast cancer risk was examined in a subset of participants from

a case-control study of postmenopausal breast cancer. A moredetailed description of this study population has been publishedelsewhere (31). Briefly, women in this study were enrolledfrom 1986 to 1991 in Erie and Niagara counties in western NewYork. Included were 439 postmenopausal breast cancer casesand 494 postmenopausal community controls between the agesof 41 and 85 years. All participants were Caucasian. Cases were

identified from most area hospitals and were interviewed within2 months of diagnosis. Of the eligible 777 women with breast

cancer, 439 (56.5%) were interviewed. The primary reason for

nonparticipation was physician refusal. Controls were 1076female residents of the two counties; names were obtained fromHealth Care Finance Administration and New York State De-partment of Motor Vehicles records. Of the eligible postmeno-pausal women, 494 (45.9%) agreed to participate. Cases andcontrols were interviewed in person by trained interviewers.

The interview took approximately 2 h to complete and includedassessment of medical history, reproductive history, occupa-tional history, exposure to exogenous hormones, family history

of cancer, and a food frequency questionnaire that assessedusual intake 2 years prior to the interview.

Of women who provided usable interviews, approximately63% agreed to donate a blood sample. Blood was drawn for 262

postmenopausal breast cancer cases and 3 19 controls. Bloodsamples from most women with breast cancer were collected

within 3 months of surgery. Blood was processed immediately

in the laboratory staffed by a trained laboratory technician and

placed in a freezer. Since collection, the saved samples haveremained frozen at -70#{176}C.A subset of the stored blood spec-

imens was used for these toxicological analyses: 154 womenwith postmenopausal breast cancer and 192 controls. Caseswere only included in the final study sample if their blood was

drawn before chemotherapy or radiation, and within 3 monthsof surgery. Controls were frequency matched to cases by date

of blood draw (±3 months) and age (±3 years).

Laboratory Methods. The laboratory analyses were per-formed by the Toxicology Research Center Analytical Labora-tory at State University of New York at Buffalo. The concen-trations of DDE, HCB, mirex and 56 PCB peaks, representing

73 congeners, in the serum samples were determined by themethod of Greizerstein et a!. (32). The procedures includestandardized extraction, clean-up, and quantification by high-

resolution gas chromatography and comprehensive quality as-surance program to minimize systematic and random errors.Trained and blinded laboratory technicians used 2 g of serum

by weight to determine levels of DDE, HCB, mirex, and PCBcongeners, measured in ng/g of serum. The sample was mixedwith solutions containing International Union of Pure and Ap-plied Chemistry (IUPAC) isomers 46 and 142 (surrogate stand-

ards). Methanol was added to precipitate the proteins, and theresulting mixture was extracted with hexane. The extract was

concentrated and then cleaned by passing through a Florisil

column. The eluate was evaporated to a small volume, andisomers 30 and 204 were added as internal standards. An

aliquot of the mixture was injected into the gas chromatographequipped with an electron capture detector. Quantification was



Cancer Epidemiology, Biomarkers & PreventIon /83

based on calibration standards and response factors calculatedusing purchased reference materials. The quality control activ-

ities consisted of analyses of samples in batches of 6-10

simultaneously with quality control samples. The quality as-

surance program checked that the procedures were under con-

trol by the use of control charts and set the criteria for data

acceptability. The LOD for each analyte was determined as themean of background noise plus 3 SDs in five reagent blank

samples. The LODs for the major analytes were 0.09 ng/g for

DDE, 0. 12 ng/g for HCB, 0.06 ng/g for mirex, and 0.09 nglg for

PCB congener 153, a congener that was detected in the sera ofall participants. The Toxicology Research Center participates inthe Great Lakes Research Program Quality Assurance/Quality

Control Program sponsored by the Agency for Toxic Sub-

stances and Disease Registry and in the Northeast/Mid-AtlanticBreast Cancer Quality Assurance/Quality Control Study.

Statistical Analyses. Risk of breast cancer was examined for

serum DDE, HCB, mirex, and PCBs. The effect of PCBs on

risk was examined using several measures of PCB exposure.

First, the detected levels of the 56 PCB peaks were added toobtain a measure of total PCBs. Second, the number of detected

PCB peaks was determined to obtain an alternative measure of

total PCB exposure. Third, three groups of PCB congeners were

determined by adding the detected levels of less chlorinated

PCB congeners (IUPAC numbers 6, 7 + 9, 18, 19, 22, 23, 33,15 + 17, 16 + 32, 25 + 50, 31 + 28, 40, 45, 49, 52, 55, 60,

64, 70, 42 + 59, 47 + 48, 66 + 95, and 77 + 1 10), moderatelychlorinated PCB congeners (IUPAC numbers 87, 97, 99, 101,118, 151 + 82, 129, 134, 135, 136, 138, 147, 149, 153, 141 +

179, 128 + 167, 171 + 156, 172, 176, 177, 180, 183, 185, 187,

188, 174 + 181), and more highly chlorinated PCB congeners

(IUPAC numbers 194, 195, 200, 203, 205, 206). Grouping wasbased on biological activity, as well as on data availability. In

our initial attempts to group these compounds (e.g. , by enzymeinduction or estrogenic activity), we discovered that mostwomen had no detectable levels for these highly reactive con-

geners. Thus, our decision to group by degree of chlorinationwas in part driven by the fact that we had available data for all

participants.

Descriptive analyses included Student’s t tests of means

for cases and controls for lifestyle, reproductive, and dietaryvariables, and �2 tests for categorical variables. Multivariate

ANOVA was used to obtain age and lipid adjusted means forthe organochlorine variables. ORs were calculated by uncon-ditional logistic regression with 95% CIs computed from the SE

of the regression coefficient. Most exposure categories were

examined in tertiles, based on the distribution of the individual

compounds in the controls. Serum mirex levels were dichoto-mized into detectable and nondetectable levels (referent) be-

cause of the small number of women with detectable levels(27.3%). For the less chlorinated congeners, women with levels

below the LOD (30.1%) served as the referent group and werecompared to women in the lower and upper halves of women

with detectable levels. ORs were adjusted for potential con-

founders, including age, education, family history of breast

cancer, parity, quetelet index, duration of lactation, age at first

birth, years since last pregnancy, fruit and vegetable intake, and

serum lipids. Covariates were only included in the final regres-sion model if they were established risk factors in these data or

changed the observed risk estimate by at least 15%. Lipidadjustment was modeled after the method described by Phillips

et a!. (33), in a study in which serum triglycerides and totalcholesterol levels were used to estimate total serum lipid con-

tent. To determine and describe effect modification by lacta-

lion, which was expected, due to differences in elimination of

organoch.lorines, women were also stratified by lactation his-

tory, excluding nulliparous women (n = 48). Subgroup analy-

ses were not performed for the less chlorinated PCB congeners

because lactation is not likely to be an important route of

excretion of these compounds in that they are metabolized

within weeks to months after exposure (34-36).

Results

Descriptive characteristics for cases and controls are shown inTable 1 for the total study population and stratified by history

of lactation. In the total study population in this nested study,cases and controls did not differ significantly with respect to

demographic, reproductive, and dietary variables. Cases weremore likely to have a positive family history of breast cancer

and less likely to reside in rural areas than controls. Amongwomen who had never lactated, cases and controls were very

similar for most of these descriptive variables, with the excep-

tion of significantly older age of menopause for cases. Simi-larly, among women who had lactated, we observed few dif-

ferences between the study groups. Cases were more likely to

have a family history of breast cancer, consumed fewer vege-

tables, and were less likely to reside in rural areas than controls.Serum concentrations, adjusted for age and serum lipids,

of DDE, HCB, mirex, and PCBs in the entire study populationand in the lactation groups are presented in Table 2. Again,nulliparous women were excluded from these subgroup analy-

ses. In the total sample, cases tended to have slightly higher

mean serum organochlorine concentrations than controls, with

the exception of HCB. Among women who had never lactated,

cases had higher serum levels of all compounds under investi-

gation. Although these differences in means were of greater

magnitude than in the total study sample, none were statisticallysignificant. In contrast, among women who had ever lactated,

cases had slightly lower levels than controls of most of thesecompounds, with the exception of total number of PCB con-

gener peaks detected and more highly chlorinated PCBs.Risk of postmenopausal breast cancer associated with en-

vironmental exposure to DDE, HCB, and mirex is shown inTable 3. In the total sample there was no evidence of greater

risk of breast cancer for women with the highest serum levels

of DDE (third tertile OR, 1.34; 95% CI, 0.71-2.55) and HCB(third tertile OR, 0.81; 95% CI, 0.43-1.53) or with detectablelevels ofmirex (OR, 1.37; 95% CI, 0.78-2.39). Among women

who had never lactated, there was a suggestion of excess risk

among women in the second and third tertile of the DDE

distribution (OR, 1.95; 95% CI, 0.58-6.67; and OR, 1.83; 95%CI, 0.63-5.33, respectively). However, further adjustment forserum PCB levels reduced the magnitude of these estimates[OR, 1.61; 95% CI, 0.61-6.01 and OR, 1.32; 95% CI, 0.59-4.08, respectively (data not shown)]. Similarly, in this group,there was some evidence for an increase in risk as a function of

increasing HCB levels (third tertile OR, I .79; 95% CI, 0.59-

5.40), but again, further adjustment for serum PCBs substan-

tially reduced this estimate (OR, I .2 1 ; 95% CI, 0.50-4. 1 1 ; data

not shown). Among women who had never lactated, those with

detectable levels of mirex had a marginally significant 2-foldincrease in risk compared to women with no detectable levels.

This effect persisted after serum PCBs and DDE levels were

entered into the model. Among women who ever lactated, therewas no association with risk and serum DDE and mirex. As for

HCB, inverse associations were observed among women with

higher HCB levels, which were most pronounced and statisti-




Table 1 Characteristics of postmenopausal breast cancer cases and community controls, westem New York, 1986-1991

Total Never lactated” Ever lactated

Characteristic Cases Controls � Cases Controls � Cases Controls

(n - 154) (n = 192) (n 46) (n 61) (n 85) (n 106)

Age” 64.1 (7.7) 63.2 (7.6) NS’ 63.9 (5.9) 60.8 (7.6) NS 64.5 (8.1) 64.8 (7.2) NS

Education (yr)” 12.5 (2.8) 12.2 (2.6) NS 12.5 (1.8) 12.4 (2.6) NS 12.5 (3.0) 12.0 (2.6) NS

Quetelet index” 25.7 (4.9) 25.9 (5.3) NS 26.1 (6.3) 26.1 (6.0) NS 25.6 (4.2) 26.1 (4.7) NS

Age at menarche5 12.9(1.6) 12.9(1.6) NS 12.8(1.5) 12.9(1.7) NS 13.0(1.6) 12.8(1.5) NS

Age at menopause5 47.5 (5.7) 47.0 (5.9) NS 47.8 (5.9) 45.5 (5.4) 0.� 47.9 (5.6) 47.5 (5.7) NS

Age at first pregnancy” 24. 1 (4.8) 23.4 (4.3) NS 24.4 (4.6) 24.3 (4.5) NS 23.7 (4.7) 22.8 (4.2) NS

Number of live births(�.b 3.2 ( 1 .9) 3.4 ( 1 .9) NS 3.4 ( 1 .7) 3.0 ( I .5) NS 3.3 (2.0) 3.7 (2.0) NS

Years since last lactation5t 33.8 (9.9) 33.9 (8.6) NS NA NA NA 33.8 (9.9) 33.9 (8.6) NS

Months of lactationbf 8.6(13.1) 10.20(12.4) NS NA NA NA 8.6(13.1) 10.2 (12.4) NS

Benign breast diseases 23% 20% NS 26% 15% NS 21% 20% NS

Family history of breast cancers 18% 9% 0.02 13% 10% NS 20% 9% 0.04

Fruit (g/month)��h 8420 (5250) 8720 (5000) NS 7010 (4460) 7290 (4130) NS 8830 (5210) 9630 (5400) NS

Vegetables (g/month)i�h 12680 (5230) 13980 (7340) NS 12160 (5300) 12200 (4960) NS 12790 (5230) 14890 (8400) 0.05

Dairy (g/month)’� 8480 (6320) 7840 (6180) NS 6970 (5140) 7720 (7440) NS 9360 (6580) 7890 (5750) NS

Fish (g/month)’� 787 (586) 833 (621) NS 766 (597) 877 (529) NS 773 (582) 791 (666) NS

Meats (g/month)55 1640 (940) 1720 (1000) NS 1540 (840) 1800 (930) NS 1710 (1040) 1680 (1040) NS

Residence”’

Urban 36% 31% 33% 39% 38% 26%

Suburban 60% 53% 56% 44% NS 61% 57%

Rural 4% 16% 0.01 1 1% 17% 1% 17% 0.001

Occupation”�

Professional 18% 17% 11% 20% 18% 11%

Sales/administration 44% 43% 41% 44% 44% 43%

Service 17% 15% 22% 7% 19% 21%

Labor 21% 25% NS 26% 29% NS 20% 25% NS

‘, Women with at least one live birth, excluding 48 nulliparous women.S Mean (SD) differences in means were examined with Student’s : test.

(, NS, not significant (P > 0.05); NA, not applicable.d Weight/height2.

, Among women with at least one pregnancy.

I Among women who had ever breastfed an infant.

g Differences between groups were examined with the s� test.h Reflects intake 2 years before the interview.

‘ Place of residence at time of interview.J Occupation 2 years before the interview.

Table 2 Serum concentrations of organochlorines in postmenopausal breast cancer cases and community controls, western New York, 1986-1991”

Total Never lactated’ Ever lactatedMeasure in ng/g of serums’

Cases (n = 154) Controls (n = 192) Cases (n = 46) Controls (n = 61) Cases (n = 85) Controls (n = 106)

DDE 1 1.47 (10.49) 10.77 (10.64) 13.16 (1 1.65) 10.82 (10.91) 10.36 (8.97) 10.44 (10.43)

HCB 0.41 (0.19) 0.42 (0.19) 0.45 (0.24) 0.39 (0.18) 0.39 (0.16) 0.44(0.19)

Mirex 0.043 (0.09) 0.037 (0.09) 0.083 (0.12) 0.046 (0.14) 0.029 (0.06) 0.036(0.08)

Total PCBs 4.29 (2.40) 4.12 (2.24) 4.63 (2.88) 4.00 (1.96) 4.27 (2.50) 4.30(2.42)

Number of peaks detected 18.38 (5.40) 17.93 (4.99) 18.68 (4.63) 17.93 (5.34) 18.49 (5.98) 18.35 (4.66)

Less chlorinated PCBs 0.31 (0.34) 0.29 (0.35) NA” NA NA NA

Moderately chlorinated PCBs 3.1 1 (1.71) 3.06 (1.73) 3.43 (1.77) 2.90 (1.48) 3.10 (1.67) 3.20(1.91)

More highly chlorinated PCBs 0.43 (0.29) 0.40 (0.25) 0.50 (0.31) 0.40 (0.24) 0.41 (0.27) 0.40(0.26)

a Mean (SD).b Adjusted for age and serum lipids.

‘ Women with at least one live birth, excluding 48 nulliparous women.d NA. not applicable.

cally significant for women in the second tertile of the distri-bution (OR, 0.32; 95% CI, 0.14-0.71).

The effect of PCB body burden on breast cancer risk ispresented in Table 4. In the entire study population, womenwith the highest serum PCB levels or the greatest number of

detected PCB congeners were not at greater risk of breastcancer compared to women with the lowest levels or fewer

detected peaks (OR, 1.14; 95% CI, 0.61-2.15; and OR, 1.34;95% CI, 0.72-2.47, respectively). However, among women

who had never lactated, there was some indication of increasing

risk with increasing serum PCBs (third tertile OR, 2.87; 95%

CI, 1.01-7.29) and increasing number of detected peaks (thirdtertile OR, 3.31; 95% CI, 1.04-1 1.3). These effects were stillobserved after adjustment for serum DDE and HCB (data not

shown). For women who had ever lactated, there was no evi-dence for an adverse effect of these exposures on risk.

As for PCB congener groups, we observed some evidenceof greater risk of breast cancer for women with detectable levels



Cancer Epidemiology, Biomarkers & Prevention 185

Table 3 Risk of postm enopausal breast cancer associated w ith enviro nmental exposure to DDE, HCB, and mirex, wes tern New York, 1986-1991

Measurements in Total (n 346) Never lactated (n = l07)� Ever lactated (n = 191)

ng/g of serum Cases Controls OR” (95% CI) Cases Controls OR’ (95% CI) Cases Controls OR” (95% Cl)

DDE”

1 (low) 54 60 1.0 15 23 1.0 29 29 1.0

2 46 69 1.01 (0.56-1.86) 13 17 1.95 (0.58-6.67) 30 44 0.76 (0.35-l.63

3 (high) 54 63 1.34 (0.71-2.55)

P=0.25

18 21 1.83 (0.63-5.33)

P=0.24

26 33 1.28 (0.54-3.05)

P=0.44

HCBdJ

I (low) 62 61 1.0 16 27 1.0 37 26 .0

2 40 66 0.56 (0.30-1.04) 12 14 1.26 (0.40-3.97) 23 44 0.32(0.14-0.71)

3 (high) 52 65 0.81 (0.43-1.53)

P=0.80

18 20 1.79 (0.59-5.40)

P=0.22

25 36 0.46(0.20-11)8)

P=0.1I

Mirex”

LOD 42 44 1.37 (0.78-2.39) 18 17 2.42 (0.98-4.32) 20 23 1.08 (0.52-2.25)

a ORs adjusted for age, education, family history ofbreast cancer, parity, quetelet index, duration oflactation, age at first birth, years since last pregnancy, fruit and vegetable

intake, and serum lipids.b Women with at least one live birth. excluding 48 nulliparous women.

� ORs adjusted for age. education, family history of breast cancer, parity, quetelet index, age at first birth, years since last pregnancy. fruit and vegetable intake, and serum

lipids.

d Tertiles based on the distribution in all controls.

, DDE tertile cutpoints (ng/g of serum): first tertile. 0.02-5.07: second tertile, 5.10-10.98: third tertile. 11.0-76.2.

1HCB tertile cutpoints (ng/g of serum): first tertile, 0-0.34; second tertile, 0.35-0.44; third tertile, 0.45-1.35.g Mirex levels > LOD ranged from 0.06 to 0.99 ng/g of serum.

of less chlorinated PCBs compared to those without detectable

levels, although this effect was most pronounced for women inthe lower half of the distribution (OR, 2.04; 95% CI, 1.09-

3.83). When all women with detectable levels were compared

to women without detectable levels, the OR was 1.61 and the95% CI was 1.07-3.51. The ORs were similar after adjustment

for serum DDE and HCB levels. No such effect was observedfor the moderately chlorinated PCBs (third tertile OR, 1.37;

95% CI, 0.73-2.59) or the more highly chlorinated PCBs (thirdtertile OR, 1.19; 95% CI, 0.60-2.36). Among women who hadnever lactated, higher levels of moderately chlorinated PCBs(third tertile OR, 3.57; 95% CI, 1.10-8.60), but not morehighly chlorinated PCBs (third tertile OR, 1 .53; 95% CI, 0.47-4.98) were associated with increased risk of breast cancer.

Among women who had ever lactated, there was no evidencefor an increase in breast cancer risk in relation to higher serumconcentrations of moderately or more highly chlorinated PCBs.

Discussion

Results from this case-control study indicated that environmen-tal exposure to organochlorine compounds was not related tobreast cancer risk among postmenopausal women who hadlactated but that higher body burden of PCBs and mirex may be

risk factor for parous women who had never lactated. In theentire sample, we did not observe increases in risk associatedwith higher serum levels of DDE, HCB, mirex, and PCBs, with

the exception of a modest increase in risk associated withhaving detectable levels of less chlorinated PCB congeners.

Elevated serum levels of DDE were not associated with

breast cancer risk, although there was some evidence for greater

risk with higher levels among women who had never lactated.Attenuation of this effect when serum PCB levels were in-

cluded into the model suggests that the observed risk waslargely driven by the association of DDE with serum PCB

levels. These findings are in contrast to those reported by Wolffet a!. (26), who observed a nearly 4-fold increase in risk forwomen with the highest levels of DDE, which was unaffectedby inclusion of PCB levels into the regression model. Similarly,

Krieger et a!. (27) observed a nonsignificant 2-fold increase in

risk of breast cancer among white women with the highest DDE

body burden. Results of three investigations indicated that

greater body burden of DDE or DDT was not associated with

greater risk of breast cancer (28-30).

The effect of HCB on breast cancer risk has not been

previously reported in studies using an epidemiological studydesign. In three studies, mammary adipose tissue levels of

breast cancer patients were compared to those of controls with

benign breast disease (24, 25) or accident fatalities (23). Noneof these studies reported a significant difference in HCB con-

centrations between breast cancer cases and controls. The ab-

sence of an adverse effect of HCB exposure in our data is

consistent with these earlier studies.

The association between body burden of mirex and breast

cancer risk has not been explored previously. In these data, we

found no increase in risk associated with having detectablelevels of mirex in the entire study population or among parous

women who had lactated. However, we observed a borderline

significant increase in risk among parous women who hadnever lactated. The latter finding needs to be considered with

caution, for the observed increase in risk was based on very

small numbers of women with detectable levels of this corn-

pound (n = 35). Future investigations should examine theassociation of mirex with breast cancer risk in populations withgreater exposure levels (e.g. , fish consumers or women residing

in the southern United States).In these data, we observed a modest increase in risk for

women with detectable levels of less chlorinated PCBs. Among

the never lactating parous women with higher levels of total

PCBs, moderately chlorinated PCBs, or greater numbers of

detected congener peaks, there was some increase in risk. The

investigation of the less chlorinated congeners in relation to risk

was problematic, because these compounds are metabolized

rapidly, and measured levels reflect only recent exposure (3).However, these congeners have been associated with greatertoxicological activity, including estrogenic activity, than someof the more highly chlorinated congeners (3). The rationale for



Table 4 Risk of postmenopausal breast cancer associated with environmental exposure to PCBs, western New York, 1986-1991

Measures in ng/g

of serum

Total (n = 346) Never lactated (n = 107)” Ever lactated (n 191)

Cases Controls OR” (95% CI) Cases Controls OR’ (95% CI) Cases Controls OR” (95% CI)

Total �Bs”

I (low) 53 2 1.0 1 1 22 1.0 32 29 1.0

2 45 68 0.70(0.37-1.29) 15 21 1.71 (0.55-5.35) 22 41 0.38 (0.17-1.03)

3 (high) 56 61 1.14(0.61-2.15)

P = 0.51

20 18 2.87(1.01-7.29)

P = 0.07

31 36 0.71(0.31-1.61)

P = 0.72

No. of peaks detected’��

I (low) 48 64 1.0 10 21 1.0 30 21 1.0

2 43 64 0.88 (0.45-1.59) 14 20 1.61 (0.41-3.56) 23 37 0.63 (0.29-1.40)

3 (high) 63 64 1.34 (0.72-2.47)

P = 0.52

22 20 3.31 (1.04-1 1.3)

P = 0.10

32 38 0.82 (0.37-1.83)

P = 0.85

Less chlorinated PCBs”5

median, 0.32-1.65.

‘ NA. not applicable.

J Moderately chlorinated PCBs tertile cutpoints (ng/g of serum): first tertile, 0.47-2.19; second tertile, 2.20-3.12; third tertile, 3.13-15.07.k More highly chlorinated PCBs tertile cutpoints (ng/g of serum): first tertile, 0.01-0.25; second tertile, 0.26-0.44; third tertile, 0.45-1.30.


years since last pregnancy. fruit and vegetable

examining these less chlorinated PCBs was based on this bio-

logical significance, as well as on the assumption that exposureto these compounds was likely to be chronic in nature, notchanging significantly over time. Therefore, serum levels at thetime of the interview were assumed to be representative of

lifetime exposure, including the time critical for tumorigenesis.In these data, a statistically significant increase in risk (OR,I .66) was observed for women with detectable levels of theseless chlorinated compounds in comparison to women with no

detectable levels. There was, however, no evidence of a dose-response relationship. These results need to be interpreted withgreat caution, because of the underlying assumption that currentserum levels of these compounds actually reflect levels at the

biologically relevant time period. Furthermore, measurement ofthese rapidly metabolized congeners was more likely to besubject to laboratory error than measurement of the more stable

and persistent congeners. On the other hand, there is no reasonto believe that measurement error was different for cases and

controls, because the laboratory technicians were blinded todisease status. Nondifferential misclassification may have at-tenuated the true risk estimate. Clearly, the effect of thesecongeners on breast cancer risk needs to be explored in aprospective study design, in which it may be possible to obtain

serum levels of these compounds from a time period that

precedes disease onset by an appropriate induction period.The observed effects of total PCBs, moderately chlori-

nated PCBs, and number of detected peaks warrants furthercomment. Number of detected peaks was examined in additionto the measure of total serum PCB levels to determine whether

prevalence of a greater variety of congeners affects risk differ-ently than total amount of these compounds. In our data, therewas no evidence of such a difference. Furthermore, the greatest

contributor to serum PCB levels were the moderately chlori-

nated PCBs; thus, all three PCB exposure measures (totalPCBs, moderately chlorinated PCBs, and number of PCBpeaks) were highly correlated, and observed risk elevations

among women who had never lactated were likely to reflect thesame effect. However, the observation that all three PCB cx-posures produced similar risk associations among parouswomen who had never lactated strengthens the notion that

environmental exposure to PCBs may be related to risk in thisgroup.

As indicated above, previous epidemiological studies did

not report an adverse effect of serum PCB levels (26, 27, 30).However, the designs of these studies differed from that of this

research in some important aspects. First, these studies corn-



Cancer Epidemiology, Biomarkers & Prevention 187

bined pre- and postmenopausal breast cancer patients in theircase group. There is some evidence that risk factors differ bymenopausal status (37-39). Nevertheless, the biological mech-

anism of an organochlorine effect on breast cancer risk is notwell understood. It may be that there is a threshold effect, that

high body burden and chronic exposure to these compounds isnecessary to produce a biological effect. Older, postmenopausalwomen would be more likely to have been exposed chronically

and to have experienced higher levels of exposure. Second, riskestimates in two of the earlier studies were not adjusted for

serum lipids (26, 27); such adjustment had some impact on themagnitude of the ORs in this study. Third, statistical adjustment

for potential confounders, such as parity (26, 27) and duration

of lactation (27), was lacking in some of these investigations.

In general, our results suggest that an adverse effect oforganochlorine body burden on breast cancer risk, if present at

all, is restricted to parous women who had never lactated. Theobserved effect modification may be explained in three ways.First, organochlorine body burden may have been measuredmore accurately among women who had never breast-fed an

infant. Serum levels in this group may represent a more validmeasure of chronic exposure, uninterrupted by elimination ofthese compounds through lactation. Second, women who had

ever breast-fed an infant may not be at increased risk of breastcancer, if they eliminated a substantial amount of organochlo-

rime body burden at a biologically relevant period of time.Third, lactation in itself may contribute to the terminal differ-

entiation of the mammary epithelium, resulting in larger corn-partments of nonproliferating cells (40, 41). Women who lac-

tated may be less susceptible to the potentially adverse effect of

PCBs. In this study population, history of lactation was asso-ciated with a slight reduction of breast cancer risk (42). We

were unable to examine the effect of organochlorines on riskamong nulliparous women, due to the small number of womenin this group (n = 48). Nulliparous women were excluded from

the subgroup analyses because of the lack of information on theeffect of pregnancy on organochlorine body burden and the

possibility that pregnancy itself may affect accumulation ofthese compounds (43). Further, when nulliparous women wereincluded in the group of women who had never lactated, the

observed increases in risk were more pronounced. Thus, in this

study, we were concerned about the appropriateness of corn-bining parous and nulliparous women without having a clear

understanding of the patterns of accumulation and kinetics inrelation to pregnancy and lactation.

There are several limitations associated with this research

that need to be considered in interpreting these findings. Thelow participation rates in the case and control group may haveintroduced error due to selection bias. Among the breast cancer

case group, nonparticipation may have resulted in a case sample

that is not representative of all women with breast cancer.However, the majority of nonparticipation of cases was due to

their physicians’ refusal to allow the patients to be contacted bythe interviewers. Thus, nonparticipation of the cases may reflect

physician characteristics, rather than patient characteristics,which may minimize the potential influence of bias on theseresults. Nonparticipation of controls was of equal concern; itmay have resulted in a more motivated and possibly morehealth-conscious control group. Results from a brief telephone

interview comparing a sample of controls who refused to par-ticipate with a sample of those who agreed to participate mdi-cated that these groups did not differ with regard to dietaryintake of fruits, vegetables, and meats, nor did they differ withrespect to cigarette use (31). It is unknown whether nonpartici-

pation among cases and controls was related to exposure to

PCBs. Selection bias may have been additionally introduced

among the participants included in the toxicological analyses.

Participants who agreed to provide a blood sample were more

likely to have breast-fed an infant than those who refused, for

both the case and control groups. Among participants who were

selected for toxicological analyses, however, this difference

with respect to lactation history was more pronounced in the

case group than in the control group. Thus, in the sample

available for this study, breast cancer cases with a history of

lactation were likely to be overrepresented.

Another concern in interpreting these findings relates to

the small number of women in both lactation groups. When

these groups of women were divided into tertiles, the numbers

in the cells became small and the corresponding risk estimates

became unstable. Despite these sample size restrictions, a sta-

tistically significant increase in risk was observed for several

PCB exposures. Lack of statistical power may have prevented

the detection of more subtle risk elevations in association with

organochlorine exposure in the entire sample.

Two further limitations of this study relate to the use of

serum levels of these compounds as a method of exposure

assessment. With regard to studying breast cancer, the obvious

tissue of choice would be mammary adipose tissue. Serum

organochlorine levels can only function as a surrogate measure

of body burden in the target tissue. Variations in serum organo-

chlorines have been observed with respect to serum lipids (44).

There is, however, good evidence that lipid adjusted serum

measures estimate the tissue levels (44-46). A more serious

concern with respect to use of serum levels involves the fact

that the blood sample in the case group was obtained after

diagnosis of breast cancer. It could not, therefore, be deter-

mined whether PCB levels were associated with the disease

process or were the result of metabolic changes associated with

disease progression. Although cancer treatment may affect or-

ganochlorine levels (47), case selection for toxicological anal-

ysis excluded those who underwent chemotherapy or radiation

therapy before the blood sample was selected. However, there

may still be an effect of the disease process on measured levels

of these compounds.

In conclusion, results from this study indicate that envi-

ronmental organochlorine exposure is a risk factor for breast

cancer among parous postmenopausal women who had never

lactated but not among women who had ever lactated. The

observed effects in the latter group should be considered with

caution, because they are based on a small number of partici-

pants. Clearly, the role of organochlorine exposure in breast

cancer etiology needs to be explored further in future research

efforts. Ideally, a long-term prospective study design should be

used to examine serum levels of these compounds that reflect

body burden of a time period preceding the onset of this

disease, thus ruling out a potential effect of the disease process

itself on measured levels of these compounds. A prospective

study could also examine the effect of the less chlorinated

congeners with more precision, avoiding the assumption that

present levels of these PCBs reflect past exposure. Future

research, case-control studies as well as cohort studies, should

also aim at examining this association among premenopausal

women. Finally, future studies should use similar epidemiolog-

ical (e.g. , determination of PCB congener group and treatment

of samples with measures below the LOD) and laboratory (e.g.,

proficiency testing and sample acquisition and storage) meth-

odologies, to ensure comparability of results across studies.




References

I. American Cancer Society. Cancer Facts and Figures, 1996, pp. 1 1-13. Atlanta:

American Cancer Society. 1996.

2. Kelsey, J. L., and Hom-Ross. P. L. Breast cancer: magnitude of the problem

and descriptive epidemiology. Epidemiol. Rev., 15: 7-16, 1993.

3. Jensen, A. A. (ed). Levels and trends of environmental chemicals in human

milk. In: Chemical Contaminants in Human Milk. pp. 199-207. Boca Raton, FL:

CRC Press, 1991.

4. IARC. DDT and associated compounds. IARC Monogr.. 53: 179-249, 1992.

5. Safe, S.. Bandiera, S., Sawyer, T., Robertson, L., Safe, L., Parkinson, A.,

Thomas, P. E., Ryan, D. E., Reik, L. M., Levin, W., Denomme, M. A., and Fujita.

T. PCBs: structure-function relationships and mechanism of action. Environ.

Health Perspect.. 60: 47-56, 1985.

6. Safe, S. Polychlorinated biphenyls (PCBs): mutagenicity and carcinogenicity.

Mutat. Res., 224: 79-88, 1989.

7. Moms, C. R., and Cabral, J. R. P. Hexachlorobenzene. Proceedings of an

Intemational Symposium. IARC Sci. PubI., 77: 13-18, 1986.

8. Waters, E. M., Huff, J. E.. and Gerstner, H. B. Mirex. An Overview. Environ.

Res., 14: 212-222, 1977.

9. Lewandowski, M., Levi, P., and Hodgson, E. Induction of cytochrome p-45O

isozymes by mirex and chordecone. J. Biochem. Toxicol., 4: 195-199, 1989.

10. Bulger. W. H., and Kupfer, D. Estrogenic action of DDT analogs. Am. J. Ind.

Med., 4: 163-173, 1983.

1 1 . McKinney. J. D., and WaIler, C. L. Polychlorinated biphenyls as hormonally

active structural analogues. Environ. Health Perspect., 102: 290-294, 1994.

12. Jansen, H. T., Cooke, P. S., Poecelli, J., Lui, T. C., and Hansen, L. G.

Estrogenic and antiestrogenic actions of PCBs in the female rat: in vitro and inviva studies. Reprod. Toxicol., 7: 237-241, 1993.

13. Foster, W. 0., McMahon, A., Villeneuve, D. C., and Jarrell, J. F. Hexachlo-robenzene (HCB) suppresses circulating progesterone concentrations during the

luteal phase in the cynomolgus monkey. J. AppI. Toxicol., 12: 13-19, 1992.

14. Foster, W. G., Pentick, J. A., McMahon, A., and Lecavalier, P. R. Ovarian

toxicity of hexachlorobenzene (HCB) in the superovulated female rat. J. Biochem.

Toxicol.. 7: 1-11, 1992.

15. Banerjee, B. D. Effects of sub-chronic DDT exposure on humoral and

cell-mediated immune responses in albino rats. Bull. Environ. Contam. Toxicol.,

39: 827-834, 1987.

16. Kimbrough, R. D. Laboratory and human studies on polychlorinated biphe-

nyls (PCBs) and related compounds. Environ. Health Perspect., 59: 99-106,1985.

I 7. Kimbrough, R. D. Human heath effects of polychlorinated biphenyls (PCBs)

and polybromated biphenyls (PBBs). Annu. Rev. Pharmacol. Toxicol., 27: 87-

Ill. 1987.

18. Norback. D. H., and Weltman, R. H. Polychionnated biphenyl induction of

hepatocellular carcinoma in the Sprague-Dawley rat. Environ. Health Perspect.,

60: 97-105. 1985.

19. Moser, G. J., Meyer. S. A., and Smart, R. C. The chlorinated pesticide Mirex

is a novel Nonphorbol ester-type tumor promoter in mice skin. Cancer Res., 52:

631-636, 1992.

20. Moser, G. J., Robinette, C. L., and Smart, R. C. Characterization of skin

tumor promotion by mirex: structure-activity relationships, sexual dimorphism

and presence of Ha-ras mutation. Carcinogenesis (Lond.), 14: 1 155-1 160, 1993.

21 . Wassermann, M., Nogueira, D. P., Tomatis, L., Mirra, A. P., Shibata, H.,

Arie. 0., Cucso, S., and Wassermann, D. Organochlorine compounds in neoplas-

tic and adjacent apparently normal breast tissue. Bull. Environ. Contam. Toxicol.,

15: 478-484, 1976.

22. Unger, M., Kiaer, H., Blichert-Toft, M., Olsen, J., and Clausen, J. Organo-

chlorine compounds in human breast fat from deceased with and without breast

cancer and in biopsy material from newly diagnosed patients undergoing breast

surgery. Environ. Res., 34: 24-28, 1984.

23. Mussalo-Rauhamaa, H., Haesaenen, E., Pyysalo. H., Antervo, K., Kauppila,R., and Pantzar, P. Occurrence of (3-hexachlorocyclohexane in breast cancer

patients. Cancer (Phila.), 66: 2124-2128, 1990.

24. Falck, F., Ricci, A., Wolff, M. S., Godbold, J., and Deckers, P. Pesticides and

polychlorinated biphenyl residues in human breast lipids and their relation to

breast cancer. Arch. Environ. Health, 47: 143-146, 1992.

25. Dewailly, E.. Dodin, S., Verreault, R., Ayotte, P., Sauve, L., Morn, J., and

Brisson, J. High organochlorine body burden in women with estrogen receptor-positive breast cancer. J. NatI. Cancer Inst.. 86: 232-234. 1994.

26. Wolff, M. S., Toniolo, P. G., Lee, E. W., Rivera, M., and Dubin, N. Blood

levels of organochlorine residues and risk of breast cancer. J. NatI. Cancer Inst.,85: 648-652, 1993.

27. Krieger, N., Wolff, M. S., Hiatt, R. A., Rivera, M., Vogelman, J., andOrentreich, N. Breast cancer and serum organochlorines: a prospective study

among white. black, and Asian women. J. NatI. Cancer Inst., 86: 589-599, 1994.

28. Van’t Veer, P., Lobbezoo, I. E., Martin-Moreno, J. M., Guallar, E., Gomez-

Aracena, J., Kardinaal, A. F. M., Kohlmeier, L., Martin, B. C., Strain, J. J.,

Thamm, M., van Zoonen, P., Baumann, B. A., Huttunen, J. K., and Kok, F. J.DDT (dicophane), and postmenopausal breast cancer in Europe:case control

study. Br. Med. J., 315: 81-85, 1997.

29. Lopez-Carrillo, L., Blair, A., Lopez-Cervantes, M., Cebrian, M., Rueda, C.,

Reyes, R., Mohar, A., and Bravo, J. Dichlorodiphenyltrichloroethane serum levels

and breast cancer risk: a case-control study. Cancer Res., 57: 3728-3732, 1997.

30. Hunter, D. J., Hankinson, S. E., Laden, F., Colditz, G. A., Manson, J. E.,

Willett, W. C., Speizer, F. E., and Wolff, M. S. Plasma organochlorine levels and

the risk of breast cancer. N. Engl. J. Med., 337: 1253-1258, 1997.

3 1. Graham, S., Hellmann, R., Marshall, J. R., Freudenheim, J., Vena, J., Swan-

son, M., Zielezny, M., Nemoto, T., Stubbe, N., and Raimondo, T. Nutritional

epidemiology of postmenopausal breast cancer in western New York. Am. J.Epidemiol., 134: 552-566, 1991.

32. Greizerstein, H., Gigliotti, P., Vena, J. E., Freudenheim, J., and Kostyniak,

P. J. Standardization of a method for routine analysis of polychlorinated biphenyl

congeners and selected pesticides in human serum and milk. J. Anal. Toxicol., 21:

558-566, 1997.

33. Phillips. D. L., Pirkle, J. L., Burse, V. W., Bemert, J. T., Henderson, L. 0.,

and Needham, L. L. Chlorinated hydrocarbon levels in human serum: effects of

fasting and feeding. Arch. Environ. Contain. Toxicol., 18: 495-500, 1989.

34. McFarland, V. A., and Clarke, J. U. Environmental occurrence, abundance,

and potential Toxicity of polychlonnated biphenyl congeners: considerations for

a congener-specific analysis. Environ. Health Perspect.. 81: 225-240, 1989.

35. Wolff, M. S., and Toniolo, P. G. Environmental organochlorine exposure as

a potential etiologic factor in breast cancer. Environ. Health Perspect., 103

(Suppl.): 141-145, 1995.

36. Woodruff, T., Wolff, M. S., Davis, D. L., and Hayward, D. Organochlorineestimation in the study of cancer etiology. Environ. Res., 65: 132-144, 1994.

37. de Waard, F. Premenopausal and postmenopausal breast cancer: one disease

or two? J. NatI. Cancer Inst., 63: 549-552, 1979.

38. Janerich, D. T., and Hoff, M. B. Evidence for a crossover in breast cancer risk

factors. Am. J. Epidemiol., 116: 737-742, 1982.

39. Velentgas, P., and Daling, J. R. Risk factors for breast cancer in younger

women. Monogr. Natl. Cancer Inst., 16: 15-24, 1994.

40. Jahnke, G. D., and Barrett, J. C. Environmental causes and molecular biology

of breast cancer: Future research directions. In: S. Huff (ed), Cellular andMolecular Mechanisms of Hormonal Carcinogenesis: Environmental Influences,

pp. 455-469. New York: Wiley-Liss, 1996.

41. Russo, J., and Russo, I. H. Biological and molecular bases of mammary

carcinogenesis. Lab. Invest., 57: 112-137, 1987.

42. Freudenheim, J. L., Marshall, J. R., Vena, J. E., Moysich, K. B., Muti, P.,Laughlin, R., Nemoto, T., and Graham, S. Lactation history and breast cancer

risk. Am. J. Epidemiol., 146: 932-938, 1997.

43. Buck, G. M., Mendola, P., Vena, J. E., Sever, L. E.. Sweeney, A. M.,

Kostyniak, P., Greizerstein, H., and Olson, J. Prospective pregnancy study, New

York State Angler Cohort. Presented at the International Scientific Conference on

the Effects of the Environment on Human Health in the Great Lakes and St.

Lawrence River Basins. Montreal, Quebec, Canada, May 1997.

44. Needham, L., Burse, V., Head, S., Liddle, J. A., Bayse, D., and Boozer, E. L.

Adipose tissue/serum partitioning of chlorinated hydrocarbon pesticides in hu-

mans. Chemosphere, 20: 975-980, 1990.

45. Brown, J. F., and Lawton, R. W. Polychlorinated biphenyl (PCB) partitioning

between adipose tissue and serum. Bull. Environ. Contain. Toxicol., 33: 277-280,1984.

46. Patterson, D. G., Needham, L. L., Pirkle, J. L., Roberts, D. W., Bagby, J., and

Garret, W. A. Correlation between serum and adipose tissue levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin in 50 persons from Missouri. Arch. Environ., Con-

tarn. Toxicol., 17: 139-143, 1988.

47. Gammon, M. D., Wolff, M. S., Neugut, A. L., Terry, M. B., Britton, J. A.,

Greenbaum, E., Hibshoosh, H., Levin, B., Wang, Q., and Santella, R. Treatmentfor breast cancer and blood levels of chlorinated hydrocarbons. Cancer Epide-miol. Biomark. Prey., 5: 467-471, 1996.



1998;7:181-188. Cancer Epidemiol Biomarkers Prev K B Moysich, C B Ambrosone, J E Vena, et al. breast cancer risk.Environmental organochlorine exposure and postmenopausal

Updated version

http://cebp.aacrjournals.org/content/7/3/181

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cebp.aacrjournals.org/content/7/3/181To request permission to re-use all or part of this article, use this link


http://cebp.aacrjournals.org/content/7/3/181http://cebp.aacrjournals.org/cgi/alertsmailto:[email protected]://cebp.aacrjournals.org/content/7/3/181http://cebp.aacrjournals.org/

Documents

Environmental Organochiorine Exposure and Postmenopausal ... · Preventive Medicine. State University of New York at Buffalo. 270 Farber Hall, Buffalo, NY 14214. Phone: (716) 829-2975;