aged; breast neoplasms; contraceptives, oral; hormone replacement therapy; perimenopause; postmenopause; premenopause; progestins

Transcription

1 American Journal of Epidemiology Copyright ª 2007 by the Johns Hopkins Bloomberg School of Public Health All rights reserved; printed in U.S.A. Vol. 165, No. 10 DOI: /aje/kwm006 Advance Access publication March 3, 2007 Original Contribution Age and Menopausal Effects of Hormonal Birth Control and Hormone Replacement Therapy in Relation to Breast Cancer Risk Sumitra Shantakumar 1, Mary Beth Terry 2, Andrea Paykin 3, Susan L. Teitelbaum 4, Julie A. Britton 4, Patricia G. Moorman 5, Stephen B. Kritchevsky 6, Alfred I. Neugut 2,7, and Marilie D. Gammon 1 1 Department of Epidemiology, University of North Carolina School of Public Health, Chapel Hill, NC. 2 Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY. 3 New York City Department of Health and Mental Hygiene, New York, NY. 4 Department of Community Medicine, Mt. Sinai School of Medicine, New York, NY. 5 Department of Community and Family Medicine, Duke University Medical Center, Durham, NC. 6 Sticht Center on Aging, Wake Forest University School of Medicine, Winston-Salem, NC. 7 Department of Medicine, College of Physicians and Surgeons, Columbia University, New York, NY. Received for publication February 8, 2006; accepted for publication November 1, It is unclear whether breast cancer risk varies by age and menopausal status in relation to use of hormonal birth control (HBC) and hormone replacement therapy (HRT), taken singly or cumulatively. The authors utilized data from 1,478 cases and 1,493 controls aged years with known menopausal status, who had participated in a population-based, case-control study conducted on Long Island during Exogenous hormone use over the lifecourse was assessed by use of memory aids. The authors examined associations among women in these subgroups: premenopausal (n ¼ 968), postmenopausal <65 years (n ¼ 1,045), and postmenopausal 65 years (n ¼ 958). Among premenopausal women, risk was increased for ever use of HBC ( (OR) ¼ 1.37, (CI): 1.04, 1.81) or HRT (OR ¼ 1.81, CI: 1.17, 2.81) and was pronounced among women reporting use of both HBC and HRT (OR ¼ 2.59, CI: 1.50, 4.46), long-term HRT use (OR ¼ 3.93, CI: 1.43, 10.84), or estrogen-plus-progestin therapy (OR ¼ 3.51, CI: 1.45, 8.49). There was no effect of ever HBC use among postmenopausal women aged less than 65 years, but risk was modestly elevated for more than 5 years of HRT use (OR ¼ 1.41, CI: 1.00, 1.99). Among postmenopausal women aged 65 years or more, odds ratios for HBC or HRT use were around the null. These results emphasize that timing of exogenous hormone use is important. Women who used these hormones before menopause had elevated risks, but the harmful effects began to decline with age after menopause. aged; breast neoplasms; contraceptives, oral; hormone replacement therapy; perimenopause; postmenopause; premenopause; progestins Abbreviations: CI, ; HBC, hormonal birth control; HRT, hormone replacement therapy; LIBCSP, Long Island Breast Cancer Study Project; OR,. Despite the large body of evidence (1 4) regarding the carcinogenic effects (5) of hormonal birth control (HBC) and hormone replacement therapy (HRT) on breast cancer risk, it still remains unclear if women are more susceptible to the effects of these exogenous sex steroids at specific points during the lifecourse. The effects of HBC use in younger women and HRT use in early postmenopausal women on breast cancer risk have been extensively studied, but because Correspondence to Dr. Sumitra Shantakumar, Worldwide Epidemiology, GlaxoSmithKline Research and Development, P.O. Box 13398, Five Moore Drive, Research Triangle Park, NC ( sumitra.y.shantakumar@gsk.com). 1187

2 1188 Shantakumar et al. of the different underlying hormonal milieu, these findings may not be generalized to other age and menopausal subgroups. Such subgroup analyses may elucidate a timing effect for current use of HRT among premenopausal women or postmenopausal women aged 65 years or more or a latency effect of prior HBC or HRT use in these older postmenopausal women. Furthermore, few studies (6 13) have investigated the combined effects of HBC and HRT, as ample numbers of women have only recently had the opportunity to be dually exposed. The analysis reported here uses data from a large population-based, case-control study to determine if women in certain age and menopausal subgroups are more susceptible to the harmful effects of exogenous hormones. Our study population provided a unique opportunity to answer these discrete questions, as there were no age restrictions for participant enrollment, which resulted in comparable numbers of younger and older participants. MATERIALS AND METHODS Study population The Long Island Breast Cancer Study Project (LIBCSP) was a population-based, case-control study of breast cancer incidence conducted among adult female residents of Nassau and Suffolk counties, Long Island, New York, between 1996 and Institutional review board approval was obtained from all participating institutions, and signed informed consent was obtained from all participants. Details of the study design have been previously published (14) and are briefly summarized below. were English speakers who were newly diagnosed between August 1, 1996, and July 31, 1997, with primary in situ or invasive breast cancer. Study personnel identified cases by rapid case ascertainment, which involved regular contact with local pathology departments. Physicians of cases were also contacted to confirm breast cancer diagnosis, and permission was obtained to contact the subject. Physician consent was given for 1,837 (90.5 percent) of the 2,030 potential cases determined eligible, but consent was generally refused if a subject had poor health. Control women were English speakers without a personal history of breast cancer, who were frequency matched to the expected age distribution of case women by 5-year age groups. aged less than 65 years (57.9 percent of controls) were identified by use of Waksberg s random digit dialing method (15), and those aged 65 years or more were randomly identified by use of Health Care Financing Administration rosters. The main questionnaire was completed for 1,508 (82.1 percent) eligible cases (235 with in situ breast cancer and 1,273 with invasive breast cancer) and 1,556 (62.7 percent) eligible controls. Subjects ages ranged from 20 to 98 years. Data collection The structured questionnaire ( gov/libcsp/projects/questionnaire.html), which averaged 101 minutes to complete, was administered by trained interviewers in the respondent s home and included questions about known and suspected risk factors for breast cancer. All cases were asked to sign a medical record release form. Medical and pathology records were abstracted to determine clinical characteristics of the primary breast cancer diagnosis. As previously reported (14), factors associated with elevated breast cancer s include lower parity, little to no breastfeeding, later age at first birth, and a positive family history of breast cancer. Exposure assessment Women were asked about HBC and HRT use as part of the questionnaire. To help identify periods of exogenous hormone use, the interviewer used a month-by-month calendar marked earlier in the interview with the participant s reproductive history and other life events. A color chart of all noncontraceptive estrogens and progestins marketed in the United States was also shown to each participant to enhance recall of HRT use. Ever use of HBC was defined as self-reported history of HBC use in the form of pills, shots, or implants for any period of time prior to the reference date (date of diagnosis for cases; date of identification for controls). Ever use of HRT was defined as self-reported history of HRT use in the form of pills, shots, skin patch, vaginal cream, and suppositories for any period of time prior to the reference date. We also created the following variables for HBC and HRT: 1) total duration of use, 2) time since last use, 3) age at first use, 4) time since first use, and 5) recency of use. These variables were first created as continuous variables but later categorized into quartiles based on their distribution among controls. To gain statistical precision, we collapsed some of the quartiled variables on the basis of the similarity of the s in each group. Information on HBC regimens, formulations, and doses was not collected at the baseline interview. The type of HRT was ascertained from the brand name reported for each of use, and we included distinctions for estrogen-only therapy and estrogen-plusprogestin therapy. A composite measure for women who used both HBC and HRT was also created. During the case-control interview, women were asked about surgeries they may have undergone involving partial or total removal of one or both ovaries (including wedge resections, cyst removal, and hysterectomies). Because the extent of ovary removal for each surgery was also queried, we could estimate the total number of ovaries removed by summing across surgeries. Separate indicator variables were created for participants who had one or both ovaries removed as opposed to those with partial removal. Using this information, we then defined unilateral oophorectomy as the complete removal of one ovary and bilateral oophorectomy as the complete removal of both ovaries. Menopausal status was based on self-reported information collected during the baseline interview, including information on a woman s last menstrual period, prior surgical information on hysterectomy or bilateral oophorectomy, cigarette smoking status, and HRT use. Postmenopausal status was defined as the age at which a woman s last menstrual period was at least 6 months from the reference date or the

3 Age, Exogenous Hormones, and Breast Cancer 1189 date of removal of both ovaries. If a subject was taking HRT or had a hysterectomy without removal of both ovaries, her menopausal status was initially classified as unknown (11.8 percent of subjects). Postmenopausal status was subsequently assigned, separately for smokers and nonsmokers, if the woman s reference age was greater than or equal to the 90th percentile age at natural menopause among controls (54.8 years among smoking controls; 55.4 years among nonsmoking controls). Ultimately, only 3.04 percent of the subjects had missing values for menopausal status. The case-control distribution of established risk factors for breast cancer among 1,478 cases and 1,493 controls with known menopausal status used in these analyses did not differ substantially from those identified among all LIBCSP participants (data not shown). Body mass index was calculated as self-reported weight (kg)/height (m) 2 at the reference age. Statistical methods Estimates of effect (s) for HBC use and HRT use were stratified by age and menopausal status to form three groups of women: premenopausal women (n ¼ 968), postmenopausal women <65 years (n ¼ 1,045), and postmenopausal women 65 years (n ¼ 958). Women who did not meet the definition of postmenopausal status were categorized as premenopausal. Unconditional logistic regression (16) was used to estimate s and 95 percent s between breast cancer and selected demographic factors and exogenous hormone use. Where applicable, tests for trend across quantiles were calculated. All models included adjustment for the frequency matching variable, age (5-year age categories). Established and suspected breast cancer risk factors found to be important in the literature and in previous LIBCSP analyses were considered as potential confounders in our analyses. Confounding was evaluated separately for premenopausal women, postmenopausal women <65 years, and postmenopausal women 65 years. Factors considered as potential confounders included body mass index at the reference age, active smoking, weight change between the reference age and age 20 years, physical activity from menarche to reference, lifetime alcohol intake, family history of breast cancer, history of benign breast disease, history of infertility, mammography use, breastfeeding, age at menarche, age at first birth, parity, type of menopause, time since menopause, race, ethnicity, income, and education. Models were also simultaneously adjusted for the other type of exogenous hormone. Multivariable logistic regression (16) was conducted to obtain the most parsimonious models. Covariates were systematically removed from each full model by use of a backwards elimination strategy. None of the potential confounders changed the estimates of either HBC use or HRT use by more than 10 percent among any of the three groups. Forward selection methods also revealed no substantial confounding by any of the covariates considered. Thus, only age-adjusted models are shown. Effect modification on the multiplicative scale by body mass index at the reference age was evaluated for premenopausal women, postmenopausal women <65 years, and postmenopausal women 65 years by using 1) stratified analyses and 2) the log likelihood ratio test to compare logistic models with and without the interaction term (16). A two-sided p value of less than 0.2 was used to assess statistically significant effect modification. Deviations from additivity were also assessed by creating two-by-two interaction tables and then calculating the relative excess risk for interaction and corresponding 95 percent s (17). All analyses were carried out with SAS, version 8.2, software (SAS Institute, Inc., Cary, North Carolina). RESULTS Table 1 presents the associations between breast cancer and demographic characteristics, stratified by age and menopausal status. Some college education was associated with decreased breast cancer risk in the premenopausal group ( (OR) ¼ 0.64, 95 percent (CI): 0.44, 0.93). However, no other statistically significant associations were apparent. We recently reported results on our analysis of reproductive factors, stratified by age and menopausal status (18), which suggest that the well-known protective effect of multiparity attenuates with older age. Moreover, breastfeeding decreased risk among older parous postmenopausal women (data not shown). As shown in table 2, the effect of ever use of HBC on breast cancer risk was evident in premenopausal women (for ever vs. never use: OR ¼ 1.37, 95 percent CI: 1.04, 1.81) but not among postmenopausal women <65 years or 65 years. Also, among premenopausal women, breast cancer risk increased with increasing duration of HBC use (p trend ¼ 0.03) and increasing age at first HBC use (p trend ¼ 0.01). No clear associations between duration of HBC use, age at first HBC use, or time since first HBC use (data not shown) and breast cancer risk were evident among postmenopausal women <65 years or 65 years, although data were sparse. Finally, among premenopausal women, those with current or recent HBC use (5 years) had a more pronounced increase in breast cancer risk (OR ¼ 1.61, 95 percent CI: 1.01, 2.57) than did past users (OR ¼ 1.31, 95 percent CI: 0.98, 1.75) (data not shown in table). As shown in table 3, postmenopausal women <65 years who reported ever using HRT had a slightly increased risk of breast cancer (OR ¼ 1.12, 95 percent CI: 0.88, 1.44), which was more pronounced with increasing months of HRT use (for 61 months of use: OR ¼ 1.41, 95 percent CI: 1.00, 1.99). Premenopausal women who reported ever use of HRT had an elevated of 1.81 (95 percent CI: 1.17, 2.81), with risk increasing to nearly fourfold among longterm users of HRT (for months of use: OR ¼ 3.93, 95 percent CI: 1.43, 10.84; for 61 months: OR ¼ 3.94, 95 percent CI: 1.10, 14.49; p trend < 0.001), although these estimates are based on small numbers of women. Further, risk was also increased more than threefold among premenopausal women who were older when they first began using HRT (OR ¼ 3.38, 95 percent CI: 1.17, 9.79). There were no clear or consistent associations between time since first HRT

4 TABLE 1. Age-adjusted s and s for breast cancer in relation to demographic risk factors among cases and controls by age and menopausal status, Long Island Breast Cancer Study Project, (n ¼ 468) Premenopausal Postmenopausal <65 years Postmenopausal 65 years (n ¼ 500) Age-adjusted (n ¼ 491) (n ¼ 554) Age-adjusted (n ¼ 519) No. % No. % No. % No. % No. % No. % (n ¼ 439) Age-adjusted Race White Black , , , 3.07 Other , , , 3.30 Missing Latina or Hispanic ethnicity No Yes , , , 4.53 Missing Marital status Ever married Never married , , , 2.45 Missing Education High school or less Some college , , , 1.30 College graduate , , , 1.83 Postcollege , , , 1.42 Missing Length of residence (years) < , , , 1.79 Missing Income ($) <20, ,000 49, , , , ,000 89, , , , , , , , 1.39 Missing Shantakumar et al.

5 TABLE 2. Age-adjusted s and s for breast cancer in relation to hormonal birth control use among cases and controls by age and menopausal status, Long Island Breast Cancer Study Project, (n ¼ 468) Premenopausal Postmenopausal <65 years Postmenopausal 65 years (n ¼ 500) Age-adjusted (n ¼ 491) (n ¼ 554) Age-adjusted (n ¼ 519) No. % No. % No. % No. % No. % No. % (n ¼ 439) Age-adjusted Hormonal birth control use Never Ever , , , 1.19 Missing Duration of hormonal birth control use (months) Never < , , , , , , , , , , , , 1.66 Missing p trend Age at first hormonal birth control use (years) Never < , , , , , , , , , , , 1.19 Missing p trend Age, Exogenous Hormones, and Breast Cancer 1191

6 TABLE 3. Age-adjusted s and s for breast cancer in relation to hormone replacement therapy use among cases and controls by age and menopausal status, Long Island Breast Cancer Study Project, (n ¼ 468) Premenopausal Postmenopausal <65 years Postmenopausal 65 years (n ¼ 500) Age-adjusted (n ¼ 491) (n ¼ 554) Age-adjusted (n ¼ 519) No. % No. % No. % No. % No. % No. % (n ¼ 439) Age-adjusted Hormone replacement therapy use Never Ever , , , 1.34 Missing Duration of hormone replacement therapy use (months) Never < , , , , , , , , , , , , 1.78 Missing p trend < Age at first hormone replacement therapy use (years) Never < , , , , , , , , , , , 1.25 Missing p trend < Type of hormone replacement therapy Never Estrogen only , , , 1.37 Estrogen þ progestin , , , 3.07 Mixture of regimens , , , 1.44 Missing Shantakumar et al.

7 Age, Exogenous Hormones, and Breast Cancer 1193 use or recency of HRT use and breast cancer risk among postmenopausal women <65 or 65 years. When specific hormone preparations were considered, there was little suggestion of an increased risk for postmenopausal women <65 years who reported either ever use of estrogen plus progestin or ever use of estrogen only. In contrast, estrogen plus progestin was associated with elevated breast cancer risk in the other two groups of women, although the effect of estrogenplus-progestin therapy was stronger in premenopausal women (OR ¼ 3.51, 95 percent CI: 1.45, 8.49) than in postmenopausal women 65 years (OR ¼ 1.56, 95 percent CI: 0.80, 3.07). Table 4 shows the age and menopausal effects of use of HBC or HRT only, or both, on breast cancer. Postmenopausal women, regardless of whether they were aged less than 65 years or aged 65 years or more, who exclusively used only HBC or HRT were not at increased breast cancer risk, although a slight nonsignificant elevation in risk was observed among those who reported use of both HBC and HRT and were aged less than 65 years at the time of diagnosis (OR ¼ 1.12, 95 percent CI: 0.80, 1.57). Cell sizes were too sparse to adequately evaluate the effects of longterm use of either product alone or jointly (data not shown). For premenopausal women, similar, modest elevations in risk were seen regardless of whether they took only HBC or only HRT, but the was increased to 2.59 (95 percent CI: 1.50, 4.46) among those who reported a history of both HBC and HRT use. Table 5 shows the relation between exogenous estrogen use and breast cancer risk stratified by body mass index. Thinner, postmenopausal women <65 years who ever used HBC had a higher risk(or ¼ 1.37, 95 percent CI: 0.93, 2.03), converse to their heavier counterparts who had a decreased risk (OR ¼ 0.71, 95 percent CI: 0.50, 0.99; p interaction ¼ 0.02). A similar, but nonstatistically significant heterogeneity of effect by body mass index was also observed among postmenopausal women <65 years (p interaction ¼ 0.23). Finally, there was no significant additive interaction between body mass index and exogenous hormone use among the three groups of women (data not shown). DISCUSSION Overall, our results emphasize that the timing of exogenous hormone use influences its subsequent impact on breast cancer risk. Women who used exogenous hormones before menopause had the greatest increase in breast cancer risk, but the harmful effects of these hormones began to attenuate with greater time since menopause. In agreement with the large body of observational data (2), our findings showed that, among postmenopausal women <65 years, ever HRT use was modestly associated with breast cancer risk, with risk increasing with duration of use. However, out of the three groups of women, premenopausal women who used HRT had the highest risk of breast cancer, with greater effects for longer durations of use, and when progestin was added to the HRT formulation. Further, breast cancer risk was higher among premenopausal women who initiated HRT use in middle age (ages years), indicating that TABLE 4. Age-adjusted s and s for breast cancer in relation to combination hormonal birth control and hormone replacement therapy use among cases and controls by age and menopausal status, Long Island Breast Cancer Study Project, Premenopausal Postmenopausal <65 years Postmenopausal 65 years (n ¼ 439) Age-adjusted (n ¼ 519) (n ¼ 554) Age-adjusted (n ¼ 491) (n ¼ 500) Age-adjusted (n ¼ 468) Combination hormonal birth control and hormone replacement therapy use No. % No. % No. % No. % No. % No. % Never Hormonal birth control only , , , 1.27 Hormone replacement therapy only , , , 1.40 Hormonal birth control and hormone replacement therapy , , , 1.55 Missing

8 TABLE 5. Age-adjusted s and s stratified by body mass index at the reference date for the association between exogenous hormone use in relation to breast cancer incidence among premenopausal women, postmenopausal women aged below 65 years, and postmenopausal women aged 65 years or older, Long Island Breast Cancer Study Project, / controls (283/278) Body mass index: <25 kg/m 2 Odds ratio Premenopausal Postmenopausal <65 years Postmenopausal 65 years Body mass index: 25.0 kg/m 2 / controls (184/215) Odds ratio / controls (203/247) Body mass index: <25 kg/m 2 Odds ratio Body mass index: 25.0 kg/m 2 / controls (281/300) Odds ratio / controls (186/194) Body mass index: <25 kg/m 2 Odds ratio Body mass index: 25.0 kg/m 2 / controls (324/232) Odds ratio Hormonal birth control use Never 70/ / / / / / Ever 213/ , / , / , / , / , / , 1.32 Hormone replacement therapy use Never 246/ / / / / / Ever 37/ , / , / , / , / , / , 1.36 Type of hormone replacement therapy Never 246/ / / / / / Estrogen only 2/ , / , / , / , / , / , 1.56 Estrogen þ progestin 16/ , / , / , / , / , / , 4.47 Mixture of regimens 19/ , / , / , / , / , / , 1.12 Combination hormonal birth control and hormone replacement therapy use Never 64/ / / / / / Hormonal birth control only 182/ , / , / , / , / , / , 1.45 Hormone replacement therapy only 6/ , / , / , / , / , / , 1.45 Hormonal birth control and hormone replacement therapy 31/ , / , / , / , / , / , Shantakumar et al.

9 Age, Exogenous Hormones, and Breast Cancer 1195 use initiated during the perimenopausal period confers an increased risk. However, the detrimental effects of HRT use were not sustained as women aged and entered postmenopause; estimates of effect were attenuated to the null value in postmenopausal women 65 years. The observation that premenopausal and older postmenopausal estrogen-plus-progestin users were at greater breast cancer risk than those who used only estrogen therapy is consistent with Key and Pike s hypothesis that adding a progestin to estrogen therapy produces a higher mitotic rate in the breast epithelium than does estrogen alone (19, 20). Out of all the HRT variables that we evaluated, estrogen-plusprogestin therapy was the only factor that increased breast cancer risk in postmenopausal women 65 years. Possibly, the long-term deprivation of ovarian hormones may have increased the sensitivity of mammary tissue to the combination of estrogen plus progestin at older ages after menopause (21 23). However, unexpectedly, we did not find any association between breast cancer and estrogen-plusprogestin therapy among postmenopausal women <65 years, which contrasts with findings from the Women s Health Initiative and other observational studies that reported higher breast cancer risks associated with estrogen-plusprogestin use (2, 4, 24 35). It was reassuring that a history of ever using HBC or longterm use of HBC (61 months) was not associated with increased breast cancer risk among postmenopausal women <65 years or 65 years. Older women (65 years) were of special concern, as they were exposed to not only firstgeneration HBC that contained the highest doses of estradiol and progestin but also the opportunity for long-term HBC use (36 38). However, our results suggest that previous HBC use is not associated with postmenopausal breast cancer. We had sufficient numbers of postmenopausal women to separate them into younger (<65 years) and older (65 years) age categories, and we were able to observe if the diminished effect of HBC use in the younger postmenopausal group was also sustained in the older group. In contrast, most other studies of HBC use and postmenopausal breast cancer (11, 37, 39 45) have considered only women aged less than 65 years. In our analysis, combination HBC and HRT use among premenopausal women was associated with a more pronounced breast cancer risk than was sole use of either hormone. However, the detrimental effect of dual use in our study was evident only prior to menopause; we did not find any substantial elevation in risk for younger or older postmenopausal women. We attempted to evaluate the influence of long-term HBC and HRT use, but our numbers were too sparse to produce meaningful estimates. Although data from other population groups indicate that women who used HBC previously are more likely to use HRT subsequently than are women who did not (6, 46, 47), only a few other studies have explored the combined effects of HBC and HRT use on breast cancer risk (6 13, 48), with one study reporting a positive interaction between HBC and HRT use (13). Our findings indicate that the later years of premenopause, namely, the perimenopause, could signify an important induction window for breast cancer, as it is a transition period when women with functioning ovaries are aging into a group with a higher incidence of breast cancer. Consequently, breast tissue in perimenopausal women that has been previously exposed to HBC may be more susceptible to subsequent carcinogenic insult from HRT. According to our results, women exposed to HRT in the perimenopause develop breast cancer soon after use, leading to an excess number of cases diagnosed before menopause. Thinner, postmenopausal women <65 years who had ever used HRT were at greater breast cancer risk than were their overweight or obese counterparts, which is consistent with findings from a recent pooled analysis (2). Compared with thinner women, postmenopausal women with a high body mass index are already estrogen swamped and may be experiencing a threshold effect, such that any subsequent exogenous hormone exposure may not further increase their breast cancer risk. The LIBCSP has the advantage of being large and population based, which increases the generalizability of study results. Furthermore, the LIBCSP did not have age restrictions, with participants ranging in age from 20 to 98 years. Consequently, we had ample numbers of premenopausal women, who are generally excluded from HRT analyses, and older postmenopausal women, an often underrepresented demographic group in case-control studies of breast cancer. We also had large numbers of baby boomers who were either perimenopausal or postmenopausal at the time of the study and had the opportunity to be dually exposed to HBC and HRT during their lifetime. However, the majority of LIBCSP participants were White, and thus our results may not be generalizable to non-white women. Assigning postmenopausal status to current HRT users is challenging, as it is difficult to ascertain if and when ovarian function ceases. For current HRT users, we utilized a conservative algorithm that, regardless of self-reported menopausal status, assigned postmenopausal status on the basis of reference age compared against the 90th percentile of natural menopausal age for smokers and nonsmokers (refer to Materials and Methods). Pike et al. (49) argue that this approach is problematic as it assigns a relatively late menopausal age to HRT users. Earlier age at menopause is associated with longer durations of HRT use, and this imputation may have resulted in an underestimation of the effect for duration of HRT use on breast cancer risk among postmenopausal women. Recall of exogenous hormone use may have been inaccurate, especially if use was intermittent and formulations were varied. However, studies have reported high agreement between interview data specific to hormone use and available physician data (50, 51), confirmed questionnaire information being concordant with medical record information on duration of use (52, 53), brand (54), dosage (54), treatment schedule (54), and time of initiation of medication (54), and shown no differential misclassification of hormone use history by case-control status (52, 55). It is possible, though, that the older postmenopausal women in this study had more difficulty with accurate recall of exogenous hormone use, leading to nondifferential misclassification in that group and, in turn, biasing results to the null. Nevertheless, studies that compared congruence of self-reported use of common medications with pharmacy records found that older women have

10 1196 Shantakumar et al. high sensitivity and high specificity when recalling ever use (56, 57). To further enhance recall in our study, a color chart of most noncontraceptive estrogens and progestins marketed in the United States up through the time of the study ( ) was also shown to each participant. Among premenopausal HRT users in our study population, 24 of 68 cases and 21 of 37 controls took HRT before the age of 40 years. HRT use in younger, premenopausal women is not uncommon and may be prescribed for the management of gynecologic issues such as primary or secondary amenorrhea (58) or endometriosis (59). However, as the etiology of amenorrhea is varied, only certain causes are managed with HRT. For example, women with polycystic ovarian syndrome who have chronic anovulation and hyperandrogenism may be prescribed cyclic progestin regimens to help maintain a normal endometrium, while patients with hypothalamic amenorrhea caused by excessive weight loss, exercise, or stress may be prescribed estrogen-plusprogestin therapy to normalize menstruation and to partially protect against bone loss (58). Finally, the older premenopausal women in our study who used HRT may have been undergoing the menopausal transition and using HRT to relieve vasomotor (e.g., sweating, hot flashes) or urogenital (e.g., vaginal dryness) symptoms (60). The increased breast cancer risk among current users of HBC and HRT has been partly attributed by some investigators to increased medical surveillance (61). Women who use exogenous hormones are also more likely to engage in healthy behaviors such as mammography (62). Moreover, our estimates of effect did not appear to be influenced by surveillance bias, as s adjusted for mammography use were similar to unadjusted estimates. In conclusion, our study shows that premenopausal women who ever used HRT or the combination of HBC and HRT are at increased breast cancer risk, with effects dissipating over time. Use of HRT in the perimenopause is also associated with a marked increase in risk. Although there are some demonstrated benefits of HRT use for severe menopausal symptoms, our results highlight that there are previously unappreciated risks of HRT use in pre- and perimenopausal women. Our findings are inconsistent with current guidelines in the United States (63), United Kingdom (64), and European Union (65) that support HRT use for the relief of perimenopausal symptoms, albeit for the shortest duration possible. It appears that the perimenopause might be an important induction window as breast tissue might be more susceptible to the carcinogenic insult (5) of exogenous sex steroids. The increased risk seen with dual HBC and HRT use is also of public health significance, as large numbers of women have by now been exposed to HBC in their reproductive years as well as HRT at older ages, but the effect of joint use has been mostly ignored. Nevertheless, this study demonstrates that the effect of past exogenous hormone use does not persist into older ages, which is reassuring since the incidence of breast cancer is increased later in life. Unfortunately, our data were too sparse to elucidate the effects of long-term current HRT use among women aged 65 years or more. Future large-scale studies are needed to support our findings and to further elucidate how HBC use and HRT use impact plasma sex hormone concentrations in premenopausal, perimenopausal, and older postmenopausal women. ACKNOWLEDGMENTS This work was supported in part by grants from the National Cancer Institute and the National Institute of Environmental Health Sciences (grants UO1CA/ES66572, CA58233, T32CA09330, P30ES09089, and P30ES10126). For their valuable contributions to the Long Island Breast Cancer Study Project, the authors thank the following: members of the Long Island Breast Cancer Network; the 31 participating institutions on Long Island and in New York City, New York; their National Institutes of Health collaborators, Dr. Gwen Collman, National Institute of Environmental Health Sciences; Dr. G. Iris Obrams, formerly of the National Cancer Institute; members of the External Advisory Committee to the population-based, case-control study: Dr. Leslie Bernstein (Committee chair), Gerald Akland, Barbara Balaban, Dr. Blake Cady, Dr. Dale Sandler, Dr. Roy Shore, and Dr. Gerald Wogan; other collaborators who assisted with various aspects of data collection efforts including Dr. Mary Wolff, Dr. Geoff Kabat, Dr. Steve Stellman, Dr. Maureen Hatch, Gail Garbowski, Dr. H. Leon Bradlow, Martin Trent, Dr. Ruby Senie, Dr. Carla Maffeo, Pat Montalvan, Dr. Gertrud Berkowitz, Dr. Margaret Kemeny, Dr. Mark Citron, Dr. Freya Schnabel, Dr. Allen Schuss, Dr. Steven Hajdu, and Dr. Vincent Vinciguerra; and for their careful review and comments on the manuscript: Dr. Robert C. Millikan and Dr. Andrew F. Olshan. Conflict of interest: none declared. REFERENCES 1. Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on women with breast cancer and women without breast cancer from 54 epidemiological studies. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 1996;347: Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 1997;350: Hulley S, Furberg C, Barrett-Connor E, et al. Noncardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/progestin Replacement Study follow-up (HERS II). JAMA 2002;288: Rossouw JE, Anderson GL, Prentice RL, et al. Risk and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women s Health Initiative randomized control trial. JAMA 2002;288: Cogliano V, Grosse Y, Baan R, et al. Carcinogenicity of combined oestrogen-progestagen contraceptives and menopausal treatment. Lancet Oncol 2005;6: Norman S, Berlin JA, Weber AL, et al. Combined effect of oral contraceptive use and hormone replacement therapy on breast

11 Age, Exogenous Hormones, and Breast Cancer 1197 cancer risk in postmenopausal women. Cancer Causes Control 2003;14: Mills P, Beeson WL, Phillips RL, et al. Prospective study of exogenous hormone use and breast cancer in Seventh-day Adventists. Cancer 1989;64: Olsson H, Bladstrom A, Ingvar C, et al. A population-based cohort study of HRT use and breast cancer in southern Sweden. Br J Cancer 2001;85: Stanford J, Brinton LA, Hoover RN. Oral contraceptives and breast cancer: results from an expanded case-control study. Br J Cancer 1989;60: Stanford J, Weiss NS, Voigt LF, et al. Combined estrogen and progestin hormone replacement therapy in relation to risk of breast cancer in middle-aged women. JAMA 1995;274: Schuurman A, van den Brandt PA, Goldbohm RA. Exogenous hormone use and the risk of postmenopausal breast cancer: results from the Netherlands Cohort Study. Cancer Causes Control 1995;6: Ursin G, Tseng CC, Paganini-Hill A, et al. Does menopausal hormone replacement therapy interact with known factors to increase risk of breast cancer? J Clin Oncol 2002;20: Brinton L, Brogan DR, Coates RJ, et al. Breast cancer risk among women under 55 years of age by joint effects of usage of oral contraceptives and hormone replacement therapy. Menopause 1998;5: Gammon M, Neugut AI, Santella RM, et al. The Long Island Breast Cancer Study Project: description of a multiinstitutional collaboration to identify environmental risk factors for breast cancer. Breast Cancer Res Treat 2002;3: Waksberg J. Sampling methods for random digit dialing. J Am Stat Assoc 1978;73: Hosmer D, Lemeshow S. Applied logistic regression. New York, NY: John Wiley & Sons, Inc, Rothman K, Greenland S. Modern epidemiology. Philadelphia, PA: Lippincott-Raven, Shantakumar S, Terry MB, Teitelbaum SL, et al. Reproductive factors and breast cancer risk among older women. Breast Cancer Res Treat (in press). 19. Hesch RD, Kenemans P. Hormonal prevention of breast cancer: proposal for a change in paradigm. Br J Obstet Gynaecol 1999;106: Key TJ, Pike MC. The role of oestrogens and progestagens in the epidemiology and prevention of breast cancer. Eur J Cancer Clin Oncol 1988;24: Manjer J, Johansson R, Berglund G, et al. Postmenopausal breast cancer risk in relation to sex steroid hormones, prolactin and SHBG (Sweden). Cancer Causes Control 2003;14: Raafat A, Hofseth LJ, Li S, et al. A mouse model to study the effects of hormone replacement therapy on normal mammary gland during menopause: enhanced proliferative response to estrogen in late postmenopausal mice. Endocrinology 1999; 140: Pasqualini J, Paris J, Sitruk-Ware R, et al. Progestins and breast cancer. J Steroid Biochem Mol Biol 1998;65: Kirsh V, Kreiger N. Estrogen and estrogen-progestin replacement therapy and risk of postmenopausal breast cancer in Canada. Cancer Causes Control 2002;13: Jernstrom H, Bendahl PO, Lidfeldt J, et al. A prospective study of different types of hormone replacement therapy use and the risk of subsequent breast cancer: the Women s Health in the Lund Area (WHILA) Study (Sweden). Cancer Causes Control 2003;14: Porch J, Lee IM, Cook NR, et al. Estrogen-progestin replacement therapy and breast cancer risk: the Women s Health Study (United States). Cancer Causes Control 2002;13: Magnusson C, Baron JA, Correia N, et al. Breast-cancer risk following long-term oestrogen- and oestrogen-progestinreplacement therapy. Int J Cancer 1999;81: Schairer C, Lubin J, Troisi R, et al. Menopausal estrogen and estrogen-progestin replacement therapy and breast cancer risk. JAMA 2000;283: Beral V, Million Women Study Collaborators. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet 2003;362: Persson I, Yuen J, Bergkvist L, et al. Cancer incidence and mortality in women receiving estrogen and estrogen-progestin replacement therapy long-term follow-up of a Swedish cohort. Int J Cancer 1996;67: Olsson H, Ingvar C, Bladstrom A. Hormone replacement therapy containing progestins and given continuously increases breast carcinoma risk in Sweden. Cancer 2003;97: Weiss L, Burkman RT, Cushing-Haugen KL, et al. Hormone replacement therapy regimens and breast cancer risk. Obstet Gynecol 2002;100: Ross R, Paganini-Hill A, Wan PC, et al. Effect of hormone replacement therapy on breast cancer risk: estrogen versus estrogen plus progestin. J Natl Cancer Inst 2000;92: Newcomb P, Titus-Ernstoff L, Egan KM, et al. Postmenopausal estrogen and progestin use in relation to breast cancer risk. Cancer Epidemiol Biomarkers Prev 2002;11: Chlebowski R, Hendrix SL, Langer RD, et al. Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women s Health Initiative Randomized Trial. JAMA 2003;289: Hormonal contraception and post-menopausal hormonal therapy. Lyon, France: International Agency for Research on Cancer, Marchbanks P, McDonald JA, Wilson HG, et al. Oral contraceptives and the risk of breast cancer. N Engl J Med 2002;346: Kelsey J. Breast cancer epidemiology: summary and future directions. Epidemiol Rev 1993;15: Rookus M, van Leeuwen FE. Oral contraceptives and risk of breast cancer in women aged years. Netherlands Oral Contraceptives and Breast Cancer Study Group. Lancet 1994;344: Rosenberg L, Palmer JR, Clarke EA, et al. A case-control study of the risk of breast cancer in relation to oral contraceptive use. Am J Epidemiol 1992;136: Rossing M, Stanford JL, Weiss NS, et al. Oral contraceptive use and risk of breast cancer in middle-aged women. Am J Epidemiol 1996;144: Magnusson C, Persson IR, Baron JA, et al. The role of reproductive factors and use of oral contraceptives in the aetiology of breast cancer in women aged 50 to 74 years. Int J Cancer 1999;80: Van Hoften C, Burger H, Peeters PH, et al. Long-term oral contraceptive use increases breast cancer risk in women over 55 years of age: the DOM cohort. Int J Cancer 2000;87: Rosenberg L, Palmer JR, Rao RS, et al. Case-control study of oral contraceptive use and risk of breast cancer. Am J Epidemiol 1996;143: Ewertz M. Oral contraceptives and breast cancer risk in Denmark. Eur J Cancer 1992;28A: Petitti DB, Perlman JA. Re: A prospective cohort study of postmenopausal hormone use and risk of breast cancer in US women. (Letter). Am J Epidemiol 1988;127:883.

12 1198 Shantakumar et al. 47. Hammar M, Brynhildsen J, Dabrosin L, et al. Hormone replacement therapy and previous use of oral contraceptives among Swedish women. Maturitas 1996;25: Dumeaux V, Fournier A, Lund E, et al. Previous oral contraceptive use and breast cancer risk according to hormone replacement therapy use among postmenopausal women. Cancer Causes Control 2005;16: Pike M, Ross RK, Spicer DV. Problems involved in including women with simple hysterectomy in epidemiologic studies measuring the effects of hormone replacement therapy on breast cancer risk. Am J Epidemiol 1998;147: Merlo J, Berglund G, Wirfalt E, et al. Self-administered questionnaire compared with a personal diary for assessment of current use of hormone therapy: an analysis of 16,060 women. Am J Epidemiol 2000;152: Hunter D, Manson JE, Colditz GA, et al. Reproducibility of oral contraceptive histories and validity of hormone composition reported in a cohort of US women. Contraception 1997;56: Goodman M, Nomura AM, Wilkens LR, et al. Agreement between interview information and physician records on history of menopausal estrogen use. Am J Epidemiol 1990;131: Paganini-Hill A, Ross RK. Reliability of recall of drug usage and other health-related information. Am J Epidemiol 1982; 116: Bergkvist L, Adami HO, Persson I, et al. The risk of breast cancer after estrogen and estrogen-progestin replacement. N Engl J Med 1989;321: Brinton L, Hoover RN, Szklo M, et al. Menopausal estrogen use and risk of breast cancer. Cancer 1981;47: Boudreau D, Daling JR, Malone KE, et al. A validation study of patient interview data and pharmacy records for antihypertensive, statin, and antidepressant medication use among older women. Am J Epidemiol 2004;159: Sjahid S, van der Linden PD, Stricker BH. Agreement between the pharmacy medication history and patient interview for cardiovascular drugs: the Rotterdam elderly study. Br J Clin Pharmacol 1988;45: Master-Hunter T, Heiman DL. Amenorrhea: evaluation and treatment. Am Fam Physician 2006;73: Pierce S, Gazvani MR, Farquharson RG. Long-term use of gonadotropin-releasing hormone analogs and hormone replacement therapy in the management of endometriosis: a randomized trial with a 6-year follow-up. Fertil Steril 2000;74: Avis N, Stellato R, Crawford S, et al. Is there a menopausal syndrome? Menopausal status and symptoms across racial/ ethnic groups. Soc Sci Med 2001;52: Shapiro S. Bias in the evaluation of low-magnitude associations: an empirical perspective. Am J Epidemiol 2000;151: Sturgeon S, Schairer C, Brinton LA, et al. Evidence of a healthy estrogen user survivor effect. Epidemiology 1995; 6: US Preventive Services Task Force. Postmenopausal hormone replacement therapy for primary prevention of chronic conditions: recommendations and rationale. Ann Intern Med 2002;137: Committee on Safety of Medicines. Review of the evidence on the long-term safety of HRT. Curr Probl Pharmacovigilance 2004;30: The European Agency for the Evaluation of Medicinal Products. Guideline for hormone replacement therapy. London, United Kingdom: European Medicines Agency, January 20, ( pdf). Accessed July 7, 2005.