Breast Cancer Risk Reduction, Version 2.2015

Familial/Genetic Risk Factors

The first step in primary assessment of breast cancer risk is a broad and flexible evaluation of the patient's personal and family history, primarily with respect to breast and/or ovarian cancers.^6,7

Genetic predispositions and the associated genes that confer a high risk for breast cancer include hereditary breast and ovarian cancer (BRCA1/2),^8,9 Li-Fraumeni syndrome (TP53),¹⁰ Peutz-Jeghers syndrome (STK11),¹¹ Cowden syndrome (PTEN),^12,13 and hereditary diffuse gastric cancer (CDH1).¹⁴ If the patient has a known genetic predisposition for breast cancer, such as mutations in BRCA1/2, TP53, PTEN, or other gene mutations associated with breast cancer risk, they must be counseled for risk reduction options.

If the familial/genetic factors are unknown, a thorough evaluation must be performed. The magnitude of risk increases with the number of affected relatives in the family, closeness of the relationship, and age at which the affected relative was diagnosed^15–17; a younger age at diagnosis of a first- or second-degree relative means that a genetic component is more likely to be present. The maternal and paternal sides of the family should be considered independently for familial patterns of cancer (see NCCN Clinical Practice Guidelines in Oncology for Genetic/Familial High-Risk Assessment: Breast and Ovarian [available online, at NCCN.org]).

Hereditary cancers are often characterized by gene mutations associated with a high probability of cancer development (ie, a high-penetrance genotype), vertical transmission through either the mother or father, and an association with other tumor types.^18,19 They often have an early age of onset and exhibit an autosomal dominant inheritance pattern (ie, occur when the individual has a germline mutation in only 1 copy of a gene).

Familial cancers share some but not all features of hereditary cancers. For example, although familial breast cancers occur in a given family more frequently than in the general population, they generally do not exhibit the inheritance patterns or early onset age consistent with hereditary cancers. Familial cancers may be associated with chance clustering of sporadic cancer cases within families, genetic variation in lower-penetrance genes, a shared environment, or combinations of these factors.^20–23

If an individual or a close family member meets one or more of the criteria described previously (and also in the NCCN Guidelines for Genetic/Familial High-Risk Assessment: Breast and Ovarian; to view the most recent version of these guidelines, visit NCCN.org), that individual may be at increased risk for familial/hereditary breast cancer, and referral for formal genetic assessment/counseling is recommended.

A cancer geneticist should be involved in determining whether the patient has a lifetime risk for breast cancer greater than 20% based on models dependent on family history (eg, Claus et al,²⁴ Tyrer et al,²⁵ and others^26–28). The Claus tables may be useful in providing breast cancer risk estimates for white women without a known cancer-associated gene mutation who have 1 or 2 first- or second-degree female relatives with breast²⁴ and ovarian cancer.²⁹

BRCAPRO³⁰ and Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA)³¹ are commonly used to estimate the risk of a BRCA mutation. Strong genetic association between breast and ovarian cancers has been shown in some families by linkage analyses. Based on a risk assessment using one of more of these models, women with a BRCA1/2, TP53, or PTEN gene mutation or a pedigree strongly suggestive of a genetic predisposition to breast cancer, may be identified. The NCCN Guidelines for Genetic/Familial High-Risk Assessment: Breast and Ovarian also describe management strategies for women with a known or suspected BRCA1/2, TP53, or PTEN mutation or a pedigree strongly suggestive of genetic predisposition to breast cancer (to view the most recent version of these guidelines, visit NCCN.org).

Other Elements of Risk

For women not considered at risk for familial/hereditary breast cancer, an evaluation of other elements of risk that contribute to increased breast cancer risk is recommended. These include demographic factors such as female gender, age, and ethnicity/race. Women of Ashkenazi Jewish decent have a reported increased incidence of BRCA1/2 mutation.³² Demographic factors also include reproductive history. Strong risk factors linked to reproductive history include nulliparity, prolonged interval between menarche and age at first live birth (eg, early menarche or late age of first live birth), and current use of menopausal hormone therapy (HT).^33–37

High body mass index (BMI) is an independent risk factor for breast cancer, especially in white women. Several studies have established the association of high BMI and adult weight gain with increased risk for breast cancer in postmenopausal women.^38–46 This increase in risk has been attributed to increase in circulating endogenous estrogen levels from fat tissue.^44–46 In addition, the association between BMI and risk for postmenopausal breast cancer is stronger for hormone-positive tumors.^40–43

Lifestyle factors such as current or prior HT,³⁷ alcohol consumption,^47–49 and, to a lesser extent, smoking^50,51 also contribute to the risk of developing breast cancer. Other factors to consider are number of breast biopsies, especially if they showed flat epithelial atypia, AH, or LCIS.

The risk for breast cancer associated with flat epithelial atypia is similar to that of benign proliferative disease without atypia. The data are not as robust with respect to the degree of risk or the benefits of risk reduction therapy in this population. AH includes both atypical ductal hyperplasia (ADH) and atypical lobular hyperplasia. AHs, especially multifocal lesions, confer a substantial increase in the risk for subsequent breast cancer.^52–54 Women with LCIS are at substantially increased risk for breast cancer.

Prior thoracic irradiation encompassing the chest/breast area received before ae 30 years of age (eg, for the treatment of Hodgkin lymphoma) is a significant risk factor for the development of breast cancer. In the Late Effects Study Group trial, the overall risk for breast cancer associated with thoracic irradiation at a young age was found to be 56.7-fold (55.5-fold for female patients) greater than the risk for breast cancer in the general population.⁵⁵ In that study, the relative risk (RR) according to follow-up interval was 0 at 5 to 9 years; 71.3 at 10 to 14 years; 90.8 at 15 to 19 years; 50.9 at 20 to 24 years; 41.2 at 25 to 29 years; and 24.5 at more than 29 years.⁵⁵ Results from a case-control study of women treated at a young age (≤30 years) for Hodgkin lymphoma with thoracic radiation indicated that the estimated, cumulative, absolute risk for breast cancer at 55 years of age was 29.0% (95% CI, 20.2%–40.1%) for a woman treated at 25 years of age with 40 Gy of radiation and no alkylating agents.⁵⁶ Women with a history of treatment with thoracic radiation for Hodgkin lymphoma are at a high risk for breast cancer based on radiation exposure alone.^55–60

Change in breast density has been suggested as a risk factor for breast cancer.⁶¹ Dense breast tissue as measured by mammography is increasingly recognized as an important risk factor for breast cancer.^62–66 For example, a report of a large case-cohort study of women 35 years and older with no history of breast cancer who underwent mammographic screening, first at baseline and then at an average of 6 years later, suggested that longitudinal changes in breast density are associated with changes in breast cancer risk.⁶⁵

Cancer Risk Assessment

Women 35 years of age or older without a BRCA1/2, TP53, or PTEN mutation; a strong family history of breast cancer; a history of thoracic radiation before age 30 years; or a history of LCIS should have their risk for breast cancer estimated according to the modified Gail model.^67–69 The modified Gail model is a computer-based, multivariate, logistic regression model that uses age, race, age at menarche, age at first live birth or nulliparity, number of first-degree relatives with breast cancer, number of previous breast biopsies, and histology of the breast biopsies to produce actuarial estimates of future breast cancer risk.^67,68,70,71 The criteria used to determine risk by the modified Gail model are described in Table 1 (available online, in these guidelines, at NCCN.org). The Gail model, as modified by the National Surgical Adjuvant Breast and Bowel Project (NSABP) investigators, is available on the NCI Web site (http://www.cancer.gov/bcrisktool/Default.aspx).

The risk threshold required for a woman to consider the use of risk reduction strategies must depend on an evaluation of the efficacy, morbidity, and expense of the proposed intervention. As a reasonable discriminating threshold, the NCCN Breast Cancer Risk Reduction Panel has adopted the 1.7% or greater 5-year actuarial breast cancer risk as defined by the modified Gail model, which was used to identify women eligible for the NSABP Breast Cancer Prevention Trial (BCPT)^72,73 and the Study of Tamoxifen and Raloxifene (STAR) trial.^74,75

The Gail model was updated using combined data from the Women's Contraceptive and Reproductive Experiences (CARE) study and the SEER database, as well as causes of death from the National Center for Health Statistics, to provide a more accurate determination of risk for African-American women.⁷⁶ The model was also updated using data from the Asian American Breast Cancer Study (AABCS) and the SEER database to provide a more accurate risk assessment for Asian and Pacific Islander women in the United States.⁷⁷ Application of the Gail model to recent immigrants from Japan or China may overestimate the risk for breast cancer.⁷⁷ Although use of the Gail model can overestimate the risk for some women, in others, notably those with AH, it can underestimate the risk, making them seem ineligible for risk reduction therapy.

The Gail model is not an appropriate breast cancer risk assessment tool for women who received thoracic radiation to treat Hodgkin lymphoma (eg, mantle radiation) or those with LCIS.⁷⁸ In addition to considering a woman's risk of a BRCA mutation, the Tyrer-Cuzick model also estimates her risk of developing breast cancer using not only family history but also epidemiologic variables including a personal history of AH or LCIS. Women with AH or a history of LCIS are also at substantially increased risk for invasive breast cancer in both the affected and contralateral breast.^{52–54,79,80} In an analysis of the Mayo Clinic cohort of more than 300 women with AH, the Gail model underestimated breast cancer risk for women with AH,⁷⁸ whereas the Tyrer-Cuzick model overestimated this risk.⁸¹ Breast density is not included in any of the commonly used risk assessment models/tools.²⁷

Women with a life expectancy of 10 years or more and no diagnosis or history of breast cancer who are considered to be at increased risk for breast cancer based on any of the previously mentioned assessments should receive counseling that is tailored to the individual to decrease breast cancer risk (eg, risk reduction surgery in BRCA1/2 mutation carriers; therapy with risk reduction agents in those without a contraindication to these agents) (see “Components of Risk Reduction Counseling,” page 905), and should undergo breast screening as detailed in the NCCN Guidelines for Breast Cancer Screening and Diagnosis (to view the most recent version of these guidelines, visit NCCN.org).

If life expectancy is less than 10 years, risk reduction therapy or screening probably provides minimal if any benefit (see NCCN Guidelines for Breast Cancer Screening and Diagnosis and NCCN Guidelines for Breast Cancer, available at NCCN.org).

Women with a diagnosis of DCIS should be managed according to recommendations outlined in the NCCN Guidelines for Breast Cancer (to view the most recent version of these guidelines, visit NCCN.org).

Lifestyle Modifications

Evidence from immigration studies indicate that in addition to family history and genetics, environmental factors play a significant role. Lifestyle modifications, such as those related to diet, body weight, exercise, and alcohol consumption, are some of the modifiable breast cancer risk factors. Although no clear evidence shows that specific dietary components can effectively reduce breast cancer risk, weight gain and obesity in adulthood are risk factors for the development of postmenopausal breast cancer.^46,82,83 Alcohol consumption, even at moderate levels, increases breast cancer risk.^48,83–86 Patients should be encouraged to maintain a healthy lifestyle and remain up-to-date with recommendations for screening and surveillance (see “Counseling Regarding Lifestyle Modifications,” page 907).

Risk Reduction Surgery

Bilateral Total Mastectomy: The lifetime risk for breast cancer in BRCA1/2 mutation carriers has been estimated to be 56% to 84%.^87–89 Retrospective analyses with median follow-up periods of 13 to 14 years have indicated that bilateral risk-reducing mastectomy (RRM) decreased the risk of developing breast cancer by at least 90% in women at moderate and high risk and in known BRCA1/2 mutation carriers.^90,91 An analysis of results from the study by Hartmann et al⁹⁰ determined that to prevent 1 case of breast cancer in women at high risk, the number of women needed to be treated with RRM was equal to 6.⁹² Results from smaller prospective studies with shorter follow-up periods have provided support for concluding that RRM provides a high degree of protection against breast cancer in women with a BRCA1/2 mutation.^93,94 A recent meta-analysis of 4 prospective studies (2635 patients) has shown a significant risk reduction of breast cancer incidence with bilateral RRM in BRCA1/2 mutation carriers (hazard ratio [HR], 0.07; 95 % CI, 0.01–0.44; P=.004).⁹⁵

The NCCN Breast Cancer Risk Reduction Panel supports the use of RRM in carefully selected women at high risk for breast cancer who desire this intervention (eg, women with a BRCA1/2, TP53, PTEN, CDH1, or STK11 mutation or, possibly, women with a history of LCIS). Although the consensus of the panel is that consideration of RRM is an option for a woman with LCIS without additional risk factors, it is not a recommended approach for most of these women. No data are available regarding RRM in women with prior mantle radiation exposure. The value of RRM in women with deleterious mutations in other genes associated with a high risk for breast cancer (based on large epidemiologic studies) in the absence of a compelling family history of breast cancer is unknown.

Women considering RRM should first have appropriate multidisciplinary consultations and a clinical breast examination and bilateral mammogram if not performed within the past 6 months. If results are normal, women who choose RRM may undergo the procedure with or without immediate breast reconstruction. Bilateral mastectomy performed for risk reduction should involve removal of all breast tissue (ie, a total mastectomy). Axillary node assessment has limited utility at the time of RRM. Women undergoing RRM do not require an axillary lymph node dissection unless breast cancer is identified on pathologic evaluation of the mastectomy specimen.⁹⁶

Some patients may be at risk for an occult primary tumor, such as those with abnormal imaging findings on either mammogram or breast MRI who do not undergo biopsy, and those with familial history who have not had a breast MRI before surgery. In these patients, a sentinel lymph node biopsy may be performed to stage the axilla for an occult cancer during the RRM, and a secondary axillary lymph node dissection could be avoided if an occult invasive cancer is discovered. This procedure has not been found to increase the risk for lymphedema.⁹⁷

After RRM, women who carry a BRCA1/2 mutation should be monitored according to the NCCN Guidelines for Genetic/Familial High-Risk Assessment: Breast and Ovarian (to view the most recent version of these guidelines, visit NCCN.org). Women found to have invasive breast cancer or DCIS at RRM should be treated according to the NCCN Guidelines for Breast Cancer (to view the most recent version of these guidelines, visit NCCN.org). All other women should be followed up with routine health maintenance after RRM. Most health maintenance recommendations are not related to the breast. For monitoring breast health, women should continue with annual examinations of the chest/reconstructed breast as there is still a small risk of developing breast cancer. Mammograms are not recommended in this situation.

Bilateral Salpingo-oophorectomy: Women with a BRCA1/2 mutation are at increased risk for both breast and ovarian cancers (including fallopian tube cancer). Although the risk for ovarian cancer is lower than the risk for breast cancer in a BRCA1/2 mutation carrier (eg, estimated lifetime risks of 36%–46% and 10%–27% in BRCA1 and BRCA2 mutation carriers, respectively^88,98–101), the absence of reliable methods of early detection and the poor prognosis associated with advanced ovarian cancer have lent support for the performance of bilateral risk-reducing salpingo-oophorectomy (RRSO) after completion of childbearing in these women. In the studies by Rebbeck et al,¹⁰² the mean age at diagnosis of ovarian cancer was 50.8 years for BRCA1/2 carriers.

The effectiveness of RRSO in reducing the risk for ovarian cancer in carriers of a BRCA1/2 mutation has been shown in several studies. For example, results of a meta-analysis involving 10 studies of BRCA1/2 mutation carriers showed an approximately 80% reduction in the risk for ovarian or fallopian cancer after RRSO.¹⁰³ However, a 1.0% to 4.3% residual risk for a primary peritoneal carcinoma has been reported in some studies.^102–107

RRSO is also reported to reduce the risk for breast cancer in carriers of a BRCA1/2 mutation by approximately 50%.^{102,103,107,108} In the case-control international study, a 56% (odds ratio [OR], 0.44; 95% CI, 0.29–0.66) and a 46% (OR, 0.57; 95% CI, 0.28–1.15) breast cancer risk reduction were reported after RRSO in carriers of a BRCA1/2 mutation, respectively.¹⁰⁸ Hazard ratios [HRs] of 0.47 (95% CI, 0.29–0.77)¹⁰² and 0.30 (95% CI, 0.11–0.84)¹⁰⁶ were reported in 2 other studies comparing breast cancer risk in women with a BRCA1/2 mutation who had undergone RRSO versus carriers of these mutations who opted for surveillance only. These studies are further supported by a meta-analysis that found similar reductions in breast cancer risk of approximately 50% for BRCA1/2 mutation carriers after RRSO.¹⁰³ Results of a prospective cohort study suggest that RRSO may be associated with a greater reduction in breast cancer risk for BRCA1 mutation carriers compared with BRCA2 mutation carriers.¹⁰⁹

Reductions in breast cancer risk for carriers of a BRCA1/2 mutation after RRSO may be associated with decreased hormonal exposure after surgical removal of the ovaries. Greater reductions in breast cancer risk were seen in women with a BRCA1 mutation who had an RRSO at age 40 years or younger (OR, 0.36; 95% CI, 0.20–0.64) relative to BRCA1 carriers aged 41 to 50 years who underwent this procedure (OR, 0.50; 95% CI, 0.27–0.92).¹⁰⁸ Nonsignificant risk reduction of developing breast cancer was found for women aged 51 years or older, although only a small number of women were included in this group.¹⁰⁸ However, results from Rebbeck et al¹⁰⁷ also suggest that RRSO after 50 years of age is not associated with a substantial decrease in breast cancer risk.

Although data are limited regarding an optimal age for RRSO, the Monte Carlo simulation model provides estimates of the survival impact of breast and ovarian risk reduction strategies (eg, mammographic/MRI breast screening; risk reduction surgery) in women who are carriers of BRCA1/2 mutations according to the type of BRCA mutation present, the specific risk reduction intervention, and the age of the women at the time of the intervention.¹¹⁰ Survival estimates generated from this model can facilitate shared decision-making regarding choice of a risk reduction approach (see Table 2; available online, in these guidelines, at NCCN.org).

A prospective multicenter study reported the benefit of risk-reducing procedures for women with strong genetic predispositions for breast cancer.¹¹¹ The study involved 2482 women diagnosed with BRCA1/2 gene mutations, almost half of whom chose either RRSO or RRM. During 3 years of follow-up, no cases of breast cancer occurred in the women who opted for RRM. During the same period, 7% of the women who used other approaches had a breast cancer diagnosis. In BRCA2 mutation carriers, no cases of ovarian cancer occurred after salpingo-oophorectomy over a 6-year follow-up period, whereas 3% of those who did not undergo the same surgery were diagnosed with ovarian cancer. None of the women who underwent RRM developed breast cancer. RRSO was associated with a reduction in overall mortality (HR, 0.40), breast cancer–specific mortality (HR, 0.44), and ovarian cancer–specific mortality (HR, 0.21). Among women who underwent RRSO, only 1.1% developed ovarian cancer. Another large prospective trial in 5783 women with BRCA1/2 gene mutations reported an 80% reduction in the risk for ovarian, fallopian tube, or peritoneal cancer with oophorectomy (HR, 0.20; 95% CI, 0.13–0.30; P<.001).¹¹² Subsequently, a meta-analysis of 3 prospective studies^109,111,112 found a significant decrease in ovarian cancer risk after RRSO (RR, 0.19; 95% CI, 0.13–0.27).¹¹³ The NCCN Breast Cancer Risk Reduction Panel recommends limiting RRSO to women with a known or strongly suspected BRCA1/2 mutation. Peritoneal washings should be performed at surgery, and pathologic assessment should include fine sectioning of the ovaries and fallopian tubes.^95,114

The additional benefit of concurrent hysterectomy is not clear at this time. Women who undergo RRSO should continue with routine health maintenance and breast screening as per the NCCN Guidelines for Breast Cancer (to view the most recent version of these guidelines, visit NCCN.org) unless they have undergone RRM.

Risk Reduction Agents

Risk reduction agents (ie, tamoxifen, raloxifene, anastrozole, exemestane) are recommended for women 35 years of age or older only, because the utility of these agents in women younger than 35 years is unknown.

Tamoxifen for Risk Reduction: The benefits of tamoxifen, a selective estrogen receptor (ER) modulator (SERM), in the treatment of breast cancer in the adjuvant and metastatic settings are well documented. Retrospective analysis of randomized, controlled, clinical trials comparing tamoxifen with no tamoxifen in the adjuvant treatment of women with breast cancer has shown a reduction in the incidence of contralateral second primary breast cancer.^115–118 The meta-analyses by Early Breast Cancer Trialists' Collaborative Group confirmed that the risk for contralateral primary breast cancer is substantially reduced (ie, a statistically significant annual recurrence rate ratio, 0.59) by 5 years of tamoxifen therapy in women with first breast cancers that are ER-positive or have an unknown ER status.¹¹⁹

NSABP Breast Cancer Prevention Trial: The effectiveness of tamoxifen in the setting of breast cancer treatment gave rise to the NSABP BCPT study, also known as the P-1 study. It was a randomized clinical trial of healthy women aged 60 years or older, aged 35 to 59 years with a 1.7% or greater cumulative 5-year risk for developing breast cancer, or with a history of LCIS.⁷² Both premenopausal and postmenopausal women were enrolled in the trial and randomized in a double-blinded fashion to treatment with tamoxifen, 20 mg daily for 5 years, or placebo. Invasive breast cancer incidence was the primary study end point; high-priority secondary end points included the occurrence of thromboembolic disease, cardiovascular disease, bone fracture, endometrial cancer, noninvasive breast cancer, and breast cancer mortality. The trial was unblinded and initial findings were reported in 1998. A subsequent report on this trial has been published, which takes into account 7 years of follow-up data subsequent to the point where the study was unblinded. However, nearly one-third of the placebo participants began taking a SERM when the study was unblinded, which decreased the proportion of women in the placebo group relative to the tamoxifen group, potentially confounding the long-term results.⁷³

The results of the P-1 study showed that treatment with tamoxifen decreased the short-term risk for breast cancer by 49% in healthy women aged 35 years or older who had an increased risk for the disease.⁷² Risk reduction benefits were shown across all age groups, in pre- and postmenopausal women. The difference in average annual rates for invasive breast cancer was 3.30 cases per 1000 women (ie, 6.76 cases per 1000 women in the placebo group and 3.43 cases per 1000 women in the group taking tamoxifen). The absolute risk reduction was 21.4 cases per 1000 women over 5 years.⁷² In terms of numbers needed to treat, this corresponds to treatment of 47 women with tamoxifen to prevent 1 case of invasive breast cancer. Updated results indicate that breast cancer risk was reduced by 43% in this population after 7 years of follow-up.⁷³ The reduction in invasive breast cancer risk in participants with AH was particularly striking (RR, 0.14; 95% CI, 0.03–0.47) in the initial study analysis, and a RR of 0.25 (95% CI, 0.10–0.52) was found after 7 years of follow-up.

An additional benefit of tamoxifen was a decrease in bone fractures (RR, 0.81; 95% CI, 0.63–1.05). However, as was anticipated from the experience in studies of women taking tamoxifen after a breast cancer diagnosis, major toxicities included hot flashes, invasive endometrial cancer in postmenopausal women, and cataracts. A significant increase in the incidence of pulmonary embolism was also seen in women 50 years of age or older taking tamoxifen. The average annual rates of pulmonary embolism per 1000 women were 1.00 versus 0.31 (RR, 3.19; 95% CI, 1.12–11.15).⁷²

No differences were seen in overall rates of mortality by treatment group with a follow-up period of up to 7 years. The initial study analysis revealed that average annual mortality from all causes in the tamoxifen group was 2.17 per 1000 women compared with 2.71 per 1000 women treated with placebo, for an RR of 0.81 (95% CI, 0.56–1.16).⁷²Annual mortality after 7 years of follow-up was 2.80 per 1000 women compared with 3.08 per 1000 women in the tamoxifen and placebo groups, respectively, for an RR of 1.10 (95% CI, 0.85–1.43).⁷³

An evaluation of the subset of patients with a BRCA1/2 mutation in the P-1 study revealed that breast cancer risk was reduced by 62% in study patients with a BRCA2 mutation receiving tamoxifen relative to placebo (RR, 0.38; 95% CI, 0.06–1.56). However, tamoxifen use was not associated with a reduction in breast cancer risk in patients with a BRCA1 mutation.¹²⁰ These findings may be related to the greater likelihood of development of ER-positive tumors in BRCA2 mutation carriers relative to BRCA1 mutation carriers. However, this analysis was limited by the very small number of patients with a BRCA1/2 mutation. Currently, no prospective studies evaluating risk are reviewing the reductive effect of tamoxifen in BRCA mutation carriers.

Based on the P-1 study results, in October 1998 the FDA approved tamoxifen for breast cancer risk reduction for women at increased risk for breast cancer.

European Studies of Tamoxifen: Three European studies comparing tamoxifen with placebo for breast cancer risk reduction have also been reported. The Royal Marsden Hospital study was a pilot trial of tamoxifen versus placebo in women aged 30 to 70 years who were at increased breast cancer risk based largely on their family history.^121,122 Women in the trial were allowed to continue or to initiate postmenopausal HT. With 2471 participants available for interim analysis, no difference in the frequency of breast cancer was seen between the 2 study groups. Moreover, the toxicity experienced by the 2 groups did not show statistically significant differences.¹²² An analysis of updated findings from this study showed a nonsignificant breast cancer risk reduction benefit with tamoxifen use (ie, 62 cases of breast cancer in 1238 women receiving tamoxifen vs 75 cases of breast cancer in 1233 women in the placebo arm).¹²¹

Most recently, an analysis of blinded results from the Royal Marsden Hospital trial at 20-year follow-up showed no difference in breast cancer incidence between the groups randomly assigned to tamoxifen or placebo (HR, 0.78; 95% CI, 0.58– 1.04; P=.10).¹²³ However, the incidence of ER-positive breast cancer was significantly lower in the tamoxifen arm (HR, 0.61; 95% CI, 0.43–0.86; P=.005). Importantly, the difference between the 2 arms became significant only in the posttreatment period (ie, after 8 years of treatment).

The Italian Tamoxifen Prevention Study randomized 5408 women aged 35 to 70 years without breast cancer, who had undergone a previous hysterectomy, to receive tamoxifen or placebo for 5 years.¹²⁴ Women in the trial were allowed to receive HT. No significant difference in breast cancer occurrence in the overall study population was identified at median follow-up of 46, 81.2 and 109.2 months.^124–126 Thromboembolic events, predominantly superficial thrombophlebitis, were increased in women treated with tamoxifen. A subset of women in the Italian Tamoxifen Prevention Study who had used HT and were classified as being at increased breast cancer risk based on reproductive and hormonal characteristics were found to have a significantly reduced risk for breast cancer with tamoxifen therapy.^126,127 However, only approximately 13% of the patients in the trial were at high risk for breast cancer.

Why no overall breast cancer risk reduction was observed in the Italian Tamoxifen Prevention Study is unclear. Possible reasons include concurrent use of HT and different study populations (ie, populations at lower risk for breast cancer).¹²⁸

The first International Breast Cancer Intervention Study (IBIS-I) randomized 7152 women aged 35 to 70 years at increased risk for breast cancer to receive either tamoxifen or placebo for 5 years.¹²⁹ Tamoxifen provided a breast cancer (invasive breast cancer or DCIS) risk reduction of 32% (95% CI, 8–50; P=.013). Thromboembolic events increased with tamoxifen (OR, 2.5; 95% CI, 1.5–4.4; P=.001), and endometrial cancer showed a nonsignificant increase (P=.20). An excess of deaths from all causes was seen in the tamoxifen-treated women (P=.028).

After a median follow-up of 8 years a significant reduction for all types of invasive breast cancer was reported (RR, 0.73; 95% CI, 0.58–0.91; P=.004) with tamoxifen.¹³⁰ Although no difference in the risk for ER-negative invasive tumors was seen between the 2 groups, those in the tamoxifen arm were found to have a 34% lower risk for ER-positive invasive breast cancer.¹³⁰ Slightly higher risk reduction with tamoxifen was observed for premenopausal patients. Importantly, the increased risk for venous thromboembolism (VTE) seen with tamoxifen during the treatment period was no longer significant in the posttreatment period. Gynecologic and vasomotor adverse effects associated with active tamoxifen treatment were not seen during the posttreatment follow-up.

A recently reported updated analysis after a median follow-up of 16 years confirmed that the preventive effects of tamoxifen continue, with a significant reduction in the first 10 years (HR, 0.72; 95% CI, 0.59–0.88; P=.001), and a slightly greater reduction in subsequent years (HR, 0.69; 0.53–0.91; P=.009).¹³¹ A similar pattern was observed regarding reduction in occurrence of invasive ER-positive breast cancer throughout the follow-up; a significant reduction in DCIS with tamoxifen was also recorded, but only in the first 10 years of follow-up. Interestingly, more ER-negative breast cancers were reported in the tamoxifen group after 10 years of follow-up than in the placebo group (HR, 2.45; 0.77–7.82; P=.13).¹³¹

The use of tamoxifen as a breast cancer risk reduction agent has been evaluated in the STAR trial^74,75 (see “The STAR Trial,” page 898).

Raloxifene for Risk Reduction: Raloxifene is a second-generation SERM that is chemically different from tamoxifen and appears to have similar antiestrogenic effects with considerably less endometrial stimulation. The efficacy of raloxifene as a breast cancer risk reduction agent has been evaluated in several clinical studies. In 2007, the FDA expanded the indications for raloxifene to include reduction in risk for invasive breast cancer in postmenopausal women with osteoporosis, and reduction in risk for invasive breast cancer in postmenopausal women at high risk for invasive breast cancer.

The MORE Trial: The Multiple Outcomes of Raloxifene Evaluation (MORE) trial was designed to determine whether 3 years of raloxifene treatment reduced the risk of fracture in postmenopausal women with osteoporosis.¹³² A total of 7705 postmenopausal women 31 to 80 years of age were randomized to receive placebo, 60 mg/d of raloxifene, or 120 mg/d of raloxifene for 3 years. At study entry, participants were required to have osteoporosis (defined as a bone density at least 2.5 standard deviations below the mean for young women) or a history of osteoporotic fracture. The study showed a reduction in vertebral fracture risk and an increase in bone mineral density (BMD) in the femoral neck and spine for the women treated with raloxifene, compared with those who received placebo.

After a median follow-up of 40 months in the MORE trial, breast cancer was reported in 40 patients: 27 cases in 2576 women receiving placebo and 13 cases in 5129 women receiving raloxifene.¹³³ The relative risk of developing invasive breast cancer on raloxifene compared with placebo was 0.24 (95% CI, 0.13–0.44). Raloxifene markedly decreased the risk for ER-positive cancers (RR, 0.10; 95% CI, 0.04–0.24) but did not appear to influence the risk of developing an ER-negative cancer (RR, 0.88; 95% CI, 0.26–3.0). Although breast cancer incidence was a secondary end point in the MORE trial, it is important to note that breast cancer risk was not a prospectively determined characteristic for the women enrolled and stratified into treatment arms in this study.¹²⁸ Furthermore, the patients enrolled in the MORE trial were, on average, at lower risk for breast cancer and older than the patients enrolled in the P-1 study.

Adverse events associated with raloxifene use included hot flashes, influenza-like syndromes, endometrial cavity fluid, peripheral edema, and leg cramps. In addition, an increased incidence of deep venous thromboses (DVTs; 0.7% for women receiving 60 mg/d of raloxifene vs 0.2% for placebo) and pulmonary emboli (0.3% for women receiving 120 mg/d of raloxifene vs 0.1% for placebo) was associated with raloxifene treatment. However, raloxifene was not associated with an increase in the risk for endometrial cancer.

The CORE Trial: The early findings related to breast cancer risk in the MORE trial led to its continuation as the Continuing Outcomes Relevant to Evista (CORE) trial. Because breast cancer incidence was a secondary end point in the MORE trial, the CORE trial was designed to assess the effect of 4 additional years of raloxifene on the incidence of invasive breast cancer in postmenopausal women with osteoporosis. A secondary end point was the incidence of invasive ER-positive breast cancer. Data from the CORE trial were reported in 2004.¹³⁴

During the CORE trial, the 4-year incidence of invasive breast cancer was reduced by 59% (HR, 0.41; 95% CI, 0.24–0.71) in the raloxifene group compared with the placebo group. Raloxifene, compared with placebo, reduced the incidence of invasive ER-positive breast cancer by 66% (HR, 0.34; 95% CI, 0.18–0.66) but had no effect on invasive ER-negative breast cancers.¹³⁴ Over the 8 years of both trials (MORE and CORE), the incidence of invasive breast cancer was reduced by 66% (HR, 0.34; 95% CI, 0.22–0.50) in the raloxifene group compared with the placebo group. Compared with placebo, 8 years of raloxifene reduced the incidence of invasive ER-positive breast cancer by 76% (HR, 0.24; 95% CI, 0.15–0.40). Interestingly, the incidence of noninvasive breast cancer was not significantly different for patients in the raloxifene and placebo arms (HR, 1.78; 95% CI, 0.37–8.61).¹³⁴

The adverse events in the CORE trial were similar to those seen in the MORE trial. A nonsignificant increase was seen in the risk for thromboembolism (RR, 2.17; 95% CI, 0.83–5.70) in the raloxifene group of the CORE trial compared with the placebo group. No statistically significant difference was seen in endometrial events (bleeding, hyperplasia, and cancer) between the raloxifene and placebo groups during the 4 years of the CORE trial or the 8 years of the MORE and CORE trials. During the 8 years of the MORE and CORE trials, raloxifene increased the risk for hot flushes and leg cramps compared with placebo; these risks were observed during the MORE trial but not during the additional 4 years of therapy in the CORE trial. Although hot flushes and leg cramps may be early events that do not persist with continued therapy, it is also possible that an increased risk for these adverse events was not observed in the CORE trial as a result of selection bias (ie, women who experienced these symptoms in the MORE trial may have chosen not to continue in the CORE trial).

The results from the CORE trial are not entirely straightforward because of the complex design of the trial. Of the 7705 patients randomized in the MORE trial, only 4011 chose to continue, blinded to therapy, in the CORE trial; this drop-off likely introduces bias in favor of the treatment group. In the CORE trial, the researchers did not randomize the patients again (1286 in the placebo arm, 2725 in the raloxifene arm), maintaining the double-blinding of the original trial.

The RUTH Trial: In the Raloxifene Use for the Heart (RUTH) trial, postmenopausal women with an increased risk for coronary heart disease were randomly assigned to raloxifene or placebo arms.^135,136 Invasive breast cancer incidence was another primary end point of the trial, although only approximately 40% of the study participants had an increased risk for breast cancer according to the Gail model. Median exposure to study drug was 5.1 years and median duration of follow-up was 5.6 years.¹³⁶ Raloxifene did not reduce risk of cardiovascular events, but there was a 44% decrease in the incidence of invasive breast cancer in the raloxifene arm (HR, 0.56; 95% CI, 0.38–0.83), with a 55% lower incidence of ER-positive breast cancer (HR, 0.45; 95% CI, 0.28–0.72). No reduction in the risk for noninvasive breast cancer was found for patients receiving raloxifene, in agreement with the initial results of the STAR trial, although only 7% of breast cancers in the RUTH trial were noninvasive.

The STAR Trial: Despite issues of trial design, the results from the CORE trial and the previous MORE study provided support for concluding that raloxifene may be an effective breast cancer risk reduction agent. However, neither of these studies was designed to directly evaluate the efficacy of raloxifene versus tamoxifen in this regard. This issue was addressed in the NSABP STAR trial (P-2), which was started in 1999; initial results became available in 2006.⁷⁴

In the STAR trial, 19,747 postmenopausal women aged 35 years or older at increased risk for invasive breast cancer as determined by the modified Gail model or with a personal history of LCIS were enrolled into 1 of 2 treatment arms (no placebo arm). The primary study end point was invasive breast cancer; secondary end points included quality of life and incidences of noninvasive breast cancer, DVT, pulmonary embolism, endometrial cancer, stroke, cataracts, and death. At an average follow-up of approximately 4 years, no statistically significant differences in patients receiving 20 mg/d of tamoxifen or 60 mg/d of raloxifene were observed with respect to invasive breast cancer risk reduction (RR, 1.02; 95% CI, 0.82–1.28). Because the study had no placebo arm, determining a raloxifene versus placebo RR for invasive breast cancer was not possible; however, tamoxifen was shown in the P-1 study to reduce breast cancer risk by nearly 50%. In addition, raloxifene was shown to be as effective as tamoxifen in reducing the risk for invasive cancer in the subset of patients with a history of LCIS or AH. However, raloxifene was not as effective as tamoxifen in reducing the risk for noninvasive breast cancer, although the observed difference was not statistically significant (RR, 1.40; 95% CI, 0.98–2.00).⁷²

At a median follow-up of nearly 8 years (81 months) involving 19,490 women, raloxifene was shown to be approximately 24% less effective than tamoxifen in reducing the risk for invasive breast cancer (RR, 1.24; 95% CI, 1.05–1.47), suggesting that tamoxifen has greater long-term benefit with respect to decreasing invasive breast cancer risk.⁷⁵ Raloxifene remained as effective as tamoxifen in reducing the risk for invasive cancer in women with LCIS (RR, 1.13; 95% CI, 0.76–1.69), but was less effective than tamoxifen in women with a history of AH (RR, 1.48; 95% CI, 1.06–2.09). Interestingly, at long-term follow-up, the risk for noninvasive cancer in the raloxifene arm was closer to that observed for the group receiving tamoxifen (RR, 1.22; 95% CI, 0.95–1.50). No significant differences in mortality were seen between the groups. In the initial analysis of the STAR trial data, invasive endometrial cancer occurred less frequently in the group receiving raloxifene compared with the tamoxifen group, although the difference did not reach statistical significance. It is important to note, however, that the incidence of endometrial hyperplasia and hysterectomy were significantly lower in the raloxifene group compared with the tamoxifen group. However, at long-term follow-up, the risk for endometrial cancer was significantly lower in the raloxifene arm (RR, 0.55; 95% CI, 0.36–0.83).

The lower incidences of thromboembolic events (RR, 0.75; 95% CI, 0.60–0.93) and cataract development (RR, 0.80; 95% CI, 0.72–0.89) observed in the raloxifene group compared to the tamoxifen group when the STAR trial results were initially analyzed were maintained at long-term follow-up.⁷⁵ The incidences of stroke, ischemic heart disease, and bone fracture were similar in the 2 groups. In the initial report, overall quality of life was reported to be similar for patients in both groups, although patients receiving tamoxifen reported better sexual function.¹³⁷

Aromatase Inhibitors for Risk Reduction: Several clinical trials testing the use of aromatase inhibitors (AIs) in the adjuvant therapy of postmenopausal women with invasive breast cancer have been reported. The first of these studies, the Arimidex, Tamoxifen, Alone or in Combination (ATAC) trial, randomized postmenopausal women with invasive breast cancer to anastrozole versus tamoxifen versus anastrozole plus tamoxifen in a double-blinded fashion.¹³⁸ The occurrence of contralateral second primary breast cancers was a study end point. At 47 months median follow-up, a nonsignificant reduction in contralateral breast cancers was observed in women treated with anastrozole alone compared with tamoxifen (OR, 0.62; 95% CI, 0.38–1.02; P=.062) and a significant reduction in contralateral breast cancers was seen in the subset of women with hormone receptor–positive first cancers (OR, 0.56; 95% CI, 0.32–0.98; P=.04).¹³⁹ Similar reductions in the risk for contralateral breast cancer have been observed with sequential tamoxifen followed by exemestane compared with tamoxifen alone and with sequential tamoxifen followed by letrozole compared with tamoxifen followed by placebo.^140,141

In the Breast International Group (BIG) 1-98 trial, postmenopausal women with early-stage breast cancer were randomized to receive 5 years of treatment with one of the following therapeutic regimens: letrozole; sequential letrozole followed by tamoxifen; tamoxifen; or sequential tamoxifen followed by letrozole. Risk for breast cancer recurrence was lower in women in the letrozole arm relative to the tamoxifen arm.¹⁴²

The results of the MAP.3 trial show promising use of exemestane in the breast cancer prevention setting. MAP.3 is a randomized, double-blind, placebo-controlled, multicenter, multinational trial in which 4560 women were randomly assigned to either exemestane (n=2285) or placebo (n=2275).⁴ The study authors reported that approximately 5% of patients in each group had discontinued the protocol treatment. The major reasons for early discontinuation of the protocol treatments were toxic effects (15.4% in the exemestane group vs 10.8% in the placebo group; P<.001) and patient refusal (6.9% vs 6.0%; P=.22). After a median follow-up of 3 years, exemestane was found to reduce the relative incidence of invasive breast cancers by 65% compared with placebo, from 0.55% to 0.19% (HR, 0.35 with exemestane; 95% CI, 0.18– 0.70).⁴

The IBIS-II study included 3864 postmenopausal women at high risk for breast cancer, defined by family history of breast cancer or prior diagnosis of DCIS, LCIS, or ADH (HR, 0.47; 95% CI, 0.32–0.68).⁵ The advantage of anastrozole was greater prevention of high-grade tumors (HR, 0.35; 95% CI, 0.16–0.74) compared with intermediate- or low-grade tumors. The follow-up period in this trial was longer than that for the MAP.3 trial. The cumulative incidence after 7 years was predicted to increase 2.8% in the anastrozole group compared with 5.6% in the placebo group.⁵

NCCN Breast Cancer Risk Reduction Panel Recommendations for Risk Reduction Agents: Based on data from the BCPT⁷² and STAR⁷⁴ trials, Freedman et al³ developed tables of benefit/risk indices for women aged 50 years and older to compare raloxifene versus no treatment (placebo) and tamoxifen versus no treatment. The risk and benefit of treatment with either tamoxifen or raloxifene depends on age, race, breast cancer risk, and history of hysterectomy. There are separate tables in the report listing the level of 5-year invasive breast cancer risk by age group for non-Hispanic white women with and without a uterus, black women with and without a uterus, and Hispanic women with and without a uterus. The NCCN Breast Cancer Risk Reduction Panel recommends using these tables³ when counseling postmenopausal women regarding use of raloxifene and tamoxifen for breast cancer risk reduction. It should be noted that these tables do not consider the greater risk reduction achieved in women with proliferative breast lesions, such as AH.

Tamoxifen Recommendations: The NCCN Breast Cancer Risk Reduction Panel recommends tamoxifen (20 mg/d) as an option to reduce breast cancer risk in healthy premenopausal and postmenopausal women aged 35 years or older whose life expectancy is 10 years or more, and who have a 1.7% or greater 5-year risk for breast cancer as determined by the modified Gail model, or who have had LCIS (category 1). The consensus of the NCCN Breast Cancer Risk Reduction Panel is that the risk/benefit ratio for tamoxifen use in premenopausal women at increased risk for breast cancer is relatively favorable (category 1), and that the risk/benefit ratio for tamoxifen use in postmenopausal women is influenced by age, presence of uterus, or other comorbid conditions (category 1). Early studies suggest that lower doses of tamoxifen over shorter treatment periods may reduce breast cancer risk in postmenopausal women, but these findings need to be validated in phase III clinical trials.¹⁴³ Only limited data are currently available regarding the efficacy of tamoxifen risk reduction in BRCA1/2 mutation carriers and women who have received prior thoracic radiation; no prospective studies evaluating the risk-reductive effect of tamoxifen in women with BRCA mutations are currently available. However, available data from a very small cohort suggest a benefit for women with a BRCA2 mutation but possibly not for women with a BRCA1 mutation.¹²⁰

The utility of tamoxifen as a breast cancer risk reduction agent in women younger than 35 years is not known. Tamoxifen is a teratogen and is contraindicated during pregnancy or in women planning a pregnancy. Data on the influence of ethnicity and race on the efficacy and safety of tamoxifen as a risk reduction agent are insufficient.

Evidence exists that certain drugs (eg, selective serotonin reuptake inhibitors [SSRIs]) interfere with the enzymatic conversion of tamoxifen to endoxifen by inhibiting a particular isoform of cytochrome P-450 2D6 (CYP2D6) enzyme involved in the metabolism of tamoxifen.¹⁴⁴ The consensus of the NCCN Breast Cancer Risk Reduction Panel is that alternative medications that have minimal or no impact on plasma levels of endoxifen should be substituted when possible.¹⁴⁴ Citalopram and venlafaxine do not disrupt tamoxifen metabolism.

Certain CYP2D6 genotypes have been reported to be markers of poor tamoxifen metabolism.^145,146 Nevertheless, the consensus of the NCCN Breast Cancer Risk Reduction Panel is that further validation of this biomarker is needed before it can be used to select patients for tamoxifen therapy.

Raloxifene Recommendations: The experts serving on the NCCN Breast Cancer Risk Reduction Panel feel strongly that tamoxifen is a superior choice of risk reduction agent for most postmenopausal women desiring nonsurgical risk reduction therapy. This is based on the updated STAR trial results that showed diminished benefits of raloxifene compared to tamoxifen after cessation of therapy.⁷⁵ However, consideration of toxicity may still lead to the choice of raloxifene over tamoxifen in some women.

If raloxifene is chosen, the NCCN Breast Cancer Risk Reduction Panel recommends use of 60 mg/d. Data regarding use of raloxifene to reduce breast cancer risk are limited to healthy postmenopausal women aged 35 years or older who have a 1.7% or greater 5-year risk for breast cancer as determined by the modified Gail model, or who have a history of LCIS. The consensus of the NCCN Breast Cancer Risk Reduction Panel is that the risk/benefit ratio for raloxifene use in postmenopausal women at increased risk for breast cancer is influenced by age and comorbid conditions (category 1). Because no data are currently available regarding the efficacy of raloxifene risk reduction in BRCA1/2 mutation carriers and women who have received prior thoracic radiation, use of raloxifene in these populations is designated as a category 2A recommendation by the NCCN Breast Cancer Risk Reduction Panel. Use of raloxifene to reduce breast cancer risk in premenopausal women is inappropriate unless part of a clinical trial. The utility of raloxifene as a breast cancer risk reduction agent in women younger than 35 years is not known. Data on the influence of ethnicity and race on the efficacy and safety of raloxifene as a risk reduction agents are insufficient.

Overall, risk reduction therapy with tamoxifen and raloxifene has been vastly underused.¹⁴⁷ Women in whom the benefits of risk reduction therapy far outweigh harms include those with AH (both ductal and lobular types) and those with LCIS.^54,72 Women with AH and LCIS have a significantly higher risk of developing invasive breast cancer. The initial and follow-up results of the P-1 study (described in previous sections) demonstrated a significant risk reduction in women with AH with tamoxifen therapy.^72,73 Despite this, a recent study has documented that only 44% of women with AH or LCIS received risk reduction therapy.⁵⁴ Considering the opportunity that exists for a significant impact of risk reduction therapy on the reducing the incidence of breast cancer, the NCCN Panel strongly recommends risk reduction therapy in women with AH.

AI Recommendations (Anastrozole and Exemestane): The experts serving on the NCCN Breast Cancer Risk Reduction Panel have included exemestane and anastrozole as choices of risk reduction agent for most postmenopausal women desiring nonsurgical risk reduction therapy (category 1). This is based on the results of the MAP.3⁴ and IBIS-II trials.⁵ The panel recommends use of 25 mg/d of exemestane or 1 mg/d of anastrozole. Data regarding the use of AIs (exemestane and anastrozole) to reduce breast cancer risk are limited to postmenopausal women 35 years of age or older with a Gail model 5-year risk score greater than 1.66% or a history of LCIS. The consensus of the NCCN Breast Cancer Risk Reduction Panel is that the risk/benefit ratio for use of an AI in postmenopausal women at increased risk for breast cancer is influenced by age, bone density, and comorbid conditions. Use of an AI to reduce breast cancer risk in premenopausal women is inappropriate unless part of a clinical trial. The utility of an AI as a breast cancer risk reduction agent in women younger than 35 years is not known. Data on the influence of ethnicity and race on the efficacy and safety of an AI as a risk reduction agent are insufficient.

Exemestane and anastrazole are not currently FDA-approved for breast cancer risk reduction. Currently, no data are available comparing the benefits and risks of AIs with those of tamoxifen or raloxifene.

Monitoring Patients on Risk Reduction Agents

Follow-up of women treated with risk reduction agents for breast cancer risk reduction should focus on the early detection of breast cancer and the management of adverse symptoms or complications. Appropriate monitoring for breast cancer and the evaluation of breast abnormalities should be performed according to the guidelines described for high-risk women in the NCCN Guidelines for Breast Cancer Screening and Diagnosis (to view the most recent version of these guidelines, visit NCCN.org). The population of women eligible for risk reduction therapy with tamoxifen, raloxifene, anastrozole, or exemestane is at sufficiently increased risk for breast cancer to warrant, at a minimum, yearly bilateral mammography, a clinical breast examination every 6 to 12 months, and encouragement of breast awareness.

Endometrial Cancer: Results from the P-1 study indicated that women aged 50 years or older treated with tamoxifen have an increased risk of developing invasive endometrial cancer compared with placebo (RR, 4.01; 95% CI, 1.70–10.90).^72,73 However, an increased risk for endometrial cancer was not observed in women aged 49 years or younger treated with tamoxifen (RR, 1.21; 95% CI, 0.41–3.60).^72,73 Although the only death from endometrial cancer in the P-1 study occurred in a placebo-treated subject,^72,73 analyses of the NSABP data have revealed a small number of uterine sarcomas among patients with an intact uterus taking tamoxifen. Uterine sarcoma is a rare form of uterine malignancy that has been reported to occur in 2% to 4% of all patients with uterine cancer.¹⁴⁸ Compared with other uterine cancers, uterine sarcomas present at a more advanced stage and thus may carry a worse prognosis in terms of disease-free and overall survivals.^149,150

Updated results from the NSABP studies have indicated that incidence of both endometrial adenocarcinoma and uterine sarcoma is increased in women taking tamoxifen when compared with placebo.¹⁵¹ Several other studies have also supported an association between tamoxifen therapy and an increased risk of developing uterine sarcoma.^{149,150,152,153} A “black box” FDA warning has been included on the package insert of tamoxifen to highlight the endometrial cancer risk (both epithelial endometrial cancer and uterine sarcoma) of tamoxifen.¹⁵⁴ Nonetheless, the absolute risk of developing endometrial cancer is low (absolute annual risk per 1000: placebo, 0.91 vs tamoxifen, 2.30). Often, in women at increased risk for breast cancer, the reduction in the number of breast cancer events exceeds that of the increase in the number of uterine cancer events.

Use of raloxifene has not been associated with an increased incidence of endometrial cancer in the MORE trial.¹³³ Long-term results from the STAR trial showed the incidence of invasive endometrial cancer to be significantly lower in the group receiving raloxifene compared with those receiving tamoxifen (RR, 0.55; 09% CI, 0.36–0.83).⁷⁵

For women with an intact uterus, a baseline gynecologic assessment is recommended before administration of tamoxifen, and follow-up gynecologic assessments should be performed at each visit.¹⁵⁵ Most women with tamoxifen-associated endometrial cancer present with vaginal spotting as an early symptom of cancer. Therefore, prompt evaluation of vaginal spotting in the postmenopausal woman is essential.

Currently, evidence is insufficient to recommend the performance of uterine ultrasonography or endometrial biopsy for routine screening in asymptomatic women.^156–158 In women diagnosed with endometrial cancer while taking a risk reduction agent, the drug should be discontinued until the endometrial cancer has been fully treated. The NCCN Breast Cancer Risk Reduction Panel believes that it is safe and reasonable to resume therapy with a risk reduction agent after completion of treatment for early-stage endometrial cancer.

Retinopathy and Cataract Formation: The association of tamoxifen with the occurrence of retinopathy has been reported, although most of this information has come from case studies.^159,160 Furthermore, these reports have not been confirmed in the randomized controlled trials of tamoxifen. A 1.14 relative risk of cataract formation (95% CI, 1.01–1.29) when compared with placebo has been reported in the P-1 study, and individuals developing cataracts while on tamoxifen have a relative risk for cataract surgery of 1.57 (95% CI, 1.16–2.14) compared with those on placebo.⁷² After 7 years of follow-up in the P-1 study, relative risks of cataract formation and cataract surgery were similar to those initially reported.⁷³ In the MORE trial, raloxifene use was not associated with an increase in the incidence of cataracts compared with placebo (RR, 0.9; 95% CI, 0.8–1.1).¹⁶¹ In the STAR trial, the incidence of cataract development and occurrence of cataract surgery was significantly higher in the group receiving tamoxifen compared with the group receiving raloxifene.^75,161 The rate of cataract development (RR, 0.80; 95% CI, 0.72–0.89) and the rate of cataract surgery (RR, 0.79; 95% CI, 0.70–0.90) were approximately 20% less in the raloxifene group than in the tamoxifen group.^75,161 Thus, patients experiencing visual symptoms while undergoing treatment with tamoxifen should seek ophthalmologic evaluation.

Bone Mineral Density: Bone is an estrogen-responsive tissue, and tamoxifen can act as either an estrogen agonist or estrogen antagonist with respect to bone, depending on the menstrual status of a woman.^{122,162–164} In premenopausal women, tamoxifen may oppose the more potent effects of estrogen on the bone and potentially increase the risk for osteoporosis, whereas tamoxifen in the presence of typically lower estrogen levels in postmenopausal women is associated with an increase in BMD.^72,73 However, the NCCN Breast Cancer Risk Reduction Panel does not recommend monitoring BMD in premenopausal patients on tamoxifen, because the development of osteopenia/osteoporosis in this population is considered unlikely. Raloxifene has been shown to increase BMD and to reduce incidence of vertebral bone fracture in postmenopausal women when compared with placebo.^132,135 Results from the STAR trial did not reveal any difference in the incidence of bone fracture in the groups of postmenopausal women on either raloxifene or tamoxifen.^74,75 Changes in BMD are of concern in women on AI therapy. Therefore, a baseline BMD scan is recommended before initiating therapy with an AI, such as anastrozole or exemestane.

Thromboembolic Disease and Strokes: Tamoxifen and raloxifene have been associated with an increased risk of thromboembolic events (ie, DVT, pulmonary embolism) and stroke.^{72–75,133,165} Increased incidences of VTE were observed in the tamoxifen arms of all the placebo-controlled, randomized, risk reduction trials. Although not statistically significant, all of these trials with the exception of the Royal Marsden Hospital trial (which enrolled only younger women) also showed an increase in risk for stroke for women receiving tamoxifen. This risk was found to be significantly elevated in 2 meta-analyses of randomized controlled trials evaluating tamoxifen for breast cancer risk reduction or treatment.^166,167 Comparison of the raloxifene and tamoxifen arms of the STAR trial did not show a difference with respect to incidence of stroke,^74,75 and the risk of fatal stroke was significantly higher for women in the RUTH trial with underlying heart disease receiving raloxifene.¹³⁶ However, evidence has shown that women with a factor V Leiden or prothrombin G20210A mutation receiving tamoxifen therapy in the P-1 study were not at increased risk of developing VTE compared with women without these mutations.¹⁶⁸

Although prospective screening of women for factor V Leiden or prothrombin mutations or intermittent screening of women for thromboembolic disease is unlikely to be of value, women taking tamoxifen or raloxifene should be educated regarding the symptoms associated with DVT and pulmonary emboli. They should also be informed that prolonged immobilization may increase risk of VTE, and should be instructed to contact their physicians immediately if they develop symptoms of DVT or pulmonary emboli. Women with documented thromboembolic disease should receive appropriate treatment for the thromboembolic condition and should permanently discontinue tamoxifen or raloxifene therapy.

Managing Side Effects of Risk Reduction Agents

Hot flashes are a common menopausal complaint. In the P-1 study, hot flashes occurred in approximately 81% of women treated with tamoxifen and 69% of women treated with placebo.⁷² In the STAR trial, women receiving tamoxifen reported a significantly increased incidence of vasomotor symptoms relative to women receiving raloxifene,¹³⁷ although raloxifene use has also been associated with an increase in hot flash severity and/or frequency when compared with placebo.¹³³ In women whose quality of life is diminished by hot flashes, an intervention to eliminate or minimize hot flashes should be undertaken. Estrogens and/or progestins have the potential to interact with SERMs and are not recommended by the NCCN Breast Cancer Risk Reduction Panel for the treatment of hot flashes for women on a risk reduction agent outside of a clinical trial.

Gabapentin, a gamma-aminobutyric acid (GABA) analog used primarily for seizure control and management of neuropathic pain, has been reported to moderate both the severity and duration of hot flashes.^169–172 It has been hypothesized that the mode of action of gabapentin is via central temperature regulatory centers.^169,170 Results from a randomized, double-blind, placebo-controlled study involving the use of gabapentin to treat hot flashes in 420 women with breast cancer have been reported. The 3 treatment arms of the trial were as follows: 300 mg/d of gabapentin; 900 mg/d of gabapentin; and placebo. Study duration was 8 weeks, and most of the women in the study (68%–75% depending on treatment arm) were taking tamoxifen as adjuvant therapy. Women in the placebo group experienced reductions in severity of hot flashes of 21% and 15% at 4 and 8 weeks, respectively, whereas those in the treatment arms reported reductions of 33% and 31% with lower-dose gabapentin, and 49% and 46% with higher-dose gabapentin at 4 and 8 weeks, respectively. Only women receiving the higher dose of gabapentin had significantly fewer and less severe hot flashes. Side effects of somnolence or fatigue were reported in a small percentage of women taking gabapentin.¹⁷²

Venlafaxine, a serotonin and norepinephrine inhibitor antidepressant, has been shown to be effective in the management of hot flash symptoms in a group of breast cancer survivors, 70% of whom were taking tamoxifen. Significant declines were observed for both hot flash frequency and severity scores for all doses of venlafaxine (37.5, 75, and 150 mg) compared with placebo; incremental improvement was seen at 75 versus 37.5 mg (P=.03).¹⁷³ Participants receiving venlafaxine reported mouth dryness, reduced appetite, nausea, and constipation with increased prevalence at increased dosages. Based on these findings, the authors suggested a starting dose of 37.5 mg with an increase as necessary after 1 week to 75 mg if a greater degree of symptom control is desired. However, this study followed subjects for only 4 weeks.

Another antidepressant, paroxetine, an SSRI, has also been studied for the relief of hot flash symptoms. A double-blind, placebo-controlled trial recruited 165 menopausal women who were randomized into 3 arms (placebo, paroxetine 12.5 mg/d, or paroxetine 25 mg/d). After 6 weeks, significant reductions in composite hot flash scores were noted for both dosages of paroxetine (12.5 mg, 62% reduction; 25 mg, 65% reduction); no significant differences were seen between dose levels.¹⁷⁴ Adverse events, reported by 54% of subjects receiving placebo and 58% receiving paroxetine, generally included nausea, dizziness, and insomnia.

In a stratified, randomized, double-blind, crossover, placebo-controlled study, 151 women reporting a history of hot flashes were randomized to one of 4 treatment arms (10 or 20 mg of paroxetine for 4 weeks followed by 4 weeks of placebo or 4 weeks of placebo followed by 4 weeks of 10 or 20 mg of paroxetine).¹⁷⁵ Hot flash frequency and composite score were reduced by 40.6% and 45.6%, respectively, for patients receiving 10 mg of paroxetine compared with reductions of 13.7% and 13.7% in the placebo group. Likewise, reductions of 51.7% and 56.1% in hot flash frequency and score were found in women receiving 20 mg of paroxetine compared with values of 26.6% and 28.8% in the placebo group. No significant differences in efficacy were observed with the lower and higher paroxetine doses. Rates of the most commonly reported side effects did not differ among the 4 arms, although nausea was significantly increased in women receiving 20 mg of paroxetine relative to the other arms, and a greater percentage of patients receiving the higher dose of paroxetine discontinued treatment.

Although these reports appear promising, further randomized studies of the use of these agents in women experiencing hot flash symptoms, especially those also taking tamoxifen, are needed to assess the long-term effectiveness and safety of these agents. In this context it should be noted that recent evidence has suggested that concomitant use of tamoxifen with certain SSRIs (eg, paroxetine and fluoxetine) may decrease plasma levels of endoxifen and 4-OH tamoxifen, active metabolites of tamoxifen, and may impact its efficacy.^144,176 These SSRIs may interfere with the enzymatic conversion of tamoxifen to its active metabolites by inhibiting a particular isoform of cytochrome P-450 (CYP2D6) enzyme involved in the metabolism of tamoxifen. Caution is advised about co-administration of these drugs with tamoxifen. Citalopram and venlafaxine appear to have only minimal effects on tamoxifen metabolism.

Of interest in this context are results of a retrospective evaluation of data from the Women's Healthy Eating and Living randomized trial, which suggest an inverse association between hot flashes and breast cancer recurrence for women with a history of breast cancer receiving tamoxifen. These results suggest that hot flashes in women receiving tamoxifen may be an indicator of the biologic availability and, thus, effectiveness of the drug. However, additional studies are needed to further elucidate whether hot flashes are predictive of benefit from tamoxifen.¹⁷⁷

A recent report of 2 nonrandomized, parallel study cohorts of women with DCIS or those at high risk for breast cancer (eg, those with LCIS, AH, or ≥1.7% 5-year breast cancer risk by the Gail model) comparing women receiving tamoxifen alone with women receiving tamoxifen concomitantly with HT (mean duration of HT at start of study was approximately 10 years) did not show a difference in the rate of tamoxifen-induced hot flashes.¹⁷⁸ The NCCN Breast Cancer Risk Reduction Panel recommends against the use of HT for women taking tamoxifen or raloxifene outside of a clinical trial.

A variety of other substances for the control of hot flashes have been described.¹⁷⁹ Both the oral and transdermal formulations of clonidine reduce hot flashes in a dose-dependent manner.^180–182 Toxicities associated with clonidine include dry mouth, constipation, and drowsiness. Anecdotal evidence suggests that the use of a number of different herbal or food supplements may alleviate hot flashes. Vitamin E may decrease the frequency and severity of hot flashes, but results from a randomized clinical trial demonstrated that only a very modest improvement in hot flashes was associated with this agent compared with placebo.¹⁸³ Results from a double-blind, randomized, placebo-controlled, crossover trial of the use of black cohosh to treat hot flashes did not show significant differences between groups with respect to improvement in hot flash symptoms.¹⁸⁴ Some herbal or food supplements contain active estrogenic compounds, the activity and safety of which are unknown. Other strategies for the management of hot flashes, such as relaxation training, acupuncture, avoidance of caffeine and alcohol, and exercise, although potentially beneficial, remain unsupported.¹⁸⁵

It should be noted that the observed placebo effect in the treatment of hot flashes is considerable, typically falling in the range 25% or more,^{169,171–175} suggesting that a considerable proportion of patients might be helped through a trial of therapy of limited duration. However, not all women who experience hot flashes require medical intervention, and the decision to intervene requires consideration of the efficacy and toxicity of the intervention. In addition, a study of women receiving tamoxifen for early-stage breast cancer showed a decrease in hot flashes over time.¹⁸⁶

Components of Risk Reduction Counseling

Women should be monitored according to the NCCN Guidelines for Breast Cancer Screening and Diagnosis (to view the most recent version of these guidelines, visit NCCN.org). Women with known or suspected BRCA1/2, TP53, PTEN, or other gene mutations associated with breast cancer risk or those with a significant family history of breast and/or ovarian cancer should also be followed according to the NCCN Guidelines for Genetic/Familial High-Risk Assessment: Breast and Ovarian (to view the most recent version of these guidelines, visit NCCN.org), despite whether they choose to undergo risk reduction therapy. Women who have abnormal results from their clinical breast examination or bilateral mammogram or those with a history of LCIS should be managed according to the NCCN Guidelines for Breast Cancer Screening and Diagnosis. All women who are appropriate candidates for breast cancer risk reduction intervention should undergo counseling that provides a description of the available strategies, including a healthy lifestyle, to decrease breast cancer risk.¹⁸⁷ Options for breast cancer risk reduction should be discussed in a shared decision-making environment. The counseling should include a discussion and consideration of (1) the individual's overall health status, including menopausal status, medical history, and medication history (eg, hysterectomy status, prior history of VTE, current use of hormones or SSRI, previous use of a SERM); (2) absolute and relative breast cancer risk reduction achieved with the risk reduction intervention; (3) risks of risk reduction therapy with an emphasis on age-dependent risks; (4) the contraindications to therapy with tamoxifen and raloxifene (eg, history of VTE, history of thrombotic stroke, history of transient ischemic attack, pregnancy or pregnancy potential without an effective nonhormonal method of contraception); and (5) the common and serious side effects of tamoxifen and raloxifene.

The 2009 ASCO guidelines comparing the effectiveness of breast cancer risk reduction agents provide some estimates of either the number needed to treat (NNT) to prevent breast cancer or the number needed to harm (NNH) by causing a specific side effect in a single patient receiving a specific risk reduction agent.¹⁸⁸ Both NNT and NNH can be useful aids in communicating risks and benefits of tamoxifen and raloxifene in this setting (eg, using long-term data from the IBIS-1 trial, NNH with respect to VTE was determined to be 73 with tamoxifen, whereas this value was 150 for patients receiving raloxifene using data from the RUTH study).

Counseling Prior to Therapy With Risk Reduction Agents: Counseling sessions with women who are considering non-surgical breast cancer risk reduction should incorporate an explanation of data from the P-1, STAR, MAP.3, and/or IBIS-II trial as appropriate.

Germline mutations in PTEN occur in 85% of patients with Cowden syndrome, an inherited condition associated with increased endometrial carcinoma risk. Therefore, increased risk for endometrial cancer in women with PTEN mutations should be discussed while considering a risk-reducing agent.

Counseling on Use of a SERM for Breast Cancer Risk Reduction: The P-1 study showed that the toxicity profile of tamoxifen is much more favorable in younger women, and the benefits in relative risk reduction are similar across all age groups and risk groups.⁷² The tamoxifen treatment risk/benefit ratio is especially favorable in women between ages 35 and 50 years. Unfortunately, individualized data regarding the risk/benefit ratio for tamoxifen are not generally available except for the broad age categories of ages 50 years and younger versus older than 50 years. Tamoxifen, unlike raloxifene, is a risk reduction agent that can be used by premenopausal women. In addition, tamoxifen may be more effective than raloxifene in reducing the incidence of noninvasive breast cancer, although the difference is not statistically significant at long-term follow-up.^74,75 Further, tamoxifen was reported by patients in the STAR trial to be associated with better sexual function than raloxifene.¹³⁷ However, tamoxifen has been associated with an increased incidence of invasive endometrial cancer relative to placebo in women 50 years of age or older,^72,73 and an increased incidence of endometrial hyperplasia and invasive endometrial cancer relative to raloxifene,^74,75 possibly making it a less attractive choice in women with a uterus. Use of raloxifene to reduce breast cancer risk may be preferred by postmenopausal women with a uterus or those at risk for developing cataracts. All women receiving a breast cancer risk reduction agent should be counseled with respect to signs and symptoms of possible side effects associated with use of these agents, and the recommended schedules for monitoring for the presence of certain adverse events. Contraindications to tamoxifen or raloxifene include history of VTE, thrombotic stroke, transient ischemic attack, current pregnancy or pregnancy potential without effective method of contraception, or known inherited clotting trait.

The optimal duration of SERM therapy for breast cancer risk reduction is not known. The P-1 and STAR trials studied 5 years of risk reduction therapy with either tamoxifen or raloxifene.^72,74 However, based on the updated STAR results, which showed that the benefits of raloxifene diminished after cessation of therapy,⁷⁵ continuing raloxifene beyond 5 years might be an approach to maintain the risk reduction activity of the agent.

The use of tamoxifen for periods longer than 5 years has been evaluated in the adjuvant treatment setting. Results of 2 randomized trials on extended adjuvant tamoxifen treatment^189,190 have demonstrated that tamoxifen for up to 10 years is more effective than shorter durations at preventing cancer recurrence and improving breast cancer survival. The option of 10 years of adjuvant tamoxifen therapy is now recommended for both premenopausal women and postmenopausal women for preventing cancer recurrence in the NCCN Guidelines for Breast Cancer (available at NCCN.org) and the recently updated ASCO guidelines.¹⁹¹ Data are limited on tamoxifen use for more than 5 years in the risk reduction setting. Until further information is available, a period of 5 years appears to be appropriate for tamoxifen therapy when the agent is used to reduce breast cancer risk.

After completing 5 years of tamoxifen therapy, women should continue to be monitored according to the NCCN Guidelines for Breast Cancer Screening and Diagnosis (to view the most recent version of these guidelines, visit NCCN.org) and should continue to undergo monitoring for late toxicity, especially for endometrial cancer and cataracts.

The prolonged effectiveness of tamoxifen as an agent to reduce breast cancer risk, particularly with respect to the development of ER-positive disease, is supported by results of several placebo-controlled, randomized trials at long-term follow-up.^73,123,130 The recent results from the STAR trial suggest that although a 5-year course of raloxifene retains considerable benefit with respect to the prevention of invasive breast cancer at a median follow-up of 81 months, the breast cancer preventive benefit of 5 years of tamoxifen therapy is significantly greater.⁷⁵

The NCCN Breast Cancer Risk Reduction Panel recommends using the tables from the Freedman et al³ publication when counseling postmenopausal women regarding use of raloxifene and tamoxifen for breast cancer risk reduction.

Counseling on Use of an AI for Breast Cancer Risk Reduction: Currently, there are no data comparing the benefits and risks of AIs (exemestane or anastrozole) with those of tamoxifen or raloxifene. Data regarding exemestane are from the single, large, randomized MAP.3 trial⁴ limited to postmenopausal women 35 years of age or older with a Gail model 5-year breast cancer risk of 1.7% or a history of LCIS, which may be used when counseling patients. The data show that exemestane has a completely different toxicity profile than the SERMs. Compared with the placebo group in the MAP.3 trial, exemestane had no increased risk of serious side effects. The incidence of osteoporosis, cardiac events, and bone fractures were identical for women in the MAP.3 trial taking exemestane and for those taking the placebo. However, follow-up was only 35 months. Women taking exemestane had a small, but not statistically significant increase in menopausal symptoms, such as hot flashes (18.3% vs 11.9%) and arthritis (6.5% vs 4.0%).⁴

Data regarding anastrozole are from a single, large, randomized trial, IBIS-II.5 The trial included postmenopausal women 40 to 70 years of age with a higher risk of developing cancer compared with the general population. Women who did not meet these criteria but had a Tyrer-Cuzick model 10-year breast cancer risk greater than 5% were also included.⁵ Musculoskeletal and vasomotor events were reported in both arms of the trial and were found to be significantly higher in the anastrazole arm (P=.0001); fracture rates were similar in both arms.⁵ The optimal duration of AI therapy is currently unknown. Changes in BMD are of concern in women receiving AI therapy. Therefore, a baseline BMD scan is recommended before initiating exemestane therapy. The role of calcium, vitamin D, and a healthy lifestyle in maintaining bone health must be emphasized in healthy postmenopausal women who are receiving exemestane.

Counseling Prior to Risk Reduction Surgery: For women at very high risk for breast cancer who are considering RRM, it is important that the potential psychosocial effects of RRM are addressed, although these effects have not been well studied.^192–194 Such surgery has the potential to negatively impact perceptions of body image, ease of forming new relationships, and the quality of existing relationships. Moreover, the procedure also eliminates the breast as a sexual organ. Multidisciplinary consultations are recommended prior to surgery, and should include a surgeon familiar with the natural history and therapy of benign and malignant breast disease¹⁹⁵ to enable the woman to become well informed regarding treatment alternatives, the risks and benefits of surgery, nipple-sparing mastectomy, and surgical breast reconstruction options. Immediate breast reconstruction is an option for many women following RRM, and early consultation with a reconstructive surgeon is recommended for those considering either immediate or delayed breast reconstruction.¹⁹⁶ Psychological consultations may also be considered.

Discussions regarding the risk for ovarian cancer and the option of RRSO for breast and ovarian cancer risk reduction should also be undertaken with women who are known carriers of a BRCA1/2 mutation. Other topics that should be addressed with respect to RRSO include the increased risk for osteoporosis and cardiovascular disease associated with premature menopause, as well as the potential effects of possible cognitive changes, accelerated bone loss, and vasomotor symptoms on quality of life. Furthermore, the surgery itself may have some associated complications.

It has been reported that short-term HT in women undergoing RRSO did not negate the reduction in breast cancer risk associated with the surgery.¹⁹⁷ In addition, results of a recent case-control study of BRCA1 mutation carriers showed no association between use of HT and increased breast cancer risk in postmenopausal BRCA1 mutation carriers.¹⁹⁷ However, the consensus of the NCCN Breast Cancer Risk Reduction Panel is that caution should be used when considering HT use in mutation carriers following RRSO, given the limitations inherent in nonrandomized studies (see “Breast Cancer Risks Associated With HT,” this page).^198,199 It is unlikely that a prospective randomized study on the use of RRSO for breast cancer risk reduction will be performed. Whether the resulting reduction in the risk for breast cancer from this procedure is preferable to a RRM is likely to remain a personal decision.²⁰⁰ Table 2 provides estimates based on a Monte Carlo simulation model of the survival impact of breast and ovarian risk reduction strategies (available online, in these guidelines, at NCCN.org). These data can be used as a tool to facilitate shared decision-making regarding choice of a risk reduction approach, particularly with respect to issues related to risk reduction surgery.

Counseling Regarding Lifestyle Modifications: Evidence indicates that certain lifestyle characteristics, such as obesity, increased alcohol consumption, and use of certain types of HT, are factors or markers for an elevated risk for breast cancer.⁸³ However, the association between a lifestyle modification and a change in breast cancer risk is not as clear. Nevertheless, a discussion of lifestyle characteristics associated with increased risk for breast cancer also provides “a teachable moment” for the promotion of overall health, and an opportunity to encourage women to make choices and changes compatible with a healthy lifestyle.

Breast Cancer Risks Associated With HT: The Women's Health Initiative (WHI) enrolled 161,809 postmenopausal women 50 to 79 years of age into a set of clinical trials from 1993 through 1998. Two of these trials were randomized controlled studies involving the use of HT (estrogen with/without progestin) in primary disease prevention: a trial involving 16,608 women with intact uteri at baseline randomized to receive estrogen plus progestin or placebo,²⁰¹ and a trial of 10,739 women with prior hysterectomy randomized to receive estrogen alone or placebo.²⁰² The former trial was terminated early due to evidence of breast cancer harm, along with a global index associated with overall harm. In that study, a 26% increased incidence of breast cancer was observed in the treatment group (HR, 1.26; 95% CI, 1.00–1.59). An increased incidence of abnormal mammograms was also observed for women in the WHI who received estrogen plus progestin, and was attributed to an increase in breast density.²⁰³ Of greater concern is that HT was associated with significant increase in rates of both breast cancer incidence and breast cancer–related mortality,²⁰⁴ although the increased risk for breast cancer rapidly declined following cessation of HT.²⁰⁵

An increased risk for breast cancer was not observed in the trial of women who had undergone hysterectomies and were receiving unopposed estrogen. In fact, the rate of breast cancer was lower in the group receiving estrogen relative to the placebo group, although this difference was not considered to be statistically significant.²⁰² The lower incidence of breast cancer seen among women randomized to estrogen alone during the intervention period became statistically significant with extended follow-up for a mean of 10.7 years.²⁰⁶ However, an increased incidence of abnormal mammograms was observed in the group of women receiving estrogen,²⁰⁷ as well as a doubling of the risk for benign proliferative breast disease.²⁰⁸ Analysis of the data from this randomized controlled WHI trial showed use of estrogen alone to significantly increase mammographic breast density compared with women receiving placebo; this effect was observed for at least a 2-year period.²⁰⁹ Contrary to the results from the WHI randomized controlled trials, results from several prospective, population-based, observational studies have shown use of estrogen-only HT to be associated with increased risks for breast cancer. These studies include the Black Women's Health Study where use of estrogen alone for a duration of 10 years or longer was associated with a nonsignificant increase in risk for invasive breast cancer (RR, 1.41; 95% CI, 0.95–2.10);²¹⁰ the Million Women Study of women 50 to 64 years of age, which showed an association between current use of estrogen-only HT and increased risk for breast cancer (RR, 1.30; 95% CI, 1.21–1.40; P<.0001);²¹¹ and the Nurses' Health Study, which demonstrated a significantly increased breast cancer risk after long-term use (≥20 years) of estrogen alone (RR, 1.42; 95% CI, 1.13–1.77).²¹²

It has been noted that there are important differences in the populations enrolled in the WHI randomized clinical trials relative to the women followed in the observational studies with respect to duration of exposure to HT and age at initiation of HT.²¹³ For example, many of the women in the WHI clinical trials did not start receiving HT until years after menopause, whereas those in the population-based studies were more likely to initiate HT at menopause and to have been exposed to such treatment for longer periods of time. One hypothesis put forward to explain the apparent contradictions in the summary of studies of HT described above is that short-term use of estrogen following a period of estrogen deprivation may decrease breast cancer risk by inducing apoptosis of occult breast cancer tumors, whereas long-term use of estrogen may initiate and promote the growth of new tumors, thereby increasing breast cancer risk.²¹⁴ However, further studies are needed to evaluate this hypothesis. Another possible explanation for the decrease in breast cancer risk observed in the first 2 years of the WHI randomized controlled trial of postmenopausal women receiving estrogen plus progestin may be related to HT effects on breast tissue and subsequent interference with the ability of mammography to detect new breast cancer tumors.²¹³

The use of estrogen/progestin therapy and estrogen therapy alone has also been associated with increased risk for cardiovascular disease (eg, stroke) and decreased risk for bone fractures.^201,202 However, a more recent secondary analysis from the WHI randomized controlled trials showed a trend for more effective reduction in the risk for cardiovascular disease with initiation of HT closer to menopause compared with administration of HT to women who experienced a greater time gap between menopause and the start of such therapy.²¹⁵ Nevertheless, recent results from a large French cohort control study show a significantly increased risk for breast cancer in women receiving short-term (ie, ≤2 years) estrogen and progesterone shortly after menopause when compared with nonusers.²¹⁶

The NCCN Breast Cancer Risk Reduction Panel recommends against the use of HT for women taking tamoxifen, raloxifene, anastrozole, or exemestane outside of a clinical trial.

Alcohol Consumption: Numerous studies have demonstrated that the intake of moderate amounts of alcohol (1–2 drinks per day) is associated with an increased risk for breast cancer.^83–86 A 10% increase in breast cancer risk for every 10 g of alcohol consumed each day was seen in analyses of 2 cohort studies.^48,86 A population-based study of 51,847 postmenopausal women provided evidence to support an association between increased alcohol consumption and an increased likelihood of development of ER-positive breast cancer.²¹⁷ A meta-analysis of epidemiologic studies shows a small but significant association between breast cancer and light alcohol intake (RR, 1.05; 95% CI, 1.02–1.08).²¹⁸ Even one drink per day modestly elevates breast cancer risk.⁸³ However, the effect of a reduction in alcohol consumption on the incidence of breast cancer has not been well studied.

The consensus of the NCCN Breast Cancer Risk Reduction Panel is that alcohol consumption should be limited to less than 1 drink per day. The panel has defined one drink as 1 ounce of liquor, 6 ounces of wine, or 8 ounces of beer.

Exercise: Increased levels of physical activity have been associated with a decreased risk for breast cancer.^83,219–222 For example, the effect of exercise on breast cancer risk was evaluated in a population-based study of 90,509 women between 40 and 65 years of age.²²² A relative risk of 0.62 (95% CI, 0.49–0.78) was observed for women who reported more than 5 hours of vigorous exercise per week compared to women who did not participate in recreational activities. These results are supported by another population-based, case-control study of 4538 case patients with newly diagnosed invasive breast cancer and control patients grouped according to race (eg, 1605 black and 2933 white patients). Both black and white women with annual lifetime exercise activity levels exceeding the median activity level for active control subjects were found to have a 20% lower risk for breast cancer when compared to inactive women (OR, 0.82; 95% CI, 0.71–0.93).²¹⁹ In addition, a prospective assessment evaluating the association of physical activity among 45,631 women showed the greatest reduction in breast cancer risk for women who reported walking/hiking for 10 hours per week or more (RR, 0.57; 95% CI, 0.34–0.95).²²⁰ A study of 320 postmenopausal sedentary women randomly assigned to 1 year of aerobic exercise or a control group showed modest but significant changes in serum levels of estradiol and sex hormone-binding globulin from baseline (ie, a decrease and an increase in these levels, respectively).²²³ However, it has been suggested that other, as yet unidentified, mechanisms are more likely to be responsible for the association between increased activity level and decreased risk for breast cancer.²²⁴

Diet: Results from the WHI controlled intervention trial of 48,835 postmenopausal women designed to test the effect of a low-fat diet (eg, fat intake limited to 20% of total caloric intake per day, and increased consumption of fruits, vegetables, and grains) on risk for breast cancer did not show a statistically significant reduction in the incidence of invasive breast cancer in women who followed a low-fat diet over an average of 8.1 years (HR, 0.91; 95% CI, 0.83–1.01).²²⁵ Limitations of this type of study include inherent difficulties in assuring compliance with dietary interventions, recall biases, the relatively short duration of the follow-up period, and the likelihood of insufficient differences between the 2 arms with respect to fat intake.²²⁶ Furthermore, it is possible that the impact of certain diets on breast cancer risk may be dependent on the age of the study population.^226,227 For example, results of several population-based studies have suggested that the effect of diet composition on breast cancer risk may be much greater during adolescence and early adulthood.^227,228 Nevertheless, diets in which the main sources of dietary fat are nonhydrogenated and unsaturated have been shown to have cardiovascular benefits.^227,229

Epidemiologic studies suggest that vitamin D (from dietary sources and the sun) may play a protective role with respect to decreasing risk of breast cancer development.^227,230 Furthermore, some evidence suggests that such protection is greatest in women who had more prolonged skin exposure to sunlight and higher dietary intake of sources of vitamin D during adolescence.^231,232 Current studies are in progress to evaluate the role of vitamin D on breast cancer risk.

Weight/BMI: A substantial amount of evidence indicates that overweight or obese women have a higher risk for postmenopausal breast cancer.^46,82,83

Recent results from the Nurses' Health Study evaluating the effect of weight change on the incidence of invasive breast cancer in 87,143 postmenopausal women suggested that women experiencing a weight gain of 25.0 kg or more since 18 years of age have an increased risk for breast cancer when compared with women who have maintained their weight (RR, 1.45; 95% CI, 1.27–1.66).⁴⁶ Furthermore, women who had never used postmenopausal HT and lost 10.0 kg or more since menopause and kept the weight off had a significantly lower risk for breast cancer than women who had maintained their weight (RR, 0.43; 95% CI, 0.21–0.86). Interestingly, evidence shows that the risk for breast cancer is lower in premenopausal women who are overweight compared with women who are not overweight.⁸³

Results from a case-control study of 1073 pairs of women with BRCA1/2 mutations indicated that a weight loss of 10 or more pounds in women with the BRCA1 mutation between the ages of 18 and 30 was associated with a decreased risk of developing breast cancer between the ages of 30 and 40 years. (OR, 0.35; 95% CI, 0.18–0.67).²³³

Breastfeeding: Breastfeeding has been shown to have a protective effect in many studies.^234–237 An analysis of 47 epidemiologic studies (50,302 women with invasive breast cancer and 96,973 controls) estimated that for every 12 months of breastfeeding, the relative risk of breast cancer decreases by 4.3%.²³⁵

Clinical Trials: Risk reduction counseling should include a discussion of breast cancer risk reduction interventions being studied in clinical trials.