NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Kinsinger L, Harris R, Lewis C, et al. Chemoprevention of Breast Cancer [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2002 Jul. (Systematic Evidence Reviews, No. 8.)

Cover of Chemoprevention of Breast Cancer

Chemoprevention of Breast Cancer [Internet].

Show details


In 2001, an estimated 192,200 women in the United States will be diagnosed with breast cancer, with almost 40,000 additional cases of in situ breast cancer, and 40,200 will die. 1 In addition to this morbidity and mortality, the fear of breast cancer alone is an important problem for many women in the United States.

Screening with mammography and clinical breast examination has been the most effective preventive care approach to reducing the burden of breast cancer. Even the best results of screening studies, however, show a reduction in breast cancer mortality of only about 30 percent. 2 In fact, screening may actually lead to an increase in the incidence of "breast cancer" by detecting some cases of ductal carcinoma in situ (DCIS) that would never have become clinically apparent. Thus, screening alone may not be adequate for controlling the problem of breast cancer.

The use of agents to prevent the development of breast cancer (primary prevention with chemoprevention) was suggested by trials of breast cancer treatment with tamoxifen, a compound with both estrogen-like and anti-estrogen properties. Tamoxifen has been found effective in reducing the likelihood of dying from breast cancer in women with early stage and advanced breast cancer; it has also been shown to reduce the risk of new malignancies in the opposite breast. 3

Therefore, in 1992 the National Cancer Institute (NCI) launched the Breast Cancer Prevention Trial (BCPT), a large randomized controlled trial (RCT) of tamoxifen as chemoprevention in women at high risk for breast cancer. 4 The results of this study, published in September 1998, 5 showed a 49 percent decrease in breast cancer incidence in women taking tamoxifen. Breast cancer mortality was not reduced, however. Compared with subjects in the placebo arm, subjects in the tamoxifen arm had increased rates of endometrial cancer, cataract surgery, stroke, pulmonary emboli, and deep venous thrombosis, but a decreased rate of some bone fractures.

Two other studies of tamoxifen chemoprevention to prevent breast cancer were published in July 1998.6,7 Although smaller than the BCPT and designed in slightly different ways, neither of these trials showed a large reduction in breast cancer incidence in women taking tamoxifen.

Given these conflicting data -- large reduction in risk for breast cancer in one but not other trials and substantial adverse effects -- the need to review rigorously the literature on chemoprevention of breast cancer to provide guidance to physicians and patients for decision making is clear. In this review, we examine the evidence of benefits, harms, and costs associated with chemoprevention of breast cancer. We will focus primarily on evidence concerning tamoxifen, as it is the most intensively studied drug. We will also examine evidence for raloxifene, a similar drug that may have some theoretical advantages over tamoxifen, and several other drugs that have also been considered.


Burden of Suffering

Excluding skin cancer, breast cancer is the most common cancer among women in the United States, accounting for 29 percent of new cancer cases in women each year. Only lung cancer accounts for more cancer deaths in women. Breast cancer is much less common in men, with an estimated 1,500 cases and 400 deaths expected in 2001, thus accounting for about 1 percent of breast cancer incidence and mortality. 1 These estimates are derived from statistical forecasting models; they use data for the number of new cases from the NCI Surveillance, Epidemiology, and End Results (SEER) database and data on the number of deaths from the National Center for Health Statistics at the Centers for Disease Control and Prevention.

The incidence of breast cancer in U.S. women increased 25 percent between 1973 and 1996, about 1.5 percent per year. 8 Most of this rise occurred in women over the age of 50 years. The sharpest rate of increase took place in the 1980s, corresponding to an increased use of mammography screening.8,9 Since 1992, incidence rates have been nearly level. According to NCI data from 1996, 111 women per 100,000 (adjusted for age) were diagnosed with invasive breast cancer. Incidence rates vary slightly by race (113/100,000 for white women vs 100/100,000 for black women) and significantly by age (75/100,000 for women under 65 vs 441/100,000 for women 65 years and older). 8 Seventy-seven percent of new cases and 84 percent of breast cancer deaths occur in women 50 years of age and older. 8

Death rates from breast cancer remained relatively stable from 1973 until the early 1990s when they began to decline modestly. This trend occurred only for white women, not for black women. In 1996, 24 white women per 100,000 died from breast cancer, whereas the rate for black women was 31 deaths per 100,000 women. 8 Women younger than age 50 years suffered 5 deaths per 100,000, whereas women ages 50 years and older accounted for 83 deaths per 100,000. 8 The decline in breast cancer mortality has been attributed both to improvements in breast cancer treatment and to the benefits of early detection through screening.

With widespread publicity about breast cancer, many women now overestimate their risk of developing the disease. A survey of women between 30 and 49 years of age in waiting rooms of 20 North Carolina primary care practices found that about 45 percent perceived their risk of breast cancer in the next 10 years as somewhat or very high, despite the fact that only 12 percent of the women had a family history of breast cancer. 10 Nearly half (48 percent) responded that they were moderately or very worried about breast cancer. Few women knew that breast cancer risk increases with age.


Effective prevention strategies for breast cancer need to be directed to women at risk for the disease. Table 1 lists more than 20 risk factors in order of the strength of their association. Many of these known risks are not easily modifiable, and they account for less than half of breast cancer incidence. Thus, primary prevention by controlling these factors will have limited impact. Additionally, the majority of women with one or more risk factors do not ever develop breast cancer.

Table 1. Risk Factors for the Development of Breast Cancer*.


Table 1. Risk Factors for the Development of Breast Cancer*.

The single-most important risk factor for developing breast cancer, other than being female, is increasing age. The risk of developing breast cancer over a woman's lifetime in the United States is 12.6 percent, a probability of one in eight. 8 However, most of this risk occurs in the later years of a woman's life; the likelihood of developing breast cancer over the next 10 years is a more relevant concern. For women in their forties, that probability is one in 67 (1.5 percent) and it increases to one in 25 (4.0 percent) for women in their seventies. 8

Family history and genetics can confer a high risk for developing breast cancer. Women with two or more first-degree relatives (mother, sister, daughter) with breast cancer diagnosed at an early age are more than four times as likely to develop breast cancer as women without affected family members. 11 Between 5 percent and 10 percent of breast cancer in the United States is attributable to inherited genetic mutations in one of two major breast cancer susceptibility genes (BRCA1, BRCA2). Women with these genetic changes are at significantly increased risk for both breast and ovarian cancer. Fortunately, less than 1 percent of women in the general population appear to carry these alterations. 8

Reproductive hormones, especially estrogen, appear to be important factors in the development of breast cancer. Up to a doubling of risk is seen in women with early onset of menarche (age 11 years or younger), late age at menopause (age 55 years or older), older age at first pregnancy (age 30 years or later), and few or no pregnancies. 12 These factors are all measures of the cumulative exposure of the breast to endogenous estrogen, which may increase the rate of cell division and mitotic activity in the breast, thus leading to a greater likelihood of DNA damage. Another theory of breast cancer carcinogenesis proposes that proliferating breast tissue is more susceptible to environmental hazards, such as ionizing radiation. 13 Reduced exposure to estrogen through premenopausal removal of ovaries or to radiation lessens the risk of breast cancer by 75 percent. 14

Levels of endogenous sex steroids may be associated with risk of breast cancer in postmenopausal women. Investigators from the Nurses' Health Study, 15 the Study of Osteoporotic Fractures, 16 and a large cohort study of New York City women 17 have all found a two- to fourfold increased relative risk of breast cancer incidence in women with the highest circulating levels of estrogen, compared to women with the lowest levels. The relationship between bone mass and breast cancer risk among postmenopausal women is similar. The Study of Osteoporotic Fractures 18 and the Framingham study 19 found relative risks of 2.7 and 3.5, respectively, for women in the highest quartile of bone mass, compared with those in the lowest quartile.

Exogenous estrogen also appears to play a role in breast cancer development. Long-term use of oral contraceptives may slightly increase the risk for breast cancer in younger, but not older, women. Evidence from many studies suggests that long-term (more than 10 years) use of postmenopausal hormone replacement therapy is associated with a small increased risk for breast cancer. 11 Three recent updates of large cohort studies have found a greater increased risk with combined estrogen-progestin hormone therapy, compared with estrogen alone.20-22 Thus, the evidence supports the hypothesis that the development of breast cancer depends on estrogen stimulation and that blocking this action may inhibit tumor growth. 14

Given the well-established role of estrogen and progesterone receptors in breast cancer tissue for predicting response to hormonal therapy and assessing prognosis, several studies have examined the role of these receptors in determining risk for breast cancer. Whether different receptor states represent different stages of one type of breast cancer or actually different types of breast cancer is not clear. About two-thirds of breast cancer cases are both estrogen receptor (ER) /progesterone receptor (PR) positive and one-sixth are both ER/PR negative; the remaining one-sixth of cases are discrepant on receptor status. 23 Older women are more likely to have ER+ tumors.

Studies analyzing the relationship between ER status and known risk factors (parity, family history, and history of benign breast disease) have not found consistent differences between ER+ and ER- cancers. The New York Women's Health Study found that, in postmenopausal women, the association of endogenous estrogen with breast cancer risk was independent of ER status of the tumor. The authors concluded that their results added weight to the hypothesis that breast cancers progress from ER+ to ER- as they evolve, rather than representing different types of cancer. 24

An observational study of both ER and PR receptor status among a cohort of women in Iowa showed that ER+/PR+ tumors were more closely associated with endogenous hormone exposure variables (parity, age at first birth, age at menarche, body mass index, and body fat distribution) than ER-/PR- cancers, which had the opposite pattern. 23 The conclusion from this study was that breast cancer might be of several different types, with separate etiologies, responses to therapy, and possibly prevention strategies.

Demographic factors associated with increased risk include being born in Northern Europe or North America (compared with Asia or Africa), high socioeconomic status, never having been married, and, for breast cancer diagnosed after age 45 years, being white. 11 Other characteristics that carry an elevated risk are a history of having had breast biopsies, proliferative breast disease on biopsy, and extensive nodular densities on mammograms. Risk factors for a modestly increased risk include alcohol intake of two or more drinks per day25,26 and possibly physical inactivity.27-30

Taken together, the well-established risk factors of later age at first birth or nulliparity, higher income, and family history of breast cancer appear to account for about 41 percent of breast cancer in the United States. 31 A study that estimated population attributable risks for breast cancer found that either later age at first birth or nulliparity was associated with about 30 percent of all breast cancer. The authors estimated that adding in the effects of earlier age of menarche and prior benign breast disease would add less than 15 percent to the overall population attributable risk. 31

Knowing and understanding the risk factors for breast cancer should aid in devising prevention efforts. However, most risk factors identified to date are not easily modified and so play a limited role in strategies to prevent the onset of this serious disease.

Health Care Interventions to Prevent Breast Cancer Occurrences or Deaths

Early Diagnosis with Screening

The effectiveness of mammography screening to reduce breast cancer mortality has been extensively studied. 2 Studies in both Europe and North American have consistently found about 30 percent fewer deaths from breast cancer in women ages 50 to 69 years who had regular (every one to two years) screening with mammograms alone or in combination with clinical breast examinations than women without screening. Screening for older women has not been as well studied, but it is likely to be effective, at least for women who have a life expectancy of 5 to 10 years and few other serious medical problems. Screening mammography in women ages 40 to 49 years has been quite controversial, but estimates suggest a modest benefit (about 15 percent reduction in breast cancer mortality).32,33

Mammography is also associated, however, with a significant burden of psychological harms because of the high false-positive rate, leading to additional studies and evaluation, especially in younger women.34,35 Breast self examinations have not been shown to reduce the chance of dying from breast cancer in women of any age. 36

Primary Prevention

Most individual risk factors for breast cancer are associated with modest elevations in risk, and few can be easily modified by environmental or behavioral changes. For example, although alcohol intake of two or more drinks per day increases a woman's risk for breast cancer, it accounts for only a small percentage of breast cancer risk overall (about 2 percent). Reduction in alcohol intake would likely have little impact on the number of breast cancer cases or deaths. 26 Weight gain after age 18 has been shown to be associated with a doubling of the risk of postmenopausal breast cancer among women who had never used hormone replacement therapy; 37 therefore, avoiding such weight gain may be important strategy to decrease risk. Prospective studies of the effectiveness of regular exercise, a lower fat diet, or dietary intake of fruits and vegetables on breast cancer incidence or mortality have shown either inconsistent or negative results.38,39 Increased physical activity and healthy eating should be encouraged for many health reasons, but improvements in these areas would likely have a small effect on breast cancer. The Women's Health Initiative, a large study from the National Institutes of Health involving approximately 70,000 postmenopausal women, includes a randomized trial of dietary modification on risks for several diseases, including breast cancer. Results from this study will not be available for several years and may not be conclusive even then, given the weak evidence for a relationship between diet and breast cancer. 11

Another approach to breast cancer prevention focuses on the hormonal milieu, which is thought to play an important role in the development of breast cancer through the proliferation of breast tissue. This process is likely to occur over a long period of time, thus creating a large window of opportunity for interventions to interfere with the promotion phase of cancer. Support for the use of hormone-based prevention strategies comes from the presence of estrogen and progesterone receptors in breast cancer tissue, the response of cancers to hormonal therapy, and risk factors related to time of estrogen exposure (age of menarche, age at first birth, age at menopause, and so on). 13

Tamoxifen, a nonsteroidal triphenylethylene derivative first synthesized in 1966 as a contraceptive agent, is now widely used to treat women with metastatic breast cancer and as adjuvant therapy in early stage breast cancer to prevent recurrences and new cancers in the opposite breast. These uses take advantage of tamoxifen's anti-estrogen properties in breast tissue. These properties place tamoxifen in the category of drugs known as selective estrogen receptor modulators (SERMs). Tamoxifen decreases cellular proliferation by binding directly to estrogen receptors and by interfering with growth factors and other proteins critical to proliferating cells. 13 Tamoxifen also has estrogen-like effects on other tissues, notably the endometrium and bones.

Numerous studies over the past 15 years have substantiated tamoxifen's effectiveness, either as a single agent or in combination with chemotherapy, for the treatment of advanced breast cancer and as adjuvant therapy in early stage disease. In 1998 the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) published a meta-analysis of 55 RCTs for early breast cancer, involving more than 37,000 women worldwide. 40 Results for the approximately 18,000 women with ER+ tumors showed that five years of tamoxifen therapy was associated with a 50 percent reduction in disease recurrence (P < 0.00001) and a 28 percent reduction in mortality (P < 0.00001) over a follow-up period of 10 years. Tamoxifen appeared to have little effect on ER- cancers. Additionally, the analysis found a 47 percent decrease in contralateral breast cancer (93 vs 159 breast cancer events, P < 0.00001), regardless of estrogen receptor status. Thus, these studies suggest that, for women who are at increased risk for new cancers in the opposite breast, tamoxifen reduces the incidence of such tumors by about half.

Recently, tamoxifen has also been found to reduce significantly the occurrence of invasive breast cancer in women with DCIS. 41 Compared with women treated only with lumpectomy and radiation therapy, women who were treated with lumpectomy, radiation therapy, and tamoxifen had a 37 percent reduction in invasive and noninvasive breast cancer events (84 vs 130; relative risk [RR], 0.63; 95% confidence interval [CI], 0.47-0.83). Contralateral breast tumors were decreased by 52 percent in the tamoxifen group (RR, 0.48; 95% CI, 0.26-0.87).

Because of its estrogen-like activity, tamoxifen may have beneficial effects on bones and lipids. Long-term studies with important health outcomes are not yet available, but breast cancer treatment trials have found improved lipid profiles and bone density measurements. Serious side effects of tamoxifen, however, have also been seen, including increased rates of endometrial cancer and venous thromoembolic events. These findings raise concerns about the safety of tamoxifen as a chemopreventive agent in healthy women.

Another drug that has been studied for its effect in breast cancer prevention is a more recently developed SERM, raloxifene, a nonsteroidal benzothiophene derivitive that binds to the estrogen receptor. Raloxifene has an estrogen-like effect on the skeletal system and has been approved for the prevention of osteoporosis in postmenopausal women. Unlike tamoxifen, it does not appear to increase the risk of endometrial cancer. 42

Other chemopreventive agents for breast cancer have been proposed but not yet fully studied. One suggested approach involves the use of luteinizing hormone-releasing hormone agonists in young women to cause a chemical premature menopause, given that premenopausal women who undergo ovariectomy have a decreased risk of breast cancer. 13 Clearly, this strategy entails alterations of a complex hormonal balance and would need strong evidence of greater benefits than harms to justify its use. Vitamin A analogs have also been considered for breast cancer prevention. Fenretinide, a synthetic retinoid, has been studied for the prevention of contralateral breast cancer in Italy but was not found to be effective. 43

Prior Recommendations and Recommendations of Other Groups

The U.S. Preventive Services Task Force (USPSTF) did not consider the topic of chemoprevention of breast cancer in previous editions of the Guide to Clinical Preventive Services.

The American Cancer Society recommends that women who are at increased risk of breast cancer should discuss the use of tamoxifen with their doctors. 8 The American College of Obstetricians and Gynecologists advises that "the decision to use tamoxifen to reduce the risk of breast cancer should be individualized" and recommends considering breast cancer risk, potential tamoxifen-related side effects and risks, patient preferences, clinician judgment, and clinician ability to monitor for and manage tamoxifen-related side effects. 44 The American Society of Clinical Oncology recently published a technology assessment on the use of tamoxifen and raloxifene as breast cancer risk reduction strategies. After reviewing the published literature, they concluded that "for women with a defined five-year projected risk of breast cancer of >1.66 percent, tamoxifen (at 20 milligrams/day [mg/d] for up to five years) may be offered to reduce their risk. On the basis of available information, use of raloxifene should currently be reserved for its approved indication to prevent bone loss in postmenopausal women." 45

In June 2001, the Canadian Task Force on Preventive Health Care published its recommendations. 46 The Canadian Task Force determined that there is "fair evidence to recommend against the use of tamoxifen to reduce the risk of breast cancer in low or normal risk women (<1.66% 5-year estimated risk) (D)" and "fair evidence to recommend counseling high risk women (>1.66% 5-year risk) about the potential benefits and risks of using tamoxifen to reduce the likelihood of breast cancer and hence support individual choice (B)." 46

An NCI workshop in July 1998 aimed to develop tools to aid physicians and patients in weighing the pros and cons of using tamoxifen. Following the workshop, Gail et al. 47 published an article extensively describing the risks and benefits of tamoxifen chemoprevention. They reported that tamoxifen is most beneficial for younger women with an elevated risk of breast cancer. They suggested, however, that decisions not be based on a single number representing a projected five-year risk of invasive breast cancer but rather should weigh the various risks and benefits of tamoxifen as they apply to individual women.

Analytic Framework and Key Questions

The RTI-University of North Carolina Evidence-based Practice Center (RTI-UNC EPC), together with members of the USPSTF, sought to clarify issues concerning chemoprevention to prevent breast cancer by performing a systematic review of the relevant scientific literature on this topic. This systematic evidence review (SER) provides the background for the first recommendation ever made by the USPSTF on this topic.

Analytic Framework

This SER examines the evidence for chemoprevention to prevent breast cancer among women who have never had breast cancer. Figure 1 presents a comprehensive analytic framework for this topic.

Figure 1. Chemoprevention of Breast Cancer: Analytic Framework.


Figure 1. Chemoprevention of Breast Cancer: Analytic Framework.

The analytic framework begins on the left side with the population at risk. Several different populations should be considered:

  1. Premenopausal women with "average" risk of breast cancer.
  2. Premenopausal women with "high" risk of breast cancer.
  3. Postmenopausal women with "average" risk of breast cancer.
  4. Postmenopausal women with "high" risk of breast cancer.

In addition, because of the effect of tamoxifen or other chemopreventive agents on potentially serious conditions other than breast cancer, women with particularly increased (or decreased) risk of thrombotic events, bone fractures, or endometrial cancer should be considered separately.

Moving to the right in the analytic framework, the "chemoprevention" arrow points to a box labeled "incidence of breast cancer." Moving downward from the chemoprevention arrow is an "adverse effects/costs" arrow. Some of the adverse effects of tamoxifen or raloxifene are deep vein thrombophlebitis, pulmonary embolus, stroke, endometrial cancer, cataracts, and psychological effects.

Possible benefits from tamoxifen or raloxifene chemoprevention include reduced risk of heart disease and bone fractures. These are listed together in a separate box: "other beneficial effects."

Farther to the right is a "treatment" arrow, leading to a box labeled "health outcomes," which includes both mortality and morbidity from breast cancer. We also consider "incidence of breast cancer" to be a health outcome worthy of consideration in its own right.

Leading downward from the "treatment" arrow is another "adverse effects/costs" arrow. Included here are adverse effects from chemotherapy, surgery, and radiation therapy used to treat breast cancer.

Finally, an "overarching" arrow for chemoprevention leads directly to reduced mortality and morbidity from breast cancer.

Key Questions

The primary, overarching question, shown as the topmost line in Figure 1, is the following:


Does tamoxifen (or other chemopreventive agents), among women without a previous diagnosis of breast cancer, reduce mortality from breast cancer? (If so, by how much?)

The secondary questions (i.e., linkage questions), shown in the body of Figure 1, are the following:


Does tamoxifen (or other chemopreventive agents) reduce the incidence of breast cancer? (If so, by how much?)


Does tamoxifen (or other chemopreventive agents) have other beneficial effects (e.g., reduce the incidence of or mortality from heart disease, reduce the incidence of specific fractures, or produce psychological benefits)? (If so, by how much?)


To what extent does tamoxifen (or other chemopreventive agents) increase the risk of adverse effects (e.g., thrombotic events, endometrial cancer, or psychological effects)?


What are the costs associated with chemoprevention of breast cancer?


Does treatment for breast cancer effectively reduce disease-specific mortality?


What are the side effects (type and magnitude) of treatment of breast cancer?


What are the costs (type and magnitude) of treatment of breast cancer?

No clinical trial has been large or long enough to examine the impact of chemoprevention on mortality from breast cancer (Key Question 1). In addition, the effectiveness of treatment for breast cancer is clear and has been examined in a continuing rigorous meta-analysis.3,40 The adverse effects of treatment are also well documented. 48

Therefore, this review focuses on Key Questions 2 through 5. We consider in these questions evidence about issues of dose and duration of chemoprevention. Key Questions 6 through 8 are included in the analytic framework for conceptual completeness but are not further addressed in this review.


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...