Breast cancer is the most frequently diagnosed cancer among American women (217,000 new cases in 2005), and incidence rates continue to rise (1). Approximately 39,000 women in the United States die of breast cancer each year, and an estimated 2.3 million breast cancer survivors in the United States alone are living with the long-term and late effects of the disease (1). Although advances in therapy have led to improvements in survival in recent years, new therapies are costly, are associated with significant side effects, and may benefit only subsets of those with breast cancer. Alternative approaches are needed to help diminish the morbidity and mortality of breast cancer, as well as its cost to society.
Multiple studies during the past 20 yrs, coming from North America, Europe, Asia, and Australia, have demonstrated that women who exercise regularly have a decreased risk of developing breast cancer compared with sedentary women (6). This association has been observed in various racial/ethnic minorities as well as in non-Hispanic white women (6). The evidence for the association between higher levels of physical activity and lower risk of breast cancer has been classified as "convincing," with the degree of protection estimated at about 30 to 40% with 3 to 4 h·wk−1 of moderate- to vigorous-intensity physical activity (6). Furthermore, recent findings from the Nurses' Health Study cohort and the Health, Eating, Activity and Lifestyle Study show that women who are physically active after a breast cancer diagnosis are at lower risk of a recurrence and death caused by breast cancer (8,9). Numerous observational studies also have demonstrated that obesity and weight gain adversely affect primary and secondary breast cancer risk, adding further evidence to the hypothesis that physical activity, one of the critical components of energy balance, influences breast cancer risk and prognosis (2,8,9).
Although observational studies have provided an important base of evidence for inferring that physical activity has a protective effect against breast cancer development, these studies cannot, by definition, show a protective "effect" of physical activity against breast cancer development. At this point, clinical trials are needed to (i) determine whether the independent effect of physical activity prevents the primary and secondary occurrence of breast cancer and (ii) elucidate the biological mechanisms by which physical activity protects healthy women from developing breast cancer and breast cancer survivors from experiencing recurrence and breast cancer-related mortality. To date, no human intervention studies have examined the effect of physical activity on primary and secondary breast cancer prevention. Conducting such a trial presents many challenges, including the need for a very large sample size, long study duration, and participation from multiple research institutes (Table 1). Nevertheless, the Women's Health Initiative Trial provides a model for such a study and suggests that such an undertaking is feasible (15). In addition, two large National Institutes of Health-funded multicenter clinical trials, the Women's Healthy Eating and Living Study and the Women's Intervention Nutrition Study, both of which are testing the effect of a change in dietary composition on recurrence and survival in breast cancer survivors, further demonstrate the feasibility of lifestyle intervention trials in this population (3,14).
Based on the epidemiological data and current understandings of how physical activity affects several specific biological processes involved in the primary and secondary occurrence of breast cancer, we hypothesize that physical activity has a direct causal role in breast cancer incidence, recurrence, and death and argue that randomized controlled physical activity trials are required to test this hypothesis, despite the challenges and limitations of this study design.
BIOLOGICAL MECHANISMS/INTERMEDIATE ENDPOINTS
A number of potential physiological responses to physical activity have been postulated to be intermediate markers or endpoints in the development and progression of breast cancer (2). The evidence for each of these effects is summarized in Table 2. We hypothesize that exercise may reduce the risk of breast cancer occurrence/recurrence through a reduction in fat mass, leading to a more beneficial metabolic and sex hormone profile in terms of breast cancer risk, as depicted in Figure 1. Other mechanisms for a protective effect of exercise on breast cancer risk could exist. For example, exercise may improve immune status, which could in turn decrease risk for breast cancer.
Obesity and Weight Control
Of the possible biological mechanisms mediating an association between physical activity and breast cancer, body fat, or weight control may be most important (2,10,12). Maintenance of normal body weight throughout a woman's adult years is one of the few known modifiable risk factors for breast cancer, and many studies have shown positive associations between obesity and breast cancer risk (2). However, the influence of obesity on breast cancer risk varies by menopausal status: obesity has a protective effect against breast cancer in premenopausal women but is associated with increased risk in postmenopausal women. The biologic rationale for this difference in effect of obesity on breast cancer risk is based on the source of endogenous estrogen before and after menopause (2). Although menopausal status may modify the effect of obesity on breast cancer risk, recent studies have shown a positive association between weight gain during premenopausal or postmenopausal years and breast cancer risk (2). Epidemiological studies have also shown that premenopausal and postmenopausal women who are overweight or obese when they are diagnosed with breast cancer are more likely to experience a recurrence or die of breast cancer than women who are of a normal weight (2). Furthermore, some, but not all, studies suggest that women who gain weight after breast cancer diagnosis, regardless of menopausal status, are at increased risk for breast cancer recurrence and death as compared with women who maintain their weight after diagnosis (2). This is especially worrisome given the fact that most women who are treated for breast cancer gain a significant amount of weight in the year after breast cancer diagnosis, and return to prediagnosis weight rarely occurs (10).
Physical activity and obesity are highly interrelated. Obese individuals generally exhibit a lower level of habitual physical activity than the nonobese, and physically active individuals may be less likely to become obese (2). In the treatment of obesity, adding exercise to a calorie-reduction program enhances weight loss, and regular exercise is a powerful predictor of long-term maintenance of weight loss. No randomized trials have examined the effect of weight loss, established via increases in physical activity and/or dietary changes, on primary and secondary breast cancer risk. Such trials are necessary to (i) establish a causal relationship between weight loss and breast cancer risk and (ii) better understand the mechanisms mediating the relationships among physical activity, weight loss, and breast cancer risk (e.g., does physical activity with or without weight loss decrease the risk of primary or secondary breast cancer?).
Insulin and Insulinlike Growth Factors
Physical activity and obesity also both influence circulating concentrations of insulin (2,7) which, in turn, may affect breast cancer risk and prognosis (2,7). In a study of women with early-stage breast cancer, higher insulin levels were associated with a two and three times higher risk of recurrence or breast cancer death, respectively (7), and in preclinical studies, insulin has a mitogenic effect in normal breast tissue and can stimulate growth of breast cancer cell lines. High insulin levels produce an increase in circulating insulinlike growth factor I (IGF-I) and a decrease in IGF-binding proteins (increasing the availability of IGF present). Insulinlike growth factor I is thought to have a major role in promoting breast cancer (2). However, the few trials that have assessed the effect of exercise on IGFs have had variable results (2). Randomized controlled trials testing the effect of exercise on insulin and IGFs are needed to elucidate the biological mechanisms by which physical activity protects against the development of primary or secondary breast cancer.
Obesity, a high insulin level, and altered IGF levels are also associated with a less favorable sex hormone profile (2). Sex steroid hormones have powerful mitogenic and proliferative influences and are strongly associated with the development of breast cancer (2,5). A recent pooled analysis of nine cohort studies showed that the risk for breast cancer in postmenopausal women increased significantly with increasing concentrations of estradiol, free estradiol, and estrone (5). Women who were in the highest quintile for these sex hormones had a twofold increased risk for breast cancer compared with women in the lowest quintile. Other findings from epidemiological studies further support the etiologic role of estrogen in breast cancer, showing that breast cancer risk is associated with early menarche, late menopause, low parity, and use of exogenous estrogens, all of which are linked to prolonged or extensive exposure of breast tissue to estrogen stimulation (2,5). Finally, a number of clinical trials show that estrogen ablation increases survival after a diagnosis of breast cancer. Changes in sex hormones are perhaps the most consistently cited potential mechanism for the association between physical activity and breast cancer.
Girls who participate in athletics tend to have a later age of menarche and a delay in establishing normal ovarian cyclicity. Later age of menarche and slowed establishment of cycling would decrease the total steroid hormone exposure to the breast (2). In adult premenopausal women, exercise has been associated with decreased levels of circulating estrogen and progesterone, shortened luteal phase, increased frequency of anovulation, and an increased incidence of oligomenorrhea and amenorrhea (2). In postmenopausal women, physical activity has been found to be associated with decreased serum estrogens and androgens (2,13). Increased physical activity also has been associated with increased sex hormone-binding globulin resulting in lower amounts of free active sex hormones in circulation (2,13).
The primary mechanism of physical activity influencing sex hormones in postmenopausal women is via decreased body fat, a substrate for estrogen, and testosterone production, which results in less tissue capable of aromatization of the adrenal androgens to estrogens. Only one randomized controlled exercise trial has been published examining the effect of exercise on sex hormone concentrations (13). Although an overall effect of exercise was significantly associated with decreased sex hormone concentrations in healthy postmenopausal women, a stronger effect was observed among women who lost body fat with exercise compared with women who did not lose body fat with exercise. Because of the paucity of randomized controlled trial data, it is not yet established that change in physical activity can affect sex hormone concentrations independently of an effect of physical activity on adiposity. More randomized trials are required to determine the associations among physical activity, body fat, sex hormones, and breast cancer risk and prognosis.
Changes in immune function may also mediate the relationship between physical activity and breast cancer risk and prognosis (2). The immune system is thought to play a role in protecting against breast cancer by recognizing and eliminating abnormal cells. A growing literature of small exercise intervention studies shows that physical activity improves immune function, both functionally and numerically (2). Physical activity appears to enhance proliferation of lymphocytes, increases the number of natural killer cells and increases lymphokine-activated killer cells activity.
Other Biological Mechanisms/Intermediate Endpoints
Other intermediate endpoints have been proposed, such as mammographic density (2,11), adipocytokines (tumor necrosis factor α [TNF-α], leptin, and adiponectin) (2), and oxidative stress (2). Strong evidence exists that the characteristics of breast tissue as seen on a mammogram, measured as mammographic density, provide information about breast cancer risk. Women with high levels of mammographic density have a fourfold to sixfold greater risk of developing breast cancer than women with lower levels of mammographic density; thus, mammographic density is a stronger predictor of breast cancer risk than most traditional risk factors. Mammographic density reflects proliferation of the breast epithelium and stroma, in response to growth factors induced by current and past circulating sex hormone levels. Mammographic density may vary throughout lifetime, with the pattern reflecting the accumulated breast cancer risk at the time the mammogram was obtained. Factors that change mammographic density may also change breast cancer risk. Physical activity may influence mammographic density by favorably changing certain hormones associated with mammographic density and breast cancer risk. Both mammographic dense area and percent density have been found to be inversely related to physical activity in obese postmenopausal women (2,14).
Adipocytokines exhibit strong associations with body mass index, abdominal fat mass, and hyperinsulinemia. In addition, several adipocytokines, including interleukin 6 (IL-6), TNF-α, and leptin, promote angiogenesis, which is essential for breast cancer development and progression and can stimulate estrogen biosynthesis by the induction of aromatase activity. C-reactive protein is not an adipocytokine per se, but its production is promoted by TNF-α and IL-6. C-reactive protein is a well-known systemic marker for inflammation that is produced by the liver and is only present during episodes of chronic inflammation. Thus far, varying effects of physical activity for the different adipocytokines have been observed (2). For postmenopausal women, the evidence most strongly supports physical activity decreasing circulating leptin, IL-6, and C-reactive protein. The evidence is mixed on TNF-α, and no studies have yet found an association between exercise and adiponectin.
Lastly, reactive oxygen species (i.e., free radicals) can play a significant role in breast cancer via their ability to produce DNA damage as well as damage to other cellular components which interact with DNA (2). Acute exercise may promote free radical production, whereas chronic exercise improves free radical defenses by up-regulating both the activities of key free radical scavenger enzymes and levels of antioxidants (2). To date, there are few studies that have examined reactive oxygen species-related damage or relevant antioxidant enzymes in the context of exercise in a cancer model.
With numerous publications showing statistically and clinically significant associations between obesity, fasting insulin levels, sex hormones, and breast cancer risk, recurrence, and death, more effective treatment strategies to reduce body fat, insulin, and sex hormone levels in healthy women and breast cancer survivors should be explored. Although much further research is needed to determine the effect of physical activity on these intermediate endpoints/biomarkers, we posit that physical activity decreases primary and secondary breast cancer risk directly and indirectly through multiple interrelated actions of body fat levels and/or hormonal concentrations and actions. Only randomized controlled exercise trials are able to determine whether physical activity directly prevents the primary and secondary occurrence of breast cancer and what the biological mechanisms are by which physical activity protects against primary or secondary breast cancer.
RANDOMIZED TRIALS OF EXERCISE AND INTERMEDIATE ENDPOINTS/BIOLOGICAL MECHANISMS
Because of the cost and logistical difficulties involved with a trial that has breast cancer as the outcome, many breast cancer researchers have advocated for trials of short-term responses that may be early disease markers (e.g., atypical benign breast disease before development of in situ or invasive breast cancer) or, more specifically, biomarkers, such as serum sex hormones, which can be measured in fresh or stored biological specimens. To the extent that these intermediate endpoints and biomarkers are valid and reliable measures of breast cancer risk, trials that examine the effects of exercise on these endpoints provide important clues about the causal pathway between physical activity and breast cancer occurrence, recurrence, and death. The few existing exercise interventions in humans that have examined these effects are summarized in Table 3. Despite the positive results of the intermediate outcome trials discussed in Table 3, favorable intermediate outcome changes may or may not convey meaningful benefits in terms of disease prevention. Thus, randomized trials with disease occurrence outcomes are still considered the gold standard for testing breast cancer prevention hypotheses.
DESIGNING AN EXERCISE TRIAL FOR PRIMARY OR SECONDARY BREAST CANCER PREVENTION
Because randomized controlled exercise trials with breast cancer prevention outcomes are very costly, only a few trials can be conducted. Consequently, a well thought out design is needed before initiating a large-scale, multisite prevention trial. Table 4 lists essential steps needed for the successful delivery of exercise and breast cancer trials. We discuss some of the essential steps in more detail below.
A team of investigators must be assembled representing a variety of scientific disciplines such as exercise science, epidemiology, and oncology. Breast cancer survivors should also be active participants in the collaborative team. Preliminary/pilot data from exercise trials in healthy women or breast cancer survivors are necessary.
A two-arm comparison of exercise versus control or two alternative types of exercise are necessary. Given the established benefits of physical activity for cardiovascular disease, the ethics of having a nonexercise control group may be questioned. Researchers also need to consider ways to make randomized assignment palatable for participants, including use of a delayed intervention or alternative intervention (e.g., health education or two different exercise prescriptions).
Recruitment and Eligibility Criteria
Because randomized trials are expensive and require a high level of staff involvement, it is most efficient to recruit individuals at high risk for developing breast cancer or breast cancer survivors. Recruitment may take 2 to 4 yrs, and approximately fewer than 10% of those approached may eventually be randomized. Pilot data on recruitment rates are critical before initiating a full-scale trial, as well as a plan for how participants will be recruited. A population-based approach is ideal; however, this approach may need to be supplemented with physician referrals and/or mass media. Whatever recruitment approached is used, demographic, behavioral, and physiological data need to be collected on all participants during a screening visit (phone call or mailing) to compare characteristics between women eventually randomized and women not randomized to the study (either because of ineligibility or lack of interest). An ethnically/racially diverse sample must be recruited that is at least similar to the ethnic/racial distribution in the general population of the study areas. To observe a maximal effect of the exercise intervention on the disease outcome, only physically inactive women would be eligible. Other eligibility criteria worth considering are age, menopausal status, body mass index, and breast cancer disease stage (if studying survivors) (Table 5).
Because of differences in incidence rates and outcomes being assessed, a trial needs to be conducted separately for healthy women and breast cancer survivors. A trial among healthy women will require a significantly larger sample size and longer study duration to examine the effect of exercise on preventing primary breast cancer compared with a trial in breast cancer survivors on preventing secondary breast cancer. However, both trials are necessary to determine the optimal dose of exercise necessary to prevent either primary or secondary breast cancer.
Choosing an appropriate exercise intervention is critical for a study's success. If null results are observed, it may be difficult to distinguish an intervention's lack of effect from inadequate exercise training or loss of power from poor adherence. Similarly, differential adherence reporting between intervention and control groups could lead to inaccurate recommendations following a positive trial result. A number of different types of exercise interventions could be implemented. A supervised program has advantages over a "lifestyle" or home-based program in that participants' adherence to exercise can be directly observed. However, home-based programs may result in better long-term adoption and maintenance of exercise, especially if the participants find cost-effective ways to incorporate exercise into their daily routine. Another approach could involve an exercise intervention consisting of a combined home exercise and supervised program. Women randomized to exercise may be taught exercise techniques and principles in an initial in-person visit at a local health club with the exercise trainer and then weekly via the telephone. A telephone counseling protocol could be the principal instrument used to promote exercise adherence in the exercise group, as was done in the Women's Health Initiative, Women's Intervention Nutrition Study, and Women's Healthy Eating and Living Studies (3,14,15). In addition to the telephone counseling and in an effort to maximize adherence to the exercise intervention and provide more variety or different options for exercising, participants could also be offered access to a health club, allowing them to exercise during inclement weather and in the evenings. The successful Diabetes Prevention Program protocol required that each clinical center offer supervised exercise sessions at least two times per week throughout the trial (4).
The type of exercise prescribed depends on the outcome. For primary or secondary breast cancer prevention, observational data recommend 3 to 4 h·wk−1 of moderate-intensity aerobic activity such as brisk walking. However, until this dose of exercise is tested using a randomized design, the optimal amount of exercise necessary to decrease breast cancer risk is unknown. The amount of exercise prescribed in other exercise trials (e.g., randomized trials of exercise and breast cancer intermediate endpoints or cardiovascular disease) may provide some guidance as to what dose to prescribe. Although many questions remain concerning the optimal duration, frequency, intensity, and type of exercise (e.g., aerobic exercise vs strength training), as well as when (e.g., after completing breast cancer treatment) and for how many years to intervene, no doubt remains that only randomized controlled exercise trials will provide direct answers as to the recommended amount of physical activity necessary for primary and secondary breast cancer prevention. However, a limitation of randomized trials is that only one or two exercise prescriptions can be prescribed per trial.
Assessment of Physical Activity and Covariates
To assess accurately the effect of physical activity on breast cancer risk and prognosis, reliable, valid, and comprehensive physical activity measures, as well as measures of other potential covariates, must be used at baseline and follow-up. All components of physical activity must be assessed (type, frequency, duration, and intensity) via multiple measures, for example, physical activity questionnaire, physical activity logs, accelerometers, and maximal oxygen uptake (VO2max) treadmill testing. After each exercise session, subjects would complete a physical activity log, recording the type of exercise, duration, average heart rate, and perceived heart rate. The physical activity log could be used as a measure of exercise adherence. For measuring study compliance, ideally, all study participants would complete physical activity questionnaires and a weekly physical activity log at baseline and follow-up visits. Participants would also wear an accelerometer for a week at baseline and follow-up visits to assess physical activity levels objectively. Lastly, participants would complete a VO2max treadmill test at baseline and follow-up as an indirect measure of physical activity. Physical activity would then be compared between participants randomized to exercise versus control.
Assessment of Breast Cancer Outcome
Outcome data must be collected at semiannual or annual follow-up visits, with medical history update questionnaires collected from participants every 6 months with questions regarding hospitalizations and procedures to detect occurrence of breast cancer or any recurrences/new primaries. Investigators would then obtain medical records for any tests or hospitalizations that occurred during the study period. An outcomes committee must be formed who will review records and decide if the event is a new primary or a recurrence or death (from any cause) has occurred.
Other Essential Steps
Other issues to consider for the successful delivery of exercise and breast cancer trials include data management, quality control, statistical analyses, and power and sample size calculations. Lastly, a Data and Safety Monitoring Committee needs to be formed. This committee is responsible for monitoring the data and safety of all study participants.
Physical activity trials with disease prevention outcomes have great potential to lead to reductions in primary and secondary breast cancer risk. One benefit of a full-scale trial is that it can focus specifically on the health benefits and risks of a prescribed change in physical activity over a specific period. However, establishing the criteria for the initiation of a full-scale disease prevention trial is itself an important methodologic goal.
If randomized controlled exercise trials demonstrate that exercise can significantly decrease primary and secondary breast cancer risk, exercise could be prescribed as an integral part of primary care and breast cancer therapy. Exercise could possibly even replace toxic and costly treatments in some patients for whom chemotherapy is not very beneficial. This would, in turn, lower both breast cancer mortality and the morbidity that many patients experience from the disease and its treatment. One can even imagine a time when health club fees and physical activity counseling for primary and secondary breast cancer prevention would be reimbursed by health insurance companies.
There are, clearly, many questions to be answered concerning who, in terms of breast cancer prevention, would benefit from increasing physical activity, when physical activity would be most beneficial, and how much physical activity would be optimal. Given the high level of physical inactivity in the population and the heavy burden that breast cancer creates for the individual and for the society, the need for well-designed randomized trials of exercise on primary and secondary breast cancer prevention is an urgent public health priority.
The author thanks the Associate Editor, Barbara Sternfeld, for her editorial assistance. This article, with a more complete list of references than that provided here, can be obtained from the author.
1. ACS. Cancer Facts and Figures 2005
, Atlanta: American Cancer Society Inc., 2005.
2. McTiernan, A. (Ed.). Cancer Prevention and Management through Exercise and Weight Control
. Boca Raton: CRC Press, Taylor & Francis Group, LLC, 2006.
3. Copeland, T., M. Grosvenor, D.C. Mitchell, et al. Designing a quality assurance system for dietary data in a multicenter clinical trial: Women's Intervention Nutrition Study. JADA
4. Diabetes Prevention Program Research Group. Achieving weight and activity goals among diabetes prevention program lifestyle participants. Obes. Res.
5. Endogenous Hormones and Breast Cancer Collaborative Group. Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies. J. Natl. Cancer Inst.
6. Friedenreich, C., and M. Orenstein. Physical activity and cancer prevention: etiologic evidence and biological mechanisms. J. Nutr.
7. Goodwin, P.J., M. Ennis, K.I. Pritchard, et al. Fasting insulin and outcome in early stage breast cancer: results of a prospective cohort study. J. Clin. Oncol.
8. Holmes, M.D., W.Y. Chen, D. Feskanich, et al. Physical activity and survival after breast cancer diagnosis. JAMA
9. Irwin, M.L., A. McTiernan, and L. Bernstein, et al. Physical activity and disease-free survival in breast cancer survivors: the Health, Eating, Activity, and Lifestyle (HEAL) Study. Journal of National Cancer Institute
. Submitted for publication.
10. Irwin, M.L., A. McTiernan, R. Baumgartner, et al. Changes in body fat and weight after a breast cancer diagnosis: influence of demographic, prognostic and lifestyle factors. J. Clin. Oncol.
11. Irwin, M.L., E.J. Aiello, A. McTiernan, et al. Pre-diagnosis physical activity and mammographic density in breast cancer survivors. Breast Cancer Res. Treat.
12. Irwin, M.L., Y. Yasui, C. Ulrich, et al. Effect of moderate- to vigorous-intensity exercise
on total and intra-abdominal body fat in postmenopausal women: a one-year randomized controlled trial. JAMA
13. McTiernan, A., S.S. Tworoger, C.M. Ulrich, et al. Effect of exercise
on serum estrogens in postmenopausal women: a 12-month randomized clinical trial. Cancer Res.
14. Pierce, J.P., S. Faerber, F.A. Wright, et al. A randomized trial of the effect of a plant-based dietary pattern on additional breast cancer events and survival: the Women's Healthy Eating and Living (WHEL) Study. Control. Clin. Trials
15. Prentice, R.L., B. Caan, R. Chlebowski, et al. Low-fat dietary pattern and risk of invasive breast cancer: the Women's Health Initiative Randomized Controlled Dietary Modification Trial. JAMA