Lifestyle and Breast Cancer Risk: The Way Forward?

Los Angeles Biomedical Research Institute at Harbor-University of California Los Angeles Medical Center, Torrance, CA

An investigative team with outstanding credentials led by Dr Friedenreich has conducted a rigorous randomized controlled trial evaluating the relationship between physical activity and reproductive hormone levels. They were able to gain extraordinary adherence from previously sedentary, mostly overweight, postmenopausal women to an extensive aerobic exercise intervention over a 12-month period requiring about 5 hours of total physical activity and about 20 met hours per week of recreational activity. At 12 months, statistically significant reductions in estradiol, free estradiol and increases in sex hormone–binding globulin were observed in the exercise group compared with the control group. However, these changes were quite modest representing about a 9% decrease in free estradiol.

While the study successfully addressed its objectives, the modest reduction in estrogen levels seen is unlikely to mediate any major reduction in breast cancer. The authors acknowledge this when they suggest that physical activity may be an acceptable means of breast cancer risk reduction only in those whose risk does not warrant use of chemoprevention intervention.

The investigators cite ongoing risk reduction trials comparing aromatase inhibitor to placebo. In contrast to the modest estrogen change seen in the current report, aromatase inhibitors reduce estrogen levels by approximately 90%. Based on aromatase inhibitor influence on contralateral breast cancers in trials with tamoxifen as comparator,the primary prevention trials evaluating aromatase inhibitors against placebo have sample sizes based on anticipated 60% to 70% reduction in breast cancer risk. While aromatase inhibitors have adverse effects, if these primary prevention trials are successful, they would represent a vigorous competitor for any lifestyle approach. So a dilemma arises. While reasonable mediators of breast cancer risk which may be under lifestyle influence have been identified, pharmacologic approaches may be both more potent and more palatable for both patients and health professionals.

Similar to others in the past, the investigative team also plans to determine the influence of their physical activity intervention on insulin-like growth factors and insulin resistance. In this regard, the Women's Intervention Nutrition Study, an adjuvant breast cancer trial evaluating a lifestyle intervention targeting fat intake reduction, also reported a five to six pound statistically significant weight loss in the intervention group. In an interim analysis, a statistically significant improvement in relapse-free survival was seen in the intervention group with a greater effect in the estrogen receptor–negative, progesterone receptor–negative population suggesting a mediator other than estrogen. This and other findings led to increased interest in a role for insulin as a potential mediator and several reports now link higher fasting-insulin levels at diagnosis to significantly worse breast cancer recurrence risk. Thus, fasting insulin and insulin resistance represents another reasonable target for lifestyle interventions. Insulin can be influenced by lifestyle change, however, a pharmacologic agent, metformin, can also influence insulin resistance. Emerging clinical studies have led to plans for a full scale, randomized adjuvant breast cancer trial with metformin as the sole treatment variable.These examples suggest that lifestyle change may find it difficult to beat pharmacologic agents once specific intervention targets are identified.

Nonetheless, given the consistent observational study evidence it seems unlikely that the influence of lifestyle change on breast cancer will be explained by the modest influence on currently identified prospective mediators of breast cancer risk including estrogen and insulin. In addition to the extensive literature cited in the report of Friedenreich and colleagues, several recent cohort analyses report that adherence to several straightforward behaviors (nonsmoking, no more than moderate alcohol intake, intake five fruit and vegetable servings per day, and moderately higher physical activity) had striking associations with favorable all-cause mortality risk. In the European Prospective Investigation Into Cancer-Norfolk (EPIC-Norfolk) cohort, those reporting adhering to none of the four behaviors had a hazard ratio for mortality that was more than 4 compared with individuals reporting that they followed all of the four behaviors. The investigators calculated the difference between complete adherers to these four behaviors and the nonadherers as representing a difference of 14 chronological years of age. More recently, comparable favorable associations with such behaviors on mortality risk were confirmed in a Japanese cohort where lack of obesity was an additional component.

In summary, we have strong signals from observational studies that lifestyle factors can potentially have major influence on overall mortality risk. However, it is recognized that such associations in observational studies can be influenced by confounding. Also, adhering and nonadhering groups to such behaviors would not likely be found to have profoundly different insulin or estrogen levels. Thus, the observational study experience suggests that lifestyle change can influence breast cancer and all-cause mortality by mechanisms currently unidentified. The question remains, can such observational study findings, largely reflecting modest, but not profound lifestyle change, be confirmed in randomized clinical trial settings?

One aspect of this question was addressed in the Women's Health Initiative Dietary Modification trial, a full scale effort to evaluate a lifestyle intervention influence on cancer outcomes. This randomized controlled trial entered more than 48,000 postmenopausal women to an intervention largely targeting fat intake reduction or a control condition. While there were fewer breast cancers in the intervention group (hazard ratio, 0.91; 95% CI, 0.83 to 1.01; P = .09), the difference was not statistically significant. Follow-up of these participants is ongoing. While the Women's Health Initiative trial targeted dietary intakes, evidence emerging in the most recent decade suggests interventions targeting body weight and physical activity would likely be more effective.

Recently, investigators at the Applied Research Program, Division of Cancer Control Population Sciences of the National Cancer Institute and consultants presented a rationale and design for a primary lifestyle intervention to evaluate as primary breast cancer prevention in women at high risk of breast cancer and in women with diagnosed breast cancer as an adjuvant therapy. The goals for both were proposed to be weight loss of 10% of body weight for those overweight and the goal for physical activity would be to achieve and maintain a moderate-intensity program for a total of 150 to 225 minutes over at least 5 days per week.

There is one ongoing adjuvant breast cancer clinical trial that meets the National Cancer Institute call, namely the Lifestyle Intervention Study Adjuvant sponsored by the Ontario Clinical Oncology Group with Pamela Goodwin as principal investigator. Postmenopausal patients with breast cancer with early-stage resected disease are being randomly assigned to a program targeting weight loss and diet and physical activity goals or a control group. This important adjuvant study deserves support. In addition, a full-scale breast cancer primary prevention trial now has more than sufficient observational study support to go forward with a high expectation of success. Now is the time to determine whether the promise of breast cancer risk reduction, consistently identified by associations in observational studies, can be supported by results in randomized trials in the clinic.