Menopause Live - IMS Updates

Date of release: 20 January, 2014

Exercise, breast size and cancer risk

At first glance, the new study by Williams [1] seemed, well, just another study showing how important physical activity is. Indeed, exercise reduces morbidity and mortality in a large list of diseases including cardiovascular, respiratory and oncological diseases. Clicking the key words 'breast cancer' and 'exercise' in PubMed yields about 2000 hits. So what's the point in commenting on this study? I guess it is because I liked its different way of looking at the issue. But first things first, and here are the core methods and findings [1]. Cox proportional hazard analyses were made of baseline pre-diagnosis MET-hours/week vs. breast cancer mortality adjusted for follow-up age, race, baseline menopause, and estrogen and oral contraceptive use in 79,124 women (32,872 walkers, 46,252 runners) from the National Walkers' and Runners' Health Studies. Women were categorized into three groups according to their level of exercise: below (< 7.5 metabolic equivalent h/week, MET-h/week), at (7.5–12.5 MET-h/week), or above (≥ 12.5 MET-h/week) recommended levels. Despite the large number of participants, only 111 women (57 walkers, 54 runners) died from breast cancer during the 11-year follow-up. The decline in mortality in women who exercised ≥ 7.5 MET-h/week was not different for walking and running (p  =  0.34), so running and walking energy expenditures were combined. The risk for breast cancer mortality was 41.5% lower for ≥ 7.5 vs. < 7.5 MET-h/week (hazard ratio (HR) 0.58, 95% confidence interval (CI) 0.38–0.92, p =  0.02), which persisted when adjusted for body mass index (BMI) (HR 0.58, 95% CI 0.37–0.96, p  =  0.03). Other than age and menopause, baseline bra cup size was the strongest predictor of breast cancer mortality, i.e. 57.9% risk increase per cup size when adjusted for MET-h/week and the other covariates (HR 1.58, 95% CI 1.27–1.97, p < 0.0001), and 70.4% greater when further adjusted for BMI (HR 1.70, 95% CI 1.34–2.2, p  =  0.0005). Breast cancer mortality was 4.0-fold greater for C-cup, and 4.7-fold greater for ≥ D-cup vs. A-cup when adjusted for BMI and other covariates. Adjustment for cup size and BMI did not eliminate the association between breast cancer mortality and ≥ 7.5 MET-h/week walked or run (HR 0.61, 95% CI 0.39–1.00, p  =  0.05). The final conclusions were that breast cancer mortality decreased in association with both meeting the exercise recommendations and smaller breast volume.


Friedenreich has summarized 73 studies in which the average reduction in breast cancer risk, when comparing the most to the least physically active women, was 25%. There was also evidence for a dose–response effect found in the majority of studies that examined this trend [2]. The strongest associations were found for recreational and household activities and for activity that was of at least moderate intensity and sustained over a lifetime. There was also some evidence for a stronger effect of physical activity amongst postmenopausal women, women who are normal weight, have no family history of breast cancer, and are parous. The suggested mechanisms by which exercise could induce this preventive effect were linked to adiposity, sex hormones, insulin resistance, adipokines, and chronic inflammation.
The current study protocol involved two interesting variables: the exercise measurement and the association with bra cup size [1]. According to the author, running and walking exercise energy expenditures can be calculated from weekly distance run or walked, which appears to be a better metric for studying dose-response than the traditional time-based calculations used by other studies [3]. The previous experience of the author demonstrated that associations of body weight, diabetes, hypertension, and high cholesterol with distance-based estimates of energy expenditure were approximately two-fold larger than those observed with time-based estimates. Thus the author favors achieving a goal of a certain total energy expenditure per week, rather than the traditional recommendation of exercising three to four times weekly for at least 30 minutes of moderate intensity. Asking women to submit their bra cup size, a simple way to estimate breast volume, and using it for the prediction of breast cancer risk, had already been mentioned in the literature many years ago [4]. In a fairly large study, women were asked to report on their bra dimensions (cup and back size) prior to a first birth or at the age of 20 for nulliparous women. Among those reporting a chest size under 34 inches, multivariate-adjusted relative risks for breast cancer were 1.34 (95% CI 1.04–1.74) for cup size B, and 1.76 (95 CI 1.04–3.01) for cup size C and larger, compared to a cup size smaller than B, and the trend for increasing cup size was statistically significant (p = 0.005). There was no relation with breast size among women reporting an average or larger back circumference (34 inches or larger). The Nurses' Health Study reported on similar results in premenopausal women [5]. Bra cup size at age 20 was assessed among 89,268 women aged 29–47 in 1993. Subsequent incident cases of invasive breast cancer were assessed until 2001. During 622,732 person-years of follow-up, 803 premenopausal women were newly diagnosed with invasive breast cancer. For women with a BMI below 25 kg/m2, those with a bra cup size of 'D or larger' had a significantly higher incidence of breast cancer than women who reported 'A or smaller' cup size (covariate adjusted HR 1.80, 95% CI 1.13–2.88, p trend = 0.01). There was no significant association among women with a BMI of 25 kg/m2 or higher.
So what did I like in the index study? Simplicity. It is easy to report on the total time spent in exercise per week and the general nature of exercise (gardening, walking, running), than keeping an exact diary for each and every bout of physical activity. Also, one may assume that any women would remember correctly her bra cup size at youth even many years later on, thus allowing a good estimation of her breast volume without a significant recall bias. Realizing that these two parameters may be important as risk factors for future breast cancer, and their potential co-associations in predicting this risk, adds another perspective to be considered while evaluating the relevant individual risk profile status.

Amos Pines
Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel