Both absolute amount of mammographic density and percent mammographic density were inversely related to recommended levels of sports/recreational physical activity among obese postmenopausal women. Specifically, we observed a 3,606 mm2 or 66% difference in the amount of dense breast area between the least active and most active obese postmenopausal women (p < .05). When we examined associations using tertiles of physical activity rather than recommended levels of physical activity, similar, but not statistically significant, associations were observed. This finding is most likely due to the fact that, when using tertiles, the highest physical activity group was only performing ~2 hr/week of physical activity compared to ~5 hr/week of physical activity when using categories of physical activity based on current physical activity recommendations. Thus, higher amounts of physical activity may be necessary to favorably affect mammographic density in obese postmenopausal women. However, until more physical activity and mammographic density studies are conducted that either confirm or contradict our findings, caution should be used in interpreting our findings. While the overall sample size was relatively large, we did have a small sample of obese postmenopausal women; thus the reduction in mammographic density in the highest activity category may be due to chance.
Inverse associations between physical activity and mammographic density were also observed among obese premenopausal women; however, because of small sample sizes, statistical significance was not reached. Conversely a positive association between physical activity and mammographic density was observed in premenopausal women with a BMI < 30. Non-obese premenopausal women had higher amounts of mammographic density compared to obese premenopausal women and both non-obese and obese postmenopausal women; these latter three groups had similar amounts of dense breast tissue. The higher amounts of mammographic density observed in lean premenopausal women are most likely a reflection of higher circulating sex hormone concentrations (1
). The finding that BMI modifies the physical activity and mammographic density association in premenopausal women is not surprising since BMI has been shown to modify the physical activity and breast cancer association (4
). Furthermore, BMI is inversely related to breast cancer risk in premenopausal women, but positively related to breast cancer risk in postmenopausal women (17
). More research examining the modifying effect of BMI and menopausal status on the physical activity and mammographic density association is necessary.
Only two previous studies have examined the association between physical activity and mammographic density, both in healthy women. Vachon et al. (5
) investigated the association between physical activity and percent mammographic density in 1900 women. Physical activity was not associated with percent mammographic density among either pre- or post-menopausal women. However, their assessment of physical activity was limited to only one question. Gram et al. (6
) examined the relationship between physical activity and mammographic density among 2720 Norwegian women. Women who reported moderate physical activity, i.e., more than two hours per week, were 20% less likely (OR = 0.80, 95% CI: 0.60 –1.10) to have high-risk mammographic patterns compared with those who reported being inactive. This relationship remained consistent when stratified by menopausal status and tertiles of BMI. One recent study examined sedentary activities and percent mammographic density among 294 healthy Hispanic Women. Lopez et al. (18
) observed a higher percent mammographic density for women who reported at least 3.5 hr/day of sedentary activities compared to women reporting less sedentary activities adjusted for age, education, BMI, parity, menopausal status, HRT use, and smoking status (p = 0.056).
One mechanism has been proposed to explain an association between physical activity and mammographic density. Physical activity may influence mammographic density by favorably changing certain hormones that may be
associated with mammographic density, such as sex steroid hormones. Many published studies have shown an effect of exogenous estrogens on increasing mammographic density (19
), however, the only published study that examined the association between mammographic density and endogenous estrogens observed no association between total estradiol and percent density in both pre- and post-menopausal women, and an inverse association between free estradiol and percent dense tissue in postmenopausal women (20
). The hypothesized mechanism of physical activity influencing mammographic density via sex hormone changes applies especially to obese postmenopausal women, who have higher levels of sex hormone concentrations than leaner postmenopausal women, due to the formation of estrogens in fatty tissue (10
). Thus, higher levels of physical activity may be associated with decreased mammographic density levels among obese postmenopausal women by decreasing sex hormone concentrations directly or indirectly by reducing body fat (7
). In a recently published yearlong randomized controlled trial, exercise had a favorable effect on decreasing circulating sex hormone concentrations among overweight postmenopausal women (7
The HEAL Study has several limitations and strengths. While the HEAL Study is a prospective cohort study, this analysis is cross-sectional in design. Physical activity was also assessed retrospectively, and is, therefore, subject to recall bias. Other limitations are that we assessed menopausal status and BMI at the baseline visit, which corresponds to approximately six months post-diagnosis or approximately 21 ± 12 months after the time of mammogram. Some women may have experienced changes in menopausal status and BMI from the time of their mammogram to the time of the baseline visit. However, when we defined menopausal status as ages 57 and over, our findings were similar to what we observed when menopausal status was defined as ages 55 and over. Another limitation is that we did not have any information on timing of the mammogram in relation to the menstrual cycle among the premenopausal women. Mammographic density is lower in the follicular phase than in the luteal phase (23
). Another limitation is sample size; while our overall sample size was relatively large, we did have a small sample of obese pre- and post-menopausal women; thus associations observed between physical activity and mammographic density among these groups should be interpreted with caution. Lastly, while not necessarily a limitation, physical activity and mammographic mammographic density were measured in women with subclinical disease (i.e., we measured physical activity and mammographic density in women approximately one year before their breast cancer diagnosis). However, we have no reason to believe that associations between physical activity and mammographic mammographic density would be different among a cohort of healthy women since we examined mammographic density in the unaffected breast. We also excluded women diagnosed with bilateral breast cancer.
Major strengths of our study are the quality of the physical activity data that were obtained from a reliable and valid 29-item interview-administered questionnaire; we measured weight and height whereas many previous studies used self report of these measures; we recruited both Non-Hispanic White and Hispanic White women; and we used a computer-assisted method to assess mammographic density.
Measuring mammographic density as a continuous scale may provide more information than the BIRADS or Wolfe categorical measures (1
), which most previous mammogram density studies have used. However, even though the computer-assisted continuous scale is considered the gold standard of assessing mammographic density, the currently available method assesses mammographic density with a 2-dimensional view. A more accurate method would involve a 3-dimensional view, or volumetric approach, that captures overall volume. The need for volumetric measures is clearly reflected in the uncertainty of results observed in this study and others related to mammographic density and BMI. With two-dimensional views, one does not get a complete picture of the overall volume of the breast. Uncertainty regarding the overall volume is a problem for women with large breasts, which is often associated with a higher BMI. While percent mammographic density may be low for heavier women (due to fatty breasts), if they have a large breast volume, they may actually have more dense breast tissue than observed with one- and two-dimensional views. Ongoing research focused on developing a more precise volumetric method of assessing mammographic density is currently being conducted by Pawluczyk and colleagues (24
In conclusion, mammographic density differs from other breast cancer risk factors in the strength of its association with breast cancer, in being present in the tissue from which breast cancer arises, and because it is modifiable (1
). Information will soon be available from randomized controlled trials (e.g., the PEPI Trial; 25) designed to determine the effects of interventions on breast cancer risk. Indirectly these trials will provide information on the value of short-term changes in mammographic density in predicting an intervention effect on breast cancer risk. Future exercise trials will then be able to observe if physical activity changes mammographic density a clinically significant amount to affect breast cancer risk. Demonstration of beneficial changes in mammographic density associated with increased physical activity levels could be used to motivate women to be more physically active. Although it remains to be determined whether the effect of increasing physical activity decreases mammographic density, it is known that both physical inactivity and mammographic density are associated with breast cancer risk (1
). Furthermore, many studies have identified obesity as an important negative factor for postmenopausal breast cancer risk and prognosis, and physical activity is associated with weight loss and maintenance (26
). Increasing physical activity among obese pre- and post-menopausal women may be a reasonable intervention approach to reduce mammographic density, thereby influencing breast cancer risk and recurrence.