Recreational Physical Activity and Leisure-Time Sitting in Relation to Postmenopausal Breast Cancer Risk

In 2007, the World Cancer Research Fund/American Institute for Cancer Research concluded that sufficient epidemiologic evidence exists for a probable inverse association between physical activity and postmenopausal breast cancer (1). More than 70 observational studies support an approximate 25% lower risk of breast cancer among physically active, relative to inactive, women (2). However, questions about the association remain unanswered. Although the evidence supports an association for regular vigorous physical activity, it is unclear whether activities of moderate intensity, such as walking, impart a benefit in the absence of vigorous exercise. In addition, whether associations differ by molecular features, such as hormone receptor status, or by individual factors such as weight status and use of postmenopausal hormones (PMH) is unclear. PMH use is known to modify associations between excess weight and breast cancer where a positive association is observed only in nonusers (3). Furthermore, while prolonged periods of sitting have been associated, independent of physical activity, with premature mortality, cardiovascular disease, type II diabetes mellitus, as well as some cancers (4), the relation between sitting time and postmenopausal breast cancer risk is not well understood (5, 6).

It has been noted that large prospective studies are needed to clarify whether associations between physical activity and breast cancer differ by hormone receptor status, body mass index (BMI, kg/m²), weight gain, or PMH use (7). An inverse association between recreational physical activity and incident postmenopausal breast cancer in the American Cancer Society Cancer Prevention Study II (CPS-II) Nutrition Cohort was previously reported (8). However, that study was not sufficiently powered to examine subgroup associations, nor was time spent sitting evaluated. With 12 years of additional follow-up time, we examined total recreational physical activity, walking, and leisure-time sitting in relation to postmenopausal breast cancer incidence. Associations also were assessed by estrogen receptor (ER) status, BMI, adult weight gain, and PMH use.

Subjects were drawn from the 97,785 women participants in the CPS-II Nutrition Cohort, a prospective study of cancer incidence established by the American Cancer Society in 1992 (9). The Nutrition Cohort is a subgroup of the CPS-II baseline mortality cohort (10). Participants of ages 50 to 74 years were enrolled in the study in 1992–93 when they completed a 10-page, self-administered questionnaire on demographic, reproductive, medical, environmental, and behavioral factors. Beginning in 1997, follow-up questionnaires have been sent to participants every 2 years to update exposure information and to ascertain newly diagnosed cancers. Response rates among living cohort members are 88% or more.

Excluded from the analysis were 3,111 women lost to follow-up (i.e., alive at first follow-up but did not return any follow-up questionnaire), who reported prevalent cancer (except non-melanoma skin cancer) at enrollment (N = 12,059), were pre/peri-menopausal (N = 4,712), or missing information on recreational physical activity, sitting time, or BMI (N = 4,222). Women who reported breast cancer on their first follow-up questionnaire, that was never verified, were also excluded (N = 66). The final cohort for analysis consisted of 73,615 postmenopausal women (i.e., whose periods had stopped permanently before enrollment, naturally or for surgical/medical reasons) with a mean age of 62.7 years and median follow-up time of 14.2 years between enrollment and June 30, 2009, which marked the end of follow-up for this analysis.

Of the 4,760 incident breast cancer cases [International Classification of Diseases for Oncology (ICD-O) topography code C50] diagnosed among the cohort during follow-up, 4,662 cases were initially identified by self-report and subsequently verified by medical records or linkage with state cancer registries. Sensitivity of self-reported cancer in this cohort has been estimated to be 93% (11). An additional 98 incident cases were identified through automated linkage with the National Death Index (12), of which 78 were subsequently verified via linkage to state registries. Of the 69% of cases with available ER status (+/−), 84% (N = 2806) were ER⁺ and 15% (N = 498) were ER⁻.

Information on weekly recreational activities was collected at enrollment as described in detail elsewhere (8). The average number of hours/week spent in each of the following activities was assessed: walking, jogging/running, lap swimming, tennis/racquetball, bicycling/stationary bike, aerobics/calisthenics, and dancing. A summary estimate of total hours/week of the metabolic equivalent (MET), which is an estimate of the ratio of the energy expenditure during a specific activity to the resting metabolic rate, was calculated for each individual (13). Women who reported no activity were categorized as “none,” and the remaining (active) women were categorized according to quintiles of MET-hours/week (>0–7.0, >7.0–17.5, >17.5–31.5, >31.5–42.0, and >42.0). Hours/week of walking was categorized consistent with questionnaire response categories (none, ≤3, 4–6, and ≥7). Because declining health or chronic conditions associated with aging or low estrogen (e.g., osteoporosis) may prevent some women from engaging in even light recreational physical activity, women reporting the lowest level of activity, as opposed to “none,” comprised the referent group for these comparisons.

Leisure-time sitting (time spent watching TV, reading, etc.) was categorized as 0–<3 (referent), 3–5, or ≥6 h/d, based on questionnaire response categories.

Physical activity and time spent sitting were requeried and updated in 1999, 2001, and 2005.

Follow-up time for each participant was calculated as person-years from the date of completion of the enrollment questionnaire to the date of: (i) diagnosis of breast cancer; (ii) death; (iii) the last cancer-free questionnaire when unverifiable breast cancer was self-reported on a subsequent questionnaire; (iv) the last completed questionnaire if no subsequent follow-up questionnaire was returned or if the follow-up questionnaire was missing information on physical activity; (v) June 30, 2009. Extended Cox regression was used to estimate age-adjusted and multivariable-adjusted hazard rate ratios for approximation of relative risks (RR) with 95% confidence intervals (CI; ref. 14). Multivariable-adjusted models included race, education, BMI, adult weight change (from age of 18 years to age at enrollment), alcohol intake, smoking status, age at menopause, number of live births/age at first live birth, personal history of breast cysts, hysterectomy or oophorectomy, family history of breast cancer, mammography, and PMH use (never, current, or former). We evaluated the influence of type and duration of PMH on the associations of interest, but found that adjustment for current hormonal status controlled for confounding equally as well as a variable also defining type and duration of use. PMH use and mammography were updated in 1997, 1999, 2001, 2003, and 2005. All models were stratified on age at enrollment.

P values for statistical significance of heterogeneity of associations between ER⁺ and ER⁻ breast cancer were derived from the Cochran Q test statistic, generated from meta-analysis procedures (15). Effect modification by BMI (<25 or ≥25 kg/m²), adult weight gain (≤35 or >35 lbs.), and PMH use (never/former or current) was assessed in stratified analyses; P values for statistical significance of interaction were generated by likelihood ratio tests comparing models of cross-product terms with models of independent main effects. Linear trend of RR in relation to MET-hours/week was tested with a variable defining the median value of each category of MET-hours/week, and the P value was derived from the Wald χ² statistic. All tests for statistical significance were two-sided.

Approximately, 9.2% (n = 6,747) of women reported no recreational physical activity at baseline (Table 1). The median MET expenditure among active women was 9.5 MET-hours/week, which is equivalent to 3.5 hours/week moderately paced walking. Physically active women, regardless of the amount, engaged primarily in activities judged to be of moderate intensity (walking, cycling, aerobics, and dancing) rather than vigorous-intensity activities (jogging/running, swimming, tennis/racquetball); 47% of women reported walking as their only recreational activity. Physically active women tended to be leaner, more likely to maintain or lose weight during adulthood, more likely to drink alcohol, and less likely to currently smoke. They were also more likely to use PMH and to have had a mammogram in the past year. Sitting time was not correlated with recreational physical activity (Pearson r = −0.05).

Recreational physical activity was inversely associated with breast cancer incidence (P_trend = 0.01, active women only; Table 2). The most active women (those reporting >42.0 MET-hours/week) had a 25% lower risk of breast cancer relative to women in the least active category (>0–7.0 MET-hours/week; 95% CI, 0.63–0.89). Walking was inversely associated with breast cancer risk. Among women who reported walking as their only activity, those walking ≥7 hours/week had a 14% lower breast cancer risk relative to women walking ≤3 hours/week (95% CI, 0.75–0.98). No statistically significant heterogeneity of associations was observed by ER status (MET hours/week, P = 0.70; walking, P = 0.99; sitting time, P = 0.08), nor was there any evidence of interaction between exposures and BMI (MET hours/week, P = 0.97; walking, P = 0.86; sitting time, P = 0.25), adult weight gain (MET hours/week, P = 0.82; walking, P = 0.56; sitting time, P = 0.44) or PMH use (MET hours/week, P = 0.97; walking, P = 0.99; sitting time, P = 0.06).

A modest positive association between time spent sitting and breast cancer was observed in age-adjusted models. However, when adjusted for other factors including physical activity, the RR was attenuated and no longer statistically significant (Table 2).

In this large prospective study of postmenopausal women, an inverse association was found between physical activity and incident breast cancer. Walking on average at least 1 hour/day was modestly associated with lower risk, even in the absence of other recreational physical activities. Associations did not differ by ER status, BMI, adult weight gain, or PMH use. Time spent sitting was not associated with breast cancer incidence in this cohort.

Physical activity is consistently associated with lower risk of postmenopausal breast cancer (1). Our findings of a 25% lower RR associated with >42 MET-hours/week and a 14% lower RR associated with the equivalent of 1 hour/day walking are notably consistent with a review of more than 70 studies that reported a 28% reduction in risk and, in particular, a 15% reduction in risk attributable to less moderate physical activity, comparing the highest to lowest categories based on frequency, duration, and intensity (2). Current guidelines for adults recommend at least 150 minutes/week of moderate-intensity or 75 minutes/week of vigorous-intensity aerobic activity for overall health (16). Yet, less than half of U.S. adult women are active at these minimum levels (17), and thus an even smaller proportion of women likely achieve the higher levels thought necessary for breast cancer risk reduction. Given that more than 60% of women report some daily walking, promotion of leisure-time walking may be an effective strategy for increasing physical activity among postmenopausal women (18).

Among the few studies that examined differences in associations by ER status (19–21), no consistent patterns of variation have emerged. Our results, drawn from a large prospective cohort with more than 4,700 incident breast cancer outcomes, suggest that the association between physical activity and lower risk does not vary by ER status, BMI, adult weight gain, or PMH use. A review of 22 studies that stratified results by BMI indicated only slightly stronger associations for physical activity in lean/normal weight (BMI <25 kg/m²), as compared with overweight/obese (BMI ≥25 kg/m²) women, and concluded that women of most body sizes likely benefited (2). However, a more recent review noted an approximate 27% lower risk for highest versus lowest levels of physical activity among women with a BMI of <22 kg/m² versus only 1% lower risk among women with BMI ≥30 kg/m² (22). To further evaluate BMI with respect to these findings, we replicated the four-level stratification scheme of the latter review and found no differences in associations between physical activity and breast cancer risk across the four levels. Whether other measures of adiposity such as weight gain or waist circumference influence associations between physical activity and breast cancer has not been extensively studied. We evaluated but did not find evidence of effect modification by adult weight gain. Unfortunately, waist circumference was not assessed at enrollment of the CPS-II Nutrition cohort; therefore, we were unable to evaluate waist size in relation to the associations of interest.

Results from two other studies suggest that sedentary time may be positively associated with breast cancer risk independent of physical activity (5, 6). Our results do not contribute evidence of such an association. However, our measure of leisure-time sitting may be capturing a different pattern of sedentary behavior than the other studies, which assessed occupational sitting time (6) and routine daily (occupational or otherwise) sitting (5).

Biologic plausibility for the inverse association between physical activity and breast cancer incidence is consistent with the well-established beneficial effects of exercise on endogenous hormone levels, weight control, glucose metabolism, insulin sensitivity, and inflammatory markers, all factors implicated in the etiology of postmenopausal breast cancer (23–25). Thus, physical activity may act favorably upon hormonal and nonhormonal pathways to lower risk of breast cancer.

The major strengths of this study are its large size, prospective design, and availability of detailed exposure information collected prior to breast cancer diagnosis and updated repeatedly during follow-up. It is important to note that the assessment of recreational or leisure-time physical activity in this study may not reflect total physical activity in those working individuals whose occupations involve manual activity. However, most CPS-II women were, or are, homemakers. Furthermore, given the age and demographic of the cohort, any additional contribution of occupational activity to these data is likely to be negligible. This study population is predominantly White, middle-aged or elderly, and well educated; therefore, our results may not be generalizable to populations with different characteristics.

These results contribute additional evidence of a potential benefit of physical activity on the risk of breast cancer in postmenopausal women. Women who engage in at least 7 hours of walking over the course of a week may reap a modest benefit, even in the absence of more vigorous exercise. Given that breast cancer is the most common cancer affecting women (26), and that walking is a common activity among postmenopausal women, the finding of a possible lower risk with an average one or more hours/day of walking is of considerable public health interest.

No potential conflicts of interest were disclosed.

Conception and design: J.S. Hildebrand, S.M. Gapstur, P.T. Campbell, A.V. Patel

Development of methodology: J.S. Hildebrand, P.T. Campbell

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.M. Gapstur, P.T. Campbell, A.V. Patel

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.S. Hildebrand, S.M. Gapstur, P.T. Campbell, M.M. Gaudet, A.V. Patel

Writing, review, and/or revision of the manuscript: J.S. Hildebrand, S.M. Gapstur, P.T. Campbell, M.M. Gaudet, A.V. Patel

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.S. Hildebrand

Study supervision: S.M. Gapstur

The authors thank the CPS-II participants and the Study Management Group of the American Cancer Society Epidemiology Research Program for their invaluable contributions to this research. The authors also thank the contribution of central cancer registries supported through the Centers for Disease Control and Prevention (CDC) National Program of Cancer Registries and cancer registries supported by the National Cancer Institute's Surveillance, Epidemiology, and End Results Program.

American Cancer Society funds the creation, maintenance, and updating of the CPS-II cohort.