A recent review of the results of more than 60 observational, epidemiological studies of physical activity and breast cancer concluded that physically active women have a lower risk of developing breast cancer than sedentary women (Physical Activity Guidelines Advisory Committee 2008). As with the investigations of colon cancer, inverse associations between physical activity and breast cancer have been seen in studies conducted in North America, Europe, Asia, and Australia. Comparison of physically active women with sedentary women showed a median risk reduction of 20% (Lee and Oguma 2006).
Approximately three-quarters of these studies classified women according to three or more levels of physical activity, allowing assessment of a dose–response relation (Lee and Oguma 2006). There appears to be an inverse dose response, with about three-fifths of these studies reporting either a significant inverse trend across levels of physical activity, or relative risks consistent with an inverse dose response that was not tested for statistical significance (Lee and Oguma 2006). A systematic review of case–control studies of recreational activity and breast cancer risk estimated that each additional hour of physical activity per week reduced
risk of postmenopausal breast cancer by 6% (95% CI: 3-8%) (Monninkhof et al. 2007).
Whether the association of physical activity with breast cancer risk differs among subgroups of women (e.g., with and without family history, nulliparous vs. parous, lean and overweight women) or by breast tumor characteristics is unclear (Physical Activity Guidelines Advisory Committee 2008). With regard to menopausal status, the inverse relation appears stronger in postmenopausal compared with premenopausal women (Physical Activity Guidelines Advisory Committee 2008; Lee and Oguma 2006). The periods of life that may be the most relevant for the protective effects of physical activity on breast cancer risk have not been established. Several studies have observed that lifetime physical activity is needed to significantly reduce risk (e.g., Bernstein et al. 1994); others have also noted that physical activity during adolescence (e.g., Maruti et al. 2008) or at various times in life (e.g., Carpenter et al. 1999) is more strongly associated with reduced risk than physical activity carried out at other times.
More than three dozen population-based case–control studies have examined the relation between physical activity and risk of developing breast cancer (Physical Activity Guidelines Advisory Committee 2008). The majority have assessed the role of recreational or leisure-time physical activity; and overall, most studies indicate that physically active women have a lower risk of developing breast cancer than sedentary women. Risk reductions ranging from 20% to 70% have been noted, with a median of 30% across all studies (Lee and Oguma 2006). Here we describe several exemplar case–control studies that have investigated the association of physical activity with breast cancer risk.
Los Angeles County
In one of the early case–control studies on this topic, Bernstein and colleagues (1994) matched 545 women (ages 40 and younger at diagnosis) who had been newly diagnosed with in situ or invasive breast cancer between 1983 and 1989 with 545 control subjects by date of birth, race, parity, and neighborhood of residence. Lifetime histories of participation in regular exercise activities were obtained during an interview. After adjustment for age at first pregnancy, age at menarche, birth date, family history, months of lactation, number of full-term pregnancies, parity, race, and oral contraceptive use, the average number of hours spent in exercise activities per week from menarche to one year prior to diagnosis was independently associated with reduced breast cancer risk. Overall, the odds ratio of breast cancer among women who participated for 3.8 h or more per week in exercise activities was 58% lower (odds ratio, 0.42; 95% CI: 0.27-0.64) than in inactive women. The inverse association appeared stronger in women with at least one full-term pregnancy.
This study showed an inverse association in Chinese women (Matthews et al. 2001) similar to that seen among U.S. women in the Los Angeles County study. Subjects for this population-based case–control study were 1459 women newly diagnosed with breast cancer and 1556 age-matched controls in urban Shanghai. Interviewers obtained information on physical activity from exercise and sports, household activities, and walking and cycling for transportation during adolescence (age 13-19 years) and adulthood (last 10 years); lifetime occupational activity also was ascertained. After the investigators controlled for age, education, income, family history of breast cancer, history of breast fibroadenoma, age at menarche, age at first live birth, and age at menopause, the risk of developing breast cancer was reduced with exercise only in adolescence (odds ratio, 0.84; 95% CI: 0.70-1.00), exercise only in adulthood (0.68; 0.53-0.88), and exercise at both times (0.47; 0.36-0.62). Household and transportation activities did not show any significant associations with breast cancer risk, while lifetime occupational activity involving more standing and walking was inversely related to risk.
Women’s Contraceptive and Reproductive Experiences Study
This was a multicenter, population-based case–control study designed to investigate the predictors of breast cancer (including physical inactivity) among white and black women in Atlanta, Detroit, Los Angeles, Philadelphia, and Seattle (Bernstein et al. 2005). Detailed histories of lifetime recreational physical activity were obtained by interviews with 1605 black and 2933 white women aged 35 to 64 years, with newly diagnosed invasive breast cancer, and 1646 black and 3033 white control women who were frequency matched to cases on age, race, and study site. Among all women, increased levels of lifetime physical activity were associated with decreased breast cancer risk. After controlling for age, study site, race, family history of breast cancer, age at menarche, menopausal status, age at first full-term pregnancy, body mass index, and months of lactation, the most active black women, expending ≥3.0 h/week on average, had an odds ratio for breast cancer of 0.75 (95% CI: 0.61-0.93) compared with inactive women. For white women, the corresponding results were similar (odds ratio, 0.83; 0.70-0.98). For both races, there were significant inverse dose–response trends. Unlike what was seen in the case–control study from Los Angeles County described earlier, the association in this study did not differ between nulliparous and parous women.
Retrospective Cohort Studies
Several retrospective cohort studies also have examined whether physical activity is related to reduced breast cancer risk. We describe two such studies below.
College Alumnae Study
Frisch and colleagues (1985) conducted one of the first major studies to examine the potential protective effects of exercise against breast cancer. This study used a retrospective cohort study design to evaluate the relation between exercise and the prevalence of female reproductive cancers in a cohort of 5398 living (as of 1981) alumnae from the classes of 1925 to 1981 from 10 U.S. colleges. A total of 69 prevalent breast cancers were reported by the women. Alumnae were classified as athletes (defined as being on at least one varsity team, house team, or other intramural team for one or more years) or nonathletes. After adjustment for age, family history of cancer, age of menarche, number of pregnancies, use of oral contraceptives, use of hormones during menopause, smoking, and leanness, the relative risk of prevalent breast cancer for the athletes, compared with nonathletes, was 0.54 (95% CI: 0.29-1.00). Questionnaires reflected that a larger percentage of the former athletes than of the former nonathletes were physically active after college (74% vs. 57%). The authors concluded that long-term athletic training establishes a lifestyle that lowers the risk of breast cancer. While this early study provided useful information, its limitations included the small number of breast cancers and the investigation only of college exercise in relation to breast cancer risk. Given that nearly 60% of the college nonathletes were currently active at the time of the study, this may have diluted an association with breast cancer. In addition, since the study included only women who were alive in 1981, a selection bias may have occurred since only prevalent cases were included and mortality from breast cancer was not evaluated. If former athletes were likely to live longer than former nonathletes, the prevalence of nonfatal breast cancer would be higher in the athletes; thus, the observed inverse association likely would have been more marked if this bias had been removed.
This retrospective cohort study took advantage of occupational information recorded on death certificates in the state of Washington to investigate the association of physical activity with breast cancer mortality (Vena et al. 1987). Vena and colleagues made use of published data from the death certificates of 25,000 white females (excluding women who were housewives) who died in that state from 1974 to 1979. Occupation titles were recorded from the death certificates related to “usual occupation during most of working life, even if retired.” Based on their job title, women were classified into five categories of occupational physical activity. Standardized proportionate mortality ratios of specific cause of death were calculated for each category of occupational physical activity through comparison of the observed number of deaths against the expected number of deaths across all occupations, based on sex and calendar year, for that specific cause of death. Women in the most sedentary category of occupation had a standardized mortality ratio for breast cancer of 115 (i.e., 15% higher than expected), while women in the three most active categories of occupational activity, combined, had a standardized mortality ratio of 85% (i.e., 15% lower than expected).
While this study made efficient use of information already collected, it also had several limitations. No adjustments were made for known risk factors of breast cancer that could be related to occupational activity. For example, it was likely that sedentary jobs were held by women with higher socioeconomic status and a history of late first pregnancy, both of which are associated with increased risk for breast cancer. In addition, occupational physical activity was based only on the job title listed on the death certificate; lifetime occupational history could not be assessed.
Prospective Cohort Studies
More than two dozen prospective cohort studies have investigated whether physical activity reduces the risk of developing breast cancer (Physical Activity Guidelines Advisory Committee 2008). As with the case–control studies, most of these examined recreational or leisure-time physical activity. The totality of evidence suggests that physically active women have a lower risk of developing breast cancer than sedentary women, with risk reductions ranging from 20% to 80%. Across all studies, the median risk reduction was 10%, a smaller magnitude than that observed in the case–control studies (Lee and Oguma 2006). Here we describe several exemplar prospective cohort studies on this topic.
This study represents one of the first detailed prospective cohort studies of physical activity and risk of developing breast cancer. Between 1974 and 1978 and again between 1977 and 1983, a total of 25,624 women, 20 to 54 years of age, answered questionnaires about leisure-time and occupational physical activity (Thune et al. 1997). Over nearly 14 years of follow-up, 351 cases of invasive breast cancer were detected. Regular leisure-time physical activity was associated with about a one-third reduction in the risk of breast cancer after adjustment for age, body mass index, height, parity, and county of residence (RR = 0.63; 95% CI: 0.42-0.95). The reduction in risk was greater in premenopausal women than in postmenopausal women, greater in younger women (≥45 years), and greater in lean women (body mass index 2) than in heavier women. Higher levels of occupational physical activity were associated with lower breast cancer risk as well, and that effect also was stronger among premenopausal women.
University of Pennsylvania Alumnae
Investigators in this study wanted to assess the association of physical activity and breast cancer risk, as well as to determine whether the association differed according to postmenopausal status and body mass index as suggested by the study from Norway just described. Approximately 1600 University of Pennsylvania alumnae who had matriculated between 1916 and 1950 (average age 45.5 years) and who were initially free of breast cancer, in 1962, were observed until 1993 for breast cancer occurrence (Sesso, Paffenbarger, and Lee 1998). At baseline, women reported their stair climbing, walking, and sport participation and were classified into three levels (≥1000 kcal/week). During 35,365 person-years of observation, 109 cases of breast cancer cases were detected. After adjustment for age and body mass index, postmenopausal women who expended ≥1000 kcal or more each week had about half the risk of developing breast cancer than women who expended
Nurses’ Health Study I and II
These two studies are among the few prospective cohort studies that updated information on physical activity over time. The earlier investigation enrolled subjects from Nurses’ Health Study II, a prospective cohort study of women’s health among approximately 117,000 women who were 25 to 42 years old in 1989 (Rockhill et al. 1998). At the baseline survey, women were asked, “While in high school and between the ages 18 and 22 years, how often did you participate in strenuous physical activity at least twice a week?” Responses to the two time periods were averaged to estimate physical activity in late adolescence. Women were also asked how many hours each week they currently spent in several leisure physical activities using questions similar to those in Nurses’ Health Study I, described earlier in the context of physical activity and colon cancer (at that time, the study was referred to as the Nurses’ Health Study). During six years of observation, 372 cases of invasive breast cancer were detected. Among these women, who were primarily premenopausal, after adjustment for age, age at menarche, age at first birth, parity, oral contraceptive use, height, alcohol intake, history of benign breast disease, and family history of breast cancer, neither physical activity in adolescence nor contemporary leisure activity was associated with the risk of breast cancer (relative risk for either time period, comparing high with low activity level, was 1.1); this finding was congruent with those seen in the University of Pennsylvania alumnae just discussed.
The second study was based on women from Nurses’ Health Study I, a cohort of older women aged 30 to 55 years at baseline in 1976 (Rockhill et al. 1999). For this analysis of physical activity and breast cancer risk, follow-up started in 1980, when women were asked about the average number of hours per week they spent in various moderate and vigorous recreational physical activities during the past year. Physical activity information was updated every two years; and starting in 1986, more detailed questions on eight different groups of recreational activities were asked. During 16 years of observation, 3137 cases of invasive breast cancer (1036 premenopausal and 2101 postmenopausal women) were detected. Women who reported participation in moderate or vigorous physical activity for 7 h or more per week had nearly 20% lower risk of breast cancer than those who averaged less than 1 h per week (RR = 0.82; 95% CI: 0.70-0.97) after adjustment for age, age at menarche, age at first birth, menopausal status, use of postmenopausal hormones, family history of breast cancer, history of benign breast disease, body mass index, and height. There also was a significant, inverse dose response across categories of physical activity. One possible explanation for the different results across these two studies might have been the higher preponderance of postmenopausal breast cancers in the latter study; the overall body of evidence suggests stronger associations for postmenopausal than for premenopausal breast cancer (Physical Activity Guidelines Advisory Committee 2008; Lee and Oguma 2006).
Women’s Health Initiative Cohort Study
All of the prospective cohort studies discussed in the preceding sections comprised primarily Caucasian women. In the Women’s Health Initiative Cohort Study, the association of physical activity with breast cancer risk was examined among approximately 74,000 women aged 50 to 70 years, 15% of whom were from race–ethnic minority groups, throughout the United States (McTiernan et al. 2003). At baseline between 1993 and 1998, women reported on their walking and on the frequency and duration spent on light-, moderate-, and vigorous-intensity physical activity. During an average follow-up of 4.7 years, 1780 women developed breast cancer. After adjustment for age, race, geographic region, income, education, body mass index, use of postmenopausal hormones, breastfeeding, hysterectomy status, family history of breast cancer, smoking, parity, age at first birth, age at menarche, age at menopause, mammography, and alcohol use, women who expended no, ≤5, 5.1 to 10, 10.1 to 20, 20.1 to 40, and >40 MET-hours/week in physical activity had relative risks (95% CI) for breast cancer of 1.00 (referent), 0.90 (0.77-1.07), 0.82 (0.68-0.97), 0.89 (0.76-1.00), 0.83 (0.70-0.98), and 0.78 (0.62-1.00), respectively; p for trend = 0.03.