Table 1.
Observational studies of physical activity and disease and mortality events in breast cancer survivors*
| First author (reference), year, country | Sample characteristics | Study design, follow-up, and outcome assessment | Disease stage and treatment data | Physical activity measure and timing | Overall results | Subgroup results |
| Studies that assessed physical activity undertaken before diagnosis | ||||||
| Rohan (23), 1995, Australia | 451 women with breast cancer, 112 breast cancer–specific deaths, No. of deaths from any cause NR. Mean age 55 y; Race or ethnicity NR; diagnosis 1982–1984; pre- and postmenopausal; interviewed at variable times after diagnosis. | Follow-up of a population-based incident cancer case–control study; median follow-up 5.5 y; vital status by record linkage | Tumor diameter, ER and PR status; no treatment data | Interviewer-administered questionnaire assessed leisure-time physical activity, in kcal/wk, and by intensity during the summer and winter seasons during the previous year. | HR of breast cancer–specific death for physical activity of >4000 vs 0 kcal/wk = 0.98 (95% CI = 0.50 to 1.94); P trend = .803. Adjusted for stage and multiple breast cancer risk factors, including BMI and energy intake, but not treatment. | No statistically significant association between total physical activity and risk of death from breast cancer overall or by menopausal status. |
| Enger (24), 2004, United States | 717 women with breast cancer, 251 breast cancer–specific deaths, 263 deaths from any cause. Age range 21–40 y; all subjects white or Hispanic; diagnosis 1983–1989; premenopausal; interviewed 12 mo before diagnosis. | Follow-up of a population-based incident cancer case–control study; median follow-up 10.4 y; vital status and cause of death by record linkage. | Tumor stage, number of lymph nodes involved; no treatment data | Interviewer-administered questionnaire assessed the frequency and duration of regular weekly leisure-time physical activity, in h/wk, during the woman’s lifetime. | HR of breast cancer–specific death for physical activity of ≥5 vs 0 h/wk = 0.78 (95% CI = 0.45 to 1.34); P trend = .31; Adjusted for stage and BMI but not treatment. | None noted |
| Abrahamson (25), 2006, United States | 1264 women with breast cancer, 246 breast cancer–specific deaths, 290 deaths from any cause. Age range 20–54 y; 25% nonwhite; diagnosis 1990–1992; pre- and postmenopausal; interviewed median 4.2 mo after diagnosis. | Follow-up of a population-based prospective incident cancer cohort study; median follow-up 8.5 y (range = 0.25–9.8 y); vital status and cause of death by record linkage. | Tumor stage, ER, and PR status; treatment (surgery, CT, RT, HT) | Interviewer-administered questionnaire assessed frequency and intensity of weekly physical activity, in relative units/wk, during the year before diagnosis, at age 12–13 y, and at age 20 y | HR of death from any cause for physical activity during year before diagnosis of 35.1–98.0 vs 1.6–3.4 units/wk = 0.78 (95% CI = 0.56 to 1.08); P trend = .10; Adjusted for stage but not treatment or BMI. | Women with BMI >25 kg/m2: HR of death from any cause for physical activity during year before diagnosis 35.1–98.0 vs 1.6–3.4 units/wk = 0.70 (95% CI = 0.49 to 0.99), P = .05 |
| Dal Maso (26), 2008, Italy | 1453 women with breast cancer, 398 breast cancer–specific deaths, 503 deaths from any cause. Median age 55 y; race or ethnicity NR; diagnosis 1991–1994; pre- and postmenopausal; interviewed ≥1 y before diagnosis | Follow-up of a multicenter incident cancer case–control study; median follow-up 12.6 y; vital status and cause of death by record linkage. | Tumor size and stage, lymph node status, ER and PR status; no treatment data | Interviewer-administered questionnaire assessed weekly leisure-time and daily occupational physical activity, in h/wk, at age 15–19 y, at age 30–39 y, and at age 50–59 y | HR of death from any cause for leisure-time physical activity before diagnosis of >2 h/wk vs <2 h/wk = 0.82 (95% CI = 0.67 to 1.01); HR of death from breast cancer for leisure-time physical activity before diagnosis of >2 h/wk vs <2 h/wk = 0.85 (95% CI = 0.68 to 1.07); P values not provided. Adjusted for stage but not BMI or treatment. | None noted |
| Friedenreich (27), 2009, Canada | 1231 women with breast cancer, 223 breast cancer–specific deaths, 341 deaths from any cause. Mean age 56 y; predominantly white; diagnosis 1995–1997; pre- and postmenopausal; interviewed median 3.4 mo after diagnosis and limited to stage I–III at diagnosis | Follow-up of a population-based incident cancer case–control study; mean follow-up 8.3 y; vital status by record linkage. | Tumor stage, ER and PR status; treatment (surgery, RT, HT) | Interviewer-administered questionnaire assessed intensity, duration, and frequency of repeated occupational, leisure-time, and household physical activity, in MET-h/wk, during the entire lifetime until diagnosis | HR of death from any cause for total physical activity of >151 vs <95 MET-h/wk/y = 0.94 (95% CI = 0.69 to 1.30); P trend = .38; HR of death from breast cancer for total physical activity of >151 vs <95 MET-h/wk/y = 0.79 (95% CI = 0.53 to 1.17); P trend = .78; Adjusted for stage, treatment, and multiple breast cancer risk factors, including BMI. | Results for leisure-time physical activity suggested stronger effects for all-cause and breast cancer–specific mortality and reduced risks of recurrence and new primary breast cancers. |
| West-Wright (28), 2009, United States; California Teachers’ Study | 3539 women with breast cancer, 221 breast cancer–specific deaths, 460 deaths from any cause. Age range 26–94 y; Race or ethnicity NR; diagnosis 1995–2004; pre- and postmenopausal; interviewed at variable times before diagnosis | Follow-up of a prospective incident cancer cohort study; median follow-up for survivors 64 mo; cause of death by record linkage. | Tumor stage, ER and PR status; no treatment data | Self-administered questionnaire assessed intensity, frequency, and duration of physical activity, in h/wk/y, at six age intervals from high school through age 54 y (long-term) and 3 y before enrollment (recent) | RR of death from any cause for long-term physical activity of >3 vs <0.5 h/wk/y = 0.73 (95% CI = 0.55 to 0.96); P trend = .03; RR of death from breast cancer for long-term physical activity of >3 vs <0.5 h/wk/y = 0.53 (95% CI = 0.35 to 0.80); P trend = .003; Adjusted for stage, comorbidity, and multiple breast cancer risk factors, including BMI and caloric intake. | Similar effects by ER status, stronger effects in localized stage disease, and reduced risk only in women with BMI ≥25 kg/m2. Recent physical activity was not strongly associated with the risk of death from breast cancer. |
| Emaus (29), 2010, Norway; Norwegian Counties Study | 1364 women with breast cancer, 355 breast cancer–specific deaths, 429 deaths from any cause. Mean age 57.5 y; Race or ethnicity NR; diagnosis 1974–2005; pre- and postmenopausal; interviewed mean 11.5 y before diagnosis | Population-based prospective incident cancer cohort study; mean follow-up 8.2 y; cancer diagnosis and cause of death by record linkage. | Tumor stage; no treatment data; used calendar time as proxy for changes in treatment regimens. | Self-administered questionnaire assessed frequency and intensity (hard consists of regular, vigorous training or at least 4 h/wk of exercise; sedentary consists of participation in reading, watching television or other sedentary activities) of usual leisure-time physical activity in the year before a screening mammogram, which occurred at variable times before cancer diagnosis | HR of death from any cause for vigorous vs sedentary physical activity = 0.74 (95% CI = 0.51 to 1.08) ; P trend = .27; HR of death from breast cancer for hard vs sedentary physical activity = 0.75 (95% CI = 0.49 to 1.15); P trend = .41; Adjusted for stage and multiple breast cancer risk factors, including BMI, but not treatment. | Stratified analyses suggested reduced risk of death from any cause associated with hard leisure-time physical activity compared with sedentary physical activity among women with a BMI ≤25 kg/m2 and age >55 y at diagnosis. |
| Hellmann (30), 2010, Denmark; Copenhagen City Heart Study | 528 women with breast cancer, 178 breast cancer–specific deaths, 323 deaths from any cause. Median age 66.9 y; Race or ethnicity NR; diagnosis 1976–2003; pre- and postmenopausal; interviewed mean 6.7 y before diagnosis | Prospective incident cancer cohort study; median follow-up 7.8 y; vital status and cause of death by record linkage. | Tumor stage; treatment (CT, RT, HT) | Self-administered questionnaire assessed leisure-time physical activity, in h/wk, before diagnosis (referent period NR) | HR of death from any cause for leisure-time physical activity of >4 vs <2 h/wk = 1.00 (95% CI = 0.69 to 1.45); P trend = .86; HR of death from breast cancer for leisure-time physical activity of >4 vs <2 h/wk = 1.01 (95% CI = 0.62 to 1.63); P trend = .51; Adjusted for stage, treatment, and multiple breast cancer risk factors, including BMI. | Stratified analyses by menopausal status and stage had limited power to detect statistically significant results. |
| Keegan (31), 2010, Multinational; Breast Cancer Family Registry | 4153 women with BC, Breast cancer–specific deaths NR, 725 deaths from any cause. Age range 18 to ≥60 y; 75% non-Hispanic white, 11% non-Hispanic Asian; diagnosis 1991–2000; pre- and postmenopausal; interviewed mean 19.2 mo after diagnosis | Population-based prospective incident cancer cohort study; median follow-up 7.8 y; vital status by record linkage. | Tumor stage factors and ER and PR status; treatment (CT, HT) | Self-administered questionnaire assessed duration and frequency of leisure-time activities throughout the lifetime at various age intervals and during the 3 y before diagnosis (recent) | HR of death from any cause for lifetime physical activity of >38.2 vs 0 MET-h/wk = 0.93 (95% CI = 0.72 to 1.21); P trend = .74; HR of death from any cause for recent physical activity of >38.2 vs 0 MET-h/wk = 0.77 (95% CI = 0.60 to 1.00); P trend = .10; Adjusted for stage and multiple breast cancer risk factors, including BMI, but not treatment. | Stratified analysis suggested stronger effects for ER-positive disease but no difference by race or ethnicity or BMI. |
| Studies that assessed physical activity undertaken after diagnosis | ||||||
| Borugian (32), 2004, Canada | 603 women with breast cancer, 112 breast cancer–specific deaths, 146 deaths from any cause. Mean age 54.5 y; predominantly white; diagnosis 1991–1992; pre- and postmenopausal; interviewed 2 mo after surgery | Prospective incident cancer cohort study; median follow-up 8.1 y; vital status and cause of death by record linkage. | Tumor size and stage, histology, ER and PR status; treatment (surgery, RT, CT, HT) | Self-administered questionnaire assessed frequency and specific types (such as walking, sports, exercise) of leisure-time physical activity, in times/wk, shortly after diagnosis | RR of death from breast cancer for exercise >1 vs 0 times/wk = 1.0 (95% CI = 0.6 to 1.6); Adjusted for stage and multiple breast cancer risk factors, including caloric intake, but not treatment or BMI. P values not provided. Frequencies of seven different types of physical activity were assessed and none were associated with breast cancer mortality. | No effect modification by menopausal status. |
| Holmes (33), 2005, United States; Nurses’ Health study | 2987 women with breast cancer, 280 breast cancer–specific deaths, 463 deaths from any cause. Age range 30–55 y; Race or ethnicity NR; diagnosis 1984–1998; predominantly postmenopausal; interviewed median 38 mo after diagnosis | Prospective incident cancer cohort study; median follow-up 8 y; vital status by record linkage. | Tumor size, number of metastatic lymph nodes, ER, and PR status; treatment (surgery, RT, CT, HT) | Self-administered questionnaire assessed frequency and intensity of leisure-time physical activity, in MET-h/wk, during the preceding year | RR of death from any cause for physical activity of ≥24 vs <3 MET-h/wk = 0.65 (95% CI = 0.48 to 0.88); P trend = .003; RR of death from breast cancer for physical activity of ≥24 vs <3 MET-h/wk = 0.60 (95% CI = 0.40 to 0.80); P trend = .004; RR of recurrence for physical activity of ≥24 vs <3 MET-h/wk = 0.74 (95% CI = 0.53 to 1.04); P trend = .05; Adjusted for stage, treatment, and multiple breast cancer risk factors, including BMI. Similar statistically significant results were seen for physical activity before diagnosis in women with BMI <25 kg/m2 | Stratified analyses suggested similar effects by menopausal status and stage. Stronger effects were seen for women with hormone receptor–positive tumors. |
| Bertram (22), 2011, United States, Women's Healthy Eating and Living Study (WHEL) update of Pierce (20), 2007, United States | 2361 women with breast cancer, 295 breast cancer recurrences, 195 deaths from any cause. Mean age 54 y; multiethnic cohort; diagnosis 1991–1994; pre- and postmenopausal; interviewed after treatment at baseline and 1 y after intervention | Prospective survivorship cohort study based on all eligible participants from WHEL, a dietary intervention trial; mean follow-up 7.1 y; vital status by record linkage | Tumor grade, stage, and ER and PR status; chemotherapy type and adjuvant therapy | Self-administered questionnaire assessed frequency, duration, and intensity of physical activity, in MET-h/wk, conducted at the time of assessment | HR of death from any cause for physical activity at baseline of 24.7–107 vs 0–2.5 MET-h/wk = 0.47 (95% CI = 0.26 to 0.84); P trend = .08; HR of recurrence for physical activity at baseline of 24.7–107 vs 0–2.5 MET-h/wk = 0.74 (95% CI = 0.50 to 1.10); P trend = .58; Adjusted for stage and multiple breast cancer risk factors, including BMI and energy intake, but not treatment. | None noted |
| Holick (34), 2008, United States; Collaborative Women's Longevity Study | 4482 women with breast cancer, 280 breast cancer–specific deaths, 463 deaths from any cause. Mean age 61.7 y; predominantly white; diagnosis 1988–2001; pre- and postmenopausal; interviewed 2 y after diagnosis | Follow-up of three population-based incident case–control studies; mean follow-up 5.6 y; vital status and cause of death by record linkage | Tumor stage and histology; treatment modality | Interviewer-administered questionnaire assessed frequency and duration of weekly leisure-time physical activity, in MET-h/wk, during the preceding year. | HR of death from any cause for physical activity of ≥21 vs <2.8 MET-h/wk = 0.44 (95% CI = 0.32 to 0.61); P trend < .001; HR of death from breast cancer for physical activity of ≥21 vs <2.8 MET-h/wk = 0.49 (95% CI = 0.27 to 0.89); P trend = .05; Adjusted for stage, treatment, and multiple breast cancer risk factors, including BMI. | No effect modification by age, stage, BMI, or time since diagnosis. |
| Irwin (35), 2008, Unites States; Health, Eating, Activity, and Lifestyle (HEAL) Study | 933 women with breast cancer, 115 breast cancer–specific deaths, 164 deaths from any cause. Mean age 55 y; multiethnic cohort; diagnosis 1995–1998; pre- and postmenopausal; interviewed at a median of 5 mo after diagnosis and again at a median of 2.5 y after diagnosis | Prospective survivorship cohort study; median follow-up 6 y; vital status by record linkage | Tumor stage, ER and PR status; adjuvant therapy and HT | Interviewer-administered questionnaire assessed leisure-time, occupational, and household physical activity, in MET-h/wk, for the year before diagnosis and again at the time of assessment (3 y after diagnosis) | HR of death from any cause for leisure-time physical activity 3 y after diagnosis for ≥9 vs 0 MET-h/wk = 0.33 (95% CI = 0.15 to 0.73); P trend = .046; HR of death from breast cancer for leisure-time physical activity 3 y after diagnosis for 9 vs 0 MET-h/wk = 0.65 (95% CI = 0.23 to 1.8); P trend = .46; HR of death from any cause for leisure-time physical activity during the year before diagnosis of ≥ 9 vs 0 MET-h/wk = 0.69 (95% CI = 0.45 to 1.06); P trend = .045; HR of death from breast cancer for leisure-time physical activity during the year before diagnosis of ≥ 9 vs 0 MET-h/wk = 0.83 (95% CI = 0.49 to 1.38); P trend = .27; Adjusted for stage, treatment, and multiple breast cancer risk, including BMI and fruit and vegetable intake. | Limited statistical power to examine subgroups, but some suggestion of a greater benefit for tumors with more advanced stage and ER-positive status. |
| Sternfeld (36), 2009, United States, Life After Cancer Epidemiology (LACE) Study | 1970 women with breast cancer, 102 breast cancer–specific deaths, 187 deaths from any cause. Age range 18–79 y; predominantly white; diagnosis 1997–2000; pre- and postmenopausal; interviewed at study entry (mean 1.9 y after diagnosis) | Prospective survivorship cohort study; mean follow-up 7.25 y; vital status and cause of death by record linkage. | Tumor size, histology, lymph node involvement, distant metastasis, and ER and PR status; treatment type | Self-administered questionnaire assessed frequency and duration of occupational, household and care giving, leisure-time, and transportation-related physical activity, in MET-h/wk, during the preceding 6 mo | HR of death from any cause for physical activity of ≥62 vs <29 MET-h/wk = 0.76 (95% CI = 0.48 to 1.19); P trend = .20; HR of death from breast cancer for physical activity of ≥62 vs <29 MET-h/wk = 0.87 (95% CI = 0.48 to 1.59); P trend = .41; HR of recurrence for physical activity of ≥62 vs <29 MET-h/wk = 0.91 (95% CI = 0.61 to 1.36); P trend = .78; Adjusted for stage, treatment, and multiple breast cancer risk factors, including weight. | None noted |
| Chen (37), 2011, China; Shanghai Breast Cancer Survival Study (SBCSS) | 4826 women with breast cancer, breast cancer–specific deaths NR, 436 deaths from any cause. Mean age 53.5 y; predominantly Asian; diagnosis 2002–2006; pre- and postmenopausal; interviewed 6, 18, 36, and 60 mo after diagnosis | Population-based cohort study; median follow-up 4.3 y; vital status by record linkage. | Tumor stage, ER and PR status; treatment type, tamoxifen use | Interviewer-administered questionnaire assessed frequency and duration of exercise on a weekly basis, in MET-h/wk, at the time of each interview (after diagnosis) | HR of death from any cause for physical activity at 6 mo after diagnosis of ≥8.3 MET-h/wk vs no exercise = 0.80 (95% CI = 0.63 to 1.02); P trend = .198; HR of death from any cause for physical activity at 36 mo after diagnosis of ≥8.3 MET-h/wk vs no exercise = 0.65 (95% CI = 0.51 to 0.84); P trend < .001; HR of death from breast cancer for physical activity at 6 mo after diagnosis of ≥8.3 MET-h/wk vs no exercise = 0.98 (95% CI = 0.78 to 1.24); P trend = .47; HR of death from breast cancer for physical activity at 36 mo after diagnosis of ≥8.3 MET-h/wk vs no exercise = 0.59 (95% CI = 0.45 to 0.76); P trend = .006; Adjusted for stage, treatment, and multiple breast cancer risk factors, including BMI. | No effect modification by menopausal status, comorbidity, quality of life, or body size. |
| Irwin (38), 2011, United States; Women's Health Initiative (WHI) | 4643 women with breast cancer, 194 breast cancer–specific deaths, 350 deaths from any cause. Mean age 63.7 y; predominantly white; diagnosed before 2005; postmenopausal; interviewed at baseline (average 4.1 y before diagnosis), and at year 3 and year 6 with respect to enrollment | Prospective longitudinal chronic disease incident cohort study; mean follow-up 3.3 y; vital status and cause of death by record linkage | Tumor stage, grade, ER and PR status, HER2/neu status; no treatment data | Self-administered questionnaire assessed frequency, duration, and intensity of physical activity, in MET-h/wk, conducted at the time of the assessment | HR of death from any cause for moderate physical activity before diagnosis of ≥9 vs 0 MET-h/wk = 0.58 (95% CI = 0.40 to 0.69); P trend < .001; HR of death from breast cancer for moderate physical activity before diagnosis of ≥9 vs 0 MET-h/wk = 0.60 (95% CI = 0.40 to 0.90); P trend = .014; Adjusted for vigorous-intensity physical activity and multiple breast cancer risk factors, including energy intake and BMI. HR of death from any cause for moderate to vigorous physical activity after diagnosis of ≥9 vs 0 MET-h/wk = 0.54 (95% CI = 0.38 to 0.79); P trend < .001; HR of death from breast cancer for moderate to vigorous physical activity after diagnosis of ≥9 vs 0 MET-h/wk = 0.61 (95% CI = 0.35 to 0.99); P trend = .049; Adjusted for stage and multiple breast cancer risk factors, including energy intake and BMI, but not treatment. | None noted |
*
ADT = androgen deprivation therapy; BMI = body mass index; CI = confidence interval; CRP =C-reactive protein; CT = chemotherapy; ECG = electrocardiography; ER = estrogen receptor; FASN = Fatty Acid Synthase Gene; FIGO = International Federation of Gynecology and Obstetrics; HEAL = Health, Eating, Activity, and Lifestyle; HEI = Healthy Eating Index; HOMA = homeostasis model assessment; HT =hormonal therapy; HR = hazard ratio; IGF = insulinlike growth factor; IGFBP = insulinlike growth factor binding protein; K-ras = Kirsten rat sarcoma 2 viral oncogene homolog; LPS = lipopolysaccharide; MET = metabolic equivalents; NK = natural killer; NR = not reported; PI3KA = Phosphatidylinositol 3-kinases; PR = progesterone receptor; PSA = prostate-specific antigen; p21 = cyclin-dependent kinase inhibitor 1; p27 = cyclin-dependent kinase inhibitor 1B; p53 = tumor protein 53; RR = relative risk; RM = repetition maximum; RT = radiation therapy; SAA = serum amyloid A; VO2 = aerobic capacity.