University of Southern California Dissertations and Theses

Recreational physical activity and risk of breast cancer

(USC Thesis Other)

Recreational physical activity and risk of breast cancer

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content RECREATIONAL PHYSICAL ACTIVITY AND RISK OF BREAST CANCER by Alpa Vipin Patel A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (EPIDEMIOLOGY) May 2003 Copyright 2003 Alpa Vipin Patel Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3103957 UMI UMI Microform 3103957 Copyright 2003 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 90007 This dissertation, written by ALPA PATEL under the direction of fu.RK Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillment of re quirements for the degree of DOCTOR OF PHILOSOPHY Dean of Graduate Studies DISSERTATION COMMITTEE Chairperson Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Dedication This dissertation is dedicated to my parents, Dr. Vipin and Mrs. Nani Patel, without whom I would have never found success, both professional and personal. Their unwavering commitment and belief in me gave me the strength and determination to fulfill my dreams. They have always been supportive of my work and offered guidance and encouragement while allowing me to find my own way. I am forever indebted and grateful to them. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Acknowledgements I would like to thank my committee members for so graciously taking their time to guide me throughout my doctoral program. I would like to give special thanks to Leslie Bernstein and Jeanne Calle for being not only co-chairwomen of my dissertation committee, but also for being the greatest of mentors. Their work, example, and encouragement continue to inspire me as I begin my own career. I would also like to thank my colleagues and friends at the American Cancer Society and University of Southern California for making a traditionally lonely dissertation process bearable, for listening to my presentations time and time again, and for always offering support and encouragement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents Dedication ii Acknowledgements iii List of Tables v Abstract vii Chapter 1: Introduction 1 Chapter 2: Literature Review-Possible Biologic Mechanisms 2 Chapter 3: Literature Review-Previous Studies 11 Chapter 4: American Cancer Society Cancer Prevention Study II Nutrition Cohort 22 Chapter 5: American Cancer Society Cancer Prevention Study II Nutrition Cohort Manuscript 25 Chapter 6: In Situ Breast Cancer Study 52 Chapter 7: In Situ Breast Cancer Study Manuscript 59 Bibliography 83 Appendices 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Tables Table 1. Creation of analytic cohort, CPS-II Nutrition Cohort, 1992-1997. Table 2. Age-adjusted percentages of various factors at baseline by recreational physical activity MET expenditure, CPS-II Nutrition Cohort, 1992-1997. Table 3. Rate ratios for recreational leisure-time physical activity at various times during a woman’s lifetime and breast cancer risk, CPS-II Nutrition Cohort, 1992-1997. Table 4. Rate ratios for recreational leisure-time physical activity at baseline and breast cancer risk, by stage, CPS-II Nutrition Cohort, 1992-1997. Table 5. Rate ratios for 1992 reported baseline recreational physical activity and postmenopausal breast cancer stratified by various factors, CPS-II Nutrition Cohort, 1992-1997. Table 6. Possible predictors of screening in control population, in situ breast cancer study. Table 7. Associations between various factors and risk of breast carcinoma in situ. Table 8. Average lifetime physical activity and relative odds of in situ breast cancer. Table 9. Hours of physical activity during various time periods and relative odds of in situ breast cancer. Table 10. Hours of physical activity during various time periods and relative odds of in situ breast cancer stratified by family history of breast cancer. Table 11. Frequencies of individual reportable activities by total MET-hrs/wk category, CPS-II Nutrition Cohort, 1992-1997. Table 12. Total MET-hrs/wk for individual reportable activities, CPS-II Nutrition Cohort, 1992-1997. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 13. Rate ratios for combined leisure-time physical activity reported for age 40 (past) and for 1992 baseline (recent) and breast cancer, CPS-II Nutrition Cohort, 1992-1997. Table 14. Rate ratios for combined leisure-time physical activity reported at baseline in 1982 (past) and 1992 (recent) and breast cancer, CPS-II Nutrition Cohort, 1992-1997. Table 15. Comparison of cases and controls on various breast cancer risk factors, in situ breast cancer study. Table 16. Average lifetime physical activity and relative odds of ductal breast carcinoma in situ. Table 17. Odds ratios and corresponding 95% confidence intervals for physical activity and in situ breast cancer by various factors. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Abstract Physical activity has been proposed as a modifiable risk factor for breast cancer because of its effects on circulating sex hormones. Overall, results from previous studies support the hypothesis that regular physical activity may reduce the risk of invasive breast cancer; however, the association between physical activity and in situ breast cancer is not well-understood. We examined the association between physical activity and breast cancer risk in two different study populations. The first study was the American Cancer Society’s Cancer Prevention Study II Nutrition Cohort, a prospective cohort followed from 1992 through 1997. Information on physical activity was obtained in 1992 via a self-administered questionnaire for 72,608 postmenopausal female participants who were cancer-free. During the five- year follow-up, 1,520 incident breast cancer cases were identified among these women. Women who were most physically active (>42.0 MET-hours/week) at baseline had 29% lower incidence rates than active women with the least activity (>0-7.0 MET-hrs/wk) (95% Cl, 0.49-1.02). The difference in risk was largest for localized breast cancer, and for women who did not use HRT at enrollment. The second study was a population-based case-control study of 567 in situ breast cancer patients and 616 properly screened control subjects from Los Angeles County between March 1, 1995 and May 31, 1998. The risk of in situ breast carcinoma (BCIS) was approximately 35 percent lower among women with ever lifetime exercise activity compared to ever inactive women. We observed no linear trend in BCIS risk with increasing levels of exercise activity, measured either by average vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. hours per week or by MET-hours per week (p-trend=0.24 and p-trend=0.27, respectively). When examining risk of BCIS associated with exercise activity during specific time periods, we observed a modest inverse relationship with exercise activity within the ten years after menarche (p-trend=0.18) and a significant inverse trend with exercise activity during the ten years prior to a woman’s reference date (p- trend=0.02) and BCIS risk. These findings suggest that exercise activity, particularly in recent years, may be important in predicting risk of both invasive and in situ breast carcinoma. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 1: Introduction Breast cancer is the leading cause of cancer and the second leading cause of cancer deaths among women in the United States. There will be an estimated 203,500 new breast cancer cases and 39,600 breast cancer deaths in the U.S. in 2002 (American Cancer Society 2002). Many risk factors for breast cancer have been identified, but there are few known modifiable risk factors, such as obesity and alcohol consumption. Many previous studies have implicated the possible role of physical activity in reducing both premenopausal and postmenopausal breast cancer risk. The main focus of this literature review is to examine the possible biologic mechanisms that may mediate an association between physical activity and postmenopausal breast cancer risk, and the results and quality of the existing literature regarding this association. The main objective of the subsequent data analyses is to examine whether physical activity at various points in a woman’s lifetime is associated with a reduced risk of postmenopausal breast cancer. The proposed data analyses will be conducted in the Cancer Prevention Study II (CPS-II) Nutrition Cohort, a prospective cohort study conducted at the American Cancer Society between 1992-1997, and in a population-based case-control study of in situ breast cancer, conducted at the University of Southern California in Los Angeles, CA. 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 2: Literature Review-Possible Biologic Mechanisms The Role of Estrogen Evidence exists that estrogen acts in promotion of breast cancer through increased proliferation of epithelial breast cells, and more recent research suggests that estrogen may also act as an initiator of breast carcinogenesis. In both in vitro (Laidlaw and others 1995; McManus and Welsch 1984) and in vivo (Chang and others 1995) studies, estradiol, a very active form of estrogen, has been shown to increase mitotic activity of breast epithelial cells. Mitosis alone does not induce carcinogenic effects; however, increased mitotic activity increases risk of DNA replication errors. If these errors go uncorrected, a malignant phenotype may result (Feigelson and Henderson 1996). By increasing mitoses, estrogens act to promote breast cell proliferation and thus increase the possibility of mutations, including those that are carcinogenic. Along with the more established role of estrogen as a proliferative agent, there is increasing evidence suggesting that indirect or direct genotoxicity originating from estrogen metabolites, such as the 4-hydroxy catechol metabolite, may contribute to the role of estrogen in breast cancer etiology. Recent animal and molecular epidemiology studies suggest that estrogen metabolites may act as initiators of breast cancer through their ability to cause oxidative DNA damage (Yager 2000). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Physical Activity and Premenopausal Estrogens Cumulative lifetime exposure to estrogen is a key factor in determining a woman’s breast cancer risk (Henderson and others 1988; Henderson and others 1985). During the reproductive years, research has shown that estrogen levels fluctuate during the menstrual cycle: they peak in the late follicular phase of the menstrual cycle, then slightly decline and plateau during the luteal phase at levels that are higher than those observed in the early follicular phase. Therefore, accumulation of estradiol is greatest during the luteal phase of the menstrual cycle (Goebelsmann and Mishell 1979; Thomeycroft and others 1971). Since estrogen exposure fluctuates during a woman’s menstrual cycle, an increase in the number of ovulatory cycles increases cumulative estrogen exposure (Henderson and others 1985). Factors that measure the cumulative exposure of estrogen over a woman’s reproductive life have subsequently been shown to be associated with breast cancer risk. Many studies have shown an association between early age at menarche and late age at menopause and increased breast cancer risk suggesting that a greater number of years of ovarian function and subsequent number of ovulatory cycles are predictive of risk (Brinton and others 1988; MacMahon and others 1973; Pike and others 1981; Trichopoulos and others 1972). For each year the onset of menses is delayed, there is approximately a 20% reduction in risk (Henderson and others 1984). Furthermore, there is a 50% reduction in risk among women whose natural 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. age at menopause is before age 45 compared to after age 55 (Trichopoulos and others 1972). Other studies have shown an association between parity and age at first pregnancy and breast cancer risk, specifically that parity and early age at first full- term pregnancy reduce breast cancer risk relative to nulliparity or late age at first full-term pregnancy (Hsieh and others 1994; Layde and others 1989; Leon 1989; MacMahon and others 1970). Women who have at least one child have been shown to have a 27% lower risk compared to nulliparous women, and women over age 30 at first pregnancy have a 67% higher risk compared to women whose age at first pregnancy was younger than age 18 (Layde and others 1989). High levels of moderate and vigorous physical activity during reproductive years impact markers of ovarian hormone exposure resulting in delayed menarche, increased likelihood of secondary amenorrhea and irregular or anovulatory menstrual cycles, and shortened luteal phases of the menstrual cycle; thus, physical activity is associated with reduced levels of estradiol, progesterone, and follicle-stimulating hormone (FSH), particularly during adolescence (Bernstein and others 1987; Bonen and others 1981; Ellison and Lager 1986; Frisch and others 1981; Shangold and others 1979; Warren 1980). These factors, in part due to the effects of physical activity, would result in a reduction in lifetime ovulatory cycles and cumulative estrogen exposure, therefore potentially reducing risk of breast cancer. However, the association between lighter recreational levels of physical activity and ovarian function is not as clear. Some studies suggest that ovarian function is altered in 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. recreational athletes through lower mean honnone levels or longer menstrual cycle length (Broocks and others 1990; Cooper and others 1996), but data are not as conclusive as those for athletes. Physical Activity and Postmenopausal Estrogens After menopause, there is negligible production of estrogen from the ovaries resulting in a dramatic decrease of circulating estrogens levels compared to premenopausal women. Studies have shown that estrogen levels decline to approximately one-third of the lowest premenopausal levels (Key and Pike 1988). The main endogenous source of postmenopausal estrogen production is from the peripheral conversion of androstenedione to estrone. Androstenedione in peripheral fatty tissue is converted to estrone via the enzyme aromatase. Estrone is then converted to the more active estrogen, estradiol, via the enzyme 17-beta- hydroxysteroid dehydrogenase (17-beta-HSD). Since aromatization of androstenedione to estrone takes place in fatty tissue, there is a larger amount of estrogen production in obese women compared to thinner women (Kelsey and Berkowitz 1988). Postmenopausal obese women have been shown to have significantly higher, as much as 50% to 100% higher, levels of circulating estrogens compared to leaner postmenopausal women (Hankinson and others 1995; Lipworth and others 1996). Furthermore, obese women have decreased levels of sex hormone-binding globulin (SHBG) resulting in higher levels of free circulating estrogen (Kelsey and Berkowitz 1988; Newcomb and others 1995; Sellers 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and others 1992; Verkasalo and others 2001). Unbound estrogen is more biologically active than SHBG-bound estrogen and is associated with increased risk of breast cancer (Hankinson and others 1998b; Toniolo and others 1995). Postmenopausal obesity has been associated with increased risk of breast cancer in previous studies (Chu and others 1991; Huang and others 1997; Petrelli and others 2002; Sonnenschein and others 1999; vandenBrandt and others 2000). Postmenopausal obese women (BMI > 30 kg/m2 ) have a 60% increased risk of breast cancer (95% 0=1.42-1.79) compared to women within normal weight ranges. Lower BMI during premenopausal years and higher BMI during postmenopausal years are associated with increased risk of breast cancer. Therefore, women who gain weight during adulthood are at an increased risk of postmenopausal breast cancer compared to women who maintain a relatively stable weight over adulthood. Women who have gained more than 20 kilograms during adulthood have been shown to have a 40% increased risk of breast cancer compared to women who have maintained a relatively constant weight during adulthood (Huang and others 1997). An exogenous source of postmenopausal estrogen is the use of hormone replacement therapy (HRT). Recent long-term use of postmenopausal HRT has been shown to increase levels of circulating estrogen, thus increasing risk of breast cancer (Colditz and others 1995; Collaborative Group on Hormonal Factors in Breast Cancer 1997; Magnusson and others 1999; Newcomb and others 2002). Risk increases by 2.3% for each year of recent HRT use, and diminishes after more than 5 years since cessation of use (Collaborative Group on Hormonal Factors in Breast 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Cancer 1997). Studies have shown that the increased risk of breast cancer from exogenous hormone use may be more pronounced in lean women (Collaborative Group on Hormonal Factors in Breast Cancer 1997), suggesting that there may be a biologic threshold of estrogen exposure. Overweight women have higher baseline levels of endogenous estrogens, so the addition of exogenous estrogens may not further increase their risk, but leaner women would experience an increase in risk with HRT use. Women who are physically active during their postmenopausal years have decreased levels of serum estrone (Cauley and others 1989; Nelson and others 1988), estradiol (Cauley and others 1989; Nagata and others 2000), and androgens (androstenedione and testosterone) that are precursors to estrogens (Cauley and others 1989; Nelson and others 1988). An association between physical activity and increased levels of SHBG has also been observed (Tymchuk and others 2000). However, the association between the various hormones and physical activity in postmenopausal women has not been as consistent as for premenopausal women, and many studies have found no clear association between hormone levels and physical activity (Nagata and others 2000; Nelson and others 1988; Newcomb and others 1995; Verkasalo and others 2001). High levels of physical activity have been more consistently associated with lower weight, lower BMI, and weight loss (Nelson and others 1988; Thune and others 1998; Tymchuk and others 2000; Verkasalo and others 2001). Thus, the effects of physical activity in postmenopausal women may be 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. due to direct suppression of hormone levels or indirect because physical activity impacts body weight. Physical Activity and Immune Response There is suggestion that the association between physical activity and reduced breast cancer risk may not be completely hormone-related. Some researchers have observed that physical activity works to increase immune response and may reduce risk of chronic disease by increasing production of natural killer (NK) cells that contribute to immune defense (Long 2002; Pedersen and others 1989). This hypothesis would not be specific to breast cancer, but would rather substantiate a protective role of physical activity in most cancers. Previous studies have shown a protective effect of physical activity in relation to overall cancer risk and to many specific cancer sites, such as colon, lung, pancreas, endometrium, thyroid, and testes, with varying degrees of consistency (Colbert and others 2001; Michaud and others 2001; Rossing and others 2001; Thune and Furberg 2001; Wannamethee and others 2001). However, the immune response hypothesis may be further associated with breast cancer because recent evidence suggests that estrogen suppresses NK cell activity (Curran and others 2001; Hanna and Schneider 1983; Seaman and Gindhart 1979); therefore, higher levels of estrogens and suppressed NK cell activity may interact to increase breast cancer risk in physically inactive women. Consequently, women who are physically active may be at a decreased risk of breast cancer because they have both a higher production of NK cells as well as lower levels of estrogen. 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Physical Activity and Insulin Sensitivity Increased levels of insulin have been associated with decreased levels of SHBG, and consequently a higher level of ffee-estradiol that may increase breast cancer risk (Nestler and others 1991). A high level of insulin has also been associated with decreased levels of insulin-like growth factor binding protein-I (IGFBP-I) that results in a higher level of unbound insulin-like growth factor (IGF-I) (Conover and others 1992). Epidemiologic studies show that high IGF levels are associated with increased risk of breast cancer (Agurs-Collins and others 1999; Hankinson and others 1998a). IGF-I has been shown in vitro to act as a mitogen in breast cell lines and synergistically with estradiol to promote mitosis (Figueroa and others 1993; Jones and Clemmons 1995; McGuire and others 1992). IGF-I also plays a role in promoting cell differentiation and transformation, and in suppressing apoptosis (Jones and Clemmons 1995). Consequently, decreased insulin sensitivity and IGFs may work to increase risk of postmenopausal breast cancer (Kazar 1995). Regular physical activity has been associated with increased insulin sensitivity and decreased level of serum insulin (Helmrich and others 1991; Tymchuk and others 2000), and may therefore reduce the risk of postmenopausal breast cancer. Current evidence does not allow clear conclusions to be drawn regarding a possible association between physical activity and IGF levels (Yu and Rohan 2000). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Physical Activity and Energy Balance Energy balance is achieved when energy intake equals energy expenditure resulting in no net change of stored energy in the body. High energy intake coupled with low expenditure leads to excess storage of adipose tissue and results in obesity (Flatt 1993). Excess adiposity, or high BMI, is associated with increased risk of breast cancer (Chu and others 1991; Huang and others 1997; Sonnenschein and others 1999; vandenBrandt and others 2000). The accumulation of adipose tissue over time, or weight gain, is also associated with increased risk of breast cancer (Huang and others 1997). Furthermore, the lack of energy balance resulting in excess adipose tissue is associated with many other potential risk factors for breast cancer such as: insulin resistance (Ballard-Barbash 1997; Bruning and others 1992), increased levels of IGFs (Ballard-Barbash 1997), increased levels of total estradiol (via increased aromatase activity) (Pike and others 1993), increased levels of free- estradiol (due to lower levels of SHBG) (Pike and others 1993; Tymchuk and others 2000), and immunosuppression (Stallone 1994). Maintenance of normal body weight is one of the few known modifiable risk factors for breast cancer, and increased physical activity has been consistently associated with lower BMI and weight maintenance. Therefore, promotion of physical activity may be an effective approach to reducing the risk of breast cancer through weight regulation (Blair 1993; Willett 1997). 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 3: Literature Review-Previous Studies Summary of Results To date, 17 reports from prospective cohort studies (Albanes and others 1989; Breslow and others 2001; Calle and others 1998; Cerhan and others 1998; Dirx and others 2001; Dorgan and others 1994; Fraser and Shavlik 1997; Lee and others 2001b; Luoto and others 2000; Moore and others 2000; Moradi and others 2002; Rockhill and others 1999; Rockhill and others 1998; Sesso and others 1998; Steenland and others 1995; Thune and others 1997; Wyshak and Frisch 2000), six from retrospective cohort studies (Frisch and others 1985; Frisch and others 1987; Moradi and others 1999; Pukkala and others 1993; Vena and others 1987; Vihko and others 1992), 23 from population-based case-control studies (Bernstein and others 1994; Carpenter and others 1999; Chen and others 1997; Coogan and Aschengrau 1999; Coogan and others 1996; Coogan and others 1997; Friedenreich and others 2001a; Friedenreich and others 2001b; Friedenreich and others 2001c; Friedenreich and Rohan 1995; Gammon and others 1998; Gilliland and others 2001; Hu and others 1997; Lee and others 2001a; Marcus and others 1999; Matthews and others 2001; McTieman and others 1996; Mittendorf and others 1995; Moradi and others 2000; Shoff and others 2000; Ueji and others 1998; Verloop and others 2000; Zheng and others 1993), and six from hospital-based case-control studies (D'Avanzo and others 1996; Dosemeci and others 1993; Hirose and others 1995; Levi and others 1999; Mezzetti and others 1998; Taioli and others 1995) have examined the 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. association between physical activity, either leisure or occupational, and breast cancer risk. Five studies yielded more than one publication ((Albanes and others 1989), (Steenland and others 1995), and (Breslow and others 2001); (Coogan and others 1996), (Coogan and others 1997), (Mittendorf and others 1995), and (Shoff and others 2000); (Friedenreich and others 2001a), (Friedenreich and others 2001c), and (Friedenreich and others 2001b); (Frisch and others 1985) and (Frisch and others 1987); (Vihko and others 1992) and (Pukkala and others 1993)); consequently, only the most recent report from each study will be presented resulting in a total of 43 original studies. Of these 43 studies, 36 specifically examined recreational leisure time activity (Bernstein and others 1994; Breslow and others 2001; Carpenter and others 1999; Cerhan and others 1998; Chen and others 1997; D'Avanzo and others 1996; Dirx and others 2001; Dorgan and others 1994; Fraser and Shavlik 1997; Friedenreich and others 2001b; Friedenreich and Rohan 1995; Frisch and others 1987; Gammon and others 1998; Gilliland and others 2001; Hirose and others 1995; Flu and others 1997; Lee and others 2001a; Lee and others 2001b; Levi and others 1999; Luoto and others 2000; Marcus and others 1999; Matthews and others 2001; McTiernan and others 1996; Mezzetti and others 1998; Moore and others 2000; Moradi and others 2002; Moradi and others 2000; Rockhill and others 1999; Rockhill and others 1998; Sesso and others 1998; Shoff and others 2000; Taioli and others 1995; Thune and others 1997; Ueji and others 1998; Verloop and others 2000; Wyshak and Frisch 2000). 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Among the 36 studies that examined recreational leisure-time physical activity, 33 independently examined risk associated with postmenopausal breast cancer. Eight of 13 prospective cohort studies found a significant reduction in postmenopausal breast cancer risk in the group with the most compared to least activity (Breslow and others 2001; Cerhan and others 1998; Dirx and others 2001; Fraser and Shavlik 1997; Moradi and others 2002; Rockhill and others 1999; Sesso and others 1998; Wyshak and Frisch 2000), and one found a non-significant increase in risk with increasing activity (Dorgan and others 1994). Dorgan et.al, in the Framingham Heart Study, found a 60% increase in risk (95% Cl 0.9-2.9) in the highest category of activity versus no activity. This finding, however, was based on one measure of physical activity at baseline over a mean follow-up time of 26 years; furthermore, there was a low correlation between baseline activity levels and activity levels towards the end of follow-up. Therefore, there is potential for a large amount of misclassification of exposure. Nine of 15 population-based case-control studies observed a significant reduction in risk of postmenopausal breast cancer with increasing physical activity (Carpenter and others 1999; Friedenreich and others 2001a; Gilliland and others 2001; Marcus and others 1999; Matthews and others 2001; McTieman and others 1996; Moradi and others 2000; Shoff and others 2000; Verloop and others 2000), and three of five hospital-based case-control studies observed a significant reduction in risk (Hirose and others 1995; Levi and others 1999; Mezzetti and others 1998). No 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. retrospective cohort studies examined the association between recreational physical activity and postmenopausal breast cancer risk. Quality of Previous Studies Overall, the existing literature supports a reduction in risk of postmenopausal breast cancer with increasing recreational physical activity, but researchers are limited in their ability to draw conclusions regarding what time period, intensity, frequency, or duration of exposure is relevant to risk reduction due to various issues related to study design, data analysis, and power. Questions regarding physical activity range from general questions such as “Do you get much exercise in things you do for recreation?” (NHANES-I questionnaire) (Breslow and others 2001) to very detailed reconstructed histories on duration, frequency, and intensity of physical activity that allows researchers to build exposure measures for different time periods (Caipenter and others 1999). In other studies, physical activity is indirectly inferred through study design comparing athletes versus non-athletes (Wyshak and Frisch 2000). Summary measures are then calculated based on the available information from questionnaire data. In some instances, investigators are only able to examine differences between self-reports of little, moderate, or much activity (Breslow and others 2001; Levi and others 1999). In other studies, the researchers examine level of activity based on metabolic equivalents (MET) calculated as a function of energy expenditure for each individual activity and the frequency of participation in each 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. activity (Carpenter and others 1999; Chen and others 1997; Friedenreich and others 2001b; Gammon and others 1998; Gilliland and others 2001; Lee and others 2001b; Matthews and others 2001; McTieman and others 1996; Sesso and others 1998; Shoff and others 2000), but MET conversion values often differ from study to study. Still in other studies, level of activity is measured by frequency of activity during specific time periods over a woman’s lifetime (D'Avanzo and others 1996; Hirose and others 1995; Moradi and others 2000; Rockhill and others 1999; Thune and others 1997). In addition to varying methods of summary measure calculations, many questionnaires only allow for limited number of activities to be reported for each time period (Matthews and others 2001), therefore, activity histories may underestimate the actual activity level of extremely active women. Not only does the existing literature differ by methods used in collecting information regarding physical activity, but studies also differ dramatically in the time frame of exposure. It is still unclear what time period of activity during a woman’s life is significant, if any, in predicting postmenopausal breast cancer risk. Women with higher estrogen levels during early adolescence (Apter and Vikho 1983), premenopausal adulthood (Bernstein and others 1990b), and postmenopausal adulthood (Bernstein and others 1990a) have been shown to have an increased risk of breast cancer. Premenopausal activity levels may work to reduce lifetime estrogen levels, and thus reduce risk of both premenopausal and postmenopausal breast cancer risk. Postmenopausal activity levels may work to suppress circulating estrogen levels or reduce adiposity, thus inferring a reduction in breast cancer risk. Combined 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. premenopausal and postmenopausal, or lifetime activity, may be the important measure if a sustained reduction in estrogen levels throughout a woman’s lifetime is necessary to infer a protection against breast cancer. Epidemiologic studies are also limited in their ability to accurately measure lifetime or past physical activity because studies are dependent on the participant’s ability to recall past activity. Studies assessing the reliability of long-term recall of physical activity have shown that agreement between baseline and recalled activity was generally about 75%. Recall of vigorous activity was more accurate than for less intensive activities, and was especially low for activities such as walking (Blair and others 1991; Falkner and others 1999). Participant characteristics, such as age, race, education, body weight, and marital status, have been shown to have little association with quality of recall of past physical activity (Falkner and others 2001). In case-control studies, cases may be more likely to accurately recall past activity; therefore, there is likely differential recall bias between cases and controls potentially resulting in a bias towards the null. In cohort studies, the inability to accurately recall past activity levels would most likely be non-differential and increase the likelihood of misclassification of exposure leading to a bias towards the null. Measures of intensity are subjective and individual physiologic response to specific activities may vary between participants. The hypothesized biologic mechanisms mediating an association between physical activity and postmenopausal breast cancer are based on observed physiologic effects of moderate to vigorous 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. exercise. However, there are no studies examining direct physiologic response to less intense activities. The effect of low-intensity to moderate-intensity (such as household activities, gardening, dancing, or walking) activity is still unclear, but may be of importance since a large portion of activity among postmenopausal and elderly women is not vigorous (Evenson and others 2002). Although previous studies examine the association between leisure physical activity, such as walking, biking, swimming, aerobics, and dancing, most studies have not examined the effects of other activities such as gardening, housework, or shopping, which may represent a large portion of postmenopausal and elderly non-vigorous physical activity levels. To date, only seven studies have included any measure of household, shopping, or other non-exercise leisure, activity (Cerhan and others 1998; D'Avanzo and others 1996; Dirx and others 2001; Gilliland and others 2001; Levi and others 1999; McTieman and others 1996; Moore and others 2000), but no study has examined these activities independent of recreational physical activity. A study examining the effect of retirement on leisure physical activity levels showed levels of recreational activity, including walking, household or gardening activity, increased after retirement regardless of their physical activity while they were working adults (Evenson and others 2002). Therefore, in studies of postmenopausal breast cancer, especially those of mostly retired persons, inclusion of activities such as gardening or other housework may be important when examining the effects of physical activity during the postmenopausal years. 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In addition to the types of activities that may be relevant, the amount of each activity needed to infer a benefit in risk against postmenopausal breast cancer is unknown. Few studies have examined the threshold of frequency and duration and found that approximately four hours of moderate-intensity activity per week is needed to provide significant protection against disease (Matthews and others 2001; McTieman and others 1996; Thune and others 1997). However, low-intensity activity, combined with moderate or vigorous intensity activities, during postmenopausal years has not been adequately examined, but may be of importance, especially in the promotion of physical activity within an aging population. Confounding and Effect Modification Energy intake has been the most commonly unmeasured potential confounder when examining the physical activity and breast cancer association. Only 11 of the 33 studies assessing postmenopausal breast cancer risk adjusted for energy intake (Breslow and others 2001; D'Avanzo and others 1996; Dirx and others 2001; Friedenreich and others 2001b; Friedenreich and Rohan 1995; Gammon and others 1998; Gilliland and others 2001; Hirose and others 1995; Levi and others 1999; Mezzetti and others 1998; Moore and others 2000). With the exception of five studies (Dirx and others 2001; Dorgan and others 1994; Marcus and others 1999; Matthews and others 2001; McTieman and others 1996), all previous studies adjusted for obesity (using body mass index). Most previous studies also adjusted for other potential confounders, such as age, parity, family history, HRT use, and 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. alcohol intake, but for the most part, the association between physical activity and postmenopausal breast cancer remained unaffected by any potentially confounding factors. However, the lack of proper control for potential confounding factors may lead to residual confounding resulting in biased risk estimates (Dorgan and others 1994; Luoto and others 2000; Taioli and others 1995). Potential effect modification by various factors has been assessed in some previous studies. Statistically significant effect modification by family history (Chen and others 1997; Friedenreich and others 2001c; Gammon and others 1998; Moore and others 2000; Rockhill and others 1999; Verloop and others 2000) or energy intake (D'Avanzo and others 1996; Dirx and others 2001; Gammon and others 1998; Levi and others 1999; Moore and others 2000) has not been observed in any previous studies. Of the eight studies that examined potential effect modification by parity (Chen and others 1997; Friedenreich and others 2001c; Gammon and others 1998; Moore and others 2000; Moradi and others 2000; Rockhill and others 1999; Verloop and others 2000; Wyshak and Frisch 2000), two studies found a greater reduction in risk with increasing physical activity among nulliparous women compared to parous women (Friedenreich and others 2001c; Moradi and others 2000). In addition, three studies examined effect modification by adult weight gain (Carpenter and others 1999; Dirx and others 2001; Rockhill and others 1999), and one study found a greater reduction in risk among women who had less than a 17% increase in weight (Carpenter and others 1999). 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. One may speculate that the influence of moderate levels of physical activity on hormone levels in women with a favorable estrogen profile (i.e. lower baseline levels of hormones) may be sufficient to reduce risk, whereas women with higher levels of baseline circulating estrogens, such as current HRT users or obese women, may not experience a reduction in risk with moderate physical activity. Women with higher baseline estrogen levels may require more vigorous and frequent activity to substantiate a reduction in risk. Only three previous studies have examined potential effect modification by HRT, but none found a significant interaction (Friedenreich and others 2001b; Gammon and others 1998; Moore and others 2000). There is mounting evidence that active lean women may experience a greater risk reduction than active overweight women (Moradi and others 2000; Thune and others 1997). Of the 17 studies that examined potential effect modification by body mass (Chen and others 1997; D'Avanzo and others 1996; Dirx and others 2001; Friedenreich and others 2001c; Gammon and others 1998; Lee and others 2001a; Levi and others 1999; Luoto and others 2000; Marcus and others 1999; Matthews and others 2001; Moore and others 2000; Moradi and others 2000; Rockhill and others 1999; Sesso and others 1998; Shoff and others 2000; Thune and others 1997; Verloop and others 2000), only two found a greater reduction in risk among leaner women (Moradi and others 2000; Thune and others 1997). In addition to the potential role of obesity as a confounder or effect modifier, there is the possibility that obesity may be on the pathway between physical activity and breast cancer development. There is a high correlation between low physical 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. activity and high BMI (Thune and others 1998; Tymchuk and others 2000; Verkasalo and others 2001), such that physical activity may be a proxy measure for BMI. For example, physical activity around the time of menopause may reduce postmenopausal obesity, and may subsequently prevent the initiation of malignant transformation of breast cells (Williamson and others 1993). Therefore, physical activity may appear to reduce breast cancer risk, albeit mediated through BMI as the etiologic factor. As previously stated, most studies of physical activity and breast cancer adjusted for BMI, and an overall suggestion of decreased risk persisted implying that BMI may not be entirely on the causal pathway; however, the role of obesity in the physical activity and breast cancer association remains unclear. Differentiating between effects of confounding, effect modification, and factors in the causal pathway is an ongoing epidemiologic challenge in examining the association between physical activity and postmenopausal breast cancer risk. Discrepancies in study results may be due to underlying differences in study population characteristics, a lack of sufficient statistical power to detect a true difference between subgroups, or chance. It should also be noted that the study publications reporting no effect modification have not provided details of analyses; therefore, it is difficult to ascertain whether or not the failure to observe significant effect modification is due to a lack of sufficient statistical power. 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 4: American Cancer Society Cancer Prevention Study II Nutrition Cohort Study Background In 1982, approximately 575,000 men and 675,000 women from all 50 states, District of Columbia, and Puerto Rico were enrolled in the American Cancer Society Cancer Prevention Study II (CPS-II), a prospective cohort study designed to examine associations between various potential risk factors and cancer and all other causes of mortality. Participants were recruited using over 77,000 American Cancer Society volunteers, and each participant completed a four-page self-administered questionnaire at baseline regarding demographic, lifestyle, height, weight, personal and family history of disease, medication and vitamin use, and reproductive factors (Garflnkel 1985). Mortality follow-up is conducted via linkage of the entire cohort to the National Death Index every two-years (Calle and Terrell 1993) and is ongoing. In 1992, a subset of the initial cohort, approximately 84,000 men and 97,000 women in 21 states, was successfully enrolled in a second study called the Cancer Prevention Study II Nutrition Cohort. Information regarding diet and other potential risk factors for cancer, such as smoking, physical activity, medications, screening practices, reproductive factors, and family history was collected. In the fall of 1992, the 10-page questionnaire was sent to 159,716 CPS-II men aged 50-69 and their 140,780 CPS-II spouses. In response to this single mailing, completed, usable questionnaires were received from 55,105 men and 50,937 women who were still residing in the 21 states. 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Questionnaires were re-mailed in the fall of 1993 to all non-respondents from the first mailing, to an additional 66,371 men who were aged 70-74, and an additional 149,804 women who were aged 70-74 or who were aged 50-69 but had not been included in 1992 because they were not living with a CPS-II male spouse. With this additional mailing, completed, usable questionnaires were received from 86,406 men and 97,788. This cohort is currently being followed for cancer incidence and all causes of mortality. Follow-up questionnaires are sent every two years (beginning in 1997) to collect updated exposure information and to ascertain newly diagnosed cancers. The first follow-up questionnaire was mailed in 1997/1998 to the 176,537 living members of the Nutrition Cohort, and approximately 91% of the participants returned the questionnaire (n=l 60,403). The most recent questionnaire was sent to the 166,773 living cohort members in 1999/2000 and yielded a response rate of approximately 90% (n=l 50,362). For the present study, an incident cancer case is defined as a cancer (other than non-melanoma skin cancer) that occurred during the interval between the time of enrollment in 1992 and August 31, 1997. Incident cancer cases were first identified by self-report on the 1997 follow-up questionnaire and then verified through medical record acquisition. Linkage with state cancer registries was used as needed for follow-up of reported cases when medical records were not available. Previous studies linking cohort participants with state cancer registries have shown that the Nutrition Cohort participants were highly accurate (93% sensitivity) in 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reporting any past cancer diagnosis; therefore, we did not exclude those cases that were self-reported for which a confirmed diagnosis was not obtained (Bergmann and others 1998). Incident breast cancers were classified according to International Classification o f Diseases, Ninth Revision, ICD-9, codes (174.0-174.9) (World Health Organization 1977). Additional breast cancer cases were identified as interval deaths through automated linkage of the cohort with the National Death Index with the death certificate listing breast cancer as a primary or contributory cause of death during the interval between the date of enrollment and August 31, 1997. Diagnosis confirmation for interval breast cancer deaths was obtained through linkage with state cancer registries. Thus far 82% of incident breast cancers have been verified. General summary stage (GSS) for all breast cancers was categorized according to the SEER Summary Staging Guide as in situ, localized, regional or distant metastasis, or unknown (National Institutes of Health 1977). All aspects of the CPS-II protocol have been reviewed and approved by the Emory University School of Medicine Human Investigations Committee. Furthermore, signed inform consent was obtained from each participant with cancer prior to obtaining medical records. 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 5: American Cancer Society Cancer Prevention Study II Nutrition Cohort- Manuscript Introduction Cumulative lifetime exposure to estrogen is a key factor in determining a woman’s risk of breast cancer (Henderson and others 1988). Studies have shown that early age at menarche, late age at menopause, nulliparity, late age at first full- term pregnancy, postmenopausal obesity, adult weight gain, and postmenopausal hormone replacement therapy (HRT) use are associated with increased breast cancer risk (Brinton and others 1988; Collaborative Group on Hormonal Factors in Breast Cancer 1997; Hsieh and others 1994; Newcomb and others 2002; Sonnenschein and others 1999; Trichopoulos and others 1972; vandenBrandt and others 2000). Physical activity has been proposed as a potential modifiable risk factor for breast cancer because of its effects on circulating sex hormones and weight gain. To date, 33 original reports from observational studies have examined the association between physical activity at various points in a woman’s lifetime and postmenopausal breast cancer risk (Breslow and others 2001; Cerhan and others 1998; Dirx and others 2001; Gilliland and others 2001; Lee and others 2001a; Lee and others 2001b; Luoto and others 2000; Matthews and others 2001; Moradi and others 2002; Vainio and Bianchini 2002). Overall, results from previous studies support the hypothesis that regular physical activity may reduce the risk of breast cancer among postmenopausal women. However, it remains unclear whether early or 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. late-life physical activity is important for postmenopausal women. Of the 33 previous studies, ten specifically examined the association of physical activity during the postmenopausal years and breast cancer risk (Cerhan and others 1998; Dirx and others 2001; Friedenreich and others 2001b; Gilliland and others 2001; Lee and others 2001b; Moore and others 2000; Moradi and others 2000; Rockhill and others 1999; Sesso and others 1998; Thune and others 1997), and all but one found that postmenopausal physical activity is associated with lower breast cancer risk (Cerhan and others 1998; Dirx and others 2001; Friedenreich and others 2001b; Gilliland and others 2001; Lee and others 2001b; Moradi and others 2000; Rockhill and others 1999; Sesso and others 1998; Thune and others 1997). It is also unclear whether non- recreational physical activity, such as housework, gardening, or shopping contribute an additional benefit to breast cancer risk. The contribution of non-recreational physical activity may be valuable since it is an important component of overall physical activity among older, retired persons (Evenson and others 2002). We examined the association of various measures of physical activity with postmenopausal breast cancer risk among women in the American Cancer Society Cancer Prevention Study II (CPS-II) Nutrition Cohort, a large prospective study in the U.S. Methods: Study Cohort and Follow-up Women in this analysis were drawn from the 97,787 female participants in the CPS-II Nutrition Cohort, which was established in 1992 by the American Cancer 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Society as a subgroup of the larger 1982 CPS-II cohort (Calle and others 2002). The great majority of participants were 50-74 years of age at enrollment in 1992, and they completed a ten-page self-administered questionnaire that included questions on demographic, medical, reproductive, behavioral, environmental, and dietary factors. A follow-up questionnaire was sent to cohort members between September 1997 and August 1998 to update exposure information and to ascertain newly diagnosed cancers. Cohort members who died during the interval were identified by routine interval linkage of the entire cohort to the National Death Index. The response rate to the 1997/1998 questionnaire among living cohort participants was 91%. We excluded from this analysis women who were lost to follow-up from 1992 to 1997-1998 (n=7,592), who were missing year of diagnosis of breast cancer (n=3), who reported prevalent cancer (except non-melanoma skin cancer) at baseline (n=l 1,599), or who were not postmenopausal in 1992 (n=4,851). Also excluded were women who left the baseline physical activity question (on all seven activities) blank (n=l,134). After all exclusions, the final analytic cohort consisted of 72,608 women with a mean age at study entry of 62.7 + 6.1 (Table 1). A total of 1,520 incident breast cancers diagnosed between the date of enrollment and August 31, 1997 were included in this analysis. Of these, 1,373 cases were identified by self-report on the 1997-1998 questionnaire and subsequently verified from medical records (n=l,138) or linkage with state cancer registries (n=235). Verified incident breast cancer cases also included a small number (n=8) identified during confirmation of another reported cancer diagnosis. Sixty-one 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. incident cases were identified as interval deaths through automated linkage of the cohort with the National Death Index. For these cases, the death certificate listed breast cancer as a primary or contributory cause of death (International Classification of Diseases, Ninth Revision, codes 174.0-174.9)(World Health Organization 1977) during the interval between the date of enrollment and August 31,1997. For 39 of the interval deaths, additional information was obtained through linkage with state cancer registries. Previous studies linking cohort participants with state cancer registries have shown that the Nutrition Cohort participants were highly accurate (93% sensitivity) in reporting any past cancer diagnoses (Bergmann and others 1998); therefore, we also included 78 self-reported breast cancers for which confirmed diagnosis was not obtained. For those cases with medical or registry records (n=l,420), we then classified by general summary stage (GSS) as recorded on the records. Cases were grouped as in situ (stage I, n=205), localized (stage II, n-880), regional and distant/systemic disease (stage III and IV, n=290), or unknown GSS (n=45). 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 1. Creation of analytic cohort, CPS-II Nutrition Cohort, 1992-1997. # of women (%) # of breast cancer cases (%) Total 97,787 1,763 Exclusions Lost to follow-up 7,592 (7.8) 0 (0.0) Missing year of diagnosis 3 (0.01) 3 (0.2) Prevalent breast cancer 6,006 (6.1) 0 (0.0) Other prevalent cancer 5,593 (5.7) 112(6.4) Pre-, peri-, or unknown menopausal status 4,851 (5.0) 106 (6.0) Missing exercise data 1,134(1.2) 22(1.2) Total exclusions 25,179 (25.7) 243 (13.8) Total available for analysis 72,608 (74.3) 1,520 (86.2) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Methods: Assessment of Physical Activity Baseline recreational physical activity information was collected using the question ‘During the past year, what was the average time per week you spent at the following kinds of activities: walking, jogging/running, lap swimming, tennis or racquetball, bicycling or stationary biking, aerobics/calisthenics, and dancing?’ Response to each activity could be ‘none’, ‘1-3 hours per week’, ‘4-6 hours per week’, or ‘7+ hours per week’. Summary MET-hours/week were calculated for each participant. A MET, or metabolic equivalent is the ratio of metabolic rate during a specific activity to resting metabolic rate (Ainsworth and others 1993). The summary MET score for each participant was calculated by multiplying the hours spent engaged in each activity (zero for ‘none’, one for ‘1-3 hours per week’, four for ‘4-6 hours per week’, and seven for ‘7+ hours per week) times the MET score estimated for each activity by Ainsworth et.al. (Ainsworth and others 1993). Due to the older age of this population, MET-hours per week were calculated using the lowest value of hours spent and moderate intensity MET values for each activity such that summary measures would be estimated conservatively. The following MET scores were used (Ainsworth and others 1993): 3.5 for walking, 7.0 for jogging/running, 7.0 for lap swimming, 6.0 for tennis or racquetball, 4.0 for bicycling/stationary biking, 4.5 for aerobics/calisthenics, and 3.5 for dancing. In addition to recreational leisure activity at baseline, non-recreational leisure activity was also examined based on information collected from the question ‘During the past year, what was the average time per week you spent at the following kinds 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of activities: gardening/mowing/planting, heavy housework/vacuuming, heavy home repair/painting, and shopping?’. The above algorithm was used to calculate MET- hrs/wk using the following values for each activity (Ainsworth and others 1993): 3.0 for gardening/mowing/planting, 2.5 for heavy housework/vacuuming, 3.0 for heavy home repair/painting, and 2.5 for shopping. In 1992, we also asked participants to recall physical activity at age 40 based on the question, ‘At age 40, what was the average time per week you spent at the following kinds of activities: walking, jogging/running, lap swimming, tennis or racquetball, bicycling or stationary biking, aerobics/calisthenics, and dancing?’, which was then summarized using the same method as baseline recreational activity mentioned above. Another measure of past physical activity was available using the original 1982 CPS-II questionnaire data, where participants estimated behavior ten- years prior to baseline. In 1982, participants were asked ‘How much exercise do you get (work or play): none, slight, moderate, heavy?’ Physical activity as recalled at age 40 and activity in 1982 were combined with baseline 1992 exposure information to assess whether risk of breast cancer was reduced among women who consistently reported being physically active. Methods: Statistical Analysis We used Cox proportional hazards modeling (Cox 1972) to calculate hazards rate ratios (RR) and corresponding 95% confidence intervals (Cl) to examine the relationship between physical activity measures and breast cancer. For each analysis, 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. we assessed risk in two models, one adjusted only for age and the other adjusting for age and potential confounding factors. Baseline recreational and age 40 physical activity were categorized in MET-hrs/wk as none, >0-7.0, >7.0-17.5, >17.5-31.5, >31.5-42.0, or >42.0. Baseline non-recreational leisure activity was categorized in MET-hrs/wk as none, >0-5.0, >5.Q-<10.0, 10.0-<18.5, or >18.5. All hazard ratios employ women reporting >0-7.0 MET-hrs/wk as the referent group. Women who reported being inactive were not used as the referent group because of the possibility that their complete inactivity may be due to underlying conditions related in some way to breast cancer risk. If inactive women suffer from other health conditions that are hormone-related and impair their ability to engage in physical activity, the association between inactivity and breast cancer risk may be confounded. Women with missing information for activity at age 40 (1,278 women), 1982 exercise (902 women), or baseline non-recreational activity (770 women) were excluded from models that included those variables. All Cox models were stratified on exact year of age at enrollment. Potential confounders included in the multivariate models were race (white, black, other/missing), education (< high school graduate, some college, college graduate, missing), family history of breast cancer in m other or sisters (yes, no), history of breast lumps and/or cysts (yes, no), recency of mammography (never had a mammogram, had mammogram within past year, 1-3 years ago, over 3 years ago, missing), smoking (never, current, former, missing), baseline alcohol intake (never, <1 drink/day, 1 drink/day, >1 drink/day, missing), parity (nulliparous, one live birth, 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2-3 live births, >3 live births, missing), age at menarche (<12, 12, 13, >13, missing), age at natural or surgical menopause (<45, 45-49, 50-54, >55, unknown), oral contraceptive use (never, < 5 years, 5-9 years, >10 years, missing), total caloric intake (kcals/day) in quartiles, HRT use (never, current, former, ever user but unknown if a current user, missing), BMI (weight(kg)/height(m) ) (<22.0, 22.0- <25.0, 25.0-<27.0, 27.0-<30.0, >30.0, missing), and weight change from age 18 to 1992 (lbs.) (>5 loss, 5 loss-5 gain, >5-15 gain, >15-25 gain, >25-35 gain, >35 gain, missing). Trend tests for physical activity models using MET-hrs/wk were obtained by assigning the mean MET value to each category, and trend for 1982 exercise used an ordinal value (1 to 4) for the four reported levels of activity. To test whether any of the potential confounders described above modified the association between baseline recreational physical activity and breast cancer risk, we constructed multiplicative interaction terms between baseline MET-hrs/wk and all other risk factors. Due to small numbers in some strata, categories of potential effect modifiers were sometimes collapsed. We also assessed whether attained age modified the association using a time-varying covariate for age until the end of follow-up (diagnosis date, death date, or August 31, 1997) for each participant. Statistical interaction was assessed in multivariate models using the likelihood ratio test and a p-value <0.05 was considered statistically significant (Kleinbaum and others 1982). 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Results Nine percent (n=6,659) of women in this study population reported no recreational physical activity at baseline, and 3% (n=l,958) reported activity >42.0 MET-hrs/wk. Among women who reported any recreational physical activity at baseline, the median MET expenditure was 9.5 MET-hr/wk, which is equivalent to approximately three hours of moderately paced walking per week. Women who were physically active were more likely to use HRT, drink any alcohol, be non-smokers, and have had a mammogram within the year prior to baseline. Active women had a lower BMI at enrollment (1992) and were less likely to have gained weight since age 18. There was also a high correlation between baseline recreational physical activity and non-recreational physical activity, activity recalled at age 40, and activity reported in 1982; active women at all levels were also more likely to engage in low versus moderate or high intensity activities (Table 2). Women in the highest category of recreational physical activity (> 42.0 MET-hr/wk) had a lower relative risk of postmenopausal breast cancer than women who reported some physical activity not exceeding 7.0 MET-hr/wk (RR=0.71, 95% Cl, 0.49-1.02), although the result was not statistically significant (Table 3). A test of trend including women who reported no recreational physical activity was of borderline statistical significance (p-trend=0.08); however, among women reported any recreational physical activity, there was a significant trend with increasing physical activity (p-trend=0.03). 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. We also examined the association of breast cancer risk with both the 1992 report of physical activity at age 40 and with exercise reported prospectively in 1982. Women who reported being the most active at age 40 (>42.0 MET-hr/wk) had a RR for breast cancer of 0.79 (95% Cl, 0.61-1.03), but there was no clear gradient of decreasing risk with greater activity at this age (p-trend=0.31, p-trend=0.36 among women reporting any physical activity). A small gradient was observed with exercise reported in 1982, but this was statistically insignificant (p-trend=0.33, p-trend=0.16 among exercisers only) (Table 3). Furthermore, the association was not stronger among women who reported being physically active both at baseline and at age 40 or ten-years prior in 1982 compared to women who reported only recreational physical activity in 1992 (data not shown). After adjustment for recreational physical activity, there was no additional change in the relative risk for breast cancer by non- recreational physical activity (RR=0.95, 95% Cl, 0.82-1.10 for >18.5 MET-hr/wk vs. >0-5.0 MET-hr/wk non-recreational activity, p-trend-0.32). We assessed the association between recreational physical activity at baseline and breast cancer, by stage of disease. The risk of localized breast cancer was most strongly associated with physical activity (RR=0.55, 95% Cl, 0.38-0.80 for >31.5 MET-hr/wk vs. >0-7.0 MET-hr/wk, p-trend=0.02 among women who reported any recreational physical activity at baseline). No inverse association was seen for in situ breast cancer (p-trend=0.94 among women reported any recreational physical activity) or regional and distant breast cancer (p-trend=0.62 among women who reported any recreational physical activity)(Table 4). 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There were no statistically significant interactions between baseline recreational physical activity levels and attained age or any of the other potential risk factors included in this analysis. The inverse association between recreational physical activity and lower risk of breast cancer was marginally stronger among women who were not currently using HRT at baseline than among those who used HRT (p-interaction=0.09). There is also a suggestion that physical activity may have a greater impact on women who do not drink alcohol and who are leaner (Table 5). 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2. Age-adjusted percentages of various factors at baseline by recreational physical activity MET expenditure, CPS-II Nutrition Cohort, 1992-1997. Recreational leisure-tim e activity M ET expenditure (total n=72,608) >17.5- >31.5- N one >0-7.0 >7.0-17.5 31.5 42.0 >42.0 V ariable N n=6.659 n=24.739 n=23.478 n=13.436 n=2.338 n=1.958 M edian M ET- hrs/w k — 0 3.5 13.5 24.5 35.5 52.5 Intensity o f activities1 Low — 0.0 97.6 92.2 89.2 77.5 60.9 M oderate — 0.0 2.4 7.8 10.8 22.5 39.1 Age at baseline <50 568 10.9 31.7 31.5 20.6 2.8 2.5 50-59 22,889 9.4 35.1 31.8 17.5 3.4 2.7 60-69 37,700 8.8 33.3 32.9 19.2 3.2 2.7 70+ 11,451 9.8 34.7 31.7 18.3 2.9 2.6 Race W hite 70,701 9.1 34.0 32.4 18.5 3.2 2.7 Black 1,053 11.0 36.8 29.4 16.4 3.6 2.8 Other 854 10.4 32.9 31.8 19.4 2.8 2.7 Education <HS grad 26,996 11.7 36.3 30.4 17.2 2.5 1.9 Some 22,778 8.5 34.2 32.5 18.6 3.4 2.7 college >college 22,338 6.8 31.2 34.5 20.0 3.9 3.7 graduate M issing 496 9.8 34.6 33.2 17.9 1.8 2.7 Age at m enarche <12 14,081 9.4 34.1 32.1 18.5 3.2 2.7 12 18,321 9.1 34.0 32.7 18.5 3.2 2.5 13 21,312 8.6 34.2 32.9 18.4 3.1 2.7 >13 17,752 9.7 34.0 31.5 18.6 3.3 2.9 M issing 1,142 9.8 33.5 31.2 17.8 4.1 3.6 Age at m enopause <45 17,330 10.2 34.1 31.5 18.5 3.1 2.6 45-49 18,349 9.6 34.5 31.7 18.5 3.1 2.6 50-54 28,674 8.4 34.1 32.9 18.4 3.3 2.7 >55 6,707 8.5 32.3 33.7 19.4 3.2 2.9 M issing 1,548 10.1 35.0 32.8 16.2 3.1 2.9 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2 (continued). BMI W eight change(age 18 to 1992) (lbs.) <22.0 16,627 7.6 30.5 32.4 21.0 4.4 4.2 22.0-<25.0 21,168 7.0 32.2 33.8 20.4 3.6 3.0 25.0-<27.0 11,481 8.3 34.6 33.2 18.7 3.0 2.2 27.0-<30.0 11,219 10.6 36.8 31.9 16.4 2.6 1.7 >30.0 11,048 15.1 39.1 29.2 13.6 1.6 1.4 Missing 1,065 11.8 38.2 30.5 14.9 2.3 2.4 > 5 loss < 5 loss - 3,838 9.0 28.3 31.4 22.1 4.4 4.8 < 5 gain 7,479 6.2 29.3 32.7 22.4 4.7 4.6 >5-15 gain 11,789 6.7 30.3 33.8 21.1 4.6 3.5 >15-25 gain 12,733 7.3 33.1 33.8 19.8 3.3 2.7 >25-35 gam 11,369 8.5 35.1 32.7 18.4 3.0 2.2 >35 gain 23,995 12.5 38.2 30.7 14.9 2.0 1.6 M issing 1,405 11.4 36.0 31.9 15.8 2.3 2.6 Family history of breast cancer No Yes 62,549 10,059 9.2 9.0 34.0 34.3 32.4 32.1 18.5 18.5 3.2 3.3 2.7 2.8 HRT use Never 35,013 10.0 34.4 31.4 18.4 3.1 2.7 Current 24,929 7.9 33.3 33.8 18.8 3.5 2.8 Form er 10,318 9.1 34.6 32.4 18.1 3.2 2.6 Ever use, unknow n 1,388 9.2 32.9 31.6 19.7 3.1 3.6 status M issing 960 12.7 35.7 29.5 17.1 2.4 2.6 # o f live births None 5,443 9.7 34.9 31.4 18.5 2.9 2.7 1 5,235 10.5 34.6 31.0 18.0 3.5 2.5 2-3 38,392 8.9 33.6 32.8 18.7 3.3 2.8 >4 22,210 9.1 34.7 32.2 18.4 3.1 2.5 M issing 1,328 10.7 32.9 29.9 19.1 3.4 3.9 ■moking N ever 39,851 8.4 35.6 32.7 18.1 2.9 2.3 Current 5,984 15.4 35.4 26.7 16.9 2.8 2.8 Former 25,776 8.9 31.4 33.2 19.5 3.8 3.2 M issing 997 10.2 35.4 29.0 18.5 2.6 4.2 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2 (continued). Alcohol intake N on-drinker 33,262 10.9 36.4 31.1 17.1 2.5 2.1 < l/d ay 20,207 7.6 32.6 34.1 19.4 3.5 2.8 1/day 10,437 6.8 30.8 34.1 20.4 4.3 3.6 > l/d ay 5,770 8.3 30.7 31.7 20.6 4.6 4.1 M issing 2,932 10.2 36.5 28.9 18.0 3.1 3.4 Caloric intake (kcals) <=1000 15,721 9.2 34.9 32.4 18.1 3.1 2.3 >1000-1300 17,783 8.6 33.3 33.8 18.5 "> 2.5 >1300-1600 15,294 8.8 34.5 32.2 18.7 3.1 2.7 >1600 17,522 9.4 33.2 31.6 19.2 3.5 3.1 M issing 6,288 10.6 35.3 30.6 17.4 2.9 3.2 Last m am m ogram Never 5,325 13.7 34.8 27.2 19.2 2.7 2.2 <1 year 47,891 7.9 33.6 33.2 18.9 3.4 2.9 1-3 years 14,816 10.1 34.7 32.7 17.2 3.0 2.1 >3 years 4,280 13.4 36.1 27.8 17.7 2.5 2.5 M issing 296 16.6 33.2 31.8 14.5 1.6 2.4 Leisure-time M ET-hrs/wk at age 40 None 10,492 30.0 32.4 25.6 10.2 1.1 0.7 >0-7.0 20,971 8.0 50.2 28.9 10.8 1.4 0.8 >7-17.5 19,522 5.3 33.2 41.7 15.9 2.5 1.5 >17.5-31.5 12,828 3.8 22.1 32.9 34.3 4.1 2.9 >31.5-42.0 3,632 2.5 16.0 30.2 33.5 11.6 6.2 >42.0 3,885 2.7 9.3 23.0 31.7 12.4 21.0 M issing 1,278 9.8 45.1 29.9 13.1 1.2 0.9 Exercise in 1982 None 1,073 28.6 40.4 22.7 7.1 0.4 0.8 Slight 18,013 13.7 41.7 30.2 12.1 1.5 0.8 M oderate 48,740 7.3 32.1 33.7 20.5 3.6 2.8 Heavy 3,880 6.6 21.6 27.4 26.4 7.8 10.2 M issing 902 8.5 32.8 35.5 17.1 4.1 2.0 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2 (continued). Baseline non- recreational M ET-hrs/wk None 1,135 35.4 32.3 21.4 8.3 1.9 0.7 >0-5 17,822 10.9 39.4 32.3 13.6 2.1 1.7 >5 -<10 16,841 6.9 35.8 35.6 16.9 2.9 1.9 10-<18.5 17,560 8.2 33.2 33.1 19.5 3.6 2.6 >18.5 18,480 8.7 28.2 29.5 24.6 4.3 4.7 Missing 770 11.2 38.4 30.0 15.0 3.4 2.0 Low intensity activities are defined as those with MET scores of 3.5-4.5 (walking, biking, aerobics/calisthenics, or dancing), and moderate intensity activities are defined as those with MET scores of 6.0-7.0 (jogging/running, swimming, or tennis/racquetball). 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 3. Rate ratios for recreational leisure-time physical activity at various times during a woman’s lifetime and breast cancer, CPS-II Nutrition Cohort, 1992-1997. # cases / person- years MET-hrs/wk in 19921 None 126/28,698 0.86 (0.71-1.04) 0.86 (0.70-1.04) >0-7.0 554/107,746 1.00 (ref.) 1.00 (ref.) >7.0-17.5 488/102,711 0.92 (0.82-1.04) 0.92 (0.81-1.04) >17.5-31.5 281/58,834 0.92 (0.80-1.07) 0.94 (0.81-1.09) >31.5-42.0 40/10,237 0.76 (0.55-1.05) 0.77 (0.56-1.06) >42.0 31/8,570 0.70 (0.49-1.01) 0.71 (0.49-1.02) p-trend =0.08 (among active women, p-trend=0.03) MET-hrs/wk at age 4& None 224/46,009 1.04(0.88-1.22) 1.03 (0.88-1.21) >0-7.0 431/91,528 1.00 (ref.) 1.00 (ref.) >7.0-17.5 428/85,495 1.06 (0.93-1.22) 1.05 (0.92-1.20) >17.5-31.5 269/55,711 1.02(0.88-1.19) 1.01 (0.87-1.18) >31.5-42.0 87/15,744 1.18(0.94-1.48) 1.16(0.92-1.46) >42.0 64/16,829 0.81 (0.62-1.05) 0.79 (0.61-1.03) p-trend =0.31 (among active women, p-trend=0.36) Exercise in 19823 None 20/4,632 0.81 (0.52-1.28) 0.80 (0.51-1.25) Slight 414/78,717 1.00 (ref.) 1.00 (ref.) Moderate 1,000/212,636 0.88 (0.78-0.98) 0.93 (0.83-1.04) Heavy 72/16,847 0.80 (0.62-1.03) 0.87 (0.68-1.13) — 7 T p-trend =0.32 (among active women, p-trend=0.16) walking, jogging/running, bicycling, swimming, aerobics/calisthenics, tennis/racquetball, and dancing. 2 MET-hrs/week calculated same as above based on recall on 1992 survey of activity at age 40 (1,278 women (17 cases) excluded for missing information). 3 Physical activity reported on 1982 CPS-II survey as “how much exercise do you get?”: none, slight, moderate, or heavy (902 women (14 cases) excluded for missing information). 4 Age-adjusted rate ratio and corresponding 95% confidence interval. 5 Multivariate-adjusted rate ratio and 95% confidence interval adjusted for: age, race, BMI, weight change from age 18 to 1992, family history of breast cancer, personal history of breast cysts, duration of OC use, HRT use, parity, age at menarche, age at menopause, smoking, alcohol intake, caloric intake, education, and mammography history. 6 Trend tests conducted in multivariate models. 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4. Rate ratios for recreational leisure-time physical activity at baseline and breast cancer risk, by stage1 , CPS-II Nutrition Cohort, 1992-1997. # cases / person- 1 & 5 (95% o T ™ RR4(95% Cl) years _____ MET -hrs/wk h7T99? ' ~ ~ ' _ In situ breast cancer None 25/28,464 1.39(0.88-2.19) 1.47 (0.93-2.34) >0-7.0 68/106,611 1.00 (ref.) 1.00 (ref.) >7.0-17.5 66/101,713 1.01 (0.72-1.42) 0.99(0.71-1.40) >17.5-31.5 33/58,254 0.89(0.59-1.35) 0.90(0.59-1.36) >31.5 13/18,679 1.09(0.60-1.97) 1.04(0.57-1.90) p-trend5 =0.38 (among active women, p-trend=0.94) Localized breast cancer None 64/28,553 0.73 (0.56-0.96) 0.74(0.56-0.96) >0-7.0 331/107,256 1.00 (ref.) 1.00 (ref.) >7.0-17.5 281/102,232 0.89(0.76-1.04) 0.88 (0.75-1.03) >17.5-31.5 172/58,559 0.95 (0.79-1.14) 0.95 (0.79-1.15) >31.5 32/18,714 0.55 (0.39-0.80) 0.55 (0.38-0.80) p-trend5 =0.10 (among active women, p-trend=0.02) Regional and distant breast cancer None 29/28,465 1.11 (0.73-1.68) 1.08(0.71-1.63) >0-7.0 98/106,687 1.00 (ref.) 1.00 (ref.) >7.0-17.5 99/101,787 1.05 (0.80-1.39) 1.08 (0.81-1.43) >17.5-31.5 50/58,274 0.93 (0.66-1.31) 0.97(0.69-1.37) >31.5 14/18,672 0.81 (0.46-1.42) 0.85 (0.49-1.50) p-trend5 =0.56 (among active women, p-trend=0.62) 45 cases with missing stage data excluded from stage analyses. 2 MET-hrs/week based on the following activities: walking, jogging/running, bicycling, swimming, aerobics/calisthenics, tennis/racquetball, and dancing. 3 Age-adjusted rate ratio and corresponding 95% confidence interval. 4 Multivariate-adjusted rate ratio and 95% confidence interval adjusted for: age, race, BMI, weight change from age 18 to 1992, family history of breast cancer, personal history of breast cysts, duration of OC use, HRT use, parity, age at menarche, age at menopause, smoking, alcohol intake, caloric intake, education, and mammography history. 5 Trend tests conducted in multivariate models. 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5. Rate ratios for 1992 reported baseline recreational physical activity and postmenopausal breast cancer risk stratified by various factors, CPS-II Nutrition Cohort, 1992-1997. # of cases and ________ RR1 (95% Cl) for 1992 baseline exercise_________ Attained >0-7.0 >7.0-17.5 >17.5-31.5 >31.5 age None mets/wk mets/wk mets/wk mets/wk 26 84 93 37 12 <60 1.16 1.00 1.21 0.87 0.79 (0.74-1.80) (ref.) (0.90-1.63) (0.59-1.28) (0.43-1.45) 27 165 120 77 23 60-64 0.62 1.00 0.79 0.92 0.85 (0.41-0.93) (ref.) (0.62-1.00) (0.70-1.20) (0.55-1.32) 35 156 149 87 18 65-69 0.88 1.00 0.92 0.93 0.62 (0.61-1.27) (ref.) (0.74-1.16) (0.71-1.21) (0.38-1.02) 38 149 126 80 18 >=70 0.92 1.00 0.90 1.01 0.73 (0.64-1.31) (ref.) >0-7.0 (0.71-1.15) >7.0-17.5 (0.77-1.32) >17.5-31.5 (0.45-1.20) >31.5 HRT use None mets/wk mets/wk mets/wk mets/wk 44 210 182 113 38 Current 0.87 1.00 0.85 0.97 0.98 (0.63-1.21) (ref.) (0.70-1.04) (0.77-1.21) (0.70-1.39) 12 77 65 24 6 Former 0.58 1.00 0.89 0.60 0.48 (0.32-1.07) (ref.) (0.64-1.23) (0.38-0.95) (0.21-1.09) 68 244 228 139 26 Never 0.98 1.00 1.02 1.08 0.64 (0.75-1.29) (ref.) (0.85-1.23) p-interaction=0.09 (0.88-1.33) (0.43-0.97) Alcohol >0-7.0 >7.0-17.5 >17.5-31.5 >31.5 intake None mets/wk mets/wk mets/wk mets/wk 62 257 208 112 19 Non 0.81 1.00 0.94 0.95 0.60 drinker (0.62-1.08) (ref.) (0.78-1.13) (0.76-1.18) (0.38-0.96) 57 275 264 160 51 Drinker 0.88 1.00 0.90 0.94 0.84 (0.66-1.17) (ref.) (0.76-1.07) p-interaction=0.74 (0.77-1.14) (0.62-1.14) 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5 (continued). Caloric >0-7.0 >7.0-17.5 >17.5-31.5 >31.5 intake None mets/wk mets/wk mets/wk mets/wk 54 252 236 127 32 <=1300 0.83 1.00 0.95 0.93 0.77 kcals (0.61-1.11) (ref.) (0.79-1.13) (0.75-1.16) (0.53-1.11) 63 251 218 132 36 >1300 0.94 1.00 0.92 0.96 0.79 kcals (0.72-1.25) (ref.) (0.77-1.10) p-interaction=0.96 (0.77-1.18) (0.56-1.12) Weight change from age >0-7.0 >7.0-17.5 >17.5-31.5 >31.5 18 to 1992 None mets/wk mets/wk mets/wk mets/wk 29 136 148 98 27 <=15 gain 0.95 1.00 0.95 0.96 0.66 (0.64-1.42) (ref.) (0.75-1.20) (0.74-1.25) (0.43-0.99) 93 403 336 181 43 >15 lbs. 0.83 1.00 0.93 0.95 0.81 gain (0.66-1.04) (ref.) (0.80-1.08) (0.80-1.13) (0.59-1.12) p -inter action=0.8 7 >0-7.0 >7.0-17.5 >17.5-31.5 >31.5 BMI None mets/wk mets/wk mets/wk mets/wk 53 258 271 152 46 <25 0.94 1.00 0.99 0.89 0.75 (0.70-1.26) (ref.) (0.83-1.17) (0.73-1.09) (0.55-1.03) 34 180 133 84 22 25-<30 0.73 1.00 0.79 0.95 0.90 (0.51-1.05) (ref.) (0.63-0.99) (0.73-1.23) (0.57-1.40) 35 103 82 46 >302 0.89 1.00 1.05 1.06 (0.61-1.31) (ref.) (0.78-1.40) (0.75-1.50) p-interaction=0.81 Multivariate-adjusted rate ratio and 95% confidence interval adjusted for: age, race, BMI, weight change from age 18 to 1992, family history of breast cancer, personal history of breast cysts, duration of OC use, HRT use, parity, age at menarche, age at menopause, smoking, alcohol intake, caloric intake, education, and mammography history. 2 , For BMI > 30, highest category of physical activity had only two cases; therefore, categories collapsed as activity > 17.5 mets/wk. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Discussion We observed lower incidence of postmenopausal breast cancer in women who reported higher levels of recreational physical activity at the time of enrollment into our study. These findings are similar to the results of six (Cerhan and others 1998; Dirx and others 2001; Lee and others 2001b; Rockhill and others 1999; Sesso and others 1998; Thune and others 1997) of the seven (Cerhan and others 1998; Dirx and others 2001; Lee and others 2001b; Moore and others 2000; Rockhill and others 1999; Sesso and others 1998; Thune and others 1997) previous prospective cohort studies and three population-based case-control studies (Friedenreich and others 2001b; Gilliland and others 2001; Moradi and others 2000) that examined physical activity during the postmenopausal years and breast cancer risk. Breast cancer incidence was approximately 29 percent lower among active women in the highest category compared to active women in the lowest category of physical activity. Compared to slightly active women, we observed a small reduction in risk of breast cancer among completely inactive women. Conditions such as osteoporosis are associated with lower levels of circulating estrogens and lower breast cancer risk; additionally, women with osteoporosis are also less likely to engage in regular physical activity. Information on osteoporosis and fractures was not available at baseline in 1992. However, using data from the 1997 questionnaire, we found that inactive women in this population were more likely to suffer from hip fractures, a consequence of osteoporosis. Since we were unable to adequately control for 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. conditions such as osteoporosis or fractures at baseline, we did not use inactive women as the referent group. Our choice of referent group is supported by the findings of Cerhan and colleagues (Cerhan and others 1998) who controlled for physical function in their analysis of physical activity and postmenopausal breast cancer and observed a lower risk among women who reported inactivity due to physical disability. Engaging in non-recreational physical activity, such as shopping, gardening, and housework, was not associated with lower risk of breast cancer in this population. Although these activities comprise a large part of total activity among older women, they may not be vigorous enough to infer any physiologic response (such as influencing hormone levels or promoting weight loss) necessary to lower breast cancer risk. We also did not observe a stronger association with prolonged physical activity in this population, although misclassification of recalled exposure information may increase the likelihood of bias towards the null. When examining risk by stage of disease, we found the greatest reduction in risk for localized breast cancer. This finding suggests that the impact of physical activity is greater among women with a more favorable prognosis; however, we did not observe a reduction in risk for in situ breast cancer. The lack of association in in situ disease may be real or may be due to residual confounding from screening. Women who are physically active are more likely to be screened (Mayer-Oakes and others 1996). Since in situ breast cancer is virtually always detected with mammography, there may be an overrepresentation of active women among in situ 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cases resulting in a bias towards the null. When examining risk of in situ breast cancer among only women screened within the year prior to baseline, we found a lower risk with increasing physical activity compared to analyses including all women (data not shown). For localized and regional/distant cancer, no differences were seen when the analysis was limited to only women who were screened. One other study examined risk of postmenopausal breast cancer by stage of disease and found a significant inverse trend with physical activity and localized breast cancer. The inverse association was not significant for distant cases; however, the study was limited in their number of distant cases (n=12), and in situ cases were not included in the analysis (Cerhan and others 1998). Although we did not find any significant effect modification by other risk factors, we did observe a stronger association between recreational physical activity and breast cancer risk among women who did not report current HRT use in 1992. One may speculate that the influence of moderate levels of physical activity on hormone levels in women with a favorable estrogen profile (i.e. lower baseline levels of hormones) may be sufficient to reduce risk, whereas women with higher levels of baseline circulating estrogens, such as current HRT users or obese women, may not experience a reduction in risk with moderate physical activity. Women with higher baseline estrogen levels may require more vigorous and frequent activity to substantiate a reduction in risk. Only three previous studies have examined potential effect modification by HRT, but none found a significant interaction (Friedenreieh and others 2001b; Gammon and others 1998; Moore and others 2000). Two (Moradi 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and others 2000; Thune and others 1997) of 17 studies (Chen and others 1997; D'Avanzo and others 1996; Dirx and others 2001; Friedenreich and others 2001b; Gammon and others 1998; Lee and others 2001a; Levi and others 1999; Luoto and others 2000; Marcus and others 1999; Matthews and others 2001; Moore and others 2000; Moradi and others 2000; Rockhill and others 1999; Sesso and others 1998; Shoff and others 2000; Thune and others 1997; Verloop and others 2000) that examined possible effect modification by body mass found that leaner women had a significantly greater reduction in risk compared to overweight women. The study publications reporting no effect modification have not provided details of analyses; therefore, it is difficult to assess whether or not this is due to a lack of statistical power. A limitation of our study was the lack of information on physical activity in adolescence and young adulthood, which may be critical to the multistage induction of invasive breast cancer. The only measures of past physical activity available are based on recalled information at age 40 and information reported prospectively in 1982. Furthermore, the 1982 question is very crude in its physical activity assessment; thus, we may have substantial misclassification of past exposure. We also lack updated information during follow-up; however, we expect this to have minimal effect due to the short (five-year) follow-up period. Other limitations of our study are that we have a limited range in the type of activities commonly done by our participants and no individual information on intensity. Most highly active women in the study engaged in walking with the addition of modest amounts of the other six 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reportable activities. The lack of information on the intensity of individual behavior increases the misclassification of true energy expenditure. Another limitation of this study is that a subset of participants originally recruited in 1982 for the CPS-II follow-up study subsequently volunteered to participate in this Nutrition Cohort in 1992. These participants who volunteered to participate in 1992 are healthier than non-respondents or the general population, and represent a select population. Although study participants are on average more affluent, educated, and health conscious than the average US population, these differences are unlikely to compromise internal validity. While the relatively homogenous nature of the women in this study reduces the likelihood of residual confounding, it also reduces the range of the physical activity exposure variable. There are many strengths of this study that should be mentioned. The prospective design reduces the likelihood of differential reporting of recalled exposure information, and eliminates the possibility of recall bias for baseline activity measures. We have a large sample size as well as the ability to test for potential confounding by most important breast cancer risk factors. Occupational physical activity could confound the association between recreational leisure-time activity and breast cancer risk if women who reported little recreational physical activity were very active in the workplace. However, there is likely to be little, if any, confounding by occupational physical activity in this population because most women reported being housewives for their ‘main lifetime’ occupation, and of those who did work outside of the home, the majority of women were in clerical 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. occupations where physical activity in the workplace would be minimal (Calle and others 1998). Finally, any protection observed with light or moderate activity during the postmenopausal years may be of public health importance. Baseline reported physical activity in this study likely reflects those women who have been consistent exercisers over their lifetime as well as women who have initiated exercising recently; therefore, it is difficult to ascertain whether our findings reflect benefit of late-life physical activity or being a long-term exerciser. We assessed risk combining physical activity measures at age 40 and in 1982 with baseline physical activity, but found no differences in risk in women who reported activity in the past and at baseline compared to women who reported only baseline physical activity. Furthermore, it is biologically plausible that the initiation of late- life physical activity may be beneficial in reducing breast cancer risk. Physical activity has been consistently associated with lower weight, lower BMI, and weight loss (Vainio and Bianchini 2002). It is, however, likely that mechanistic pathways other than the effects of physical activity on body weight explain, at least in part, the relationship between physical activity and postmenopausal breast cancer. Studies have shown that after adjustment for body mass, physical activity during postmenopausal years is still associated with lower levels of serum estrone, estradiol, and androgens, and higher levels of SHBG (Vainio and Bianchini 2002). Furthermore, independent of its effects on body mass, studies have shown that postmenopausal women with low to moderate levels of physical activity have increased insulin sensitivity and decreased plasma insulin levels 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Vainio and Bianchini 2002). Insulin sensitivity may impact breast cancer risk because higher levels of insulin are associated with decreased levels of SHBG, and consequently a higher level of free-estradiol (Nestler and others 1991). Current evidence does not allow clear conclusions to be drawn regarding the possible association between physical activity and IGF levels (Vainio and Bianchini 2002). Thus, physical activity after menopause may directly suppress sex hormones or increase insulin sensitivity. The lack of understanding of minimal dose of physical activity necessary to cause any hormonal change is a major limitation of the existing literature; however, regular moderate intensity exercise, such as that in our highest physical activity category, is thought to be sufficient to induce some physiologic responses (Vainio and Bianchini 2002). In summary, postmenopausal women who engaged in high amounts of recreational physical activity at baseline were at a lower risk of breast cancer than those engaged in low levels of physical activity. There is sufficient biologic plausibility for this association to warrant further research on late-life activity and primary prevention of breast cancer in postmenopausal women. 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 6: In Situ Breast Cancer Study Study Background All English-speaking, US-born, white (including Hispanic) and African American female residents of Los Angeles County between the ages of 35 and 64 years with no prior diagnosis of breast cancer were eligible to participate in the study. Eligible case patients were diagnosed with histologically confirmed in situ breast carcinoma between March 1, 1995 and May 31, 1998. Classification of in situ breast carcinoma was based on the International Classification of Diseases for Oncology, ICD-O, coding (C50.0-C50.9) with a 5th digit behavior code of two (indicating in situ cancer). Histologic subtype was based on ICD-O morphologic codes: 8500-8504, 8522, 8543, and 8573 for ductal carcinoma in situ (DCIS), 8520 for lobular carcinoma in situ (LCIS), and remaining subtypes were classified as ‘other’ (World Health Organization 1990). The University of Southern California Cancer Surveillance Program, the population-based cancer registry for Los Angeles County, using rapid case ascertainment procedures, identified eligible case patients. For comparability with control subjects, case patients were also required to have a working residential telephone at reference date (i.e., date of diagnosis). In an effort to further maintain comparability by age and race with selection of control subjects, the following selection fractions were used to select case patients: all black case patients were recruited, all white case patients between the ages of 35-39 years were selected, 75% of white case patients ages 40 to 44 years, 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67% ages 45 to 49 years, 50% ages 50 to 54 years and 55 to 59 years, and 75% ages 60 to 64 years. The Cancer Surveillance Program identified 720 in situ breast cancer patients (597 whites and 123 African Americans) who satisfied the eligibility criteria for the study and were selected for participation. Interviews were completed with 567 in situ breast carcinoma patients. Response rates were 80% for white patients (n=475) and 75% for African American patients (n=92). Eligible patients were first contacted by letter and then telephoned to schedule an interview. Of the 153 women who did not participate, one woman (1%) was too ill to complete the interview, 54 women (35%) could not be contacted due to physician refusal, nine women (6%) had relocated outside of Los Angeles County, and 89 (58%) refused to be interviewed. Control subjects were selected from the Women’s Contraceptive and Reproductive Experiences Study (CARE), which is a multi-center study that started in mid-1994. The Women’s CARE study is a population-based case-control study that was designed to study various reproductive factors, such as oral contraceptive use, and invasive breast cancer among white and black women. The five sites participating in the Women’s CARE study were Los Angeles County, Atlanta, Detroit, Philadelphia, and Seattle. Control subjects were identified using random digit dialing (RDD) from the same counties as case patients for the Women’s CARE study. Control subjects were frequency-matched to Women’s CARE case patients within strata of geographic site, race, and five-year age group. Once contacted, standardized eligibility questions were asked by the interviewer to establish eligibility. Control subject eligibility was identical to that for case patients with the 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. exception that controls did not have a current diagnosis of invasive or in situ breast cancer. A more detailed description of the Women’s CARE study and subject selection has been previously reported (Marchbanks and others 2002), and it should be noted that the in situ study followed the same eligibility guidelines and study procedures as the Women’s CARE study except that the in situ study was conducted solely in Los Angeles County. Signed informed consent was obtained from each subject, and study procedures were approved by the University of Southern California Research Committee, in accord with assurances approved by the U.S. Department of Health and Human Services. In-person interviews were conducted with each study subject to obtain information on demographic and reproductive factors, personal and family history of diseases, lifestyle, and screening behavior. Reproductive information included oral contraceptive use, other hormone use, and complete histories of pregnancy, lactation, and surgeries to reproductive organs. Behavioral and other modifiable factors, such as physical activity, alcohol consumption, and smoking behaviors, were recorded up to the reference date for all study subjects. For case patients, this date was the date of diagnosis of BCIS. For control subjects, this date was the date of initial contact with the subject’s household in the random digit dialing process. All subjects were interviewed within 18 months of the reference date. For the present analysis, we used all control subjects recruited in the Los Angeles County site of the Women’s CARE study who were interviewed between March 1, 1995 and May 1, 1998 (n=l,026). The response rate for control subjects 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. across all Women’s CARE study sites was approximately 78% (Marchbanks and others 2002). Of the 1,026 potential control subjects from Los Angeles County who were eligible for this study, 410 (40%) reported having had no mammogram within the two years prior to the reference date. Since the vast majority of newly diagnosed BCIS patients have no palpable lesions and are diagnosed by mammography (Millikan and others 1995; Winchester and others 2000), we limited control subject eligibility to those who reported a mammogram within the past two years (n=616). Of the 567 case patients, 56 (10%) reported having had no mammogram within two years of their diagnosis date; however, upon review of pathology reports and tumor registry abstracts, all 56 women had record of a mammogram during their initial work-up and were subsequently included in the analysis. Measures of Physical Activity Lifetime physical activity was recorded following completion of a calendar of life events upon which was recorded all pregnancy information and use of oral contraceptives and other hormonal products. We documented all episodes of physical activity in which a participant engaged, but restricted analysis to episodes that occurred from the woman’s menarche until her reference date. For each activity in which a woman participated, the type of activity, the age at which the woman started and stopped the activity, the number of months per year in which she participated in the activity, and the average duration per week that she engaged in the activity during that episode were recorded. Based on these data, we calculated the average 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. number of hours of exercise activity per week from menarche to reference date for each participant. We also estimated lifetime metabolic-equivalents-of-energy expenditure (MET) hours per week by multiplying the number of hours per week a woman engaged in a particular activity by the estimated MET score for that activity based on the Ainsworth Compendium of Physical Activities (Ainsworth and others 1993). Average hours and MET-hrs were also calculated for the following time periods: the 10 years after menarche, ages 20 to 34 years, and the 10 years prior to reference date. Screening and Restriction of Control Population Since the vast majority of in situ breast carcinoma is initially identified via mammography, appropriate selection of a properly screened control group is essential. Women who have not been screened are technically not “at risk” since identification of the outcome is very unlikely; consequently, they are not an adequate comparison group. Inclusion of unscreened women increases the likelihood of false negatives in the control population. Furthermore, if the hypothesized risk factor is associated with screening behavior, inadequate adjustment for screening would confound the association being examined. Current screening guidelines by the American Cancer Society recommend women age 40 and older should have an annual mammogram [Society, 2001 #303]. Engaging in routine screening practice, especially in younger women, is associated with many breast cancer risk factors. Women who are routinely screened are more 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. likely to be highly educated, have a family history of breast cancer, use of have used oral contraceptives and/or hormone replacement therapy, be normal weight, and be physically active (Calle and others 1993; Fontaine and others 1998; Fontaine and others 2001; Frazier and others 1996; Ruffin and others 2000). We restricted our analysis to women who have been screened within two years of the reference date. When restricting the analysis to women screened within the last year or within the last six months, risk estimates for the main exposure and other important risk factors did not change; therefore, we used the two-year cutoff to maximize the available sample size while reducing the likelihood of false negatives. Alternatively, we considered indirectly adjusted for screening by including all controls and adjusting for those risk factors that we know to be predictive of screening behavior into models. However, other factors, such as perceived risk of developing breast cancer, general adherence to preventive practices (such as seatbelt use), and physician recommendations have also be associated with likelihood of being screened (Friedman and others 1995; Mayer-Oakes and others 1996), and since we were not able to adjust for these, or other potential, predictors of screening, we chose the option to restrict the control population a priori. Furthermore, by including unscreened women, we still risked the inclusion of false negatives. Although restricting the controls by screening creates a select population and is not representative of the general population, it appears to be the most adequate method of selecting a control population that is representative of the base population from 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. where cases arise, and we believe that internal validity of study results will not be compromised. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 7: In Situ Breast Cancer Study-Manuscript Introduction Incidence rates of breast carcinoma in situ (BCIS) have dramatically increased over the past two decades primarily due to increased mammography screening. Ductal carcinoma in situ (DCIS) accounts for approximately 85% of all BCIS, lobular carcinoma in situ (LCIS) accounts for approximately 5-10%, and the remainder are mixed forms of the disease. DCIS is generally recognized as the final step in the progression to invasive disease and accounts for approximately 10-20% of all breast cancer diagnoses (American Cancer Society 2001; Habel and others 1998; Millikan and others 1995). The study of in situ carcinomas as an endpoint in epidemiologic studies is important because it will help identify the common risk factors between in situ and invasive breast cancers, as well as provide information on risk factors that are likely to be on the pathway to invasive disease, such as reproductive and lifestyle factors (Millikan and others 1995). Few studies have been conducted to examine BCIS risk factors, thus the epidemiology of BCIS remains unclear. Family history o f breast cancer is generally recognized as a risk factor for both in situ and invasive breast cancer, and parity has also been shown to be inversely associated with in situ risk (Claus and others 2001; Kerlikowske and others 1997; Lambe and others 1998; Trentham-Dietz and others 2000). However, few other risk factors for breast carcinoma in situ are well-established. 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Physical activity has been proposed as a modifiable risk factor for invasive breast cancer because of its effects on circulating sex hormones. Overall, results from previous studies support the hypothesis that regular physical activity may reduce the risk of invasive breast cancer (Vainio and Bianchini 2002). However, the relationship between physical activity and BCIS risk is poorly understood. Many studies examining invasive breast cancer have included in situ carcinomas, but they were limited by the number of in situ cases or data on frequency of mammographic screening to conduct thorough, separate analyses of BCIS (Carpenter and others 1999; Friedenreich and others 2001b; Gilliland and others 2001; Hu and others 1997; Lee and others 2001b; McTieman and others 1996; Thune and others 1997; Ueji and others 1998). We examined the association of lifetime recreational physical activity and in situ breast cancer risk among screened women in a population-based case-control study conducted at the University of Southern California in Los Angeles County, CA. Methods: Study Background All English-speaking, US-born, white (including Hispanic) and African American female residents of Los Angeles County between the ages of 35 and 64 years with no prior diagnosis of breast cancer were eligible to participate in the study. Eligible case patients were diagnosed with histologically confirmed in situ breast carcinoma between March 1, 1995 and May 31, 1998. Classification of in situ 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. breast carcinoma was based on the International Classification of Diseases for Oncology, ICD-O, coding (C50.0-C50.9) with a 5th digit behavior code of two (indicating in situ cancer). Histologic subtype was based on ICD-O morphologic codes: 8500-8504, 8522, 8543, and 8573 for ductal carcinoma in situ (DCIS), 8520 for lobular carcinoma in situ (LCIS), and remaining subtypes were classified as ‘other’ (World Health Organization 1990). The University of Southern California Cancer Surveillance Program, the population-based cancer registry for Los Angeles County, using rapid case ascertainment procedures, identified eligible case patients. For comparability with control subjects, case patients were also required to have a working residential telephone at reference date (i.e., date of diagnosis). In an effort to further maintain comparability by age and race with selection of control subjects, the following selection fractions were used to select case patients: all black case patients were recruited, all white case patients between the ages of 35-39 years were selected, 75% of white case patients ages 40 to 44 years, 67% ages 45 to 49 years, 50% ages 50 to 54 years and 55 to 59 years, and 75% ages 60 to 64 years. The Cancer Surveillance Program identified 720 in situ breast cancer patients (597 whites and 123 African Americans) who satisfied the eligibility criteria for the study and were selected for participation. Interviews were completed with 567 in situ breast carcinoma patients. Response rates were 80% for white patients (n=475) and 75% for African American patients (n=92). Eligible patients were first contacted by letter and then telephoned to schedule an interview. Of the 153 women who did not participate, one woman (1%) was too ill to complete the interview, 54 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. women (35%) could not be contacted due to physician refusal, nine women (6%) had relocated outside of Los Angeles County, and 89 (58%) refused to be interviewed. Control subjects were selected from the Women’s Contraceptive and Reproductive Experiences Study (CARE), which is a multi-center study that started in mid-1994. The Women’s CARE study is a population-based case-control study that was designed to study various reproductive factors, such as oral contraceptive use, and invasive breast cancer among white and black women. The five sites participating in the Women’s CARE study were Los Angeles County, Atlanta, Detroit, Philadelphia, and Seattle. Control subjects were identified using random digit dialing (RDD) from the same counties as case patients for the Women’s CARE study. Control subjects were frequency-matched to Women’s CARE case patients within strata of geographic site, race, and five-year age group. Once contacted, standardized eligibility questions were asked by the interviewer to establish eligibility. Control subject eligibility was identical to that for case patients with the exception that controls did not have a current diagnosis of invasive or in situ breast cancer. A more detailed description of the Women’s CARE study and subject selection has been previously reported (Marchbanks and others 2002), and it should be noted that the in situ study followed the same eligibility guidelines and study procedures as the Women’s CARE study except that the in situ study was conducted solely in Los Angeles County. Signed informed consent was obtained from each subject, and study procedures were approved by the University of Southern 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. California Research Committee, in accord with assurances approved by the U.S. Department of Health and Human Services. In-person interviews were conducted with each study subject to obtain information on demographic and reproductive factors, personal and family history of diseases, lifestyle, and screening behavior. Reproductive information included oral contraceptive use, other hormone use, and complete histories of pregnancy, lactation, and surgeries to reproductive organs. Behavioral and other modifiable factors, such as physical activity, alcohol consumption, and smoking behaviors, were recorded up to the reference date for all study subjects. For case patients, this date was the date of diagnosis of BCIS. For control subjects, this date was the date of initial contact with the subject’s household in the random digit dialing process. All subjects were interviewed within 18 months of the reference date. For the present analysis, we used all control subjects recruited in the Los Angeles County site of the Women’s CARE study who were interviewed between March 1, 1995 and May 1, 1998 (n=l,026). The response rate for control subjects across all Women’s CARE study sites was approximately 78% (Marchbanks and others 2002). Of the 1,026 potential control subjects from Los Angeles County who were eligible for this study, 410 (40%) reported having had no mammogram within the two years prior to the reference date. Since the vast majority of newly diagnosed BCIS patients have no palpable lesions and are diagnosed by mammography (Millikan and others 1995; Winchester and others 2000), we limited control subject eligibility to those who reported a mammogram within the past two years (n=616). 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Of the 567 case patients, 56 (10%) reported having had no mammogram within two years of their diagnosis date; however, upon review of pathology reports and tumor registry abstracts, all 56 women had record of a mammogram during their initial work-up and were subsequently included in the analysis. Methods: Measures of Physical Activity Lifetime physical activity was recorded following completion of a calendar of life events upon which was recorded all pregnancy information and use of oral contraceptives and other hormonal products. We documented all episodes of physical activity in which a participant engaged, but restricted analysis to episodes that occurred from the woman’s menarche until her reference date. For each activity in which a woman participated, the type of activity, the age at which the woman started and stopped the activity, the number of months per year in which she participated in the activity, and the average duration per week that she engaged in the activity during that episode were recorded. Based on these data, we calculated the average number of hours of exercise activity per week from menarche to reference date for each participant. We also estimated lifetime metabolic-equivalents-of-energy expenditure (MET) hours per week by multiplying the number of hours per week a woman engaged in a particular activity by the estimated MET score for that activity based on the Ainsworth Compendium of Physical Activities (Ainsworth and others 1993). Average hours and MET-hrs were also calculated for the following time 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. periods: the 10 years after menarche, ages 20 to 34 years, and the 10 years prior to reference date. Methods: Statistical Analyses We used unconditional logistic regression modeling (Kleinbaum 1994) to calculate odds ratios (OR) and corresponding 95% confidence intervals (Cl) to examine the relationship between exercise activity measures and in situ breast cancer risk. Women who reported no lifetime exercise activity were used as the referent group. For each analysis, we assessed risk in two models, one adjusted for age and race, and the other adjusted for age, race, and potential confounding factors. All models were stratified on age (in five-year age groups) and race (white, black). Other potential confounders assessed for inclusion in the multivariate models were education (< high school graduate, some college, > college graduate), income (<$15,000, $15-35,000, >$35-70,000, >$70,000, missing), family history of breast cancer in mother, sisters, or daughters (yes, no, not known), age at menarche (<12, 12-13, >13), smoking status (never, current, former), body mass index (weight(kg)/height(m)2 ) (<25.0, 25.0-<30.0, >30.0), oral contraceptive use (never, <2 years, 2-5 years, >5 years, unknown), number of pregnancies with greater than 26 weeks gestation (none, 1-2, >2), and menopausal status (premenopausal, perimenopausal, postmenopausal, unknown), age at menopause (<50, 50-54, >55, unknown), postmenopausal HRT use (never, ever estrogen use (unopposed or opposed), ever other hormone use). Women were considered post-menopausal if 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. they reported being post-menopausal and had no menstrual periods in the twelve months prior to their reference date, if they reported a bilateral oophorectomy, if their menstrual periods stopped due to radiation or chemotherapy at least six months prior to their reference date, or if they had reached the age of 55 and had not designated their menstrual status as premenopausal (for these women, age at menopause was considered unknown). Potential confounders were forced into the model if they were significant predictors of screening behavior in an effort to adjust for possible residual confounding by screening (income, family history, smoking status, menopausal status, age at menopause, and HRT use) (Table 6). Other confounders were retained in the multivariate models if they changed physical activity parameter estimates by more than 10% (BMI and number of pregnancies). Trend tests for exercise activity models were obtained by assigning the mean number of hours or MET-hours within each category to each category. To assess effect modification by any of the potential confounders described above, we used the likelihood ratio test comparing two multivariate logistic models (Kleinbaum and others 1982). Due to small numbers in some strata, categories of potential effect modifiers were collapsed dichotomously, and one model included a single physical activity variable, and the other included two sets of exercise activity variables, one for each category of the potential effect modifier (test for homogeneity of trends). All p-values reported are two-sided and a p-value <0.05 was considered statistically significant. 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6. Possible predictors of screening in control population, in situ breast cancer study. ______________________________ ________________________________ Covariate Controls screened w/in 2 years (n=616) Controls not screened w/in 2 years (n=410) OR1 (95% Cl) Reference age2 (yrs) <45 139 250 0.33 (0.23-0.49) 45-49 100 60 1.00 (ref.) 50-54 122 47 1.56 (0.98-2.48) 55-59 133 28 2.84(1.69-4.77) >60 122 25 2.87(1.68-4.92) Race White 364 222 1.00 (ref.) Black 252 188 0.88(0.67-1.16) Educational level < High school graduate 210 155 1.00 (ref.) Some college 218 152 1.06 (0.79-1.42) College graduate 188 103 1.35 (0.98-1.85) Income2 <$15,000 74 73 1.00 (ref.) $15-35,000 134 107 1.24 (0.82-1.86) $35-70,000 215 125 1.70 (1.15-2.51) >$70,000 182 98 1.83 (1.22-2.75) Unknown 11 7 1.55 (0.57-4.22) Age at menarche (vrs) <12 169 111 1.00 (ref.) 12-13 321 213 0.99 (0.74-1.33) >13 126 86 0.96 (0.67-1.39) 2 Smoking status Never smoked 284 183 1.00 (ref.) Current smoker 121 119 0.66 (0.48-0.90) Former smoker 211 108 1.26 (0.94-1.69) Bodv mass index (kg/m2) <25.0 318 216 1.00 (ref.) 25.0-<30.0 172 119 0.98 (0.73-1.31) >30.0 126 75 1.14(0.82-1.59) 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6 (continued). Family history of breast cancer" No 552 Yes 60 Not known 4 OC use (yrs) None 133 < 2 181 2-5 99 5-9 109 > 10 92 Unknown 2 Age at menopause2 Premenopausal 121 Perimenopausal 87 Postmenopausal <50 years 160 50-54 years 63 55+years 8 Unknown age 83 Unknown status 94 HRT use2’ 3 Never 67 Ever, ERT and/or EPRT 114 Ever, other use 50 Average hours of physical activity (menarche to reference)2 (hrs/wk) None 115 <1 hr/week 192 1 -4 hrs/week 206 >4 hrs/week 103 391 1.00 (ref.) 19 2.28 (1.34-3.87) 0 88 1.00 (ref.) 128 0.94(0.66-1.33) 61 1.07(0.71-1.63) 73 0.99(0.66-1.46) 60 1.02 (0.67-1.55) 0 207 1.00 (ref.) 64 2.33 (1.57-3.45) 70 3.91 (2.73-5.60) 20 5.39(3.11-9.35) 2 6.84(1.43-32.7) 10 14.2 (7.10-28.4) 37 4.35 (2.80-6.76) 49 1.00 (ref.) 28 2.98(1.71-5.18) 15 2.44(1.23-4.83) 104 1.00 (ref.) 96 1.81 (1.26-2.60) 142 1.31 (0.93-1.84) 68 1.37(0.91-2.05) 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6 (continued). Number of pregnancies (gestation >26 weeks) None 104 87 1.00 (ref.) 1-2 282 177 1.33 (0.95-1.88) >3 230 146 1.32 (0.93-1.87) Odds ratios and corresponding 95% confidence intervals comparing control subjects with a screening mammogram within two years of reference date to those with not such screening mammogram obtained in multivariate logistic regression models that include age and race. 2 Significant predictors of screening in multivariate model. 3 Restricted to postmenopausal women with a known age at menopause. 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Results The average age of diagnosis for case patients was 51.6 years (+ 7.2) and the average age at reference date for control subjects was 51.6 years (+ 7.7). The associations between potential confounding factors considered for inclusion in the multivariate logistic regression models and risk of BCIS are summarized in Table 7. Family history of breast cancer was significantly associated with an increase in risk of BCIS (OR=2.02, 95% Cl 1.45-2.96). BMI was inversely associated with BCIS risk. Parity and number of pregnancies were also inversely associated with risk (OR=0.45, 95% Cl 0.31-0.64, > 3 pregnancies vs. nulliparous). The distributions of average number of hours per week of exercise activity between menarche and reference date and MET-hours per week over the same time period are shown in Table 8. 117 case patients (20.6%) and 115 (18.7%) control subjects reported no exercise activity during the time interval between menarche and the participant’s reference date. After adjustment for age, race, income, BMI, family history of breast cancer, menopausal status, age at menopause, postmenopausal hormone use, smoking status, age at menarche, and parity, the risk of in situ breast carcinoma was approximately 35 percent lower among women with ever lifetime exercise activity compared to ever inactive women. We observed no linear trend in BCIS risk with increasing levels of exercise activity, measured either by average 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. hours per week or by MET-hours per week (p-trend=0.24 and p-trend=0.27, respectively). When examining specific time periods during a woman’s reproductive life, exercise activity during the ten years after menarche appeared to be modestly associated with lower risk of BCIS , but no clear trend was observed (p-trend=0.18). Exercise activity during the age period 20-34 years was unrelated to BCIS risk. Exercise activity during the ten years preceding a woman’s study reference date was associated with a reduced risk of BCIS (OR=0.62, 95% Cl 0.42-0.92, >4 hrs/wk vs. no activity). We observed a significant decline in risk with increasing levels of exercise activity during this time period (p-trend=0.02) (Table 9). When restricting analyses to only DCIS cases, we found no differences in risk estimates for lifetime exercise activity or exercise activity during any specific time period (data not shown). The effects of exercise activity on BCIS risk differed according to family history of breast cancer. Among women with no family history of breast cancer, the relative odds of in situ breast cancer was 0.53 (95% Cl 0.34-0.82) comparing women with >4 hours of lifetime exercise activity per week to inactive women, whereas no comparable reduction in risk was observed among active women with a family history of breast cancer (OR=2.29, 95% Cl 0.62-8.44) (homogeneity of trends p- vale=0.02). Among women with no family history of breast cancer, we also observed an inverse trend in BCIS risk with increasing exercise activity during the ten years after menarche and the ten years prior to women’s reference date, whereas no 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. comparable trends were observed among women with a positive family history and exercise activity during these time periods (homogeneity of trends p-value=0.05, for both time periods) (Table 10). We found no evidence of effect modification by age, race, parity, HRT use, menopausal status, smoking status, body mass, oral contraceptive use, education, income, or age at menarche and lifetime exercise activity or exercise activity at any specific time period (data not shown). Although we conducted all analyses restricting the population to women who had been screened via mammography within two years of reference date, we attempted to further control for possible confounding by screening history and screening regularity in an effort to maximize the comparability between case patients and control subjects. We repeated analyses restricting the population to women who had been screened within one year of reference date and to women who had been screened within six months of reference date. We also repeated analyses adjusting for and stratifying by the number of mammograms within five years of reference date (<3, >3 mammograms) in an effort to compare women who were more likely to be screened regularly versus inconsistently screened women. We found no significant differences in risk estimates when comparing any of these models to the initial models restricting the population by screening within two years of reference date (data not shown). 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 7. Association between various factors and risk of breast carcinoma in situ. Covariate Cases (n=567) Controls (n=616) OR1 (95% Cl) OR2 (95% Cl) .......................... .......................2 --- ---------- --------- — - --- -- ------------ --- - ............ ......... ..............................— ............. Reference age (vrs) <45 106 139 — — 45-49 107 100 — — 50-54 145 122 — 55-59 112 133 — — 60+ 97 122 — — Race3 White 475 364 — Black 92 252 — — Educational level < High school graduate 167 210 1.00 (ref.) 1.00 (ref) Some college 180 218 1.04 (0.78-1.38) 0.89 (0.65-1.21) College graduate 220 188 1.47 (1.11-1.95) 1.02 (0.73-1.43) Income <$15,000 49 74 1.00 (ref.) 1.00 (ref.) $15-35,000 111 134 1.25 (0.81-1.94) 1.10(0.69-1.75) $35-70,000 162 215 1.14(0.75-1.72) 0.89 (0.57-1.40) >$70,000 225 182 1.87 (1.24-2.82) 1.30 (0.81-2.07) Unknown 20 11 2.75 (1.21-6.23) 2.31 (0.97-5.53) Age at menarche (yrs) <12 145 169 1.00 (ref.) 1.00 (ref.) 12-13 308 321 1.12 (0.85-1.47) 0.93 (0.69-1.25) >13 114 126 1.05 (0.75-1.48) 0.84 (0.59-1.21) Smoking status Never smoked 258 284 1.00 (ref.) 1.00 (ref.) Current smoker 86 121 0.78 (0.57-1.08) 0.78 (0.55-1.12) Former smoker 223 211 1.16(0.90-1.50) 1.19(0.91-1.57) Body mass index (kg/m2) <25.0 378 318 1.00 (ref.) 1.00 (ref.) 25.0-<30.0 125 172 0.61 (0.47-0.80) 0.67 (0.50-0.90) >30.0 64 126 0.43 (0.31-0.60) 0.45 (0.31-0.65) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 7 (continued). Family history of breast cancer No 450 551 1.00 (ref.) 1.00 (ref.) Yes 110 61 2.21 (1.58-3.09) 2.02(1.45-2.96) Not known 7 4 2.14 (0.62-7.37) 2.92 (0.76-11.2) Aee at menopause Premenopausal 147 121 1.00 (ref.) 1.00 (ref.) Perimenopausal 85 87 0.80 (0.55-1.18) 0.91 (0.61-1.36) Postmenopausal <50 years 115 160 0.59 (0.42-0.83) 0.74 (0.51-1.08) 50-54 years 63 63 0.82 (0.54-1.26) 0.97(0.61-1.53) 55+ years 10 8 1.03 (0.39-2.67) 1.16(0.41-3.23) Unknown age 64 83 0.64 (0.42-0.95) 0.89 (0.57-1.40) Unknown status 83 94 0.73 (0.50-1.06) 0.85 (0.57-1.28) HRT use4 Never 53 67 1.00 (ref.) 1.00 (ref.) Ever, ERT and/or EPRT 89 114 0.99 (0.63-1.56) 0.92 (0.56-1.50) Ever, other type 46 50 1.16(0.68-1.99) 1.00 (0.56-1.80) OC use (vrs) None 124 133 1.00 (ref.) 1.00 (ref.) <2 156 181 0.92 (0.67-1.28) 0.84 (0.59-1.20) 2-5 98 99 1.06 (0.73-1.54) 1.00 (0.66-1.50) 5-9 92 109 0.91 (0.63-1.31) 0.73 (0.49-1.10) > 10 96 92 1.12 (0.77-1.63) 1.00 (0.66-1.52) Unknown 1 2 --- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 7 (continued). Number of pregnancies (gestation >26 weeks) None 148 104 1.00 (ref.) 1.00 (ref.) 1-2 283 282 0.71 (0.52-0.95) 0.75 (0.55-1.03) >3 136 230 0.42 (0.30-0.58) 0.45 (0.31-0.64) Odds ratios and corresponding 95% confidence intervals obtained in multivariate logistic regression models that include age and race. 2 , Odds ratios and corresponding 95% confidence intervals obtained m multivariate logistic regression models that include all covariates in the table and average lifetime hours/week of physical activity (measured from menarche to reference date). 3 Age and race odds ratios are not presented since selection was based on distribution of these two characteristics in the Women’s CARE Study (see methods section). Restricted to postmenopausal women with a known age at menopause. 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 8. Average lifetime physical activity and relative odds of in situ breast cancer. Cases Controls Covariate (n=567) (n=616) OR1 (95% Cl) OR2 (95% Cl) Ever physical activity No 117 115 1.00 (ref.) 1.00 (ref.) Yes 450 501 0.88 (0.66-1.18) 0.65 (0.48-0.90) Average hours/week of lifetime physical activity None 117 115 1.00 (ref.) 1.00 (ref.) <1 hour/week 157 192 0.80 (0.58-1.12) 0.66 (0.46-0.94) 1-4 hours/week 193 206 0.92(0.67-1.27) 0.66 (0.46-0.94) >4 hours/week 100 103 0.95 (0.66-1.39) 0.64 (0.42-0.96) trend3p=0.24 (among exercisers only, trend p=Q.86) Average MET- hrs/wk of lifetime physical activity None 117 115 1.00 (ref.) 1.00 (ref.) >0-3.0 117 137 0.84 (0.59-1.20) 0.70 (0.48-1.03) >3.0-8.0 107 126 0.84 (0.58-1.20) 0.65 (0.44-0.96) >8.0-16.0 100 111 0.89 (0.61-1.29) 0.61 (0.41-0.92) >16.0-32.0 75 77 0.96 (0.64-1.44) 0.63 (0.40-0.98) >32.0 51 50 1.00 (0.63-1.60) 0.65 (0.39-1.08) trend p=0.27 (among exercisers only, trend p=0.81) Adjusted for age and race. 2 Adjusted for age, race, income, body mass index, family history of breast cancer, menopausal status, age at menopause, postmenopausal hormone use, smoking status, and number of pregnancies. 3 Trend tests based on multivariate models that include all covariates listed in footnote 2. 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 9. Hours o f physical activity during various time periods and relative odds of in situ breast cancer. Covariate Cases (n=567) Controls (n=616) OR1 (95% Cl) OR2 (95% Cl) Ten vears after menarche No activity, at any age 117 115 1.00 (ref.) 1.00 (ref.) Activity only at other ages 160 176 0.89 (0.64-1.25) 0.72 (0.50-1.05) <1 hour/week 58 73 0.78 (0.51-1.20) 0.55 (0.35-0.89) 1-4 hours/week 127 132 0.95 (0.66-1.35) 0.71 (0.48-1.06) >4 hours/week 105 120 0.86 (0.60-1.24) 0.58 (0.36-0.91) trend3p=0.18 (among exercisers only, trend p=0.55) Age 20-34 No activity, at any age 117 115 1.00 (ref.) 1.00 (ref.) Activity only at 143 172 0.82 (0.58-1.15) 0.69 (0.47-1.00) other ages <1 hour/week 90 97 0.91 (0.62-1.34) 0.69 (0.45-1.06) 1-4 hours/week 133 147 0.89 (0.63-1.26) 0.59 (0.39-0.88) >4 hours/week 84 85 0.97 (0.65-1.44) 0.63 (0.36-1.11) trend3p=0.53 (among exercisers only, trend p=0.77) Ten vears prior to reference date No activity, at any age 117 115 1.00 (ref.) 1.00 (ref.) Activity only at other ages 74 88 0.83 (0.55-1.24) 0.68 (0.44-1.06) <1 hour/week 75 78 0.95 (0.63-1.42) 0.75 (0.48-1.16) 1-4 hours/week 193 215 0.88 (0.64-1.22) 0.61 (0.43-0.87) >4 hours/week 108 120 0.89 (0.61-1.28) 0.52 (0.33-0.80) trend3p=0.02 (among exercisers only, trend p=0.13) Odds ratios and corresponding 95% confidence intervals obtained in multivariate logistic regression models that include age and race. 2 Adjusted for age, race, income, body mass index, family history of breast cancer, menopausal status, age at menopause, postmenopausal hormone use, smoking status, and number of pregnancies. 3 Trend tests based on multivariate models that include all covariates listed in footnote 2; women with acti vity only in other time periods excluded from the analysis. 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 10. Hours of physical activity during various time periods and relative odds of in situ breast cancer stratified by family history of breast cancer. No family history Family history Average physical # cases/ # cases/ activity controls OR1 (95% Cl) controls OR1 (95% Cl) Lifetime None 95/100 1.00 (ref.) 22/15 1.00 (ref.) <1 hour/week 131/170 0.66 (0.45-0.97) 26/22 0.74 (0.30-1.83) 1-4 hours/week 151/186 0.60 (0.41-0.88) 42/20 1.01 (0.47-2.44) >4 hours/week 80/99 0.53 (0.34-0.82) 20/4 2.29 (0.62-8.44) Trend p 0.04 0.10 Homogeneity of trends p-value=0.02 Ten vears after menarche No activity, at any age 95/100 1.00 (ref.) 22/15 1.00 (ref.) Activity only at other ages 130/154 0.72 (0.48-1.07) 30/22 0.86 (0.35-2.12) <1 hour/week 46/69 0.46 (0.28-0.77) 12/4 1.64 (0.42-6.36) 1-4 hours/week 105/119 0.69 (0.45-1.06) 22/13 0.74 (0.27-1.99) >4 hours/week 81/113 0.48 (0.29-0.78) 24/7 1.67(0.54-5.22) Trend p 0.05 0.26 Homogeneity of trends p-value=0.05 Ten vears prior to reference date No activity, at any age 95/100 1.00 (ref.) 22/15 1.00 (ref.) Activity only at other ages 61/83 0.62 (0.39-0.99) 13/5 1.41 (0.39-5.04) <1 hour/week 63/70 0.70 (0.44-1.13) 12/8 1.05 (0.33-3.31) 1-4 hours/week 159/188 0.61 (0.42-0.90) 34/27 0.59 (0.25-1.42) >4 hours/week 79/114 0.43 (0.26-0.69) 29/6 2.20 (0.70-6.96) Trend p 0.01 0.16 Homogeneity of trends p-value=0.05 Adjusted for age, race, income, body mass index, family history of breast cancer, menopausal status, age at menopause, postmenopausal hormone use, smoking status, and number of pregnancies. 2 Trend tests based on multivariate models that include all covariates listed in footnote 1; women with activity only in other time periods excluded from the analysis. 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Discussion We observed a reduction in BCIS risk among women reporting some exercise activity in the time period of interest (menarche to reference date); however, the odds ratio estimates remained relatively constant with increasing level of exercise activity. Limiting assessment of activity to the ten years after menarche, we observed a modest inverse relationship with exercise activity. We observed a similar relationship for exercise activity during the ten years prior to reference date. Family history of breast cancer modified the association between exercise activity and BCIS risk with no reduction in risk observed among women reporting a first-degree family history of breast cancer. We observed a similar association for invasive breast cancer among postmenopausal women (Carpenter and others in press). These findings suggest that exercise activity, during adolescence and in recent years, may impact BCIS risk, especially among women without a family history of breast cancer. One mechanism by which physical activity may impact breast cancer risk is through hormonal pathways. Other factors that represent hormonal exposures (age at menarche, parity, and HRT) have also been associated with BCIS risk (Claus and others 2001). High levels of moderate and vigorous physical activity prior to menarche may delay the onset of menses (Bernstein and others 1987; Frisch and others 1981); during a woman’s reproductive years, strenuous exercise activity is associated with increased likelihood of secondary amenorrhea, irregular or anovulatory menstrual cycles, and shortened luteal phases of the menstrual cycle 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Shangold and others 1979; Warren and Perlrith 2001). These alterations in menstrual function have been associated with reduced levels of estradiol, progesterone, and follicle-stimulating hormone (FSH), particularly during adolescence (Bonen and others 1981; Ellison and Lager 1986; Shangold and others 1979). Recent physical activity may also act through hormonal mechanisms by continuing to impact ovarian hormone levels among premenopausal women, or may alter circulating estrogen levels directly (Vainio and Bianchini 2002), by lowering BMI or maintaining weight among postmenopausal women (Nelson and others 1988; Thune and others 1998; Tymchuk and others 2000; Verkasalo and others 2001). Other mechanisms that have been proposed for the protective effect of exercise on invasive breast cancer risk may also account for the decrease in risk of BCIS; these include alterations in immune response (Long 2002; Pedersen and others 1989) and increased insulin sensitivity (Helmrich and others 1991; Tymchuk and others 2000). The insulin sensitivity hypothesis is related to the previously mentioned hormonal pathway because higher levels of insulin and decreased insulin sensitivity have been associated with decreased levels of sex hormone-binding globulin (SHBG), and consequently a higher level of free-estradiol, that may increase breast cancer risk (Nestler and others 1991). Since the vast majority of in situ breast carcinoma is initially identified via mammography, appropriate selection of a properly screened control group was essential in our analysis. Thus, we restricted our analysis to women who had been screened within two years of their reference dates to capture annual screening since 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. many women get annual mammograms more than 12 months apart. Further restriction of our control group to women screened within 12 months or within six months of their reference dates did not impact our findings; therefore, we used the two-year cutoff to maximize the available sample size and minimize the likelihood of false negatives among control subjects. Restricting our control group to recently screened women creates a select population that is not representative of the general population; nevertheless, this group more likely represents the base population from which our case patients arose and does not compromise the internal validity of our study results. A strength of our study is the very detailed measure of lifetime exercise activity; further, we have little missing exposure data since in-person interviews were conducted. We also have detailed screening history information, the ability to control for many potential confounding factors, and information on histologic subtype of BCIS which permit separate analyses of DCIS. The examination of DCIS may be the most important histologic subtype of BCIS since it is generally recognized as the final step in the progression to invasive disease (Habel and others 1998; Millikan and others 1995). Our study also has several limitations. We have insufficient sample size to conduct analyses of potential effect modification in more than two categories of an exposure, to assess risk of lobular carcinoma in situ, or to examine extreme levels of physical activity. Finally, since the control population was originally recruited for the Women’s CARE Study, control subjects were not matched exactly to case patients 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on screening; therefore, we reduced our control subject pool by excluding women based on screening history. Nevertheless, it is possible that our observed reduction in risk associated with exercise activity is an artifact related to differences in screening history. Differentiating between an independent effect of physical activity on BCIS risk and artifacts due to screening practices is a challenge in the study of in situ breast carcinomas. Our results suggest that regular physical activity, particularly that done in recent years, is beneficial in lowering risk of breast carcinoma in situ; however, further studies of in situ breast cancer are needed to understand the role of physical activity independently of its relationship to screening practices. 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bibliography Agurs-Collins T, Adams-Campbell L, Sook K, Culten K. 1999. Insulin-like growth factor-I and breast cancer risk in postmenopausal African American women [abstract]. Proceedings of the American Association of Cancer Research 40:152. Ainsworth B, Haskell W, Leon A, Jr. DJ, Montoye H, Sallis J, et.al. 1993. Compendium of physical activities: classification of energy costs of human physical activities. Medicine and Science in Sports and Medicine 25:71-80. Albanes D, Blair A, Taylor P. 1989. Physical activity and risk of cancer in the NHANES I population. American Journal of Public Health 79(6):744-750. American Cancer Society. 2001. Breast Cancer Facts & Figures 2001-2002.. Atlanta, GA: American Cancer Society. American Cancer Society. 2002. Cancer Facts & Figures. . Atlanta, GA: American Cancer Society. Apter D, Vikho R. 1983. Early menarche, a risk factor for breast cancer, indicates early onset of ovulatory cycles. Journal of Clinical Endocrinology & Metabolism 57:82-86. Ballard-Barbash R. 1997. Perspectives on integrating experimental and epidemiologic research on diet, anthropometry, and breast cancer. Journal of Nutrition 127 (Supp. 5)(936S-939S). Bergmann M, Calle E, Mervis C, Miracle-McMahill H, Thun M, Heath C. 1998. Validity of self-reported cancers in a prospective cohort study in comparison to data from state cancer registries. American Journal of Epidemiology 147:556-562. Bernstein L, Henderson B, Hanisch R, Sullivan-Halley J, Ross R. 1994. Physical exercise and reduced risk of breast cancer in young women. Journal of the National Cancer Institute 86(18): 1403-1408. Bernstein L, Ross R, Lobo R, Hanisch R, Krailo M, Henderson B. 1987. The effects of moderate physical activity on menstrual cycle patterns in adolescence: implications for breast cancer prevention. British Journal of Cancer 55:681- 685. 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bernstein L, Ross R, Pike M, Brown J, Henderson B. 1990a. Hormone levels in older women: a study of post-menopausal breast cancer patients and healthy populations controls. British Journal of Cancer 61:298-302. Bernstein L, Yuan J, Ross R, Pike M, Hanisch R, Lobo R, Stanczyk F, Gao Y, Henderson B. 1990b. Serum hormone levels in pre-menopausal Chinese women in Shanghai and white women in Los Angeles from two breast cancer case-control studies. Cancer Causes & Control 1:51-58. Blair S. 1993. Evidence for success of exercise in weight loss and control. Annals of Internal Medicine 119(7):702-706. Blair S, Dowda M, Pate R, Kronenfeld J, Jr. HH, Parker G, Blair A, Fridinger F. 1991. Reliability of long-term recall of participation in physical activity by middle-aged men and women. American Journal of Epidemiology 133(3):266-275. Bonen A, Belcastro A, Ling W, Simpson A. 1981. Profiles of selected hormones during menstrual cycles of teenage athletes. Journal of Applied Physiology 50(3):545-551. Breslow R, Ballard-Barbash R, Munoz K, Graubard B. 2001. Long-term recreational physical activity and breast cancer in the National Health and Nutrition Examination Survey I Epidemiologic Follow-up Study. Cancer Epidemiology, Biomarkers & Prevention 10:805-808. Brinton L, Schairer C, Hoover R, et.al. 1988. Menstrual factors and risk of breast cancer. Cancer Investigations 6:245-254. Broocks A, Pirke K, Schweeiger U, Tuschl R, Laessle R, Strowitzki T, Horl E, Horl T, Haas W, Jeschke D. 1990. Cyclic ovarian function in recreational athletes. Journal of Applied Physiology 68:2083-2086. Burning P, Bonfree J, vanNoord P, deJong-Bakker M, Nooijen W. 1992. Insulin resistance and breast cancer risk. International Journal of Cancer 52:511-516. Calle E, Flanders W, Thun M, Martin L. 1993. Demographic predictors of mammography and pap smear screening in US women. American Journal of Public Health 83:53-60. Calle E, Murphy T, Rodriguez C, Thun M, Heath C. 1998. Occupation and breast cancer mortality in a prospective cohort of US women. American Journal of Epidemiology 148(2): 191-197. 84 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Calle E, Rodriguez C, Jacobs E, Almon M, Chao A, McCullough M, Feigelson H, Thun M. 2002. The American Cancer Society Cancer Prevention Study II Nutrition Cohort-rational, study design, and baseline characteristics. Cancer 94:500-511. Calle E, Terrell D. 1993. Utility of the National Death Index for ascertainment of mortality among Cancer Prevention Study II participants. American Journal of Epidemiology 137:235-241. Carpenter C, Ross R, Paganini-Hill A, Bernstein L. 1999. Lifetime exercise activity and breast cancer risk among post-menopausal women. British Journal of Cancer 80(11):1852-1858. Carpenter C, Ross R, Paganini-Hill A, Bernstein L. in press. Effect of family history, obesity, and exercise on breast cancer risk among postmenopausal women. International Journal of Cancer. Cauley J, Gutai J, Kuller L, LeDonne D, Powell J. 1989. The epidemiology of serum sex hormones in postmenopausal women. American Journal of Epidemiology 129(8): 1120-1131. Cerhan J, Chiu B, Wallace R, Lemke J, Lynch C, Tomer J, Rubenstein L. 1998. Physical activity, physical function, and the risk of breast cancer in a prospective study among elderly women. Journal of Gerontology 53A:M251- M256. Chang K, Lee T, Linares-Cruz G, Fournier S, Lignieres Bd. 1995. Influence of percutaneous administration of estradiol and progesterone on human breast epithelial cell in vivo. Fertility & Sterility 63:785-791. Chen C, White E, Malone K, Daling J. 1997. Leisure-time physical activity in relation to breast cancer among young women (Washington, United States). Cancer Causes and Control 8:77-84. Chu S, Lee N, Wingo P, Senie R, Greenberg R, Peterson H. 1991. The relationship between body mass and breast cancer among women enrolled in the Cancer and Steroid Hormone Study. Journal of Clinical Epidemiology 44:1197-1206. Claus E, Stowe M, Carter D. 2001. Breast carcinoma in situ: risk factors and screening patterns. Journal of the National Cancer Institute 93(23):1811- 1817. 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Colbert L, Hartman T, Malila N, Limburg P, Pietinen P, Virtamo J, Taylor P, Albanes D. 2001. Physical activity in relation to cancer of the colon and rectum in a cohort of male smokers. Cancer Epidemiology, Biomarkers & Prevention 10:265-268. Colditz G, Hankinson S, Hunter D, Willett W, Manson J, Stampfer M, Hennekens C, Rosner B, Speizer F. 1995. The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. New England Journal of Medicine 332(24):1589-1593. Collaborative Group on Hormonal Factors in Breast Cancer. 1997. Breast cancer and hormone replacement therapy: combined reanalysis of data from 51 epidemiological studies involving 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet 350:1047-1059. Conover C, Lee P, Kanaley J, Clarkson J, Jensen M. 1992. Insulin regulation of insulin-like growth factor binding protein-1 in obese and nonobese humans. Journal of Clinical Endocrinology & Metabolism 74:1355-1360. Coogan P, Aschengrau A. 1999. Occupational physical activity and breast cancer risk in the Upper Cape Cod Cancer Incidence Study. American Journal of Industrial Medicine 36:279-285. Coogan P, Clapp R, Newcomb P, Mittendorf R, Bogdan G, Baron J, Longnecker M. 1996. Variation in female breast cancer risk by occupation. American Journal of Industrial Medicine 30:430-437. Coogan P, Newcomb P, Clapp R, Trentham-Dietz A, Baron J, Longnecker M. 1997. Physical activity in usual occupation and risk of breast cancer (United States). Cancer Causes and Control 8:626-631. Cooper G, Sandler D, Whelan E, Smith K. 1996; Association of physical and behavioral characteristics with menstrual cycle patterns in women age 29-31 years. Epidemiology 7:624-628. Cox D. 1972. Regression models and life tables. Journal of the Royal Statistical Society 34:187-220. Curran E, Berghaus L, Vemetti N, Saporita A, Lubahn D, Estes D. 2001. Natural killer cells express estrogen receptor-alpha and estrogen receptor-beta and can respond via a non-estrogen receptor-alpha-mediated pathway. Cellular Immunology 214:12-20. 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. D'Avanzo B, Nanni 0 , LaVecchia C, Franceschi S, Negri E, Giacosa A, Conti E, Montella M, Talamani R, Decarli A. 1996. Physical activity and breast cancer risk. Cancer Epidemiology, Biomarkers & Prevention 5:155-160. Dirx M, Voorrips L, Goldbohm R, vandenBrant P. 2001. Baseline recreational physical activity, history of sports participation, and postmenopausal breast carcinoma risk in the Netherlands Cohort Study. Cancer 92:1638-1649. Dorgan J, Brown C, Barrett M, Splansky G, Kreger B, D'Agostino R, Albanes D, Schatzkin A. 1994. Physical activity and risk of breast cancer in the Framingham Heart Study. American Journal of Epidemiology 139(7):662- 669. Dosemeci M, Hayes R, Vetter R, Hoover R, Tucker M, Engin K, Unsal M, Blair A. 1993. Occupational physical activity, socioeconomic status, and risks of 15 cancer sites in Turkey. Cancer Causes and Control 4:313-321. Ellison P, Lager C. 1986. Moderate recreational running is associated with lowered salivary progesterone profdes in women. American Journal of Obstetrics and Gynecology 154(5):100-1003. Evenson K, Rosamond W, Cai J, Diez-Roux A, Brancati F. 2002. Influence of retirement on leisure-time physical activity: the Atherosclerosis Risk in Communities Study. American Journal of Epidemiology 155:692-699. Falkner K, McCann S, Trevisan M. 2001. Participant characteristics and quality of recall of physical activity in the distant past. American Journal of Epidemiology 154(9):865-872. Falkner K, Trevisan M, McCann S. 1999. Reliability of recall of physical activity in the distant past. American Journal of Epidemiology 150(2): 195-205. Feigelson H, Henderson B. 1996. Estrogens and breast cancer. Carcinogenesis 17:2279-2284. Figueroa J, Sharma J, Jackson J, McDermott M, Hilsenbeck S, Yee D. 1993. Recombinant insulin-like growth factor binding protein-1 inhibits IGF-1, serum, and estrogen dependent growth of MCF-7 human breast cancer cells. Journal of Cellular Physiology 157:229-236. Flatt J. 1993. Dietary fat, carbohydrate balance, and weight maintenance. Annals of the New York Academy of Sciences 683:122-140. 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fontaine K, Faith M, Allison D, Cheskin L. 1998. Body weight and health care among women in the general population. Archives of Family Medicine 7:381-384. Fontaine K, Fleo M, Allison D. 2001. Body weight and cancer screening among women. Journal of Women's Health and Gender Based Medicine 10:463-470. Fraser G, Shavlik D. 1997. Risk factors, lifetime risk, and age at onset of breast cancer. Annals of Epidemiology 7:375-382. Frazier E, Jiles R, Mayberry R. 1996. Use of screening mammography and clinical breast examinations among Black, Hispanic, and White women. Preventive Medicine 25:118-125. Friedenreich C, Bryant H, Coumeya K. 2001a. Case-control study of lifetime physical activity and breast cancer risk. American Journal of Epidemiology 154:336-347. Friedenreich C, Coumeya K, Bryant H. 2001b. Influence of physical activity in different age and life periods on the risk of breast cancer. Epidemiology 12:604-612. Friedenreich C, Coumeya K, Bryant H. 2001c. Relation between intensity of physical activity and breast cancer risk reduction. Medicine & Science in Sports & Exercise 33(9): 1538-1545. Friedenreich C, Rohan T. 1995. Physical activity and risk of breast cancer. European Journal of Cancer Prevention 4:145-151. Friedman L, Woodruff A, Lane M, Weinberg A, Cooper H, Webb J. 1995. Breast cancer screening behaviors and intentions among asymptomatic women 50 years of age and older. American Journal of Preventive Medicine 11:218- 223. Frisch R, Goetz-Welbergen A, McArthur J, et.al. 1981. Delayed menarche and amenorrhea of college athletes in relation to age at onset of training. JAMA 246:1559-1563. Frisch R, Wyshak G, Albright N, Albright T, Schiff I, Jones K, Witschi J, Shiang E, Koff E, Marguglio M. 1985. Lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to non athletes. British Journal of Cancer 52:885-891. 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Frisch R, Wyshak G, Albright N, Albright T, Schiff I, Witschi J, Marguglio M. 1987. Lower lifetime occurrence of breast cancer and cancers of the reproductive system among former college athletes. American Journal of Clinical Nutrition 45:328-335. Gammon M, Schoenberg J, Britton J, Kelsey J, Coates R, Brogan D, Potischman N, Swanson C, Daling J, Stanford J and others. 1998. Recreational physical activity and breast cancer risk among women under age 45. American Journal of Epidemiology 147(3):273-280. Garflnkel L. 1985. Selection, follow-up, and analysis in the American Cancer Society prospective studies. National Cancer Institute Monograph 67:49-52. Gilliland F, Li Y, Baumgartner K, Crumley D, Samet J. 2001. Physical activity and breast cancer risk in Hispanic and non-Hispanic white women. American Journal of Epidemiology 154(5):442-450. Goebelsmann U, Mishell D. 1979. The menstrual cycle. Mishell D, Davajan V, editors. Philadelphia, PA: FA Davis Company. Habel L, Daling J, Newcomb P, Self S, Porter P, Stanford J, Seidel K, Weiss N. 1998. Risk of recurrence after ductal carcinoma in situ of the breast. Cancer Epidemiology, Biomarkers & Prevention 7:689-696. Hankinson S, Willett W, Colditz G, Hunter D, Michaud D, Deroo B, Rosner B, Speizer F, Poliak M. 1998a. Circulating concentrations of insulin-like growth factor-I and risk of breast cancer. Lancet 351:1393-1396. Hankinson S, Willett W, Manson J, Colditz G, Hunter D, Spiegelman D, Barbieri R, Speizer F. 1998b. Plasma sex steroid hormone levels and risk of cancer is postmenopausal women. Journal of the National Cancer Institute 90:1292- 1299. Hankinson S, Willett W, Manson J, et.al. 1995. Alcohol, height, and adiposity in relation to estrogen and prolactin levels in postmenopausal women. Journal of the National Cancer Institute 87:1297-1302. Hanna N, Schneider M. 1983. Enhancement of tumor metastasis and suppression of natural killer cell activity by B-estradiol treatment. The Journal of Immunology 130(2):974-980. Helmrich S, Ragland D, Leung R. 1991. Physical activity and reduced occurrence of non-insulin-dependent diabetes mellitus. New England Journal of Medicine 325:147-152. 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Henderson B, Pike M, Ross R. 1984. Epidemiology and risk factors. In: Bonadonna G, editor. Breast Cancer: diagnosis and management. New York: John Wiley Sons Ltd. p 1-17. Henderson B, Ross R, Bernstein L. 1988. Estrogens as a cause of human cancer: the Richard and Hinda Rosenthal Foundation award lecture. Cancer Research 48:246-253. Henderson B, Ross R, Judd H, Krailo M, Pike M. 1985. Do regular ovulatory cycles increase breast cancer risk? Cancer 56:1206-1208. Hirose K, Tajima K, Hamajima N, Inoue M, Takezaki T, Kuroishi T, Yoshida M, Tokudome S. 1995. A large-scale, hospital-based case-control study of risk factors of breast cancer according to menopausal status. Japanese Journal of Cancer Research 86:146-154. Hsieh C, Lambe P, Lan S, et.al. 1994. Dual effect of parity on breast cancer risk. European Journal of Cancer 7:969-973. Hu Y, Nagata C, Shimizu H, Kaneda N, Kashiki Y. 1997. Association of body mass index, physical activity, and reproductive histories with breast cancer: a case- control study in Gifu, Japan. Breast Cancer Research and Treatment 43:65- 72. Huang Z, Hankinson S, Colditz G, Stampfer M, Hunter D, Manson J, Hennekens C, Rosner B, Speizer F, Willett W. 1997. Dual effects of weight and weight gain on breast cancer risk. JAMA 278:1407-1411. Jones J, Clemmons D. 1995. Insulin-like growth factors and their binding proteins: biological actions. Endocrinology Reviews 16:3-34. Kazar R. 1995. Insulin-resistance, insulin-like growth factor-I and breast cancer: a hypothesis. International Journal of Cancer 62:403-406. Kelsey J, Berkowitz G. 1988. Breast cancer epidemiology. Cancer Research 48:5615-5623. Kerlikowske K, Barclay J, Grady D, Sickles E, Emster V. 1997. Comparison of risk factors for ductal carcinoma in situ and invasive breast cancer. Journal of the National Cancer Institute 89:77-82. Key T, Pike M. 1988. The dose-effect relationship between 'unopposed' oestrogen and endometrial mitotic rate: its central role in explaining and predicting endometrial cancer risk. British Journal of Cancer 57:205-212. 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kleinbaum D. 1994. Logistic Regression. Dietz K, Gail M, Krickeberg K, Singer B, editors. New York: Springer. Kleinbaum G, Kupper L, Morgenstem H. 1982. Epidemiologic Research: Principles and Quantitative Methods. New York: Van Nostrand Reinhold Co. Laidlaw I, Clarke R, Howell A, Owen A, Potten C, Anderson E. 1995. The proliferation of normal human breast tissue implanted into athymic nude mice is stimulated by estrogen but not progesterone. Endocrinology 136:164- 171. Lambe M, Hsieh C, Tsaih S, Ekbom A, Trickopoulos D, Adami H. 1998. Parity, age at first birth, and the risk of carcinoma in situ of the breast. International Journal of Cancer 77:330-332. Layde P, Webster L, Baughman A, et.al. 1989. The independent associations of parity, age at first full-term pregnancy, and duration of breast feeding with risk of breast cancer. Journal of Clinical Epidemiology 42:963-973. Lee I, Cook N, Rexrode K, Buring J. 2001a. Lifetime physical activity and risk of breast cancer. British Journal of Cancer 85(7):962-965. Lee I, Rexrode K, Cook N, Hennekens C, Buring J. 2001b. Physical activity and breast cancer risk: the Women's Health Study (United States). Cancer Causes and Control 12:137-145. Leon D. 1989. A prospective study of the independent effects of parity and age at first birth on breast cancer incidence in England and Wales. International Journal of Cancer 43:986-991. Levi F, Pasche C, Lucchini F, LaVecchia C. 1999. Occupational and leisure time physical activity and the risk of breast cancer. European Journal of Cancer 35(5):775-778. Lipworth L, Adami H, Trichopoulos D, Carlstrom K, Mantzoros C. 1996. Serum steroid hormone levels, sex hormone-binding globulin, and body mass index in the etiology of postmenopausal breast cancer. Epidemiology 7:96-100. Long E. 2002. Tumor cell recognition by natural killer cells. Seminars in Cancer Biology 12:57-61. Luoto R, Latikka P, Pukkala E, Hakulinen T, Vikho V. 2000. The effect of physical activity on breast cancer risk: a cohort study of 30,548 women. European Journal of Epidemiology 16(10):973-980. 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. MacMahon B, Cole P, Brown J. 1973. Etiology of human breast cancer: a review. Journal of the National Cancer Institute 50:21-42. MacMahon B, Cole P, Lin T, Lowe C, Mirra A, Ravnihar B, Salber E, Valaoras V, Yuasa S. 1970. Age at first birth and breast cancer. Bulletin of the World Health Organization 43:209-221. Magnusson C, Baron J, Correia N, Bergstrom R, Adami H, Persson I. 1999. Breast cancer risk following long-term oestrogen- and oestrogen-progestin replacement therapy. International Journal of Cancer 81:339-344. Marchbanks P, McDonald J, Wilson H, Burnett N, Daling J, Bernstein L, Malone K, Strom B, Norman S, Weiss L and others. 2002. The NICHD Women's Contraceptive and Reproductive Experiences Study: methods and operational results. Annals of Epidemiology 12:213-221. Marcus P, Newman B, Moorman P, Millikan R, Baird D, Qaqish B, Stemfeld B. 1999. Physical activity at age 12 and adult breast cancer risk (United States). Cancer Causes and Control 10:293-302. Matthews C, Shu X, Jin F, Dai Q, Herbert J, Ruan Z, Gao Y, Zheng W. 2001. Lifetime physical activity and breast cancer risk in the Shanghai Breast Cancer Study. British Journal of Cancer 84(7):994-1001. Mayer-Oakes S, Atchison K, Matthias R, DeJong F, Lubben J, Schweitzer S. 1996. Mammography use in older women with regular physicians: what are the predictors? American Journal of Preventive Medicine 12:44-50. McGuire W, Jackson J, Figueroa J, Shimasaki S, Powell D, Yee D. 1992. Regulation of insulin-like growth factor binding protein (IGFBP) expression by breast cancer cells: use of IGFBP-1 as an inhibitor of insulin-like growth factor action. Journal of the National Cancer Institute 84:1336-1341. McManus M, Welsch C. 1984. The effects of estrogen, progesterone, and human placental lactogen on DNA synthesis on human breast ductal epithelium uaintained in athymic mice. Cancer 54:1920-1927. McTieman A, Stanford J, Weiss N, Daling J, Voigt L. 1996. Occurrence of breast cancer in relation to recreational exercise in women age 50-64 years. Epidemiology 7:598-604. Mezzetti M, LaVecchia C, Decarli A, Boyle P, Talamini R, Franceschi S. 1998. Population attributable risk for breast cancer: diet, nutrition, and physical exercise. Journal of the National Cancer Institute 90(5):389-394. 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Michaud D, Giovannucci E, Willett W, Colditz G, Stampfer M, Fuchs C. 2001. Physical activity, obesity, and height, and the risk of pancreatic cancer. JAMA 286:921-929. Millikan R, Dressier L, Geradts J, Graham M. 1995. The need for epidemiologic studies of in-situ carcinoma of the breast. Breast Cancer Research and Treatment 35:65-77. Mittendorf R, Longnecker M, Newcomb P, Dietz A, Greenberg E, Bogdan G, Clapp R, Willett W. 1995. Strenuous physical activity in young adulthood and risk of breast cancer (United States). Cancer Causes and Control 6:347-353. Moore D, Folsom A, Mink P, Flong C, Anderson K, Kushi L. 2000. Physical activity and incidence of postmenopausal breast cancer. Epidemiology 11:292-296. Moradi T, Adami H, Bergstrom R, Gridley G, Wolk A, Gerhardsson M, Dosemeci M, Nyren O. 1999. Occupational physical activity and risk for breast cancer in a nationwide cohort study in Sweden. Cancer Causes and Control 10:423- 430. Moradi T, Adami H, Ekbom A, Wedren S, Terry P, Floderus B, Lichtenstein P. 2002. Physical activity and risk for breast cancer: a prospective cohort study among Swedish twins. International Journal of Cancer 100:76-81. Moradi T, Nyren O, Zack M, Magnusson C, Persson I, Adami H. 2000. Breast cancer risk and lifetime leisure-time and occupational physical activity (Sweden). Cancer Causes and Control 11:523-531. Nagata C, Shimizu H, Takami R, Flayashi M, Takeda N, Yasuda K. 2000. Relations of insulin resistance and serum concentrations of estradiol and sex hormone- bonding globulin to potential breast cancer risk factors. Japanese Journal of Cancer Research 91:948-953. National Institutes of Health. 1977. Summary Staging Guide. . Cancer Surveillance Epidemiology and End Results Reporting. Bethesda, MD: NIH. Nelson M, Meredith C, Dawson-Hughes B, Evans W. 1988. Hormone and bone mineral status in endurance-trained and sedentary postmenopausal women. Journal of Clinical Endocrinology and Metabolism 66(5):927-933. Nestler J, Powers L, Matt D, Steingold K, Plymate S, Rittmaster R, Clore J, Blackard W. 1991. A direct effect of hyperinsulinemia on serum sex hormone-binding globulin levels in obese women with the polycystic ovary syndrome. Journal of Clinical Endocrinology & Metabolism 72:83-89. 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Newcomb P, Klein R, Klein B, Haffher S, Mares-Perlman J, Cruickshanks K, Marcus P. 1995. Association of dietary and life-style factors with sex hormones in postmenopausal women. Epidemiology 6:318-321. Newcomb P, Titus-Emstoff L, Egan K, Trentham-Dietz A, Baron J, Storer B, Willett W, Stampfer M. 2002. Postmenopausal estrogen and progestin use in relation to breast cancer risk. Cancer Epidemiology, Biomarkers & Prevention 11:593-600. Pedersen B, Tvede N, Christensen L, Lkarlund K, Kragbak S, Halkjr-Kristensen J. 1989. Natural killer cell activity in peripheral blood of highly trained and untrained persons. International Journal of Sports Medicine 10(2): 129-131. Petrelli J, Calle E, Rodriguez C, Thun M. 2002. Body mass index, height, and postmenopausal breast cancer mortality in a prospective cohort of US women. Cancer Causes and Control in press. Pike M, Henderson B, Casagrande J, Rosario I, Gray G. 1981. Oral contraceptive use and early abortion as risk factors for breast cancer in young women. British Journal of Cancer 43:72-76. Pike M, Spicer D, Dahmoush L, Press M. 1993. Estrogens, progestogens, normal breast cell proliferation, and breast cancer risk. Epidemiologic Reviews 15(1): 17-35. Pukkala E, Poskiparta M, Apter D, Vihko V. 1993. Life-long physical activity and cancer risk among Finnish female teachers. European Journal of Cancer Prevention 2:369-376. Rockhill B, Willett W, Hunter D, Manson J, Hankinson S, Colditz G. 1999. A prospective study of recreational physical activity and breast cancer risk. Archives of Internal Medicine 159:2290-2296. Rockhill B, Willett W, Hunter D, Manson J, Hankinson S, Spiegelman D, Colditz G. 1998. Physical activity and breast cancer risk in a cohort of young women. Journal ofthe National Cancer Institute 90(15):1155-1160. Rossing M, Remler R, Voigt L, Wicklund K, Daling J. 2001. Recreational physical activity and risk of papillary thyroid cancer (United States). Cancer Causes and Control 12:881-885. Ruffin M, Gorenflo D, Woodman B. 2000. Predictors of screening for breast, cervical, colorectal, and prostate cancer among community-based primary care practices. Journal of the American Board of Family Practice 13:1-10. 94 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Seaman W, Gindhart T. 1979. Effect of estrogen on natural killer cells. Arthritis and Rheumatology 22:1234-1240. Sellers T, Kushi L, Potter J, Kaye S, Nelson C, McGovern P, Folsom A. 1992. Effect of family history, body-fat distribution, and reproductive factors on the risk of postmenopausal breast cancer. New England Journal of Medicine 326:1323-1329. Sesso H, Paffenbarger R, Lee I. 1998. Physical activity and breast cancer risk in the College Alumni Health Study (United States). Cancer Causes and Control 9:433-439. Shangold M, Freeman R, Thysen B, Gatz M. 1979. The relationship between long distance running, plasma progesterone, and luteal phase length. Fertility and Sterility 31:130-133. Shoff S, Newcomb P, Trentham-Dietz A, Remington P, Mittendorf R, Greenberg E, Willett W. 2000. Early-life physical activity and postmenopausal breast cancer: effect of body size and weight change. Cancer Epidemiology, Biomarkers & Prevention 9:591-595. Sonnenschein E, Toniolo P, Terry M, Briming P, Kato I, Koenig K, Shore R. 1999. Body fat distribution and obesity in pre- and postmenopausal breast cancer. International Journal of Epidemiology 28:1026-1031. Stallone D. 1994. The influence of obesity and its treatment on the immune system. Nutritional Reviews 52:37-50. Steenland K, Nowlin S, Palu S. 1995. Cancer incidence in the National Health and Nutrition Survey I follow-up data: diabetes, cholesterol, pulse, and physical activity. Cancer Epidemiology, Biomarkers & Prevention 4:807-811. Taioli E, Barone J, Wynder E. 1995. A case-control study on breast cancer and body mass. European Journal of Cancer 31 A(5):723-728. Thomeycroft I, Jr. DM, Stone S, Kharma K, Nakamura R. 1971. The relation of serum 17-hydroxyprogesterone and estradiol-17-beta levels during the human menstrual cycle. American Journal of Obstetrics & Gynecology 111(7):947- 951. Thune I, Brenn T, Lund E, Gaard M. 1997. Physical activity and the risk of breast cancer. New England Journal of Medicine 336(18): 1269-1275. 95 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Thune I, Furberg A. 2001. Physical activity and cancer risk: dose-response and cancer, all sites and site-specific. Medicine & Science in Sports & Exercise 33(Supp 6):S530-S550. Thune I, Njolstad I, Lochen M, Forde O. 1998. Physical activity improves the metabolic risk profiles in men and women. Archives of Internal Medicine 158:1633-1640. Toniolo P, Levitz M, Zeleniuch-Jacquotte A, Baneijee S, Koenig K, Shore R, Strax P, Pasternack B. 1995. A prospective study of endogenous estrogens and breast cancer in postmenopausal women. Journal of the National Cancer Institute 87:190-197. Trentham-Dietz A, Newcomb P, Storer B, Remington P. 2000. Risk factors for carcinoma in situ of the breast. Cancer Epidemiology, Biomarkers & Prevention 9:697-703. Trichopoulos D, MacMahon B, Cole P. 1972. Menopause and breast cancer risk. Journal of the National Cancer Institute 48:605-613. Tymchuk C, Tessler S, Barnard R. 2000. Changes in sex hormone-binding globulin, insulin, and serum lipids in postmenopausal women on a low-fat, high-fiber diet combined with exercise. Nutrition and Cancer 38(2): 158-162. Ueji M, Ueno E, Osei-Hyiaman D, Takahashi H, Kano K. 1998. Physical activity and the risk of breast cancer: a case-control study of Japanese women. Journal of Epidemiology 8:116-122. Vainio H, Bianchini F, editors. 2002. LARC Handbooks on Cancer Prevention: Weight Control and Physical Activity. Lyon: IARC Press. vandenBrandt P, Spiegelman D, Yaun S, Adami H, Beeson L, Folsom A, Fraser G, Goldbohm R, Graham S, Kushi L and others. 2000. Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. American Journal of Epidemiology 152:514-527. Vena J, Graham S, Zielezny M, Erasure J, Swanson M. 1987. Occupational exercise and risk of cancer. American Journal of Clinical Nutrition 45:318-327. Verkasalo P, Thomas H, Appleby P, Davey G, Key T. 2001. Circulating levels of sex hormones and their relation to risk factors for breast cancer: a cross-sectional study in 1092 pre- and postmenopausal women (United Kingdom). Cancer Causes and Control 12:47-59. 96 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Verloop J, Rookus M, Kooy KVD, Leeuwen FV. 2000. Physical activity and breast cancer risk in women aged 20-54 years. Journal of the National Cancer Institute 92(2):128-135. Vihko V, Apter D, Pukkala E, Oinonen M, Hakulinen T, Vihko R. 1992. Risk of breast cancer among female teachers of physical education and languages. Acta Oncologica 31(2):201-204. Wannamethee S, Shaper A, Walker M. 2001. Physical activity and risk of cancer in middle-aged men. British Journal of Cancer 85:1311-1316. Warren M. 1980. The effects of exercise on pubertal progression and reproductive function in girls. Journal of Clinical Endocrinology and Metabolism 51:1150- 1157. Warren M, Perlrith N. 2001. The effects of intense exercise on the female reproductive system. Journal of Endocrinology 170:3-11. Willett W. 1997. Fat, energy, and cancer. Journal of Nutrition 127:921S-923S. Williamson D, Mandans J, Anda R, Kleinman J, Kahn H, Byers T. 1993. Recreational physical activity and ten-year weight change in a US national cohort. International Journal of Obesity 17:279-286. Winchester D, Jeske J, Goldschmidt R. 2000. The diagnosis and management of ductal carcinoma in-situ of the breast. Cancer Journal for Clinicians 50:184- 200 . World Health Organization. 1977. International Classification of Diseases Ninth Revision. . Manual of the international statistical classification of disease, injuries, and cases of death. 9th rev. ed. Geneva: WHO. World Health Organization. 1990. International classification of diseases for oncology. . 2nd ed. Geneva: WHO. Wyshak G, Frisch R. 2000. Breast cancer among former college athletes compared to non-athletes: a 15-year follow-up. British Journal of Cancer 82(3):726-730. Yager J. 2000. Endogenous estrogens as carcinogens through metabolic activation. Journal of the National Cancer Institute Monographs 27:67-73. Yu H, Rohan T. 2000. Role of the insulin-like growth factor family in cancer development and progression. Journal of the National Cancer Institute 92:1472-1489. 97 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Zheng W, Shu X, McLaughlin J, Chow W, Gao Y, Blot W. 1993. Occupational physical activity and the incidence of cancer of the breast, corpus uteri, and ovary in Shanghai. Cancer 71:3620-3624. 98 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix A Table 11. Frequencies of individual reportable activities by total MET-hrs/wk category, CPS-II Nutrition Cohort, 1992-1997. MET-hrs/wk in 1992 >0-7 >7-17.5 >17.5- >31.5- ... >42.0 31.5 - - - - - - 42.0 # of women 24,379 .... 23.478 2,338.... 1,958 Median MET-hrs/wk 3.5 13.5 24.5 35.5 52.5 Walking 1 -3 hours 92.0 54.0 15.9 19.8 18.9 4-6 hours - 43.0 42.7 43.8 29.3 7+ hours - - 37.4 32.3 46.3 Jogging/running 1 -3 hours 0.2 1.5 2.6 6.0 8.1 4-6 hours - - 0.3 2.5 4.6 7+ hours - - - - 4.6 Swimming 1-3 hours 0.9 8.1 10.7 14.1 12.4 4-6 hours - - 1.6 9.5 14.5 7+ hours - - - - 9.5 Biking 1-3 hours 3.2 31.2 30.3 32.8 22.2 4-6 hours - 0.5 6.8 16.6 13.8 7+ hours - - 0.8 3.8 13.9 Tennis 1 -3 hours 0.3 3.9 4.9 6.3 5.2 4-6 hours - - 6.3 10.7 11.0 7+ hours - - - 1.0 14.0 Aerobics 1-3 hours 2.4 21.7 21.7 31.0 20.2 4-6 hours - - 5.1 21.0 22.5 7+ hours - - 0.1 2.2 12.9 Dance 1 -3 hours 7.1 8.8 9.4 23.1 18.7 4-6 hours - 1.0 1.6 6.1 5.7 7+ hours 0.7 0.7 2.5 6.9 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix B Table 12. Total MET-hrs/wk for individual reportable activities, CPS-II Nutrition Cohort, 1992-1997. _ _ _ _ _ ' ' — —— 1992 >0-7 >7-17.5 >17.5- 31.5 >31.5- 42.0 >42.0 Total # of women Median MET-hrs/wk 24,379 3.5 23,478 13.5 13,436 24.5 2,338 35.5 1,958 52.5 Total MET-hrs/wk (%) 93,952.5 (100.0) 271,648.5 (100.0) 313,955.0 (100.0 84,387.0 (100.0) 112,975.5 (100.0) % total MET-hrs/wk from... Walking 84.8 68.4 67.2 40.9 28.1 Jogging/running 0.3 0.9 1.2 3.1 7.2 Swimming 1.6 4.9 5.1 10.1 16.6 Biking 3.3 11.5 10.8 14.0 12.1 T ennis/racquetball 0.5 2.0 4.5 9.3 15.3 Aerobics/ calisthenics 2.9 8.5 8.2 16.3 15.1 Dance 6.6 _ J b 8 _ 3.0 6.3 5.5 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix C Table 13. Rate ratios for combined leisure-time physical activity reported for age 40 (past) and for 1992 baseline (recent) and breast cancer, CPS-II Nutrition Cohort, 1992-1997. MET-hrs/week1 # cases/# ^ ( 9 5 % Cl) RJT(95% Cl) women2 Long-term, non-exercisers 0.94 0.95 61/3,147 (0.72-1.23) (0.73-1.24) Long-term, light exercisers 484/23,042 1.00 (ref.) 1.00 (ref.) Long-term, moderate 1.03 1.03 exercisers 430/19,862 (0.90-1.17) (0.90-1.17) Long-term, heavy exercisers 0.95 0.95 105/5,280 (0.77-1.17) (0.77-1.18) Past, light/moderate 0.98 0.95 exercisers 55/2,699 (0.74-1.30) (0.72-1.26) Past, moderate/heavy 0.63 0.62 exercisers 9/690 (0.33-1.23) (0.32-1.20) Recent, light/moderate 1.07 1.06 exercisers 138/6,081 (0.88-1.29) (0.87-1.28) Recent, moderate/heavy 0.92 0.94 exercisers 25/1,264 (0.62-1.38) (0.63-1.41) Past light, recent heavy 1.07 1.04 exercisers 67/3,495 (0.88-1.30) (0.86-1.27) Recent light, past heavy 0.90 0.92 exercisers 129/5,770 (0.69-1.16) (0.71-1.18) 1 MET-hrs/week for 1992 baselineand recalled at age 40 based on the following activities: walking, jogging/running, bicycling, swimming, aerobics/calisthenics, tennis/racquetball, and dancing. 2 1,278 women (17 cases) excluded for missing information on activity at age 40. 3 Age-adjusted rate ratio and corresponding 95% confidence interval. 4 Multivariate-adjusted rate ratio and 95% confidence interval adjusted for: age, race, BMI, weight change from age 18 to 1992, family history of breast cancer, personal history of breast cysts, duration of OC use, HRT use, parity, age at menarche, age at menopause, smoking, alcohol intake, caloric intake, education, and mammography history. 101 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix D Table 14. Rate ratios for combined leisure-time physical activity reported at baseline in 1982 (past) and 1992 (recent) and breast cancer, CPS-II Nutrition Cohort, 1992- 1997. Average exercise level1 ” "leases” / # " r R5 (95% Cl) RR4 (95% Cl) women2 Long-term, non-exercisers 0.58 0.57 4/305 (0.22-1.57) (0.21-1.52) Long-term, light exercisers 294/12,964 1.00 (ref.) LOO (ref.) Long-term, moderate 0.87 0.92 exercisers 530/26,397 (0.75-1.00) (0.80-1.06) Long-term, heavy exercisers 28/1,718 0.72 0.78 (0.49-1.06) (0.53-1.16) Past, light/moderate 1.08 1.07 exercisers 59/2,453 (0.82-1.43) (0.81-1.42) Past, moderate/heavy 0.73 0.79 exercisers 63/3,824 (0.55-0.96) (0.60-1.03) Recent exercisers (any level)5 0.92 0.91 16/768 (0.56-1.53) (0.55-1.51) Past light, recent heavy 0.99 1.05 exercisers 115/5,714 (0.85-1.15) (0.91-1.23) Recent light, past heavy 0.88 0.91 exercisers 397/17,563 (0.71-1.09) (0.74-1.14) 1982 exercise measured as: none, slight, moderate, heavy exercise; 1992 exercise measured as MET-hrs/week based on the following activities: walking, jogging/running, bicycling, swimming, aerobics/calisthenics, tennis/racquetball, and dancing. 2 902 women (14 cases) excluded for missing information on activity at age 40. 3 Age-adjusted rate ratio and corresponding 95% confidence interval. 4 Multivariate-adjusted rate ratio and 95% confidence interval adjusted for: age, race, BM1, weight change from age 18 to 1992, family history of breast cancer, personal history of breast cysts, duration of OC use, HRT use, parity, age at menarche, age at menopause, smoking, alcohol intake, caloric intake, education, and mammography history. 5 A 1 1 women with only recent exercise grouped together due to small number of cases. 102 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix E Table 15. Comparison of cases and controls on various breast cancer risk factors, in situ breast cancer study.____________________________ __________________ Covariate Cases (n=567) Controls (n=616) p-value Reference age 51.6 ±7.2 51.6 ± 7.7 0.71 Race (% black) 16.2 40.9 <0.001 Educational level (%) < high school graduate 29.5 34.1 Some college 31.8 35.4 > college graduate 38.8 30.5 0.01 Income (%) <15K 9.0 12.2 15K-35K 20.3 22.2 35K-70K 29.6 35.5 >7 OK 41.3 30.1 0.001 Age at menarche 12.5 ± 1.5 12.4 ±1.6 0.20 Smoking status (%) Never 45.5 46.1 Current 15.2 19.6 Former 39.3 34.3 0.06 Body mass index (BMI) 24.7 ±7.5 26.7 ± 10.4 <0.001 Family history of breast cancer 19.6 10.0 <0.001 (% yes) Nulliparous (%) 26.1 16.9 <0.001 Number of pregnancies (>26 2.3 ± 1.3 2.7 ± 1.6 <0.001 weeks)2 OC use (% ever use) 78.1 78.3 0.92 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Table 15 (continued). Menopausal status Premenopausal 25.9 19.6 Perimenopausal 15.0 14.1 Postmenopausal 44.5 51.0 Unknown status3 14.6 15.3 HRT use4 (% ever use) 47.3 49.4 Unless otherwise specified, mean + SD presented. Subjects with missing values were excluded for that specific variable. 2Among parous women only. 3 Unknown menopausal status if bilateral oophorectomy unknown. 4Among postmenopausal women only. 104 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix F Table 16. Average lifetime physical activity and relative odds of ductal breast carcinoma in situ. Covariate Cases (n=567) Controls (n=616) OR1 (95% Cl) OR2 (95% Cl) Ever physical activity No 103 115 1.00 (ref.) 1.00 (ref.) Yes 376 501 0.84 (0.62-1.13) 0.63 (0.46-0.88) Average hours/week of lifetime physical activity None 103 115 1.00 (ref.) 1.00 (ref.) <1 hour/week 132 192 0.77 (0.54-1.09) 0.63 (0.44-0.92) 1-4 hours/week 161 206 0.87 (0.62-1.22) 0.64 (0.44-0.92) >4 hours/week 83 103 0.90 (0.61-1.33) 0.62 (0.40-0.95) p-trend3 =0.24 (among exercisers only, p-trend=0.92) Average MET- hrs/wk of lifetime physical activitv None 103 115 1.00 (ref.) 1.00 (ref.) >0-3.0 99 137 0.81 (0.56-1.17) 0.68 (0.46-1.02) >3.0-8.0 88 126 0.78 (0.53-1.14) 0.62 (0.41-0.92) >8.0-16.0 83 111 0.84 (0.57-1.23) 0.59 (0.39-0.91) >16.0-32.0 66 77 0.96(0.63-1.46) 0.65 (0.41-1.03) >32.0 40 50 0.89 (0.55-1.46) 0.59(0.35-1.01) p-trend3 =0.21 (among exercisers only, p-trend=0.73) Adjusted for age and race. 2 Adjusted for the following factors: age, race, income, body mass index, family history of breast cancer, menopausal status, age at menopause, postmenopausal hormone use, smoking status, age at menarche, and number of pregnancies. 3 Trend tests based on multivariate models that included all covariates listed in footnote 2. 105 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Appendix G Table 17. Odds ratios' and corresponding 95% confidence intervals for physical activity and in situ breast cancer by various factors. Average lifetime physical activity (# cases/#controls and OR (95% CD) 1-4 >4 Covariate None <1 hour/week hours/week hours/week Income <35,000 47/61 57/64 50/65 26/29 1.00 0.80 0.62 0.73 (ref.) (0.45-1.41) (0.34-1.11) (0.36-1.47) >35,000 70/54 100/128 143/141 74/74 1.00 0.57 0.65 0.61 (ref) (0.36-0.90) (0.41-1.01) (0.37-1.01) p2=0.68 Education 90/101 100/132 103/134 54/61 < college 1.00 0.76 0.66 0.74 graduate (ref.) (0.50-1.14) (0.43-0.99) (0.45-1.21) 27/14 57/60 90/72 46/42 > college 1.00 0.42 0.50 0.40 graduate (ref.) (0.19-0.92) (0.24-1.07) (0.18-0.91) p2=0.46 BMI (kg/m2 ) 65/49 100/91 137/112 76/66 <25.0 1.00 0.72 0.69 0.65 (ref.) (0.44-1.17) (0.43-1.10) (0.38-1.11) 52/66 57/101 56/94 24/37 >25.0 1.00 0.61 0.63 0.67 (ref.) (0.37-1.01) (0.38-1.06) (0.35-1.30) p2 =0.95 Family history of breast cancer 95/100 131/170 151/186 80/99 No 1.00 0.66 0.60 0.53 (ref.) (0.45-0.97) (0.41-0.88) (0.34-0.82) 22/15 26/22 42/20 20/4 Yes 1.00 0.74 1.01 2.29 (ref.) (0.30-1.83) (0.47-2.44) (0.62-8.44) p2 =0.1 1 Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission. Table 17 (continued). Menopausal status 36/25 56/68 88/77 52/38 Premenopausal 1.00 0.50 0.61 0.69 (ref.) (0.26-0.96) (0.33-1.13) (0.35-1.39) 61/70 79/99 75/95 37/50 Postmenopausal 1.00 0.78 0.67 0.60 (ref.) (0.48-1.26) (0.41-1.09) (0.34-1.08) p2 =0.46 HRT use 15/20 21/17 17/34 7/12 Never 1.00 1.38 0.57 0.55 (ref) (0.52-3.70) (0.22-1.46) (0.16-1.88) 32/37 34/58 42/47 21/25 Ever 1.00 0.61 0.78 0.74 (ref.) (0.31-1.21) (0.40-1.53) (0.33-1.64) p2 =0.20 Oral contraceptive use 28/29 32/41 37/46 27/17 Never 1.00 0.64 0.64 1.10 (ref.) (0.30-1.34) (0.31-1.31) (0.47-2.57) 89/85 125/150 155/160 73/86 Ever 1.00 0.67 0.63 0.53 (ref.) (0.44-1.00) (0.42-0.95) (0.33-0.85) p2 =0.33 Parity 23/13 30/27 56/43 39/21 Nulliparous 1.00 0.49 0.53 0.72 (ref) (0.20-1.19) (0.24-1.26) (0.29-1.77) 94/102 127/165 137/163 61/82 Parous 1.00 0.72 0.72 0.62 (ref.) (0.49-1.05) (0.49-1.06) (0.39-0.99) p2 =0.63 Smoking status 54/49 79/91 86/94 39/50 Never 1.00 0.73 0.63 0.55 (ref.) (0.43-1.22) (0.37-1.05) (0.30-1.00) 63/66 78/101 107/112 61/53 Ever 1.00 0.69 0.75 0.80 (ref.) (0.43-1.12) (0.47-1.19) (0.47-1.37) p2=0.66 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 17 (continued). Age at menarche 36/27 40/57 51/59 18/26 1.00 0.41 0.45 0.36 (ref.) (0.21-0.81) (0.23-0.87) (0.16-0.81) 81/88 117/135 142/147 82/77 1.00 0.80 0.75 0.77 (ref.) (0.53-1.21) (0.50-1.13) (0.48-1.23) p2 =0.31 Adjusted for the following factors: age, race, income, body mass index, family history of breast cancer, menopausal status, age at menopause, postmenopausal hormone use, smoking status, age at menarche, and number of pregnancies. 2 Tests of homogeneity of trends based on multivariate models that included all covariates listed in footnote 1. Reproduced with permission ofthe copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Recreational physical activity and risk of breast cancer: The California Teachers Study

PDF

Physical activity and adenomatous polyps: Measures of association and impact

PDF

Genetic risk factors in breast cancer susceptibility: The multiethnic cohort

PDF

Dietary fats, fat metabolizing genes, and the risk of breast cancer

PDF

Risk factors for breast cancer according to estrogen and progesterone receptor status

PDF

The effects of dietary and other lifestyle behaviors on the risk of colorectal cancer among the Mexican -origin Latino population

PDF

Insulin-like growth factor 1 genotype, phenotype and breast cancer risk, by racial/ethnic group

PDF

Androgens and breast cancer

PDF

The role of genetic variation in insulin -like growth factors in prostate cancer risk: The multiethnic cohort

PDF

Determinants of the age at natural menopause: The multiethnic cohort

PDF

Identifying predictors of cytogenetic relapse and death for patients undergoing allogeneic blood or marrow transplantation for treatment of chronic myeloid leukemia in chronic phase: The City of...

PDF

Single-nucleotide polymorphisms in 17beta-hydroxysteroid dehydrogenase type III and prostate cancer risk

PDF

The association between recreational physical activity and mammographic density

PDF

Androgen receptor gene and prostate -specific antigen gene in breast cancer

PDF

Risk factors, sequellae and intervention studies for the treatment of diabetes mellitus in Native American populations

PDF

beta3-adrenergic receptor gene Trp64Arg polymorphism and obesity-related characteristics among African American women with breast cancer: An analysis of USC HEAL Study

PDF

A case-control study of passive smoking and bladder cancer risk in Los Angeles

PDF

Breast cancer in the multiethnic cohort study: Genetic (prolactin pathway genes) and environmental (hormone therapy) factors

PDF

Assessment of fatigue as a late effect of therapy among survivors of childhood leukemia

PDF

Lymphedema: Impact on breast cancer survivors

Asset Metadata

Creator Patel, Alpa Vipin (author)

Core Title Recreational physical activity and risk of breast cancer

Degree Doctor of Philosophy

Degree Program Epidemiology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag health sciences, oncology,health sciences, public health,health sciences, recreation,OAI-PMH Harvest,women's studies

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-376631

Unique identifier UC11340271

Identifier 3103957.pdf (filename),usctheses-c16-376631 (legacy record id)

Legacy Identifier 3103957.pdf

Dmrecord 376631

Document Type Dissertation

Rights Patel, Alpa Vipin

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA