Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Risk factors for breast cancer according to estrogen and progesterone receptor status
(USC Thesis Other)
Risk factors for breast cancer according to estrogen and progesterone receptor status
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
RISK FACTORS FOR BREAST CANCER ACCORDING TO ESTROGEN AND PROGESTERONE RECEPTOR STATUS by Huiyan Ma 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) August 2006 Copyright 2006 Huiyan Ma UMI Number: 3238309 3238309 2007 UMI Microform Copyright 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, MI 48106-1346 by ProQuest Information and Learning Company. ii Dedication To my husband, Dongchang Yang, my son, Shawn X. Yang, and my daughter, Alice Y. Yang for their support and understanding iii Acknowledgements This work was done under the direction and supervision of my guidance committee chair, Dr. Giske Ursin. I would like to give my gratitude and appreciation to Dr. Ursin for her invaluable guidance and support provided throughout my doctoral study. Furthermore, I would also like to extend my special thanks and gratitude to Dr. Leslie Bernstein for her exceptional guidance, her insight, and her knowledge throughout the course of my research and the writing of my papers. I am especially grateful to Dr. Chi-Ping Chou for his endless guidance and support since 1998, when I was just a student working toward a Master’s Degree. I would also like to extend my appreciation to other members of my guidance committee, Dr. Roberta McKean- Cowdin and Dr. Sue Ellen Martin for their advice and suggestions throughout the course of my research and in the preparation of this manuscript. I would like to express my most sincere appreciation to Dr. Ronald K. Ross for his comments on my data analysis paper – “Hormone-related risk factors for breast cancer in women under age 50 years by estrogen and progesterone receptor status”. Special appreciation also goes to Dr. Malcolm C. Pike for his comments and methodological help with my meta-analysis paper. I would also like to thank Dr. Colin K. Hill for his comments on my data analysis paper – “Low-dose medical iv radiation and breast cancer risk in women under age 50 years by estrogen and progesterone receptor status”. v Table of Contents Dedication ii Acknowledgements iii List of Tables vii List of Figures viii Abstract ix Chapter 1 Breast cancer and steroid hormone receptors 1 1.1 Introduction 1 1.2 Brief history of the role of hormones in breast cancer 2 1.3 What are ER and PR? 3 1.4 Measurement of ER and PR 4 1.5 ER and PR expression in normal and malignant breast tissue 6 1.6 Breast cancer subtypes and their etiologies 8 Chapter 2 Reproductive factors and breast cancer risk according to joint estrogen and progesterone receptor status: a meta-analysis of epidemiological studies 10 2.1 Summary 10 2.2 Introduction 11 2.3 Materials and Methods 12 2.4 Results 17 2.5 Discussion 36 2.6 Conclusions 40 Chapter 3 Hormone-related risk factors for breast cancer in women under age 50 years by estrogen and progesterone receptor status - results from a case-control and a case-case comparison 41 3.1 Summary 41 3.2 Introduction 42 3.3 Materials and Methods 43 vi 3.4 Results 51 3.5 Discussion 57 3.6 Conclusions 68 Chapter 4 Low-dose medical radiation and breast cancer risk in women under age 50 years overall and by estrogen and progesterone receptor status - results from a case-control and a case-case comparison 70 4.1 Summary 70 4.2 Introduction 71 4.3 Materials and Methods 73 4.4 Results 80 4.5 Discussion 93 4.6 Conclusions 98 Bibliography 99 vii List of Tables Table 1.1 American women’s probability of developing breast cancer during the next 10 years by current age 1 Table 1.2 Estrogen expression status in normal breast tissue during different days of the normal menstrual cycle 7 Table 2.1 Basic characteristics of ten studies 19 Table 2.2 Parity and breast cancer risk by ER/PR status 23 Table 2.3 Age at first birth and breast cancer risk by ER/PR status 26 Table 2.4 Breastfeeding and breast cancer risk by ER/PR status 30 Table 2.5 Age at menarche and breast cancer risk by ER/PR status 33 Table 3.1 Demographic and hormone-related risk factors for 1725 case patients and 440 control subjects 52 Table 3.2 Adjusted ORs and 95% CIs of breast cancer associated with hormone-related risk factors by ER/PR status 58 Table 4.1 The characteristics of demography and medical radiation exposure 81 Table 4.2 Adjusted ORs and 95% CIs of breast cancer associated with medical radiation exposure 83 Table 4.3 Adjusted ORs and 95% CIs of breast cancer associated with medical radiation exposure by age at first exposure 87 Table 4.4 Adjusted ORs and 95% CIs of breast cancer associated with medical radiation exposure by parity status 89 Table 4.5 Adjusted ORs and 95% CIs of breast cancer associated with medical radiation exposure by a family history of breast or ovarian cancer 91 viii List of Figures Figure 2.1 Parity and breast cancer risk by ER/PR status: RR with 95% CI per birth 25 Figure 2.2 Age at first birth and breast cancer risk by ER/PR status: RR with 95% CI for oldest versus youngest age category 28 Figure 2.3 Breastfeeding and breast cancer risk by ER/PR status: RR with 95% CI for highest versus lowest category 32 Figure 2.4 Age at menarche and breast cancer risk by ER/PR status: RR with 95% CI for oldest versus youngest age category 35 ix ABSTRACT It has been hypothesized that subtypes of breast cancer defined by estrogen and progesterone receptor (ER/PR) status may have different etiologies. However, the findings on this topic have been inconsistent, especially in younger women. This dissertation presents the first meta-analysis of results from epidemiological studies that have investigated associations of parity, age at first birth, breastfeeding and age at menarche with risk of ER+PR+ and ER-PR- breast cancer. Each birth reduced the risk of ER+PR+ cancer by 11% (relative risk [RR] = 0.89, 95% confidence interval [CI] = 0.84-0.94). Women who were in the highest age at first birth category had, on average, 27% higher risk of ER+PR+ cancer than women who were in the youngest age at first birth category (RR = 1.27, 95% CI = 1.07-1.50). Neither parity nor age at first birth was associated with the risk of ER-PR- cancer. Breastfeeding and late age at menarche were associated with decreased risk of both receptor subtypes of breast cancer. The relationship between hormone-related factors and breast cancer risk overall and by ER/PR status was examined in a case-control study of women under age 50 years. The findings for parity, age at first birth, breastfeeding and age at menarche are consistent with the results from the meta-analysis. In addition, recent average x alcohol consumption was associated with an increased risk of ER+PR+ cancer (P trend = 0.03), but not with risk of ER-PR- cancer (P trend = 0.42). The relationship between low-dose medical radiation and breast cancer risk overall and by ER/PR status was examined in the same case-control study. An elevated breast cancer risk was observed among women who reported having had multiple chest X-rays (P trend = 0.0007), 7 or more mammograms (OR = 1.80, 95% CI = 0.95- 3.42), or dental X-rays without lead apron protection before age 20 years (OR = 1.81, 95% CI = 1.13-2.90). Age at first exposure, parity, and breast or ovarian cancer family history appeared to modify the radiation effect on breast cancer risk, but we found no effect modification by ER/PR status. 1 Chapter 1 Breast cancer and steroid hormone receptors 1.1 Introduction Based on recent Surveillance, Epidemiology and End Results (SEER) Cancer statistics, breast cancer is still the most common cancer among American women (American Cancer Society, 2005). Among U.S. women, the lifetime probability of developing breast cancer is 1 out of 8 (or 13.2%). The probability of developing breast cancer increases with age (Table 1.1) and the rate of increase is greater during the premenopausal period than the postmenopausal period. Data from the American Cancer Society indicate that approximately 269,730 American women were newly diagnosed with and 40,410 women died from breast cancer in 2005 (American Cancer Society, 2005). Table 1.1 American women’s probability of developing breast cancer during the next 10 years by current age* Current age Probability of developing breast cancer during the next 10 years 20 0.05% 1/1985 30 0.44% 1/229 40 1.5% 1/68 50 2.7% 1/37 60 3.8% 1/26 70 4.1% 1/24 Lifetime 13.2% 1/8 *These estimated probabilities are based on cancer incidence from 2000 through 2002 (American Cancer Society, 2005). 2 Although the etiology of breast cancer is not fully understood, animal and human data have shown that ovarian hormones, estrogen and progesterone, are important for breast cancer development. These hormones act through their respective receptors, estrogen receptor (ER) and progesterone receptor (PR). To understand the role of these hormones and their receptors better, I have included in this chapter a brief history of the role of hormones in breast cancer. Subsequently I describe the definition of ER and PR and their function, and explain how ER and PR are measured and what the potential concerns are with these measurements. I also briefly describe what is known about ER and PR expression in normal and malignant breast tissue. Finally, I summarize the research questions that are addressed in this dissertation. 1.2 Brief history of the role of hormones in breast cancer In 1896, Beatson first determined that ovarian function is involved in breast cancer by performing oophorectomies on two women with advanced breast cancer, which resulted in tumor regression (Beatson, 1896). In 1932, Lacassagne discovered a causal relationship between the ovarian hormone, estrogen, and breast tumorigenesis with an experiment, which showed that estrone induced breast tumors in mice (Lacassagne, 1932). In 1962, Huggins et al. reported that another ovarian hormone, progesterone, accelerated the growth of breast cancer and increased the number of tumors in rats (Huggins et al., 1962). Since then, many experimental, clinical, and 3 epidemiologic studies have demonstrated that estrogen and progesterone play important roles in breast tumorigenesis (Bernstein and Ross, 1993; Henderson et al., 1982; Key and Pike, 1988), and the effects are mediated by their respective receptors. 1.3 What are ER and PR? The estrogen and progesterone receptors were identified in late 1960s (Jensen, 1968; Jensen et al., 1967; Jensen et al., 1969) and in early 1970s (Milgrom et al., 1970; Milgrom and Baulieu, 1970), respectively (Milgrom et al., 1970; Milgrom and Baulieu, 1970). They are protein molecules located in the nucleus of hormone target cells. Both receptors contain specific ligand sites to which only their corresponding hormone or closely related molecules can bind. When a hormone (estrogen or progesterone) binds to its receptor, the conformation of the receptor changes and a hormone-receptor complex forms. The hormone-receptor complex then binds to a specific DNA sequence in target gene regulatory regions. With recruitment of coactivator proteins, this complex activates target genes to synthesize messenger RNA (mRNA). Translation of this mRNA generates proteins that further influence cellular behavior, such as inducing mitosis of breast epithelia cells (Gao and Nawaz, 2002; Hanstein et al., 2004; Klinge, 2000; McDonnell and Norris, 2002). 4 1.4 Measurement of ER and PR Historically, ER and PR expression status in breast tissue was defined by the dextran-coated charcoal (DCC) biochemical assay (Leclercq et al., 1975; Leclercq et al., 1973). The DCC biochemical assay measures the amount of radio-labeled specific binding estradiol/progesterone in tissue homogenates (Elwood and Godolphin, 1980). The result is expressed in femtomoles per milligram of cytosol protein (fmol/mg). Based on a specific fmol/mg cutoff point, the breast tissue receptor status can be defined as either positive or negative. The commonly used cutoff point in published literature is 10 fmol/mg (Cooper et al., 1989; Cotterchio et al., 2003; Hislop et al., 1986; Kreiger et al., 1991; McTiernan et al., 1986; Palmer et al., 1991; Wohlfahrt et al., 1999; Yoo et al., 1997; Yoo et al., 1993), with positive receptor status defined as a value greater than or equal to 10 fmol/mg (Cooper et al., 1989; Cotterchio et al., 2003; Kreiger et al., 1991; Palmer et al., 1991; Wohlfahrt et al., 1999; Yoo et al., 1997; Yoo et al., 1993). The DCC assay only detects unoccupied receptors; therefore, endogenous/exogenous hormones and endocrine treatment may interfere with assay results and theoretically result in false negative results (Pichon and Milgrom, 1992; Robertson, 1996; Thorpe, 1988; Wittliff, 1984). In addition, the determination of ER status or PR status with the DCC assay requires at least 0.5 gram of tissue. 5 With the development of antibodies to ER and PR, the DCC assay has been increasingly replaced by monoclonal assays (Greene and Jensen, 1982). Two types of monoclonal assays are used for ER and PR measurement – the quantitative enzyme immunoassay (EIA) (Blankenstein et al., 1987; Nicholson et al., 1986) and the semiquantitative immunocytochemistry assay (ICA), also known as the immunohistochemistry assay (IHC) (Charpin et al., 1988; King and Greene, 1984). EIA measures the concentration of ER or PR in tissue homogenates. ICA localizes the steroid receptor in the tissue structure through labeling with specific monoclonal antibody (Clarke et al., 1997; Markopoulos et al., 1988). ICA results are usually expressed on two scales - the intensity of cell nuclear staining and the percentage of positively stained cells. The intensity of nuclear staining is scored for individual tumor cell nuclei as negative (-)/no staining, plus one (+1)/weak intensity, plus two (+2)/intermediate intensity, or plus three (+3)/strong intensity (Press et al., 2002). Compared to the DCC assay, monoclonal assays are more sensitive and need less breast tissue (Donegan, 1992; Nicholson et al., 1986). This increases the likelihood of accurate results for ER and PR, especially for small tumors. An additional advantage of monoclonal assays is that they measure both unoccupied and occupied receptors, and therefore the results will not be influenced by endogenous/exogenous hormones or endocrine treatment. In general more recent studies have tended to use monoclonal assays. 6 1.5 ER and PR expression in normal and malignant breast tissue ER and PR expression in normal breast tissue is affected by menopausal status. In premenopausal adult breast tissue, the average proportion of epithelial cells that express hormone receptors is approximately 1.8-6.0% for ER and 12.1-29.0% for PR using ICA (Jacquemier et al., 1990; Khan et al., 1999; Williams et al., 1991). ER expression is lower when endogenous estrogen levels are high and vice versa. The level of ER expression in the follicular phase of the menstrual cycle is approximately twice the level in the luteal phase (Battersby et al., 1992; Khan et al., 1998; Ricketts et al., 1991; Williams et al., 1991) (Table 1.2). Such an “inverse” pattern is not seen for PR. PR expression shows a small increase in the luteal phase over follicular phase levels in two studies (Ricketts et al., 1991; Williams et al., 1991), but not in a third study (Battersby et al., 1992). Compared to normal premenopausal breast tissue, the average proportion of epithelial cells that express hormone receptors in normal postmenopausal breast is higher for ER (14.6-26.0%) and lower for PR (2.0- 7.6%) (Jacquemier et al., 1990; Khan et al., 1999). Breast cancer in postmenopausal women is more likely to express ER and PR than breast cancer in premenopausal women. Among premenopausal women, the average proportion of malignant epithelial cells that express receptors is 50-70% for ER and 10-30% for PR (Colleoni et al., 2002). Among postmenopausal women, the mean proportion of malignant epithelial cells that express receptors were reported as 7 Table 1.2 Estrogen expression status* in normal breast tissue during different days of the normal menstrual cycle Days of menstrual cycle Study Measures 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 Ricketts, 1991 % ER+ † women 22% Day1-14 10% Day15-28 Battersby, 1992 % ER+ ‡ women 61%§ Day1-13 34%§ Day14-28 Williams, 1991 Mean % ER+ cells 4.3 Day1-7 4.3 Day8-14 3.7 Day15-21 2.2 Day22-28 Khan, 1998 Median % ER+ cells 5.4 Day1-4 16.9 Day5-9 0.15 Day10-14 2.4 Day15-18 2.6 Day19-23 1.6 Day24-28 *Determined by ICA. † Percentage of women who had 50% or more tumor cells that expressed ER. ‡ Percentage of women who had 5% or more tumor cells that expressed ER. §ER staining in terminal duct lobular unit cells. 7 8 73.3% for ER and 42.7% for PR (Geisler et al., 2001). The expression of steroid hormone receptors is associated with tumor grade and histological types. ER+ and PR+ cells are more common in well-differentiated tumors than in poorly differentiated or undifferentiated tumors (Donegan, 1992). Compared to ductal tumors, lobular tumors are more often ER positive (Stanford et al., 1986). The relationship between the receptor status in normal breast cells and those in malignant breast cells is still unclear. Based on a recent stem cell model (Dontu et al., 2004), receptor negative cancers are proposed to arise from transformation of the most primitive ER- stem/early progenitor cells, while receptor positive cancer arises from one of two origins. One origin is the transformation of a receptor negative stem/early progenitor cell, but, the mutation in the stem cell allows for the differentiation of a subset of cancer cells into receptor positive cells; the alternative origin is through malignant transformation of receptor positive progenitors. 1.6 Breast cancer subtypes and their etiologies It has been hypothesized that estrogen and progesterone exposures may predominantly associate with breast cancers that express ER and PR, but not with those that do not express ER and PR (Elwood and Godolphin, 1980; Hildreth et al., 1983; Huang et al., 2000; Li et al., 2003; Potter et al., 1995; Yasui and Potter, 1999; Zhu et al., 2002) while the effect of non-hormonal risk factors such as radiation 9 exposure may be confined to those that do not express ER and PR (Huang et al., 2000). In other words, subtypes of breast cancer defined by ER and PR status may have different etiologies. A number of epidemiological studies have examined this hypothesis by ER and PR status separately or jointly (Althuis et al., 2004; Habel and Stanford, 1993; Stanford and Greenberg, 1989). However, the findings have been inconsistent, especially in young women. Part of the reason may have been that a number of previous studies were too small to find effects of risk factors on ER/PR negative tumors. To help shed light on this issue, first, this dissertation quantitatively summarizes all epidemiological studies that investigated the association between reproductive factors in relation to ER+PR+ and ER-PR- breast cancer. Second, this dissertation presents data from a large study of early onset (under age 50) breast cancer, both on the association between hormone-related factors and breast cancer risk overall and by estrogen and progesterone receptor status. Third, this dissertation reports the association between low-dose medical radiation and breast cancer risk overall and by estrogen and progesterone receptor status in women under age 50. 10 Chapter 2 Reproductive factors and breast cancer risk according to joint estrogen and progesterone receptor status: a meta-analysis of epidemiological studies 2.1 Summary Although reproductive factors have been known for decades to be associated with breast cancer risk, it is unclear to what extent these associations differ by estrogen and progesterone receptor (ER/PR) status. This report presents the first meta- analysis of results from epidemiological studies that have investigated parity, age at first birth, breastfeeding and age at menarche in relation to ER+PR+ and ER-PR- cancer risk. We calculated summary relative risks (RRs) and corresponding 95% confidence intervals (95% CIs) using a fixed effects model. Each birth reduced the risk of ER+PR+ cancer by 11% (RR per birth = 0.89, 95% CI = 0.84-0.94) while women who were in the highest age at first birth category had, on average, 27% higher risk of ER+PR+ cancer compared to women who were in the youngest age at first birth category (RR = 1.27, 95% CI = 1.07-1.50). Neither parity nor age at first birth was associated with the risk of ER-PR- cancer (RR per birth = 0.99, 95% CI = 0.94-1.05; RR of oldest versus youngest age at first birth category = 1.01, 95% CI = 0.85-1.20). Breastfeeding and late age at menarche decreased the risk of both receptor subtypes of breast cancer. The protective effect of late age at menarche was statistically significantly stronger for ER+PR+ than ER-PR- cancer (RR = 0.72 for 11 ER+PR+ cancer; RR = 0.84 for ER-PR- cancer, P for homogeneity = 0.006). Our findings suggest that breastfeeding (and age at menarche) may act through different hormonal mechanisms than parity and age at first birth. 2.2 Introduction Although it is well known that reproductive factors are associated with breast cancer risk (Bernstein and Ross, 1993; Henderson et al., 1982; Key and Pike, 1988); it is unclear to what extent these associations differ across subtypes of breast cancer defined by estrogen and progesterone receptor (ER/PR) status. There have been three narrative reviews of this topic (Althuis et al., 2004; Habel and Stanford, 1993; Stanford et al., 1986). The review published in 1986 (Stanford et al., 1986) summarized the results from seven clinical case-series and one hospital-based case- control study and did not find convincing evidence for any difference in effects of reproductive factors by ER status. The review published in 1993 (Habel and Stanford, 1993) summarized the results from three population-based and four hospital-based case-control studies and concluded that nulliparity was positively associated with risk of ER+ breast cancer, but not with ER- breast cancer. A review published in 2004 (Althuis et al., 2004) summarized the epidemiological studies published by 2004 and concluded that nulliparity, delayed childbearing, and early age at menarche were consistently associated with increased risk of ER/PR positive 12 cancer, but not with ER/PR negative cancer. They also stated that the protection from breastfeeding did not differ by ER/PR status, but no data were given. The majority of the epidemiological studies reviewed had a small number of cases with receptor negative cancer; and several important questions could not be addressed in these reviews. We have recently conducted two large studies addressing these issues (Ma et al., under review; Ursin et al., 2005), and we have now conducted and report here a meta-analysis to quantitatively summarize all these studies that investigated the association between parity, age at first birth, breastfeeding or age at menarche in relation to ER+PR+ and ER-PR- breast cancer. 2.3 Materials and Methods 2.3.1 Literature search strategy We identified epidemiological studies (cohort or case-control studies) in MEDLINE from the year 1966 to Dec. 1, 2005 by running searches with the key words “Breast Neoplasm/ep [Epidemiology]” and “(ER or PR).mp [mp=title, abstract, name of substance, mesh subject heading]”. We identified additional studies by tracking the references in all identified articles. We noticed that the studies published before 1995 all defined their receptor subtypes according to either ER or PR status and most of them had a hospital-based study design, while most of studies published since 13 1995 used joint ER/PR status to define receptor subtypes and had a population-based study design. Using joint ER/PR status could reduce the chance of including any tumors where one of the receptor statuses was mislabeled. Therefore, for inclusion into this meta-analysis, the identified articles have to have estimates of relative risk for ER+PR+ and ER-PR- breast cancer. We did not summarize the data for the two rare subtypes (ER-PR+ and ER+PR- breast cancer) because few studies reported estimates of relative risk for them. For exposure variables, we focused on the summary of results for reproductive factors that had been more frequently tested across studies, although some studies had also examined other factors such as body mass index, hormone replacement therapy, etc. We totally identified two cohort (Colditz et al., 2004; Potter et al., 1995), five population-based case-control (Britton et al., 2002; Cotterchio et al., 2003; Huang et al., 2000; McCredie et al., 2003; Ursin et al., 2005), and two hospital-based case-control studies (Rusiecki et al., 2005; Yoo et al., 1997) that investigated these issues. We also included one population-based case-control analysis under review (Ma et al., under review). 2.3.2 Meta-analysis We extracted study-specific estimates of relative risk (odds ratios, rate ratios, and risk ratios) and their 95% confidence intervals (95% CIs) for highest versus lowest category of parity, age at first birth, breastfeeding and age at menarche. For this 14 analysis we used RR to refer to any relative risk measure. Two studies presented their results by subgroups only, either by menopausal status (Cotterchio et al., 2003) or by age at birth (Colditz et al., 2004). To obtain one summary estimate, we combined the RRs for the subgroups through weighting their log (base e) RRs by the inverse of their variances. For RRs of parity, seven studies (Britton et al., 2002; Colditz et al., 2004; Cotterchio et al., 2003; Ma et al., under review; McCredie et al., 2003; Potter et al., 1995; Yoo et al., 1997) used nulliparous women as the reference group, while one of our own studies (Ursin et al., 2005) used never pregnant women as the reference group. We re-computed our results to have nulliparous women as the reference group using our original data (Ursin et al., 2005) and changed the lower limit in the highest category from 5 to 3 births for consistency with other studies. For the same reason, we also changed the lower limit in the highest category from 24 months of breastfeeding to 7 months for two of our own studies (Ma et al., under review; Ursin et al., 2005). For parity, we also extracted or computed study-specific trend estimates. One study provided the estimates in the original publication (Yoo et al., 1997); we computed trend estimates using original data for two of our own studies (Ma et al., ; Ursin et al., 2005), and computed trend estimates using RRs (and 95% CIs) for categorical variables in four studies (Britton et al., 2002; Cotterchio et al., 2003; McCredie et al., 2003; Potter et al., 1995) using the method described by Greenland and Longnecker 15 (Greenland and Longnecker, 1992). We used the midpoint of each category in these calculations; while for the open-ended highest category we used its lower limit plus one as an estimate of the mean number of births. For the study of women under age 45 years from Georgia, Washington, or New Jersey, that reported RRs for ever versus never having given birth (Britton et al., 2002), we used two as our estimate of the number of births, since that was close to the mean number of births for parous women under age 45 years in two of our own studies (Ma et al., under review; Ursin et al., 2005). To calculate the estimate for one study, we combined the RRs given for one birth at age 20 and one birth at age 35, weighting their log RRs by the inverse of their variances (Colditz et al., 2004). To summarize the results by menopausal status, we did as follows. We accepted the definitions used in three studies (Cotterchio et al., 2003; Huang et al., 2000; Potter et al., 1995). Three studies restricted eligibility to young women (under age 40 (McCredie et al., 2003), under age 45 (Britton et al., 2002), and under age 50 (Ma et al., under review)). We combined these young women with premenopausal women to form a group representing women who were premenopausal or young. We used fixed effects models to calculate summary RRs for all studies combined, by type of study design and menopausal status (premenopausal or young and 16 postmenopausal status), since we did not detect statistically significant heterogeneity of effects between studies (P ≥ 0.10) (DerSimonian and Laird, 1986). In order to ensure that each particular study contributed the same weight to the summary log RRs for ER+PR+ and ER-PR- cancers, we used the inverse of the average variances of the two subtypes as the weighting variable. This is necessary for the calculated summary values to be directly comparable. In Tables 2 through 5, we list the study- specific RRs (95% CIs), then the summary RRs (95% CIs) by type of study design and menopausal status, and the overall summary estimates for all studies. To test for potential heterogeneity in risk by receptor status, we first calculated the average variance weighted difference in the log relative risks for the two receptor status groups. The variance for each difference in log relative risk is estimated by the sum of their variances. The differences are divided by their pooled variance, these weighted values are summed across studies, and the total is divided by the sum of the inverses of these pooled variances. To create a test statistic, χ 2 (with 1 degree of freedom), this value is squared and then divided by its variance which is the inverse of the sum of the inverses of the pooled variances across studies. Potential heterogeneity in effect by study design and by menopausal status was examined by using standard homogeneity tests. All P values reported for homogeneity are two- sided, and P values less than 0.05 were considered statistically significant. 17 We assessed the possibility of publication bias using Egger’s regression asymmetry test (Egger et al., 1997). This analysis is based on a regression model, where the standard normal deviate is regressed against the study specific estimate of the precision of log RR. When no publication bias is present, the points will scatter around a regression line that runs through the origin. We considered there is publication bias if the intercept of the Egger’s regression line deviated from zero with a two-sided P value of less than 0.10. All analyses were performed using the STATA statistical software (Version 8, College station, Texas). 2.4 Results 2.4.1 Characteristics of Studies Table 2.1 presents the basic characteristics of the ten studies that we reviewed. Nine of the ten studies were published between 1995 and 2005 (Britton et al., 2002; Colditz et al., 2004; Cotterchio et al., 2003; Huang et al., 2000; McCredie et al., 2003; Potter et al., 1995; Rusiecki et al., 2005; Ursin et al., 2005; Yoo et al., 1997) and the other is currently under review (Ma et al., under review). Seven of the ten studies are from the US (Britton et al., 2002; Colditz et al., 2004; Huang et al., 2000; Ma et al., under review; Potter et al., 1995; Rusiecki et al., 2005; Ursin et al., 2005) and the others are from Canada (Cotterchio et al., 2003), Australia (McCredie et al., 18 2003) and Japan (Yoo et al., 1997). These studies include women of all ages. The main source of hormone receptor information was medical records (Britton et al., 2002; Colditz et al., 2004; Cotterchio et al., 2003; Huang et al., 2000; Ma et al., under review; McCredie et al., 2003; Potter et al., 1995; Ursin et al., 2005); in one study a single laboratory provided the data on receptor status (Rusiecki et al., 2005); and one study did not specify the source of their receptor data (Yoo et al., 1997). The percentage of participating cases with available ER/PR status was at least 65% in nine studies (Britton et al., 2002; Colditz et al., 2004; Cotterchio et al., 2003; Huang et al., 2000; Ma et al., under review; McCredie et al., 2003; Potter et al., 1995; Rusiecki et al., 2005; Ursin et al., 2005), while it was only 40% in one hospital-based case-control study (Yoo et al., 1997). The number of subjects included in these analyses ranged from 104 to 2130 for ER+PR+, and 80 to 1081 for ER-PR-. On average, the number of ER-PR- breast cancer cases involved in the analyses was about 46% of that of ER+PR+ cancer cases. All studies considered confounding in their analyses although the confounders included in the models varied by study (Table 2.1). 19 Table 2.1 Basic characteristics of ten studies Source of study subjects Number of subjects Study Cases Controls Age (yrs) Source of receptor information (% subjects with the information) ER+PR+/ ER-PR- cases Controls Adjustment for potential confounders in data analyses Cohort studies Potter, 1995 Iowa Women’s Health Study, US (7 years follow-up, 1986-1992, 241627 person-yrs) 55-69 at baseline Medical record (65%) 414/80 - BMI, BMI at age 18 yrs, WHR, age at menarche, type of MP, age at MP, oophorectomy history, FHBC, parity, age at first live birth, contraceptive, non- contraceptive estrogen use, ALC Colditz, 2004 Nurses’ Health Study Cohort, US (20 yrs follow-up, 1980- 2000, 1029414 person-yrs) 30-55 at baseline in 1976 Medical record (74%) 1281/417 - Age, age at menarche, time since menopause, parity at age, age at birth, FHBC, HBBD, years on ET, years on EPT, BMI, height, ALC Population-based case-control studies Huang, 2000 North Carolina, US DMV/HCF A 1 20-74 Majority from medical record (91%) 381/262 790 Age at menarche, nullyparity/age at first full- term pregnancy, breastfeeding, abortion or miscarriage, BMI, WHR, OC, HT, FHBC, medical radiation to the chest, SMK, ALC, education, age, race Britton, 2002 Georgia, Washington, and New Jersey, US RDD 2 20-44 Medical record (78%) 616/360 1397 Age, race, education, BMI, WHR, parity, age at first birth, breastfeeding, OC, SMK, ALC, recreational exercise at age 12-13 and 1 yr prior to interview, age at menarche, FHBC, MP, geographic site 19 20 Table 2.1 continued. Source of study subjects Number of subjects Study Cases Controls Age (yrs) Source of receptor information (% subjects with the information) ER+PR+/ ER-PR- cases Controls Adjustment for potential confounders in data analyses Population-based case-control studies Cotterchio, 2003 Ontario, Canada Assessment roll of the Ministry of Finance 25-74 Hospital laboratories and medical record (87%) 1901/737 3691 Age, age at menarche, parity, age at first live birth, OC, BMI, ALC, SMK, breastfeeding, HBBD, FHBC, current strenuous activity for pre-MP women; age at MP, HT, and oophorectomy history for post-MP women McCredie, 2003 Victoria and New South Wales, Australia Electoral roll <40 Medical record (81%) 323/181 564 Age, study center, study period, education, country of birth, marital status, FHBC, BMI, age at menarche, number of live births, OC use Ursin, 2005 Georgia, Washington, Michigan, Pennsylvania, and California, US RDD 2 35-64 Medical record (82%) 2130/1081 4668 Age, race, education, FHBC, age at menarche, study site, number of full-term pregnancies and age at first full-term pregnancy only for models of parous women Ma, under review California, US Neighbours 20-49 Medical record (84%) 854/385 440 Age, race, education, FHBC, age at menarche, gravidity, number of full-term pregnancies, OC use, BMI, ALC, MP and HT, age at first full-term pregnancy and breastfeeding only for parous women 20 21 Table 2.1 continued. Source of study subjects Number of subjects Study Cases Controls Age (yrs) Source of receptor information (% subjects with the information) ER+PR+/ ER-PR- cases Controls Adjustment for potential confounders in data analyses Hospital-based case-control studies Yoo, 1997 Aichi Cancer Hospital, Nagoya, Japan Hospital ≥25 Not specified (40%) 176/141 21714 Age, occupation, FHBC, age at menarche, menstrual regularity as a teenager, age at MP, age at first full-term pregnancy, number of full-term pregnancies, breastfeeding, ALC, SMK Rusiecki, 2005 Yale, New Haven Hospital Hospital 40-80 Single laboratory (76%) 104/107 401 Age, race, FHBC, age at menarche, nulliparity/age at first full-term pregnancy, breastfeeding, MP, BMI, ever estrogen use, ALC, SMK Note: MP = menopausal or menopause, BMI = body mass index, WHR = waist:hip ratio, OC = oral contraceptive use, HT = hormone therapy, ET = estrogen therapy, EPT = estrogen and progestin therapy, FHBC= family history of breast cancer, SMK = cigarette smoking , ALC = alcohol drinking, HBBD = history of breast benign diseases, ER+ = estrogen receptor positive, ER- = estrogen receptor negative, PR+ = progesterone receptor positive, PR-= progesterone receptor negative. 1 DMV/HCFA: Division of Motor Vehicles for women under 65 years/Health Care Financing Administration for women aged 65 years or older. 2 RDD: Random digit dialing. 21 22 2.4.2 Parity Eight studies were included in the meta-analysis of parity and breast cancer risk by ER/PR status (Table 2.2). Both the summary RRs for the highest versus the lowest category and the summary RRs per birth indicated that the protective effect of parity was confined to ER+PR+ cancer. Each birth reduced the risk of ER+PR+ cancer by 11% (RR per birth = 0.89, 95% CI = 0.84-0.94), and the P-value for homogeneity between ER+PR+ versus ER-PR- cancer was less than 0.001 (Figure 2.1). 2.4.3 Age at First Birth Nine studies were included in the meta-analysis of age at first birth and breast cancer risk by ER/PR status (Table 2.3, Figure 2.2). Women in the oldest age at first birth category were on average at a 27% greater risk (summary RR = 1.27, 95% CI = 1.07-1.50) for ER+PR+ cancer than women in the youngest age category, but age at first birth was not associated with risk of ER-PR- cancer (summary RR=1.01, 95% CI = 0.85-1.20). The difference in effects between ER+PR+ and ER-PR- was statistically significant (P for homogeneity between ER+PR+ and ER-PR- cancer = 0.010). The summary RR for ER+PR+ cancer appeared greater among postmenopausal than premenopausal or young women (postmenopausal women: summary RR=1.65, 95% CI = 1.15-2.38; premenopausal or young women: summary RR = 1.24, 95% CI= 0.96-1.62), but the difference was not statistically significant (P = 0.211). 23 Table 2.2 Parity and breast cancer risk by ER/PR status RR (95% CI) Study Subgroups by MP status No. of exposure categories Highest/Lowest exposure category ER+PR+ ER-PR- Cohort studies Potter, 1995 Post-MP 3 ≥3/Nulliparous 0.75 (0.52-1.06) 2.24 (0.69-7.24) Per birth 0.96 (0.89-1.03) 1.09 (0.91-1.30) Colditz, 2004 All women 3 4/Nulliparous 0.74 (0.61-0.89) 1.17 (0.80-1.70) Per birth 0.88 (0.79-0.98) 1.13 (0.92-1.41) Summary RRs for cohort studies Highest/Nulliparous 0.74 (0.50-1.10) 1.25 (0.84-1.87) Per birth 0.93 (0.80-1.08) 1.11 (0.95-1.29) Population-based case-control studies Britton, 2002 Young 2 Ever/Nulliparous 0.83 (0.60-1.15) 0.82 (0.55-1.23) Per birth 0.91 (0.77-1.07) 0.91 (0.74-1.11) All women 4 ≥3/Nulliparous 0.62 (0.43-0.90) 0.77 (0.46-1.27) Per birth 0.90 (0.82-0.97) 0.94 (0.84-1.06) Pre-MP 4 ≥3/Nulliparous 0.44 (0.26-0.75) 0.90 (0.46-1.76) Per birth 0.83 (0.73-0.94) 0.97 (0.83-1.14) Post-MP 4 ≥3/Nulliparous 0.71 (0.53-0.97) 0.72 (0.46-1.12) Cotterchio, 2003 Per birth 0.92 (0.86-0.98) 0.93 (0.84-1.03) McCredie, 2003 Young 4 ≥3/Nulliparous 1.0 (0.5-1.2) 0.8 (0.4-1.3) Per birth 1.00 (0.90-1.11) 0.94 (0.81-1.09) Ursin, 2005 All women 4 ≥3/Nulliparous 0.63 (0.54-0.73) 1.07 (0.86-1.32) Per birth 0.86 (0.82-0.90) 1.01 (0.95-1.08) 23 24 Table 2.2 continued. RR (95% CI) Study Subgroups by MP status No. of exposure categories Highest/Lowest exposure category ER+PR+ ER-PR- Ma, under review Young 4 ≥3/ Nulliparous 0.61 (0.42-0.88) 0.93 (0.60-1.44) Per birth 0.86 (0.78-0.95) 0.95 (0.84, 1.06) Summary RRs for population-based case-control studies Highest/Nulliparous 0.67 (0.55-0.82) 0.96 (0.79-1.17) Per birth 0.88 (0.83-0.94) 0.98 (0.92-1.04) Hospital-based case-control study Yoo, 1997 Per 1 unit ↑ Per birth 0.96 (0.79-1.17) 1.00 (0.81-1.23) Pre-MP/Young Highest/Nulliparous Per birth 0.72 (0.52-0.98) 0.90 (0.82-0.98) 0.86 (0.63-1.18) 0.95 (0.86-1.04) Summary RRs by MP status Post-MP Highest/Nulliparous Per birth 0.72 (0.44-1.17) 0.93 (0.84-1.03) 0.86 (0.53-1.42) 0.97 (0.88-1.08) Highest/Nulliparous 0.75 (0.65-0.88) 1.01 (0.87-1.17) Summary RRs for all studies Per birth 0.89 (0.84-0.94) 0.99 (0.94-1.05) P for homogeneity between ER+PR+ and ER-PR- cancer Highest/Nulliparous Per birth P < 0.001 P < 0.001 Note: ER/PR = estrogen and progesterone receptor, MP = menopausal or menopause, ER+ = estrogen receptor positive, ER- = estrogen receptor negative, PR+ = progesterone receptor positive, PR-= progesterone receptor negative, RR = Relative risk, 95% CI = 95% confidence interval. Test for homogeneity across all studies: P ER+PR+ =0.52 and P ER-PR- = 0.77 for highest/nulliparous, P ER+PR+ = 0.84 and P ER-PR- = 0.80 per birth. Egger’s test for publication bias for all studies: P ER+PR+ = 0.98 and P ER-PR- = 0.87. 24 25 Figure 2.1 Parity and breast cancer risk by ER/PR status: RR with 95% CI per birth 25 26 Table 2.3 Age at first birth and breast cancer risk by ER/PR status RR (95% CI) for oldest vs. youngest age category Study Subgroups by MP status No. of exposure categories Oldest/Youngest age category ER+PR+ ER-PR- Cohort study Potter, 1995 Post-MP 2 ≥30/<30 1.76 (1.21-2.56) 1.71 (0.76-3.85) Population-based case-control studies All women 2 >25/ ≤25 1.3 (0.9-1.8) 0.8 (0.5-1.3) Pre-MP 2 >25/ ≤25 1.0 (0.6-1.7) 0.9 (0.5-1.5) Huang, 2000 Post-MP 2 >25/ ≤25 1.6 (1.0-2.7) 0.9 (0.4-1.7) Britton, 2002 Young 2 >24.3/ ≤24.3 1.21 (0.94-1.57) 1.03 (0.75-1.41) All women 3 ≥28/<24 1.43 (1.06-1.92) 1.19 (0.78-1.81) Pre-MP 3 ≥28/<24 1.08 (0.73-1.60) 1.00 (0.60-1.65) Cotterchio, 2003 Post-MP 3 ≥28/<24 1.64 (1.28-2.10) 1.30 (0.89-1.89) McCredie, 2003 Young 2 ≥25/<25 1.7 (1.1-2.5) 0.9 (0.6-1.4) Ursin, 2005 All women 4 ≥30/<20 1.22 (0.97-1.54) 0.91 (0.68-1.22) Ma, under review Young 4 ≥32/<22 1.23 (0.72-2.10) 0.56 (0.30-1.07) Summary RRs for population-based case-control studies 1.31 (1.07-1.60) 0.94 (0.76-1.15) Hospital-based case-control studies Yoo, 1997 All women Per 5 years ↑ 1.19 (0.93-1.51) 1.19 (0.91-1.55) 26 27 Table 2.3 continued. RR (95% CI) for oldest vs. youngest age category Study Subgroups by MP status No. of exposure categories Oldest/Youngest age category ER+PR+ ER-PR- Rusiecki, 2005 All women 2 ≥30/<30 0.5 (0.2-1.2) 1.0 (0.5-1.9) Summary RRs for hospital-based case-control studies 1.10 (0.78-1.55) 1.17 (0.83-1.65) Pre-MP/Young 1.24 (0.96-1.62) 0.92 (0.71-1.20) Summary RRs by MP status Post-MP 1.65 (1.15-2.38) 1.28 (0.89-1.84) Summary RRs for all studies 1.27 (1.07-1.50) 1.01 (0.85-1.20) P for homogeneity between ER+PR+ and ER-PR- cancer P = 0.010 Note: ER/PR = estrogen and progesterone receptor, MP = menopausal or menopause, ER+ = estrogen receptor positive, ER- = estrogen receptor negative, PR+ = progesterone receptor positive, PR-= progesterone receptor negative, RR = relative risk, 95% CI = 95% confidence interval. Test for homogeneity across all studies: P ER+PR+ = 0.80 and P ER-PR- = 0.70. Egger’s test for publication bias for all studies: P ER+PR+ = 0.83 and P ER-PR- = 0.67. 27 28 Figure 2.2 Age at first birth and breast cancer risk by ER/PR status: RR with 95% CI for oldest versus youngest age category 28 29 2.4.4 Breastfeeding The summary RRs from the seven studies of breastfeeding show that breastfeeding was associated with reduced RRs of both ER+PR+ and ER-PR- cancer (summary RRs (95% CIs): 0.95 (0.87-1.05) and 0.91 (0.83-1.00) for ER+PR+ and ER-PR- cancers, respectively) (Table 2.4, Figure 2.3). The protective effect from breastfeeding was observed among population-based case-control studies, but not among hospital-based case-control studies. This difference is marginally significant (P = 0.071 for ER+PR+ and P = 0.054 for ER-PR- cancer). One reason for this discrepancy could be that the two hospital-based studies included small number of cases and therefore had insufficient power to find any effect. 2.4.5 Age at Menarche Nine studies were included in the meta-analysis of age at menarche and breast cancer risk by ER/PR status (Table 2.5, Figure 2.4). The overall summary RRs of the oldest versus the youngest age at menarche category show that age at menarche was negatively associated with the risk of both ER+PR+ and ER-PR- cancer (summary RRs (95% CIs): 0.72 (0.64-0.80) and 0.84 (0.75-0.94) for ER+PR+ and ER-PR- cancers, respectively). The protective effect of late age at menarche was statistically significantly stronger for ER+PR+ than ER-PR- cancer (P for homogeneity between ER+PR+ and ER-PR- cancer = 0.006). 30 Table 2.4 Breastfeeding and breast cancer risk by ER/PR status RR (95% CI) for highest vs. lowest category Study Subgroups by MP status No. of exposure categories Highest/Lowest exposure category ER+PR+ ER-PR- Population-based case-control studies All women 2 Ever/Never 0.7 (0.5-1.0) 0.8 (0.5-1.1) Pre-MP 2 Ever/Never 0.7 (0.4-1.1) 0.7 (0.4-1.1) Huang, 2000 Post-MP 2 Ever/Never 0.8 (0.5-1.2) 1.1 (0.6-1.8) Britton, 2002 Young 3 >12 months/Never 0.80 (0.58-1.12) 0.75 (0.50-1.12) All women >6 months/Never or nulliparous 1.05 (0.79-1.39) 1.01 (0.69-1.48) Pre-MP 3 >6 months/Never or nulliparous 1.41 (0.96-2.08) 0.86 (0.53-1.39) Cotterchio, 2003 Post-MP 3 >6 months/Never or nulliparous 0.92 (0.73-1.16) 1.09 (0.77-1.52) Ursin, 2005 All women 4 >6 months/Never 0.76 (0.65-0.89) 0.68 (0.55-0.83) Ma, under review Young 4 >6 months/Never 0.51 (0.31- 0.84) 0.57 (0.33-1.00) Summary RRs for population-based case-control studies 0.78 (0.64-0.94) 0.74 (0.61-0.89) P for effect modification by receptor status P = 0.80 Hospital-based case-control studies Yoo, 1997 All women Per 3 months ↑ 3 months increase 1.02 (0.95-1.09) 0.98 (0.90-1.07) 30 31 Table 2.4 continued. RR (95% CI) for highest vs. lowest category Study Subgroups by MP status No. of exposure categories Highest/Lowest exposure category ER+PR+ ER-PR- Rusiecki, 2005 40-80 3 ≥12 months/Never 1.25 (0.67-2.50) 0.91 (0.48-1.67) Summary RRs for hospital-based case-control studies 1.02 (0.92-1.14) 0.98 (0.88-1.09) Pre-MP/Young 0.83 (0.61-1.14) 0.73 (0.53-1.00) Summary RRs by MP status Post-MP 0.89 (0.62-1.27) 1.09 (0.77-1.56) Summary RRs for all studies 0.95 (0.87-1.05) 0.91 (0.83-1.00) P for homogeneity between ER+PR+ and ER-PR- cancer P =0.38 Note: ER/PR = estrogen and progesterone receptor, MP = menopausal or menopause, ER+ = estrogen receptor positive, ER- = estrogen receptor negative, PR+ = progesterone receptor positive, PR-= progesterone receptor negative, RR = relative risk, 95% CI = 95% confidence interval. Test for homogeneity across all studies: P ER+PR+ = 0.15 and P ER-PR- = 0.17. Egger’s test for publication bias for all studies: P ER+PR+ = 0.18 and P ER-PR- = 0.18. 31 32 Figure 2.3 Breastfeeding and breast cancer risk by ER/PR status: RR with 95% CI for highest versus lowest category 32 33 Table 2.5 Age at menarche and breast cancer risk defined by ER/PR status RR (95% CI) for oldest vs. youngest age category Study Subgroups by MP status No. of exposure categories Oldest/Youngest age category ER+PR+ ER-PR- Cohort studies Potter, 1995 Post-MP 2 ≥13/<13 0.69 (0.56-0.85) 1.07 (0.67-1.71) Colditz, 2004 All women 2 15/11 0.68 (0.62-0.73) 0.78 (0.68-0.89) Summary RRs for cohort studies 0.68 (0.59-0.79) 0.80 (0.69-0.93) Population-based case-control studies All women 2 ≥12/<12 0.67 (0.50-0.91) 0.91 (0.67-1.43) Pre-MP 2 ≥12/<12 0.67 (0.4-1.0) 1.11 (0.67-1.67) Huang, 2000 Post-MP 2 ≥12/<12 0.63 (0.42-1.0) 0.77 (0.45-1.43) Britton, 2002 Young 2 ≥13/<13 0.77 (0.63-0.94) 0.78 (0.61-1.00) All women ≥14/<12 0.72 (0.52-0.99) 0.99 (0.63-1.53) Pre-MP 4 ≥14/<12 0.49 (0.31-0.76) 1.12 (0.62-2.03) Cotterchio 2 , 2003 Post-MP 4 ≥14/<12 0.84 (0.64-1.10) 0.94 (0.64-1.37) McCredie, 2003 Young ≥13/<13 0.8 (0.6-1.1) 0.6 (0.4-0.9) 33 34 Table 2.5 continued. RR (95% CI) for oldest vs. youngest age category Study Subgroups by MP status No. of exposure categories Oldest/Youngest age category ER+PR+ ER-PR- Ma, under review Young 4 ≥14/<12 0.60 (0.42-0.86) 0.59 (0.38-0.92) Summary RRs for population-based case-control studies 0.73 (0.59-0.89) 0.77 (0.63-0.94) Hospital-based case-controls studies Yoo, 1997 All women Per 2 years ↑ 0.83 (0.68-1.01) 1.11 (0.91-1.36) Rusiecki, 2005 All women 3 ≥14/<12 1.0 (0.53–2.00) 1.43 (0.71-2.5) Summary RRs for hospital-based case-control studies 0.84 (0.64-1.10) 1.14 (0.87-1.49) Pre-MP/Young 0.71 (0.57-0.88) 0.77 (0.62-0.95) Summary RRs by MP status Post-MP 0.74 (0.54-1.01) 0.95 (0.70-1.30) Summary RRs for all studies 0.72 (0.64-0.80) 0.84 (0.75-0.94) P for homogeneity between ER+PR+ and ER-PR- cancer P = 0.006 Note: ER/PR = estrogen and progesterone receptor, MP = menopausal or menopause, ER+ = estrogen receptor positive, ER- = estrogen receptor negative, PR+ = progesterone receptor positive, PR-= progesterone receptor negative, RR = relative risk, 95% CI = 95% confidence interval. Test for homogeneity across all studies: P ER+PR+ = 0.95 and P ER-PR- = 0.23. Egger’s test for publication bias for all studies: P ER+PR+ = 0.42 and P ER-PR- = 0.55. 34 35 Figure 2.4 Age at menarche and breast cancer risk by ER/PR status: RR with 95% CI for oldest versus youngest age category 35 36 2.4.6 Publication Bias We found no evidence of publication bias in results across all studies for the factors we reviewed (Egger’s test: all P > 0.10). 2.5 Discussion This quantitative overview estimates that each birth reduces the risk of ER+PR+ breast cancer by 11%, and that women who were in the oldest age at first birth category were, on average, at 27% higher risk of ER+PR+ cancer than those who were in the youngest age at first birth category after adjustment for parity. Furthermore, we found that neither parity nor age at first birth was associated with reduced risk of ER-PR- cancer. Breastfeeding and late age at menarche decreased the risk of both the subtypes of breast cancer. The protective effect of late age at menarche was statistically significantly stronger for ER+PR+ than ER-PR- cancer. All the studies we summarized have been published except one (Ma et al., under review), which is currently under review. If studies that detected a difference in association by ER/PR status were more likely to be published, our results for parity and age at first birth could be biased. However, we found no evidence for publication bias for either parity or age at first birth results. If survival among cases depends on the two reproductive factors and differs between receptor positive and receptor negative tumors, then this could result in bias among case-control studies. 37 However, cohort studies also observed that the protective effect of parity (Colditz et al., 2004; Potter et al., 1995) and early first birth (Potter et al., 1995) was restricted to ER+PR+ cancer. We therefore think it is unlikely that survival bias explains why parity and age at first birth are associated with ER+PR+ tumors, but not ER-PR- tumors. We also considered whether the different association by receptor status could be caused by residual confounding by age since the ratio of ER+PR+ to ER-PR- cancers increases with age (Chu and Anderson, 2002; Yasui and Potter, 1999). We therefore examined the effect of parity using stratified analyses by age (5 years) using our own data (Ursin et al., 2005). We found that the protective effect from parity was still confined to ER+PR+ cancers (results not shown). We therefore think it is unlikely that the difference in association by receptor status is due to residual confounding by age. The main source of hormone receptor information for studies that we reviewed was medical records. Although we assume that the majority of laboratories have used immunoassays since 1995, we could not exclude the possibility that the assays and cutoffs for determining ER and PR status differed across studies. However, we believe that any such inconsistencies would be unlikely to cause the observed 38 associations, and if anything, that they would bias the relative risk estimates towards the null value. Some data suggest that compared to white, African-American women are more likely to develop ER-PR- cancer (Chu and Anderson, 2002). We were unable to address whether race modifies these associations as only one study provided results by race (Ursin et al., 2005). However, in this study, we found the associations for parity or breastfeeding was similar in white and African-American women. The differences between the comparison and reference categories for age at first birth varied substantially across the studies, ranging from 1 (Britton et al., 2002; Huang et al., 2000; McCredie et al., 2003; Potter et al., 1995; Rusiecki et al., 2005) to 11 years (Ma et al., under review; Ursin et al., 2005) with a 4 year average. One would expect that the effects would be greater for the studies with the greater difference or gap, but this was not the case. We were unable to pursue this further since we did not know the underlying distribution of age at first birth in each specific category from each study. The protective effects of increasing number of births and an early age at first birth against ER+PR+ but not ER-PR- breast cancer suggest that their effects influence 39 risk predominantly through hormonal mechanisms that involve estrogen and progesterone. The effects of these hormones on breast tissue depend upon the amount of both hormones and their specific receptors (Anderson, 2002; Dickson and Stancel, 2000; Evans, 1988; Gorski and Gannon, 1976). A greater number of births and an early first birth may protect against receptor positive breast cancer through several mechanisms: (1) by reducing estrogen and progesterone in plasma (Bernstein et al., 1985; Dorgan et al., 1995; Garcia-Closas et al., 2002); (2) by increasing levels of sex hormone-binding globulin (Bernstein et al., 1985); or (3) by causing further differentiation of the breast epithelium, which may reduce the susceptibility to estrogen and progesterone (Kelsey et al., 1993). Contrary to expectations, breastfeeding and late age at menarche protected against both ER+PR+ and ER-PR- cancers, although menarche had stronger protective effects against ER+PR+ than ER-PR- cancer. This seems to be inconsistent with the hypothesis that these factors act through estrogen and progesterone mediated by their respective receptors (Anderson, 2002; Dickson and Stancel, 2000; Evans, 1988; Gorski and Gannon, 1976). However, evidence shows that when ER+ progenitor cells are exposed to estrogen, they will produce paracrine signals that cause neighbouring populations of ER- cells to proliferate (Dontu et al., 2004). Thus, our findings do not preclude a hormonal mechanism for breastfeeding and late age at 40 menarche, but suggest that the mechanism differs from that involved in parity and age at first birth. 2.6 Conclusions Our quantitative overview shows that parity and early age at first birth only protect against ER+PR+ breast cancer while breastfeeding and late age at menarche protect against both ER+PR+ and ER-PR- breast cancer. Our findings suggest that breastfeeding (and age at menarche) may act through different hormonal mechanisms than parity and age at first birth. 41 Chapter 3 Hormone-related risk factors for breast cancer in women under age 50 years by estrogen and progesterone receptor status - results from a case-control and a case-case comparison 3.1 Summary It has been suggested that hormonal risk factors act predominantly on estrogen and progesterone receptor (ER/PR) positive breast cancers. However the data have been inconsistent, especially in younger women. We evaluated the impact of age at menarche, pregnancy history, duration of breastfeeding, body mass index (BMI), combined oral contraceptive (COC) use, and alcohol consumption on breast cancer risk by ER/PR status in 1725 population-based case patients and 440 control subjects aged 20-49 years identified within neighbourhoods of case patients. We used unconditional logistic regression methods to conduct case-control comparisons overall as well as by ER/PR status of the cases, and to compare ER+PR+ with ER- PR- case patients. Number of full-term pregnancies was inversely associated with risk of ER+PR+ cancer (P trend = 0.005), whereas recent average alcohol consumption was associated with an increased risk of ER+PR+ cancer (P trend = 0.03). Neither of these two factors was associated with risk of ER-PR- cancer. There was some indication that the effect of late age at first full-term pregnancy and higher BMI differed for ER+PR+ versus ER-PR- cancers, although neither case-control 42 comparison was statistically significant. Late age at menarche and a longer duration of breastfeeding were both associated with decreased breast cancer risk, irrespective of receptor status (all P trend ≤ 0.03). Our results suggest that number of full-term pregnancies and recent alcohol consumption impact breast cancer risk in younger women predominantly through progesterone and estrogen mediated by their respective receptors. Late age at menarche and breastfeeding may act through different hormonal mechanisms. 3.2 Introduction It has been well documented that estrogen and progesterone play important roles in breast tumorigenesis (Bernstein and Ross, 1993; Henderson et al., 1982; Key and Pike, 1988) and their effects on the breast are mediated by their respective receptors, the estrogen receptor (ER) and the progesterone receptor (PR) (Anderson, 2002; Dickson and Stancel, 2000; Evans, 1988; Gorski and Gannon, 1976). Furthermore, it has been hypothesized that hormone-related risk factors which reflect estrogen and progesterone exposure may be predominantly associated with breast cancers that express ER and PR, but not with those lacking ER and PR expression (Elwood and Godolphin, 1980; Hildreth et al., 1983; Huang et al., 2000; Li et al., 2003; Potter et al., 1995; Yasui and Potter, 1999; Zhu et al., 2002). A number of epidemiological studies have examined this hypothesis by ER and PR status separately or jointly (Althuis et al., 2004; Habel and Stanford, 1993; Stanford and Greenberg, 1989), and 43 a review from 2004 (Althuis et al., 2004) concluded that early age at menarche, nulliparity, and delayed childbearing were associated with increased risk of ER+PR+ cancer, but not with ER-PR- cancer. However, the findings have been inconsistent. In the prospective data from the Nurses’ Health Study (Colditz et al., 2004), the adverse effect of nulliparity was confined to ER+PR+ cancer, but early age at menarche was associated with increased risk of both ER+PR+ and ER-PR- cancer and the adverse effect of delayed childbearing was limited to ER-PR- instead of ER+PR+ cancer. Results from studies of young women under age 50 years (Britton et al., 2002; McCredie et al., 2003; Ursin et al., 2005) or premenopausal women (Cotterchio et al., 2003; Huang et al., 2000) are even less consistent, with two (Britton et al., 2002; McCredie et al., 2003) of the five studies finding no effect of parity on receptor subtypes. Part of the inconsistency may be because these effects are weaker in younger women or the studies were not large enough to find an effect. To help shed light on the issue, we evaluated hormone-related risk factors for breast cancer by receptor subtypes (ER+PR+, ER-PR- cancer) in a large study of women aged 20-49 years, using both case-control and case-case comparisons. 3.3 Materials and Methods 3.3.1 Case Patients Case patients were identified through the Los Angeles County Cancer Surveillance Program (CSP), the population-based cancer registry that is part of the National 44 Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) cancer registry program. Eligible case patients were U.S.-born English-speaking, white (including Hispanic) or African-American, female residents of Los Angeles County aged 20-49 years when diagnosed with histologically confirmed first primary invasive breast cancer. We identified 2882 eligible case patients (2534 whites and 348 African Americans). White patients were diagnosed between February 1998 and May 2003 and African-American patients were diagnosed between January 2000 and May 2003. We were unable to interview 1088 of the 2882 eligible case patients (38%) due to patient refusal (n = 428), no longer living in Los Angeles County (n = 37), inability to be located (n = 88), death (n = 38), serious illness or disability (n = 18), physician refusal (n = 50), or inability to schedule the interview within 18 months of diagnosis (n = 429). We successfully interviewed 1794 (62%) eligible case patients (1585 whites, 209 African-Americans). 3.3.2 Control Subjects Since this study was originally designed as a case-case study to examine genetic risk factors for breast cancer, we did not collect control subjects for all case patients. The control subjects who were recruited were matched by race and age (within 5-years and aged 20-49 years) to a subset of case patients who were diagnosed between July 2000 and March 2003. Control subjects were U.S.-born English-speaking white or African-American women who had never been diagnosed with invasive or in situ 45 breast cancer. They were identified using a neighbourhood walk algorithm that we have used in previous case-control studies (Bernstein et al., 1994; Ursin et al., 1994). Field staff conducted walks according to a predefined pattern in the neighbourhoods where case patients lived at the time of their diagnoses. The houses on the immediate blocks surrounding the home of the case patients were excluded from the walk. Residences were visited sequentially and information was obtained on potentially eligible women. If no one was home at the time of the visit, we left a request for information at the door, and we sought further information from neighbours so that we could contact the residents later. If we received no response to our written request, we sent additional letters until we were able to determine whether an eligible woman lived at the address in question. Detailed records were maintained to determine the number of housing units contacted in order to identify and interview a control subject. By study end, we had identified 603 eligible control subjects for the 1108 case patients (1018 whites, 90 African-Americans). We were unable to interview 159 of the 603 control subjects (26%) due to subject refusal (n = 77), no longer living in Los Angeles County (n = 18), death (n = 1), serious illness (n = 2), or inability to schedule the interview within 18 months from the date of initial household contact (n = 61). We successfully interviewed 444 (74%) of eligible control subjects (409 whites, 35 African-Americans). On average, 20 houses were canvassed to find an eligible control subject who agreed to be interviewed. 46 3.3.3 Data Collection All participants were interviewed in-person using a structured questionnaire which was a modified version of the questionnaire used for the Women's Contraceptive and Reproductive Experiences (CARE) Study (Marchbanks et al., 2002). Our questionnaire included reproductive history (including breastfeeding), detailed histories of oral contraceptive use and alcohol consumption, family breast cancer history, demographics, and other factors. Information was recorded up to a predetermined reference date for each participant. The reference date was the date of diagnosis for case patients and the date of initial household contact for control subjects. All participants signed informed consent documents prior to interview. The study protocol was approved by the federally approved Institutional Review Board at the University of Southern California Medical Center. Information on ER and PR status for interviewed case patients was obtained from the CSP. Among the 1794 case patients, 1510 (84%) had information on both ER and PR status; 91 of these were reported as weakly or borderline positive (84 cases) or undecided (7 cases) for either ER or PR. Among the remaining 1419 case patients, 881 (62%) were ER+PR+; 92 (6%) were ER+PR-; 41 (3%) were ER-PR+; and 405 (29%) were ER-PR-. 47 3.3.4 Data Analyses We compared demographic and hormone-related risk factors among case patients with known ER and PR information, borderline positive or undecided results, and patients without ER or PR information using F tests for differences in means and Pearson χ 2 tests for differences in frequency distributions. When the two-sided P value comparing all three groups was less than 0.05, we further did pairwise comparisons using Bonferroni t-tests or Pearson χ 2 tests imposing a Bonferroni correction to the p-value, restricting the overall type I error to 5% by setting as statistically significant only two-sided P values < 0.017 for each pairwise comparison (Pagano and Gauvreau, 2000). Analyses were conducted to assess the association between breast cancer and the following factors: age at menarche, number of full-term (greater than 26-week gestation) pregnancies, age at first full-term pregnancy (defined for each woman as the age at which that pregnancy ended), duration of breastfeeding, body mass index (BMI, kg/m 2 ) one year before the participant’s reference date, duration of combination oral contraceptive (COC) use, alcohol drinking status during reference age (never; former; and current, i.e., drinking alcohol during reference age), and the average number of alcoholic drinks per week in the five year period that ended two years before the reference age. One alcoholic drink was defined as 12 oz of beer, 4 oz of wine, or 1.5 oz of liquor. 48 We conducted case-control comparisons for overall, ER+PR+, and ER-PR- case patients with control subjects and also compared ER+PR+ with ER-PR- case patients. We used polytomous logistic regression (Hosmer and Lemeshow, 2000) to simultaneously compare ER+PR+ and ER-PR- case patients with control subjects. We used a multivariable unconditional logistic regression approach (Hosmer and Lemeshow, 2000) for the comparisons of all case patients with controls and ER+PR+ with ER-PR- case patients. We estimated multivariable adjusted odds ratios (ORs) and 95% confidence intervals (CIs). Tests for trend were conducted by fitting ordinal values corresponding to categories of exposure in our models and testing whether the coefficient (slope of the dose response) differed from zero. Adjustment was made for race (white, African-American), age (<30, 30-34, 35-39, 40-44, 45-49 years), and education (<high school, technical school or some college, college graduate) in all our models. Additionally, multivariable models included variables selected a priori as potential confounders: first-degree family history of breast cancer (no first-degree family history, mother or sister with breast cancer, unknown first-degree family history), age at menarche ( ≤11, 12, 13, ≥14 years), gravidity (never, ever pregnant), number of full-term pregnancies (never full-term pregnant, 1, 2, 3, >4 full-term pregnancies), BMI one year prior to reference date (<25, 25-29, 30-34, >35 kg/m 2 ), COC use (never, <1, 1-4, 5-9, ≥10 years), the 49 average number of alcoholic drinks per week in the recent five years (never, <3, 3-5, 6-11, ≥12 drinks, drinking alcohol but not within the five years of interest), and a three-category variable combining menopausal status and hormone therapy use (premenopausal or, among postmenopausal women: never used hormone therapy, had used estrogen therapy or estrogen plus progestin therapy). When estimating the effects of parity or restricting analyses to parous women, we did not include gravidity in our models. A single model was fit to assess the joint effects of age at first full-term pregnancy (<22, 22-27, 28-31, >32 years) and breastfeeding duration (0, <1, 1-6, 7-23, >24 months) among parous women. All variables were included as categorical variables in the models. In reporting the results of trend tests, we considered a two-sided P value less than 0.05 as statistically significant. All analyses were performed using the SAS statistical package (Version 9.0, SAS Institute, Cary, NC). To maintain a constant sample size for all analyses, we excluded 69 case patients and 4 control subjects for the following reasons: missing information on educational attainment (15 cases and 1 control), age at menarche (3 cases), parity (4 cases and 1 control), duration of breastfeeding (6 cases), BMI (4 cases), duration of COC use (17 cases and 1 control), recent alcohol consumption (14 cases); missing information on two or more of these factors (6 cases and 1 control). This resulted in 1725 case patients and 440 control subjects available for the overall case-control analyses. 50 Among the 1725 case patients, 1449 (84%) had information on both ER and PR status; 83 of these were reported as weakly, borderline positive or undecided for either ER or PR and were excluded from the analysis by receptor subtypes. Among the other 1366 case patients, 63% were ER+PR+; 6% were ER+PR-, 3% were ER- PR+; and 28% were ER-PR-. The frequency distribution across receptor subtypes are similar to those reported by prior studies conducted within the SEER registries (Chu and Anderson, 2002; Ursin et al., 2005). We further excluded ER+PR- and ER- PR+ subtypes from analysis by receptor status because they were rare. Therefore, there were 1239 remaining for the analysis by ER and PR status (854 ER+PR+ and 385 ER-PR-). As described above, our control subjects were identified through matching to a subset of the cases. An alternative to conditional logistic regression in matched studies with disproportionate numbers of cases and controls is to break the match and conduct unconditional logistic regression with detailed adjustment for the matching factors (Rothman and Greenland, 1998). We conducted both detailed stratified analyses, with strata defined by race, age (5 years) and education as a proxy of socioeconomic status/neighbourhood as well as standard unconditional logistic regression with adjustment for the same factors. Since the results remained the same, we chose to use unconditional logistic regression for all our analyses. We also repeated all the analyses with the subset of the cases used to identify the controls. 51 Again the results were essentially identical to the overall analyses, and we therefore present results based on all case patients. 3.4 Results Table 3.1 shows the distributions of demographic and hormone-related risk factors among case patients stratified by availability of ER/PR information and among control subjects. Age at breast cancer diagnosis (P F-test = 0.0001), education level (P χ 2 = 0.01), age at first full-term pregnancy (P F-test = 0.002), and percentage of case patients who ever breastfed (P χ 2 = 0.01) differed across the three case groups. Pairwise analyses revealed that the differences were between case patients with known ER/PR status and those without ER/PR information. Compared to case patients without ER/PR information, case patients with known ER/PR status were on average 1.5 years younger at diagnosis (P t-test < 0.0001), better educated (P χ 2 = 0.001), on average 1.8 years older at first full-term pregnancy (P t-test = 0.001), and had a higher percentage of case patients who had breastfed (P χ 2 = 0.004). 3.4.1 Age at Menarche Age at menarche was negatively associated with breast cancer risk regardless of ER/PR status (all P trend ≤ 0.008) (Table 3.2). Compared to women who had 52 Table 3.1 Demographic and hormone-related risk factors for 1725 case patients and 440 control subjects Cases by availability of ER/PR information Variables Cases with known receptor status Cases with borderline positive/undecid ed status Cases without information P value for comparison across the three case groups Controls No. of subjects (%) 1366 (79.2) 83 (4.8) 276 (16.0) 440 Mean age at reference date (SD), years 42.5 (5.4) 42.8 (5.2) 44.0 (4.6) 0.0001*, ‡ 42.6 (4.9) Race White African-American 1206 160 77 6 238 38 0.26† 405 35 Education < High school Technical school or some college College graduate 262 489 615 19 28 36 76 103 97 0.01†, § 60 150 230 First-degree breast cancer family history No Yes Unknown|| 1091 229 46 63 19 1 224 46 6 0.39† 388 41 11 Mean age at menarche (SD), years 12.4 (1.5) 12.6 (1.7) 12.4 (1.4) 0.51* 12.7 (1.5) Ever had a full-term (>26 week) pregnancy, % 66.6 74.7 68.5 0.28† 71.6 Mean number of full-term pregnancies (SD) 2.1 (1.0) 2.1 (1.1) 2.1 (0.9) 0.92* 2.2 (1.1) 52 53 Table 3.1 continued Cases by availability of ER/PR information Variables Cases with complete results Cases with borderline positive/undecid ed status Cases without information P value for comparison across the three case groups Controls Mean age at first full-term pregnancy (SD) 26.7 (6.4) 26.2 (6.2) 24.9 (5.8) 0.002*, ‡ 27.1 (6.2) Ever breastfed, % of parous women 80.4 77.4 70.9 0.01†, § 87.9 Mean duration of breastfeeding (all pregnancies) among those who breastfed (SD), months 14.2 (16.3) 13.7 (13.6) 12.6 (14.6) 0.56* 16.6 (17.3) Mean BMI one year prior to reference date (SD), kg/m 2 25.8 (6.1) 26.2 (6.4) 26.6 (6.5) 0.12* 25.7 (6.4) Ever used COC, % 85.9 84.3 85.9 0.92† 89.3 Mean duration of COC use among users (SD), years 6.9 (6.3) 5.8 (6.6) 6.4 (6.4) 0.23* 6.7 (6.2) Alcohol drinking status during reference age Never Former Current 428 259 679 34 16 33 96 53 127 0.30† 141 90 209 53 54 Table 3.1 continued. Cases by ER/PR information collection status Variables Cases with complete results Cases with borderline positive/undecid ed status Cases without information P value for comparison across the three case groups Controls Alcohol drinking status in recent 5 year period No Yes Yes, but not in the five year period 428 759 179 34 40 9 96 137 43 0.17† 141 228 71 Mean number of alcoholic drinks per week in recent 5 years for drinkers (SD) 6.6 (20.1) 4.9 (6.1) 6.0 (8.4) 0.80* 4.6 (9.1) Note: ER/PR = estrogen and progesterone receptor, SD = standard deviation, BMI = body mass index, COC = combined oral contraceptive *F-test. † Pearson χ 2 test. ‡Bonferroni t-tests for pairwise comparisons: P values for cases with complete ER/PR information versus cases without ER/PR information ≤ 0.001, and P values for the other comparisons ≥ 0.20. §Pearson χ 2 tests for pairwise comparisons: P values for cases with complete ER/PR information versus cases without ER/PR information ≤ 0.004, and P values for the other comparisons ≥ 0.32. ||Unknown category was excluded from Pearson χ 2 tests. 54 55 menarche before age 12 years, later age at menarche ( ≥14 years) was associated with an approximately 40% reduced risk of breast cancer among the ER+PR+ case patients, ER-PR- case patients, and all case patients combined. The associations with age at menarche did not differ between ER+PR+ and ER-PR- case patients (P trend = 0.85). 3.4.2 Parity A protective effect of parity was confined to women with ER+PR+ cancer (Table 3.2). The OR of ER+PR+ cancer decreased with increasing number of full-term pregnancies (P trend = 0.005). Parous women who had 4 or more full-term pregnancies had about a 50% reduction in the risk of ER+PR+ cancer compared to women who never had a full-term pregnancy. Comparing categories representing numbers of full-term pregnancies, ER+PR+ case patients were less likely to have had many full-term pregnancies than ER-PR- case patients (P trend = 0.09). 3.4.3 Age at first Full-term Pregnancy A slight increase in risk for ER+PR+ cancer and a reduced risk for ER-PR- cancer was observed with increasing age at first full-term pregnancy, but none of the confidence limits for the risk estimates excluded 1.0 and no linear trend in risk was observed for either cancer type (ER+PR+: P trend = 0.49, ER-PR-: P trend = 0.08) (Table 56 3.2). ER+PR+ case patients were more likely to have had a late first full-term pregnancy than ER-PR- case patients (P trend = 0.009). 3.4.4 Breastfeeding Duration of breastfeeding was negatively associated with breast cancer risk regardless of ER/PR status (all P trend ≤ 0.03) (Table 3.2). In the case-case analysis duration of breastfeeding was not associated with ER/PR status (P trend = 0.63). 3.4.5 Body Mass Index (BMI) one year prior to reference date Increasing level of BMI was associated with a non-statistically significant decreasing risk of ER+PR+ cancer (P trend = 0.20), but was not associated with ER-PR- cancer (Table 3.2). Moreover, among premenopausal women, increasing BMI was marginally statistically significantly associated with a decreasing risk of ER+PR+ (P trend = 0.08), but not ER-PR- cancer (P trend = 0.54). Compared to premenopausal women who had low BMI (<25 kg/m 2 ), the OR among premenopausal women in the highest BMI category ( ≥35 kg/m 2 ) was 0.58 (95% CI = 0.34-1.00) for ER+PR+ cancer and 1.07 (95% CI = 0.58-1.97) for ER-PR- cancer. Examining the trend in BMI across increasing categories, ER+PR+ case patients were less likely to have had higher BMI than ER-PR- case patients (P trend = 0.005). 57 3.4.6 Combined Oral Contraceptive (COC) Use COC use was not associated with either risk of ER+PR+ cancer or ER-PR- cancer (Table 3.2). Women who had used COC for 10 years or longer had a slightly higher OR of ER-PR- cancer (OR = 1.27, 95% CI = 0.75-2.14) but a lower OR of ER+PR+ cancer (OR = 0.76, 95% CI = 0.49-1.18) compared to never users. Comparing categories of increasing COC use, ER+PR+ case patients were less likely to have had longer duration of COC use than ER-PR- case patients (P trend = 0.008). 3.4.7 Alcohol Consumption Alcohol drinking status during the reference year was not associated with breast cancer risk (Table 3.2). However, the average number of alcoholic drinks per week in the recent 5 year time period ending two years prior to reference date was positively associated with ER+PR+ cancer (P trend = 0.03), weakly associated with all types of cancer together (P trend = 0.12), and not associated with ER-PR- cancer (P trend = 0.42). Overall, ER+PR+ case patients appeared more likely to have drunk larger quantities of alcohol than ER-PR- case patients, but the difference was not statistically significant (P trend = 0.23). 3.5 Discussion Overall in this study of women under age 50 years, we found associations that differed by ER/PR status for number of full-term pregnancies, recent alcohol 58 Table 3.2 Adjusted* ORs and 95% CIs of breast cancer associated with hormone-related risk factors by ER/PR status All cases vs. controls ER+PR+ cases vs. controls ER-PR- cases vs. controls Variables No. controls No. cases OR (95% CI) No. cases OR (95% CI) No. cases OR (95% CI) ER+PR+ cases vs. ER-PR- cases OR (95% CI) Age at menarche, years ≤11 12 13 ≥14 90 109 121 120 400 532 475 318 1.00 1.13 (0.83-1.55) 0.88 (0.65-1.21) 0.61 (0.44-0.85) 188 278 231 157 1.00 1.24 (0.88-1.75) 0.87 (0.62-1.24) 0.60 (0.42-0.86) 92 119 110 64 1.00 1.18 (0.79-1.77) 0.94 (0.62-1.41) 0.59 (0.38-0.92) 1.00 1.04 (0.74-1.47) 0.93 (0.65-1.33) 1.02 (0.67-1.53) Trend P-value 0.0007 0.0006 0.008 0.85 Number of full-term pregnancies† None 1 2 3 ≥4 125 85 139 56 35 564 347 492 224 98 1.00 0.91 (0.66-1.25) 0.81 (0.61-1.07) 0.81 (0.56-1.18) 0.49 (0.31-0.78) 301 157 251 100 45 1.00 0.77 (0.55-1.10) 0.77 (0.57-1.06) 0.69 (0.45-1.04) 0.47 (0.28-0.80) 113 82 106 54 30 1.00 1.06 (0.70-1.62) 0.95 (0.64-1.39) 1.06 (0.65-1.73) 0.72 (0.39-1.33) 1.00 0.70 (0.49-1.02) 0.84 (0.60-1.18) 0.68 (0.44-1.05) 0.65 (0.37-1.15) Trend P-value 0.008 0.005 0.50 0.09 Age at first full-term pregnancy for parous women, years†, ‡ <22 22-27 28-31 ≥32 68 89 78 80 316 333 265 247 1.00 1.07 (0.72-1.58) 1.03 (0.66-1.60) 0.92 (0.57-1.48) 126 149 134 144 1.00 1.14 (0.73-1.78) 1.17 (0.71-1.93) 1.23 (0.72-2.10) 96 81 57 38 1.00 0.92 (0.56-1.50) 0.82 (0.46-1.45) 0.56 (0.30-1.07) 1.00 1.25 (0.81-1.94) 1.42 (0.85-2.37) 2.22 (1.24-3.98) Trend p-value 0.65 0.49 0.08 0.009 58 59 Table 3.2 continued. All cases vs. controls ER+PR+ cases vs. controls ER-PR- cases vs. controls Variables No. controls No. cases OR (95% CI) No. cases OR (95% CI) No. cases OR (95% CI) ER+PR+ cases vs. ER-PR- cases OR (95% CI) Duration of breastfeeding for parous women, months†, ‡ 0 <1 1-6 7-23 ≥24 38 23 80 110 64 247 144 273 322 175 1.00 0.99 (0.56-1.77) 0.58 (0.37-0.91) 0.52 (0.33-0.82) 0.51 (0.30-0.86) 102 65 131 165 90 1.00 1.01 (0.53-1.90) 0.57 (0.34-0.94) 0.52 (0.31-0.87) 0.49 (0.27-0.87) 63 38 68 63 40 1.00 1.19 (0.59-2.39) 0.72 (0.41-1.27) 0.55 (0.31-0.98) 0.62 (0.32-1.21) 1.00 0.82 (0.47-1.44) 0.77 (0.47-1.26) 0.97 (0.59-1.61) 0.77 (0.42-1.39) Trend P-value 0.001 0.002 0.03 0.63 BMI one year prior to reference date, kg/m 2 <25 25-29 30-34 ≥35 257 95 51 37 939 425 221 140 1.00 1.18 (0.89-1.55) 1.01 (0.70-1.44) 0.88 (0.58-1.34) 495 209 94 56 1.00 1.11 (0.82-1.50) 0.88 (0.59-1.30) 0.69 (0.43-1.11) 183 101 61 40 1.00 1.41 (0.99-2.02) 1.43 (0.91-2.23) 1.18 (0.70-2.01) 1.00 0.81 (0.59-1.10) 0.60 (0.40-0.89) 0.61 (0.38-0.99) Trend P-value 0.82 0.20 0.16 0.005 Duration of COC use, years Never <1 1-4 5-9 ≥10 47 73 108 115 97 244 295 429 362 395 1.00 0.78 (0.51-1.18) 0.80 (0.54-1.19) 0.62 (0.42-0.93) 0.84 (0.56-1.26) 128 133 227 173 193 1.00 0.70 (0.44-1.12) 0.81 (0.53-1.25) 0.57 (0.37-0.88) 0.76 (0.49-1.18) 47 59 90 90 99 1.00 0.76 (0.43-1.33) 0.87 (0.52-1.46) 0.82 (0.49-1.37) 1.27 (0.75-2.14) 1.00 0.93 (0.57-1.52) 0.92 (0.59-1.43) 0.71 (0.45-1.12) 0.61 (0.39-0.96) Trend P-value 0.30 0.16 0.18 0.008 59 60 Table 3.2 continued. All cases vs. controls ER+PR+ cases vs. controls ER-PR- cases vs. controls Variables No. controls No. cases OR (95% CI) No. cases OR (95% CI) No. cases OR (95% CI) ER+PR+ cases vs. ER-PR- cases OR (95% CI) Alcohol drinking status during reference age § Never Former Current 141 90 209 558 328 839 1.00 0.96 (0.71-1.31) 1.05 (0.81-1.36) 251 162 441 1.00 1.02 (0.72-1.44) 1.17 (0.88-1.55) 138 74 173 1.00 0.88 (0.59-1.33) 0.93 (0.67-1.30) 1.00 1.14 (0.79-1.64) 1.25 (0.93-1.68) Average number of alcoholic drinks per week in recent 5 years Never <3 3-5 6-11 ≥12 141 130 49 32 17 558 481 171 157 127 1.00 1.01 (0.76-1.35) 0.93 (0.63-1.37) 1.16 (0.75-1.81) 1.77 (1.01-3.08) 251 248 90 82 68 1.00 1.11 (0.81-1.53) 1.01 (0.66-1.54) 1.26 (0.78-2.03) 2.10 (1.17-3.79) 138 101 34 34 27 1.00 0.89 (0.61-1.30) 0.76 (0.45-1.28) 1.06 (0.60-1.86) 1.71 (0.87-3.38) 1.00 1.22 (0.87-1.71) 1.31 (0.81-2.11) 1.20 (0.74-1.94) 1.24 (0.73-2.09) Trend P-value 0.12 0.03 0.42 0.23 Drinking alcohol but not in the recent five years 71 231 0.88 (0.63-1.22) 115 0.92 (0.63-1.34) 51 0.80 (0.51-1.26) 1.14 (0.76-1.73) Note: OR = odds ratio, CI = confidence interval, ER/PR = estrogen and progesterone receptor, BMI = body mass index, COC = combined oral contraceptive *Models included race, age, education, first-degree breast cancer family history, age at menarche, gravidity, number of full-term pregnancies, BMI one year prior to reference date, COC use, the average number of alcoholic drinks per week in the recent five years, a variable combining menopausal status and hormone therapy use except where otherwise noted. †Models did not include gravidity. ‡Age at first full-term pregnancy and duration of breastfeeding mutually adjusted for each other. §Alcohol drinking status during reference age replaced the number of alcoholic drinks per week in the recent five years in the models. 60 61 consumption, and possibly age at first full-term pregnancy and BMI. Risk reductions associated with late age at menarche and long duration of breastfeeding did not differ by ER/PR status. COC use was not associated with breast cancer risk in this study. The risk of ER+PR+ cancer decreased with increasing number of full-term pregnancies while the risk of ER-PR- cancer was not associated with number of full- term pregnancies. Our results are consistent with two previous studies of young (Ursin et al., 2005) or premenopausal women (Cotterchio et al., 2003), but not with two studies of women under 45 years of age that did not find any associations with receptor subtype (Britton et al., 2002; McCredie et al., 2003). When we restricted our analysis to women under 45 years, results were essentially identical to our overall findings (results not shown). The differing results between our study and the two previous studies may be due to the larger sample size in our study. A greater number of full-term pregnancies may protect against receptor positive breast cancer by causing the full differentiation of breast epithelium (Russo et al., 1992), thereby reducing cyclical morphological change in breast tissue during the menstrual cycles (Russo et al., 2000). We observed that, relative to women whose first pregnancy occurred before age 22 years, a first full-term pregnancy at age 32 years or later was associated with a 62 modestly elevated risk for ER+PR+ cancer and a reduced risk for ER-PR- cancer, but confidence intervals for both OR estimates included 1.0. Of four previous studies that examined the effect of late age at first full-term pregnancy among young (Britton et al., 2002; McCredie et al., 2003) or premenopausal (Cotterchio et al., 2003; Huang et al., 2000) women, three did not observe a statistically significant association with any receptor subtype of breast cancer (Britton et al., 2002; Cotterchio et al., 2003; Huang et al., 2000); the fourth study found that late age at first full-term pregnancy increased the risk of ER+PR+ cancer among women under 40 years, but was not associated with ER-PR- cancer (McCredie et al., 2003). Our results are consistent with the bulk of the evidence showing that late age at first full- term pregnancy does not appear to be a strong risk factor in young women, but if it increases the risk of any subtype of breast cancer, our results suggest that it would be the receptor positive cancer. We found that higher BMI was inversely associated, although marginally, with ER+PR+ cancer, but had no effect on ER-PR- cancer. Among the five previous studies that examined the effect of BMI according to ERPR status among young (Britton et al., 2002; McCredie et al., 2003) or premenopausal (Cotterchio et al., 2003; Enger et al., 2000; Huang et al., 2000) women, two found that higher BMI only protected against ER+PR+ cancer (Britton et al., 2002; Cotterchio et al., 2003), while the other three studies found no association with any receptor subtype (Enger 63 et al., 2000; Huang et al., 2000; McCredie et al., 2003). A possible mechanism is that high BMI results in anovulatory menstrual cycles with reduced exposure to ovarian hormones (Pike, 1990), especially in the latter half of the cycle (Westhoff et al., 1996). We also found that average weekly alcohol consumption in the five year period ending two years before the reference date was positively associated with the risk of ER+PR+ cancer, but not ER-PR- cancer. Among five previous epidemiological studies that have examined the effect of alcohol consumption according to ER/PR status among young women (Britton et al., 2002; McDonald et al., 2004) or premenopausal women (Cotterchio et al., 2003; Enger et al., 1999; Huang et al., 2000), two found a non-statistically significant increase in risk of ER+PR+ cancer (Cotterchio et al., 2003; Enger et al., 1999), one found a increase in risk of ER+PR- subtypes (McDonald et al., 2004), and another found a non-statistically significant increase in risk of both receptor subtypes (Britton et al., 2002). Experimental (Mendelson et al., 1988) and cross-sectional (Reichman et al., 1993) data have shown that alcohol consumption may result in an increase of blood estrogen levels among premenopausal women. Our results are consistent with the hypothesis that the effect of alcohol on the premenopausal breast is via estrogen. 64 Late age at menarche and duration of breastfeeding were negatively associated with the breast cancer risk regardless of ER/PR status. Our findings for age at menarche are consistent with two (Britton et al., 2002; McCredie et al., 2003) of four previous case-control studies conducted among young women under 45 years (Britton et al., 2002; McCredie et al., 2003) or premenopausal women (Cotterchio et al., 2003; Huang et al., 2000). The protective effects of breastfeeding on both ER+PR+ and ER-PR- cancer were also observed in previous studies among young (Britton et al., 2002; Ursin et al., 2005) or premenopausal (Huang et al., 2000) parous women, although the associations were not statistically significant in two of these studies (Britton et al., 2002; Huang et al., 2000). COC use was not associated with breast cancer risk in this study. Our results are consistent with three (Cotterchio et al., 2003; Huang et al., 2000; McCredie et al., 2003) of four previous studies by ER/PR status among young (Britton et al., 2002; McCredie et al., 2003) or premenopausal women (Cotterchio et al., 2003; Huang et al., 2000); the only inconsistent study found a marginally statistical significant increased risk of ER-PR- cancer among women who had ever used oral contraceptives (Britton et al., 2002). If certain hormone-related factors predominantly act through estrogen and progesterone mediated by their respective receptors, then it has been argued (Huang 65 et al., 2000; Potter et al., 1995) that these hormone-related factors will be associated with hormone receptor positive, but not with receptor negative cancer. Our findings that increasing number of full-term pregnancies is associated with lower risk, while increasing recent average weekly alcohol consumption is associated with greater risk of ER+PR+ cancer, and that ER-PR- cancer are not impacted by these risk factors, support the hypothesis that these factors predominantly act through this type of hormonal mechanism. On the other hand, it could be hypothesized that hormonal factors should impact receptor positive and receptor negative cancer similarly. It has been hypothesized that an ER- stem cell population gives rise to ER+ progenitor cells (Dontu et al., 2004), which will proliferate when exposed to estrogen, but can also send paracrine signals that will cause neighbouring populations of ER- cells to proliferate. Thus, late age at menarche and breastfeeding may still act through hormonal mechanisms that involve ER and PR, but our results suggest that the exact mechanism differs from that involved in the parity and alcohol consumption. One strength of our study is the large number of case patients included. Our analysis included more young breast cancer patients than six of the eight previously published studies with results for ER/PR status among women under age 50 years (Britton et al., 2002; McCredie et al., 2003; McDonald et al., 2004; Ursin et al., 2005) or among 66 premenopausal women (Cotterchio et al., 2003; Enger et al., 2000; Enger et al., 1999; Huang et al., 2000). For the only two studies with a larger sample size, one exclusively focused on alcohol consumption (McDonald et al., 2004), the other one focused on reproductive factors including parity, age at first full-term pregnancy, and breastfeeding (Ursin et al., 2005). Several limitations of this study must also be considered. The number of control subjects in this study was relatively small. This could explain why we detected similar magnitude of effect for late first full-term pregnancy on ER+PR+ cancer as did Women’s CARE Study (Ursin et al., 2005), but in our study it did not reach statistically significance because of our limited statistical power. This also could explain why our case-case analyses suggested that the effect of late age at first full- term pregnancy and BMI significantly differed between ER+PR+ and the ER-PR- case patients while we did not detect any associations when comparing each of the subtypes to the control subjects. Because of our decision not to retain the case-control match during data analyses so that we could maximize the number of cases included in the analyses, we used unconditional instead of conditional logistic regression approach. This could have biased our relative risk estimates towards the null value per Rothman and Greenland (Rothman and Greenland, 1998). However, compared to the data from the Women’s 67 CARE Study (Ursin et al., 2005), which, thus far, has been the largest population- based case-control study of women aged 35-64 years, our results were consistent for parity and breastfeeding, while the positive association between late first full-term pregnancy and ER+PR+ cancer was statistically significant in the Women’s CARE Study, but not in our study. Compared to the prospective data from the Nurses’ Health Study (Colditz et al., 2004) for both pre- and post menopausal women, our results were consistent for age at menarche and parity, but inconsistent for late first full-term pregnancy. Furthermore, our overall findings for all the hormone-related factors we examined are quite similar to those in the literature for young or premenopausal women. We have no data on the methods and cutoff points for receptor status used by each laboratory as we obtained this information from the CSP which bases its classifications on information in pathology reports from a variety of laboratories. Although we assume that the majority of laboratories used immunohistochemistry assays and consistent cutoff points, it is possible that some laboratories used different methods, or different cutoff points. However, we believe that any such inconsistencies would be unlikely to cause the observed associations, and if anything, that they would bias our relative risk estimates towards the null value. Another limitation is that in our analyses by receptor subtypes, we excluded 21% of our cases patients, since 16% of patients had no ER or PR status information and 5% 68 were borderline positives or undecided for either ER or PR. The percentage of case patients without information from the cancer registry in this study (16%) is similar to that reported by prior studies conducted within the SEER registries (Chu and Anderson, 2002; Ursin et al., 2005). The reason some case patients do not have known ER/PR status is unclear. We observed that in this study, the case patients with known ER/PR status information were somewhat younger, were better educated, gave birth later, and were more likely to breastfeed than those whose ER/PR status had not been determined. These differences between case patients with known ER/PR status and those without ER/PR information from CSP could have biased our case-control comparison by receptor subtypes and caused us to find an effect of late age at first full-term pregnancy even if none existed, but would tend to underestimate any protective effect of breastfeeding on breast cancer risk. It is unlikely that this bias would be different for ER+PR+ and ER-PR- cancer. Furthermore, this would not have influenced our case-control analysis using all case patients combined. Since our results from the case-case and the case-control analysis by receptor subtypes or using all case patients combined were generally consistent, we think it is unlikely that these issues caused important bias in this study. 3.6 Conclusions Our results suggest that number of full-term pregnancies and recent alcohol consumption, and possibly age at first full-term pregnancy and BMI, impact breast 69 cancer risk in young women predominantly through progesterone and estrogen mediated by their respective receptors. Our results further suggest that late age at menarche and breastfeeding may protect against breast cancer through a different hormonal mechanism. 70 Chapter 4 Low-dose medical radiation exposure and breast cancer risk in women under age 50 years overall and by estrogen and progesterone receptor status - results from a case-control and a case-case comparison 4.1 Summary Although moderate to high-dose ionizing radiation exposure is an established risk factor for breast cancer, the effect of low-dose exposure is unclear. We evaluated the effect of low-dose radiation from medical procedures on risk of breast cancer overall and by subtypes defined by estrogen and progesterone receptor (ER/PR) status in 1742 population-based case patients aged 20-49 years and 441 control subjects identified from neighbourhoods of case patients in Los Angeles County. We found an elevated breast cancer risk among women who reported having had multiple chest X-rays (P trend =0.0007), 7 or more mammograms (OR=1.80, 95% CI=0.95-3.42) 5 or more years before diagnosis/reference date, or dental X-rays without lead apron protection before age 20 years (OR=1.81, 95% CI=1.13-2.90). Women who had their first exposure to these medical radiation procedures at an early age had a greater increased risk compared to those who were first exposed at a later age; nulliparous women had a greater increased risk associated with these radiation exposures compared to parous women; and women with a family history of breast or ovarian cancer had a greater risk associated with multiple chest X-rays or mammograms than 71 those without a family history; but none of the effect modifications reached statistical significance. Some results suggested that radiation had a greater impact on ER-PR- than ER+PR+ breast cancer, but these differences were not fully consistent. In conclusion, our results support that low-dose ionizing radiation is associated with the increased risk of breast cancer. Age at first exposure, parity, and breast or ovarian cancer family history appeared to modify the radiation effect, but we found no consistent effect modification by ER/PR status. 4.2 Introduction Moderate to high-dose (above ~200 mSv) ionizing radiation exposure is a well established risk factor for breast cancer (Ronckers et al., 2005; Wakeford, 2004). However, the effect of low-dose ionizing radiation exposure is unclear. Previous results on this topic have been inconsistent. One population-based case-control study found no statistically significant association between low-dose medical radiation exposure and breast cancer among women aged 30-80 years (Zheng et al., 2002). However, a population-based case-control study of women aged 40 years or younger found that low-dose medical radiation exposure before age 20 years was associated with an elevated breast cancer risk, particularly among nulliparous women or those with a breast or ovarian cancer family history (Hill et al., 2002). Another case- control study suggested that the increased risk among young women may be 72 confined to estrogen and progesterone receptor negative (ER-PR-) cancer (Huang et al., 2000). Radiation exposure from commonly performed medical procedures is low. Based on the literature, the mean radiation dose to breast tissue of a chest X-ray is 0.2 mSv (John and Kelsey, 1993) while the radiation dose from a mammogram is on average 3 mSv (Brenner et al., 2002). However, some women may have had a large number of procedures and therefore possibly a substantial cumulative dose. In addition, women may be first exposed to radiation from medical procedures such as chest X- ray and dental X-ray at a very young age. Results from studies of moderate to high- dose radiation exposures have found that age at exposure is an important effect modifier (Land et al., 2003; Preston et al., 2002; Tokunaga et al., 1994). Women who were under age 10 years when first exposed to moderate to high-dose radiation are at higher risk of breast cancer than those first exposed at an older age (Land et al., 2003; Tokunaga et al., 1994). However, whether this is the case for low-dose medical exposure is unknown. Clarifying this issue will enhance our understanding of radiation carcinogenesis and clarify any potential risks of low-dose radiation. Therefore, we evaluated the effect of low-dose radiation from medical procedures on risk of breast cancer overall and by subtypes defined by ER/PR status in women under age 50. 73 4.3 Materials and Methods 4.3.1 Case Patients The details of this study have previously been described (Ma et al., under review). In brief, we identified case patients through the Los Angeles County Cancer Surveillance Program (CSP), the population-based cancer registry that is part of the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) registry program. Eligible case patients were U.S.-born English-speaking, white (including Hispanic) or African-American, female residents of Los Angeles County aged 20-49 years when diagnosed with histologically confirmed first primary invasive breast cancer. We identified a total of 2882 eligible case patients; 2534 were white women diagnosed between February 1998 and May 2003 and 348 were African-American women diagnosed between January 2000 and May 2003. We completed interviews with 1794 (62%) eligible case patients (1585 whites, 209 African-Americans). 4.3.2 Control Subjects Since this study was originally designed as a case-case study to examine genetic risk factors for breast cancer, we did not collect control subjects for all case patients. The control subjects recruited were matched by race and age (within 5-years and aged 20- 49 years) to a subset of case patients who were diagnosed between July 2000 and March 2003. Control subjects were U.S.-born English-speaking white or African- 74 American women who had never been diagnosed with invasive or in situ breast cancer. They were identified by using a neighbourhood walk algorithm that we have used in previous case-control studies (Bernstein et al., 1994). Field staff conducted walks according to a predefined pattern in the neighbourhoods where case patients lived at the time of their diagnoses. By study end, we had identified 603 eligible control subjects for the 1108 case patients (1018 whites, 90 African-Americans). We completed interviews with 444 (74%) eligible control subjects (409 whites, 35 African-Americans). On average, 20 houses were canvassed to find an eligible control subject who agreed to be interviewed. 4.3.3 Data Collection All participants were interviewed in-person using a structured questionnaire which was a modified version of the questionnaire used for the Women's Contraceptive and Reproductive Experiences (CARE) Study (Marchbanks et al., 2002). We expanded questions to collect detailed information on medical radiation exposures prior to a predetermined reference date. This reference date was the date of diagnosis for case patients and the date of initial household contact for control subjects. We collected the information regarding medical X-ray exposures including diagnostic X-rays of chest, mammograms, dental X-rays, spinal X-rays, upper gastrointestinal (GI) exams, computerized tomography (CT) exams above the waist, X-rays of the heart or arteries, chest fluoroscopy, and radiation therapy. We also collected information on 75 age at first exposure, frequency of exposure, and for dental X-rays, we additionally obtained information on the use of a lead apron (“always” and “not always”). Because the induction period for breast cancer associated with radiation exposure is known to be long (Darby et al., 1987; Land et al., 2003; Preston et al., 2002), we only included exposure information up to 5 years prior to the diagnosis/reference date. We also used 10 years as a cut-off point; however, our results were almost identical to those obtained using the 5 year cut off. To maximize our power we present the results obtained using the 5 year cut off. In addition to the radiation exposure data, we recorded reproductive history, alcohol consumption history, family breast and ovarian cancer history, history of benign breast disease, demographics, and information on other factors up to the reference date for each participant. All participants provided informed consent prior to interview. The study protocol was approved by the federally approved Institutional Review Board at the University of Southern California. Information on ER and PR status was obtained from the CSP. This information is routinely collected for newly diagnosed breast cancer patients by all SEER registries. Among the 1794 case patients, 1510 (84%) had information on both ER and PR status. Of these, 1419 had known ER and PR status (62% ER+PR+; 6% ER+PR-; 76 3% ER-PR+; 27% ER-PR-) while 91 were reported as weakly or borderline positive (84 cases) or undecided (7 cases) for either ER or PR. 4.3.4 Data Analyses We used Pearson χ 2 tests to evaluate case-control differences in the frequency distributions of categorical variables and t tests to evaluate case-control differences in continuous variables. Using multivariable approaches, we examined the effects of all medical procedures when 2% or more participants reported having ever had the exposure 5 or more years before diagnosis/reference date. We conducted case-control comparisons for breast cancer overall, and then for ER-PR-, and ER+PR+ subtypes; we also compared ER- PR- case patients with ER+PR+ case patients. We used a multivariable polytomous logistic regression approach (Hosmer and Lemeshow, 2000) to simultaneously compare ER+PR+ and ER-PR- case patients with control subjects. We used a multivariable unconditional logistic regression approach (Hosmer and Lemeshow, 2000) to compare all case patients with control subjects and ER-PR- with ER+PR+ case patients, and estimated odds ratios (ORs) and 95% confidence intervals (CIs). As described above, our control subjects were matched to a subset of the cases. An alternative to conditional logistic regression in matched studies with disproportionate numbers of cases and controls is to break the match and conduct unconditional 77 logistic regression with detailed adjustment or stratification by the matching factors (Rothman and Greenland, 1998). We conducted both stratified analyses, with strata defined by race, 5 year age groups and education as a proxy of socioeconomic status/neighbourhood as well as standard unconditional logistic regression with adjustment for the same factors. Since the results did not differ using these two approaches, we chose to present unconditional logistic regression results. Tests for trend were conducted by fitting ordinal values corresponding to categories of exposure in our models and testing whether the coefficient (slope of the dose response) differed from zero. Multivariable models included variables selected a priori as potential confounders: race (white, African-American), age (<30, 30-34, 35-39, 40-44, 45-49 years), education (<high school, technical school or some college, college graduate), family history of breast or ovarian cancer [no, breast or ovarian cancer in a mother or a sister (first degree), breast or ovarian cancer in an aunt or grandmother (second degree), unknown first-degree family history], age at menarche (<13, ≥13 years), parity (never/ever had a full-term (>26 week) pregnancy), age at first full-term pregnancy (<30, ≥30 years), BMI one year prior to reference date (continuous, kg/m 2 ), a three-category variable combining menopausal status and hormone therapy use (premenopausal or, among postmenopausal women: never used hormone therapy, used estrogen or estrogen plus progestin therapy), the average number of 78 alcoholic drinks per week in the five year period ending two years prior to the reference age (never, <3, 3-5, ≥6 drinks, drinking alcohol but not within the five years of interest; one alcoholic drink was defined as 12 oz of beer, 4 oz of wine, or 1.5 oz of liquor), and history of benign breast diseases (never, ever). We considered women who did not have any first-degree family history, but had no information regarding breast or ovarian cancer among second-degree relatives as having no family history of breast or ovarian cancer. Since mutually adjusting for chest X-ray, mammogram, and dental X-ray history in our multivariable models altered the OR estimates less than 10%, we concluded that these exposures were independent exposures (little evidence for confounding), and we modeled each exposure in a separate model. For all medical procedures that were associated with breast cancer risk in our data, we further examined the potential effect modification by age at first exposure, parity, and family history. To test potential for effect modification by age at first exposure, we used z test to compare Log ORs of ever exposure or each corresponding level of medical radiation exposures (when there were more than two categories involved) between early and late first exposure categories. We only report the P-value comparing highest to the lowest category of radiation exposure when none of the odds ratios differ statistically. For the potential effect modification by parity, or breast or ovarian cancer family history we estimated the coefficient associated with 79 an interaction term using dichotomous or ordinal variables of medical radiation exposure and assessed whether this coefficient differed statistically from zero using the Wald test. To maintain a constant sample size for all analyses, we excluded 52 case patients and 3 control subjects for the following reasons: missing information on educational attainment (15 cases and 1 control), age at menarche (3 cases), parity (5 cases and 1 control), recent alcohol consumption (14 cases), BMI (4 cases), history of benign breast disease (6 cases), or menopausal status (1 case) or missing information on two or more of these factors (4 cases and 1 control). This resulted in 1742 case patients and 441 control subjects available for the overall case-control analyses. Among the 1742 case patients, 1462 (84%) had information on both ER and PR status; 83 of these were reported as weakly, borderline positive or undecided for either ER or PR and were excluded from the analyses by receptor subtypes. Among the other 1379 case patients, 62% were ER+PR+; 6% were ER+PR-, 3% were ER-PR+; and 29% were ER-PR-. The frequency distribution across receptor subtypes was similar to those reported by prior studies conducted within the SEER registries (Chu and Anderson, 2002; Ursin et al., 2005). We further excluded ER+PR- and ER-PR+ subtypes from the analyses by receptor status because they were rare. Therefore, there were 1251 remaining for the analyses by ER and PR status (857 ER+PR+ and 394 ER-PR-). 80 All statistical significance levels (P values) reported are two-sided. All analyses were performed using the SAS statistical package (Version 9.0, SAS Institute, Cary, NC). 4.4 Results Case subjects were more likely to have a first-degree family history of breast cancer (P < 0.0001), an early menarche (P = 0.0002), and a history of benign breast diseases (P < 0.0001) (Table 4.1). They were also less likely to be parous (P = 0.09) than control subjects. The Case subjects were more likely than control subjects to report having ever had a chest X-ray (P = 0.002) and having dental X-rays without the protection of a lead apron both before age 20 years (P = 0.006) as well as after age 20 years (P = 0.001). Table 4.2 shows the adjusted ORs of breast cancer associated with medical procedures where at least 2% of the participants reported being exposed. Compared to women who never had a chest X-ray, women who had received 9 or more chest X- rays were at two-fold increased risk of breast cancer (OR=2.08, 95% CI = 1.20– 3.60, P trend = 0.0007). Women who had 7 or more mammograms had 80% greater odds of breast cancer than women who had not had a mammogram (1.80, 95% CI = 0.95-3.42). Women who had dental X-rays without lead apron protection before age 81 Table 4.1 The characteristics of demography and medical radiation exposure Variables Controls All cases P -values* No. of subjects 441 1742 Mean age (SD), yrs † 42.6 (4.8) 42.8 (5.3) 0.67 Mean BMI (SD), kg/m 2† 25.7 (6.4) 25.9 (6.2) 0.39 First-degree breast cancer family history 9.3% 17.1% <0.0001 Mean age at menarche (SD), yrs † 12.7 (1.5) 12.4 (1.5) 0.0002 Parous 71.7% 67.5% 0.09 Menopause and HRT use Premenopausal Postmenopausal and Never ET/EPT Ever ET/EPT 78.5% 9.5% 12.0% 80.8% 9.6% 9.6% 0.32 Ever alcohol drinking status in recent 5 year period,% 51.9% 54.2% 0.39 History of benign breast diseases 33.3% 44.2% <0.0001 Ever had the following radiation exposure 5 or more years prior to reference date Chest X-ray 53.3% 61.4% 0.002 Mammogram 49.0% 51.3% 0.38 81 82 Table 4.1 continued. Variables Controls All cases P -values* Dental X-ray Dental X-ray, before age 20 years 88.4% 88.1% 0.85 Did not always wear lead apron, before age 20 years 9.5% 15.6% 0.006 Dental X-ray, after age 20 years 98.6% 97.9% 0.30 Did not always wear lead apron, after age 20 years 7.4% 12.1% 0.001 Spine X-ray 31.3% 28.7% 0.27 Upper G.I exam 22.0% 19.9% 0.33 Rib X-ray 6.4% 5.9% 0.73 CT exam of the trunk, above the waist 2.3% 3.0% 0.42 Chest fluoroscopy 1.8% 2.1% 0.74 Angiogram for heart or large arteries 0.7% 0.8% 0.99 Radiation treatment, above the waist 0.7% 1.3% 0.27 Note: SD = standard deviation, ET = estrogen therapy, EPT = estrogen therapy plus progestin therapy *P ascertained from Pearson χ 2 test, except where otherwise noted. † P ascertained from t test. 82 83 Table 4.2 Adjusted * ORs and 95% CIs of breast cancer associated with medical radiation exposure All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Frequency of chest X-ray Never 1-2 3-4 5-8 ≥9 199 129 54 31 17 650 552 238 145 114 1.00 1.34 (1.04-1.74) 1.39 (0.98-1.97) 1.66 (1.07-2.56) 2.08 (1.20-3.60) 160 123 49 30 26 1.00 1.42 (1.01-1.99) 1.39 (0.88-2.21) 1.67 (0.95-2.96) 2.50 (1.28-4.90) 304 275 125 74 58 1.00 1.35 (1.02-1.80) 1.53 (1.04-2.25) 1.66 (1.03-2.67) 2.17 (1.20-3.93) 1.00 1.07 (0.79-1.45) 0.93 (0.62-1.40) 0.97 (0.59-1.59) 1.18 (0.69-2.02) Trend P-value 0.0007 0.002 0.0009 0.83 Frequency of mammogram Never 1-2 3-4 5-6 ≥7 225 137 47 19 13 848 498 179 92 115 1.00 0.96 (0.73-1.27) 0.94 (0.63-1.41) 1.09 (0.63-1.91) 1.80 (0.95-3.42) 228 97 32 15 20 1.00 0.99 (0.68-1.43) 1.00 (0.58-1.75) 1.09 (0.51-2.32) 2.11 (0.95-4.69) 396 251 96 49 57 1.00 0.95 (0.70-1.29) 0.96 (0.62-1.50) 1.10 (0.60-2.02) 1.62 (0.81-3.21) 1.00 1.00 (0.72-1.40) 1.02 (0.62-1.67) 0.98 (0.51-1.89) 1.25 (0.68-2.32) Trend P-value 0.20 0.19 0.32 0.62 Used lead apron during dental X-ray Always Did not always wore Before age 20 After age 20 Both before and after Lead apron use unknown 320 22 12 11 70 1117 112 60 118 303 1.00 1.52 (0.94-2.47) 1.42 (0.75-2.71) 2.86 (1.51-5.41) 1.27 (0.94-1.71) 255 28 21 26 58 1.00 1.88 (1.03, 3.41) 2.11 (0.99-4.46) 3.13 (1.49-6.55) 1.16 (0.78-1.74) 540 63 25 65 149 1.00 1.62 (0.96-2.73) 1.20 (0.58-2.47) 3.17 (1.63-6.17) 1.27 (0.92-1.77) 1.00 1.19 (0.73-1.94) 1.82 (0.97-3.40) 1.00 (0.61-1.65) 0.91 (0.64-1.30) 83 84 Table 4.2 continued. All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Spine X-ray Never Ever 300 138 1228 499 1.00 0.94 (0.74-1.18) 290 104 1.00 0.87 (0.64-1.19) 593 254 1.00 0.92 (0.71-1.19) 1.00 0.95 (0.72-1.27) Upper G.I. or barium swallow Never Ever 342 97 1388 347 1.00 0.85 (0.65-1.11) 319 74 1.00 0.93 (0.65-1.33) 683 171 1.00 0.82 (0.61-1.11) 1.00 1.14 (0.82-1.58) Rib X-rays Never Ever 412 28 1625 103 1.00 0.98 (0.63-1.53) 371 22 1.00 1.03 (0.57-1.87) 801 50 1.00 0.97 (0.59-1.59) 1.00 1.05 (0.61-1.80) CT exam, above waist Never Ever 431 10 1689 52 1.00 1.32 (0.65-2.66) 381 12 1.00 1.52 (0.63-3.65) 831 26 1.00 1.41 (0.66-3.02) 1.00 1.05 (0.51-2.17) Chest fluoroscopy Never Ever 428 8 1693 36 1.00 1.07 (0.48-2.38) 386 5 1.00 0.60 (0.19-1.93) 830 20 1.00 1.32 (0.56-3.15) 1.00 0.42 (0.15-1.19) Note: OR = odds ratio, CI = confidence interval, ER = estrogen receptor, PR = progesterone receptor, BMI = body mass index * Models included age (<30, 30-34, 35-39, 40-44, 45-49 years), race, education, breast or ovarian cancer family history [none, breast or ovarian cancer in a mother or a sister (first degree), breast or ovarian cancer in an aunt or grandmother (second degree), unknown], age at menarche (<13, ≥13), parity (Never, Ever), age at first full-term pregnancy (<30, ≥30), BMI one year prior to reference date (continuous), average number of alcoholic drinks per week in the five years ending 2 years prior to the reference age (Never, <3, 3-5, ≥6), a variable combining menopausal status and hormone therapy use, history of benign breast diseases (Never, ever). † Excluded the subjects who missed information for the corresponding medical radiation exposure. ‡Exposure before age 20 years and after age 20 years are mutually adjusted. 84 85 20 years had a higher OR (1.81, 95% CI = 1.13-2.90), and women who had dental X- rays without lead apron protection both before and after age 20 had the highest OR (2.86, 95% CI = 1.51-5.41). Although the associations with chest X-rays and mammograms appeared slightly stronger for ER-PR- than ER+PR+ breast cancer, none of these differences were statistically significant (both P trend > 0.60). There were no statistically significant associations between breast cancer risk and other less common medical procedures. The ORs of breast cancer appeared higher when the first exposure to each of the types of radiations had occurred at an early age (Table 4.3). The OR of breast cancer associated with 5 or more chest X-rays was 2.78 (95% CI = 1.29-5.98) when the first chest X-ray was at age 10 years or younger and 1.62 (1.08-2.43) when the first chest X-ray was taken after age 10 years. The OR of breast cancer associated with at least 5 mammograms was 1.58 (95% CI = 0.95-2.64) when the first mammogram was obtained at age 35 years or younger while it was 0.86 (95% CI = 0.39-1.91) when the first mammogram was obtained after age 35 years. The OR of breast cancer associated with dental X-rays taken without the protection of a lead apron was 2.12 (95% CI = 1.35-3.33) when the first dental X-ray was done before or at age 10 years while it was 1.47, 95% CI = 0.90-2.40) when the first dental X-ray was after age 10 years. However, none of the differential effects observed by age were statistically significant (all Ps for effect modification ≥ 0.11). 86 The effects of the three common radiation exposures on breast cancer risk were consistently stronger among nulliparous women than among parous women (Table 4.4). The OR of breast cancer associated with 5 or more chest X-rays was 2.17 (95% CI = 1.10-4.30) for nulliparous women and 1.72 (1.11-2.67) for parous women. The OR of breast cancer associated with 5 or more mammograms was 2.20 (95% CI = 0.80-6.07) for nulliparous and 1.20 (95% CI = 0.71-2.01) for parous women. The OR of breast cancer associated with dental X-ray without the protection of lead apron was 3.06 (95% CI = 1.45-6.43) for nulliparous women and 1.53 (95% CI = 1.03 – 2.27) for parous women. However, none of these differences in OR by parity reached statistically significance (all Ps for effect modification ≥ 0.10). The association of breast cancer risk with mammograms and chest X-rays was stronger among women with a first- or second-degree breast or ovarian cancer family history than among women without a family history (Table 4.5). The effect modification by family history did not reach statistically significance for any of the three common radiation exposures (all Ps for effect modification ≥ 0.37). 87 Table 4.3 Adjusted * ORs and 95% CIs of breast cancer associated with medical radiation exposure by age at first exposure All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Age at first and frequency of chest X-ray Never Age ≤10 + (1-4) Age ≤10 + ( ≥5) Age >10 + (1-4) Age >10 + ( ≥5) Age or frequency unknown 199 17 8 160 38 12 650 70 66 680 182 71 1.00 1.27 (0.72-2.24) 2.78 (1.29-5.98) 1.34 (1.05-1.72) 1.62 (1.08-2.43) 1.85 (0.97-3.54) 160 16 16 149 37 13 1.00 1.41 (0.67-2.97) 3.37 (1.37-8.29) 1.39 (1.01-1.92) 1.67 (0.99-2.83) 1.77 (0.76-4.11) 304 38 27 345 100 32 1.00 1.38 (0.75-2.57) 2.44 (1.06-5.61) 1.39 (1.06-1.83) 1.74 (1.12-2.70) 1.78 (0.87-3.62) 1.00 0.98 (0.51-1.87) 1.38 (0.71-2.71) 1.03 (0.77-1.38) 0.95 (0.60-1.50) 0.98 (0.48-1.99) Test effect modification by age at first chest X-ray, P- value 0.22 0.19 0.48 0.37 Age at first and Frequency of mammogram Never Age ≤35 + (1-4) Age ≤35 + ( ≥5) Age >35 + (1-4) Age >35 + ( ≥5) 225 95 23 89 9 848 342 169 332 36 1.00 0.93 (0.69-1.25) 1.58 (0.95-2.64) 0.99 (0.70-1.39) 0.86 (0.39-1.91) 228 73 29 55 6 1.00 1.01 (0.68-1.50) 1.75 (0.91-3.36) 0.95 (0.59-1.51) 0.90 (0.29-2.74) 396 170 89 175 15 1.00 0.89 (0.64-1.25) 1.51 (0.87-2.62) 1.01 (0.69-1.47) 0.74 (0.30-1.84) 1.00 1.08 (0.76-1.54) 1.11 (0.66-1.87) 0.91 (0.59-1.38) 1.28 (0.46-3.61) Test effect modification by age at first mammogram, P-value 0.21 0.31 0.19 0.81 87 88 Table 4.3 continued. All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Use of lead apron during dental X-ray Always Age at first dental X-ray of subjects who did not always wear lead apron Age ≤10 Age >10 Lead apron use or age at first use unknown 320 24 21 70 1117 176 112 305 1.00 2.12 (1.35-3.33) 1.47 (0.90-2.40) 1.28 (0.95-1.72) 255 45 30 58 1.00 2.68 (1.57-4.57) 1.78 (0.98-3.25) 1.16 (0.78-1.73) 540 100 51 151 1.00 2.36 (1.46-3.81) 1.31 (0.76-2.25) 1.29 (0.93-1.79) 1.00 1.16 (0.78-1.74) 1.39 (0.85-2.29) 0.89 (0.63-1.28) Test effect modification by age at first dental X-ray, P- value 0.28 0.32 0.11 0.58 Note: OR = odds ratio, CI = confidence interval, ER = estrogen receptor, PR = progesterone receptor, BMI = body mass index * Models included age (<30, 30-34, 35-39, 40-44, 45-49 years), race, education, breast or ovarian cancer family history [none, breast or ovarian cancer in a mother or a sister (first degree), breast or ovarian cancer in an aunt or grandmother (second degree), unknown], age at menarche (<13, ≥13), parity (Never, Ever), age at first full-term pregnancy (<30, ≥30), BMI one year prior to reference date (continuous), average number of alcoholic drinks per week in the recent five years (Never, <3, 3-5, ≥6), a variable combining menopausal status and hormone therapy use, history of benign breast diseases (Never, ever). † Excluded the subjects who missed information for the corresponding medical radiation exposure. ‡Exposure before 20 and since 20 mutually adjusted for each other. 88 89 Table 4.4 Adjusted * ORs and 95% CIs of breast cancer associated with medical radiation exposure by parity status All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Frequency of chest X-ray Nulliparous women Never 1-4 ≥5 57 50 15 209 256 88 1.00 1.57 (0.99-2.51) 2.17 (1.10-4.30) 41 51 21 1.00 2.17 (1.16-4.08) 4.10 (1.70-9.89) 104 143 45 1.00 1.77 (1.06-2.96) 2.17 (1.02-4.63) 1.00 1.22 (0.71-2.08) 1.77 (0.87-3.59) Trend P-value 0.01 0.0008 0.02 0.13 Parous women Never 1-4 ≥5 142 133 33 441 534 171 1.00 1.28 (0.97-1.70) 1.72 (1.11-2.67) 119 121 35 1.00 1.25 (0.87-1.81) 1.59 (0.90-2.79) 200 257 87 1.00 1.29 (0.94-1.77) 1.81 (1.12-2.92) 1.00 1.01 (0.72-1.41) 0.89 (0.54-1.45) Trend P-value 0.009 0.08 0.01 0.70 Test effect modification by parity status, P-value 0.69 0.18 0.67 0.32 Frequency of mammogram Nulliparous women Never 1-4 ≥5 67 52 6 297 193 75 1.00 0.79 (0.47-1.32) 2.20 (0.80-6.07) 70 33 12 1.00 1.03 (0.51-2.06) 3.44 (0.98-12.09) 154 105 40 1.00 0.75 (0.43-1.32) 1.85 (0.63-5.40) 1.00 1.20 (0.67-2.16) 1.65 (0.68-4.05) Trend P-value 0.51 0.17 0.76 0.28 Parous women Never 1-4 ≥5 158 132 26 551 484 132 1.00 1.01 (0.74-1.38) 1.20 (0.71-2.01) 158 96 23 1.00 0.96 (0.64-1.46) 1.15 (0.58-2.29) 242 242 66 1.00 1.04 (0.73-1.48) 1.19 (0.67-2.11) 1.00 0.90 (0.62-1.31) 0.94 (0.52-1.72) 89 90 Table 4.4 continued All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Trend P-value 0.60 0.84 0.60 0.71 Test effect modification by parity status, P-value 0.77 0.58 0.94 0.47 Use of lead apron during dental X-ray Nulliparous women Always Did not always wear Lead apron use unknown 95 9 20 367 110 81 1.00 3.06 (1.45-6.43) 1.14 (0.64-2.03) 74 25 16 1.00 3.63 (1.53-8.62) 1.26 (0.58-2.74) 197 63 36 1.00 3.01 (1.39-6.53) 0.89 (0.47-1.69) 1.00 1.23 (0.69-2.21) 1.56 (0.77-3.14) Parous women Always Did not always wear Lead apron use unknown 225 36 50 750 180 222 1.00 1.53 (1.03-2.27) 1.26 (0.88-1.79) 181 50 42 1.00 1.93 (1.19-3.15) 1.07 (0.66-1.72) 343 90 113 1.00 1.64 (1.06-2.54) 1.39 (0.94-2.05) 1.00 1.22 (0.80-1.84) 0.75 (0.49-1.15) Test effect modification by parity status, P-value 0.10 0.18 0.18 0.80 Note: OR = odds ratio, CI = confidence interval, ER = estrogen receptor, PR = progesterone receptor, BMI = body mass index * Models included age (<30, 30-34, 35-39, 40-44, 45-49 years), race, education, breast or ovarian cancer family history [none, breast or ovarian cancer in a mother or a sister (first degree), breast or ovarian cancer in an aunt or grandmother (second degree), unknown], age at menarche (<13, ≥13), BMI one year prior to reference date (continuous), average number of alcoholic drinks per week in the recent five years (Never, <3, 3-5, ≥6), a variable combining menopausal status and hormone therapy use, history of benign breast diseases (Never, ever). † Excluded the subjects who missed information for the corresponding medical radiation exposure ‡Exposure before 20 and since 20 mutually adjusted for each other. 90 91 Table 4.5 Adjusted * ORs and 95% CIs of breast cancer associated with medical radiation exposure by a family history of breast or ovarian cancer All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Frequency of chest X-ray Women with breast or ovarian cancer family history Never 1-4 ≥5 63 74 15 277 356 119 1.00 1.23 (0.83-1.84) 2.14 (1.14-4.04) 65 71 26 1.00 1.34 (0.80-2.25) 2.71 (1.24-5.93) 132 181 60 1.00 1.22 (0.79-1.88) 1.92 (0.97-3.80) 1.00 1.13 (0.72-1.76) 1.47 (0.80-2.69) Trend P-value 0.02 0.02 0.07 0.23 Women without breast or ovarian cancer family history Never 1-4 ≥5 131 105 33 357 405 131 1.00 1.44 (1.06-1.96) 1.60 (1.01-2.53) 91 92 29 1.00 1.43 (0.95-2.14) 1.57 (0.86-2.87) 165 206 65 1.00 1.56 (1.10-2.20) 1.75 (1.05-2.94) 1.00 0.93 (0.63-1.37) 0.82 (0.47-1.44) Trend P-value 0.01 0.06 0.008 0.50 Test effect modification by parity status, P-value 0.96 0.86 0.87 0.56 Frequency of mammogram Women with breast or ovarian cancer family history Never 1-4 ≥5 62 80 15 334 299 134 1.00 0.81 (0.53-1.24) 1.81 (0.93-3.52) 84 59 21 1.00 0.99 (0.56-1.76) 2.01 (0.85-4.73) 162 149 68 1.00 0.74 (0.46-1.20) 1.54 (0.75-3.17) 1.00 1.29 (0.80-2.10) 1.30 (0.67-2.54) Trend P-value 0.22 0.19 0.50 0.35 Women without breast or ovarian cancer family history Never 1-4 ≥5 158 99 17 485 355 68 1.00 1.10 (0.77-1.56) 1.13 (0.60-2.12) 133 67 14 1.00 1.03 (0.64-1.65) 1.32 (0.57-3.06) 223 185 34 1.00 1.16 (0.78-1.71) 1.04 (0.51-2.12) 1.00 0.85 (0.55-1.32) 1.20 (0.55-2.60) 91 92 Table 4.5 continued All cases vs. controls ER-PR- cases vs. controls ER+PR+ cases vs. controls Variables Controls † Cases † OR (95% CI) Cases † OR (95% CI) Cases † OR (95% CI) ER-PR- vs. ER+PR+ cases OR (95% CI) Trend P-value 0.60 0.62 0.65 0.94 Test effect modification by parity status, P-value 0.48 0.41 0.37 0.92 Use of lead apron during dental X-ray Women with breast or ovarian cancer family history Always Did not always wear Lead apron use unknown 115 19 22 497 134 125 1.00 1.66 (0.98-2.83) 1.36 (0.82-2.27) 103 38 23 1.00 2.53 (1.33-4.79) 1.31 (0.67-2.57) 247 68 59 1.00 1.64 (0.93-2.91) 1.33 (0.76-2.33) 1.00 1.55 (0.95-2.53) 0.99 (0.56-1.73) Women without breast or ovarian cancer family history Always Did not always wear Lead apron use unknown 197 26 46 586 143 167 1.00 1.86 (1.18-2.95) 1.20 (0.82-1.75) 141 36 33 1.00 2.21 (1.26-3.90) 1.11 (0.66-1.87) 278 74 87 1.00 1.93 (1.17-3.19) 1.26 (0.82-1.92) 1.00 1.16 (0.72-1.88) 0.83 (0.51-1.34) Test effect modification by parity status, P-value 0.70 0.86 0.65 0.46 Note: OR = odds ratio, CI = confidence interval, ER = estrogen receptor, PR = progesterone receptor, BMI = body mass index * Models included age (<30, 30-34, 35-39, 40-44, 45-49 years), race, education, breast or ovarian cancer family history [none, breast or ovarian cancer in a mother or a sister (first degree), breast or ovarian cancer in an aunt or grandmother (second degree), unknown], age at menarche (<13, ≥13), BMI one year prior to reference date (continuous), average number of alcoholic drinks per week in the recent five years (Never, <3, 3-5, ≥6), a variable combining menopausal status and hormone therapy use, history of benign breast diseases (Never, ever). † Excluded the subjects who missed information for the corresponding medical radiation exposure. ‡Exposure before 20 and since 20 mutually adjusted for each other. 92 93 4.5 Discussion We found an elevated breast cancer risk among women who reported having had multiple chest X-rays or mammograms 7 or more years before diagnosis, or who had had dental X-rays without lead apron protection before age 20 years. Age at first exposure, parity, and breast or ovarian cancer family history appeared to modify the effects of medical radiation on breast cancer risk although none of the statistical tests for effect modification were statistical significant. Although it appears that the effect of radiation on ER-PR- cancer was greater than it was on ER+PR+ cancer, the differences in the effects by receptor status were not fully consistent, and none of them were statistically significant. We found that the frequency of chest X-rays was positively associated with ORs of breast cancer. The effect magnitude of chest X-rays was similar to that from mammograms. This seemed contrary to the radiation dose-response relationship since a mammogram has a higher mean radiation dose to breast tissue (about 3 mSv (Brenner et al., 2002)) than that from a chest X-ray (about 0.2 mSv (John and Kelsey, 1993)). However, the magnitude of risk per unit of radiation depends strongly on when radiation exposure occur: exposure before the age of 20 years carries the greatest risk (Boice and Monson, 1977; Ronckers et al., 2005). In our data, the median age at first exposure was 15 years for chest X-ray and 35 years for mammogram. Previously, one case-control study from Los Angeles reported a 94 positive association between number of chest X-rays before age 20 and breast cancer, but no such association for exposure after age 20 (Hill et al., 2002); another case-control study did not detect any association between chest X-rays and breast cancer risk among women aged 30-80 years (Zhu et al., 2002). We found that the frequency of dental X-rays was not associated with breast cancer risk among women who reported having always worn a lead apron when having dental X-rays, whereas the OR of breast cancer increased among women who reported having had dental X-rays without lead apron protection before age 20 years. These results are consistent with a previous case-control study conducted among Los Angeles women aged 40 or younger (Hill et al., 2002). Another case-control study conducted among Connecticut women aged 30-80 years reported that the frequency of dental X-ray was not associated with increased risk of breast cancer among premenopausal women, but this study did not report their results by lead apron use or by age at first exposure (Zheng et al., 2002). In addition, we found the high OR among those every having dental X-rays without the protection of a lead apron both before age 20 years as well as after age 20 years. If we assume these women received higher cumulative dose of radiation from dental X-rays, our findings indicate a dose-response relationship between dental X-rays and breast cancer risk. 95 We observed that younger age at first exposure of chest X-ray ( ≤10 years), dental X- ray without lead apron protection ( ≤10 years) or mammogram ( ≤ 35 years), was associated with higher ORs than those initially exposed at older age, which is consistent with the pattern seen in previous studies for both relatively low-dose (Boice and Monson, 1977) and moderate to high-dose (Land et al., 2003; Mattsson et al., 1993; Preston et al., 2002; Swerdlow et al., 2000; Tokunaga et al., 1994) radiation. In our data, nulliparous women had a higher OR following radiation exposures than parous women. A similar effect modification by parity has been observed in women who had been exposed to multiple fluoroscopies (Ronckers et al., 2005) and among women with a history of benign disease who had been exposed to low-dose medical radiation (Hill et al., 2002). The findings are also consistent with the observed decrease in breast cancer risk with increasing parity among atomic bomb survivors (Land et al., 1994; Tokunaga et al., 1994). However, a modifying effect by parity was not observed among women who had been treated with various radiation treatments for skin hemangioma during their infancy (Holmberg et al., 2001). Parity may protect against receptor positive breast cancer by causing the full-differentiation of breast epithelium which reduces the susceptibility to carcinogenesis (Russo et al., 2005; Russo et al., 1992). 96 In addition, we found that women with breast or ovarian cancer family history had higher ORs following the exposure of chest X-rays or multiple mammograms compared to those without family history. This result is consistent with a previous study conducted among Los Angeles women aged 40 years or younger (Hill et al., 2002). We did not find any increased breast cancer risk related to a history of spine X-rays, upper GI exams, rib X-rays, CT exams above the waist, or chest fluoroscopy. Women who reported having had CT exams above their waist or chest fluoroscopy received a high dose of radiation, but there was only a small number of subjects in our study (both ≤3%), and thus, this resulted in limited statistical power to detect any association. Spine X-rays, upper GI exams, and rib X-rays were more frequently used than CT exams or chest fluoroscopies. However, these exams were all relatively rare (median number among women who ever had: 2 exams for spine X- rays, 1 for upper GI exams and rib X-rays), and age at exposure was relative older (all at 25 or older). In our data, there was some suggestion of a stronger effect on ER-PR- than ER+PR+ cancer. However, the differences between the two subtypes were not fully consistent. Huang et al. reported that ever medical radiation exposure to chest possibly increased the risk for ER-PR- cancer (OR = 1.8, 95% CI = 0.5-5.9), but not 97 for ER+PR+ (OR = 0.6, 95% CI: = 0.1-3.4) breast cancer among pre/peri- menopausal women (Huang et al., 2000). We found no other published report for radiation exposure by ER/PR status. Several limitations of this study must be considered. Because of the small number of controls, we used an unconditional instead of a conditional logistic regression approach. Although the results of our analyses were essentially similar to that of a detailed stratified analysis, there is some possibility that this would have biased our results. However, if at all, this bias would have been towards the null value (Rothman and Greenland, 1998). Another limitation is that the information of radiation exposure from different medical procedures was obtained from the participants’ reports. It is possible that case patients recalled medical radiation exposure more carefully than control subjects since they were concerned about the cause of their diseases. This could have resulted in recall bias and caused the apparent effects. However, we think recall bias is unlikely to have accounted for the stronger association we observed among nulliparous than among parous women. We also considered the possibility that early stages of breast cancer, such as certain types of benign proliferative breast conditions could have caused women to have a larger number of radiation procedures several years prior to their diagnosis. However, when we repeated our 98 analyses among women without benign breast disease the results were essentially the same (results not shown). Similarly, when we restricted analyses to women who had never undergone radiation treatment the associations were essentially identical to those we observed among all women (results not shown). 4.6 Conclusions Our results support previous reports that low-dose ionizing radiation is associated with increased risk of breast cancer. Age at first exposure, parity, and breast or ovarian cancer family history may modify the radiation effect, but we found no consistent effect modification by ER/PR status. 99 Bibliography Althuis, M. D., Fergenbaum, J. H., Garcia-Closas, M., Brinton, L. A., Madigan, M. P., and Sherman, M. E. (2004). Etiology of hormone receptor-defined breast cancer: a systematic review of the literature. Cancer Epidemiology, Biomarkers & Prevention 13, 1558-1568. American Cancer Society (2005). Brest Cancer Facts and Figure 2005-2006 (American Cancer Society). Anderson, E. (2002). The role of oestrogen and progesterone receptors in human mammary development and tumorigenesis. Breast Cancer Research 4, 197-201. Battersby, S., Robertson, B. J., Anderson, T. J., King, R. J., and McPherson, K. (1992). Influence of menstrual cycle, parity and oral contraceptive use on steroid hormone receptors in normal breast. British Journal of Cancer 65, 601-607. Bernstein, L., Henderson, B. E., Hanisch, R., Sullivan-Halley, J., and Ross, R. K. (1994). Physical exercise and reduced risk of breast cancer in young women. Journal of the National Cancer Institute 86, 1403-1408. Bernstein, L., Pike, M. C., Ross, R. K., Judd, H. L., Brown, J. B., and Henderson, B. E. (1985). Estrogen and sex hormone-binding globulin levels in nulliparous and parous women. Journal of the National Cancer Institute 74, 741-745. Bernstein, L., and Ross, R. K. (1993). Endogenous hormones and breast cancer risk. Epidemiologic Reviews 15, 48-65. Boice, J. D., Jr., and Monson, R. R. (1977). Breast cancer in women after repeated fluoroscopic examinations of the chest. Journal of the National Cancer Institute 59, 823-832. Brenner, D. J., Sawant, S. G., Hande, M. P., Miller, R. C., Elliston, C. D., Fu, Z., Randers-Pehrson, G., and Marino, S. A. (2002). Routine screening mammography: how important is the radiation-risk side of the benefit-risk equation? International Journal of Radiation Biology 78, 1065-1067. 100 Britton, J. A., Gammon, M. D., Schoenberg, J. B., Stanford, J. L., Coates, R. J., Swanson, C. A., Potischman, N., Malone, K. E., Brogan, D. J., Daling, J. R., and Brinton, L. A. (2002). Risk of breast cancer classified by joint estrogen receptor and progesterone receptor status among women 20-44 years of age. American Journal of Epidemiology 156, 507-516. Chu, K. C., and Anderson, W. F. (2002). Rates for breast cancer characteristics by estrogen and progesterone receptor status in the major racial/ethnic groups. Breast Cancer Research & Treatment 74, 199-211. Colditz, G. A., Rosner, B. A., Chen, W. Y., Holmes, M. D., and Hankinson, S. E. (2004). Risk factors for breast cancer according to estrogen and progesterone receptor status. Journal of the National Cancer Institute 96, 218-228. Colleoni, M., Rotmensz, N., Robertson, C., Orlando, L., Viale, G., Renne, G., Luini, A., Veronesi, P., Intra, M., Orecchia, R., et al. (2002). Very young women (<35 years) with operable breast cancer: features of disease at presentation. Annals of Oncology 13, 273-279. Cotterchio, M., Kreiger, N., Theis, B., Sloan, M., and Bahl, S. (2003). Hormonal factors and the risk of breast cancer according to estrogen- and progesterone-receptor subgroup. Cancer Epidemiology, Biomarkers & Prevention 12, 1053-1060. Darby, S. C., Doll, R., Gill, S. K., and Smith, P. G. (1987). Long term mortality after a single treatment course with X-rays in patients treated for ankylosing spondylitis. British Journal of Cancer 55, 179-190. DerSimonian, R., and Laird, N. (1986). Meta-analysis in clinical trials. Controlled Clinical Trials 7, 177-188. Dickson, R. B., and Stancel, G. M. (2000). Estrogen receptor-mediated processes in normal and cancer cells. Journal of the National Cancer Institute Monographs 27, 135-145. Donegan, W. L. (1992). Prognostic factors. Stage and receptor status in breast cancer. Cancer 70, 1755-1764. Dontu, G., El-Ashry, D., and Wicha, M. S. (2004). Breast cancer, stem/progenitor cells and the estrogen receptor. Trends in Endocrinology & Metabolism 15, 193-197. 101 Dorgan, J. F., Reichman, M. E., Judd, J. T., Brown, C., Longcope, C., Schatzkin, A., Campbell, W. S., Franz, C., Kahle, L., and Taylor, P. R. (1995). Relationships of age and reproductive characteristics with plasma estrogens and androgens in premenopausal women. Cancer Epidemiology, Biomarkers & Prevention 4, 381-386. Egger, M., Davey Smith, G., Schneider, M., and Minder, C. (1997). Bias in meta- analysis detected by a simple, graphical test. BMJ 315, 629-634. Elwood, J. M., and Godolphin, W. (1980). Oestrogen receptors in breast tumours: associations with age, menopausal status and epidemiological and clinical features in 735 patients. British Journal of Cancer 42, 635-644. Enger, S. M., Ross, R. K., Paganini-Hill, A., Carpenter, C. L., and Bernstein, L. (2000). Body size, physical activity, and breast cancer hormone receptor status: results from two case-control studies. Cancer Epidemiology, Biomarkers & Prevention 9, 681-687. Enger, S. M., Ross, R. K., Paganini-Hill, A., Longnecker, M. P., and Bernstein, L. (1999). Alcohol consumption and breast cancer oestrogen and progesterone receptor status. British Journal of Cancer 79, 1308-1314. Evans, R. M. (1988). The steroid and thyroid hormone receptor superfamily. Science 240, 889-895 . Gao, X., and Nawaz, Z. (2002). Progesterone receptors - animal models and cell signaling in breast cancer: Role of steroid receptor coactivators and corepressors of progesterone receptors in breast cancer. Breast Cancer Research 4, 182-186. Garcia-Closas, M., Herbstman, J., Schiffman, M., Glass, A., and Dorgan, J. F. (2002). Relationship between serum hormone concentrations, reproductive history, alcohol consumption and genetic polymorphisms in pre-menopausal women. International Journal of Cancer 102, 172-178. Geisler, J., Detre, S., Berntsen, H., Ottestad, L., Lindtjorn, B., Dowsett, M., and Einstein Lonning, P. (2001). Influence of neoadjuvant anastrozole (Arimidex) on intratumoral estrogen levels and proliferation markers in patients with locally advanced breast cancer. Clinical Cancer Research 7, 1230-1236. Gorski, J., and Gannon, F. (1976). Current models of steroid hormone action: a critique. Annual Review of Physiology 38, 425-450. 102 Greenland, S., and Longnecker, M. P. (1992). Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. American Journal of Epidemiology 135, 1301-1309. Habel, L. A., and Stanford, J. L. (1993). Hormone receptors and breast cancer. Epidemiologic Reviews 15, 209-219. Hanstein, B., Djahansouzi, S., Dall, P., Beckmann, M. W., and Bender, H. G. (2004). Insights into the molecular biology of the estrogen receptor define novel therapeutic targets for breast cancer. European Journal of Endocrinology 150, 243-255. Henderson, B. E., Ross, R. K., Pike, M. C., and Casagrande, J. T. (1982). Endogenous hormones as a major factor in human cancer. Cancer Research 42, 3232-3239. Hildreth, N. G., Kelsey, J. L., Eisenfeld, A. J., LiVolsi, V. A., Holford, T. R., and Fischer, D. B. (1983). Differences in breast cancer risk factors according to the estrogen receptor level of the tumor. Journal of the National Cancer Institute 70, 1027-1031. Hill, D. A., Preston-Martin, S., Ross, R. K., and Bernstein, L. (2002). Medical radiation, family history of cancer, and benign breast disease in relation to breast cancer risk in young women, USA. Cancer Causes & Control 13, 711-718. Holmberg, E., Holm, L. E., Lundell, M., Mattsson, A., Wallgren, A., and Karlsson, P. (2001). Excess breast cancer risk and the role of parity, age at first childbirth and exposure to radiation in infancy. British Journal of Cancer 85, 362-366. Hosmer, D. W., and Lemeshow, S. (2000). Applied logistic regression (New York, NY: A Wiley-Interscience Publication). Huang, W. Y., Newman, B., Millikan, R. C., Schell, M. J., Hulka, B. S., and Moorman, P. G. (2000). Hormone-related factors and risk of breast cancer in relation to estrogen receptor and progesterone receptor status. American Journal of Epidemiology 151, 703-714. Jacquemier, J. D., Hassoun, J., Torrente, M., and Martin, P. M. (1990). Distribution of estrogen and progesterone receptors in healthy tissue adjacent to breast lesions at various stages--immunohistochemical study of 107 cases. Breast Cancer Research & Treatment 15, 109-117. 103 Jensen, E. V. (1968). Estrogen receptor: ambiguities in the use of this term. Science 159, 15. Jensen, E. V., Desombre, E. R., Hurst, D. J., Kawashima, T., and Jungblut, P. W. (1967). Estrogen-receptor interactions in target tissues. Archives d Anatomie Microscopique et de Morphologie Experimentale 56, 547-569. Jensen, E. V., Suzuki, T., Numata, M., Smith, S., and DeSombre, E. R. (1969). Estrogen-binding substances of target tissues. Steroids 13, 417-427. John, E. M., and Kelsey, J. L. (1993). Radiation and other environmental exposures and breast cancer. Epidemiologic Reviews 15, 157-162. Kelsey, J. L., Gammon, M. D., and John, E. M. (1993). Reproductive factors and breast cancer. Epidemiologic Reviews 15, 36-47. Key, T. J., and Pike, M. C. (1988). The role of oestrogens and progestagens in the epidemiology and prevention of breast cancer. European Journal of Cancer & Clinical Oncology 24, 29-43. Khan, S. A., Rogers, M. A., Khurana, K. K., Meguid, M. M., and Numann, P. J. (1998). Estrogen receptor expression in benign breast epithelium and breast cancer risk. Journal of the National Cancer Institute 90, 37-42. Khan, S. A., Sachdeva, A., Naim, S., Meguid, M. M., Marx, W., Simon, H., Halverson, J. D., and Numann, P. J. (1999). The normal breast epithelium of women with breast cancer displays an aberrant response to estradiol. Cancer Epidemiology, Biomarkers & Prevention 8, 867-872. Klinge, C. M. (2000). Estrogen receptor interaction with co-activators and co- repressors. Steroids 65, 227-251. Land, C. E., Hayakawa, N., Machado, S. G., Yamada, Y., Pike, M. C., Akiba, S., and Tokunaga, M. (1994). A case-control interview study of breast cancer among Japanese A-bomb survivors. II. Interactions with radiation dose. Cancer Causes & Control 5, 167-176. Land, C. E., Tokunaga, M., Koyama, K., Soda, M., Preston, D. L., Nishimori, I., and Tokuoka, S. (2003). Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1990. Radiation Research 160, 707-717. 104 Li, C. I., Daling, J. R., and Malone, K. E. (2003). Incidence of invasive breast cancer by hormone receptor status from 1992 to 1998. J Clin Oncol 21, 28-34. Ma, H., Bernstein, L., Rose, R. K., and Ursin, G. (under review). Hormone-related factors and breast cancer risk of women under age 50 years by estrogen and progesterone receptor status. Marchbanks, P. A., McDonald, J. A., Wilson, H. G., Burnett, N. M., Daling, J. R., Bernstein, L., Malone, K. E., Strom, B. L., Norman, S. A., Weiss, L. K., et al. (2002). The NICHD Women's Contraceptive and Reproductive Experiences Study: methods and operational results. Annals of Epidemiology 12, 213-221. Mattsson, A., Ruden, B. I., Hall, P., Wilking, N., and Rutqvist, L. E. (1993). Radiation-induced breast cancer: long-term follow-up of radiation therapy for benign breast disease. Journal of the National Cancer Institute 85, 1679-1685. McCredie, M. R., Dite, G. S., Southey, M. C., Venter, D. J., Giles, G. G., and Hopper, J. L. (2003). Risk factors for breast cancer in young women by oestrogen receptor and progesterone receptor status. British Journal of Cancer 89, 1661-1663. McDonald, J. A., Mandel, M. G., Marchbanks, P. A., Folger, S. G., Daling, J. R., Ursin, G., Simon, M. S., Bernstein, L., Strom, B. L., Norman, S. A., et al. (2004). Alcohol exposure and breast cancer: results of the women's contraceptive and reproductive experiences study. Cancer Epidemiology, Biomarkers & Prevention 13, 2106-2116. McDonnell, D. P., and Norris, J. D. (2002). Connections and regulation of the human estrogen receptor. Science 296, 1642-1644. Mendelson, J. H., Lukas, S. E., Mello, N. K., Amass, L., Ellingboe, J., and Skupny, A. (1988). Acute alcohol effects on plasma estradiol levels in women. Psychopharmacology 94, 464-467. Milgrom, E., Atger, M., and Baulieu, E. E. (1970). Progesterone binding plasma protein (PBP). Nature 228, 1205-1206. Milgrom, E., and Baulieu, E. E. (1970). Progesterone in uterus and plasma. I. Binding in rat uterus 105,000 g supernatant. Endocrinology 87, 276-286. Pagano, M., and Gauvreau, K. (2000). Principles of biostatistics, 2nd. edn (Pacific Grove, CA: Duxbury Press). 105 Potter, J. D., Cerhan, J. R., Sellers, T. A., McGovern, P. G., Drinkard, C., Kushi, L. R., and Folsom, A. R. (1995). Progesterone and estrogen receptors and mammary neoplasia in the Iowa Women's Health Study: how many kinds of breast cancer are there? Cancer Epidemiology, Biomarkers & Prevention 4, 319-326. Preston, D. L., Mattsson, A., Holmberg, E., Shore, R., Hildreth, N. G., and Boice, J. D., Jr. (2002). Radiation effects on breast cancer risk: a pooled analysis of eight cohorts. Radiation Research 158, 220-235. Reichman, M. E., Judd, J. T., Longcope, C., Schatzkin, A., Clevidence, B. A., Nair, P. P., Campbell, W. S., and Taylor, P. R. (1993). Effects of alcohol consumption on plasma and urinary hormone concentrations in premenopausal women. Journal of the National Cancer Institute 85, 722-727. Ricketts, D., Turnbull, L., Ryall, G., Bakhshi, R., Rawson, N. S., Gazet, J. C., Nolan, C., and Coombes, R. C. (1991). Estrogen and progesterone receptors in the normal female breast. Cancer Research 51, 1817-1822. Ronckers, C. M., Erdmann, C. A., and Land, C. E. (2005). Radiation and breast cancer: a review of current evidence. Breast Cancer Research 7, 21-32. Rothman, K. J., and Greenland, S. (1998). Modern epidemiology (Philadelphia: Lippincott-Raven). Rusiecki, J., Holford, T., Zahm, S., and Zheng, T. (2005). Breast cancer risk factors according to joint estrogen receptor and progesterone receptor status. Cancer Detection and Prevention 29, 419-426. Russo, J., Hu, Y. F., Yang, X., and Russo, I. H. (2000). Developmental, cellular, and molecular basis of human breast cancer. Journal of the National Cancer Institute Monographs 27, 17-37. Russo, J., Mailo, D., Hu, Y. F., Balogh, G., Sheriff, F., and Russo, I. H. (2005). Breast differentiation and its implication in cancer prevention. Clinical Cancer Research 11, 931s-936s. Russo, J., Rivera, R., and Russo, I. H. (1992). Influence of age and parity on the development of the human breast. Breast Cancer Research & Treatment 23, 211-218. 106 Stanford, J. L., and Greenberg, R. S. (1989). Breast cancer incidence in young women by estrogen receptor status and race. American Journal of Public Health 79, 71-73. Stanford, J. L., Szklo, M., and Brinton, L. A. (1986). Estrogen receptors and breast cancer. Epidemiologic Reviews 8, 42-59. Swerdlow, A. J., Barber, J. A., Hudson, G. V., Cunningham, D., Gupta, R. K., Hancock, B. W., Horwich, A., Lister, T. A., and Linch, D. C. (2000). Risk of second malignancy after Hodgkin's disease in a collaborative British cohort: the relation to age at treatment. Journal of Clinical Oncology 18, 498-509. Tokunaga, M., Land, C. E., Tokuoka, S., Nishimori, I., Soda, M., and Akiba, S. (1994). Incidence of female breast cancer among atomic bomb survivors, 1950-1985. Radiation Research 138, 209-223. Ursin, G., Bernstein, L., Lord, S., Karim1, J. R., Deapen1, D., Press, M. F., Daling, J. R., Norman, S. A., Liff, J. M., Marchbanks, P. A., et al. (2005). Reproductive factors and subtypes of breast cancer defined by hormone receptor and histology. British Journal of Cancer 93, 364-371. Ursin, G., Peters, R. K., Henderson, B. E., d'Ablaing, G., 3rd, Monroe, K. R., and Pike, M. C. (1994). Oral contraceptive use and adenocarcinoma of cervix. Lancet 344, 1390-1394. Wakeford, R. (2004). The cancer epidemiology of radiation. Oncogene 23, 6404- 6428. Williams, G., Anderson, E., Howell, A., Watson, R., Coyne, J., Roberts, S. A., and Potten, C. S. (1991). Oral contraceptive (OCP) use increases proliferation and decreases oestrogen receptor content of epithelial cells in the normal human breast. International Journal of Cancer 48, 206-210. Yasui, Y., and Potter, J. D. (1999). The shape of age-incidence curves of female breast cancer by hormone-receptor status. Cancer Causes & Control 10, 431-437. Yoo, K. Y., Tajima, K., Miura, S., Takeuchi, T., Hirose, K., Risch, H., and Dubrow, R. (1997). Breast cancer risk factors according to combined estrogen and progesterone receptor status: a case-control analysis. American Journal of Epidemiology 146, 307-314. 107 Zheng, T., Holford, T. R., Mayne, S. T., Luo, J., Owens, P. H., Zhang, B., Zhang, W., and Zhang, Y. (2002). Radiation exposure from diagnostic and therapeutic treatments and risk of breast cancer. European Journal of Cancer Prevention 11, 229- 235. Zhu, K., Beiler, J., Hunter, S., Payne-Wilks, K., Roland, C. L., Forbes, D. S., Chinchilli, V. M., Bernard, L. J., Jacobsen, K. H., and Levine, R. S. (2002). The relationship between menstrual factors and breast cancer according to estrogen receptor status of tumor: a case-control study in African-American women. Ethnicity & Disease 12, 23-29.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Antioxidants and risk of myocardial infarction and cancer in a cohort of middle-aged finnish men: the kuopio ischemic heart disease risk factor study
PDF
Colorectal cancer: genomic variations in insulin-like growth factor-1 and -2
PDF
Recreational physical activity and risk of breast cancer
PDF
Risk factors associated with moderate and serious injuries attributable to the 1994 Northridge earthquake, Los Angeles County
PDF
Genetic risk factors in breast cancer susceptibility: The multiethnic cohort
PDF
Dietary fats, fat metabolizing genes, and the risk of breast cancer
PDF
The lifestyle and genetic influence on the progression of atherosclerosis
PDF
Breast cancer in the multiethnic cohort study: Genetic (prolactin pathway genes) and environmental (hormone therapy) factors
PDF
A prospective cohort study of morbidity and mortality among middle-aged men in Shanghai, China
PDF
Insulin-like growth factor 1 genotype, phenotype and breast cancer risk, by racial/ethnic group
PDF
Colorectal cancer risks in Singapore Chinese: Polymorphisms in the insulin-like growth factor-1 and the vitamin D receptor
PDF
The role of genetic variation in insulin -like growth factors in prostate cancer risk: The multiethnic cohort
PDF
Androgen receptor gene and prostate -specific antigen gene in breast cancer
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
Lymphedema: Impact on breast cancer survivors
PDF
Risk factors, sequellae and intervention studies for the treatment of diabetes mellitus in Native American populations
PDF
A case-control study of passive smoking and bladder cancer risk in Los Angeles
PDF
Physical activity and adenomatous polyps: Measures of association and impact
PDF
The effects of dietary and other lifestyle behaviors on the risk of colorectal cancer among the Mexican -origin Latino population
PDF
Androgens and breast cancer
Asset Metadata
Creator
Ma, Huiyan
(author)
Core Title
Risk factors for breast cancer according to estrogen and progesterone receptor status
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,OAI-PMH Harvest
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c17-129846
Unique identifier
UC11351900
Identifier
3238309.pdf (filename),usctheses-c17-129846 (legacy record id)
Legacy Identifier
3238309.pdf
Dmrecord
129846
Document Type
Dissertation
Rights
Ma, Huiyan
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
Tags
health sciences, oncology
health sciences, public health