Magnetic Fields, Night Shift Work and the Risk of Breast Cancer among Female Textile Workers in Shanghai, China
TABLE OF CONTENTS Page List of Tables ii Chapter 1: Introduction 1 Background 1 Risk Factor of Interest 2 Aims and Structure of the Dissertation 9 Notes to Chapter 1 11 Chapter 2: Occupational Exposure to Magnetic Fields in the Shanghai Textile Industry Background 18 Materials and Methods 19 Results 24 Discussion 26 Notes to Chapter 2 35 Chapter 3: Occupational Exposure to Magnetic Fields and Breast Cancer Among Textile Workers in Shanghai, China Background 39 Materials and Methods 40 Results 45 Discussion 46 Notes to Chapter 3 57 Chapter 4: Shift Work and Breast Cancer Among Textile Workers in Shanghai, China Background 67 Materials and Methods 68 Results 72 Discussion 75 Notes to Chapter 4 88 Chapter 5: Summary 91 Bibliography 96 l
LIST OF TABLES Table Number Page 2.1 Textile Job Grouping 31 2.2 Magnetic Field Exposures of Textile Job Categories 32 2.3 Magnetic Field Eexposures of Subgroups of Three Job Categories 34 2.4 Variance Component Estimates from Mixed-Effect Model 34 3.1 Characteristics of Cases and Non-cases 50 3.2 Reproductive, Lifestyle Factors in Relations to Risk of Breast Cancer 51 3.3 Magnetic Fields in Relations to Risk of Breast Cancer 53 3.4.a Magnetic Fields in Relations to Risk of Breast Cancer for Women < 50 years ....54 3.4.b Magnetic Fields in Relations to Risk of Breast Cancer for Women >50 years ...55 4.1 Characteristics of Cases and Non-cases 79 4.2 Years on Shift Work in Relations to Risk of Breast Cancer 80 4.3.a Years on Shift Work and Risk of Breast Cancer for Women < 50 years 81 4.3.b Years on Shift Work and Risk of Breast Cancer for Women >50 years 82 4.4 Number of Night Shifts in Relations to Risk of Breast Cancer 83 4.5.a Number of Night Shifts and Risk of Breast Cancer for Women < 50 years 84 4.5.b Number of Night Shifts and Risk of Breast Cancer for Women >50 years 85 4.6 Years on Frequent Shift Rotation Cycle and Risk of Breast Cancer 86 4.7 Number of Night Shifts on Frequent Shift Rotation Cycle and Risk of Breast Cancer 87 4.8 Joint Effects of Shift Work and Magnetic Field Exposure 88 n
ACKNOWLEDGEMENTS I wish to thank a number of people who have been involved with this project and who have been incredibly generous with their time, support, expertise, and encouragement. My deepest gratitude is to my advisor, Dr. Harvey Checkoway. I have been amazingly fortunate to have an advisor who gave me the freedom to explore on my own, and at the same time provided the clear guidance on conducting high quality research. His patience and support helped me overcome many obstacles and finish this dissertation. Many thanks also go to my co-advisor, Dr. David Thomas. Through twelve years working with him on numeriou epidemiological studies, Dr. Thomas has inspired and encouraged me to become a researcher. He sets high standards for his students and he encourages and guides them to meet those standards. I am particularly grateful for his intense feedback and detailed review to my writings. I would like to acknowledge Dr. Mike Yost for the long discussions and his insightful comments on my research. His incredible expertise in research on electromagnetic fields is essential in completion of my dissertation. I am grateful to Dr. Scott Davis for his encouragement, insightful feedback, and always making time for me whenever I had requests. I am also indebted to Dr. Norman Breslow, the best teacher and biostatisican, for numerous discussions and lectures on related topics that helped me improve my knowledge in the area. This project would never have been accomplished without the tenacious effort of an amazing research team. I wish to acknowledge the Seattle staff and researchers including Roberta Ray, Janice Camp, Dawn Ritzgibbons, Dr. Karen Wernli, Dr. Eva Wong, in
Georgia Green, and Dr. Angelar Carden; and the Shanghai staff and researchers including Dr. Dao-Li Gao, Wen Wan Wang, the STUB hygienists and the field workers. This project would never have been accomplished without their tenacious effort. My success would not have been possible without the love and patience of my family. My heart-felt thanks to my immediate family for a constant source of love, concern, support and strength all these years. I would like to express my heart-felt gratitude to mom and dad for their unconditional love and support. IV
DEDICATION To Jisheng, for inspiration v
1 CHAPTER 1: INTRODUCTION Background Breast cancer is the most common non-skin cancer among women worldwide. The incidence rates are high in most of the affluent countries but low, although increasing, in developing countries (1). Intensive research, traditionally focused on reproductive and lifestyle risk factors, has been conducted to improve the understanding of the etiology of breast cancer. Although a number of risk factors have been established, they only explain a moderate portion of breast cancer cases. The etiologic contributions of occupational risk factors have not been adequately studied, especially in view of the large numbers of women in the workforce worldwide with potentially hazardous exposures. The purpose of this dissertation project was to quantify breast cancer risks related to two very widespread occupational exposures, magnetic fields and shiftwork, for which a shared biological mechanism of melatonin suppression is hypothesized. Established Risk Factors Incidence rates of breast cancer increase dramatically with age. While the rate of increase in breast cancer incidence is greatest during women's reproductive years under about age 50, the majority of cases occur after age 50. A substantial body of experimental, epidemiologic, and clinical evidence suggests that breast cancer risk is influenced by endogenous hormones (2). Estrogens have been extensively studied in this regard because of their important role in the growth of breast epithelium. Animal studies have repeatedly demonstrated that estrogens can induce and promote mammary tumors in rodents (3), and there is strong epidemiologic evidence that increased estrogen levels are important in determining risk of breast cancer in humans (4, 5). Most of the early epidemiological research on breast cancer etiology focused on factors that are closely related to endogenous hormones. To date, the most consistently observed risk factors in this regard that have been established are: young age at menarche, late age
2 at menopause, late age at first birth, nulliparity, use of oral contraceptives, and hormone replacement therapy for menopause. Long total duration of breastfeeding may have protective effects for breast cancer. In addition, observational studies have repeatedly shown that having first degree relatives with breast cancer, history of benign breast disease, exposure to ionizing radiation, and alcohol consumption are associated with moderately increased risks of breast cancer, whereas consumption of vegetables and moderate physical activity may confer protection (6). A small portion of breast cancers can be explained by genetic mutations. To date, at least five germline mutations that predispose to breast cancer have been identified or localized (7). Risk Factors of Interest Despite extensive research into the causes of breast cancer, only a moderate proportion of breast cancer can be explained by known risk factors. Numerous efforts have been initiated in recent years to expand the scope of epidemiologic investigation into breast cancer etiology. Of particular interest has been the possible influence of environmental and occupational factors. Two factors in this regard that have received great attention during the past two decades are long-term exposure to extremely low frequency electromagnetic fields generated by the use of electricity, and shift work. These two factors have been suggested to share the same underlying mechanism of action; that is, by altering normal pineal function and suppressing the normal nocturnal rise in melatonin production and release (8). Lately, hypothesized mechanisms for the effect of shift work have expanded to circadian disruption. Melatonin is synthesized and secreted by the pineal gland. Its release is stimulated by darkness and suppressed by light perceived by the retina. Hence, circulating melatonin concentrations are low in daylight and higher at night, exhibiting a characteristic rise in concentration after darkness and a peak near the midpoint of the dark interval (9). Melatonin appears to be involved in the regulation of gonadal function by influencing the hypothalamic-pituitary-gonadal axis. Animal studies indicate that melatonin can modify the frequency of the hypothalamic gonadotropin releasing hormone (GnRH) pulse
3 generator, thereby reducing the release of lutinizing hormone (LH) and follicle stimulating hormone (FSH) from the pituitary. LH and FSH, in turn, are critical in the biosynthesis of steroid hormones in the ovary, including estradiol (10, 11). Thus, melatonin may have an inhibitory effect on ovarian hormone-dependent tumors, such as breast cancer in women, and factors that reduce melatonin production, such as light at night during shift work, and magnetic fields, may enhance risk (12-14). Melatonin may also have a more direct effect on the development of tumor growth. In rodent models, pinealectomy has been found to enhance the growth of chemically- induced tumors (15), and exogenous melatonin administration has demonstrated anti- initiating (16) and oncostatic activity (17-19) in various chemically-induced cancers. Since the discovery of melatonin as the main pineal hormone (20), substantial effort has been devoted to testing the possible anti-neoplastic activity of melatonin. The overwhelming number of experimental studies has shown circadian stage-dependent effects of melatonin on growth of breast or other tumors. A clear inhibitory effect was observed if melatonin was administered in the evening, but not if given in the morning (21,22). Another possible effect of melatonin on tumors is through the modulation of the immune system (23). The immunostimulatory and antiapoptotic effects of melatonin are exerted mainly through its action on T-helper lymphocytes (Thl). One of the mechanisms that tumors use for evading the immune system is the production of factors which suppress Thl responses—cell mediated immunity against tumor cells. In vivo studies indicate that melatonin enhances T-cell immunity by stimulating interleukin 12 (IL-12) production. IL-12 drives T-cell differentiation towards Thl phenotype (24-26). It has become increasingly clear that melatonin also acts on natural killer (NK) cells, which are widely established killers of virus-infected cells and a variety of tumor cells. Animal studies have demonstrated that feeding endogenous melatonin to mice enhances the quantity of NK cells in the bone marrow and spleen.
4 The idea that electric power use might relate to risk of breast cancer derives from a hypothesis articulated by Stevens in 1987 that exposure to light at night or exposure to extremely low frequency of electromagnetic fields might lead to increased risk of breast cancer via impairment of pineal secretion of melatonin. (8). Electromagnetic fields are comprised of electric fields and magnetic fields. Magnetic fields have been studied extensively for effects on health. The ability of 60-Hz magnetic fields to suppress melatonin is the necessary first step in the proposed pathway leading to mammary carcinoma. Evidence for this was first provided by Yellon (27), who reported that a short exposure (15 min, 2 h before darkness onset) to 100 uT sinusoidal 60-Hz magnetic fields suppressed the normal nocturnal rise in pineal melatonin production in Djungarian hamsters. The same changes have been reported by Wilson et al. (28). In search for similar effects in other animal species, Kato et al. failed to find an effect on melatonin levels in Wistar-King rats exposed to various doses of magnetic fields including the dose used in Yellon and Wilson's studies (29, 30). Selmaoui and Tauitou have compared melatonin effects by short-term and long-term magnetic field exposures in Wistar rats. They found short-term exposure (once for 12 hours) had no effect on plasma concentration of melatonin levels, but that 30 days of repeated exposure with both 10 uT and 100 uT intensities resulted in an approximately 42 % decrease of plasma melatonin levels, thus suggesting that magnetic field may have a cumulative effect upon pineal function (31). The evidence for a suppressive effect of magnetic field exposure on melatonin production in human studies is also conflicting (32). Most experimental studies on humans have found no effects on serum melatonin production when subjects are exposed to magnetic fields during sleep in the laboratory (33), whereas some evidence of reductions in nocturnal urinary excretion of the melatonin metabolite, 6-sulfatoxymelatonin was observed from observational studies in occupational or residential settings (34-37). Urinary excretion of 6-hydroxymelatonin sulfate was also found to be lower in sewing
5 machine operators in the Finnish garment industry relative to office workers, although the magnitude of the reduction was not related to EMF dose over the work week (38). During the past two decades, there has been a considerable amount of epidemiologic research on possible relations between EMF exposure and breast cancer, including studies of residential and occupational exposures (39). The evidence is mixed. Early studies on residential exposures have focused on power transmission lines, which is probably the least specific indicator of personal exposure. Moderately elevated breast cancer risks (odds ratios=1.58, 95% CI: 1.30-1.92) were detected in a case-control study among Norwegian women (40), but were not found in other studies that relied on residential proximity to power transmission lines (41-44). Vena and colleagues reported that the use of electric blankets was associated with a moderately increased risk of breast cancer in both postmenopausal (45) and premenopausal (46) women, although findings from the Nurses' Health Study (47) and from a multi-center study in the US (48) indicate no relation with electric blanket use in either pre- or post-menopausal women. Davis et al. (49) reported no associations with risk of breast cancer and residential exposures, based on magnetic field measurements, home electrical wiring configuration, or self- reported electrical appliance use, including electric blankets, in a large population-based case-control study in Seattle. Most epidemiologic data on occupational magnetic field exposure and breast cancer have indicated little or no overall association (50-53), but some studies have suggested moderately elevated risks (54) or associations restricted to premenopausal women, particularly for breast tumors that are rich in estrogen receptors (52, 55) , which fits the hypothesized biologic mechanism. Early studies focused on "electric occupations," which presumably experienced high intensity of EMF exposure. The results from such studies in men have suggested a possible association between male breast cancer and electrical occupations (56-59). There have been relatively few women working in electric occupations in most countries, and findings were based on small numbers of cases. Recently, several population based case-control studies with large number of cases and
6 improved exposure assessment have been published. In a population based case-control study of 843 incident breast cancer cases identified through North Carolina Central Cancer Registry and 773 race and age frequency-matched controls (14), exposure was assessed through measurements of the magnetic fields made on a convenience sample of 200 women. The time weighted average exposure was calculated across occupational categories. These categories were, however, very broad. The only groups with elevated fields were industrial workers (0.54 uT) and "miscellaneous" occupations (0.23 uT). Overall, no associations were found between magnetic field exposure and breast cancer risk. For pre-menopausal women, or estrogen receptor positive breast cancer with more than 10 years of exposure, an indication of increased risk was found (Odds Ratios (OR) = 1.7, 95%CI: 1.1-2.7 for pre-menopause; OR = 2.1, 95%CI: 1.1-4.0 for estrogen-receptor positive breast tumor), but with no clear dose-response pattern. A similar overall pattern of results was seen in a record linkage study from Sweden in which occupation was determined from census data and exposure estimates were inferred from a job/exposure matrix devised for a previous study of leukemia and brain cancer (52). The same authors recently published a very large case-control study of breast cancer in which exposures were estimated from a job-exposure matrix based on an extensive magnetic field measurement program. This study found no effects of the exposure, regardless of age at diagnosis or estrogen receptor status of the tumor (50). Results from two other case- control studies also detected no associations with EMF exposure (40, 53). There have only been a few cohort studies on occupational EMF exposure in relation to female breast cancer risk. In a follow-up study of Norwegian female radio and telegraph operators, cumulative exposure was estimated based on ship type and number of years of employment (60). Increased risk of breast cancer was found in the exposed occupations as compared to the general population (Standardized Incidence Ratio [SIR] = 1.3, 95% CI: 1.1-1.6), but results were unstable due to a small number of exposed cases and possible confounding by exposure to light at night.
7 Evidence supporting the suppression of pineal production of melatonin by constant exposure to light at night, such as shiftwork, is consistent in both experimental and observational studies (34, 37). Even relatively small changes in ordinary light exposure during the late evening hours, e.g. experienced during evening or night work, can significantly reduce the plasma melatonin concentrations and affect the entrained phase of the circadian pacemaker (61). Recently, the hypothesized mechanisms for exposure to light at night increasing breast cancer risk have expanded from suppression of melatonin to phase shifting of endogenous circadian rhythm, and sleep disruption resulted from desynchrony of the master circadian pacemaker with sleep cycle and with peripheral oscillators in tissues throughout the body, collectively called "circadian disruption." In addition to melatonin suppression, circadian disruption may be linked to expression of clock genes, which are involved in cell-cycle regulation, DNA repair, and apoptosis (62). Exposure to light at night, by working at night or from home, in relation to risk of breast cancer has been studied extensively in the past two decades. In 2007, an International Agency for Research on Cancer (IRAC) Monographs Working Group reviewed the current evidence at the time and concluded that shift-work that involves circadian disruption is probably carcinogenic to humans, Group 2A, based on sufficient evidence in experimental animals for the carcinogenicity of light during the daily dark period and limited evidence in humans for the carcinogenicity of shift work that involves night work (63). The most convincing evidence in humans comes from two cohort studies (13, 64) and one nested case-control study (65), in which significantly increased risks of breast cancer were observed for long-term night shift work (more than 20-30 years), although no effects were found for shorter durations of shift work. All three studies were conducted in nurses. Recently, three population-based studies were published, which reported no association between shift work and breast cancer. In a register-based cohort study of the Swedish working population (66) recruited from 1960 and 1970 censuses, 1,148,661 women were followed up for all major cancers diagnoses from 1971 to 1989,
8 and 70 breast cancer cases were identified. Occupations were obtained from the censuses in 1960 and 1970. The classification of shift work relied on a job-exposure matrix constructed from a survey of living conditions in 1977-1981. Shift workers were defined as workers in occupation-industry combinations in which at least 40% of the workers either worked on rotating shifts with three or more possible shifts per week or worked any hour between 1300 and 0400 at least 1 day per week. The reference group was women who worked in occupation-industry combinations in which <30% of the workers experienced either of these employment patterns. The standard Incidence Ratio for breast cancer was 0.94, 95% CI: 0.74-1.18. Small number of cases, exposure misclassification, and lack of controlling for confounders were the major limitations of the study. Pesch et al. conducted a population-based case-control study in Germany (67), in which 857 breast cases and 892 controls were selected. Shift work experience, based on occupational history obtained from personal interviews, was defined as working full time between 2400 and 0500 for at least one year. No increased risk was associated with having ever worked on night shift (OR=0.96, 95% CI: 0.67-1.38). Odds ratio for night work for >= 20 years was 2.48 (95% CI: 0.62-9.99) based on only 12 exposed cases and 5 exposed controls. Pronk et al. (67), reported results from a prospective study that was conducted to investigate night shift work and breast cancer risk in a cohort of 73,049 Chinese women, recruited from seven Shanghai urban communities between 1996 and 2000 (68). A total of 717 incident breast cancer cases were identified through 2007. The assessment of shift work was based on two approaches, a job-exposure matrix and self-reported shift work histories. Relative risks were adjusted for a number of important risk factors for breast cancer. Breast cancer risk was not associated with ever working the night shift. (Based on JEM: hazard ratio=1.0, 95%: 0.9-1.2; based on self-reported history: hazard ratio=0.9, 95%: 0.7-1.1). Risk was not associated with frequency, duration, or cumulative amount
of night-shift work. This was the first study of night-shift work and breast cancer risk to have been carried out in a non-Western population. To date, the evidence from epidemiological studies of night shift work and breast cancer risks remain inconclusive. The conflicting results may be due to inconsistent definition and measurement of shift work among studies, to variation in duration of shift work across studies, or to true biological differences in the influence of various aspects of shift work on carcinogenesis. Possible variation in suppression of melatonin by light at night between different ethnic populations has been suggested as a possible explanation of the conflicting findings between the study in China and some studies in Caucasians. In an experimental study comparing the influence of eye color on melatonin suppression among Caucasians and Asians in Japan, Higuchi et al. (69) observed that the percentage of suppression of melatonin as a result of nocturnal light exposure was significantly smaller in the Asian (dark-eye color) compared with the Caucasian subjects (light-eye color). Aims and Structure of the Dissertation The purpose of this dissertation is to investigate two important occupational exposures, low frequency MF and night shift work, in relation to breast cancer risk in a large well- defined cohort of women textile workers in Shanghai, China. The textile industry is one of the world's largest employers of women. This industry often requires workers to perform tasks in close proximity to large machinery (such as spinning machines) or electric motors (such as sewing machine motors). Sewing machine operators have been reported repeatedly to experience high MF exposures (70). This study is based on a cohort of 267,400 women textile workers from 526 factories in Shanghai, China, who had originally been enrolled in a study of breast-self examination, and on women in the cohort who were subsequently included in a series of case-cohort studies of occupational exposures in relation to various cancers including breast cancer. All women in the cohort completed a baseline questionnaire eliciting data on their
10 reproductive and menstrual histories, lifestyle factors (e.g., tobacco and alcohol use), and other risk factors for breast cancer (e.g., family history of breast cancer). All breast cancer cases were identified by both active and passive case finding procedures, and their diagnoses were verified by review of pathological report or histological review of tissue slides as part of the trial. We collected work history data in the textile industry primarily from factory personnel records. Occupational MF exposure was characterized by constructing a quantitative job- exposure matrix, based on detailed job assignment data, coupled with exposure measurements that we made. Assessment of shift work was based on detailed historical shift work information obtained from 502 textile factories (where the cases and non-cases in this study worked) for various textile jobs. Dose-response relations for associations between breast cancer risk and the two exposures were estimated separately. Potential confounding of the associations by age, number of live birth, age at first live birth, and alcohol consumption was examined. A possible interaction between the MF and shiftwork exposures on risk of breast cancer was also assessed. The following chapters describe and summarize in detail the results of the studies that were conducted. Assessment of occupational exposure to magnetic fields in the Shanghai textile industry is described in Chapter 2. The possible associations between risk of breast cancer and occupational exposures to magnetic fields and shift work in the Shanghai textile industry are presented in chapters 3 and 4, respectively. The final chapter summarizes the findings from all analyses, and relates the results to current research on these topics.
11 Notes to Chapter 1: 1. Althuis MD, Dozier JM, Anderson WF, Devesa SS, Brinton LA. Global trends in breast cancer incidence and mortality 1973-1997. Int J Epidemiol 2005;34(2):405-12. 2. Bernstein L, Ross RK. Endogenous hormones and breast cancer risk. Epidemiol Rev 1993;15(l):48-65. 3. Dao TL. The role of ovarian steroid hormones in mammary carcinogenesis. In: Pike MC Siiteri PK, Welsch CW (eds). Hormones and breast cancer. (Banbury Report No. 8). Cold Spring Harbor, NY: Cold Spring Harbor Laboratory 1981;pp 218-295. 4. Kabuto M, Akiba S, Stevens RG, Neriishi K, Land CE. A prospective study of estradiol and breast cancer in Japanese women. Cancer Epidemiol Biomarkers Prev 2000;9(6):575-9. 5. Dorgan JF, Longcope C, Stephenson HE, Jr., Falk RT, Miller R, Franz C, et al. Relation of prediagnostic serum estrogen and androgen levels to breast cancer risk. Cancer Epidemiol Biomarkers Prev 1996;5(7):533-9. 6. Key TJ, Verkasalo PK, Banks E. Epidemiology of breast cancer. Lancet Oncol 2001 ;2(3): 133-40. 7. Easton DF. How many more breast cancer predisposition genes are there? Breast Cancer Res 1999;1(1): 14-7. 8. Stevens RG. Electric power use and breast cancer: a hypothesis. Am J Epidemiol 1987;125(4):556-61. 9. Wurtman RJ, Axelrod J. The Pineal Gland. Sci Am 1965;213:50-60. 10. Cart K, Dafau M. Gonadotropic hormones: biosynthesis, secretion, receptors, and actions. In SSC Yen and RB Jaffe (eds). Reproductive Endocrinology, Third edition. W.B. Saunders, Philadelphia 1991:pp 144-151. 11. Adashi E. The ovarian life cycle. In SSC Yen and RB Jaffe (eds).. Reproductive Endocrinology, Third edition. W.B. Saunders, Philadelphia 1991 :pp 202-204. 12. Davis S, Mirick DK, Stevens RG. Night shift work, light at night, and risk of breast cancer. J Natl Cancer Inst 2001 ;93(20): 1557-62. 13. Schernhammer ES, Kroenke CH, Laden F, Hankinson SE. Night work and risk of breast cancer. Epidemiology 2006; 17( 1): 108-11.