University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content P3-ADRENERGIC RECEPTOR GENE TRP64ARG POLYM ORPHISM AND OBESITY-RELATED CHARACTERISTICS AMONG AFRICAN AMERICAN W OM EN W ITH BREAST CANCER — AN ANALYSIS OF USC HEAL STUDY by Yaling Teng A Thesis Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment o f the Requirements for the Degree M ASTER OF SCIENCE (APPLIED BIOSTATISTICS AND EPIDEM IOLOGY) August 2002 Copyright 2002 Yaling Teng Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 1414890 UMI UMI Microform 1414890 Copyright 2003 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY O F S O U T H E R N CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 9 0 0 0 7 This thesis, written by YALffs/G* TlbJCr under the direction of h.9J:.....Thesis Committee, and approved by all its members, has been pre sented to and accepted by the Dean of The Graduate School, in partial fulfillment of the requirements for the degree of M oUex o f Science. D tax THESIS COMMITTEE Chair i Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Acknowledgements My greatest thanks are dedicated to Dr. Frank Gilliland, my advisor and thesis committee chair, for his guidance and support. Special thanks are given to other committee members, including Dr. Leslie Bernstein for her insightful comments on this thesis; and Dr. James W. Gauderman, for the sharing o f his expertise in statistical methods. I would also like to extend my gratitude to Drs. Azen and van Tournaut, who introduced me into the exciting field o f bio statistics. Ms. Jane Sullivan-Halley excellently managed the data for this study. Mary Trujillo is warmly acknowledged for her administrative support during the past years. This thesis is dedicated to my family for their love and encouragement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents Acknowledgements 1 1 List o f Tables iv Abstract vi Introduction 1 Methods 7 Results 15 Discussion 35 References 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Tables Table 1. Population Characteristics o f HEAL Study Subjects at 30 months after diagnosis. 16 Table 2. Pairwise correlation coefficients among various body composition metrics o f HEAL study subjects at 30 months after diagnosis. 18 Table 3-1. Univariate relationship between weight, BMI, waist, hip, waist- hip-ratio o f HEAL study subjects at 30 months after diagnosis and selected categorical variables. 19 Table 3-2. Univariate relationships between body fat mass, body fat percent, internal and external abdominal fat, and the ratio o f internal to external abdominal fat o f HEAL study subjects at 30 months after diagnosis (to be continued). 21 Table 4-1. Frequency o f obesity among HEAL subjects at 30 months after diagnosis and ADR/53 Trp64Arg polymorphism. 23 Table 4-2. Association o f the ADR/53 Trp64Arg polymorphism, sports activity and smoking with obesity (BMI>30Kg/m2) among HEAL patients at 30 months after diagnosis. 25 Table 4-3. Association o f the ADR/53 Trp64Arg polymorphism, sports activity and smoking with obesity (body fat>45.52%) among HEAL patients at 30 months after diagnosis. 25 Table 5. Weight and BMI change from adult usual status to 30 months after diagnosis in relation to ADR/53 Trp64Arg genotype and selected characteristics. 27 Table 6. The ADR/53 Trp64Arg polymorphism, smoking, chemotherapy and weight at 30 months post diagnosis. 29 Table 7. The ADR/53 Trp64Arg polymorphism, smoking, chemotherapy and BMI at 30 months post diagnosis. 29 Table 8-1. The ADR/53 Trp64Arg polymorphism, smoking and waist circumference (cm) at 30 months post diagnosis. 30 Table 8-2. The ADR/33 Trp64Arg polymorphism, smoking status and hip circumference (cm) at 30 months post diagnosis. 30 IV Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 9. The ADR/33 Trp64Arg polymorphism, smoking status and Waist- hip ratio at 30 months post diagnosis. Table 10-1. The ADR/33 Trp64Arg polymorphism, smoking and Body fat mass (Kg) at 30 months post diagnosis. Table 10-2. The ADR/33 Trp64Arg polymorphism, smoking and Body fat percent at 30 months post diagnosis. Table 11. The ADR/33 Trp64Arg genotype, vigorous activity and internal abdominal fat (IAF) (cm2) at 30 months post diagnosis. Table 12. The ADR|33 Trp64Arg genotype, chemotherapy, current tamoxifen use, smoking and external abdominal fat (EAF) (cm2) at 30 months post diagnosis. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Abstract We examined the effects o f p3-adrenergic receptor gene Trp64Arg polymorphism on multiple obesity-related characteristics among African American women with breast cancer. We included 176 women who were diagnosed with first primary breast cancer (stage 0-IIIa) while residing in Los Angeles County, who participated in an ongoing case-control study and who had no recurrence and secondary cancer up until about 30 months post diagnosis. W e found that at 30 months post diagnosis, obese patients (BMI>30 kg/m2) had significantly lower prevalence o f the Arg64 (variant) allele than non-obese patients (14.29% versus 27.36%, p<0.05). After adjustment for adult usual weight or BMI, women carrying an Arg64 allele were 6.04Kg lighter and had 2.17Kg/m2 lower BM I than those not carrying Arg64 allele (p<0.01 for both). W e conclude that ADRfS3 Arg64 variant allele is associated with decreased indexes o f obesity among African American women with breast cancer and may be associated with better prognosis. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. INTRODUCTION Breast cancer is the leading site o f cancer among women worldwide (Parkin, 1999). In the United States, breast cancer accounts for 29% o f all newly diagnosed cancers in women and is second only to lung cancer as cause o f cancer death in women (Thompson, 2000). Although mortality is declining in some developed countries including the United States due to early detection and improved adjuvant therapy (Brewster, 2001), a growing number o f women with resected breast cancer live in continuous threat o f recurrence. Tamoxifen and aromatase inhibitor are available for long-term prevention o f recurrence, however, innovative therapies are needed to further improve the breast cancer prognosis. Modification o f lifestyle factors that affect survival may be an important approach to prolong disease-free survival and reduce mortality, as well as to improve quality o f life during and after standard treatment regimes o f breast cancer. Obesity is an increasingly prevalent life style factor that appears to be an important prognostic factor for breast cancer (Chlebowski, 2002). Some studies have identified an association between increased BM I and advanced diagnostic stage of breast cancer (Ingram, 1989, Verreault, 1989, Daling, 2001). Higher BMI at diagnosis and weight gain after diagnosis have also been associated with higher recurrence rates as well as higher mortality in several studies (Herbert, 1988, Goodwin, 1990, Senie, 1992, Newman, 1997, Zhang, 1995, Demark-Wahnefried, 1997, Daling, 2001). Obesity was also found to be related to increased comorbidity from wound complexion and lymphedema development after breast cancer surgery 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Forouki, 1995), and endometrial cancer development after tamoxifen therapy (Bernstein, 1999). Obesity also has an unfavorable effect on the quality o f life. In a systematic review o f the current evidence correlating obesity and breast cancer outcome, Chlebowski (2002) concluded that body weight was strongly associated with clinical outcome in women with breast cancer, so he recommended clinical care targeting on weight control for obese breast cancer patients. It is o f special interest to study obesity and weight gain among African American women with breast cancer. Obesity is more common in African American women with breast cancer than that among white breast cancer patients (Rose, 2001), and they have the highest breast cancer mortality rates o f any population subgroup (Ries, 1997). Many factors may contribute to this high mortality, such as lower social economic status, younger age at diagnosis, higher diagnostic stage (Hunter, 1993) and more aggressive tum or characteristics (Elledge, 1994), as compared with Caucasians. Jones (1997) reported that almost one third o f the racial difference in stage at diagnosis could be explained by the greater prevalence o f severe obesity among African American women. In addition to affecting the diagnostic stage, the high obesity prevalence o f African American women at diagnosis and weight gain after diagnosis may act as independent factors affecting their recurrence rates and overall survival. The high prevalence o f obesity among African American patients may result either from higher weight at diagnosis or more weight gain after diagnosis or both. Weight gain commonly occurs in breast cancer patients during the treatment period, 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. particularly among those who receive adjuvant chemotherapy (Demark-Wahnefried, 2001). Other factors that may contribute to weight gain in the early years after diagnosis include tamoxifen use, premature menopause due to treatment, and increased energy intake and/or reduced energy expenditure (Chlebowski, 2002). It is not clear whether African American women with breast cancer tend to gain more weight after diagnosis than white patients under the ‘obesity stress’ associated with diagnosis and treatment o f breast cancer. Under a similar ‘obesity stress’ in midlife, African American women are more likely to become overweight (Matthews, 2001). Further weight gain after breast cancer diagnosis is amenable to change, which may constitute a new therapy approach to prevent recurrence and prolong survival. From a public health perspective, it is important to find the sub-population who is predisposed to more weight gain than others under the same treatment conditions, if weight control becomes a target o f medical care o f breast cancer patients. Both life style and genetic factors may explain part o f the ethnic difference in weight gain under ‘obesity stress’. It is believed that obesity is the result o f an interaction between genetic predisposition and environmental factors, although the evidence for these interactions is weak (Sorensen, 2001). The P3-Adrenergic receptor gene has been proposed as a candidate “obesity” gene, because it is predominantly expressed in adipose tissue and regulates lipid metabolism and thermogenesis. Its functional impairment may lead to obesity through its effect on metabolism o f fat tissue (Oizumi, 2001). A miss-sense mutation in codon 64 o f this gene with a replacement o f trytophane to arginine (Trp64Arg) has been identified 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and studied in relation to obesity-related phenotypes (Walston, 1995, Gagnon, 1996, Corella, 2001), noninsulin-dependent diabetes mellitus characteristics (Walston, 1995, Yoshioka, 1996, Oizumi, 2001) and breast cancer risk (Huang, 2001). Pima Indians and Japanese populations have been most extensively studied because o f their higher frequency o f this polymorphism. A review about the relationship between this polymorphism and obesity-related phenotypes identified a link between obesity and this polymorphism in 13 o f 28 studies (Arner, 1999). A meta-analysis by Allison (1998) failed to find a significant association between this polymorphism and BMI, or to find any significant evidence o f effect heterogeneity by ethnicity (not including African American) or diabetic status. A recent large cross-sectional study o f the Japanese population showed that the Arg/Arg genotype, but not Trp/Arg genotype, was associated with the risk o f Type 2 diabetes and obesity (Oizumi, 2001). Possible explanations for the conflicting results include population difference in ethnicity, obesity status, age structure and exclusion criteria, unknown confounding and/or modification effects o f other environmental factors or other genetic factors, sampling error and measurement error. Clearly more studies are needed to clarify this association. Limited information is available regarding the functional importance o f the ADRf33 Trp64Arg polymorphism. In the first study on this polymorphism (Walston, 1995), subjects with the Arg64 allele were found to have a lower adjusted resting metabolic rate (P - 0.14, by analysis o f covariance). In rodents, the principal function o f the 03-adrenergic receptor was found to mediate lipolysis in adipocytes 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and thermogenesis in brown adipose tissue (Amer, 1999). In vitro experimental studies o f this polymorphism have evaluated the ligand binding and adenyl cyclase activation characteristics among the 3 genotypes (Allison, 1998), and have used specific adrenergic receptor agonists (Umekawa, 1999). It has been reported that the wild type allele (Trp64) and variant allele (Arg64) do not differ in the aforementioned two characteristics in adipose tissue (Candelore, 1996). However, in Chinese hamster ovary cells and human embryonic kidney cells, the stimulation of the Trp64Arg variant 133-Adrenergic receptor resulted in a lower maximal level o f adenyl cyclase relative to the wild type polymorphism (Pietri-Rouxsel, 1997). It is not clear whether this result can be generalized to human fat tissues, the chief location where (33-ADR functions. It was suggested that P3-Adrenergic receptor was functionally more active in omental than in subcutaneously adiposities (Tavenier, 1996). Umekawa et al (1999) investigated whether or not lipolysis in human omental adiposities induced by three potent and selective human P3-ADR agonists ( isoproterenol, CGP12177 and L755,507) was affected by the ADRJ33 Trp64Arg polymorphism. It was found that both the heterozygous and homozygous variant genotype had a lower median effective concentration (EC50) and lower responsiveness to L755, 507 compared with the wild-type. However, the other two agonists did not show any effect (isoproterenol more effective, CGP12177 less effective than L755, 507). In summary, the experimental evidence tends to support the hypothesis that Arg64 allele is associated with more efficient energy expenditure, or higher risk o f obesity under constant energy intake. However, these results 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. warrant careful interpretation because they are based on small sample sizes and the measured physiologic reactions have large variance. In addition, the specificity o f the available agonists o f the 3P-adrenergic receptor may have affected the study results. Thus more convincing evidence (such as in vivo results) is needed to unequivocally explain the observed relationship between the ADRf33 Trp64Arg polymorphism and obesity-related phenotypes. There has been only one previous study o f the ADR/33 Trp64Arg polymorphism and obesity-related traits among African American women (Lowe, 2001). A negative association was found between recessive inheritance o f the Arg64 allele and BM I and leptin level, which alluded some difference on this polymorphism’s effect among different ethnicity. This polymorphism has not been studied in relation to breast cancer prognosis. Based on the above studies, we hypothesize that the ADR/33 Trp64Arg might affect obesity status, body fat distribution and weight change among African American women with breast cancer, and thereby modulate breast cancer survival. To test this hypothesis, we examined data from the HEAL (Health, Eating, Activity and Life style) study, a population- based cohort study o f women with breast cancer. We assessed the relationship between the ADR/33 Trp64Arg genotype and weight, BMI, several metrics o f central obesity (waist, hip, waist-hip ratio, external and internal abdominal fat), obesity status, weight and BM I change between usual adult status at 30 months post diagnosis o f breast cancer among African American women, with consideration of the effects o f dietary intake o f energy, physical activity, smoking and treatment. 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. METHODS The HEAL (Health, Eating, Activity and Lifestyle) study is a multi-center, longitudinal study on breast cancer patients. Participants at the University o f Southern California (USC) center were patients who participated in one o f two concurrent case-control studies o f breast cancer risk factors, the W omen’s Contraceptive and Reproductive Experiences (CARE) study o f invasive breast cancer and the study o f in situ breast cancer (INSITU) diagnosed during the same time period. Subjects eligible for the HEAL Study were African-American women aged 35 to 64 years at breast cancer diagnosis, who were diagnosed with a first primary breast cancer (stage 0 to IIIA) between 5/95 and 5/98, and resided in Los Angeles County at the time o f diagnosis. A total o f 367 African American women who participated in the CARE or INSITU study satisfied the above criteria. Each of the participants in the HEAL Study signed an informed consent and completed an interview. • Baseline data collection Baseline information was collected in both the CARE and INSITU study during in-person interviews using a comprehensive questionnaire. All baseline interviews were finished within 18 months o f diagnosis, with mean interval o f 6.0 months and a median o f 4.8 months. Questionnaire included information about age, age at diagnosis, height, menstrual history and menopausal status, reproductive history, family history o f breast, ovarian and endometrial cancer, smoking history, and lifetime physical activity. Weight at ages o f 18, 35 and 50 and 5 years prior to 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. diagnosis was obtained by self report during the interview. Adult usual weight was defined as the most recent weight among the several collected weights (at 18, 35 and 50 years old). Adult usual Body Mass Index (BMI) was defined as adult usual weight (kg) divided by squared adult height (M2). • 30-month Follow up The follow-up interview was targeted to take place 24 months (+/-3 months) after the initial interview for the W omen’s CARE Study or the in situ study, which occurred at approximately 6 months post diagnosis. O f 367 eligible women, 262 completed the 30-month follow-up interview. Reasons for nonparticipation included: inability to locate participants after CARE and INSITU baseline interview (64 women), death or severe illness (21 women), and refusal at time o f follow-up contact (20 women). In-person interviews were conducted at home or another site that the participant preferred. Additional data were collected on the wom an’s breast disease and treatment, other medical conditions and hospitalizations since her breast cancer diagnosis, and her current menstrual status, medication use, HRT use, recent diet, physical activity patterns, and tobacco use during this interview. Following the interview, anthropometric characteristics were measured and blood samples were drawn. Tobacco smoke exposure data Structured questions were asked about current and past smoking habits. Each woman was classified into one o f the 3 categories: never smoker, former smoker and a current smoker. Using the self-reported date o f smoking initiation and cessation, 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. smoking status was also estimated for the time when adult usual weight was assessed (usual adult smoking status). M ost women did not change their smoking status from the usual adult smoking status” to 30 months post diagnosis; however, 27 women who were defined as current smokers at quit smoking during the interval, and became “former smokers” at 30 months post diagnosis. We categorized women into four categories o f smoking status before and after diagnosis, i.e., never-never, former-former, current-current and current-former. Physical activity The HEAL questionnaire was designed to assess physical activity o f the study subjects during the past month (at month 30). Study subjects were asked to report the type, frequency and duration o f physical activities performed regularly using a list o f 30 activities; they also reported other activities that were not included on the list. The 30 activities listed included common sports like running, jogging, walking, dancing, swimming and Yoga, and household activities like housecleaning, and home repair Metabolic equivalents (METs) in Kcal/kg/hour was assigned to each physical activity according to recent “Ainsworth Compendium” (1993). Weekly average energy expenditure was calculated for each subject for sports activities, vigorous activities and all activities by summing average M ET-hour for each component activity per week. Weekly average energy expenditure on sports and vigorous activity were coded as categorical variables, with their corresponding quartiles or tertiles as the cutpoints. The physical activity one year before diagnosis 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that was collected at baseline was included in the analysis in a similar way as the measure at the 30 month follow-up interview. Dietary information Dietary intake o f various macro- and micronutrients during the past month (at month 30) was collected using a modified version o f a valid quantitative food frequency questionnaire (FFQ) that was developed for the W omen’s Health Initiative (WHI) study by Fred Hutchinson Cancer Research Center (Patterson RE, 1999). The FFQs were mailed to the participants to be completed prior to the interview. The returned FFQ data was optically scaned and reviewed for errors and outliers. Percent daily calories from fat, percent daily calories from carbohydrates, percent calories from protein, and daily caloric intake was assessed. Treatment data Diagnostic and treatment information were collected from self-reports during the 30 month follow-up interview, medical record abstraction by the USC Cancer Surveillance Program for Los Angeles County, a member o f the SEER program o f National Cancer Institute (NCI). Treatment type for primary breast cancer was categorized as surgery, radiation and/or chemotherapy. Information on tamoxifen use was obtained from 2 sources including the medical abstract, and information collected by reviewing medication containers during the 30-month follow-up interview. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Body composition assessment The body composition assessment battery included the following standardized measurements and procedures. Weight, waist and hip circumference were measured at the 30-month interview. Weight was measured to the nearest 0.1 kg. Waist and hip circumferences were measured using a steel tape to the nearest 0.1 cm. Waist circumference was measured 1 cm superior to the iliac crests, hip circumference at the level o f the maximal posterior protrusion o f the buttocks. All women were measured twice, the mean o f tw o measurements was used in the analysis. Current BM I and waist-hip ratio (WHR) were calculated. Following the methods described in Segal’s paper (1988), bioelectrical impedance analysis (BIA) was used to measure the body fat-free component. BIA measurements were taken at the time o f the MRI (Magnetic Resonance Image) appointment at the Los Angeles County - University o f Southern California (LAC- USC) Imaging Science Center or at the home setting, if the respondent refused to do an MRI. BIA measurements were made using a single frequency (50 kHz) BCA analyzer and standard electrode placements after a 4-hour fast and the subjects have emptied their bladders. Each woman was measured twice, the mean o f two measurements was used to calculate lean body mass with a cross-validated equation as follows. Lean body mass (Kg) = 14.59453 + 0.00108*height (cm)**2 + 0.23199*weight (kg) -0.06777*age - 0.02090*resistance. 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Body fat mass was obtained by subtracting lean body mass from weight, and body fat percent was the percent o f body fat mass compared to weight. External and internal abdominal fat was measured by M RI technique at the LAC-USC Image Science Center. Single slice Tl-w eighted images with 1 cm thickness was acquired axially at the umbilicus using a 1.5 Tesla General Electric (Milwaukee, WI) LX Echospeed M RI scanner. M RI parameters included: TE (repeat time), 400; TE (time to echo), 16; 192 Phase encodes; and NEX (number o f excitations), 2. Intra-abdominal fat and extra-abdominal fat regions were outlined electonically by technicians on the MRI viewing monitor, reported in mm2, and photographed. Blood sample collection, DNA extraction, and polymorphism determination Blood samples were obtained from 233 o f 262 women who completed 30- month interview and processed by standardized procedures. Genomic DNA was prepared from buffy coat pellets using Qiagen kit (QIAGEN, Valencia, CA). The amount and concentration o f extracted DNA was measured and a standard dilution was prepared for polymorphism detecting. The ADRP3 Trp64Arg polymorphism is a T to C transversion at codon 64. Genotypes were determined using 5’ nuclease assay that uses specific fluorogenic TaqMan probes (Thetagen, Seattle, Washington). The PCR primers were: (sense) 5’-CGCCCAATACCGCCAACAC3’; and (anti-sense) 5’-CCACCAGGAGTCCCATCACC-3’. The sense and anti-sense PCR primers were used at a final concentration o f 900nM in 25pi PCR solution. Both fluorescence fluorogenic wild-type and variant allele-specific probes were complementary to their 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. corresponding anti-sense strands and were labeled with the TAM RA quencher at the 3’ end, with the 6-FAM reporter dye and the VIC reporter dye at the 5’ end. The wild type and variant fluorogenic probes were used at a final concentration o f lOOnM and detected alleles from approximately 40ng o f genomic DNA template. Temperature cycling was performed as follows: 50°C for 2 minutes and 95°C for 10 minutes, followed by 40 cycles o f 95°C for 15 s and 62°C for 1 minute. The data were collected and automatically analyzed with ABI Prism 7700 Sequence Detection System version 1.6.5. All Taqman genotyping calls were confirmed by sequencing several samples o f each allele. Statistical analysis Descriptive statistics were first calculated for variables that were characteristic o f the study population. Pair-wise correlation analyses were performed among multiple obesity-related variables we collected. General linear and logistic regression methods were used to test the main effects o f the ADRfi3 Trp64Arg polymorphism on obesity-related characteristics (weight, BMI, waist and hip circumference, WHR, total body fat mass, body fat percent, external and internal abdominal obesity, weight change and BM I change from adult usual status) at the significance level o f 0.05. Two genetic models were fitted for the ADR/33 Trp64Arg allele. As in most o f the studies on this polymorphism, we used the dominant model for the variant allele (Arg64 allele). For this model, women were assigned to the variant genotype if one Arg64 allele was present. We were unable to fit the recessive 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. model because the frequency o f the homozygous variant genotype was low. The additive model did not fit the data because there was not a significant consistent change o f the effect with the number o f copies o f the variant allele. Potential confounding covariates we considered included, adult usual weight and BMI, age, diagnostic stage, menopausal status, treatment type, current tamoxifen use, smoking status, daily energy intake, percent o f daily energy intake from fat, total leisure time activity 1 year before diagnosis, sports activity and vigorous activity during the past year. Variables for energy intake and physical activity were categorized using their corresponding quartiles or tertiles as outpoints. Potential confounding variables entered into the model based on review o f the literature, univariate analyses, and Likelihood Ratio Test. For the covariates that were coded as categorical variables, the null hypothesis that obesity-related traits did not differ across levels o f these variables were tested, adjusted Least Square (LS) means and corresponding 95% confidence intervals were also reported. Linear trend was tested for selected covariates. Modification effect o f the covariates on gene-obesity related traits relationship was tested by comparing the model included the corresponding interaction term and the model without it at the significance level o f 0.10. Analyses were performed with the Statistical Analysis Software (SAS) 8.0 (SAS Institute, Cary, NC) (1999). 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. RESULTS O f 262 women who completed the 30 month interview, 233 subjects provided a blood sample. 20 subjects were excluded from the statistical analysis for recurrence or second primary cancer. W e also excluded 34 subjects who had diabetes mellitus, 2 patients with liver disease and 1 patient who had no weight measurement at the 30 month follow-up interview. Our final sample size was 176 for the analysis o f the effect o f ADRfi3 Trp64Arg polymorphism on weight and BMI. The analyses on other obesity-related endpoint variables were based on various smaller sample sizes due to incomplete measurement o f these variables. Population characteristics The characteristics o f the 176 subjects are displayed in Table 1. At 30 months after diagnosis, these women were aged 51.7±7.4 years old, weighed 80.8±20.9 kg, and had BM I o f 29.6±7.3 kg/m2. According to the corresponding definitions o f W ord Health Organization (WHO), the prevalence o f overweight (BMI>25Kg/m2), obesity (BMI>30Kg/m2 ), severe obesity (BMI>35Kg/m2) and morbid obesity (BMI>40Kg/m2) among the study subjects were 73.9%, 39.8%, 18.8% and 10.2%, respectively. Their adult usual weight was 68.2±15.0 kg, and the average BMI was 25.0±5.0 kg/m2. Over the period between adult weight and 30 months after diagnosis, these women experienced an average weight gain o f 12.6±13.3 kg. G enotype and allele frequencies Among the 176 women genotyped, 3 women had the homozygous variant type (Arg64Arg) and 36 women had the heterozygous genotype (Trp64Arg), with a 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. variant frequency o f 11.9%. The genotypes didn’t show significant departure from Hardy-Weinberg equilibrium. Table 1. Population Characteristics of HEAL Study Subjects at 30 months after diagnosis. Continuous variable N Mean (stddev.) Continuous Variable N Mean (stddev.) Age (years) 176 51.7(7.4) Body fat mass (Kg) 169 34.3(13.9) Weight (Kg) 176 80.8 (20.9) Body Fat Percent (%) 169 41.1 (6.6) Height (cm) 176 165.0 (7.1) Intra abdominal fat (cm2) 115 91.6(46.5) BMI (kg/m2 ) 176 29.6 (7.3) Extra abdominal fat (cm2 ) 115 385.3(187.4) Waist (cm) 171 94.8(15.2) Total abdominal fat (cm2 ) 115 476.9(215.3) Hip (cm) 171 112.9(15.4) Ratio of IAF to EAF A 115 27.7 (24.5) Waist-hip ratio 171 0.84 (0.06) Categorical variable N % Categorical variable N % Overweight ADRfi3 gene No (BMI < 25Kg/m2 ) 46 26.14 Trp64Trp 137 77.84 Yes(BMI > 25Kg/m2) 130 73.86 Trp64Arg, Arg64Arg 39 22.16 Obese Age group (Yrs) No (BMI < 30Kg/m2 ) 106 60.23 38-45 44 25.00 Yes(BMI > 30Kg/m2) 70 39.77 46-50 39 22.16 51-55 37 21.02 Severe Obese 56-60 30 17.05 No (BMI < 35Kg/m2 ) 143 71.25 61-66 26 14.77 Yes(BMI > 35Kg/m2) 33 18.75 Still having periods Morbid Obese Yes 34 19.32 No (BMI < 40Kg/m2 ) 158 89.77 No 142 80.68 Yes(BMI > 40Kg/m2) 18 10.23 Tamoxifen Use Diagnostic stage No 117 66.48 In situ 30 17.05 Yes 59 33.52 local 81 46.02 regional 65 36.93 Treatment Surgery only 59 33.52 Smoking status Surgery+chemotherapy 47 26.70 Never smoke 94 53.41 Surgery+radiation 37 21.02 Formerly smoking 54 30.68 Surgery+chem# +rad* 31 17.61 Currently smoking 28 15.91 No surgery, chem# or rad* 2 1.14 Notes: A : IAF and EAF indicate intra and extra abdominal fat; * and # indicate radiation and chemotherapy, j g Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Correlation coefficients among various obesity-related characteristics Table 2 shows the pairwise correlation coefficients among the 12 continuous endpoint variables representative o f body mass, body fat percent and central obesity. BMI, total abdominal fat, body fat percent, external abdominal fat (EAF), waist circumference, and hip circumference were highly correlated with each other (|R|>0.60). Internal abdominal fat (IAF) was highly correlated with total abdominal fat, body fat percent, waist, and BM I (|R|>0.60). Waist to hip ratio (WHR) was most highly correlated with waist circumference and IAF. The correlation between W HR and IAF was not as high as those among BMI, body fat percent, waist and EAF. Body composition at 30 month post diagnosis The ADR/33 Trp64Arg genotype, and potential covariates (age, diagnostic stage, treatment, menopausal status, smoking, energy intake and physical activity) were studied in univariate analyses in relation to the multiple obesity-related variables at 30 months post diagnosis (Table 3-1 and 3-2). The ADR/33 Trp64Arg genotype was not a significant predictor o f any one o f the studied endpoint variables in these analyses. W omen who currently smoked had lower value o f weight and BMI than those who had never smoked. Age and menopausal status were significant univariate predictors for waist and hip circumference, body fat mass, body fat percent and EAF, but not for weight, BMI, W HR or IAF. Diagnostic stage and current tamoxifen use were not associated with any one o f the endpoint variables in univariate analyses. Chemotherapy was associated with higher external abdominal fat. Total energy intake was not associated with any one o f the endpoint variables, 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 2. Pairwise correlation coefficients among various body composition metrics of HEAL study subjects at 30 months after diagnosis. MRIIAF MRIEAF IAF EAF+IAF EAF IAF IAF+EAF Waist Hip WHR FAT% Body fat mass BMI Weight MRIIAF 1 MRIEAF 0.52 *** 1 EAF+IAF 0.67 *** 0.98 *** 1 IAF/EAF 0.38 *** -0.33 *** -0.20 * 1 IAF/(IAF+EAF) 0.48 *** -0.40 *** -0.24 ** 0.90 *** 1 Waist 0.68 *** 0.88*** 0.92*** -0.12 -0.13 1 Hip 0.57 *** 0.91 *** 0.91 *** -0.19 * -0.27 ** 0.90 *** 1 WHR 0.44 *** 0.24 * 0.30 ** 0.11 0.23 * 0.51 *** 0.09 1 FAT% 0.61 *** 0.87 *** 0.89 *** -0.17 -0.20 * 0.82 *** 0.84 *** 0.24 * 1 Body fat mass 0.62 *** 0.92 *** 0.94 *** -0.16 -0.23 * 0.91 *** 0.95 *** 0.22 * 0.90 *** 1 BMI 0.62 *** 0.92 *** 0.94 *** -0.17 -0.22 * 0.90 *** 0.92 *** 0.24 * 0.91 *** 0.98 *** 1 Weight 0.61 *** 0.89 *** 0.91 *** -0.14 -0.21 * 0.90 *** 0.95 *** 0.19 * 0.83 *** 0.98 *** 0.94 *** 1 Notes: 1. * 0..01<p<0.05, ** 0.001<p<0.01, *** P<0.001 for pairwise correlations; 2. MRIIAF, MRIEAF: internal and external abdominal fat measured by MRI; EAF+IAF, total abdominal fat; IAF/EAF, the ratio of IAF to EAF; IAF/(IAF+EAF), the ratio of IAF to total abdominal fat; Body fat mass: total body fat measured by Bioelectric Impedance Assay (BIA); FAT%, the percent of body fat mass in total weight; BMI, body mass index (Kg/m2); WHR, waist hip ratio. oo Table 3-1. Univariate relationship between weight, BMI, waist, hip, waist-hip-ratio of HEAL study subjects at 30 months after diagnosis and selected categorical variables (to be continued). Variable N Weight (Kg) BMI(kg/m2 ) N Waist(cm) Hip (cm) Waist-hip Mean±Stdde Mean±Stdde V V Mean±Stddev Mean±Stddev ratio ADR03 gene Trp64Trp 137 81.6 ± 21.7 30.0 ±7.5 132 95.6 ± 15.6 113.7 ± 15.8 0.84 ± 0.07 Trp64Arg, Arg64Arg 39 77.9 ± 17.4 28.2 ± 6.0 39 92.0 ± 13.5 110.2 ± 13.9 0.83 ± 0.06 P 0.34 0.17 0.19 0.21 0.63 Age group (Years) 38-45 44 79.2 ± 19.1 29.1 ±6.6 41 91.3 ± 12.9 109.2 ± 13.0 0.84 ± 0.07 46-50 39 83.0 ± 22.2 30.8 ±8.1 38 94.8 ± 16.3 115.4 ± 16.9 0.82 ± 0.06 51-55 37 83.6 ±21.7 29.7 ±7.2 37 97.8 ± 16.5 115.7 ± 16.4 0.84 ±0.06 56-60 30 85.1 ±23.0 31.5 ± 7.9 29 99.7 ± 16.4* 116.9 ±16.5* 0.85 ±0.07 61-66 26 71.2 ± 15.6 26.6 ±5.3 26 90.4 ± 11.5 106.6 ± 11.4 0.85 ± 0.06 P for global test 0.08 0.10 0.06 0.03 0.30 Diagnostic stage In situ 30 79.9 ±21.2 29.0 ±7.5 29 93.5 ± 15.9 113.2 ± 17.4 0.83 ±0.05 local 81 79.7 ± 19.3 29.5 ±7.0 77 94.2 ± 14.5 111.7 ± 13.7 0.84 ± 0.07 regional 65 82.5 ± 22.8 30.1 ±7.6 65 96.0 ± 15.8 114.2 ± 16.5 0.84 ± 0.07 P for global test 0.70 0.77 0.68 0.62 0.47 Still having periods Yes 34 75.9 ± 17.9 28.5 ±6.2 34 89.5 ±12.8 107.2 ± 12.7 0.84 ± 0.07 No 142 81.9 ± 21.4 29.9 ±7.5 137 96.1 ± 15.5* 114.3 ± 15.7* 0.84 ± 0.06 P 0.13 0.29 0.02 0.02 0.69 Tamoxifen Use No 117 80.9±21.3 29.7 ±7.2 112 94.6 ± 14.9 113.0 ± 15.6 0.84 ± 0.06 Yes 59 80.6 ± 20.2 29.5 ±7.4 59 95.2 ± 15.8 112.8 ± 15.3 0.84 ± 0.07 P 0.93 0.86 0.80 0.95 0.59 Treatment Surgery only 59 80.1 ±20.5 29.4 ± 7.2 57 94.0 ± 15.0 112.1 ± 14.5 0.84 ± 0.07 Surgery+chem* 47 81.1 ±25.5 29.1 ±6.7 45 94.7 ± 14.4 114.2 ± 19.4 0.84 ± 0.07 Surgery+radiation 37 80.3 ± 18.9 30.1 ±8.6 37 95.4 ± 17.9 112.7 ± 14.0 0.84 ± 0.06 Surgery+chem# +radA 31 82.2 ± 16.5 30.4 ±5.9 30 96.3 ±17.8 113.5 ± 12.3 0.85 ± 0.07 No surgery ,chem#,radA 2 63.1 ± 10.7 24.3 ±4.7 2 81.8 ± 19.4 101.5 ± 18.4 0.80 ±0.05 P for global test 0.78 0.77 0.74 0.81 0.79 Smoking status Never smoke 94 83.2 ±22.5 30.5 ±7.6 92 95.8 ± 15.8 114.1 ± 16.0 0.84 ± 0.06 Formerly smoking 54 81.0 ± 19.8 29.6 ±7.3 52 95.6 ± 15.0 113.9 ± 14.5 0.84 ± 0.06 Currently smoking 28 72.2 ± 14.7* 26.8 ±4.9* 27 89.6 ± 12.4 107.0 ± 14.0* 0.84± 0.08 P for global test 0.05 0.06 0.16 0.09 0.99 P for trend test 0.02 0.02 0.11 0.07 0.94 Notes: 1. Mean values for each category were compared to the lowest level of the same categorical variable, * indicates a p value of equal or lower than 0.05. 2. M and A indicate chemotherapy and radiation, respectively. 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 3-1. Univariate relationships between weight, BMI, waist, hip, waist-hip-ratio of HEAL study subjects at 30 months after diagnosis and selected categorical variables (Cont’d). Variable N Weight (Kg) BMI(kg/m2 ) N Waist(cm) Hip(cm) Waist-hip Mean±Stdde Mean±Stdde v v Mean±Stddev Mean±Stddev ratio Total energy intake(kcal) Q1 [198.8, 874.1) 42 82.2 ± 20.0 29.9 ±6.6 39 93.9 ± 13.1 113.3 ± 14.4 0.83 ± 0.07 Q2 [874.1, 1306.7) 45 79.1 ±22.9 29.9 ±8.1 44 94.2 ± 16.5 111.7 ± 16.9 0.84 ± 0.06 Q3 [1306.7,1732.2) 43 83.2 ±22.3 29.8 ±7.3 43 96.7 ± 16.3 115.3 ± 16.5 0.84 ± 0.06 Q4 [1732.2, 12364.1) 46 78.8 ± 18.4 29.0 ±7.0 45 94.2 ± 14.6 111.4 ± 13.8 0.84 ± 0.07 P for global test 0.69 0.93 0.81 0.63 0.78 P for trend test 0.66 0.58 0.78 0.83 0.44 Energy from fat (%) Q1 [18.4, 33.6) 44 77.6 ± 18.3 28.5 ±6.8 42 91.8 ± 13.5 109.3 ±12.5 0.84 ± 0.07 Q2 [33.6, 38.7) 44 79.9 ±20.2 29.4 ±6.6 42 93.5 ± 12.8 114.0 ± 13.5 0.82 ± 0.06 Q3 [38.7, 43.4) 44 79.2 ± 20.0 29.1 ±6.9 44 94.7 ± 16.4 110.7 ± 15.2 0.85 ±0.07 Q4 [43.4,61.4) 44 86.3 ±24.3 31.6 ±8.4* 43 99.0 ± 17.0* 117.5 ± 18.8* 0.84 ± 0.06 P for global test 0.23 0.19 0.15 0.06 0.11 P for trend test 0.07 0.06 0.03 0.05 0.28 Total leisure time Activity 1 Yr before Diagnosis (MET.hrs/week) 0 99 83.5 ±23.7 30.5 ±8.1 97 96.7 ± 16.8 114.6 ± 17.7 0.84 ±0.07 (0,9] 26 75.8 ± 17.7 27.9 ± 5.7 26 91.0 ± 13.0 110.8 ±13.7 0.82 ± 0.05 (9,20] 22 85.2 ± 16.0 31.6 ± 6.6 22 97.0 ± 13.5 114.4 ± 11.7 0.85 ± 0.07 (20, 27 73.1 ± 12.7* 26.5 ±4.3* 26 89.5 ± 10.0* 107.6 ±7.9* 0.83 ±0.07 P for global test 0.06 0.02 0.08 0.18 0.35 trend 0.08 0.05 0.07 0.08 0.47 Sports activity level After Diagnosis (MET.hrs/week) <3.0 84 81.9 ± 23.8 30.4 ± 8.2 82 96.0 ± 17.1 113.2 ± 17.9 0.85 ± 0.07 3.0-8.0 34 80.2 ± 15.0 29.5 ± 5.7 32 95.6 ± 11.6 113.1 ± 11.0 0.84 ± 0.05 8.0-20.0 29 81.8 ±22.6 29.7 ±7.5 29 94.3 ± 15.0 112.6 ± 16.4 0.84 ± 0.07 >=20.0 29 77.0 ± 15.8 27.6 ±5.3 28 90.6 ± 12.5 112.1 ± 11.0 0.81± 0.05* P for global test 0.74 0.38 0.42 0.99 0.03 P for trend test 0.37 0.11 0.12 0.72 0.01 Vigorous activity level After Diagnosis (MET.hrs/week) 0 99 82.4 ± 22.9 30.3 ±7.9 97 96.4 ± 16.7 113.6 ± 17.3 0.85 ±0.07 <5.0 40 79.6 ± 18.8 29.5 ±6.5 38 94.9 ± 12.9 112.0 ± 12.8 0.84 ± 0.06 >5.0 37 77.6 ± 16.9 28.1 ±6.0 36 90.7 ± 12.2 111.9± 12.4 0.81± 0.05* P for global test 0.45 0.27 0.16 0.78 0.01 P for trend test 0.21 0.11 0.06 0.51 0.004 Notes: Mean values for each category were compared to the lowest level of the same categorical variable, * indicates a p value of equal or lower than 0.05. _ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 3-2. Univariate relationships between body fat mass, body fat percent, internal and external abdominal fat, and the ratio of internal to external abdominal fat of HEAL study subjects at 30 months after diagnosis (to be continued). Variable N Body Fat Body Fat N IAF EAF IAF/EAF Mass (Kg) (% ) (cm2 ) (cm2 ) (% ) ADR/13 gene Trp64Trp 130 35.1 ± 14.6 41.4 ±6.9 88 95.3 ±49.8 391.6 ± 190.6 28.5 ±27.3 Trp64Arg, Arg64Arg 39 31.9 ± 11.4 40.0 ±5.5 27 79.7 ±31.7 365.0 ± 178.7 24.8 ± 11.3 P 0 .2 2 0.23 0.08 0.49 0.35 Age group (Yrs) 38-45 40 32.2 ±12.6 39.5 ±7.0 26 89.2 ± 4 8 .9 ' 344.4 ± 186.4 35.3 ±45.2 46-50 38 35.9 ± 15.0 42.1 ±6.5 27 82.8 ±46.8 405.7 ± 181.1 21.3 ±7.6* 51-55 36 35.7 ± 15.0 40.9 ±7.1 24 94.2 ±41.8 436.7 ±223.1 23.3 ±9.6 56-60 29 38.6 ± 15.0 43.5 ±6.0* 19 97.6 ± 60.2 410.7 ± 188.4 24.4 ± 14.2 61-66 26 28.6 ±9.4 39.6 ± 5.6 19 98.2 ±34.1 322.2 ± 127.3 35.0 ± 18.1 P for global test 0.06 0.07 0.84 0.17 0.13 Diagnostic stage In situ 29 33.8 ± 14.7 40.3 ±7.7 2 0 83.8 ±44.8 350.6 ± 195.3 34.8 ±51.6 local 76 33.4 ± 12.7 41.0 ±6.2 52 93.5 ±45.3 365.8 ± 161.6 27.8 ±13.9 regional 64 35.7 ± 15.1 41.6 ± 6 .8 43 92.9 ±49.4 425.1 ±209.1 24.2 ± 12.3 P for global test 0.60 0 .6 8 0.72 0.18 0.29 Still having periods Yes 34 30.7 ±11.6 39.4 ±5.9 27 84.7 ±46.2 318.9 ± 160.2 34.0 ± 44.0 No 135 35.2 ± 14.4 41.5 ± 6 .8 88 93.7 ±46.7 405.7 ± 191.2* 25.7 ± 13.8 P 0.09 0.09 0.38 0.04 0.35 Tamoxifen Use No 1 10 34.5 ± 14.1 41.2 ±6.7 75 88.3 ±42.2 369.7 ± 191.2 28.9 ±28.9 Yes 59 34.0 ± 13.8 40.8 ± 6 .6 40 97.8 ± 53.7 414.6 ± 178.8 25.3 ± 12.8 P 0.80 0.70 0.30 0.28 0.40 Treatment Surgery only 56 33.9 ± 13.4 40.9 ±7.2 36 92.8 ± 50.6 362.3 ± 195.3 28.5 ±15.5 Surgery+chem# 44 35.4 ± 17.2 41.4 ±7.2 27 92.4 ±48.9 422.7 ±216.0 23.5 ± 10.7 Surgery+radiation 37 33.5 ± 12.6 40.7 ±5.6 28 88.9 ±46.3 341.0 ± 143.9 33.8 ±44.6 Surgery+chem#+radiA . 30 35.3 ± 11.6 42.0 ±6.0 2 2 95.7 ±39.3 445.5 ± 171.4 24.3 ± 12.6 No surg, chem.# , radA 2 2 2 .8 ± 8 .6 35.5 ±7.6 2 50.0 ± 15.0 213.9 ±60.8 23.3 ±0.36 0.75 0.70 0.76 0.14 0.56 Smoking status Never smoke 91 35.6 ± 15.0 41.4 ±7.2 61 91.4 ±43.8 417.3 ± 211.2 27.1 ±30.9 Formerly smoking 51 35.2 ± 13.6 41.6 ±6.3 35 89.3 ±44.7 373.6 ± 160.8 25.1 ±9.9 Currently smoking 27 28.4 ± 8 .8 * 39.0 ±5.1 19 96.6 ±59.2 304.3 ±119.4* 34.2 ± 19.5 P for global test 0.05 0.19 0 .8 6 0.08 0.43 trend 0.04 0.16 0.77 0 .0 2 0.43 Notes: IAF and EAF indicate internal and external abdominal fat, respectively. Mean values for each category were compared to the lowest level of the same categorical variable, * indicates a p value of ‘<=0.05’; # and A indicate chemotherapy and radiation, respectively. 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 3-2. Univariate relationships between body fat mass, body fat percent, internal and external abdominal fat, and the ratio of internal to external abdominal fat of HEAL study subjects at 30 months after diagnosis (Cont’d). Variable N Body Fat Body Fat N IAF EAF IAF/EAF Mass (Kg) (% ) (cm2 ) (cm2 ) (% ) Total energy intake (kcal) Q1 [198.8, 874.1) 38 34.6 ± 13.0 41.3 ±5.9 31 88.1 ±40.4 410.1 ± 163.7 23.0 ±9.4 Q2 [874.1,1306.7) 43 34.3 ±15.3 41.6 ± 6 .8 30 98.1 ±48.0 336.7 ± 176.2 32.8 ±15.8 Q3 [1306.7, 1732.2) 43 35.4 ± 15.1 40.9 ±6.7 29 86.1 ±53.6 404.7 ±225.9 22.9 ± 12.9 Q4 [1732.2,12364.1) 45 33.1 ± 12.5 40.6 ± 7.3 25 94.5 ±44.7 390.5 ± 179.0 32.8 ±46.1 P for global test 0.90 0.89 0 .6 6 0.25 0.77 P for trend test 0.72 0.52 0 .8 6 0.98 0.36 Energy from fat (% ) Q1 [18.4, 33.6) 40 32.0 ± 12.5 39.8 ±7.1 29 93.1 ±51.5 351.7 ± 178.4 28.8 ± 16.5 Q2 [33.6, 38.7) 42 33.9 ± 13.0 41.5 ±5.9 26 84.6 ± 37.9 364.9 ±137.5 23.7 ±9.6 Q3 [38.7, 43.4) 44 33.1 ± 13.3 40.5 ±6.3 27 99.3 ±46.7 394.2 ± 202.2 28.3 ±12.7 Q4 [43.4,61.4) 43 38.2 ±16.3* 42.5 ±7.1 31 89.1 ±50.0 419.3 ±221.5 29.8 ±42.2 P for global test 0.19 0.29 0.70 0.52 0.81 P for trend test 0.07 0.13 0.98 0.14 0.71 Total leisure time Activity 1 Y r before Diagnosis (MET.hrs/week) 0 96 35.8 ± 15.8 41.5 ± 6.9 58 97.1 ±52.7 403.8 ±210.4 30.0 ±32.1 (0,9] 25 31.6 ± 11.1 40.3 ± 6.4 2 0 92.4 ±42.5 367.1 ± 174.0 27.1 ± 12.2 (9,20] 2 2 37.8 ±11.5 43.4 ±6.0 2 0 88.3 ±39.7 426.9 ± 158.0 21.3 ±8.2 (2 0 , 26 28.6 ±8.3* 38.3 ±5.6* 17 75.7 ±33.7 295.0 ± 120.2* 28.2 ± 16.4 P for global test 0.05 0.04 0.46 0 .1 1 0.57 P for trend test 0.08 0.16 0 .1 0 0.14 0.42 Sports activity level A fter Diagnosis (MET.hrs/week) <3.0 82 34.9 ±15.9 41.4 ±7.6 56 91.7 ±48.2 396.1 ±208.3 26.2 ± 13.8 3.0-8.0 32 34.2 ±10.0 41.6 ± 4.9 21 110.1 ±43.4 382.2 ± 155.6 38.8 ±49.8* 8 .0 -2 0 .0 28 35.7 ± 14.6 42.0 ±6.3 19 80.2 ± 44.4 369.3 ±200.2 25.3 ± 15.0 >=2 0 .0 27 31.4 ± 10.8 39.4 ±5.6 19 82.3 ±43.8 373.0 ± 148.8 22.0 ± 7.3 P for global test 0.65 0.47 0 .2 1 0.91 0 .1 1 P for trend test 0.42 0.51 0.32 0.55 0.56 Vigorous activity level A fter Diagnosis (MET.hrs/week) 0 97 35.4 ± 15.3 41.4 ±7.1 65 96.4 ±48.6 403.8 ±201.0 27.3 ± 14.8 <5.0 37 33.8 ±11.7 41.5 ±5.3 23 92.0 ±43.6 371.2 ± 169.8 33.6 ±48.1 >5.0 35 31.9 ± 12.0 39.7 ± 6 .6 27 79.8 ±43.2 353.0 ± 167.4 23.6 ±9.4 P for global test 0.44 0.40 0.32 0.48 0.30 P for trend test 0 .2 0 0.24 0.13 0 .2 2 0.70 Notes: IAF, EAF indicate internal and external abdominal fat, respectively. Mean values for each category were compared to the lowest level of the same categorical variable, * indicates a p value of ‘<=0.05’. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. however, women in the fourth quartile o f ‘percent o f energy taken from fat’ had significantly higher values o f BMI, waist, hip and body fat mass, and marginally significantly higher values for weight and body fat percent than women who were in the first quartile o f this variable. There were marginally significant linear trends between weight, BMI, waist, hip, body fat mass, body fat percent and the levels o f percent energy taken from fat. Women who had the highest level o f total leisure time activity 1 year before diagnosis had significantly lower value for nearly all the endpoint variables except for W HR and the ratio o f IAF to EAF, compared with those who reported no leisure time activity. A significant linear trend existed across categories o f sports activity, vigorous activity level and waist hip ratio (p=0.01 and 0.004, respectively). Women who had the highest level o f sports or vigorous activity had significantly smaller W HR than women who didn’t participate in these activities after diagnosis. ADRfi3 Trp64Arg polymorphism and obesity status at 30 months post diagnosis Table 4-1 shows that there was a disproportional distribution o f variant genotypes between non-obese and obese subjects at 30 months post diagnosis. 27.36% o f non-obese subjects versus 14.29% o f obese subjects had ADR03 Trp64Arg variant genotype (p<0.05). Table 4-1. Frequency of obesity among HEAL subjects at 30 months after diagnosis and ADRp3 Trp64Arg polymorphism._______________________________________________________________ Genotypes Total Trp/Trp Trp/Arg Arg/Arg N on-obese(BMI<3 Okg/m2 ) 77 (72.64%) 26 (24.53%) 3 (2.83%) 106(100%) Obese (BMI>=30kg/m2 ) 60 (85.71%) 10 (14.29%) 0 70(100%) Notes: Assuming dominant effect of the Arg variant allele, the variant group showed significantly lower 2 3 frequency of obesity than the Trp64Trp subgroup (p<0.05). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Multivariate logistic regression models were fitted to estimate the risk of obesity associated with the ADR/33 Trp64Arg genotype at 30 months post diagnosis (Table 4-2). The risk o f being obese at 30 months was reduced by 59% among subjects with Arg64 allele compared to subjects without Arg64 allele (Odds ratio: 0.41, with 95% confidence interval: 0.18-0.94), after adjusted for sports activity level and smoking status. The risk o f being obese among women who had smoked since pre-adult usual status was reduced by 68% compared with those who has never smoked (Odds ratio: 0.32 with 95% confidence interval: 0.12-0.84), after adjusted for the ADR/33 genotype and sports activity level. The women who had the highest level o f sports activity had reduced risk o f obesity relative to those who had the lowest level o f sports activity (odds ratio =0.35; 95% confidence interval: 0.11-0.88). Similar effects o f the variant genotype and smoking on obesity status at 30 months post diagnosis were found after adjustment for vigorous activity level, instead o f sports activity level. In addition to BMI, other obesity-related variables were studied in similar logistic regression models; the only significant genotype effect was found for “Body fat percent” (Table 4-3). The odds o f being in the fourth quartile versus in the other 3 quartiles o f ‘body fat percent’ among women with variant genotype at 30 months post diagnosis was reduced by 68% (odds ratio=0.32, 95% confidence interval: 0.11-0.92), compared to those with wild genotype. 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4-2. Association of the ADRp3 Trp64Arg polymorphism, sports activity and smoking with obesity (BMI>30Kg/m2 ) among HEAL patients at 30 months after diagnosis. Unadjusted Analysis Adjusted Analysis+ Variable N Odds Ratio P value Odds Ratio 95% Confidence Interval PValu ADRfi3 genotype Trp/Trp 137 reference reference Trp/Arg, Arg/Arg 39 0.44 0.04 0.41 (0.18, 0.94) 0.03 Sports Activity (MET’ *hrs/week) <3.0 82 reference reference 3.0-8.0 32 1.05 0.90 1.01 (0.44, 2.33) 0.99 8 .0 -2 0 .0 29 1.08 0.85 1.11 (0.46, 2.68) 0.82 >=2 0 .0 28 0.35 0.04 0.31 (0 .1 1 , 0 .8 8 ) 0.03 Smoking status change* Never-Never 94 reference reference Former-Former 27 0.46 0 .1 0 0.50 (0.19,1.28) 0.15 Current-current 28 0.36 0.04 0.32 (0.12, 0.84) 0 .0 2 Current-Former 27 0.64 0.32 0.70 (0.28, 1.74) 0.44 Notes: 1. # indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. 2 . + : adjusted for the other variables in the table. 3. When the sports activity is replaced by vigorous activity level, the ADRf}3’s effect is similar and still significant (p=0.03). Table 4-3. Association of the ADRp3 Trp64Arg polymorphism, sports activity, chemotherapy and smoking with obesity (body fat>45.52%) among HEAL patients at 30 months after diagnosis. Unadjusted Analysis Adjusted Analysis+ Variable N Odds Ratio P value Odds Ratio 95% Confidence Interval P Value ADRfi3 genotype Trp/Trp 137 reference reference Trp/Arg, Arg/Arg 39 0.38 0.06 0.32 (0.11,0.92) 0.03 Sports Activity (MET*hrs/week) <3.0 82 reference reference 3.0-8.0 32 1.08 0.87 1.16 (0.44, 3.02) 0.77 8 .0 -2 0 .0 29 1.35 0.53 1.46 (0.54, 3.92) 0.46 >=2 0 .0 28 1.48 0 .2 2 0.44 (0.13, 1.48) 0.18 Chemotherapy No 98 reference reference Yes 78 1.85 0.08 1.96 (0.94,4.11) 0.07 Smoking status change Never-Never 94 reference reference Former-Former 27 1 .1 0 0.84 1.31 (0.48, 3.59) 0.60 Current-current 28 0.31 0.08 0.29 (0.08, 1.06) 0.06 Current-Former 27 0.75 0.57 0.92 (0.32, 2.68) 0 .8 8 Notes: 1. # indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. 2 . + : adjusted for the other variables in the table. 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ADRP3 Trp64Arg polymorphism and weight / BMI gain from adult usual status to 30 months post diagnosis Table 5 shows the univariate relationships between the ADRft3 Trp64Arg polymorphism, multiple covariates and the change o f weight and BM I from adult usual status to 30 months post diagnosis. Women carrying the Arg64 allele had significantly lower weight and BMI gain since adult usual status (8.1 versus 13.8Kg, 3.0 versus 5.1Kg/m2, p=0.02 for both comparisons). The multivariate model for weight at 30 months is shown in Table 6. After controlling for adult usual weight, chemotherapy, smoking status change and time between “adult usual status” and 30 months post diagnosis, the ADRf}3 Arg64 allele was associated with 6.04Kg lighter weight at 30 months (p<0.01). After controlling for the other variables in the model, chemotherapy was associated with 3.61 Kg higher weight at 30 months (p=0.06), while women who quit smoking during the studied time interval weighted 5.19Kg more at 30 months than those had never smoked (p=0.06). A similar multivariate model for the BMI at 30 months post diagnosis is shown in Table 7. After controlling for adult usual BMI, chemotherapy, smoking status change and time between the two time points o f interest, the Arg64 allele o f the ADRp3 was associated with 2.17Kg/m2 lower BM I at 30 months (p=0.001). Subjects who had chemotherapy had 1.30Kg/m2 higher BMI at 30 months than those had no chemotherapy (p=0.06). Women who quit smoking during the interval had 1.97Kg/m2 higher BM I at 30 months than those had never smoked (p=0.06). 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 5. Weight and BMI change from adult usual status to 30 months after diagnosis in relation to ADRp3 Trp64Arg genotype and selected characteristics (to be continued)._________________________________________________________________________________________________ N W eight (Kg) BM I (Kg/m2) Adult usual + 30 months After Dx + Change + P value for Change * Adult usual + 30 months After Dx + Change + P value for Change * ADRft3 genotype Trp64Trp 137 64.8(1.3) 81.6(1.9) 13.8(1.2) 25.0 (0.4) 30.0 (0.6) 5.1 (0.4) Trp64Arg & Arg64Arg 39 69.8 (2.2) 77.9 (2.8) 8.1 (1.7) 0 .0 2 25.3 (0.7) 28.2(1.0) 3.0 (0.6) 0 .0 2 Age group (years) 38-45 44 64.7(1.7) 79.2 (2.9) 14.5(1.9) 23.7 (0.6) 29.1(1.0) 5.3 (0.7) 46-50 39 64.0(1.8) 83.0(3.6) 19.0(2.5) 0 .1 1 23.8(0.7) 30.8(1.3) 7.0 (0.9) 0 .1 0 51-55 37 74.9 (3.5) 83.6 (3.6) 8.7 (1.7) 0.04 26.6(1.2) 29.7(1.2) 3.1 (0.6) 0.04 56-60 30 74.7 (2.4) 85.1 (4.2) 10.4 (2.5) 0.17 27.7 (0.8) 31.5(1.4) 3.8 (0.9) 0.17 61-66 26 63.6 (2.1) 71.2 (3.1) 7.6 (1.7) 0.03 23.8(0.6) 26.6(1.0) 2.8 (0.7) 0.03 P for global test 0 .0 0 1 0 .0 0 1 Diagnostic stage In situ 30 69.1 (2.6) 79.9 (3.9) 10.9(2.3) 25.1 (1.0) 29.0(1.4) 3.9 (0.8) local 81 6 8 .2 ( 1.6 ) 79.7(2.1) 11.5(1.5) 0.84 25.2 (0.5) 29.5 (0.8) 4.3 (0.5) 0.73 regional 65 67.8 (2.0) 82.5 (2.8) 14.7(1.7) 0 .2 0 24.8 (0.6) 30.1 (0.9) 5.3 (0.6) 0.19 P for global test 0.26 0.30 Still having periods Yes 34 64.7 (2.1) 75.9(3.1) 1 1 .2 (2 .0 ) 24.3 (0.7) 28.5 (1.1) 4.2 (0.7) No 142 69.1(1.3) 81.9(1.8) 12.9(1.2) 0.52 25.2 (0.4) 29.9 (0.6) 4.7 (0.4) 0.59 Treatm ent Surgery only 59 68.5(2.1) 80.1 (2.7) 11.6(1.5) 25.1 (0.7) 29.4 (0.9) 4.3 (0.6) Surgery +radiation 37 69.1 (2.1) 80.3(3.1) 11.2 (2.5) 0.99 25.0 (0.7) 29.1 (1.1) 4.1 (0.9) 0.95 Surgery+chemotherapy 47 66.9 (2.3) 81.8(3.7) 14.9(2.3) 0.16 24.6 (0.7) 30.1 (1.3) 5.5 (0.8) 0.18 Surgery+radiation+chemotherapy 31 69.1 (2.8) 82.3 (3.0) 13.1 (1.6) 0.53 25.5(1.0) 30.4(1.1) 4.9 (0.6) 0.52 P for global test 0.48 0.49 Smoking status change Never - never smoking 94 71.1 (1.7) 83.2 (2.3) 12.1 (1.4) 26.0 (0.5) 30.5 (0.8) 4.5 (0.5) Formerly-formerly smoking 27 69.3 (2.7) 80.9 (4.5) 1 1 .6 (2 .6 ) 0.87 25.2 (0.9) 29.5(1.6) 4.2 (0.9) 0.84 Currently-currently smoking 28 62.3(1.9) 72.2 (2.8) 9.9 (2.1) 0.45 23.1 (0.6) 26.8 (0.9) 3.6 (0.8) 0.44 Currently-formerly smoking 27 63.4 (2.3) 81.0 (3.0) 17.7(2.8) 0.06 23.3 (0.8) 29.8(1.2) 6 .6 ( 1.1) 0.05 P for global test 0.28 0.28 Notes: + : values are mean(standard error); * : p values are based on the comparison of weight or BMI change with the lowest level as the reference. to < i Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. (cont’d). N Weight (Kg) BMI (Kg/m2) Adult usual + 30 months After Dx + Change + P value for Change * Adult usual + 30 months After Dx + Change + P value for Change * Tamoxifen Use Non user 117 68.2(1.5) 80.9 (2.0) 1 2 .6 ( 1.2 ) 25.1 (0.5) 29.7 (0.7) 4.6 (0.4) User 59 68.2(1.7) 80.6 (2 .6 ) 12.4(1.8) 0.89 25.0 (0.6) 29.5(1.0) 4.6 (0.7) 0.91 Total energy intake (Kcal) Quartile 1 (<874.1) 42 69.6 (2.2) 82.2 (3.1) 1 2 .6 (2 .2 ) 25.3 (0.7) 29.9(1.0) 4.6 (0.8) Quartile 2 (874.1-1306.7) 45 67.7 (2.0) 79.1 (3.4) 11.4(2.2) 0 .6 8 25.6 (0.6) 29.9(1.2) 4.3 (0.8) 0.77 Quartile 3 (1306.7-1732.2) 43 69.5 (2.7) 82.3 (3.4) 13.7(2.0) 0.69 24.9 (0.9) 29.8(1.1) 4.9 (0.7) 0.78 Quartile 4 (>1732.2) 46 66.3(2.1) 78.8 (2.7) 12.5(1.7) 0.98 24.3 (0.7) 29.0(1.0) 4.7 (0.7) 0.92 P for global test 0 .8 8 0.95 Energy intake from fat (%) Quartilel [18.4, 33.6) 44 6 6 .6 (2.3) 77.6 (2.8) 11.0(1.9) 24.4 (0.8) 28.5(1.0) 4.1 (0.7) Quartile 2 [33.6, 38.7) 44 66.8(1.7) 79.9 (3.0) 13.1 (2.1) 0.46 24.6 (0.5) 29.4(1.0) 4.8 (0.8) 0.53 Quartile 3 [38.7,43.4) 44 6 8 .1 (2 .2 ) 79.2 (3.0) 1 1 .1 ( 1.8 ) 0.98 25.0 (0.8) 29.1 (1.0) 4.0 (0.6) 0.94 Quartile 4 [43.4,61.4) 44 71.3(2.7) 86.3 (3.7) 15.0 (2.2) 0.16 26.1 (0 .8 ) 31.6(1.3) 5.5 (0.8) 0.18 P for global test 0.44 0.45 Leisure time activity lYr before diagnosis (met*hour/week) 0 101 69.2(1.5) 83.1 (2.3) 13.9(1.5) 25.4 (0.5) 30.5 (0.8) 5.1 (0.5) 0-9 26 6 6 .1 (2 .6 ) 75.8 (3.4) 9.7 (1.8) 0.15 24.3 (0.8) 27.9(1.1) 3.6 (0.7) 0.15 9-20 2 2 74.8 (3.4) 85.2 (3.4) 10.4 (2.9) 0.26 27.6(1.2) 31.6(1.4) 4.0 (1.1) 0.26 > 2 0 27 61.0 (2 .0 ) 73.1 (2.4) 1 2 .1 (2 .0 ) 0.54 2 2 .1 (0 .6 ) 26.5 (0.8) 4.4 (0.7) 0.54 P for global test 0.42 0.49 Sports activity after diagnosis (met*hour/week) <3.0 84 68.6(1.7) 81.9(2.6) 13.3 (1.6) 25.4 (0.6) 30.4 (0.9) 4.9 (0.6) 3.0-8.0 34 68.5 (2.2) 80.2 (2 .6 ) 11.8(1.9) 0.58 25.1 (0.8) 29.5(1.0) 4.3 (0.7) 0.56 8 .0 -2 0 .0 29 69.4 (2.9) 81.8(4.2) 12.4 (2.8) 0.77 25.2 (0.9) 29.7(1.4) 4.5 (1.0) 0 .6 6 >2 0 .0 29 65.6 (2.6) 77.0 (2.9) 11.5(1.7) 0.54 23.5 (0.9) 27.6(1.0) 4.1 (0.6) 0.45 P for global test 0.91 0 .8 6 Vigorous activity after diagnosis (met'hour/week) 0 99 69.6(1.6) 82.4 (2.3) 12.8(1.5) 25.6 (0.5) 30.3 (0.8) 4.7 (0.5) <5.0 40 66.7 (2.0) 79.6 (3.0) 13.0(2.1) 0.94 24.7 (0.7) 29.5(1.0) 4.8 (0.8) 0.92 >5.0 37 66.1 (2.4) 77.6 (2.8) 11.5(1.6) 0.62 23.9 (0.8) 28.1 ( 1.0 ) 4.2 (0.6) 0.59 0 .8 6 0.84 Notes: + : values are mean(standard error); * : p values are based on the comparison of weight or BMI change with the lowest level as the reference. ro co Table 6. The ADR[i3 Trp64Arg polymorphism, smoking, chemotherapy and weight at 30 months post diagnosis. Variable N Adjusted Least Square 95% Cl for LS means P Value for (LS) means of weight at of weight at 30 months global test ______________________________________ 30 months (Kg)____________ (Kg)__________________ A D R p3 genotype Trp64Trp 137 82.9 (80.4, 85.4) Trp64Arg, Arg64Arg 39 76.7 (72.5, 81.0) <0 .0 1 Smoking status change* Never smoker-never smoker 94 78.4 (77.4,81.4) Former smoker-former smoker 27 78.2 (73.1, 83.3) Current smoker-current smoker 28 77.2 (72.1,82.2) Current smoker-former smoker 27 85.5 (80.4, 90.6) 0.06 Treatment No chemotherapy 98 78.1 (75.0,81.1) Had Chemotherapy 78 81.6 (78.2, 84.9) 0.08 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘adult usual weight’. 2. * indicates the smoking status at the ‘usual adult weight’ and 30 months post diagnosis. Table 7. The A D R p3 Trp64Arg polymorphism, smoking, chemotherapy and BMI at 30 months post diagnosis. Variable N Adjusted LS means of BMI at 30 months (Kg/m2 ) 95% Cl for LS means P Value for of BMI at 30 months global test (Kg/m2 ) AD RP3 genotype Trp64Trp 137 30.4 (29.5, 31.3) Trp64Arg, Arg64Arg 39 28.2 (26.6,29.7) 0.01 Smoking status change* Never smoker-never smoker 94 28.8 (27.7, 29.9) Former smoker-former smoker 27 28.7 (26.8, 30.6) Current smoker-current smoker 28 28.3 (26.4, 30.1) Current smoker-former smoker 27 31.4 (29.6,33.3) 0.05 Treatment No chemotherapy 98 28.7 (27.5,29.8) Had Chemotherapy 78 29.9 (28.7,31.2) 0.08 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘adult usual BMI’. 2. * indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. ADRfi3 Trp64Arg polymorphism and waist and hip circumference at 30 months Table 8-1 and 8-2 show that, the Arg64 allele of the ADR/33 genotype was associated with 4.23cm smaller waist and 3.98cm smaller hip circumference at 30 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. months after diagnosis (p=0.03 and 0.06, respectively), after controlling for adult usual BMI, smoking status change and time during the studied interval. Women who quit smoking during the time interval had 5.12cm larger waist, but non-significantly larger hip circumference than those who had never smoked (p=0.04 and 0.28, respectively). Table 8-1. The ADR/33 Trp64Arg polymorphism, smoking and waist circumference (cm) at 30 months post diagnosis. Variable N Adjusted LS means of 95% Cl for LS P Value waist at 30 months (cm) means of waist at 30 for global months (cm) test A D R p3 genotype Trp64Trp 132 96.4 (94.3, 98.5) Trp64Arg, Arg64Arg 39 91.9 (88.4, 95.5) 0.03 Smoking status change* Never smoker-never smoker 92 92.6 (90.0, 95.1) Former smoker-former smoker 26 92.9 (8 8 .6 , 97.2) Current smoker-current smoker 27 92.5 (88.1,96.8) Current smoker-former smoker 26 98.8 (94.5,103.1) 0.07 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘adult usual BMI’. 2. * indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. Table 8-2. The ADR/33 Trp64Arg polymorphism, smoking status and hip circumference (cm) a 30 months post diagnosis. Variable N Adjusted LS means of 95% Cl for LS means P Value for hip circumference at 30 of hip circumference at global test months (cm) 30 months (cm) ADR/33 genotype Trp64Trp 132 114.4 (112.1,116.6) Trp64Arg, Arg64Arg 39 1 1 0 .1 (106.3, 113.9) 0.04 Smoking status change* Never smoker-never smoker 92 111.0 (108.3, 113.7) Former smoker-former smoker 26 113.2 (108.6, 117.8) Current smoker-current smoker 27 109.7 (105.1,114.2) Current smoker-former smoker 26 115.2 (110.6, 119.8) 0.27 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘ adult usual BMI’. 3 q 2. * indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ADR/33 Trp64Arg polymorphism and W HR at 30 months post diagnosis The ADR/33 Trp64Arg polymorphism was not associated with W HR at 30 months post diagnosis. The highest level o f sports or vigorous activity was associated with significantly lower W HR (Table 9, the model including vigorous activity was not shown). W omen who stopped smoking after diagnosis had marginally significantly lower W HR than those who had never smoked (p=0.09). Table 9. The A D R P3 Trp64Arg polymorphism, smoking status and Waist-hip ratio (WHR) at 30 months post diagnosis. Variable N Adjusted LS means of WHR at 30 months 95% C l for LS means P Value for of WHR at 30 months global test A D R /i3 genotype Trp64Trp 132 0.84 (0.83, 0.85) Trp64Arg, Arg64Arg 39 0.83 (0.81,0.85) 0.49 Smoking status change* Never smoker-never smoker 92 reference (0.81,0.84) Former smoker-former smoker 26 0.83 (0.79, 0.84) Current smoker-current smoker 27 0.82 (0.81,0.86) Current smoker-former smoker 26 0.85 (0.83,0.88) 0.22 Sports Activity (MET*hrs/week) <3.0 82 0.85 (0.83, 0.86) 3.0-8.0 32 0.84 (0.82, 0.87) 8 .0 -2 0 .0 29 0.84 (0.82, 0 .8 6 ) >=2 0 .0 28 0.81 (0.79,0.84) 0.10 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘adult usual BMI’. 2. # indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. 3. Genotype’s effect keeps non-significant, ‘sports activity level’ is replaced by ‘vigorous activity level’. ADRfi3 Trp64Arg polymorphism and body fat mass and body fat percent at 30 months post diagnosis Table 10-1 and 10-2 show that the Arg64 allele o f th& ADR/33 genotype was associated with 3.75% Kg less body fat mass and 1.69 lower body fat percent at 30 months after diagnosis (p=0.03 and 0.054, respectively), after controlling for 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. adult usual BMI, smoking status change and time between adult usual weight and 30 months post diagnosis. Women who quit smoking during the studied time interval had 2.73Kg more body fat mass, and 2.03 lower body fat percent than those who had never smoked (p=0.09 and 0.07, respectively). Table 10-1. The ADR/33 Trp64Arg polymorphism, smoking and Body fat mass (BFM) (Kg) at 30 months post diagnosis. Variable N Adjusted LS means of BFM at 30 months (Kg) 95% Cl for LS means P Value for of BFM at 30 months global test (Kg) A I)R p3 genotype Trp64Trp 130 35.7 (33.8, 37.5) Trp64Arg, Arg64Arg 39 31.8 (28.7,34.8) 0.02 Smoking status change* Never smoker-never smoker 91 32.6 (30.4, 34.8) Former smoker-former smoker 25 33.6 (29.8, 37.4) Current smoker-current smoker 27 31.3 (27.5, 35.0) Current smoker-former smoker 26 37.4 (33.7,41.1) 0.09 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘adult usual BMI’. 2. * indicates the smoking status at the usual adult status and 30 months post diagnosis. Table 10-2. The ADR/33 Trp64Arg polymorphism, smoking and Body fat percent (FP) (%) at 30 months post diagnosis. Variable N Adjusted LS means of FP at 30 months 95% Cl for LS means of FP at 30 months P Value for global test A D R p3 genotype Trp64Trp 130 41.8 (40.8, 42.8) Trp64Arg, Arg64Arg 39 40.0 (38.4,41.6) 0.05 Smoking status change* Never smoker-never smoker 91 40.1 (39.0,41.3) Former smoker-former smoker 25 40.5 (38.5,42.5) Current smoker-current smoker 27 40.2 (38.2,42.1) Current smoker-former smoker 26 42.9 (40.9,44.9) 0 .1 0 Notes: 1. All the effects have been adjusted for the other variables listed in the table and ‘adult usual BMI’. 2. * indicates the smoking status at ‘adult usual weight’ and 30 months post diagnosis. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ADRp3 Trp64Arg polymorphism and abdominal fat measured by M RI 115 out o f 176 study subjects who completed 30 months interview and had blood draw had abdominal fat measured using MRI. With this smaller sample, we examined the effect o f the ADR/}3 Trp64Arg polymorphism on internal and external abdominal fat (IAF and EAF, respectively), as well as their sum (total abdominal fat) and ratio (IAF to EAF). Women with the ADRJ33 Trp64Arg variant genotype had lower internal abdominal fat at 30 months post diagnosis, after adjusting for the adult usual BM I and vigorous activity level (Table 11). Table 11. The ADRP3 Trp64Arg genotype, vigorous activity and internal abdominal fat (IAF) (cm2 ) at 30 months post diagnosis. N Adjusted LS means of IAF at 30 months (cm2 ) 95% Cl for LS means of IAF at 30 months (cm2 ) P Value for global test A D R p3 genotype Trp/Trp, 8 8 94.7 (84.9, 104.6) Trp/Arg or Arg/Arg 27 77.9 (60.9, 95.0) 0.08 Vigorous activity(MET*hrs/wk) 0 (MET.hrs/wk) 65 89.2 (77.5, 101.0) <5.0 (MET.hrs/wk) 23 92.9 (74.5, 111.3) >5.0 (MET.hrs/wk) 27 76.9 (59.3, 94.5) 0.37 Notes: The genotype’s effect becomes even non-significant, if vigorous activity was replaced by sports activity. There was no main effect o f the ADRfi3 genotype on any o f the other variables representative o f abdominal fat above mentioned. It was shown that, current tamoxifen users had higher EAF than the non-users, after adjustment for the other significant predictors o f EAF (chemotherapy and smoking status). W e further identified a significant modification effect o f the ADRJ33 Trp64Arg genotype on the 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Tamoxifen use and external abdominal fat relationship. Tamoxifen’s effect o f increasing EAF was limited to women who did not carry the ADR/53 Arg64 allele (p=0.06) (Table 12). Table 12. The ADR|13 Trp64Arg genotype, chemotherapy, current Tamoxifen use, smoking and external abdominal fat (EAF) (cm2 ) at 30 months post diagnosis. N Adjusted LS means of EAF at 30 months (cm2 ) 95% Cl for LS means of EAF at 30 months (cm2 ) P Value for global test A D R p3 genotype and tamoxifen use+ Trp/Trp, no tamoxifen 57 380.1 (344.2,415.9) VariantA and no tamox 18 374.0 (316.3,431.7) Trp/Trp and using Tamoxifen 31 437.8 (391.9, 483.8) VariantA and using Tamox 9 325.4 (240.3,410.5) 0.06 Treatment No Chemotherapy 66 348.0 (311.1,384.8) Had Chemotherapy 49 410.7 (370.3,451.0) 0.01 Smoking status change* Never -Never 61 375.8 (338.4,413.3) Former-Former 20 374.0 (316.9,431.0) Current-Current 19 341.0 (282.8, 399.2) Current-Former 15 426.5 (362.3, 490.7) 0.26 Notes: 1. + means tamoxifen use at 30 months interview. Tamoxifen-use at initial treatment didn’t show a significant effect in this model, A indicates Trp/Arg or Arg/Arg genotypes. 2. * indicates smoking status at adult usual status and at 30 months post diagnosis. 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DISCUSSION The HEAL study is a population-based longitudinal study o f women with breast cancer. W e examined the effects o f the ADR/33 Trp64Arg polymorphism on multiple obesity-related variables among 176 African American women with breast cancer, with adjustment for life style and treatment. W e found that the ADR/33 Arg64 (variant) allele was associated with reduced risk o f obesity (defined as BMI>30Kg/m2 or body fat percent in the fourth quartile) at 30 months after diagnosis o f breast cancer. W omen bearing an Arg64 allele experienced smaller increases in weight and BMI between “usual adult weight” and 30 months post diagnosis, compared with those with no copies o f that allele. The Arg64 allele was also associated, with lower body fat mass, body fat percent, lower waist and hip circumferences and decreased internal abdominal fat at 30 months post diagnosis. The ADR/33 Trp64Arg polymorphism has been studied in a number o f ethnic populations, including Caucasians, African Americans, Japanese, and Pima Indians (Arner, 1999). The present study, to our best knowledge, is the first to investigate the effects o f the ADR/33 Trp64Arg polymorphism on obesity-related phenotypes among African American women with breast cancer. The frequency o f the ADR/33 Arg64 allele was 11.9% in our sampled population, consistent with previously reported 9.5% in African American women in Chicago (Lowe, 2001), and 10.5% in Jamicans, a population o f African descent (McFarlane-Anderson, 1998). Because the allele frequency is the same in this breast cancer population as controls from the general population, this polymorphism is not strongly correlated with the occurrence o f 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. breast cancer among African American women, a finding consistent with a previous study by Huang (2001) among Japanese women. Our study demonstrated a negative association between inheritance o f an ADR/33 Arg64 allele and obesity, which is consistent with the study by Lowe et al (2001) showing a negative correlation between recessive inheritance o f the Arg64 allele and obesity-related traits among African American women, but not among African American men. In contrast to the present study and Lowe et al., McFarlane- Anderson (1998) found that Jamaican women carrying Arg64 allele had a significantly increased BM I and hip circumference compared to non-carriers. Studies among other ethnic populations reported either increased risk o f obesity associated with the Arg64 allele or a null effect o f this polymorphism (Allison, 1998, Arner, 1999). Our study suggests that the ADR/!3 Trp64Arg polymorphism may limit weight gain during the adult life including during the pre- and post-breast cancer diagnosis time period, and after breast cancer treatment. Two biological mechanisms may explain the observed results. First, the Arg64 allele may have directly or indirectly modulated the expression o f adrenergic receptor with mildly stronger lipolysis activity than that from the Trp64 allele. Second, the variant (Arg64) allele may be in linkage disequilibrium with a genetic factor associated with higher energy expenditure and/or lower fat deposition. Thus the carriers o f the Arg64 allele may be predisposed to reduced risk o f obesity under “obesity stress” in midlife as well as during breast cancer treatment. 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The available experimental studies on the functional importance o f the ADR/33 Trp64Arg polymorphism, all conducted in Caucasians or Japanese population, indicate that the Arg64 allele may have physiological properties that would lead to increased risk o f obesity (Sipilainen, 1997, Hoffstedt, 1999, Tchernof, 1999,Umekawa, 1999). The inconsistency o f studies may result, in part, from different effects o f this polymorphism among African American women from other races; difference in population characteristics between our study and those o f others; or unrecognized bias, confounding, or effect modification in our study or in the other studies. Difference in population characteristics Among a number o f population characteristics, we found important differences in the following three factors: A. Breast cancer status Our study is unique in that it is on the effect o f the ADR/33 Trp64Arg polymorphism on African American women with breast cancer. The closest study was the one by Lowe et al (2001), who found similar results among African American women without breast cancer. Therefore, the breast cancer status o f our study subjects is unlikely to explain the opposite effects o f the Arg64 allele on obesity compared with the studies among other races. B. Overall obesity status A large percent o f study subjects were obese in our study and that by Lowe et al (2001), with mean BM I o f 29.6±7.3 kg/m2 and 31.5±8.2 kg/m2, respectively. In 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. contrast, many studies on Japanese women (Oizumi, 2001) and Jamaican women (McFarlane-Anderson, 1998) that have reported a risk effect o f the Arg64 allele on obesity were based on relatively lean subjects. For example, the Jamaican women in the study by McFarlane-Anderson (1998) had a mean BMI o f 27.2±6.2 kg/m2. The experiment by Tchernof et al (1999), which suggested that the obese state might mask a moderate effect o f the ADR/33 Arg64 allele on energy expenditure, may explain how the difference in overall obese status among different studies has led to the inconsistency in the observed effect o f the Arg64 allele on obesity. Based on this finding, the effect o f the Arg64 allele on weight gain in an obese population should be weaker than that in a non-obese population. W e examined the effect o f the ADRj33 Arg64 allele on the weight/BMI gain between “adult usual weight/BMI” and 30 months post diagnosis among women who were obese at “adult usual weight”, as well as among those who were not obese at that time. The effects o f the Arg64 allele did not differ significantly between these two groups. Therefore, the difference in the overall obesity status between our study and other studies that reported a risk effect o f the Arg64 allele on obesity is unlikely to explain the inconsistency in the effects of the Arg64 allele. C. Ratio o f internal and external abdominal fat Previous studies showed that African American women have lower proportion o f visceral fat (Conway, 1995, Albu, 1997), which is more closely related to the ADR/33 Trp64Arg polymorphism than subcutaneous fat (Tavenier et al 1996), due to their higher adult weight gain compared with Caucasian women (Perry AC, 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2000). This difference in the fat distribution may lead to the difference in the effects o f the ADR/33 Trp64Arg polymorphism on the total body mass and body fat between African American women and women o f other ethnicity, even if there is no racial difference on the effect o f this polymorphism to visceral obesity. D. Difference in exclusion criteria Our study has excluded women with Diabetes Mellitus (DM) because DM status may be a confounding factor in the relationship between the studied polymorphism and obesity, while many previous studies did not. This difference in exclusion criteria may be another source o f the inconsistency in the observed effect of the ADR/33 Trp64Arg polymorphism on obesity. However, prevalence o f DM in other studies is likely to be low and is unlikely to explain the difference in the effects o f the ADR/33 Trp64Arg polymorphism. Bias, confounding and effect modification In addition to the differences on study population, inconsistency among results from different studies on the ADR/33 Trp64Arg polymorphism may arise from bias, confounding, gene-gene and gene-environment interaction, and chance. A. Bias Selection bias is not likely to be a serious problem o f our study because firstly, our study is population-based; at the same time, we did not find significant unbalanced distribution o f diagnostic stage between the ADR/33 Trp64Arg genotypes. Thus we believe that non-participation or lost-follow-up due to serious disease or 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. early morbid should not be more serious in one genotype defined by the ADR/33 Trp64Arg than the other. Recall bias could be a problem for the self-reported “adult usual weight” and the “adult usual BM I” derived therein. This recall bias could be random or differential by current obese status, i.e., current obese women tend to underreport their previous weight. The latter situation, if present, is more destructive to the validity o f our study. However, there is no evidence supporting a deferential recall bias based on the ADR/33 Trp64Arg genotype, so the protecting effect o f the Arg64 allele to obesity we observed is unlikely to be attributable to the recall bias o f “adult usual weight” . In addition, our study is susceptible to measurement bias on some covariates, especially those about the energy intake and physical activities. But the influence o f this bias on the results should have been minimized by categorization. B. Confounding Serious confounding by population admixture is unlikely to affect our study results based on previous studies that showed no substantial difference on the frequency o f this polymorphism between the two parent populations o f African American (8-10% in American Caucasian, Arner, 1999). In the control o f potential confounding factors, our study may have performed better than previous studies. M ost previous studies, as reviewed by Allison et al. (1998), examined the effect o f th q ADR/33 Trp64Arg polymorphism by comparing the mean value o f obesity-related phenotypes between the group with variant genotype and that with wild genotype without sufficient adjustment o f potential 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. confounding. W e studied the effect o f this polymorphism on weight / BM I gain in a duration o f obesity development, i.e., between “adult usual weight” and 30 months after diagnosis o f breast cancer, as well as comparing the prevalence o f obesity between variant and wild genotypes. After adjusting for all available potential confounding factors (energy intake, physical activity, smoking, treatment and time between tw o study points), we found a protective effect o f this polymorphism on weight / BMI gain. C. Effect modification The inheritance pattern o f obesity strongly suggests that the effect is polygenic, with each variant o f many different genes contributing to a small difference (Sorensen TA, 2001). The strong environmental effects that have been observed led to the belief that obesity is the result o f an interaction between genetic predisposition and environmental influences (World Health Organization, 1997). Among several approaches that have been frequently used in the studies o f low penetrance genes, a feasible one to evaluate the main effect o f a polymorphism on phenotypes as long as the information on other genetic and environmental modification factors are unavailable. Similar to previous studies, we tested the main effect o f the ADRfS3 Trp64Arg polymorphism on obesity-related characteristics without considering other genetic factors, which may also contribute to the inconsistency between our results and those o f others. W e did test potential modification effects from other environmental factors that may affect the effect o f the ADRfi3 Trp64Arg 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. polymorphism on obesity, e.g. energy intake, physical activity, smoking and breast cancer treatment (chemotherapy and tamoxifen use). Conclusions and future studies Defining the genetic factors that influence obesity in breast cancer patients is important because obesity is associated with poor prognosis. W e found that the ADR/33 Arg64 variant allele was associated with reduced risk o f obesity among African American women with breast cancer. Although previous studies in other ethnic populations have found an increased obesity risk associated with the ADR/33 Arg64 variant allele, our findings are consistent with a previous report by Lowe et al (2001), who observed an inverse association among African American women in Chicago. Notably, Lowe et al. found no association o f the ADR/33 Arg64 variant allele in African American men in their study. Taken together, the available studies suggest that the ADR/33 Arg64 variant allele may have ethnic- and sex-specific effects on obesity. Furthermore, the sex and ethnic differences suggest that a gene- environment (including treatment) and/or gene-gene interactions may be important determinants o f the variant allele’s effects. For example, weight gain after chemotherapy was larger among women who did not carrying the ADR/33 Arg64 allele. Future in vivo and in vitro studies on the function o f this polymorphism need to be designed that consider the influences o f environmental and genetic background on ADR/33 gene expression, regulation or function. Because our results may have clinical implications in the treatment decision-making for breast cancer patients, additional experimental and epidemiological studies on the ADR/33 Trp64Arg polymorphism 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are needed among breast cancer patients, especially those involving its interaction with multiple genetic and treatment factors. 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCE Ainsworth BE, Haskell WL, Leon AS, Jacobs DR Jr., M ontoye HJ, Sallis, JF, and Paffenbarger RS Jr. Compendium o f physical activities: classification o f energy costs o f human physical activities. Medicine and Science in Sports and Exercise, 25: 71-80, 1993. Albu JB, M urphy L, Frager DH, et al: Visceral and race-dependent health risks in obese nondiabetic premenopausal women. Diabetes 46: 456-462, 1997. Allison DB, Heo M, Faith MS, Pietrobelli A. Meta-analysis o f the association o f the Trp64Arg polymorphism in the 03 adrenergic receptor with body mass index. International Journal o f Obesity (1998) 22, 559-566. Aggel-Leijssen DPCV, Saris WHM, Homan M, Baak MAV. The effect o f exercise training on 0-adrenergic stimulation o f fat metabolism in obese men. International Journal o f Obesity (2001) 25, 16-23. Arner P, Hoffstedt J. Adrenoceptor genes in human obesity. Journal o f Internal Medicine 1999; Volume 245(6), June 1999, 667-672. Aslani A, Smith RC, Allen BJ, Pavlakis N, Levi JA. Changes in body composition during breast cancer chemotherapy with the CMF-regimen. Breast Cancer Research and Treatment 1999 Oct; 57(3): 285-90. Bernstein L, Deapen D, Cerhan JR, et al: Tamoxifen therapy for breast cancer and endometrial cancer risk. Journal o f National Cancer Institute 91: 1654-1662, 1999. Brewster A, Helzlsouer K. Breast cancer epidemiology, prevention, and early detection. Current Opinion in Oncology 2001, 13: 420-425. Candelore MR, Deng L, Tota LM, Kelly LJ, Cascieri MA, Strader CD. Pharmacological characterization o f a recently described human 03 adrenergic receptor mutation. Endocrinology 1996; 137:2638-2641. Chlebowski T., Aiello E., McTiernan A. weight Loss in Breast Cancer Patient Management. Journal o f Clinical Oncology, Vol 20, no 4 (February 15), 2002: pp 1128-1143. Cold S, Hansen S, Overvad K, Rose C. A woman’s build and the risk o f breast cancer. European Journal o f Cancer 1998; 34: 1163-74. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Conway JM, Yanovski SZ, Avila NO, et al: Visceral adipose tissue differences in black and white women. American Journal o f Clinical Nutrition 61: 765-771, 1995. Corella D, Guillen M, Portoles O, Sorll JV, Alonso V, Folch J, Saiz C. Gender specific associations o f the Trp64Arg mutation in the |33-adrenergic receptor gene with obesity-related phenotypes in a Mediterranean population: interaction with a common lipoprotein lipase gene variation. Journal o f Internal Medicine 2001; 250: 348-360. Daling JR, Malone KE, Doody DR, Johnson LG, Gralow JR, Porter PL. Relation o f Body Mass Index to Tumor Markers and Survival among Young Women with Invasive Ductal Breast Carcinoma. Cancer 2001, Volume 92, Number 4. Demark-Wahnefried W, Peterson BL, Winer EP, Marks L, Aziz N, M arcom PK, Blackwell K, Rimer BK. Changes in weight, body composition, and factors influencing energy balance among premenopausal breast cancer patients receiving adjuvant chemotherapy. Journal o f Clinical Oncology, Vol 19, N o 9 (May 1), 2001: pp 2381-2389. Demark-Wahnefried W, Rimer BK, Winer EP. Weight gain in women diagnosed with breast cancer. Journal o f American D iet Association 1997; 97: 519-26. Elledge RM, Clark GM, Chamness GC, et al. Tumor biologic factors and breast cancer prognosis among white, Hispanic, and black women in the United States. Journal o f National Institute 1994; 86; 705-12. Ferrara CM, Kumar M, Nicklas B, M cCrone S and Goldberg AP. Weight gain and adipose tissue metabolism after smoking cessation in women. International Journal o f Obesity (2001) 25, 1322-1326. Forouki P, Dixon JM, Leonard RC, et al: Prospective randomized study o f surgical morbidity following primary systemic therapy for breast cancer. British Journal o f Surgery 82: 79-82, 1995. Friedenreich CM. Review o f anthropometric factors and breast cancer risk. European Journal o f Cancer Prevention 2001, 10, 15-32. Gagnon J, Mauriege P, Roy S, Sjostrom D, Chagnon YC, Dionne FT, Oppert JM, Perusse L, Sjostrom L, Bouchard C. The Trp64Arg M utation o f the 1 3 3 Adrenergic Receptor Gene Has no Effect on Obesity Phenotypes in the Quebec Family Study and Swedish Obese Subjects Cohorts. The American Society fo r Clinical Investigation, Inc. Volume 98, Number 1996, 2086-2093. 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Goodwin PJ, Boyd NF. Body size and breast cancer prognosis: a critical review o f the evidence. Breast Cancer Research and treatment 1990; 16: 205-14. Herbert J, Augustine A, Barone J, Kabat G, Kinne D, Winder E: Weight, height and body mass index in the prognosis o f breast cancer: early results o f a prospective study. International Journal o f Cancer 42: 315-318, 1988. Hoffstedt J, Poirier O, Thorne A, Lonnqvist F, Herrmann SM, Cabien F, Arner P. Polymorphism o f the human beta3-Adrenoceptor gene forms a well-conserved haplotype that is associated with moderate obesity and altered receptor function. D iabetes volume 48(1), January 1999, pp 203-205. Huang X, Hamajima N, Saito T, M atsuo K, Mizutani M, Iwata H, Iwase T, M iura S, Mizuno T, Tokudome S, Tajima K. Possible association o f [32- and 03- adrenergic receptor gene polymorphisms with susceptibility to breast cancer. Breast Cancer Research 2001; 3 (4): 264-269. Hunter CP, Redmond CK, Chen VW, et al. Breast Cancer: factors associated with stage at diagnosis in black and white women. Black/White Cancer Survival Study Group. Journal o f National Cancer Institute 1993; 85: 1129-37. Ingram D, N ottage E, Siobhan N, Sparrow L, Roberts A Willcox D: Obesity and breast cancer- the role o f the female sex hormones. Cancer 64: 1049-1053, 1989. Kumar NB, Cantor A, Allen K, Cox CE. Android obesity at diagnosis and breast carcinoma survival. Cancer Volume 88, Issue 12, 2000, Pg 2751-2757. Li LS, Lonnqvist F, Luthman H, Arner P. Phenotypic characterization o f the Trp64Arg polymorphism in the 03-adrenergic receptor gene in normal weight and obese subjects. Diabetologia 1996; 39: 857-860. Lowe W L Jr, Rotimi CN, Luke A, Guo X, Zhu X, Comuzzie AG, Schuh TS, Halbach S, Kotlar TJ, Cooper RS. The 03-adrenergic receptor gene and obesity in a population sample o f African Americans. International Journal o f Obesity 2001; 25: 54-60. M atthews KA, Abrams B, Crawford S, Miles T, N eer R, Powell LH and Wesley D. Body mass index in mild-life women: relative influence o f menopause, hormone use, and ethnicity. International Journal o f Obesity (2001) 25, 863-873. McCann J. Obesity, Cancer Links Prompt N ew Recommendations. Journal o f the National Cancer Institute, Vol.93, No. 12, 901-902, June 20, 2001. 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. McFarlanne-Anderson N, Bennett F, Wilks R, Howell S, Newsome C, Cruickshank K, Forrester T. The Trp64Arg mutation o f the beta 3-adrenergic receptor is associated with hyperglycemia and current body mass index in Jamaican women. Metabolism 1998; 47: 617-621. M oorman P, Jones B, Millikan R, Hall I, Newman B. Race, anthropometric factors, and stage at diagnosis o f breast cancer. American Journal o f Epidemiology 2001 Feb 1; 153 (3): 284-91. NCCDPHP website: http://www.cdc.gov/nccdphp/dnpa/obesity/epidemin.htm. Newman SC, Lees AW, Jenkins H. The effect o f body mass index and oestrogen receptor level on survival o f breast cancer patients. International Journal o f Epidemiology 1997; 26: 484-90. Oizumi T, Daimon M, Saitoh T, Kameda W, Yamaguchi H, Ohnuma H, Igarashi M, Eguchi H, Manaka H, Tominaga M, Kato T. Genotype Arg/Arg, but not Trp/Arg, o f the Trp64Arg polymorphism o f the [beta]3-Adrenergic Receptor Is Associated With Type 2 Diabetes and Obesity in a Large Japanese Sample. D iabetes Care, Volume 24(9), September 2001, 1579-1583. Parkin DM, Pisani P, Ferlay J. Estimates o f the worldwide incidence o f 25 major cancers in 1990. International Journal o f Cancer 1999, 80: 827-841. Patterson RE, Kristal AR, Tinker LF, Carter RA, Bolton MP, Agurs-Collins T. Measurement characteristics o f the Women's Health Initiative food frequency questionnaire. Annals o f Epidemiology 1999 Apr; 9(3): 178-87. Perry AC, Applegate EB, Jackson ML, Deprima S, Goldberg RB, Ross r, Kempner L, and Feldman BB. Racial differences in visceral adipose tissue but not anthropometric markers o f health-related variables. Journal o f Applied physiology 89: 636-643, 2000. Pietri-Rouxsel F, St. John-Manning B, Gros, J, Strosberg AD. The biochemical effect o f the naturally occurring Trp64-Arg mutation on human 133-adrenoreceptor activity. European Journal o f Biochemistry 1997; 247: 1174-1179. Ries LA, Kosary CL, Hankey BF, et al. SEER Cancer Statistics Review, 1973-1994: tables and graphs. Bethesda, MD: National Cancer Institute, 1997. (NIH publication no. 97-2789). Rose DP, Royak-Schaler R. Tumor biology and prognosis in black breast cancer patients: a review. Cancer Detection Prevention 2001; 25. 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sakane N, Yoshida T, Umekawa T, Sakai Y, Takahashi T, Kondo M: Effects of Trp64Arg mutation in the beta3-adrenergic receptor gene on weight loss, body fat distribution, glycemic control, and insulin resistance in obese type 2 diabetic patients. D iabetes Care 20: 1887-1890, 1997. SAS Institute, Inc. SAS/STAT U ser’s Guide, Version 8. Cary, NC: SAS Institute, Inc, 1999. Segal KR, Van Loan M, Fitzgerald PI, Hodgdon JA and Van Itallie TB. Lean body mass estimation by bioelectrical impedance analysis: a four-site cross-validation study. American Journal o f Clinical Nutrition 1988; 47: 7-14. Senie R, Rosen P, Rhodes P, Lesser M, Kinne D: Obesity at diagnosis o f breast carcinoma influences duration o f disease-free survival. Annals o f Internal Medicine 116: 23-32, 1992. Sipilainen RM, Uusitupa M, Heikkinen S, Rissanen A, Laakso M. Polymorphism o f the [beta]3-adrenergic receptor gene affects basal metabolic rate in obese Finns. D iabetes 1997; 46: 77-80. Sorensen TIA, Echwald SM. Obesity genes— Identifying single genes involved in polygenic inheritance is not easy. BM J Volume 322, 17 M arch 2001. Tannenbaum A (1942). Genesis and growth o f tumors. II. Effects o f caloric restriction per se. Cancer Research 2; 460-7. Tavenier G, Barbe P, Galizky J, Berlan M, Caput D, Lafontan M, Langinl D: Expression o f beta3-adrenoceptors with low lipolytic and human subcutaneous white adipocytes. Journal o f Lipid Research 37: 87-97, 1996. Tchernof A, Starling RD, Walston JD, Shuldiner A, Dvorak RV, Silver K, M atthews DE, Poehlman ET. Obesity-related phenotypes and the beta3-adrenoceptor gene variant in postmenopausal women. D iabetes 1999 Volume 48(7), July 1999, pp 1425-1428. Thompson PA, Ambrosone C. Chapter 7: Molecular Epidemiology o f Genetic Polymorphisms in Estrogen Metabolizing Enzymes in Human Breast Cancer. Journal o f the National Cancer Institute Monographs No. 27, 2000. Toshihide O, M akoto D, Tamotsu S, W ataru K, Hiroshi Y, Hiroshi O, Masahiko I, Hideyuki E, Hideo M, M akoto T, Takeo K. Genotype Arg/Arg, but not Trp/Arg, o f the Trp64Arg polymorphism o f the [beta] 3-Adrenergic Receptor Is Associated With Type 2 Diabetes and Obesity in a Large Japanese Sample. Diabetes Care. Volume 24(9), September 2001, pp 1579-1583. 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Umekawa T, Yoshida T, Sakane N, Kogure A, Kondo M, Honjyo H. Trp64Arg mutation o f beta 3-adrenoceptor gene deteriorates lipolysis induced by beta 3- adrenoceptor agonist in human mental adipocytes. Diabetes. Volume 48 (1): 117-120, January 1999. Van den Brandt PA, Spiegelman D, Yaun S-S, et al. Pooled analysis o f prospective cohort studies on height, weight, and breast cancer risk. American Journal o f Epidemiology 2000; 152: 514-27. Verreault R, Brisson J, Deschenes L, Naud F: Body weight and prognostic indicators in breast cancer- modifying effect o f estrogen receptors. American Journal o f Epidemiology 129: 260-268, 1989. Walston J, Silver K, Bogardus C, Knowler WC, Cell FS, Austin S, Manning B, Strosberg D, Stern MP, Raben N, Sorkin JD, Roth J, Shuldiner AR: Time o f onset o f non-insulin-dependent diabetes mellitus and genetic variation in the [beta]3-adrenergic receptor. New England Journal o f Medicine 333:343-347, 1995 Widen E, Lehto M, Kanninen T, W alston J, Shuldiner AR, Groop LC: Association of a polymorphism in the [beta] 3-adrenergic receptor gene with features o f the insulin resistance syndrome in Finns. New England Journal o f Medicine 333:348-351, 1995. World Health Organization. Obesity: preventing and managing the global epidemic. Genava: WHO, 1998: 1-276. Yoshioka K, Yoshida T, Sakane N, Umekawa T, Takahashi T, Sakai Y, Kondo M. Association o f Trp64Arg mutation o f the beta 3-adrenergic receptor gene with NIDDM, current and maximal body mass index. D iabetologia 1996, 39:1410- 1411. Zhang S, Folsom AR, Sellers TA, Kushi LH, Potter JD. Better breast cancer survival for postmenopausal women who are less overweight and eat less fat. Cancer 1995; 76: 275-83. 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Determinants of mammographic density in African-American, non-Hispanic white and Hispanic white women before and after the diagnosis with breast cancer

PDF

BRCA1 mutations and polymorphisms in African American women with a family history of breast cancer identified through high throughput sequencing

PDF

Family history, hormone replacement therapy and breast cancer risk on Hispanic and non-Hispanic women, The New Mexico Women's Health Study

PDF

A descriptive analysis of medication use by asthmatics in the Children's Health Study, 1993

PDF

Association between body mass and benign prostatic hyperplasia in Hispanics: Role of steroid 5-alpha reductase type 2 (SRD5A2) gene

PDF

Descriptive epidemiology of thyroid cancer in Los Angeles County, 1972-1995

PDF

Dietary fiber intake and atherosclerosis progression: The Los Angeles Atherosclerosis Study

PDF

Comparisons of metabolic factors among gestational diabetes mellitus probands, siblings and cousins

PDF

Evaluation of the accuracy and reliability of self-reported breast, cervical, and ovarian cancer incidence in a large population-based cohort of native California twins

PDF

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

PDF

Androgens and breast cancer

PDF

A pilot survey of medical abortion knowledge and practices among obstetrician/gynecologists and family practitioners in Los Angeles County

PDF

Lymphedema: Impact on breast cancer survivors

PDF

Cigarettes and alcohol in relation to colorectal cancer within the Singapore Chinese Health Study

PDF

A linear model for measurement errors in oligonucleotide microarray experiment

PDF

Association of vitamin D receptor gene polymorphisms with colorectal adenoma

PDF

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

PDF

P53 and bladder cancer outcome: A combined analysis from the Keck School of Medicine

PDF

Development and evaluation of standardized stroke outcome measures in a population of stroke patients in rural China

PDF

Association between latchkey status and smoking behavior in middle school children

Asset Metadata

Creator Teng, Yaling (author)

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

School Graduate School

Degree Master of Science

Degree Program Applied Biostatistics and Epidemiology

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

Tag biology, biostatistics,Black studies,health sciences, oncology,OAI-PMH Harvest,women's studies

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Gilliland, Frank (committee chair), Bernstein, Leslie (committee member), Gauderman, William James (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-302160

Unique identifier UC11341294

Identifier 1414890.pdf (filename),usctheses-c16-302160 (legacy record id)

Legacy Identifier 1414890.pdf

Dmrecord 302160

Document Type Thesis

Rights Teng, Yaling

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