Gut microbial differences in breast and prostate cancer cases from two randomised controlled trials compared to matched cancer-free controls.

Implicated in several chronic diseases, the gastrointestinal microbiome is hypothesised to influence carcinogenesis. We compared faecal microbiota of newly diagnosed treatment-naïve overweight and obese cancer patients and matched controls. Cases were enrolled in presurgical weight-loss trials for breast (NCT02224807) and prostate (NCT01886677) cancers and had a body mass index (BMI) ≥25 kg/m2. Cancer-free controls were matched 1:1 by age (±5 years), race, gender, and BMI (±5 kg/m2). All participants provided faecal samples; isolated bacterial DNA were PCR amplified at the V4 region of the 16S rRNA gene and analysed using the QIIME pipeline. Tests compared cases versus controls, then separately by gender. Microbial alpha-diversity and beta-diversity were assessed, and relative abundance of Operational Taxonomic Units (OTU's) were compared at the genus level, with false discovery rate (FDR) correction. 22 overweight and obese cancer patients were matched with 22 cancer-free controls, with an average BMI of 30.5±4.3 kg/m2, age 54.4±5.3 years, and 54.5% were black. Fourteen matches were made between breast cancer cases and healthy female controls, and 8 matches were made with prostate cancer cases and healthy male controls. Comparison of all cases and controls revealed no differences in alpha diversity, though prostate cancer patients had higher Chao1 (P=0.006) and Observed Species (P=0.036) than cancer-free males. Beta-diversity metrics were significantly different between cases and controls (P<0.03 for all tests in whole sample and in men), though only unweighted Unifrac was different in women (P=0.005). Kruskal Wallis tests indicated significant differences among 16 genera in all matches, 9 in female, and 51 in male. This study suggests the faecal microbiota of treatment-naive breast and prostate cancer patients differs from controls, though larger samples are needed to substantiate these findings. Trial registration: NIH Clinical Trials, NCT01886677, NCT02224807, registered 26 June 2013, 25 Aug 2014 (respectively) - retrospectively registered, https://clinicaltrials.gov/ct2/show/NCT01886677; https://clinicaltrials.gov/ct2/show/NCT02224807.

Disparate results between proliferation rates of surgically excised prostate tumors and an in vitro bioassay using sera from a positive randomized controlled trial.

In vitro bioassay has been used extensively to test the effects of culturing cancer cells in sera from humans participating in dietary interventions, i.e, studies of modified intake of nutrients for the purpose of reducing cancer risk or progression. It has been hypothesized that cell proliferation rates determined by the in vitro bioassay indicate whether modification of dietary intake could decrease cancer cell growth in vivo. It has been suggested, however, that the in vitro bioassay may not correlate with tumor cell proliferation rates in prostate cancer. We investigated the concordance of cell proliferation rates from surgically excised prostate tumor tissue with the in vitro bioassay using sera from matched patients. We used samples from an earlier randomized clinical trial that showed that supplementation with flaxseed significantly inhibited prostate cancer cell proliferation rates in vivo as indicated by Ki67 staining in tumor specimens. Proliferation rates of LNCaP, DU145 and PC3 cell lines cultured in 10% human sera from participants in the flaxseed trial were determined using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Spearman's Rho correlation coefficients (ρ) indicated no association between Ki67 staining in prostate tumors and the in vitro bioassay for the three cell lines. These disparate findings suggest that the in vitro bioassay may not provide an accurate assessment of the environment in vivo.

Weight loss intervention trials in women with breast cancer: a systematic review.

Obesity has been associated with poor health outcomes in breast cancer survivors. Thus, weight loss is recommended for overweight and obese survivors. We systematically reviewed studies (published up to July 2013) that evaluated behaviourally based, weight loss interventions in women with breast cancer exclusively. Completed randomized trials, single-arm trials and ongoing trials were reviewed. Within-group and between-group differences for weight loss were extracted, as was data on secondary outcomes, i.e. clinical biomarkers, patient-reported outcomes, adverse events. Ten completed randomized trials, four single-arm trials and five ongoing trials were identified. Statistically significant within-group weight loss was observed over periods of 2 to 18 months in 13 of the 14 trials, with six randomized and two single-arm trials observing mean weight loss ≥5%. Clinical biomarkers, psychosocial and patient-reported outcomes were measured in a small number of studies. No serious adverse events were reported. Only two trials assessed maintenance of intervention effects after the end-of-intervention and none reported on cost-effectiveness. The studies included in this review suggest that weight loss is feasible to achieve and is safe in women following treatment for breast cancer. Future studies should assess (and be powered for) a range of biomarker and patient-reported outcomes, and be designed to inform translation into practice.

The effects of depression and use of antidepressive medicines during pregnancy on the methylation status of the IGF2 imprinted control regions in the offspring.

In utero exposures to environmental factors may result in persistent epigenetic modifications affecting normal development and susceptibility to chronic diseases in later life. We explored the relationship between exposure of the growing fetus to maternal depression or antidepressants and DNA methylation at two differentially methylated regions (DMRs) of the imprinted Insulin-like Growth Factor 2 (IGF2) gene. Aberrant DNA methylation at the IGF2 and neighboring H19 DMRs has been associated with deregulated IGF2 expression, childhood cancers and several chronic diseases during adulthood. Our study population is comprised of pregnant mothers and their newborns (n = 436), as part of the Newborn Epigenetics Study (NEST). A standardized questionnaire was completed and medical record data were abstracted to ascertain maternal depression and antidepressive drug use. DMR methylation levels in umbilical cord blood leukocytes were quantified using pyrosequencing. From the 436 newborns, laboratory data were obtained for 356 individuals at the IGF2 DMRs, and for 411 individuals at the H19 DMRs; about half of each group was African American or Caucasian. While overall no association between depression and methylation profiles was found, we observed a significant hypermethylation of the H19 DMRs in newborns of African American (n = 177) but not Caucasian (n = 168) mothers who reported the use of antidepressive drugs during pregnancy (ß = +6.89, p = 0.01). Of note, our data reveal a race-independent association between smoking during pregnancy and methylation at the IGF2 DMR (+3.05%, p = 0.01). In conclusion, our findings suggest a race-dependent response related to maternal use of antidepressants at one of the IGF2 DMRs in the offspring.

Pilot study of dietary fat restriction and flaxseed supplementation in men with prostate cancer before surgery: exploring the effects on hormonal levels, prostate-specific antigen, and histopathologic features.

OBJECTIVES: Dietary fat and fiber affect hormonal levels and may influence cancer progression. Flaxseed is a rich source of lignan and omega-3 fatty acids and may thwart prostate cancer. The potential effects of flaxseed may be enhanced with concomitant fat restriction. We undertook a pilot study to explore whether a flaxseed-supplemented, fat-restricted diet could affect the biomarkers of prostatic neoplasia. METHODS: Twenty-five patients with prostate cancer who were awaiting prostatectomy were instructed on a low-fat (20% of kilocalories or less), flaxseed-supplemented (30 g/day) diet. The baseline and follow-up levels of prostate-specific antigen, testosterone, free androgen index, and total serum cholesterol were determined. The tumors of diet-treated patients were compared with those of historic cases (matched by age, race, prostate-specific antigen level at diagnosis, and biopsy Gleason sum) with respect to apoptosis (terminal deoxynucleotidyl transferase [TdT]-mediated dUTP-biotin nick end-labeling [TUNEL]) and proliferation (MIB-1). RESULTS: The average duration on the diet was 34 days (range 21 to 77), during which time significant decreases were observed in total serum cholesterol (201 +/- 39 mg/dL to 174 +/- 42 mg/dL), total testosterone (422 +/- 122 ng/dL to 360 +/- 128 ng/dL), and free androgen index (36.3% +/- 18.9% to 29.3% +/- 16.8%) (all P <0.05). The baseline and follow-up levels of prostate-specific antigen were 8.1 +/- 5.2 ng/mL and 8.5 +/- 7.7 ng/mL, respectively, for the entire sample (P = 0.58); however, among men with Gleason sums of 6 or less (n = 19), the PSA values were 7.1 +/- 3.9 ng/mL and 6.4 +/- 4.1 ng/mL (P = 0.10). The mean proliferation index was 7.4 +/- 7.8 for the historic controls versus 5.0 +/- 4.9 for the diet-treated patients (P = 0.05). The distribution of the apoptotic indexes differed significantly (P = 0.01) between groups, with most historic controls exhibiting TUNEL categorical scores of 0; diet-treated patients largely exhibited scores of 1. Both the proliferation rate and apoptosis were significantly associated with the number of days on the diet (P = 0.049 and P = 0.017, respectively). CONCLUSIONS: These pilot data suggest that a flaxseed-supplemented, fat-restricted diet may affect prostate cancer biology and associated biomarkers. Further study is needed to determine the benefit of this dietary regimen as either a complementary or preventive therapy.

Changes in weight, body composition, and factors influencing energy balance among premenopausal breast cancer patients receiving adjuvant chemotherapy.

PURPOSE: Weight gain is a common problem among breast cancer patients who receive adjuvant chemotherapy (CT). We undertook a study to determine the causes of this energy imbalance. PATIENTS AND METHODS: Factors related to energy balance were assessed at baseline (within 3 weeks of diagnosis) and throughout 1 year postdiagnosis among 53 premenopausal women with operable breast carcinoma. Thirty-six patients received CT and 17 received only localized treatment (LT). Measures included body composition (dual energy x-ray absorptiometry), resting energy expenditure (REE; indirect calorimetry), dietary intake (2-day dietary recalls and food frequency questionnaires) and physical activity (physical activity records). RESULTS: Mean weight gain in the LT patients was 1.0 kg versus 2.1 kg in the CT group (P =.02). No significant differences between groups in trend over time were observed for REE and energy intake; however, a significant difference was noted for physical activity (P =.01). Several differences between groups in 1-year change scores were detected. The mean change (+/- SE) in LT versus CT groups and P values for uncontrolled/controlled (age, race, radiation therapy, baseline body mass index, and end point under consideration) analysis are as follows: percentage of body fat (-0.1 +/- 0.4 v +2.2 +/- 0.6%; P =.001/0.04); fat mass (+0.1 +/- 0.3 v +2.3 +/- 0.7 kg; P =.002/0.04); lean body mass (+0.8 +/- 0.2 v -0.4 +/- 0.3 kg; P =.02/0.30); and leg lean mass (+0.5 +/- 0.1 v -0.2 +/- 0.1 kg; P =.01/0.11). CONCLUSION: These data do not support overeating as a cause of weight gain among breast cancer patients who receive CT. The data suggest, however, that CT-induced weight gain is distinctive and indicative of sarcopenic obesity (weight gain in the presence of lean tissue loss or absence of lean tissue gain). The development of sarcopenic obesity with evidence of reduced physical activity supports the need for interventions focused on exercise, especially resistance training in the lower body, to prevent weight gain.

Zinc alpha-2-glycoprotein is expressed by malignant prostatic epithelium and may serve as a potential serum marker for prostate cancer.

Zinc alpha-2-glycoprotein (ZAG) is a M(r) 41,000 glycoprotein secreted by a variety of normal epithelia. ZAG was recently shown to stimulate lipolysis in adipocytes, leading to the development of cachexia in animals with ZAG-producing tumors. To understand the possible contribution of ZAG to the development of cachexia in men with prostate cancer, ZAG production by normal and malignant prostate tissue was investigated using immunohistochemical assays. Anti-ZAG monoclonal antibodies reacted strongly with normal prostate epithelium but not with other components of prostate or seminal vesicles. The majority of prostate cancers tested (35 of 48; 73%) also reacted with anti-ZAG antibodies. High-grade tumors expressed significantly less ZAG than moderate-grade tumors (mean ZAG score 1.1 versus 1.9; P < 0.01). Men with ZAG-producing prostate carcinomas had elevated levels of serum ZAG relative to their normal age- and race-matched controls (P < 0.02). Furthermore, s.c. growth of human ZAG-producing murine tumors in syngeneic mice and orthotopic growth of ZAG-producing human prostate carcinomas in nude rats resulted in readily detectable levels of human ZAG in the serum. Taken together, these studies show that ZAG production by prostate cancer can lead to systemically elevated serum ZAG levels that may be useful diagnostically. The effects of elevated systemic ZAG on cachexia-associated complications in patients with advanced prostate cancer deserves additional investigation.

Effect of mammalian lignans on the growth of prostate cancer cell lines.

BACKGROUND: Mammalian lignans, enterolactone (EL) and enterodiol (ED), have been shown to inhibit breast and colon carcinoma. To date, there have been no reports of the effect of lignans on prostatic carcinoma. We investigated the effects of ED and EL on three human prostate cancer cell lines (PC-3, DU-145 and LNCaP). MATERIALS AND METHODS: Cells were treated with either 0.1% (v/v) DMSO (vehicle) or 10-100 microM of EL, ED or genistein (positive control) for 72 hours. Cell viability was measured by the propidium iodide nuclei staining fluorometric assay with each assay performed in triplicate. RESULTS: At 10-100 microM, EL significantly inhibited the growth of all cell lines, whereas ED only inhibited PC-3 and LNCaP cells. While EL was a more potent growth inhibitor than ED, both were less potent than genistein. The dose for 50% growth inhibition of LNCaP cells (IC50) by EL was 57 microM, whereas IC50 was 100 microM for ED, (the observed IC50 for genistein was 25 microM). CONCLUSION: ED and EL suppress the growth of prostate cancer cells, and may do so via hormonally-dependent and independent mechanisms.

Partnering with African American churches to achieve better health: lessons learned during the Black Churches United for Better Health 5 a day project.

African Americans (AAs) are at increased risk for many diseases, including cancer, but health promotion efforts often fail to reach them. Effective partnerships can be established with African American churches to deliver health-based interventions. In an NCI-funded study aimed at increasing fruit and vegetable consumption among rural AAs, investigators at three academic institutions and the North Carolina State Health Department partnered successfully with 50 churches to promote dietary change. This study adds to the increasing body of research in support of the African American church as an able partner in delivering health-based interventions. In conducting interventions and research through this channel, the health professional should gain support from regional secular associations; respect the power of the pastor; incorporate the strengths of the congregation; respect the mission of the church; establish open communication and develop trust; provide ample support and training to assure fidelity to interventions and integrity of data; and be patient and persevere.

Psychological impact of diagnosis and risk reduction among cancer survivors.

Life-threatening health events prompt psychological distress that may motivate individuals to reduce health risks. If so, interventions timed to take advantage of these 'teachable moments' could be particularly effective. To explore this association, early stage prostate and breast cancer patients were identified from a hospital-based tumor registry within 6 years of diagnosis. These patients (n=920) completed a mailed survey assessing the Horowitz impact of events scale, risk behaviors and readiness to change the behaviors. Breast cancer patients, younger patients and those reporting poor health status reported the greatest impact of the cancer diagnosis. Impact was inversely associated with time from diagnosis for prostate, but not breast cancer patients. Prostate patients who reported exercising regularly had lower impact scores than those who were not exercising (medians: 0.13 vs 0.56, respectively; p=0.02). Breast patients who were eating five or more fruits and vegetables reported lower impact scores than those who were not eating the recommended servings (0.75 vs 1.06, respectively; p=0.03). Breast patients who were non-smokers reported lower impact scores than smokers (0.88 vs 1.31, respectively; p=0. 02). Prospective studies are needed to understand the psychological impact of cancer diagnosis and how it might facilitate or impede the adoption of health promoting behaviors.

The North Carolina Black Churches United for Better Health Project: intervention and process evaluation.

The North Carolina Black Churches United for Better Health project was a 4-year intervention trial that successfully increased fruit and vegetable (F&V) consumption among rural African American adults, for cancer and chronic disease prevention. The multicomponent intervention was based on an ecological model of change. A process evaluation that included participant surveys, church reports, and qualitative interviews was conducted to assess exposure to, and relative impact of, interventions. Participants were 1,198 members of 24 intervention churches who responded to the 2-year follow-up survey. In addition, reports and interviews were obtained from 23 and 22 churches, respectively. Serving more F&V at church functions was the most frequently reported activity and had the highest perceived impact, followed by the personalized tailored bulletins, pastor sermons, and printed materials. Women, older individuals, and members of smaller churches reported higher impact of certain activities. Exposure to interventions was associated with greater F&V intake. A major limitation was reliance on church volunteers to collect process data.

Current health behaviors and readiness to pursue life-style changes among men and women diagnosed with early stage prostate and breast carcinomas.

BACKGROUND: Cancer survivors are at increased risk for developing secondary tumors, diabetes, osteoporosis, and cardiovascular disease, thus making them an important target population for health-related interventions. However, little is known regarding cancer survivors' behaviors and their interest in pursuing healthier diets, exercise, and smoking cessation. METHODS: A 65-item survey was mailed to 1667 patients diagnosed with early stage carcinoma of the breast or prostate that requested data regarding demographics, health behaviors, stage of readiness for smoking cessation, exercise, increased fruit and vegetable (F&V) consumption, and decreased fat intake, as well as interest in specific health programs and when, relative to diagnosis, these programs should be offered. RESULTS: Responses from 978 cancer survivors suggested that the overwhelming majority (85%) reported their health as good to excellent. The majority of respondents (55%) ate fewer than 5 daily servings of F&Vs, with prostate carcinoma patients reporting eating significantly fewer servings than patients with breast carcinoma (P < 0.001). In contrast, 69% of respondents reported adherence to a low fat diet, again with lower practice rates among prostate carcinoma patients than patients with breast carcinoma (P = 0.006). The majority of respondents (58%) reported routine exercise and 8% were current smokers. Significantly more breast carcinoma patients than prostate carcinoma patients were current smokers (P = 0.03). There was strong interest in health promotion programs across behaviors, and a fair amount of concordance between low fat dietary behaviors and exercise, as well as consumption of five daily servings of F&Vs. The majority of patients expressed a preference for programs that could be delivered via mailed brochure, with 57% of respondents indicating that programs should be initiated at the time of diagnosis or soon after (within 6 months). CONCLUSIONS: The results of the current study suggest that although many cancer survivors already practice healthy life-style behaviors, there is a substantial proportion who do not. Interventions especially are needed to increase daily F&V intake; however, given the fair degree of clustering between dietary and exercise behaviors, multiple risk factor interventions also may be warranted. Among cancer survivors, receptivity is high for health promotion programs, especially those that can be delivered by mail and soon after diagnosis.

Baseline fruit and vegetable intake among adults in seven 5 a day study centers located in diverse geographic areas.

OBJECTIVE: To examine baseline rates of fruit and vegetable consumption among adults in the 5 A Day research trials in order to identify any regional and sociodemographic differences associated with daily servings. DESIGN: The main outcome measure was the frequency of fruits and vegetables consumed within 1 month of the baseline survey as assessed by a 7-item food frequency questionnaire (FFQ). SUBJECTS/SETTING: Participants (N = 15,060) were from 7 study centers. Study centers included schools (N = 48), worksites (N = 60), churches (N = 50), or the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) clinics (N = 15) in interventions to increase fruit and vegetable consumption. STATISTICAL ANALYSES: Means and standard errors, adjusting for clusters, were calculated. A mixed linear model analyzed relationships between fruit and vegetable consumption and regional center, gender, age, race, education, income, marital status, food-shopping responsibility, and whether one lives with children. RESULTS: Results indicate an overall mean intake of 3.6 daily servings of fruits and vegetables. Significant differences in mean daily servings were found among the regional study centers (low of 3.0 to high of 4.1). There were significant differences in mean daily consumption by age (< 30 years = 3.7 servings per day; 30 to 49 years = 3.4; > or = 50 years = 3.7), education (> high school = 3.4 servings per day; high school graduate = 3.4; some college = 3.5; college graduate = 3.9), race (black = 3.7 servings per day; Hispanic = 3.0; white = 3.6; other = 3.7), marital status (married = 3.6 servings per day; single = 3.5), and food-shopping responsibilities (little = 3.2 servings per day; about half = 3.6; most = 3.8). Only 17% of respondents ate 5 or more servings of fruits and vegetables per day. CONCLUSIONS: The 7 regions showed significant variability in daily fruit and vegetable consumption, suggesting that a single national message to increase fruit and vegetable consumption may not reach the population segments most in need of changing. It is advisable to spend more time understanding the food consumption habits of the population under investigation to develop messages to foster behavior change.

Fruit and vegetable consumption and prevention of cancer: the Black Churches United for Better Health project.

OBJECTIVES: This study assessed the effects of the Black Churches United for Better Health project on increasing fruit and vegetable consumption among rural African American church members in North Carolina. METHODS: Ten counties comprising 50 churches were pair matched and randomly assigned to either intervention or delayed intervention (no program until after the follow-up survey) conditions. A multicomponent intervention was conducted over approximately 20 months. A total of 2519 adults (77.3% response rate) completed both the baseline and 2-year follow-up interviews. RESULTS: The 2 study groups consumed similar amounts of fruits and vegetables at baseline. AT the 2-year follow-up, the intervention group consumed 0.85 (SE = 0.12) servings more than the delayed intervention group (P < .0001). The largest increases were observed among people 66 years or older (1 serving), those with education beyond high school (0.92 servings), those widowed or divorced (0.96 servings), and those attending church frequently (1.3 servings). The last improvement occurred among those aged 18 to 37 years and those who were single. CONCLUSIONS: The project was a successful model for achieving dietary change among rural African Americans.

Prostate cancer: demographic and behavioral correlates of stage at diagnosis among blacks and whites in North Carolina.

OBJECTIVES: Although stage at diagnosis is one of the most important predictors of survival from prostate cancer, demographic factors, screening practices, and knowledge and beliefs associated with stage at diagnosis have not been well documented, particularly by race. METHODS: We conducted telephone interviews with 117 black and 114 white men diagnosed with prostate cancer to identify the demographic factors, healthcare-seeking behaviors, and prostate cancer-related knowledge, attitudes, and practices associated with stage. The sample was stratified by stage at diagnosis and was composed of men 50 to 74 years old who resided in a contiguous 63-county region in North Carolina and who were diagnosed at 1 of 16 participating hospitals. RESULTS: Among blacks, stage was inversely correlated with income (P = 0.04) and health insurance status (P < or = 0.001); among whites, stage was not associated with income or health insurance status, but approached significance with marital status (P = 0.06). Awareness of prostate cancer before diagnosis tended to decline with advancing stage among black men (P = 0.07), but was high for all stages (greater than 93%) among whites. Report of a prostate-specific antigen screen was inversely correlated with stage among black men (P = 0.01); a trend was observed among whites but was not significant (P = 0.20). Knowledge of prostate cancer risk factors was not significantly associated with stage for blacks or whites. Less than one third of men in each race and stage group knew that black men are at increased risk of prostate cancer. CONCLUSIONS: Demographic and other factors vary with stage and should be considered when designing and targeting interventions to reduce late diagnosis of prostate cancer.

Utility of produce ratios to track fruit and vegetable consumption in a rural community, church-based 5 A Day intervention project.

Previous research suggests that grocery store characteristics may be useful in evaluating community-based dietary interventions. We undertook a study to determine whether produce ratios (ratios of produce sales to total grocery sales) were a useful indicator of fruit and vegetable (F & V) consumption in a church-based, community intervention trial that promoted 5 A Day guidelines within 10 rural counties of North Carolina. Produce ratios were collected from stores identified by participants in the Black Churches United for Better Health Project. Baseline and study period data for 21 stores in intervention counties and 18 stores in nonintervention counties were compared using repeated-measures analysis of variance. Produce ratios were significantly associated with seasonality (p < 0.0001), but no differences were seen between the two groups of stores. These findings do not support data from individual telephone surveys, which showed significant differences in F & V consumption between participants in the two groups. Our inability to detect differences at the store level may have been due to 1) the incapacity of produce ratios to capture F & V purchases that were juice, frozen, or canned products; 2) shifts in procuring F & Vs from grocery stores to other sources (i.e., gleaning and produce cooperatives); 3) the modest proportion of shoppers that received the full intervention dose; and 4) a general lack of power to detect differences at the store level. Therefore, although produce ratios did not serve as a valid measure for this project, if their limitations are recognized and compensated for, they may have applicability for future investigations that monitor F & V consumption.

Environmental contaminants and body fat distribution.

The effect of body mass index (BMI) and waist:hip ratio (WHR) on plasma levels of organochlorines [i.e., 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDE)] was investigated in a sample of black and white women drawn from a population-based study in North Carolina. Organochlorine levels were determined in plasma samples from 99 women selected on the basis of race (black versus white) and quartile of the WHR (1st versus 4th). Of a panel of 20 organochlorine compounds tested, only DDE was detectable in most study subjects. Measurements of height, weight, and waist and hip circumferences were taken during an in-person interview. Information was elicited regarding dietary, residential, and breast-feeding histories. Results of multiple regression analyses indicate that black women had significantly higher plasma levels of DDE than white women. These levels were independent of BMI and WHR. BMI but not WHR was also found to be an independent predictor of DDE plasma level. These results suggest that black/white differences should be considered in studies that explore the relationship between environmental contaminants and various disease outcomes, such as breast cancer risk. In addition, BMI may affect circulating levels of contaminants and should also be considered a potentially important modifying factor for exposure to lipophilic substances.