Etiologic factors such as vitamin D and estrogen are potentially related to breast cancer development, although details of their mechanisms are not completely understood. We prospectively investigated correlates of breast cancer risk among postmenopausal women in the Study of Osteoporotic Fractures (SOF). First, we undertook a case-cohort study to test the hypothesis that low serum 25-hydroxyvitamin D [25(OH)D] will be associated with an increased risk of ER+ breast cancer (N=502). Low 25(OH)D levels were not associated with an increased risk of breast cancer and do not support an association between 25(OH)D and ER+ breast cancer development. Second, we utilized fractional calcium absorption (FCA) as a marker of tissue resistance to vitamin D to test the hypothesis that low FCA will be associated with an increased risk of breast cancer (N=5035). To the contrary, over a mean 9.6 years, increasing rates of FCA were associated with a higher risk of invasive breast cancer. A stronger positive relationship was noted among women with low dietary calcium intake. The findings support a modestly increased risk of breast cancer with higher FCA rates particularly among those who have low calcium intake. Finally, we examined the long-term association of an initial bone mineral density (BMD) measure and change in BMD (annual percent change assessed 3.5 years later) on breast cancer risk (N=5385). Furthermore, we tested the hypothesis that the risk associated with an initial BMD measure would be strengthened by the addition of the change variable. Over a mean 9.5 years, there was no association between increasing levels of BMD, change in BMD, or a combined model and breast cancer. The effect of BMD was found to be dependent upon family history of breast cancer. Among women with a positive family history, high BMD was associated with a 3-fold higher risk of breast cancer compared to low BMD. Through our investigations of two etiologic factors and their association with breast cancer development, we have enhanced our knowledge regarding the interdependence of vitamin D, calcium, and estrogen. These findings may lead to improved opportunities for prevention and early detection and are of significant public health relevance.

Background: Major aims of the psychological research in PCOS were: (1) to compare the prevalence of depressive symptoms in women with PCOS to controls; (2) to determine whether depression and psychological traits (anger, anxiety, hostility/cynicism) and satisfaction with life are associated with PCOS (3) to determine the prevalence of Metabolic Syndrome (MS) in PCOS cases and controls (4) to determine if psychological factors are risk factors for Metabolic Syndrome independent of age, marital status, education, and parity (5) to determine if psychological factors, independent of baseline cardiovascular risk factors, are a risk factor for IMT in women with PCOS. Design: Prospective and cross sectional. Methods: Cases (n=161) and controls (n=161) matched on age, race, and neighborhood a subset of the Cardiovascular Health and Risk Measurement study (CHARM) investigating coronary heart disease risk factors in women with PCOS. Psychological measures were Beck Depression Inventory I (BDI I), Speilberger Trait Anger and Anxiety Scales, Cook-Medley Scale, Diener Satisfaction with Life Scale. Results: PCOS women had a higher prevalence of depression (BDI scores > 9; predominately a mild level of depression): 31% vs. 17% in controls (P=.016; OR 1.9; CI 1.55-2.16). Within cases, BMI, education, and parity were statistically significant predictors of depression, p<.05. The odds of being depressed (at least mild severity) increased by 6% for each unit increase of BMI, the odds of being depressed decreased by 20% for each year of education, and the odds of being depressed increased by 44% for parity (per live birth). The odds of having PCOS increased with each unit of BDI score by 1.06 times, adjusting for marital status, BMI, smoking, and education (entire sample). Results from MS, 27% (n=40) of the cases and 9.9% (n=15) of the controls had MS (p<.05; OR 3.4; CI 1.726-6.400. The odds of having MS increased by 10% for each year of age and threefold for PCOS cases. There were no effects from the psychological factors independent baseline cardiovascular risk factors on IMT. Conclusions: Depression is a major psychological concern in PCOS. Public Health Significance: Women with PCOS should be screened on diagnosis and monitored closely for depression.

Prostate cancer incidence and mortality rates continue to be higher among African Americans than Caucasians. While psychosocial factors may explain some of the disparities, the role played by genetic differences in the two racial groups is not so clear. Emerging evidence suggests an important role of chronic or recurrent inflammation in prostate carcinogenesis. Interleukin-1 (IL-1) and IL-6 are inflammatory genes reported to be associated with prostate cancer risk. Interleukin-1 and IL-6 cytokines also decrease bone mineral density (BMD) by inducing osteoclasts to resorb bone matrix. We sought to determine if genotypes of IL-1A, IL-1B, IL-1RN, IL-6 and IL-6R were associated with prostate cancer risk, as well as with selected risk factors, in the two racial groups. We examined allele frequency distributions of polymorphisms in IL-1A, IL-1B, IL-1RN, IL-6 and IL-6R genes in a cross-sectional study of African American and Caucasian men ages 40 to 80 years old. We also assessed the associations of genotypes of these inflammatory genes and the risk of prostate cancer in a case-control study of the two racial groups. Additionally, we evaluated the associations of bone mineral density and prostate cancer in our sample. We found racial differences in minor allele frequencies, as well as in the associations of single nucleotide polymorphisms of inflammatory genes IL-1 and IL-6 and prostate cancer. We also found associations of IL-1 and IL-6 genotypes and prostate cancer. Additionally, we found an inverse association of BMD and prostate cancer in both racial groups. Our findings support a growing body of evidence that chronic or recurrent inflammation play an important role in prostate carcinogenesis, and the possibility of ethnic based differences in susceptibility. Understanding the role of IL-1 and IL-6 genes in the development of prostate cancer is of great public health significance because it will enable their possible use as biomarkers for early detection and prompt intervention, increase our understanding of the molecular biology of the disease, open up new avenues for prevention and treatment, as well as explain some of the observed disparities in the disease.

Estrogen metabolites may play an important role in breast carcinogenesis. Animal and in vitro studies suggest differing biological effects of the 2-hydroxyestrone (2-OHE1) and 16&alpha;-hydroxyestrone (16&alpha;-OHE1) metabolites, lending support to the use of 2:16&alpha;-OHE1 as a measure of estrogen balance. Although previous studies have evaluated the association between these specific metabolites and breast cancer among pre- and postmenopausal women, the results have been inconclusive. The sample size of individual studies is often small and lacks the statistical power to draw conclusions or to adequately assess the relationships within subgroups. Furthermore, the relationship between various lifestyle factors and personal characteristics and estrogen metabolites remains unclear. We evaluated the association between the 2-OHE1, 16&alpha;-OHE1 and 2:16&alpha;-OHE1 metabolites and breast cancer among premenopausal (183 cases/548 controls) and postmenopausal (319 cases/647 controls) women using a combined analysis of individual level data from previously published research studies. In separate study adjusted conditional logistic regression models matched on 5-year age groups, higher levels of 2:16&alpha;-OHE1 were not associated with breast cancer among pre- or postmenopausal women [Premenopausal: OR&ge;2.67 vs. <1.76=0.81 (95% CI: 0.49, 1.32); Postmenopausal: OR&ge;2.46 vs. <1.53=0.87 (95% CI: 0.58, 1.29)]. Using multivariable regression analyses adjusted for study, we evaluated various predictors of estrogen metabolites among the control populations of the participating studies (544 premenopausal/720 postmenopausal). Among premenopausal women, BMI was negatively associated with 2-OHE1 and 2:16&alpha;-OHE1 (p < 0.05). Analyses among postmenopausal women revealed significant associations (p<0.05) between age, age at menopausal status, and history of benign breast disease. In summary, this combined analysis does not support an association between urinary estrogen metabolites and breast cancer among pre- or post menopausal women. However, our results do suggest potential differences in factors related to estrogen metabolite levels among pre- and postmenopausal women. Enhancing our knowledge of estrogen metabolites among breast cancer patients and among healthy populations of women is a significant contribution to public health. By improving our understanding of estrogen metabolites we may be able to identify women at higher risk of breast cancer as well as increase our understanding of breast cancer etiology.

Diabetes is a complex chronic disease with many causes, complications and management needs. It affects a large proportion of people of varying ages, income levels, races/ethnicities and geographic areas. Approximately 7.0% of Pennsylvanians have been diagnosed with diabetes. Diabetes is a major public health challenge due to the enormous impact on the affected individual, their families and the health care system. However, recent research has shown that diabetes related mortality and morbidity can be prevented or delayed by controlling risk factors. Certain environmental aspects play an important role in the prevention and treatment of chronic diseases such as diabetes. In order to provide the public health community with another tool to enhance our understanding of the factors that affect the numbers and types of diabetes cases in Pennsylvania, it is important that we undertake a project that will support the analysis of geographic in terms of associated risk factors. This study proposed to investigate geographical patterns of diabetes hospitalizations, risk factors for diabetes complications and glycemic control among individuals with type 2 diabetes in rural regions. Residents of more rural counties are 11% more likely to be hospitalized for uncontrolled diabetes compared to those living in areas that are less rural for every increase in rurality ranking. Furthermore, we demonstrated that there is a clear association between the presence of food stores, food service places, and health care locations with risk factors for diabetes complications among individuals with diabetes. Our findings also indicated that those who live more than ten miles from their diabetes management center are 88% more likely to have an HbA1c level greater than 7.0% compared to those who live less than ten miles from their center, adjusted for individual-level and community level factors. Results demonstrated that for every mile the subjects live from their diabetes management center, they are 2% more likely to have an HbA1c level greater than 7.0%. These findings are of public health significance because they demostrate a clear association between the built envirnment and diabetes hospitalizations, risk factors for diabetes complications and glycemic control among individuals with diabetes in rural regions.

There is a general misperception that sexuality and sexual orientation are not important in the lives of older adults. Older gay and bisexual men remain an invisible population in which health behaviors and health status are less well-known. While an enormous body of research has documented the disparities of substance use, depression, and HIV prevalence among Men who have Sex with Men (MSM) compared to heterosexual men, less is known about how patterns of substance use, depression, and risky sexual behavior evolve as MSM age. Using a semi-parametric, group-based approach, this study investigated the effect of aging on risky sexual behavior, depression, and illicit drug use in a cohort of middle-aged and older MSM in the Pitt Mens Study. In the three presented manuscripts, results indicate that the majority of participants did not have unprotected anal intercourse partners or use illicit drugs, and reported very low depressive symptoms across the age range of the study. At the same time, the analysis identified trajectory groups with respect to midlife exposure to risky sexual behavior, chronic depressive symptoms, and multiple illicit drug use. Several correlates such as socio-demographic variables, health behaviors (smoking and binge drinking), and psychological variables (HIV-related attitudes) were differentially associated with the trajectory groups. The results of the study refuted the negative stereotypes of older gay and bisexual men as sad, undesirable, and depressed, yet the study managed to provide a more realistic picture of the development of mental health and HIV risk behaviors of this cohort from middle to early old age. This confirms the diversity and heterogeneity of this population noted by previous researchers. Public Health Significance: Using this epidemiologic approach, future researchers can identify subgroups of men who are at most risk for depression, substance abuse, and risky sexual behaviors over an extended period of time. As a result, public health resources such as prevention and intervention programs can be allocated to these men in a cost-effective way. In addition, investigations of the risk factors associated to each trajectory group may give clues to different etiologies of group characteristics among a cohort of aging sexual minorities.

Breast density affects mammographic sensitivity and is predictive of breast cancer risk. Factors that increase breast density may compromise the reduction in mortality gained by mammographic screening. Understanding these factors is crucial as it may help us improve mammographic screening and reduce breast cancer risk. Prolactin, an endogenous hormone that acts as a mitogen and differentiating agent in the breast, may be one such factor. To our knowledge, this is the first study to examine the association between prolactin and mammographic breast density in a cross-sectional study of healthy, cancer-free postmenopausal women. A weak, but statistically significant correlation was observed between prolactin and percent breast density (spearman correlation coefficient of 0.1197; p-value 0.013) after adjusting for every being pregnant and ever breast feeding. Prolactin is most likely one of several factors that contribute to increased mammographic breast density, and further analyses are needed to determine its full contribution. No statistically significant associations were observed for the prolactin gene single nucleotide polymorphisms (SNPs) examined in relation to prolactin, percent breast density, or proportion of dense breast area. However, two SNPs in the prolactin receptor gene (rs7734558 and rs7705216) were significantly associated with serum prolactin level at the 0.10 significance level. Women with the G allele (AG and GG) at SNP rs7734558 have a slightly elevated level of prolactin when compared with women homozygous for the A allele (AG 10.76 ± 6.40 ng/mL, GG 10.77 ± 4.60 ng/mL vs. AA 9.86 ± 6.32 ng/mL); and those with the GG allele at SNP rs7705216 have a slightly elevated prolactin level when compared with individuals with the C allele (GG 11.71 ± 2.78 ng/mL vs. CG 11.15 ± 6.22 ng/mL, CC 10.27 ± 5.99 ng/mL). These SNPs need to be further investigated to determine their full contribution in relation to serum prolactin levels. Having an understanding of factors that affect breast density is an important public health issue has it may lead to improvements in breast cancer screening and help identify not only women at an increased risk for breast cancer, but women who may benefit from prevention strategies.

Bacterial vaginosis (BV) is a highly prevalent vaginal condition that has been associated with a number of pregnancy complications, including spontaneous preterm births (sPTB). A number of risk factors for BV have been identified yet its etiology is not understood. Few studies have investigated the role of nutrition in the etiology of BV. Maternal iron and folate status may be important in the development of BV as they play key roles in host immunity. The purpose of this review is to critically evaluate the literature and provide support to the hypothesis that certain micronutrients, iron, and folate, are risk factors for BV early in pregnancy. Iron and folate may be important in both humoral and cell-mediated innate immunity, respectively. Both facets of innate immunity are important in the control of BV. Preliminary data suggests that vitamin D is associated with BV early in pregnancy. In a prospective cohort study, pregnant women enrolled at <16 weeks were followed through delivery. Serum collected at enrollment was analyzed for 25-hydroxyvitamin D (25(OH)D) and Gram-stained vaginal smears were evaluated for BV using Nugent criteria. Multivariable logistic regression was used to determine the association between 25(OH)D and BV while adjusting for a number of confounders. The association between vitamin D and BV varied by race (likelihood ratio test, p=0.09). A 50-nmol/l decrease in 25OHD was associated with a 4.2-fold (95% CI, 2.1, 8.1) increase in the odds of BV in black women. Among white women, there was no association between maternal 25(OH)D and BV. These results indicate that maternal vitamin D deficiency is associated with BV in early pregnancy among black, but not white women. Studies such as these are of great public health significance because maternal nutrition is modifiable, and interventions to improve maternal nutritional status can be safe, inexpensive, and readily acceptable to patients.

Insulin resistance (IR) is well documented in type 1 diabetes (T1D) and is theorized to relate to diabetes complications, including renal and coronary artery disease (CAD). The hyperinsulinemic-euglycemic clamp technique provides accurate assessment of IR, yet the laborious, costly, and invasive nature of this technique is often inappropriate for large investigations. Increasing use of the Estimated Glucose Disposal (eGDR) equation in T1D makes further examination of this equation desirable as it may be improved with additional assessments. Leg adiposity has been favorably associated with IR and cardiovascular risk, but whether this protective tendency is similar in T1D populations is unknown. This dissertation examines whether diabetes complications or additional clinical factors (i.e. regional adiposity distribution) contributes to the estimation of IR in T1D. Differences in regional adiposity, and the extent to which these differences influence IR, were examined in T1D and individuals without diabetes. Associations between CAD risk factors and regional adiposity were also investigated in individuals with T1D. No differences in IR were observed between T1D individuals with CAD or renal disease. All adiposity measures were detrimentally associated with IR, however, general obesity most strongly predicted IR in this population. Despite lower levels of adiposity, more severe IR was observed in individuals with T1D compared to non-diabetic individuals. Leg adiposity was favorably associated with presence of CAD, even after controlling for general obesity, but this association was only observed in non-diabetes and in T1D individuals who were obese. Trunk and leg fat displayed equal yet opposite associations with CAD risk factors and increasing leg adiposity was associated with decreased risk for the presence of CAD in females with T1D. This dissertation thus yields significant Public Health findings by providing evidence that IR is a prominent feature in T1D, is largely driven by adiposity, and can be estimated using clinical measures. Furthermore, the finding that leg adiposity was favorably associated with presence of CAD in individuals with T1D provides impetus to further study and underscores the complex association of adiposity with morbidity in T1D.

Globally, cardiovascular disease (CVD) is the leading cause of death. Increased risk for CVD can be attributed to smoking, high blood pressure, poor lipid profiles, obesity and psychosocial factors. Markers of subclinical CVD are non-invasive measures that detect early atherosclerotic changes. The purpose of this dissertation was to evaluate novel risk factor associations for subclinical CVD in three distinct populations. The protective effect of HDL-c for subclinical CVD was diminished in a population of postmenopausal women compared to premenopausal women. Furthermore, the concentration of small HDL particles was higher among postmenopausal women. Lipid profile changes with the menopausal transition may in part explain the increased risk of CVD seen after menopause. The protective effect of education for subclinical CVD was evident only among females from an Afro-Caribbean population. Educational differences in SBP and lipids varied for males and females providing insight into potential mechanisms for the education-subclinical CVD relationship observed on the island of Tobago. Tonic cardiac sympathetic activity and parasympathetic reactivity were independent predictors of subclinical CVD in a population of overweight and obese young adults. The effect of C-reactive protein (CRP) on subclinical CVD is potentially explained by the autonomic anti-inflammatory mechanisms linking heart rate variability and CRP. Identifying novel risk factor associations for subclinical CVD in various populations supports the important public health objective of reducing the global burden of CVD morbidity and mortality through early detection of atherosclerosis.