Periodontitis is a multifactorial disease with microbial dental plaque as the etiological agent. The manifestation and progress of periodontitis is influenced by a wide variety of determinants and factors, including subject characteristics, social and behavioral factors, systemic factors, genetic factors, the microbial composition of dental plaque, and others. The pathogenesis of periodontal disease results in resorption of alveolar bone and loss of the attachment apparatus to the teeth. There is a biological potential that periodontal destruction may be influenced by systemic bone loss. Since alveolar bone loss is a prominent feature of periodontal disease, disturbances in bone mineral density (BMD), especially in the jaws, are suspected of being an aggravating factor in periodontal disease. In previously published research, the severity of osteoporosis may be related to tooth loss in post-menopausal women. Considering the relationship among bone mineral density, osteoporosis, and periodontitis in men, it is known that men completing androgen ablation therapy for control of prostate cancer are at higher risk for osteoporosis. As androgen deprivation therapy is the recommended treatment for men with metastatic or locally-advanced nonmetastatic prostate carcinoma, and as prostate carcinoma is the most common visceral malignancy and the second leading cause of death from cancer in men, the relationship between androgen deprivation therapy and loss of bone mineral density is a matter of public health importance. This dissertation assesses the association between bone mineral density, the presence of periodontal disease, and the possible subsequent onset of clinical osteoporosis, as seen among a population of older women followed longitudinally; a set of men with prostate carcinoma undergoing androgen ablation therapy; and those men in the same set not receiving androgen ablation for prostate cancer. We believe our research, using the model of periodontal bone density and oral bone loss, shows additional clear empirical evidence pointing to a cause-and-effect relationship between androgen deprivation therapy and loss of bone mineral density.

Cardiovascular disease (CVD) and osteoporosis are common age-related conditions with major public health impact. Mounting evidence suggests a link between the two diseases. The purpose of this research was to investigate the association of BMD measures (areal and volumetric) with prevalent CVD, incident CVD, and subclinical measures of atherosclerosis. We utilized data from two prospective epidemiological studies: a) the Health, Aging, and Body Composition (Health ABC) study that enrolled a population of older men and women (age 68-80 years, 51% women, 42% black), and b) the Study of Womens Health Across the Nation (SWAN) that followed a cohort of women through the menopause transition (age 45-58 years, 61% white, 64% peri-menopausal). In the cross-sectional Health ABC analysis, lower volumetric BMD (vBMD) measures of the spine were associated with higher CVD prevalence in men and women, and areal BMD (aBMD) of the trochanter was related to CVD in women. Additionally, aBMD of the total hip was related to subclinical peripheral arterial disease in men. In the SWAN analysis, we observed an inverse cross-sectional association between trabecular vBMD of the spine and aortic calcification. Meanwhile, no associations with coronary artery calcification were noted after adjusting for age. In the longitudinal Health ABC analysis, lower vBMD measures of the spine were associated with higher CVD incidence in white men, but not in blacks. In women, aBMD of the femoral neck was associated with incident CVD in the full cohort. In race-specific analyses, aBMD measures of the total hip, femoral neck, and trochanter exhibited significant relationships with incident CVD in black women, but not in whites. These relationships were independent of age and shared risk factors between osteoporosis and CVD. The inflammatory cytokines interleukin-6 and tumor necrosis factor-alpha, oxidized LDL, and endogenous estradiol did not explain these associations. Overall, our findings provide epidemiological evidence for the presence of an inverse association between BMD and CVD. An understanding of the common mechanisms underlying bone loss and atherogenesis has significant public health implications as it may set the stage for dual-purpose preventive and therapeutic interventions that target both osteoporosis and CVD.

Cardiovascular disease increases with age and menopause. Atherosclerosis directly contributes to cardiovascular disease and subclinical markers of atherosclerosis are noninvasive methods that help to detect early vascular changes. Thus, risk factors associated with markers of subclinical atherosclerosis may be targeted for interventions in individuals at high risk of developing cardiovascular disease. C-Reactive Protein (CRP), a marker of inflammation, has been found to be associated with cardiovascular events in a large number of populations. However, studies examining the association between CRP and markers of subclinical atherosclerosis have been limited. The cross-sectional association between CRP and central arterial stiffness, assessed by carotid-femoral pulse wave velocity (PWV), was tested in a biracial (Caucasian and Black) cohort of 154 women transitioning through menopause within the Study of Womens Health Across the Nation (mean age 50.8 ± 2.6; 44.2% Black). After adjustment for age, systolic blood pressure, ethnicity, study site, waist circumference, diastolic blood pressure, and physical activity, the mean pulse wave velocity increased with increasing CRP tertiles (758.5 cm/s, 784.5 cm/s, 861.7 cm/s; p for trend = .01). Furthermore, the association was stronger in women later in their transition compared to women earlier in their transition (p for interaction = .02). The cross-sectional association between CRP and systemic arterial stiffness, assessed by brachial artery distensibility, was tested in 1069 women of the same cohort (mean age 53.6 ± 2.6 years). After adjustment for confounders, the mean distensibility decreased with increasing tertiles of CRP (p for trend = .001). This pattern was similar in women of different ethnic groups and stages of the menopausal transition. The association between CRP and three year incident peripheral arterial disease (PAD), assessed by the ankle-brachial index (ABI), was tested in a biracial (Caucasian and Black) cohort of 1918 older adults within the Health, Aging, Body, and Composition Study (mean age 73.6±2.9; 40.3% Black). Participants in the top tertile of CRP had an increased odds of developing PAD compared to those in the bottom tertile (OR=1.87, 95% CI= 1.22 to 2.88). Given the high prevalence of cardiovascular disease, finding risk factors associated with early vascular changes in high risk populations is of public health importance.

Building research capacity in health services has been recognized internationally as an important pillar for the production of a sound evidence base for decision-making in policy and practice. The developing countries are currently facing an increasing epidemic of non-communicable diseases in addition to non-resolving problems of infections, malnutrition and health problems of reproductive health. Clinical research is the link between advances in research and innovations in medical practice. Physicianscientists, trained in patient care and epidemiological research, are crucial in developing and performing cutting-edge clinical research in developing countries. Due to lack of local research capacity, these challenges have not been matched by the ability and capability of developing countries to carry out appropriate studies, the results of which will enable them deal with the health problems in their national contexts. An effort was made to build and strengthen local research capacity in Pakistan and conducted a 9-day workshop on epidemiology research methods to train the trainers. Study objectives: (a) To assess the short and long-term effectiveness in terms of knowledge gain from the epidemiologic research training workshop offered to participants by face-to-face (F2F) and Video-teleconferencing (VTC) methods in Pakistan. (b) to assess the impact of the workshop on students future career goals in both F2F and VTC groups and (C) to assess the cost-effectiveness of VTC relative to F2F instruction of training. Methods: This was a prospective study on 40 F2F and 18 VTC health care professional with post-graduate degrees. A 9- day epidemiological research training workshop was conducted by 5 research faculty from University of Pittsburgh who developed course contents. Pretest and post-test1 were on 1st and last day of the workshop respectively. Post-test 2 was conducted after one year of the workshop. Cost of both teaching methods were obtained using ingredient method and cost effective ratios were calculated Results: The total study sample included 56 and 49 for the short-term and long-term workshop assessment. Within each group, paired sample t-test showed significant improvement in scores after the completion of workshop (P<0.001 for F2F and VTC). In F2F, mean scores increased from 11.13 (pre-test) to 15.08 (post-test1) and in VTC scores increased from 10.67 (pre-test) to 13.22(post-test1). After one year, post-test2 scores remained higher than pretest scores in both the groups (2-sample T-test P=0.11) and were not statistically significant. On 2-way repeated measure ANOVA, both groups showed significant changes in mean scores over time (P<0.001), and no interaction was seen between time and groups (P=0.31). The between subject-effect of groups was found significant (P=0.013). The total incremental cost per score gained was higher for VTC group for both short-term ($166 incremental cost /score gained) and long-term ($458 incremental cost / score gained). Conclusion: The epidemiology research training workshop was found to be effective in terms of knowledge gain in both the groups. This study has public health significance and has presented a model for training doctors and other health care professional in research methods by providing in-house training to reduce increasing problems of brain drain.

Diabetes mellitus is a prevalent, costly condition associated with substantial morbidity and mortality. It is an important public health problem as it is equally burdensome to individuals and to society, and disproportionately affects disadvantaged people and nations. Despite the number of possibilities for reducing much of this burden, the incidence of diabetes continues to grow. New approaches to how diabetes care is delivered are needed to improve care at the patient, provider, community, and health systems levels. It was therefore our objective to explore the effectiveness of implementing a multifaceted diabetes care intervention, based on the Chronic Care Model, into an underserved community, with the goal of changing the way diabetes care is delivered to improve outcomes in patients who receive their diabetes care in the primary care setting. A marked decline in HbA1c was observed in the multifaceted intervention group (-0.6%, p=0.008) but not in the other groups. The magnitude of the association remained strong after adjustment for clustering (p=0.01). The same pattern was observed for a decline in Non-HDL-c and for the proportion of participants who self-monitor blood glucose (SMBG) in the multifaceted intervention group (Non-HDL-c: -10.4 mg/dl, p=0.24; SMBG: +22.2%, p<0.0001) with statistically significant between group differences in improvement (Non-HDL-c: p=0.05; SMBG: p=0.03) after adjustment. The multifaceted intervention group also showed improvement in diabetes knowledge test scores (+6.7%, p=0.07), and empowerment scores (+2, p=0.02). Secondary analyses revealed that psychosocial and psychological factors accounted for a greater amount of the variation in HbA1c, Non-HDLc, and blood pressure values than clinical factors, and are important in contributing to improvement. The improvements observed in HbA1c, systolic and diastolic blood pressure, and the proportion of participants who self monitor their blood glucose at 12-month follow-up were sustained at 36-month follow-up in all study groups. Additional improvements occurred in Non-HDLc levels in all study groups, and quality of well-being scores in the multifaceted intervention group, but not the other groups. These findings are important as they help to close a gap in the literature on improving the quality of care for people with diabetes through redesigning the process of diabetes care delivery.

Diabetes is associated with significant morbidity and mortality. The majority of disease burden is attributed to long-term complications. Screening tests to detect and therapies to treat early forms of diabetes complications are available, but few diabetes patients receive screening at the recommended levels. This report investigated the prevalence and correlates of screening in a cohort of type 1 diabetes patients. The study population was the Pittsburgh Epidemiology of Diabetes Complications study cohort. Screening tests assessed included the HbA1c test, dilated eye exam, foot exam, fasting lipid profile, and urine protein screen. The aims were to: 1) identify the frequency and trends in screening; 2) identify general correlates of screening as well as to evaluate the influence of patient behavior and health care access factors on receipt of screening tests and examine the association between clinical risk of developing complications and receipt of screening tests to detect complications. Reported screening rates varied widely between individual tests, and optimal screening, the use of all tests, was reported by the fewest subjects. Overall, screening in this population is improving over time. The strongest general correlates of screening were specialist care, weekly blood sugar testing, and gender. A more in depth analysis of screening predictors was aimed at determining whether patient or health care access level factors have a stronger influence on screening was conducted. Health care access factors that specifically included specialist care, intensive insulin therapy, and number of physician visits were found to have a stronger influence on screening compared to patient level factors. Finally, this study found that overall, screening does not appear to be associated with clinical risk of developing complications. Based upon this research, areas in need of improvement include optimal screening rates and targeting screening endeavors towards patients at clinical risk for developing complications, and interventions that incorporate access factors may have the strongest impact. The findings of this report have public health significance and have implications for diabetes preventive care. The data from this research can be used to design interventions and policies that improve screening rates, and reduce subsequent morbidity and mortality associated with chronic complications.

Dramatic advances have been demonstrated over the past decade in the prevention and treatment of cardiovascular disease (CVD). Despite these major strides, CVD continues to be our nations most significant cause of morbidity and mortality. The risk status of persons without known CVD varies greatly and thus requires a range of intense screening and interventions. This dissertation focuses on subclinical CVD measures as well as a new methodology that will improve the evaluation of CVD in clinical trials and eventually improve primary prevention of CVD. There are three related projects in this dissertation, each of which uses noninvasive subclinical methodologies to assess cardiovascular risk. The first project focuses on a high-risk population, the elderly, and evaluates the association calcified carotid plaques with cardiovascular outcomes. Carotid plaque characterization is a new focus of research across the nation and what makes one plaque more dangerous than another is unclear. We do know that as plaques age, the plaques often become more complicated and often calcify. However, the significance of calcification in the carotid arteries is poorly understood. In this project, I assess if carotid calcification is predictive of cardiovascular outcomes. The second project focuses on another high cardiovascular risk population, women systemic lupus erythematosus (SLE). Women with SLE have a significantly high risk of myocardial infarction compared to women without SLE. The role that lupus-related risk factors have in cardiovascular disease progression above the traditional risk factors is unclear. Using carotid ultrasound, associations are evaluated between intima-media thickness and plaque with both cardiovascular and SLE-specific risk factors. The third and final project is the development of a protocol that will allow new computerized assessment of carotid artery plaques. Over the past decade both ultrasound technology and computerized assessment tools have improved. This creates opportunity for improved plaque assessment in vivo. This methodology characterized plaque components, possibly identifying plaques that may be dangerous. Plaque characterization software is now available for use with ultrasound and I have developed the protocols to execute this technique in the Ultrasound Research Laboratory. The final project outlines the software and testing process development, staff training, worksheet design, quality control processes, and a pilot study to evaluate the reproducibility of the measure. This research will contribute to public health through new cardiovascular risk assessment techniques and may lead to improved primary prevention and research methods.

Mammographic density is one of the strongest risk factors for breast cancer. Exactly how breast density increases breast cancer risk is unknown, although it is believed that dense breast areas may reflect exposure to estrogen. Breast cancer incidence is highest in postmenopause, when most estrogens are produced in non-ovarian tissues. Cyclooxygenase (COX)-2 and the cytokine tumor necrosis factor (TNF)-alpha may play a role in regulating estrogen synthesis in postmenopausal women. The aim of the present study was to explore the association between inflammation and breast cancer risk in two populations of postmenopausal women. Different exposures associated with inflammation (non-steroidal anti-inflammatory drug (NSAID) use, circulating receptors for TNF-alpha, and a polymorphism in the TNF receptor-II gene) were measured and tested for their association with incident breast cancer or mammographic density. In the first study, the Study of Osteoporotic Fractures (SOF), complete NSAID medication and breast cancer risk factor information was available for 6695 women, mean (SD) age 73 (5) years. During a mean (SD) of 13.2 (3.8) years of follow-up, 372 women were diagnosed with primary breast cancer. There were no differences in incident breast cancer by NSAID use, either before or after adjusting for covariates. In the second study, Mammograms and Masses (MAMS), mean mammographic density was lower among women in the highest quartiles of circulating soluble TNF receptor levels. After adjustment for body mass index, the inverse association disappeared. In evaluating the TNFR2 196 M/R polymorphism (T>G), the unadjusted mean (SD) mammographic density was higher in women with the TT genotype (32.3% (21.0)) as compared to women with the TG/GG genotypes (26.6% (17.2)), p=0.003. The association remained statistically significant after adjustment for age and BMI (p=0.03); however, inclusion of additional covariates reduced the level of statistical significance (p=0.08). There was no observable difference in circulating sTNFR2 levels between the TNFR2 genotypes. An increased understanding of factors that affect mammographic density and their underlying mechanisms is needed, and inflammation may be involved. An association between breast cancer risk and inflammation would have important pubic health implications for screening and primary prevention of breast cancer.

BACKGROUND: Estrogen metabolites, sex-steroid hormones, and breast density are associated with breast carcinogenesis. OBJECTIVE: Complete a systematic study of the contribution of two biological measures (breast density and hormone metabolism) to an endocrine-based model of breast cancer risk. METHODS: The study groups included breast cancer-free participants (N=282) in the Study of Osteoporotic Fractures (SOF), and participants in the Mammogram and Masses Study (MAMS), inclusive of 176 cases (55 pre-menopausal, 121 post-menopausal) and 380 controls (124 pre-menopausal, 256 postmenopausal). Sex-steroid hormones, percent breast density, serum concentrations of 2-hydroxyestrone (2-OH) and 16 alpha-hydroxyestrone (16alfa-OH), and breast cancer risk factors were evaluated to determine associations. RESULTS: In SOF,16alfa-OH was positively associated with body mass index (BMI) (r=0.162); however, this association was not significant in multivariate analyses that controlled for the serum sex-steroid hormone concentrations (total estradiol, total testosterone, SHBG). Women who reported a surgical menopause were significantly more likely to have higher levels of 16alfa-OH (OR=(tertile 3 vs tertile 1) 7.37, 95% Confidence Interval (CI) 2.20-24.70), but there was no type of menopause difference with respect to 2-OH tertile. In all MAMS control subjects (N=380), breast density correlated weakly with log-transformed serum concentrations of 16alfa-OH (Pearson correlation coefficient = 0.10, p-value < 0.1). Stratification according to menopausal status substantially reduced or eliminated associations between breast density and the estrogen metabolite concentrations. Logistic regression analyses showed a 3-4 fold increased risk of breast cancer among pre-menopausal women in the highest tertile of breast density compared with those in the lowest tertile of density, even with adjustment for the estrogen metabolites. A statistically non-significant 1.5-fold increased risk of breast cancer in high vs. low tertile of density was observed among post-menopausal women taking hormone therapy (HT) after adjusting for estrogen metabolites, BMI, and age. Breast density did not appear to substantially increase breast cancer risk among post-menopausal women not taking HT. CONCLUSION: In SOF, results did not show consistent associations between risk factors and estrogen metabolites except for a positive association between BMI and 16alfa-OH and surgical menopause and 16alfa-OH. With respect to MAMS, menopausal status may influence substrate estrogen hormone levels primarily, and, estrogen hormone levels may influence breast density secondarily, through pathways not involving the estrogen metabolites. The breast density-breast cancer association remains significant even with adjustment for the estrogen metabolites, at least in pre-menopausal women, suggesting that breast density may relate to breast cancer risk through pathways not involving estrogen metabolism. PUBLIC HEALTH SIGNIFICANCE: Understanding factors that affect breast density and their underlying mechanism is an important public health issue. Such an understanding will help us improve breast cancer screening and may help us identify women who are at an increased risk of breast cancer and for whom prevention strategies may be useful.

Lung cancer remains a significant public health problem in 2006, despite efforts aimed at educating individuals about the dangers of tobacco and the successes of smoking cessation programs. No screening methods aimed at high risk individuals are currently supported by the National Cancer Institute or other medical organizations. Since the risk for lung cancer persists even after smoking cessation, studies of methods to detect lung cancer in its early stages, when it is amenable to cure, are clearly needed. This study examined a novel method of sputum collection and processing to determine whether the adequacy of the samples collected was improved over conventional preparation methods. We also examined lung function as a possible predictor of cytologic abnormality in the sputum of individuals at high risk for lung cancer, and studied the potential of a molecular biomarker for the early detection of lung cancer. This study employed a cross-sectional design and utilized quantitative methods for exploring the relationships between the variables, particularly those of lung function, cytologic diagnosis and gene mutation status, with personal risk factors for developing lung cancer. This study demonstrated an association between lung function and cytologic diagnosis of moderate or worse atypia in the sputum collected from these participants and examined at the University of Colorado. We also demonstrated a higher rate of sputum specimen adequacy than has been previously reported from conventional clinical experience. We examined the feasibility of conducting somatic gene mutation analysis on the samples collected by this novel method. We achieved some success in studying K-ras mutations, but were unsuccessful in our analysis of DNA methylation. From these experiences we have gathered information upon which to base further studies. The importance of these findings from a public health perspective is that there is an opportunity for early detection of lung cancer via analysis of cellular and molecular changes in sputum. We have demonstrated that the Thin Prep® methodology, applied to sputum, produces material suitable for cytologic examination and, under certain circumstances, material suitable for molecular analysis. Patients at high risk for lung cancer will benefit from continued research into novel screening methods.