Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Epidemiology and Biostatistics

Major Professor

Janice Zgibor, Ph.D.

Co-Major Professor

Kevin Kip, Ph.D.

Committee Member

Amy Alman, Ph.D.

Committee Member

Wei Wang, Ph.D.

Committee Member

Zachary Pruitt, Ph.D.


Coronary Heart Disease, Controlled Direct Effect, MACE, Health disparities, Metabolomics


Background: Despite declining mortality in cardiovascular diseases (CVD), racial disparities between non-Hispanic Blacks (NHB) and to non-Hispanic Whites (NHW) persist. Although the prevalence of traditional risk factors of CVD such as hypertension, is higher in NHB compared to NHW, adjusting for this difference does not eliminate the disparity completely. This suggests other factors might explain the persisting disparities. Thus, the purpose of this dissertation is to quantify the impact of chronic stress in explaining the racial disparities in cardiovascular diseases (CVD). This dissertation contains three studies that addressed the following Specific Aims: Specific aims: 1) To create and assess the reliability of various measures (definitions) of chronic stress and examine their incremental predictive benefit with incident cardiovascular disease. 2) To quantify the effect of chronic stress in explaining the racial disparities of incident CVD. 3) To identify chronic stress related metabolites associated with incident CVD Methods: This dissertation leveraged data from the Heart Strategies Concentrating on Risk Evaluation (Heart SCORE) study. Heart SCORE is an ongoing, prospective cohort and community-based study of 2,000 community dwelling males and females, aged 45-75 years in Pittsburgh, Pennsylvania. Chronic stress was defined by counting the number of times an individual was exposed to various stressors such as perceived discrimination, financial difficulties, caregiving, job difficulties, and residing in a neighborhood with high depravity (CRCS). It was also defined using perceived stress using Cohen’s Perceived stress scale (PSS-4) and allostatic load (AL). Coronary heart disease- a type of cardiovascular disease- was as a composite outcome defined as the first occurrence of cardiac death, myocardial infarction or coronary revascularization. In Aim 1, reliability analyses using Cronbach Alpha, Weighted Kappa and Spearman correlations were used to assess the reliability of the derived measures of stress. Cox proportional hazard regression models were subsequently used with each measure of chronic stress to determine the incremental benefit in risk prediction with cardiovascular risk scores such as the Framingham risk score (FRS) and pooled cohort equation (PCE) risk score, when applicable. The increase in C-statistic, likelihood ratio test was used to determine predictive benefit and the net reclassification among events and non-events were used to provide a summary measure to quantify this effect. These analyses were repeated among various demographic subgroups. In Aim 2, marginal structural weighted Cox proportional hazards regression models were used to calculate the controlled direct effect of Race on CVD and the percentage of the disparity that would be eliminated. Finally, in Aim 3, multiple logistic and linear regression models were used to identify stress-related metabolites while controlling for multiplicity error using false discovery rate. Subsequently, ordinal and Cox proportional hazard regression models were used to identify stress related metabolites associated with ideal cardiovascular health and CVD. Results: Among the 1,825 individuals who met the eligibility criteria in Aim 1, 17.3%, 20.1%, 31.4% of the population were classified as having high chronic stress according to CRCS, allostatic load and perceived stress, respectively. Allostatic load had the weakest agreement with the other measures of chronic stress (AL vs PSS, κ =0.02; AL vs CRCS, κ =0.11) while CRCS and PSS had a slight agreement (κ =0.2). All measures of chronic stress did not improve CVD risk prediction in the overall population, however, CRCS improved the risk prediction among low-income Blacks (p=0.08). The net reclassification was 0.455 and -0.237 among low-income Blacks with and without CVD, respectively. In Aim 2, the cumulative incidence of CVD was 5% among the 1,735 individuals who met the eligibility criteria. This resulted in an incidence rate of 5.07/1,000 individuals/year among non-Hispanic Blacks and 4.79/1,000 individuals/year among non-Hispanic Whites (incidence rate ratio: 1.06 (0.54, 2.09)). However, this was much higher among individuals aged 45 – 55 (4.29 (1.22, 15.06)). Among 1,443 individuals with complete data on all relevant study variables, the controlled direct effects using CRCS as the mediator were 1.45 (0.70 , 3.01) and 1.39 (0.64, 3) before and after adjusting for traditional risk factors of CVD. This equated in a 43% and 12.6% elimination of the racial disparity. The effect of CRCS was largely driven by perceived discrimination which completely eliminated the disparity after adjusting for traditional risk factors. The effect of CRCS on the racial disparity was attenuated when missing data were imputed. Finally, results from Aim 3 identified 36 metabolites associated with chronic stress. Of these, 14 were associated with ideal cardiovascular health (ICH) and one was associated with incident CVD. However, this association was driven by its association with CVD among non-Hispanic Whites. Conclusion: In conclusion, chronic stress, defined as a count of exposure to multiple stressors provided incremental predictive benefit beyond the Framingham risk score (FRS) and pooled cohort equation (PCE) among low-income Blacks. Chronic stress also plays a modest role as a mediator in the racial disparities in CVD. Finally, the results from this dissertation suggests other biomarkers of chronic stress might exist and these biomarkers could elucidate the physiological relationship between chronic stress and CVD. All three conclusions need to be validated in a larger, biracial cohort. Future research should examine the drivers of the disparities and the impact of stress-related epigenetics in young NHB.

Included in

Epidemiology Commons