PREDICTION OF CORONARY HEART DISEASE WITHIN THE AEROBICS CENTER LONGITUDINAL STUDY POPULATION

This dissertation is a compilation of three studies that were conducted to better 1) Further validate a thoroughly tested Framingham Risk Score (FRS) on a unique cohort with comprehensive measures available, 2) Update and improve the predictability of the FRS through the addition of cardiorespiratory fitness (CRF) while resolving limitations in previous studies, and 3) Assess the predictability of non-exercise estimated CRF (e-CRF) and FRS on CHD. A manuscript was generated for each study utilizing data from the Aerobics Center Longitudinal Study. To validate the FRS, a multivariable Cox Proportional Hazard Model was used to determine the association between FRS component and CHD. The Area Under the Curve (c-statistic) from the receiver operating characteristic (ROC) curve was used to determine predictability of the FRS model on ACLS. The FRS’ components were significantly associated with CHD and the c-statistic was statistically significant. The second study’s goal was to update the FRS by adding CRF. This study included 29,854 men from ACLS that completed a baseline examination from 1979-2002. FRS was defined as a composite score and modeled as a continuous and categorical variable. CRF was defined as a continuous variable through maximally achieved metabolic equivalent of task (METs) and categorical: low, moderate, or high CRF. Multivariable survival analysis showed a significant association between CRF, FRS and CHD. Although the second study found there was a significant relationship with CRF, FRS, and CHD, traditionally measured CRF is not a clinically viable tool. The third study aim was to use a non-exercise estimated CRF (e-CRF) to determine the relationship between e-CRF, FRS, and CHD. Estimated CRF was defined through a 5-item questionnaire and the same data from study #2 was utilized for the multivariable Cox Proportional Hazard modeling. The relationship between e-CRF and CHD was investigated in subset populations based on age, smoking, hypertension, and diabetes diagnosis. Our study found that among men with ‘moderate or high’ risk for CHD, men with moderate or high fitness had a decreased risk for CHD compared to men with low fitness. CHD is one of the leading causes of death in the U.S. and early establishment of CHD risk is important for primary and secondary prevention. The series of papers presented in this dissertation provide the evidence needed to begin establishing a more comprehensive and clinically feasible risk prediction tool. Clinicians may want to consider capturing their patients’ medical history, CHD risk factors, and their e-CRF so they can take advantage of CRF’s improved prediction of CHD. This comprehensive approach can help physicians predict adverse events for their patients while also counseling them on how to improve their overall health through improvement of CRF