Abstract 27: Volume and Density of Calcium in the Ascending Thoracic Aorta, When Present, Predict Incident Coronary Heart Disease Beyond Coronary Artery Calcium: the Multi-Ethnic Study of Atherosclerosis

Abstract

Introduction: Coronary artery calcium (CAC) volume and density differentially predict coronary heart disease (CHD) beyond traditional CVD risk factors. Whether similar associations exist for ascending thoracic aortic calcium (ATAC) volume and density are unknown. We hypothesized that ATAC volume and density predict CHD independent of the pooled cohort equations (PCE) and CAC. Methods: The Multi-Ethnic Study of Atherosclerosis followed 6,814 participants free of clinical CVD at recruitment. Cardiac CT was performed at baseline. For this analysis, only those with prevalent ATAC were included (necessary for evaluation of density). Cox regression was used to estimate associations of ATAC volume and density with incident CHD (fatal/non-fatal MI and cardiac arrest), with adjustment for the PCE and CAC volume and density. The incremental predictive value of ATAC volume and density was evaluated by area under receiver operating characteristic curves (AUC). Results: Of the total cohort, 233 (3.4%) had prevalent ATAC. In these participants, 29 CHD events occurred over 10.3 years. After simultaneous adjustment for the PCE, ATAC and CAC volumes and densities, ATAC alone was associated with CHD, with one standard deviation (SD) higher natural log ATAC volume associated with a 76% greater risk [HR 1.76 (95% CI 1.07-2.87)], and one SD lower ATAC density associated with a 57% lower risk [0.43 (0.26-0.71)]. The PCE, CAC volume and CAC density were not significantly associated with CHD. A model containing the PCE, CAC volume and density had an AUC of 0.584 (0.467-0.701) for incident CHD prediction. Adding ATAC volume and density to the model improved the AUC to 0.703 (0.588-0.817) (see table). Conclusions: ATAC was uncommon in a cohort free of baseline clinical CVD. However, when ATAC was present, both ATAC volume and density were independently associated with incident CHD. Furthermore, in these participants, the PCE and CAC poorly discriminated incident CHD, and ATAC volume and density substantially improved risk prediction.