Introduction
Coronary artery calcium (CAC), a measure of atherosclerosis predicts mortality in both diabetes and the general population. The incremental utility of measuring atherosclerosis across multiple vascular beds beyond CAC for the prediction of mortality has not been reported. In this study we derived a composite atherosclerosis score (CAS) from multiple vascular beds and compared it with CAC in predicting all-cause mortality. We hypothesized that this composite score will be significantly better than CAC in predicting mortality.
Methods
A total of 803 participants, ages 39–86, with complete data on diabetes and vascular imaging in the Diabetes Heart Study (DHS) were followed for an average of 7.4 years. Computed tomography (CT) scans were performed at baseline to obtain measures of carotid (CAAC), coronary (CAC) and abdominal aorta (AAC) Agatston scores. A principal component analysis using studentized residuals of log transformed (CAAC+1), (CAC+1), and (AAC+1) adjusting for age, race, and gender was performed. We selected the first principal component as the CAS. Seven-year risk estimates for mortality were obtained using logistic regression models. Model 1 included age, gender, smoking, systolic blood pressure, antihypertensive medication use, total and high-density lipoprotein cholesterol, and ethnicity and CAS. Model 2 included these risk factors plus CAC. We compared the estimation of mortality in Model 1 with CAS vs. Model 2 with CAC using chi-square values.
Results
Overall, 14% (116/803) of participants died during follow-up. CAS explained 70% of the variance, (eigenvalue of 2.1 with loading, CAAC, 0.57; CAC, 0.57; and AAC, 0.59). After adjusting for potential confounders, the odds ratio (95% CI) of all-cause mortality for 1-standard deviation (SD) increment was 2.12 (1.64–2.78) for CAS and 2.38 (1.77–3.35) for CAC. The area under the curve (chi-square value) with CAS or CAC to predict mortality was 0.76 (36) vs 0.76 (37) respectively.
Conclusion
Subclinical atherosclerosis, as measured by CT determined calcified plaque burden has increasing evidence supporting its role as a tool to stratify future risk for mortality. Here we demonstrated that the diagnostic accuracy between CAS and CAC are comparable and the predictive value of CAC alone for mortality is not further enhanced by inclusion of calcified plaque burden in carotid or abdominal aortic territories.
Ethnic Differences in the Relationship between Pericardial Adipose Tissue and Coronary Artery Calcified Plaque: African-American-Diabetes Heart Study
