Coronary Calcium Coverage Score: Determination, Correlates, and Predictive Accuracy in the Multi-Ethnic Study of Atherosclerosis

PURPOSE: To develop a new calcium score for use with unenhanced cardiac computed tomography (CT) that can be used to define the percentage of coronary arteries affected by calcium and to correlate this score with risk factors and cardiovascular events. MATERIALS AND METHODS: Institutional review boards at all participating centers approved this HIPAA-compliant study, and all participants gave written informed consent. Calcium coverage score (CCS), which represents the percentage of coronary arteries affected by calcific plaque, was calculated for 3252 participants in the Multi-Ethnic Study of Atherosclerosis in whom calcific plaque was detected with CT. Quasi-Poisson models were used to estimate associations (assessed by using t tests with robust standard errors) between CCS and risk factors. Associations between the CCS, Agatston, and calcium mass scores (hereafter, mass scores) and outcomes were estimated and assessed by using Cox proportional hazards models with Wald tests. The predictive ability of these models was assessed by using area under the receiver operating characteristic curves and bootstrap t tests. RESULTS: After adjustments were made for age, race, ethnicity, and sex in the quasi-Poisson model, CCS was associated with hypertension, dyslipidemia, and diabetes (P < .001 for all diseases). After adjustments for age and sex, a twofold increase in CCS was associated with a 52% (95% confidence interval: 34%, 72%) increase in risk for any coronary heart disease (CHD) event. When Agatston or mass scores were included with CCS in a Cox model for prediction of CHD events, neither Agatston score nor mass score was a significant predictor, whereas CCS remained significantly associated with CHD events. Although receiver operating characteristic curves suggested that there was a difference between CCS score and Agatston and mass scores in prediction of a cardiac event, no differences in prediction of hard cardiac events (myocardial infarction, death) were found. CONCLUSION: Both spatial distribution and amount of calcified plaque contribute to risk for CHD.