CAC Volume Positively Associated with CHD and CVD Risk

A recent study sought to determine the independent associations of coronary artery calcium (CAC) volume and CAC density with incident cardiovascular disease (CVD), including coronary heart disease (CHD) [JAMA. 2014;311(3):271-278]. CAC density was inversely associated with CHD and CVD risk, indicating that the role of CAC density should be considered when evaluating current CAC scoring systems. 

Using data from the MESA (Multi-Ethnic Study of Atherosclerosis) trial, the amount of CAC from computed tomography (CT) was quantified from the individual Agatston scores and the volume scores.

The study was limited to the 3398 MESA participants (1964 men [58%] and 1434 women [42%]) with a CAC score of >0 at baseline. The median follow-up was 7.6 years with an average of 7 years; 4 individuals were lost to follow-up, leaving 3394 participants for analysis. In the first quartile of volume scores, the density score was lower; in the upper 3 quartiles the density score was larger and similar. There were a total of 265 CVD events during follow-up (175 CHD events and 90 other CVD events) with rates increasing monotonically across the 4 quartiles. 

When analyzing the data by quartile of density score, Agatston, area, and volume scores increased across quartile 1, 2, and 3 but were slightly lower in the fourth quartile compared with the third. Spearman correlation coefficients for the various CAC scores were used to evaluate the degree of association between the Agatston, volume, area, and density scores. The Agatston, volume, and area scores were all highly correlated (coefficient for bivariate analysis, r>0.99), while the density score was only moderately correlated with the other 3 scores (correlations of 0.623, 0.560, and 0.539, respectively). The density upweighing the Agatston score is indicated in the slightly higher correlation of 0.623 between the 2 scores.

Based on previous analyses in MESA, the natural log (ln) of Agatston, volume, and area scores were used in performing Cox proportional hazard models to estimate the hazard ratios (HRs), reflecting 1 standard deviation of the independent variable. The standard deviations for lnAgatston, lnvolume, lnarea, and density were 1.8, 1.6, 1.6, and 0.7, respectively. In a single age, sex, and ethnicity adjusted model of the data, the Agatston, volume, and area scores all showed similar and significant associations for both P values and 95% confidence interval (CI). The similarity in associations reflects the high correlation (r>0.99) of these 3 scores. Additionally, the associations were modestly stronger for CHD than for CVD, while the density score showed no association for CHD or for CVD.  

In a fully adjusted Cox model containing volume and density scores, the density score caused the HRs, reflecting a 1 standard deviation difference, for the volume score to increase for both end points from 1.52 (95% CI, 1.29-1.80) to 1.81 (95% CI, 1.47-2.23), absolute risk increase 6.1/1000 person-years for CHD, and from 1.38 (95% CI, 1.21-1.58) to 1.68 (95% CI, 1.42-1.98), absolute risk increase 7.9/1000 person-years for CVD. The density score became significantly and inversely associated with CHD, generating an HR value of 0.73 (95% CI, 0.58-0.91), absolute risk decrease 5.5/1000 person-years, and with CVD, generating an HR value of 0.71 (95% CI, 0.60-0.85), absolute risk decrease 8.2/1000 person-years.

Overall, when accounting for the density score in the model the strength of the volume score increased and, at any given CAC volume score, a 1 standard deviation higher CAC density score was associated with a significantly lower risk for CHD and CVD.

Due to the independent and inverse association between CAC density, at any given CAC volume, and risk, the Agatston or volume scores alone are not optimal measurements to use in CVD risk prediction, according to the researchers. An increase in CAC may be a result of an increase in volume, an increase in density, or both.