Rationale and design of the coronary artery calcium consortium: A multicenter cohort study

Background

Although coronary artery calcium (CAC) has been investigated for over two decades, there is very limited data on the association of CAC with cause of death. The CAC Consortium is a large ongoing multi-center observational cohort of individuals who underwent non-contrast cardiac-gated CAC testing and systematic, prospective, long-term follow-up for mortality with ascertainment of cause of death.

Methods

Four participating institutions from three states within the US (California, Minnesota, and Ohio) have contributed individual-level patient data to the CAC Consortium (spanning years 1991–2010). All CAC scans were clinically indicated and physician-referred in patients without a known history of coronary heart disease. Using strict inclusion and exclusion criteria to minimize missing data and to eliminate non-dedicated CAC scans (i.e. concomitant CT angiography), a sharply defined and well-characterized cohort of 66,636 patients was assembled. Mortality status was ascertained using the Social Security Administration Death Master File and a validated algorithm. In addition, death certificates were obtained from the National Death Index and categorized using ICD (International Classification of Diseases) codes into common causes of death.

Results

Mean patient age was 54 ± 11 years and the majority were male (67%). Prevalence of CVD risk factors was similar across sites and 55% had a <5% estimated 10-year atherosclerotic cardiovascular disease (ASCVD) risk. Approximately 45% had a Calcium score of 0 and 11% had an Agatston Score ≥400. Over a mean follow-up of 12 ± 4 years, there were 3158 deaths (4.15 per 1000 person-years). The majority of deaths were due to cancer (37%) and CVD (32%). Most CVD deaths were due to CHD (54%) followed by stroke (17%). In general, CAC score distributions were similar across sites, and there were similar cause of death patterns.

Conclusions

The CAC Consortium is large and highly generalizable data set that is uniquely positioned to expand the understanding of CAC as a predictor of mortality risk across the spectrum of disease states, allowing innovative modeling of the competing risks of cardiovascular and non-cardiovascular death.