Incidence of New Coronary Calcification: Time to Conversion From CAC = 0

Although high coronary artery calcium (CAC) scores predict high risk, absence of CAC has been shown to be a powerful negative risk marker for the development of coronary heart disease (CHD) (1). Accordingly, the 2018 American College of Cardiology/American Heart Association/Multisociety guidelines assigned a Class IIa recommendation for CAC testing in select borderline- to intermediate-risk patients, acknowledging that intensive statin therapy is of limited value among patients with CAC = 0 and may be safely avoided over the next 5 to 10 years.

Given the prognostic power of CAC = 0 and the clinical significance of CAC development, it is important to know when an initial CAC = 0 scan should be updated to reassess risk in these patients. We sought to characterize the time period to incident CAC (“CAC conversion”) of a CAC = 0 scan, thus defining the time frame in which the probability of developing CAC >0 is reasonably low and in which repeating a CAC scan would likely result in a low testing yield.

We studied 3,116 participants from the MESA (Multi-Ethnic Study of Atherosclerosis) study with baseline CAC = 0 and at least 1 follow-up scan (≤10 years after baseline). The cumulative prevalence of CAC >0 was calculated. The time period to CAC conversion was modeled using a Weibull parametric survival model as the time needed to progress to detectable CAC, as a function of baseline risk, sex, and desired yield of testing. We then assessed what percentage of the participants with baseline CAC = 0 developed coronary events during follow-up. Theoretically, these individuals could have been identified before the occurrence of the event by detecting conversion to CAC >0 with repeat testing at different intervals, and the event(s) could have been prevented with CAC-guided timely preventive therapy.

Mean age of study participants was 58 ± 9 years (63% women, mean 10-year atherosclerotic cardiovascular disease risk 14 ± 13%). The prevalence of CAC >0 increased with time, from 11% at 2 years to approximately 50% at 10 years. Using a testing yield of 25% (number needed to scan = 4 to detect CAC >0), the estimated time period to CAC conversion of low, intermediate, and high estimated risk men was 7, 4, and 3 years, respectively, and for women was 8, 5, and 3 years, respectively. In the overall population, a rescanning interval of 3 to 5 years resulted in missing 7% to 14% of all participants developing future 10-year events, with a single incident CAC >0 scan potentially picking up to 44% of all participants developing future 10-year events.

Our findings suggest that the average time period to CAC conversion in individuals with CAC = 0 and mean age of 58 years lies between 3 and 7 years and is dependent on age, sex, baseline risk profile, and desired testing yield, and retesting at this interval (and subsequent tailored preventive management) appears to miss very few events over a 10-year time horizon (Figure 1). A few prior studies have investigated the incidence of CAC >0 after initial CAC = 0. Min et al. (2) conducted a prospective 5-year study in 422 participants (mean 49 ± 9 years of age) with a baseline CAC = 0 scan. The authors concluded that rescan should happen at 4 years (2). However, the relatively small study population lacked ethnic diversity, and patients were referred, potentially introducing bias. Gopal et al. (3) conducted a retro-spective study among 710 physician-referred individuals (mean 56 ± 9 years of age) with initial CAC = 0 and a follow-up scan at least 12 months apart. The authors concluded that a repeat scan should be conducted no sooner than 5 years (3). This study was limited because only about one-third of the subjects had a >5-year rescan. Importantly, none of the studies mentioned previously considered age, sex, and baseline risk for their recommendations, and none formally considered desired yield of testing in their calculations.

FIGURE 1. Time to Conversion From CAC = 0 in the MESA Study.

Time period to incident coronary artery calcium (CAC) conversion (in years) from CAC = 0 as a function of estimated 10-year atherosclerotic cardiovascular disease risk (A), age (B), sex and desired yield of testing (number needed to scan [NNS]). MESA = Multi-Ethnic Study of Atherosclerosis.

Our study has some limitations. In the MESA study, not all participants were scanned at every visit. This precluded a calculation of a true annual CAC incidence, and we had to rely on cumulative yearly incidence as a measure of incident CAC over time. Although the MESA study is relatively large, the total number of CHD events was still low, not unexpectedly limiting power, as low CHD event rates are to be expected in a healthy study population with baseline CAC = 0.

In summary, we show that, overall, a 3- to 5-year time frame seems reasonable for repeat scanning in CAC = 0 individuals of age similar to those included in the MESA study. However, we have also demonstrated that this varies considerably depending on sex, age, baseline atherosclerotic cardiovascular disease risk, and the preferred yield of retesting.