Predictive Utility of Coronary Artery Calcium Score Added to the PREVENT Atherosclerotic Cardiovascular Disease Equations

Importance: The 2026 American College of Cardiology/American Heart Association/Multisociety Dyslipidemia Guideline newly recommends selective use of coronary artery calcium (CAC) scoring based on a 10-year risk of 3% to less than 10% estimated with the Predicting Risk of Cardiovascular Disease EVENTS (PREVENT) atherosclerotic cardiovascular disease (ASCVD) equations. However, the utility of CAC scoring alongside the PREVENT equations for ASCVD risk estimation is not known. Objective: To determine the change in predictive utility when adding CAC scores to the PREVENT-ASCVD equations. Design, Setting, and Participants: Longitudinal observational cohort study conducted at 6 sites in the United States and enrolling adult participants aged 45 to 79 years without ASCVD at baseline in the Multi-Ethnic Study of Atherosclerosis. Exposures: Ten-year ASCVD risk with the PREVENT-ASCVD equations before and after adding CAC scores. Main Outcomes: Performance metrics for estimation of risk of fatal and nonfatal ASCVD, assessed using the Harrell C statistic, calibration, and net reclassification improvement (NRI) for newly defined ASCVD risk categories: low (<3%); borderline (3% to <5%); intermediate (5% to <10%); and high (≥10%). Results: A total of 6098 participants were included. At baseline, participant mean age was 61.4 (SD, 9.6) years, 52% were female, mean estimated 10-year ASCVD risk was 6.4% (SD, 4.8%) with the PREVENT-ASCVD base equations, and 49% had a CAC score greater than 0. During 10 years of follow-up, 366 (6%) experienced an ASCVD event. The Harrell C statistic for the PREVENT-ASCVD base equations was 0.73 (95% CI, 0.70-0.75). After adding CAC to the PREVENT-ASCVD base equations, the change in Harrell C statistic was 0.02 (95% CI, 0.01-0.03), and the categorical NRI was 0.095 (95% CI, 0.053-0.137). The calibration slope for the PREVENT-ASCVD base equations was 1.09 (95% CI, 0.93-1.25), and for the PREVENT-ASCVD base plus CAC equations was 0.92 (95% CI, 0.81-1.03). Among those at borderline risk, incident ASCVD occurred in 1.9% of those with CAC score of 0, 3.9% with CAC greater than 0 and less than 100, 7.4% with CAC 100 or greater and less than 300, and 14.3% with CAC 300 or greater. Conclusions and Relevance: Among US adults aged 45 to 79 years, addition of CAC score to the PREVENT-ASCVD equations across all risk groups overall only modestly improved and, in some groups, did not change model discrimination and reclassification. Reclassification results support selective use of CAC in those with borderline to intermediate risk estimated by the PREVENT-ASCVD equations.