Abstract 10435: Coronary Artery Calcium Scores and Coronary Risk - Coronary Artery Calcium Thresholds Are Not Consistent Across Baseline Risk Levels: The Multi-Ethnic Study of Atherosclerosis
Introduction: By convention, coronary artery calcium (CAC) scores have been stratified, with 0 low, 1–100 average, 101–399 moderate, and >400 high risk. How these strata correspond to estimated coronary risk when continuous CAC scores are incorporated into prediction models has not been studied. Aim: To assess whether CAC cut-points for risk classification can be defined when CAC score is included in a prediction model with traditional risk factors.
Methods: The Multi-Ethnic Study of Atherosclerosis (MESA) is a population-based cohort of asymptomatic adults ages 45–84 without known CVD. Agatston CAC score was assessed by CT. Coronary heart disease (CHD) was defined as MI, death from CHD, resuscitated cardiac arrest, definite angina +/− revascularization, or probable angina with revascularization. We excluded persons with diabetes because it is a CHD risk-equivalent. Risk estimates, using a Cox proportional hazards model, were categorized as 0–2.5% (low), >2.5%−5.0% (low-intermediate), >5.0%−10.0% (intermediate), and >10% (high) for 5-yr incident CHD. Model 1 used age, gender, tobacco use, SBP, antihypertensive drug use, total and HDL cholesterol, and race/ethnicity. Model 2 used these variables plus continuous CAC [expressed as ln(CAC+1)]. Cross-tabulations of risk categories based on the model +/− CAC were performed.
Results: Inclusion of continuous CAC in the model was associated with a net reclassification improvement (NRI 0.31, P<0.001). However, CAC scores required for reclassification depended greatly on baseline risk in the model without CAC (see Table). Substantially higher CAC was needed to classify persons as high-risk when the baseline risk was low (median CAC 1788) compared to intermediate (median CAC 673). Very low scores (median CAC 0) were needed to reclassify persons from high to low risk.
Conclusions: These findings suggest that CAC results should be interpreted in the context of clinical risk factors, rather than predetermined strata.
- © 2010 by American Heart Association, Inc.