Abstract 9891: Are CAC Scores Less Valuable for Patients of Low Socioeconomic Status? A Dallas Heart Study Analysis

Abstract

Background: Low socioeconomic status (SES) is associated with atherosclerotic cardiovascular disease (ASCVD) and possible underestimation of risk by the pooled cohort equations (PCE). Whether coronary artery calcium (CAC) scores can improve risk discrimination in those with low SES is unknown. Methods: Individuals from the Dallas Heart Study free of ASCVD and with CAC scanning were evaluated. Low SES was defined as an annual income <$16,000 or an educational attainment level ≤11 years. Unadjusted and adjusted analyses assessing the relationships between SES and 1) CAC scores and 2) ASCVD events (fatal or non-fatal MI or stroke) were performed. The incremental predictive value of CAC in those at low and higher SES was assessed using c-statistics (PCE alone vs PCE + CAC). Results: Among 2246 individuals (mean age 44 yrs, 55% women, 45% Black), 115 ASCVD events occurred over a median of 12.5 years. Low SES by income (n=451) was associated with increased risk of ASCVD after adjustment for risk factors (HR 2.16, 95% CI 1.37-3.42), but was not independently associated with CAC. Higher CAC scores (0, 1-99, ≥100) tracked with increased 10-year ASCVD event rates in those at low (5, 10, and 22%) and higher SES (0.002, 4, and 11%), with low SES groups carrying greater ASCVD risk for every stratum of CAC, including with CAC=0. In multivariable models compared with CAC=0, those with CAC 1-99 and ≥100 had a greater risk of ASCVD events for individuals at higher SES [HR 3.40 (1.5-7.8), and 4.8 (1.9-12.6)], but not for those with low SES [HR 1.00 (0.4-2.4), and 1.6 (0.6-4.6)], (p-interaction CAC x SES=0.002). Similar results were found for low SES defined by education. The addition of CAC to PCE improved risk discrimination for individuals at higher SES, but not for individuals at low SES (Table 1). Conclusion: Individuals at low SES have significantly increased risk of ASCVD, however, CAC scores are not independently associated with ASCVD in this population and may be less useful in risk discrimination.