REGARDS: Validation of Pooled Cohorts 10-Year Atherosclerotic CVD Risk Equations
Observed and predicted five-year atherosclerotic cardiovascular disease (ASCVD) risks were similar among the population studied in the Reasons for the Geographic And Racial Differences in Stroke (REGARDS) study, indicating that risk equations were well calibrated in the population for which they were designed to be used, and that they demonstrated moderate to good discrimination, according to a study presented March 29 as part of ACC.14 in Washington, DC, and simultaneously published in the Journal of the American Medical Association (JAMA).
The study assessed outcomes in the cohort of U.S. adults for whom statin initiation is considered, based on the 2013 ACC/American Heart Association Pooled Cohort risk equations that estimate 10-year cardiovascular disease risk. Today’s study assessed the calibration and discrimination in 10,997 adults aged 45 to 79 between January 2003 and October 2007, following up through December 2010.
The observed and predicted five-year ASCVD incidence per 1,000 person-years for participants with a 10-year predicted risk of less than 5 percent was 1.9 (95 percent CI, 1.3-2.7) and 1.9, respectively; risk of 5 percent to less than 7.5 percent was 4.8 (95 percent CI, 3.4-6.7) and 4.8; risk of 7.5 percent to less than 10 percent was 6.1 (95 percent CI, 4.4-8.6) and 6.9; and risk of 10 percent or greater was 12.0 (95 percent CI 10.6-13.6) and 15.1. Researchers used a modified Hosmer-Lemeshow test, where a chi-square >20 or p-value <0.05 indicates poor calibration, and analyzed discrimination using the C-index of 0.72 (95 percent CI, 0.70-0.75).
There were 234 ASCVD events (120 coronary heart disease events, 114 strokes) among Medicare-linked participants and the observed and predicted five-year ASCVD incidence per 1,000 person years for participants with a predicted risk of less than 7.5 percent was 5.3 (95 percent CI, 2.8-10.1) and 4.0, respectively; risk of 7.5 percent to less than 10 percent was 7.9 (95 percent CI, 4.6-13.5) and 6.4; and risk of greater than 10 percent was 17.4 (95 percent CI, 15.3-19.8) and 16.4. The C-index was 0.67 (95 percent CI, 0.64-0.71).
"Differences in the observed and predicted ASCVD risk were small when limited to participants without diabetes, with a LDL-C level between 70 and 189 mg/dL, and who were not already taking statins. Calibration in this group is particularly important because it represents the population for whom high predicted risk is intended to trigger a discussion about statin initiation," notes Paul Muntner, PhD, of the University of Alabama at Birmingham.
Keywords: Stroke, Atherosclerosis, Calibration, Coronary Disease, Medicare, Diabetes Mellitus
< Back to Listings