This article is more than 5 years old. Information may no longer be current.
Existing CVD risk equations appropriate for black adults
Click Here to Manage Email Alerts
The American College of Cardiology/American Heart Association pooled-cohort risk equation and the Framingham risk score work well in black adults, and a separate risk calculator for them may not be needed, according to an analysis in JAMA Cardiology.
The most frequently used risk prediction models were primarily derived from white populations, so researchers assessed whether two of the most common ones, the ACC/AHA pooled-cohort risk equation and the Framingham risk score, could be improved upon in black adults.
Ervin R. Fox, MD, MPH, from the division of cardiovascular diseases, department of medicine, University of Mississippi Medical Center, Jackson, and colleagues constructed models that included age, sex, BMI, systolic BP, diastolic BP, ratio of fasting total cholesterol to HDL, estimated glomerular filtration rate, BP medications taken, diabetes, smoking, blood biomarkers, ankle-brachial index, carotid intima-media thickness, left ventricular hypertrophy and systolic dysfunction.
The researchers compared the performance of the models with the ACC/AHA pooled-cohort risk equation and the Framingham risk score in 3,689 participants from the Jackson Heart Study (mean age, 53 years; 64.8% women; 100% black) and evaluated how well they discriminated in cohorts from ARIC and MESA.
The outcome of interest was incident CVD event, defined as first incidence of MI, CHD death, congestive HF, stroke, incident angina or intermittent claudication. Median follow-up was 9.1 years.
Reclassification improvement modest
During the study period, 270 participants had a first CVD event.
A model with standard CVD risk factors, B-type natriuretic peptide and ankle-brachial index modestly improved on a model that did not include B-type natriuretic peptide and ankle-brachial index (C-statistic, 0.79; 95% CI, 0.75-0.83; relative integrated discrimination improvement, 0.22; 95% CI, 0.15-0.3).
However, the researchers found that reclassification improvement with the standard CVD risk factors/B-type natriuretic peptide/ankle-brachial index model was not substantial compared with the ACC/AHA pooled-cohort risk equation or the Framingham risk score.
David C. Goff Jr
When the models were calibrated to events in the ARIC and MESA cohorts, the C-statistics ranged from 0.7 (95% CI, 0.66-0.75) to 0.77 (95% CI, 0.71-0.85), according to Fox and colleagues.
Extending the information
In a related editorial, David C. Goff Jr., MD, PhD, from the department of epidemiology, Colorado School of Public Health, Aurora, and Donald M. Lloyd-Jones, MD, ScM,
Donald M. Lloyd-Jones
According to Goff and Lloyd-Jones, the finding that newer potential risk markers did not improve risk prediction “is consistent with a growing body of evidence underscoring the primacy of the traditional risk factors and the difficulty of improving short-term risk prediction with newer biomarkers.” – by Erik Swain
Disclosure: Fox, Goff and Lloyd-Jones report no relevant financial disclosures. Please see the full study for a list of the other researchers’ relevant financial disclosures.
Perspective
Back to Top
Roxana Mehran, MD, FACC, FACP, FCCP, FESC, FAHA, FSCAI
The authors should be congratulated for their incredible work. African Americans are often underrepresented in clinical studies, and we know very little about the natural history of CVD in this population. This study helps identify those at risk and is an important addition to the knowledge base.
I am not certain that this study was definitive enough to make the important conclusions that the Framingham and Pooled Cohort equations appear to be appropriate for African Americans. First, there were many exclusions, and the final analyzable population may not be representative of the entire African American community at risk for CVD. Of the 5,301 individuals in the original cohort, only 2,802 ended up in the standard-risk and fewer in the important-risk model with subclinical disease. The evaluation of subclinical disease is incredibly important in other studies (such as BIOIMAGE) in reclassifying the risk for CV events. I am not sure that we can conclude definitively that a risk model specific for the black population is not warranted. The population with analyzable data is too small for such an important conclusion.
I would like to see this studied in larger cohorts from more areas across the United States. The fact that 64.5% of the population was female is great, but it is also a limitation. We need more data on young black males.
Click Here to Manage Email Alerts