In a large contemporary, multiethnic population, the American College of Cardiology/American Heart Association (ACC/AHA) Pooled Cohort Risk Equation overestimated the overall 5-year risk of atherosclerotic cardiovascular disease (ASCVD) events across sociodemographic subgroups.
Study researchers sought to evaluate the 2013 ACC/AHA risk equation within a large, multiethnic population in clinical care to determine the accuracy of predicted ASCVD events.
The study included 307 591 patients without diabetes between 40 and 75 years of age. Of these, 22 283 patients were black, 52 917 were Asian/Pacific Islander, and 18 745 were Hispanic.
The cohort was selected from the Kaiser Permanente Northern California integrated health care delivery system in 2008, and followed the patients until 2013. Patients with known ASCVD, diabetes, low-density lipoprotein cholesterol levels <70 or ≥190 mg/dL, prior lipid-lowering therapy, or incomplete 5-year follow-up data were excluded from the study.
ASCVD event rates were calculated using validated algorithms for hospitalization databases and death certificates. Predicted 5-year ASCVD risk (overall and according to sex and race/ethnicity) was compared against observed risk. Predicted and observed risk was also compared in patients with diabetes.
There were 2061 observed ASCVD events during 1 515 142 person-years, including 1464 (0.5%) acute myocardial infarctions, 525 (0.2%) coronary heart disease-related deaths, and 71 (0.02%) ischemic strokes.
The results showed significantly lower observed 5-year ASCVD incidence compared with predicted risk in each category. The observed incidence was 0.20% (95% confidence interval [CI]: 0.20-0.25) among patients who had a predicted risk <2.50%, 0.65% (95% CI: 0.55-0.70) for a predicted risk of 2.50% to <3.75%, 0.90% (95% CI: 0.75-1.00) for a predicted risk of 3.75% to 5.00%, and 1.85% (95% CI: 1.75-1.95) for predicted risk ≥5.00% (C statistic: 0.74).
In addition to systematic overestimation of 5-year incidence, researchers also observed poor calibration in both sexes. The discrimination was better among women (C statistic: 0.72) compared to men (C statistic: 0.68).
“Given that Asian/Pacific Islander and Hispanic subjects are considered under ‘white or other’ in the ASCVD risk equation, we compared the observed vs predicted 5-year ASCVD incidence separately among non-Hispanic white, non-Hispanic black, Asian/Pacific Islander, and Hispanic subjects,” the authors explained.
They found systematic overestimation and poor calibration of actual ASCVD risk in each ethnic group, although discrimination varied in each group (C statistic: 0.70 to 0.74).
Researchers also conducted a risk analysis among 4242 adults with diabetes, and found that calibration was improved, but discrimination was worse (C statistic: 0.64).
“Our study provides evidence to support calibration of the ASCVD Pooled Cohort Risk Equation in adults without diabetes, especially given the public health implications of widespread application of this risk calculator,” they concluded. “Ongoing research and dialogue in this area remains crucial and should be encouraged to provide more rigorous, valid evidence in contemporary, diverse populations.”
Rana JS, Tabada GH, Solomon MD, et al. Accuracy of the atherosclerotic cardiovascular risk equation in a large contemporary, multiethnic population. J Am Coll Cardiol. 2016;67(18): 2118-2130. doi: 10.1016/j.jacc.2016.02.055.