Why is Atherosclerotic Cardiovascular Disease Risk Overestimated by the ACC/AHA Pooled Cohort Equation?

The 2013 ACC/AHA Cholesterol Guidelines expand the recommendations for statin use to populations previously felt to be at lower risk. Central to risk-estimation in these guidelines is a new equation for determination of atherosclerotic cardiovascular disease (ASCVD) risk. However, this risk model has been criticized overestimating ASCVD risk in validation studies of the model.  Using the Women’s Health Study, Cook et al. sought to determine the reasons for risk-overestimation by the ACC/AHA model. Among 27,542 women, 632 experienced an ASCVD event, defined as any myocardial infarction, any stroke, or death from a cardiovascular cause. The average 10-year predicted risk was 3.6% in comparison to an observed risk of 2.2%. When stratified by risk, the ratio between predicted to actual rates was greater for lower risk groups (less than 7.5% risk; ratio 1.90 or higher) than higher risk groups (greater than or equal to 7.5% risk; ratio over 1.4). Statin use and revascularization rates increased over the period of the cohort study, particularly among patients at highest ASCVD risk. Sequentially controlling for statin use, revascularization procedures, and confounding by indication did not attenuate the overestimation of risk by the model.  Failure to capture events was unlikely to explain these differences given the given the excellent follow-up (97.2%) in the Women’s Health Study.


Changing trends in statin use, revascularization, and underascertainment of events failed to explain the overestimation of ASCVD risk by the ACC/AHA risk prediction model.  Further efforts are needed to recalibrate the model using contemporary data to ensure optimal guidance in ASCVD risk reduction recommendations.

Summarized by Jehu S. Mathew and Steven M. Bradley

  •  Cook NR, Ridker PM. Further Insight Into the Cardiovascular Risk Calculator: The Roles of Statins, Revascularizations, and Underascertainment in the Women’s Health Study. JAMA Intern Med. 2014 Dec 1;174(12):1964-71.