Accuracy of the ACC/AHA Cardiovascular Risk Calculator Is Challenged

March 25, 2015

Accuracy of the ACC/AHA Cardiovascular Risk Calculator Is Challenged

  1. Allan S. Brett, MD

When applied to a contemporary cohort, the calculator overestimated 10-year risk by 78%.

  1. Allan S. Brett, MD

Since the 2013 release of the American College of Cardiology and American Heart Association (ACC/AHA) 10-year cardiovascular risk calculator (NEJM JW Gen Med Dec 15 2013 and J Am Coll Cardiol 2013 Nov 12 [e-pub]), several analyses have challenged its accuracy (Lancet 2013; 382:1762 and JAMA Intern Med 2014; 174:1964). Because the calculator is used to select patients for statin therapy, the implications of inaccuracy are important.

In this new analysis, researchers used data from the MESA study — a contemporary, multiethnic, prospective epidemiologic study — to examine the accuracy of five risk-score calculators, including the ACC/AHA calculator. The cohort included 4227 people (age range, 50–74) without clinical cardiovascular disease or diabetes at baseline; average follow-up was 10 years. One risk score (Reynolds) performed relatively accurately in the MESA population. In contrast, three Framingham-based risk scoring systems and the ACC/AHA calculator overestimated risk substantially.

ACC/AHA scores predicted an event rate of 9.16%, but the actual observed rate was only 5.16%; thus, risk was overestimated by 78%. In MESA participants whose ACC/AHA-predicted 10-year risk was 7.5% to 10% (the threshold at which the ACC/AHA cholesterol guideline recommends statin therapy), the actual observed event rate was only 3.0% in men and 5.1% in women.


This study (in conjunction with its predecessors that reached similar conclusions) poses a serious problem for the ACC/AHA calculator. Some commentators have glossed over the issue of overestimated risk, implying that some overtreatment with statins isn't important. But that reaction is not valid, in my view. The rationale for risk prediction is precisely to distinguish between patients who are reasonably likely to benefit from drug therapy and patients whose probability of benefit is marginal or absent.

Editor Disclosures at Time of Publication

  • Disclosures for Allan S. Brett, MD at time of publication Nothing to disclose


Reader Comments (3)

Gail MD Physician, Family Medicine/General Practice

While not perfect, I find that I am putting fewer patients on statins, and poking them for followup lipids, than I did in the days when it seemed like everyone w/and LDL>100 had an indication to start statins. I really like not treating to a specific level any longer. So do my patients.

Herbert Ross MD Physician, Endocrinology, Retired

A meta analysis of RCTs with statins reporting cognitive test results in over 29000 patients over 10 years found no evidence of cognitive impairment with statin use. Ott et al: J. Gen. Int. Med March 2015. The studies included both cognitively normal patients and those with Alzheimer's Disease.

Thomas Smith Physician, Family Medicine/General Practice, Small town, single specialty group

I keep both the new CV risk calculator and the Reynolds risk calculator on my phone and "crunch the numbers" whenever considering starting a patient on a statin. The 78% overcalculation of the new risk calculator can be overcome, by not starting a statin unless that person has coronary disease or has 15% or higher risk. The recommendation to start a statin at 7.5% ten year risk is a huge over reach. I start far fewer patients on statins than I use to and have taken many low risk people off statins. I think the risk of cognitive impairment and peripheral neuropathy due to statins has been undercalculated vs. the benefit of CV reduction.

Your Comment

(will not be published)

Filtered HTML

  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Do you have any conflict of interest to disclose?
This question is for testing whether you are a human visitor and to prevent automated spam submissions.

Vertical Tabs

* Required

Reader comments are intended to encourage lively discussion of clinical topics with your peers in the medical community. We ask that you keep your remarks to a reasonable length, and we reserve the right to withhold publication of remarks that do not meet this standard.

PRIVACY: We will not use your email address, submitted for a comment, for any other purpose nor sell, rent, or share your e-mail address with any third parties. Please see our Privacy Policy.