BIRMINGHAM, AL and BOSTON, MA — At the 2013 American Heart Association Scientific Sessions, the launch of the new calculator designed to assess the 10-year risk of atherosclerotic cardiovascular disease (ASCVD) took a hit when two researchers suggested it might be flawed.
As reported by heartwire at the time, Drs Paul Ridker and Nancy Cook (Brigham and Women's Hospital, Boston, MA) calculated the 10-year risk of cardiovascular events in three large-scale primary-prevention cohorts—the Women's Health Study (WHS), the Physicians' Health Study (PHS), and the Women's Health Initiative Observational Study (WHI-OS)—and found the new algorithm overestimated the risk by 75% to 150%.
In addition, the algorithm also significantly overestimated risk in the Multiethnic Study of Atherosclerosis (MESA) and Reasons for Geographic and Racial Differences in Stroke (REGARDS) studies, two validation cohorts that were assessed by the new guideline committee.
Now, in a paper published online December 11, 2013 in Circulation, the REGARDS researchers provide some clarification on their trial and suggest the new cardiovascular disease risk algorithm might be more valid than suggested by Ridker and Cook[1]. In the letter to the editor, Drs Paul Muntner, Monika Safford, and George Howard (University of Alabama, Birmingham), along with Dr Mary Cushman (University of Vermont, Burlington), suggest four possible reasons why the risk calculator might appear to overestimate risk.
REGARDS and MESA did not include costly surveillance procedures of clinical events and were unable to retrieve the medical records of all patients. By possibly not ascertaining all clinical events, this might have had a large impact on the absolute CVD-incidence rates in Ridker and Cook's analysis.
Statin use has increased significantly over the past decade, and this could contribute to the overestimation of CVD risk in comparison with the observed rates, particularly in patients who start the drugs after their baseline examination visit.
Increased use of PCI since the time of studies used to develop the risk equation might contribute to the overestimation of event rates in the validation studies.
And finally, MESA and REGARDS had relatively short follow-up during which there were a limited number of events.
"In our opinion, these four considerations lead us to the conclusion that it is premature to draw firm conclusions about potential overestimation of risk using the new ASCVD risk equation," write Muntner et al.
In response, Ridker and Cook say that inadequate ascertainment might have occurred in MESA and REGARDS, but it was unlikely to have occurred to the same extent in WHS and PHS, two studies with excellent follow-up[2]. In addition, they say that while "incident statin use could explain some of the discrepancy and lack of model fit seen in the more contemporary external cohorts, we believe this effect to be small." In WHS, 22% of women were taking statins at 10 years, and assuming a 25% reduction in risk with statin therapy, the new risk model still overestimates cardiovascular disease events by 40% to 100%.
The increasing of revascularization, they concede, is an important issue, because it is reflective of clinical practice. In fact, "a strong argument can be made for including revascularization as an end point for risk calculators as well, an approach we took in our own prediction modeling for the Reynolds Risk score," write Ridker and Cook.
Regarding the relatively short-term follow-up in REGARDS, they agree it might have led to some instability in the risk estimates. Further follow-up will provide valuable information. "Given that the new risk calculator has overestimated risk in the five contemporary external validation cohorts so far evaluated, we look forward to seeing additional data from other cohorts and working toward clinical resolution," conclude Ridker and Cook.
Muntner has served on advisory boards and has received research grants from Amgen. Safford has received research funding from Amgen and diaDexus and has consulted for diaDexus. Cushman has received research funding from diaDexus. Howard has no disclosures. Ridker receives research support from Amgen, AstraZeneca, Novartis, and Pfizer. Ridker is also listed as a coinventor on patents held by the Brigham and Women's Hospital related to inflammatory biomarkers in cardiovascular disease and diabetes that have been licensed to Siemens and AstraZeneca. Cook has no conflicts of interest.
Heartwire from Medscape © 2013 Medscape, LLC
Cite this: New Cardiovascular Disease Risk Calculator Debated Again - Medscape - Dec 13, 2013.
Comments