February 16th, 2015

Cardiovascular Risk Calculators Overestimate

Calculating cardiovascular risk has become a central and highly controversial component of cardiovascular guidelines. Now a study in the Annals of Internal Medicine finds that most of the commonly employed tools seriously overestimate the risk of people today.

Researchers used data from 4,227 people enrolled in the Multi-Ethnic Study of Atherosclerosis (MESA) between 2000 and 2002 to assess the predictive accuracy of the new AHA/ACC atherosclerotic cardiovascular disease (ASCVD) tool and four alternative risk scores: FRS-CHD, the ATP-III-FRS-CHD, the Reynolds Risk Score (RRS), and the FRS-CVD.

They found that four out of the five scores overestimated risk for men by as little as 37% and by as much as 154% and for women by as little as 8% and as much as 67%. The RRS (Reynolds Risk Score) overestimated risk in men by 9% but underestimated risk in women by 21%.

Current AHA-ACC guidelines recommend that statin therapy should be considered for adults with a 10-year risk greater than 7.5%. But the researchers found that the actual event rate was only 3% in men and 5.1% in women among people with an AHA-ACC-ASCVD risk score of 7.5%-10%. Additional analyses found that this discrepancy was not explained by higher than expected use of aspirin, statins, or blood pressure drugs.

The authors noted that the cohorts used to develop the risk scores are “decades old” and quite different from “the more modern MESA cohort.” The implications of the findings are not just academic. “Overestimation of risk,” the authors write, “would likely result in increased use of preventive medications, potentially exposing some patients to the unnecessary risks of these drugs. In particular, aspirin use in primary prevention is known to be a delicate balance between cardiovascular risk reduction and increased incidence of bleeding.” In addition, they note, “overestimation of risk has implications for public health planning with accompanying financial ramifications for a heavily burdened health care system.”

In an accompanying editorial, Paul Ridker and Nancy Cook write that the study “serves as a crucial test of external validation and demands the attention of both clinicians and policymakers.”

Speaking on behalf of the AHA, David Goff pointed out two significant reasons that may explain in part why the AHA risk calculator overestimates events in the MESA cohort. First is the inclusion of data from the lower-risk Chinese-American and Hispanic-American participants. These data had been excluded from the AHA’s analysis and, for this reason, the new guidelines include a caution about using the calculator in this population. Second, said Goff, the AHA equations are meant to be used in untreated people but over 80% of the MESA cohort received some sort of preventive treatment during the study period. Although the Annals investigators sought to correct for this, Goff said that it is extremely difficult to fully take into account the downstream effect of treatment.

Finally, Goff emphasized that although “there’s been a lot of attention paid to just exactly how well these equations work” we shouldn’t “lose sight of our goal” of reducing the burden of cardiovascular disease. “It’s really important not to let the perfect become the enemy of the good,” he said.




Comments are closed.