American Association for Clinical Chemistry
Better health through laboratory medicine
March 29, 2007

In this Issue...

Predicting Cardiovascular Risk in Women: The Reynolds Risk Score
By Julie McDowell

Even though up to 20% of all coronary events in women occur in the absence of major risk factors such as age, hypertension, smoking, diabetes, and hyperlipidemia, current algorithms still rely on these variables to predict risk. While the clinical utility of additional risk markers—like high-sensitivity C-reactive protein (hs-CRP), plasma creatinine, and homocysteine levels—has been established, their role in improving risk prediction algorithms remains unclear. This issue of Strategies looks at recent research by renowned cardiology investigator Paul M. Ridker, MD, MPH, and his colleagues, who have developed the Reynolds Risk Score, which evaluates a woman’s chance of developing cardiovascular disease based on family history, hs-CRP, HDL-C, total cholesterol, systolic blood pressure, current smoking habits, and other risk markers.

Since its launch in 1991 and conclusion in 2004, the Women’s Health Study (WHS) has revealed not only key findings on the role of aspirin in cardiovascular disease prevention, but also the relationship between risk factors and coronary events. Recently, leading Boston-based cardiologist Paul M. Ridker, MD, MPH and his colleagues mined data on almost 25,000 WHS participants followed over 10 years to assess 35 risk factors for coronary events (JAMA 2007; 297:611–619). Based on their analysis, they developed two clinical algorithms—a best-fitting model and a clinically-simplified version called the Reynolds Risk Score—that both predict 10-year cardiovascular disease risk with improved accuracy. In fact, these algorithms reclassified 40%–50% of WHS participants at intermediate risk into higher or lower-risk categories.

“In the set where we validated the score—of which there were about 8,000 women—we found that the Reynolds score predicted risk better than the standard risk scores previously developed,” explained Samia Mora, MD, one of the study’s co-investigators, a cardiologist at Brigham and Women’s Hospital in Boston, and Instructor in Medicine at Harvard Medical School.

Current guidelines from the third National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP-III) recommend that all adults should be evaluated for their individual risk of coronary heart disease using the Framingham risk score, which is based on data from Framingham Heart Study, which primarily involved men. Among the women included in the Framingham study, few had cardiovascular disease events, said Mora. The Reynolds Risk Score study is unique because it looked only at women, in contrast to other research that looks at both men and women. “The other unique aspect about this research is that this study found that two markers of risk appear to be very important,” she explained. “These are hs-CRP and the other is the parental history of whether a parent had a heart attack before age 60. The other factors that came out that were important were the standard factors, which were cholesterol, blood pressure, whether or not the woman smoked, whether or not she had diabetes, and her age.”

hs-CRP: A Standout Marker

Through their analysis of the WHS data, Ridker and his co-investigators examined the role of numerous markers in risk prediction. In addition to hs-CRP, these included apolipoproteins A-I and B-100, high density lipoprotein cholesterol (HDL-C), as well as inflammatory biomarkers such as soluble intercellular adhesion molecule 1 (sICAM-1) and fibrinogen. They also looked at markers of glycemic control such as hemoglobin A1c, plasma creatinine, and homocysteine levels.

Out of all of these biomarkers, the researchers found hs-CRP to be the strongest, said Mora. In addition, they found that while alternative cholesterol measurements—such as apolipoproteins B-100 and A-1—are important predictors of risk, HDL and total cholesterol measurements would suffice. “These lipid measures appear to be important as well, but could be substituted by putting in the HDL cholesterol, without losing that much information,” said Mora. “This means that those lipid measures are important. But from a clinical perspective, it can be simplified into HDL and total cholesterol.”

Improving Prediction

The findings in the Ridker group’s study and the Reynolds Risk Score are prompting many to emphasize the value of both hs-CRP and family history in evaluating women for heart-disease risk. In an accompany editorial in JAMA, Johns Hopkins University cardiologists Roger S. Blumenthal, MD, and Erin D. Michos, MD, espoused the potential improvements of the Reynolds Risk Score over the Framingham score (2007; 297:641–643). Unlike the Reynolds algorithm, the Framingham risk score does not incorporate hs-CRP or family history.

“When you use the Framingham Risk score, the majority of women get low risk scores,” said Michos, who is a clinical cardiology fellow at Hopkins’ Ciccarone Preventive Cardiology Center. “We are concerned that some of these women are not really low risk, and might not be treated soon enough. By putting hs-CRP and family history into this new model, we feel the advantage of the Reynolds Risk Score is that it helps predict a bit more accurately which women are going to have cardiovascular disease in the next 10 years.”

It’s important to note that Ridker’s findings don’t necessarily mean that more women should be on a statin drug, said Michos. However, based on the Reynolds model, she and Blumenthal estimate that approximately 20% of women will have different lipid treatment goals. By incorporating hs-CRP and a family history of premature coronary disease into the risk equation, clinicians can better target which low- to intermediate-risk women really need the statins. “Since 8–10 million women fall in to this intermediate risk group, by having a better predictive model, it has important impact on cardiovascular prevention in figuring out which women really need a statin,” she explained.

Moving forward, these findings need to be evaluated in other populations, including men and different female ethnic populations, since 95% of the participants in the WHS were white, said Mora. Ridker’s research group is also examining what genetic risk factors are important in predicting cardiovascular disease in women.

“Maybe there is a simple test that could be developed that could incorporate genetics, hs-CRP, and family history, as well as some other risk factors, such as age and diabetes state, to predict accurately which women are at risk,” she said.

Please help us improve Strategies by rating how useful this article is to your job.