A new, much more accurate clinical algorithm for predicting cardiovascular risk in women has been developed and validated by investigators in the Women's Health Study.
The new method reclassified approximately half of women who had been previously categorized as intermediate risk into either low-risk or high-risk categories. If the new algorithm were applied to a representative population of 100,000 U.S. women who are now considered to be at intermediate risk, it would recategorize 13,500 of them as low risk, 48,500 as low to moderate risk, 32,500 as moderate to high risk, and 5,400 as high risk, researchers reported.
“As 8–10 million U.S. women have an estimated 1-year risk between 5% and 20%, application of these data could have an immediate effect on cardiovascular prevention,” allowing more accurate targeting of statin and other therapies, wrote Dr. Paul M. Ridker of Brigham and Women's Hospital, Boston, and his associates in the WHS.
The investigators used data on a subgroup of 16,400 healthy WHS subjects to assess which of 35 possible risk markers would best predict CV risk, then developed a model (model A) that included the nine most valuable predictors. Next, they modified the model to create a simplified version (model B) that would be more practical for clinical use. They tested the validity of both models in another subgroup of 8,158 WHS subjects.
During a median follow-up of 10 years, 504 CV events occurred in the first cohort and 262 occurred in the validation cohort.
With model A, 43% of women who had been classified as being at intermediate risk by traditional criteria were reclassified as being at either lower or higher risk. Of these 681 reclassified subjects, all but 93 were placed into more accurate risk categories, based on their 10-year clinical outcomes.
Among subjects who did not have diabetes, approximately 50% who had been classified as being at intermediate risk by traditional criteria were reclassified as being at either lower or higher risk. Of these 722 reclassified subjects, all but 2 were placed into more accurate risk categories (JAMA 2007;297:611–9).
Similar results were obtained using the simplified model B, which the researchers have termed the Reynolds Risk Score. For the 647 subjects without diabetes who were reclassified using model B, all but 6 were placed into more accurate risk categories, they said.
The Reynolds Risk Score uses these eight clinical markers to predict risk: age, systolic blood pressure, hemoglobin A1c if the patient is diabetic, current smoking status, total and HDL cholesterol levels, high-sensitivity C-reactive protein, and parental history of MI before the age of 60.
“A user-friendly calculator for the Reynolds Risk Score can be freely accessed at www.reynoldsriskscore.org
Homocysteine, fibrinogen, soluble intercellular adhesion molecule 1, and creatinine measures did not add to the accuracy of risk prediction, nor did body mass index or exercise frequency, Dr. Ridker and his associates said.
The researchers cautioned that since the study subjects were predominantly white, well-educated women, these findings may not be generalizable to other populations.