M3. Calculate 10-Year Risk Score for CVD

objective

Determine short-term risk (i.e., over ten years) as the basis for determining the type and intensity of interventions.

background

The magnitude of efficacy of interventions for dyslipidemia in preventing CVD is dependent upon the absolute risk for coronary heart disease as determined by a compilation of major risk factors.  The higher the risk, the greater the absolute risk reduction associated with dyslipidemia interventions.  Determining short-term risk (i.e., over ten years) serves as the basis for determining the type and intensity of interventions.

recommendations

1. A global 10-year risk for CVD should be calculated to assess the short-term (10-year) absolute risk of a CVD event.  [A]
2. The Framingham Risk Calculator should be used, as it is the most commonly used and readily available calculator validated in numerous populations.  [I]
   http://hin.nhlbi.nih.gov/atpiii/calculator.asp?usertype=prof
3. Other risk markers or measure of atherosclerotic burden may be useful to adjust the risk category, if they have been validated to have independent prognostic value.  [C]

discussion

Using conventional risk factors (age, male gender, blood pressure, tobacco smoking, and cholesterol level) to derive a composite measure of absolute risk for CVD in the subsequent 10 years is now recommended by the AHA and NCEP as an initial step to determine the type and intensity of lipid interventions (NCEP ATP-III, 2002; Sheridan et al., 2003; Grundy et al., 1999; Wilson et al., 1998; Bethesda Conference, 1996; Grundy et al., 2001; Grundy et al., 1999).  There are several risk prediction tools, all of which use conventional risk factors (Sheridan et al., 2003).  The most commonly used calculator is derived from the Framingham Study that has been validated in multiple U.S. and international populations (Sheridan et al., 2003; Grundy et al., 1999; Wilson et al., 1998).  Its limitations include overestimation of risk in younger (<age 40) populations and certain ethnic groups (e.g., Japanese and Hispanic) as well as potentially creating a false sense of reassurance in young populations with high relative risk but low absolute risk, since age is the strongest variable in predicting CVD risk.  There are no clinical trials that have determined the clinical outcomes impact of using a risk calculator for lipid intervention decision-making.  However, since all of the lipid trials have proven efficacy only in patients at high absolute risk, it is rational to use a systematic tool to accurately define whether a patient meets the characteristics of the populations in whom lipid-lowering therapies have proven effective.

Risk Score Calculation and Validity of Scoring Tools

Strategies that explicitly consider CVD risk factors in addition to lipid levels are more accurate than those that measure only lipid levels.  Grover et al. found that a Framingham-based coronary risk model was a better predictor of CVD mortality compared to LDL-C/HDL-C ratio, TC/HDL-C ratio, or TC alone (Grover et al., 1995).

In a later study, Grover et al. (2000) found that the Framingham risk equations were more accurate than counting risk factors for predicting coronary artery disease (CAD) risk.  Risk counting was a particularly poor method for predicting risk for women.  Calculating risk using risk equations is a more accurate method to identify people at high-risk for CVD than counting the number of risk factors present, especially for women.

There is emerging evidence that genetic, serologic, physiologic, psychosocial, and anatomic markers of CHD risk can add prognostic value to the Framingham Risk Score (FRS) (Ridker, 2001; Pearson et al., 2003; O’Donnel et al., 2004; Ford et al., 1998; Greenland et al., 2000 & 2004; Pletcher et al., 2004).  The risk markers with proven independent prognostic value are: High Sensitivity C-reactive Protein (hsCRP) (>3mg/dL), first degree family history of premature CAD, metabolic syndrome, elevated carotid intima-media thickness, decreased brachial artery reactivity, history of major depressive disorder, coronary artery calcification (>75percent for age and gender), and microalbuminuria with impaired renal function.

Although there is insufficient evidence at this time to recommend routine screening for these risk markers, it may be useful in the intermediate risk patient for whom it is less convincing that drug therapy would have a meaningful impact on outcomes.  If any of this information is in the abnormal range, it would be reasonable to multiply the predicted ten-year risk calculated from the FRS by the adjusted relative risk associated with the abnormal risk marker, and then base dyslipidemia management on the resulting category of risk. (See Figure 1 for a schematic on how to adjust risk based on new information from new independent prognostic information).  However, at this time such a strategy of risk determination for the purposes of guiding lipid management has never been proven to be associated with improved outcomes.

Schematic to help adjust the 10-year risk associated with new test information (e.g., C-reactive protein [CRP] or high coronary artery calcium [CAC] scores on electron beam computed tomography [EBCT]) that has incremental prognostic value, independent of conventional risk factors (i.e., FRS).  One should first determine the 10-year risk using the FRS to determine the x-axis point on the identity line, and then use the adjusted relative risk associated with the new information to multiply the 10-year FRS.  For example, if the FRS indicates a 10-year risk of 13 percent, and the patient has a high CRP (>3mg/L) which has an approximate adjusted relative risk of 2, then the adjusted 10yr risk would be 13 percent x 2 = 26 percent (i.e., a CHD risk equivalent).  While a rational approach, such a strategy of refining risk has not been sufficiently validated in prospective studies, but can be helpful in guiding care in the face of uncertainty surrounding new prognostic test information.