Predicting Cardiovascular Risk:  So What Do We Do Now?

Clinicians who see patients for primary care or for primary prevention of cardiovascular disease (CVD) are fortunate to have multivariable models^1- 2 that can predict, with reasonable reliability, the absolute risk for future CVD events in large segments of the population.^3- 4 Even in 2006, the absolute risk for future development of disease can be predicted for very few diseases, much less with the precision afforded by current CVD risk prediction models. These “risk scores” are a major advance over clinical risk prediction using relative risk estimates, and CVD prevention is one of the few areas in clinical practice to incorporate the use of absolute risk prediction into clinical practice guidelines.² However, it has been widely recognized that CVD risk prediction models are imperfect. Appropriately, many investigators are attempting to improve CVD risk prediction to refine our ability to estimate absolute risk and to find those individuals who appear to be at low risk but are actually at high risk and merit preventive therapies. There is intense ongoing debate about which factors, if any, should be considered for addition to existing CVD risk prediction algorithms.