C-Reactive Protein in the Prediction of Cardiovascular Disease

Publisher Website
Google Scholar
Abstract
Ridker et al. provided a stimulating article on inflammatory markers and cardiovascular disease in women (March 23 issue).1 Not surprisingly, the popular press picked up on the article and gave their findings prominent coverage. It is not clear that Ridker et al. wanted this to happen. Their findings are cast in terms of relative risk only, not in terms of traditional predictive value; it is the latter that is more relevant to the practicing physician.2 That is, we learn that subjects in the highest quartile for high-sensitivity C-reactive protein (hs-CRP), relative to those in the lowest quartile, had a 4.4-fold risk of cardiovascular events. However, the overall risk was just 0.4 percent (122 events in 28,263 subjects over a period of three years). We suspect that the positive predictive value (the proportion of all subjects with “elevated” levels of hs-CRP who had cardiac events) in this population was low.