Is measuring C-reactive protein useful for guiding treatment in women > or = 60 years and men > or = 50 years of age?

Using the results of the JUPITER trial, a recent report estimated that up to 11 million older United States (US) adults with C-reactive protein (CRP) levels > or =2 mg/L not currently recommended statins may benefit from treatment. However, the need to measure CRP in making this treatment decision has not been evaluated. Using data from 887 older US men and women (men > or =50 years old, women > or =60 years old) not currently on or recommended statin therapy participating in the National Health and Nutrition Examination Survey 2003 to 2006, we determined the sensitivity, specificity, and positive and negative predictive values of patient characteristics in identifying the presence of CRP > or =2 mg/L. If CRP > or =2 mg/L were included as an indication for statin therapy, then 90% of older US adults would be recommended treatment. Patients with CRP > or =2 mg/L were more likely (p <0.05) to be current smokers, obese, and have chronic kidney disease. However, characteristics (including demographics, cigarette smoking, obesity, chronic kidney disease, and metabolic syndrome) had low positive predictive values (<70%) for identifying patients with CRP > or =2 mg/L and negative predictive values (<60%) for those with CRP <2 mg/L. In conclusion, these findings suggest patient characteristics cannot be easily used to identify patients with CRP > or =2 mg/L. Given the demonstrated benefits of statin therapy, cost of measuring CRP, and large percentage of older US adults with high CRP, universal statin therapy for older US adults warrants investigation.