Malik et al. (1) report on the usefulness of C-reactive protein (CRP) in stratifying risk in patients with the metabolic syndrome and diabetes.
Studies on normal volunteers showed intraindividual variability, which in ∼50% of individuals was sufficient to change their CRP-related risk category (2).
Bogaty et al. (3) have also demonstrated what seems to be spontaneous fluctuation of CRP in stable patients with coronary artery disease (CAD).
Intraindividual biological variation data for high-sensitivity CRP (hsCRP) needs to be established for each individual before using the level to estimate risk and prognosis.
Reliance on single or even the mean of two measurements 2–4 weeks apart is clearly unacceptable, and there is conflict in the published literature regarding the number of samples that should be tested and the time span (3).
Many laboratories use conventional CRP assays, which report levels <5 mg/l as normal. These assays are obviously unsuitable for risk assessment, as it now seems that levels as low as 2 mg/l confer additional risk (3,4). Laboratories should be able to provide hsCRP assays for the purpose of cardiovascular risk stratification.
CRP seems to be an important player in the inflammatory component of atherosclerosis and an independent predictor of adverse CAD outcomes (5,6). Adding it formally to risk stratification scoring methods would improve our ability to identify high-risk patients in both primary and secondary prevention. Before adopting this strategy, it is important to decide how many measurements should be checked, over what interval, and under what conditions.
Also, what weight should be given to the presence of an elevated CRP? Should it influence risk scores qualitatively, like diabetes, or quantitatively, like systolic blood pressure and cholesterol; the higher the level, the higher the risk? How should CRP level be incorporated into standard scoring systems such as the Framingham risk score?
Until these issues are addressed, CRP measurements should perhaps be reserved for problematic or borderline cases, where the decision to use an intervention, whether medical or procedural, is difficult.
