Background: C-reactive protein (CRP), a sensitive marker of the acute-phase response, has been associated with future cardiovascular endpoints independently of other risk factors. A joint analysis of the role of risk factors in predicting mean concentrations and variation of high-sensitivity CRP (hsCRP) in serum has not been carried out previously.
Methods: We used data from 1003 myocardial infarction (MI) survivors who had hsCRP measured monthly up to 8 times and multivariate mixed effects statistical models to study the role of time-variant and -invariant factors on the geometric mean of and the intraindividual variation in hsCRP concentrations.
Results: Patients with > or =6.5% glycosylated hemoglobin (HbA1c) had 26.2% higher hsCRP concentrations (95% CI, 7.2%-48.6%) and 20.7% greater variation in hsCRP values (P = 0.0034) than patients with lower baseline Hb A(1c) values (<6.5%). Similar but less pronounced differences were seen in patients with a self-reported diagnosis of type 2 diabetes. hsCRP concentrations showed less variation in patients who reported angina pectoris, congestive heart failure, or emphysema (-11.0%, -24.9%, and -41.6%, respectively, vs patients without these conditions) but greater variation in males and smokers (+24.8% and +27.3%, respectively, vs females and nonsmokers). Exposures in the 24 h before blood sampling, including exposure to environmental tobacco smoke, alcohol consumption, and extreme stress, did not have a major impact.
Conclusions: One or 2 hsCRP measurements may not be sufficient to adequately characterize different patient groups after MI with similar precisions. We found hsCRP concentrations to be especially variable in males, smokers, and patients with increased Hb A(1c) values.