High-sensitivity C-reactive protein (CRP), proposed as a new coronary risk marker, may reflect either an acute phase reaction or the level of chronic inflammation. Thus, CRP may be less predictive of long-term outcomes when measured after acute myocardial infarction (AMI) than after unstable angina pectoris (UAP) or stable angina pectoris (SAP). A total of 1,360 patients with severe coronary artery disease (>/=1 stenosis >/=70%) had CRP levels obtained at angiography. Presenting diagnoses were SAP (n = 599), UAP (n = 442), or AMI (n = 319). During follow-up (mean 2.8 years), death or nonfatal AMI (D/AMI) occurred in 19.5%, 16.1%, and 17.2% (p = NS) with SAP, UAP, and AMI, respectively. Corresponding median CRP levels were 1.31, 1.27, and 2.50 mg/dl (p <0.001). For the overall cohort, increasing age, low ejection fraction, revascularization, and elevated CRP were the strongest of 6 independent predictors for D/AMI. Among those presenting with SAP, CRP levels above the first tertile were associated with an adjusted hazard ratio of 1.8 (95% confidence interval [CI] 1.2 to 2.8, p <0.009) for D/AMI. After UAP, the hazard ratio was 2.7 (95% CI 1.4 to 5.0, p <0.002). However, when measured during hospitalization for AMI, CRP was not predictive of long-term outcome (hazard ratio 1.0 [95 % CI 0.5 to 1.7] p = 0.86). In conclusion, predischarge CRP levels are higher after AMI than after UAP or SAP. However, whereas CRP is strongly predictive of long-term D/AMI for patients presenting with SAP or UAP, it is not predictive shortly after AMI, suggesting that measurements should be delayed until the acute phase reaction is over and levels have returned to baseline.