Correlation of postpercutaneous coronary intervention creatine kinase-MB and troponin I elevation in predicting...

Creatine kinase-MB (CK-MB) and troponin I elevations after successful percutaneous coronary intervention (PCI) are common, and different gradations have been correlated with mortality. To establish which of these 2 markers of myonecrosis, CK-MB and troponin I, accurately predicts mortality after successful PCI, we analyzed 2,873 patients without acute myocardial infarction who underwent PCI for in-hospital events and mid-term mortality. Patients were stratified into 4 groups based on peak post-PCI cardiac markers values: group I: normal CK-MB (<16 U/L) or troponin I (<2 ng/ml); group II: CK-MB or troponin I levels 1 to 3 times normal; group III: >3 to 5 times normal; and group IV: >5 times normal. CK-MB elevation occurred in 16.1% of patients, with 12.2%, 2.3%, and 1.6% in groups II to IV, respectively. Troponin I elevation was detected in 38.9% of patients, with 16.4%, 8.4%, and 14.1% in groups II to IV, respectively. There was poor correlation between postprocedural CK-MB and troponin I values (r = 0.10) and in their individual subgroups. Kaplan-Meier estimates of death for postprocedure CK-MB were 2.1%, 2.7%, 1.7%, and 10.3% (p = 0.002) for groups I to IV, respectively; for troponin I, these estimates were 2.2%, 2.3%, 2.9%, and 2.1% for groups I to IV, respectively (p = 0.58). A Cox proportional hazards model showed that CK-MB >5 times normal was the strongest predictor of mortality (hazard ratio 6.7, 95% confidence interval 1.9 to 22.9; p = 0.002), although heart failure, peripheral vascular disease, pre-PCI digoxin therapy, and post-PCI renal failure also predicted mortality. However, neither troponin I peak elevation nor any subgroup predicted mortality. Troponin I is frequently elevated after PCI, but does not predict mortality. Periprocedural CK-MB elevation >5 times normal remains an independent predictor of mid-term mortality and a valuable marker for PCI prognosis in low-to-medium risk patients.