The threshold of creatine kinase elevation after coronary interventions has been set at levels ranging in different studies from 2 to > 5 times the laboratory's upper limit of normal. This high variability is caused by the absence of any systematic evaluation of the prognostic implications of cardiac-enzyme elevation in this setting. This study was undertaken to evaluate the clinical, morphologic, and procedural correlates, and the long-term follow-up of two commonly used thresholds of creatine kinase (CK) elevation after successful percutaneous coronary interventions, in an attempt to define the level of postprocedural cardiac enzymes that correlates with adverse clinical outcome. We examined 4664 consecutive patients who underwent successful coronary angioplasty or directional atherectomy at the Cleveland Clinic. Group I (4480 patients) had CK > or = 2 times control levels after the procedure (i.e., < or = 360 IU/L). Group II (123 patients) had a peak level between 361 and 900 IU/L, and group III (61 patients) had a peak level >900 IU/L with positive myocardial isoenzymes (CK-MB > 4%). Elevation of cardiac enzymes was associated with distinct clinical, morphologic, and procedural characteristics, including coronary embolism, recent infarction, transient in-laboratory closure, hemodynamic instability, vein graft procedures, and large dissections. Clinical follow-up was available in 4644 (99.6%) patients, with a mean follow-up of 36 +/- 22 months. Kaplan-Meier survival analysis adjusted with Cox proportional hazards regression model showed that cardiac-enzyme elevation was an important correlate of cardiac death (risk ratio, 2.19; p < 0.0001). The groups with elevated cardiac enzymes had a higher incidence of cardiac death compared with group I (p < 0.0001). There was also a trend toward more cardiac hospitalizations in the same groups (p = 0.15). The incidence of cardiac death and cardiac hospitalization on follow-up was not different between groups II and III, This study shows that CK elevations between 2 and 5 times control values after successful coronary interventions are associated with an adverse long-term outcome. The findings suggest that an appropriate CK threshold that has prognostic implications would be twice the upper limit of normal.