After collecting baseline blood samples for CK-MB, TnI, and aspirin responsiveness, all patients received an oral loading dose of 300-mg clopidogrel 12 to 24 h before the procedure. The PCI procedure was performed on the following day according to standard practice, after the patients received an additional 75-mg maintenance dose of clopidogrel. Unfractionated heparin 70 U/kg or enoxaparin 1 mg/kg was used for procedural anticoagulation at operator discretion. Following the procedure, blood samples for TnI were collected at 12 to 24 h, whereas those for CK-MB were collected at 6 to 8 h. If CK-MB was elevated, serial measurements every 8 h were obtained and the peak level was recorded. The CK-MB was considered elevated if ≥16 U/l, which was further subdivided into 1 to 3 × (16 to 48 U/l), 3 to 5 × (49 to 80 U/l), and >5 × (>80 U/l) normal. A TnI value of ≥2.0 ng/ml was considered elevated. Aspirin-induced platelet inhibition was measured using a commercially available point-of-care assay, the Ultegra Rapid Platelet Function Assay-ASA (RPFA-ASA) (Accumetrics Inc., San Diego, California). Citrate-anticoagulated blood 2 ml was added to RPFA-ASA cartridges, which contain fibrinogen-coated beads and platelet agonists. If aspirin has produced the expected antiplatelet effect, fibrinogen-coated beads will not agglutinate, and light transmission will not increase. The result is expressed as aspirin reaction unit (ARU). An ARU ≥550 indicates the absence of aspirin-induced platelet dysfunction, based on correlation with epinephrine-induced light transmission aggregometry in aspirin-naive patient tested prior to and between 2 to 30 h after aspirin (325 mg) ingestion (15), and is defined as aspirin-resistant. From this study, both the sensitivity (92%) and the specificity (85%) of this assay were determined. The coefficient of variance was 2.5% on repeated measures within patients. The between-patient coefficient of variance was 12.5% for baseline samples and 15.0% for post-aspirin samples. Digital angiograms were analyzed off-line using a computer-based edge-detection program (CMS-GFT, MEDIS, Leiden, The Netherlands) by experienced cardiologists who were unaware of the patient characteristics and outcomes.