BACKGROUND: Platelet glycoprotein IIb/IIIa inhibitor (GPI) and thrombus aspiration (TA) can reduce microembolic incidence in primary percutaneous coronary intervention (PCI) for ST-segment elevation myocardial infarction (STEMI) with heavy thrombotic burden. However, optimal timing of starting GPI and the impact of adding TA is unclear. We performed a prospective, randomized study to compare the upstream versus bail-out use of GPI with the selective combination of TA. METHODS: Total 73 consecutive patients with STEMI were randomly assigned to the upstream (n=37) versus bail-out use (n=36) of abciximab before undergoing coronary angiography. TA was performed using Thrombuster II (Kaneka Corp. Japan) on infarct related artery (IRA) diameter ���2.5 mm with thrombotic occlusion or evidence of thrombi. Primary endpoint was the flow restoration of the IRA using TIMI grade and ST-segment resolution. Secondary endpoint was the myocardial perfusion assessment using TMP grade. RESULTS: Baseline characteristics of both groups were similar. Bail-out GPI was associated with significantly worse initial TIMI flow compared with upstream use (TIMI grade 0; 70% vs. 43%, p<0.01), with no difference in ST-segment resolution (complete or partial resolution; 92% vs. 88%, p=0.50), final TIMI (grade 3; 88% vs. 95%, p=0.32) or TMP (grade 3; 61% vs. 65%, p=0.36). Eight patients from bail-out group (8/36, 22%) were crossed over to the late use of GPI due to significantly lower rate of achieving final TIMI grade 3, compared with upstream use and no use of GPI patients (63% vs. 95% vs. 96%, p<0.01). TA was added to the substantial portion of patients (55/73, 75%), showing the tendency of being used more frequently in bail-out group (83% vs. 68%, p=0.09) to overcome the worse initial TIMI flow. TA resulted in significant shortening of time to reach the peak CK-MB level after PCI (average 5.7 vs. 8.6 hrs, p<0.05), with similar peak CK-MB level, ST-segment resolution, final TIMI and TMP grade, compared to patients without TA. There were no significant differences in major adverse cardiac event rate between two groups at 1 and 9 months. CONCLUSIONS: Upstream GPI use before primary PCI for STEMI is associated with better initial TIMI flow (p<0.01) and shows the tendency to decrease the necessity of TA. TA in selected cases with thrombotic burden shortens the time to reach the peak CK-MB level (p<0.05), and appears to contribute to achieving comparable angiographic outcome with or without upstream use of GPI. However, GPI treatment is still recommended for bail-out strategy in case of thrombotic occlusion or poor TIMI flow in angiography.