We compared the effect of bivalirudin or heparin and use or nonuse of glycoprotein IIb/IIIa inhibitors (GPI) on the outcome of left main coronary artery (LMCA) percutaneous coronary intervention (PCI) in the randomized EXCEL trial. The optimal antithrombotic regimen to support PCI of the LMCA remains controversial because of low representation of this subset in clinical trials. The PCI cohort (n = 928) in EXCEL was divided according to bivalirudin versus heparin antithrombin treatment and compared for the primary composite endpoint of death, myocardial infarction (MI), or stroke at 30 days and 5 years. Bivalirudin was used in 319 patients (34.4%). The composite endpoint at 30 days occurred in 7.2% versus 3.8% bivalirudin and heparin patients, respectively, p = .02; at 5 years, the composite endpoint occurred in 26.3% versus 19.9% bivalirudin and heparin patients, respectively, p = .02. Major bleeding was more frequent in bivalirudin patients (4.1% versus 1.3%, p = .008). There were no differences in stent thrombosis between the groups. Bivalirudin use was an independent predictor of the 30-day composite endpoint (OR 2.88, 95% CI 1.28-6.48, p = .01) but not of the 5-year composite endpoint (OR 1.30, 95% CI 0.84-2.02, p = .23). GPI use was infrequent (n = 67, 7.2%) and was not associated with adverse outcomes. Among patients undergoing LMCA PCI in the EXCEL trial, procedural use of bivalirudin was associated with greater rates of periprocedural MI and the 30-day composite endpoint without reducing bleeding complications. Five-year outcomes were similar. GPIs were used infrequently and were not associated with clinical outcomes.