Randomized trials of platelet glycoprotein (GP) IIb/IIIa receptor inhibitors in patients who have undergone percutaneous coronary intervention (PCI) have shown a modest increase in bleeding risk associated with GP IIb/IIIa inhibitor use. Because clinical trials often enroll highly selected patient populations and are performed in high-volume experienced centers, these results may not apply to a nonclinical trial population, thus altering the risk-benefit ratio of the drugs. Given the widespread use of these agents, we sought to determine bleeding risks in a broad-based population of patients who underwent PCI. We performed a retrospective cohort study of GP IIb/IIIa inhibitors and bleeding in 18,821 procedures from June 1, 1996 to December 31, 1998 using the Society for Cardiac Angiography and Interventions Registry. The primary outcome was bleeding events, defined as clinically significant hematoma formation or hemorrhage. Bleeding risk was 1.9% in the 2,525 patients who received GP IIb/IIIa inhibitors compared with 1.0% in the 16,296 who did not (unadjusted odds ratio [OR] 1.87, 95% confidence interval [CI] 1.35 to 2.59, p <0.001). After adjustment for multiple clinical and procedural variables, the effect was attenuated, with at most a twofold bleeding risk associated with GP IIb/IIIa inhibitor use (adjusted OR 1.39, 95% CI 0.96 to 2.03, p = 0.083). The small increase in absolute risk of bleeding from GP IIb/IIIa inhibitor use in this study is similar to the risk observed in clinical trials. Assuming these agents are as effective as shown in these trials, the risk-benefit ratio of GP IIb/IIIa inhibitors in broad-based PCI practice should be favorable.