OBJECTIVE: To evaluate access site and other bleeding complications associated with radial versus femoral access in patients receiving oral anticoagulation (OAC) with warfarin. BACKGROUND: Patients receiving OAC with warfarin undergoing coronary angiography and percutaneous coronary intervention (PCI) may have OAC continued despite the risk of bleeding. To what extent arterial access site impacts bleeding in such patients is not well studied. METHODS: Over 6 years, we identified 255 consecutive patients in whom warfarin was continued who underwent coronary angiography with an international normalized ratio >1.8. A total of 97 patients also underwent PCI at the same setting (27% femoral vs 73% radial). The primary outcome was Bleeding Academic Research Consortium bleeds; a secondary end point was frequency of access site complications in the 72 hours post-PCI. Complications were evaluated based on the initial access site attempted. RESULTS: Minimal baseline clinical characteristics differences existed between the groups. International normalized ratio was significantly higher in the radial group (2.42 ± 0.67 vs 2.24 ± 0.49, P = .02). Bivalirudin use was greater during radial PCI than femoral (76% vs 42%, P < .05), whereas unfractionated heparin alone was greater during femoral PCI than radial (46% vs 18%, P < .05). No significant difference was seen in the primary end point between femoral (2.8%) and radial (1.6%, P = .54) during coronary angiography alone. However, PCI via the femoral artery had significantly more Bleeding Academic Research Consortium bleeding (19.2% vs 1.4%, P = .005) and transfusions (15% vs 0%, P = .004) than via the radial artery. Patients who underwent PCI using radial access were less likely to have any vascular or bleeding complications (1% vs 23%, P = .001). CONCLUSIONS: Patients who underwent coronary angiography during uninterrupted OAC had similar bleeding rates regardless of access site. However, when PCI was performed, radial access was associated with fewer bleeding and vascular complications than the femoral approach. CONDENSED ABSTRACT: We retrospectively identified 255 consecutive patients on warfarin who underwent coronary angiography, 97 of whom underwent a percutaneous coronary intervention. The data reveal a reduction in Bleeding Academic Research Consortium bleeds (1.6% vs 8.1%, P = .02) with radial versus femoral access. The radial approach was associated with an overall lower rate of any vascular or bleeding complication than the femoral approach during percutaneous coronary intervention (1% vs 23%, P = .001).