Background:Radial access for PCI is increasing but its use in emergency still remains confidential. We report our single center experience where all interventional cardiologists use the radial access as default strategy for primary PCI. Methods and Results: STEMI patients (n = 671) were evaluated for bleeding complications using a web-based registry (e-PARIS). In-hospital bleeding was adjudicated using various definitions (TIMI, GUSTO, STEEPLE). MACE was the composite of death, MI and stroke. In this non-selected, high risk population, 6.1% had cardiogenic shock on admission, 3.9% out-of-hospital cardiac arrest and 51.2% multivessel disease. Radial access (88%) was the default strategy as was abciximab (78%). Clopidogrel loading dose ranged from 300 to 900 mg. Pre-hospital fibrinolysis was used in 7.1%. Hemodynamic support devices (IABP, ECMO, Tandem Heart) were needed in 7.0%. In-hospital bleeding rates varied widely according to the definitions used: 2.5%, 1.5%, 5.7%, 9.2% and 10.9% with TIMI Major, GUSTO Severe, TIMI Major/minor, GUSTO Severe/moderate or STEEPLE Major, respectively. In-hospital death rate was 5.5%. One-year mortality (8.2%) was seriously impacted by in-hospital bleeding (31.6% vs 6.8%, p < 0.001). The most frequent bleeding site was gastro-intestinal (figure). GUSTO Severe/moderate bleeding was independently correlated with MACE (OR 2.60; 95%CI 1.21−5.59; p = 0.01) and radial access was a strong predictor of survival (OR 0.33; p = 0.002). Conclusions: The gastro-intestinal tract is the most frequent site of bleeding when the radial artery is the predominant vascular access site for primary PCI. GUSTO Severe/moderate bleeding is an independent predictor of MACE while radial access predicts survival.