Transradial approach is an established procedure in percutaneous coronary intervention (PCI). However, long term clinical outcomes of transradial PCI compared to transfemoral PCI have not been fully elucidated. Among 13087 patients undergoing first PCI in the CREDO-Kyoto registry Cohort-2 from January 2005 to December 2007, we identified 2736 patients with transradial approach and 4092 patients with transfemoral approach, excluding patients with acute myocardial infarction, patients on dialysis and patients treated with transbrachial approach. Using propensity score methodology, 2701 patients with transfemoral approach were randomly matched to 2701 patients with transradial approach based on clinical, angiographic, and procedural characteristics. The rates for procedural success of PCI were high in both transradial and transfemoral PCI (99 vs. 98%, P = 0.57). At 30 days, there was no significant difference in the incidence of all-cause death between the 2 groups (radial group 0.2% vs. femoral group 0.2%, P = 0.73). Incidence of bleeding event tended to be lower in the radial group than in the femoral group (1.0 vs. 1.6%, P = 0.09), and incidence of puncture site bleeding was significantly lower in the radial group than in the femoral group (0.2 vs. 0.6%, P = 0.005). Through 3-year follow up, cumulative incidence of all-cause death was not significantly different between the 2 groups (5.6 vs. 6.7%, hazard ratio 0.90 (95% confidence interval 0.71-1.13), P = 0.35]. In conclusion, transradial PCI reduced 30-day puncture site bleeding event compared to transfemoral PCI, with similar PCI success rates. In contrast, 3-year mortality rate was comparable between transradial and transfemoral PCI.