In patients with prior CABG requiring subsequent PCI there is uncertainty whether bypass grafts or native coronary arteries should be targeted. We analysed data from 2,764 patients with prior CABG in the Melbourne Interventional Group registry (2005-2018), divided into two groups: those undergoing PCI to a native vessel (n=1,928) and those with PCI to a graft vessel (n=836).Patients with a graft vessel PCI were older, had more high-risk clinical characteristics (prior MI, heart failure, ejection fraction <50%, renal impairment, peripheral and cerebrovascular disease), and high-risk procedural features (ACC/AHA types B2/C lesions). However, patients in the native vessel group were more likely to have PCI to chronic total occlusions. The majority of graft PCI were to saphenous vein grafts (84%), with 10% to radial and 6% to LIMA/RIMA grafts. Distal embolic protection devices were used in 30% of graft PCI. Patients with graft PCI had higher rates of no reflow (6.3 vs. 1.5%; p<0.001), coronary perforation (p=0.02), and inpatient stent thrombosis (p=0.03). However, 30-day mortality and major adverse cardiovascular and cerebrovascular events (MACCE) were similar. Unadjusted long-term mortality (median follow up 4.8 years) was higher in patients who had undergone a graft PCI (44 vs. 32%, p<0.001), but following Cox proportional hazards modelling, PCI vessel type was not a predictor of long-term mortality (HR 1.13; 95% CI 0.96-1.33, p=0.14). In conclusion, early clinical outcomes and risk-adjusted long-term mortality are similar for patients with prior CABG undergoing PCI to a native vessel or a bypass graft.