Abstract. Background: The aim of this study was to determine the outcomes of primary bypass graft surgery (BGS) compared to BGS after failed angioplasty (PTA). Patients and methods: Between January 2007 and January 2014, we performed 136 BGSs exclusively for the treatment of critical limb ischaemia. Two cohorts were identified: 1) primary BGS (n = 102; group I), and 2) BGS after prior PTA (n = 34; group II). Data were analysed retrospectively and the primary endpoints were the rates of secondary patency, amputation-free survival, freedom from major adverse outcomes (graft occlusion, amputation, or death), and overall survival, which were assessed with the Kaplan-Meier method. Results: Both groups were comparable with a predominance of Rutherford’s category 5 ischaemic lesions (73.3 %). Most patients had extensive TASC D athe-rosclerotic disease (83.6 %), and the main conduit was the greater saphenous vein (58.1 %). The mean follow-up time was 36.2 months. The 3-year secondary patency rates were better for group I (64.3 % vs 49.6 %; P = 0.04). During the same period, the amputation-free survival rates were similar between the groups (77.4 % vs 74.5 %; P = 0.59). For multivariate Cox regression analysis, BGS after prior PTA was the only factor associated with re-intervention for limb salvage (hazard ratio = 2.39; CI 95 % = 1.19 - 4.80; P = 0.02). At the 3-year point, there were no differences in the overall survival rates (72.6 % vs 70 %; P = 0.97), but the proportion of patients without adverse outcomes was higher in group I (37.3 % vs 13.4 %; P = 0.007). Conclusions: Although secondary patency was better after primary BGS, the amputation-free and overall survival rates support the use of BGS after prior PTA.