The increased complexity of redo infrainguinal bypass procedures can result in prolonged operative time and increased morbidity. This review was undertaken to compare outcomes from primary and redo bypass procedures and to identify factors predictive of graft failure and limb loss after redo bypass. All infrainguinal bypasses ( n = 468) from 1995 to 1999 were reviewed. A total of 367 primary bypasses in 317 patients were compared to 101 redo grafts in 84 patients with previously failed bypasses. Risk factors and types of procedures were compared using Student's t-test and the chi(2) test. Patency and limb salvage were compared using life-table analysis. Patients requiring redo bypasses were less likely to have diabetes and end-stage renal disease. Two-year patency (66 +/- 4% primary vs. 55 +/- 7% redo, p = 0.13) and limb salvage (75 +/- 3% primary vs. 72 +/- 6% secondary, p = 0.43) were comparable between primary and redo bypass groups. Female gender was predictive of redo graft failure (2-year patency 73 +/- 8% male vs. 39 +/- 9% female, p = 0.01). Clinical indications that predicted failure of a redo bypass included thrombosis of an autologous graft (1-year patency 71 +/- 7% previous prosthetic vs. 49 +/- 10% previous autologous, p = 0.004), thrombosis of an infrageniculate bypass (2-year patency 65 +/- 10% suprageniculate vs. 46 +/- 9% infrageniculate, p = 0.044), and a limb salvage indication for the primary operation (2-year patency 86 +/- 9% claudication vs. 44 +/- 8% limb salvage, p = 0.008). When a primary bypass fails despite the use of optimal conduit (autologous vein) and an infrageniculate target vessel, the redo bypass has a higher risk of failure, particularly in female patients. Nonetheless, patency and limb salvage rates justify an attempt at revascularization after failed primary bypass.