As existing arteriovenous grafts (AVGs) fail, the National Kidney Foundation KDOQI Guidelines and the AV Fistula First Breakthrough Initiative ("Fistula First") project recommend that each patient be re-evaluated for conversion to an arteriovenous fistula (AVF). AVFs created following failure of an AVG have been termed secondary fistulas (SAVF). We review our experience and outcomes converting AVGs to SAVFs, utilizing the mature outflow vein of the AVG when possible, otherwise creating a new AVF at a remote site. We reviewed two groups of consecutive patients undergoing operations for vascular access at different centers. Group 1 had a SAVF protocol in place during the study period with specific criteria for timing SAVF construction. Patients from group 2 were referred for evaluation by nephrologists or dialysis nurses as access problems were recognized, without a formal protocol in place. All patients had preoperative ultrasound or contrast imaging in addition to physical examination. Indications for creating a SAVF were AVG thrombosis, dysfunction, erosion, bleeding, or steal syndrome involving the existing AVG. The simple presence of a functional AVG without evidence of dysfunction was not an indication for conversion to a SAVF. SAVFs were classified according to location and the potential for utilizing the existing mature AVG outflow vein. Group 1: 40 consecutive patients, age 26-78 (mean = 62), 42% were female; 55% were diabetic. These patients had 1-22 previous access operations (mean = 3). 92.5% underwent SAVF surgery prior to loss of the AVG, minimizing catheter use. Cumulative patency was 92.5% at 1 year and 87.5% at 2 years. Group 2: 102 consecutive patients, age 24-87 (mean = 55), 52% were female; 50% were diabetic. These patients had 1-50 previous access operations (mean = 3). Only 19.3% were referred for SAVF surgery prior to loss of the AVG or outflow vein. Cumulative patency was 94.4% at 1 year and 91.6% at 2 years. Failure, dysfunction, or complications of AVGs may be resolved by conversion to a SAVF. Further, the limited lifespan of AVGs and the superiority of AVFs dictates that a plan be in place to transition the AVG patient to an AVF. Most, if not all, hemodialysis patients whose access is an AVG will have one or more anatomic sites and vessels suitable for an autogenous SAVF. Vessel mapping is critical in the evaluation of failing AVGs and in preparation for a SAVF. Cumulative patency rates exceeded 90% at 12 months for SAVFs in both patient groups in this report. The need for catheters was dramatically less in the patient group with an established SAVF conversion plan.