PURPOSE To compare the SMART (shape memory alloy recoverable technology) stent with percutaneous transluminal angioplasty (PTA) alone in hemodialysis access venous stenoses. MATERIALS AND METHODS A prospective, nonrandomized study was undertaken in 60 patients with dysfunctional polytetrafluoroethylene dialysis grafts. Indications for stent placement were acute PTA failure, rapid restenosis, and vessel perforation. The primary endpoint was improved graft patency in patients treated with stents compared with that in patients whose disease responded to PTA alone. The secondary endpoints were lower postprocedural midgraft pressures and similar complication rates compared with PTA alone. RESULTS The key venous stenosis was at the graft-to-vein anastomosis in all but two patients. Thirty-five patients showed a response to PTA alone. Sixteen patients received stents for stenoses greater than 30% after angioplasty, six for rapidly recurrent stenosis, and three for venous rupture. Nine patients received stents across the level of the elbow joint. Stenosis after intervention was significantly less frequent in the stent group (7% vs 16%; P = .001), but the midgraft systolic pressure ratios did not significantly differ. The clinical success rates were 100% after stent implantation and 97% after PTA alone. Except for venous rupture, there were no procedure-related complications, and, excluding early graft thrombosis, there were no complications at 30 days. A single stent fracture was found on follow-up. The mean primary graft patency times were 5.6 months after PTA and 8.2 months after stent treatment (P = .050). When stents were placed across the level of the elbow joint, the mean primary graft patency time was 8.9 months. CONCLUSION Implantation of the SMART stent is safe and effective for the treatment of residual or rapidly recurrent dialysis access venous stenoses and is associated with better patency than PTA alone.