OBJECTIVE This study was undertaken to compare repeat percutaneous transluminal angioplasty (rPTA), arterial reconstruction, and noninvasive therapy for treatment of patients with recurrent stenosis after PTA of the superficial femoral or popliteal artery. METHODS From 1983 to 1993, 93 patients were treated for recurrent femoropopliteal stenosis. Indication for treatment was claudication in 72 patients, rest pain in 9, and ischemic ulcer in 12. Thirty-six patients (38%) were treated with arterial bypass, 35 (38%) with rPTA, and 22 (24%) with exercise and medication. Patients were monitored with clinical examination, ankle-brachial indexes, and duplex scanning. Follow-up ranged from 6 to 110 months (mean 42 months). RESULTS With life-table analysis, the clinical and hemodynamic success of patients treated with rPTA was 41% at 1 year, 20% at 2 years, and 11% at 3 years. For patients treated with arterial bypass, the primary graft patency rate was 84%, 72%, and 72% at 1, 2, and 3 years, respectively. The secondary graft patency rate was 94%, 88%, and 88% at the same intervals. All patients with patent grafts were symptom free. All 22 patients treated with noninvasive therapy continued to have symptoms, but none required amputation during follow-up (range 6 to 108 months). Overall, patients with claudication did better than those treated for rest pain or an ischemic lesion after either rPTA or arterial bypass, but no other variable was statistically significant in predicting outcome. CONCLUSIONS This study finds that arterial bypass is safe and more effective than rPTA in treating patients with recurrent stenosis. Preoperative evaluation is unable to select patients likely to benefit from rPTA. Repeat PTA should be reserved for patients with limited life expectancy or contraindications to operation.