OBJECTIVE The performance of a graft created from the upper arm basilic and cephalic veins in continuity was investigated. METHODS Retrospective analysis of 50 patients, who underwent 54 distal reconstructions with an upper arm vein loop graft between February 1989 and October 1993 (male-to-female ratio of 30/20; mean age of 69.2 years, range 39 to 87; 74% had diabetes) was undertaken. Vein grafts were harvested through a near continuous incision, leaving a skin bridge in the cubita. Intraoperative angioscopy was used to exclude endoluminal disease and to directly observe valvulotomy of the nonreversed part of the graft. RESULTS Operations were performed for limb salvage in 98.2% of 17 primary and 37 reoperative procedures. Eleven femoropopliteal, 33 femorotibial-pedal, seven popliteal-distal, and two outflow jump grafts were performed. The ipsilateral saphenous vein was unavailable because of previous infrainguinal bypass in 35, coronary artery bypass grafting in 14, and unsuitable quality in 5 cases. Thirty-eight grafts were used in continuity, and 16 grafts required repair or splicing with additional vein segments. Primary 30-day patency rate was 92.6% (n = 4 occlusions). No operative deaths occurred. The cumulative patency rate at 1 year was 74.4%, the limb salvage rate 90.7%. CONCLUSIONS The upper arm vein loop is a durable graft with excellent short-term and midterm patency rates. Sufficient vein length can be obtained to reach the below-knee and midtibial levels. Angioscopic quality assessment is a valuable adjunct to exclude endoluminal disease most commonly occurring in the median cubital vein. Straightening the curve of the median cubital vein and valvulotomy do not influence patency rates. This is a valuable technique for vascular surgeons that enables rescue of ischemic limbs under otherwise difficult circumstances.