© 2015 by the American Society of Nephrology. Arteriovenous grafts (AVGs) are prone to frequent thrombosis that is superimposed on underlying hemodynamically significant stenosis, most commonly at the graft-vein anastomosis. There has been great interest in detecting AVG stenosis in a timely fashion and performing preemptive angioplasty, in the belief that this will prevent AVG thrombosis. Three surveillance methods (static dialysis venous pressure, flow monitoring, and duplex ultrasound) can detect AVG stenosis. Whereas observational studies have reported that surveillance with preemptive angioplasty substantially reduces AVG thrombosis, randomized clinical trials have failed to confirm such a benefit. There is a high frequency of early AVG restenosis after angioplasty caused by aggressive neointimal hyperplasia resulting from vascular injury. Stent grafts prevent AVG restenosis better than balloon angioplasty, but they do not prevent AVG thrombosis. Several pharmacologic interventions to prevent AVG failure have been evaluated in randomized clinical trials. Anticoagulation or aspirin plus clopidogrel do not prevent AVG thrombosis, but increase hemorrhagic events. Treatment of hyperhomocysteinemia does not prevent AVG thrombosis. Dipyridamole plus aspirin modestly decreases AVG stenosis or thrombosis. Fish oil substantially decreases the frequency of AVG stenosis and thrombosis. In patients who have exhausted all options for vascular access in the upper extremities, thigh AVGs are a superior option to tunneled internal jugular vein central vein catheters (CVCs). An immediate-use AVG is a reasonable option in patients with recurrent CVC dysfunction or infection. Tunneled femoral CVCs have much worse survival than internal jugular CVCs.