It is widely accepted that surveillance with correction of stenosis before thrombosis prolongs hemodialysis graft survival. However, few prospective randomized controlled trials have confirmed the efficacy of this approach. Most studies have been nonrandomized or have included historic rather than concurrent controls. Although graft dysfunction as assessed by blood flow determination has been widely recognized as the surveillance method of choice, all fully randomized prospective studies that have found a benefit in terms of decreased thrombosis or increased survival have evaluated stenosis and the benefit has only been in new grafts. Finally, the accuracy of blood flow measurements as determined by ultrasound dilution in predicting thrombosis is controversial. Some have reported that predictive accuracy is adequate, whereas others have reached the opposite conclusion. Thus, further studies are needed to explore the effectiveness of surveillance with intervention in improving graft survival. Until such studies establish the use of surveillance programs, surveillance of all grafts using blood-flow monitoring cannot be supported.