The relationship between vascular access recirculation and access blood flow has been obscured by measurement techniques that overestimate recirculation. We measured brachial artery blood flow (a surrogate for access blood flow) using Doppler ultrasound and access recirculation using a two-needle slow/stop flow method and the standard peripheral vein method in 77 chronic hemodialysis patients. These patients had 25 native arteriovenous fistulae and 52 polytetrafluoroethylene grafts. Access recirculation by the slow/stop flow method was uniformly absent in these patients unless their access blood flow rate was less than the dialyzer blood flow rate. The peripheral vein method averaged 10.7% in these patients; many values were considerably higher. We conclude that access recirculation is a function of access blood flow and does not occur (in the absence of or reversed needle cannulation) unless access blood flow is markedly impaired.