Background: Measurement of vascular access flow (Qa) has been proposed as the ideal method for surveillance of native fistulae. However, debate exists about the influence of blood pressure (mean arterial pressure [MAP]) on Qa during dialysis. Methods: During three consecutive dialysis treatments, 10 patients had paired measurements of Qa and MAP performed at 30, 60, 120, 180, 210, and 240 minutes. Access resistance (AR; in peripheral resistance units, PRUs) was calculated from MAP and Qa values. Results: Overall pooled coefficients of variation (CVs) for MAP, Qa, and AR were 8.4%, 12.3%, and 12.9%, respectively. A significant reduction in Qa and MAP occurred throughout the dialysis treatment (Qa, 104 mL/min; P = 0.008; MAP, 10.4 mm Hg; P = 0.007). Mean percentages of change in Qa for the first third compared with the middle and last thirds of the session were −4.6% ± 11.15% (SD) and −9.6% ± 10.5%, respectively. Thus, Qa varied between 11.4% and −30.6% from baseline during the last hour of dialysis treatments. A stronger correlation between MAP and Qa was seen in radiocephalic (r2 = 0.55; P < 0.0001) compared with brachiocephalic fistulae (r2 = 0.06; P = 0.023). Mean AR was unchanged during the dialysis session (0.23 PRU; P = 0.358). AR for radiocephalic fistulae was significantly greater compared with brachiocephalic fistulae (6.03 ± 3.90 versus 3.00 ± 1.11 PRU; P < 0.0001). Conclusion: Qa could decrease up to 30% from baseline, potentially impairing the ability of Qa to predict impending vascular access failure. AR remained stable during the treatment and may be a more useful measure of vascular access performance as part of an access surveillance program. Am J Kidney Dis 41:132-139. © 2003 by the National Kidney Foundation, Inc.