Access flow (QACC) is a major determinant of patency. Access recirculation (AR > 2%), normalized venous intra-access pressure (vPIA/MAP), and QACC are used to detect access dysfunction. We compared these three measures of access function (ultrasound dilution to measure AR and QACC). A total of 779 measurements were performed on 58 arteriovenous fistulas (AVFs) and 114 polytetrafluoroethylene (PTFE) grafts (1-8/access) over 13 months, and the access parameters at the beginning of each period were related to access events within that period. Pump blood flow averaged > 420 ml/min. AR occurred uncommonly (3.8%), and in half the cases, resulted from technical error by staff. In accesses that thrombosed or underwent intervention for stenosis, AR was present in only 3 of 11 AVFs and 8 of 57 PTFE accesses. When AR was present in grafts, QACC averaged 270 +/- 23, and access thrombosis followed unless intervention occurred. In grafts, vPIA/MAP averaged 0.34 +/- 0.01 in those remaining patent, 0.52 +/- 0.08 in those that had undergone intervention, and 0.54 +/- 0.04 in those that had thrombosed. QACC averaged 1,121 +/- 26, 605 +/- 45, and 550 +/- 65 ml/min, respectively, in the three groups. By contrast, QACC differed significantly in patent AVFs (1,053 +/- 35) compared with failing AVFs (363 +/- 48), but vPIA/MAP did not. AR is thus a late manifestation of access failure. QACC is the best diagnostic test of access dysfunction in AVFs. Interpretation of vPIA/MAP in grafts is enhanced by periodic QACC measurements.