Estimation of hemodialysis access blood flow rates by a urea method is a poor predictor of access outcome

Blood flow in peripheral arteriovenous fistulae and grafts as used for hemodialysis access can be derived from simultaneous measurements of 1) the amount of access recirculation (AR) induced by reversing the dialysis blood lines, and 2) the dialyzer blood flow rate (Qb). The hemodynamic monitor (HDM) uses magnetic principles to measure AR. The measurement is based on differential conductivity between arterial (A) and venous (V) blood flows in the dialysis blood tubing sets after the injection of hypertonic saline into the V line as a conductivity tracer. Access blood flow rates (Qa) derived from AR measurements by the HDM are predictive of access outcome. The measurement of AR is traditionally done from the comparison of urea levels simultaneously taken from the A and V blood lines and from the systemic circulation. Thus, the urea method can also be used to estimate access blood flow rates. The purpose of this study was to determine whether urea based Qa values are also predictive of outcome. Forty-one patients with arteriovenous fistulae (n = 25) or Gore-Tex grafts (n = 16) were studied by a standard protocol. The protocol involved temporarily reversing the A and V lines, taking three blood samples for urea estimation, performing an HDM recirculation test, and recording Qb as per the machine blood pump setting. The data allowed calculation of Qa by the HDM (Qa [HDM]) and urea (Qa [urea]) methods. Qa (HDM) was 1,177 +/- 887 ml/min (mean +/- standard deviation) and Qa (urea) 964 +/- 793 ml/min, a statistically significant difference (paired t-test p < 0.001). There was a significant linear correlation between the results (r = 0.94, p < 0.0001), but the regression equation also showed that Qa (urea) values were less than Qa (HDM). The influence of the Qa value on access outcome was determined after an 8 month follow-up. Nine of the 41 accesses were lost to clotting. Chi-square and discriminate analyses showed that Qa (HDM) significantly (p = 0.005) predicted access outcome, whereas Qa (urea) did not (p = 0.164). The specificity of a low Qa (HDM) in predicting access clotting was 0.78, compared with 0.62 for Qa (urea). The data show that although Qa can be estimated by the urea method, the finding of a low Qa (urea) is a poor predictor of access outcome and may lead to cost ineffective investigations.