A comparison of methods for the measurement of hemodialysis access recirculation and access blood flow rate

The ability to accurately measure access recirculation (AR) is of importance because its presence indicates access dysfunction and may explain why a prescribed Kt/V (urea) has not been delivered. The ability to measure access flow (Qa) allows access monitoring and the detection of impending access dysfunction. AR can be measured by indicator dilution or conductivity tracer techniques. Qa calculation is simple if AR can be detected. The previous techniques are used while the patient's blood lines are reversed to induce AR, and the Krivitski equation gives: Qa = Qb [1-r/r] where Qb = dialyzer blood flow and r = proportion of AR induced. Three methods for AR and Qa measurements were directly compared: 1) ultrasound dilution (Transonics Hemodialysis Monitor, Transonics Systems Incorporated) (TRANS); 2) hematocrit dilution (Crit-Line Monitor, In-Line Diagnostics) (CRIT); and 3) differential conductivity (Hemodynamic Monitor, GAMBRO Healthcare Incorporated) (HDM). Patients were cannulated in a standard fashion and dialysis commenced with lines in normal configuration. A HDM test was performed and, if AR = zero, the lines were reversed to induce AR. HDM, TRANS, and CRIT tests for AR were next done in rapid succession for direct comparison. Each test was repeated three times in succession, the device in random order, to assess test repeatability. Qb was taken from the 1) dialysis machine pump, and 2) directly from TRANS and Qa calculated, using 1) and 2) AR results. In comparison to TRANS, AR results were virtually identical for HDM (TRANS AR = 1.04 HDM-AR + 0.02, r = 0.98, p = 0.0000), and good for CRIT (CRIT-AR = 0.84 TRANS-AR - 0.2, r = 0.81, p = 0.001), but CRIT underestimated the values. Repeatability was assessed by normalizing (%) the SD of repeated measurements; values were 7.5% (HDM), 9.1 % (TRANS), and 17.4% (CRIT). Qa value comparisons were similar (minimal r = 0.83) regardless of Qb source, but CRIT overestimated the value; repeatability data showed 10.6% (HDM), 13.0% (TRANS), and 25.2% (CRIT) (n ranged from 15-64). In summary, TRANS and HDM appear equal as far as accuracy and repeatability of measurements; CRIT results correlated well, but tended to underestimate AR and overestimate Qa, and was less reproducible.