A comparison of intra-session and inter-session variation in hemodialysis access flow - Commentary

BACKGROUND Monitoring of blood flow through hemodialysis shunts, using ultrasound dilution or duplex ultrasound, can provide early warning of a thrombotic blockage. Few studies have investigated the reproducibility of shunt flow measurements between hemodialysis sessions. OBJECTIVES To assess whether shunt flow measurements by ultrasound dilution vary more between hemodialysis sessions than within sessions, and to compare the within-session variability of flow measurements obtained with ultrasound dilution and duplex ultrasound. DESIGN AND INTERVENTION All hemodialysis patients who underwent shunt flow monitoring at the Dialyse Centrum Groningen or the Groningen University Medical Center, The Netherlands, between October 2003 and December 2003 were eligible to enter this study. If patients were receiving dialysis through a single needle, they were excluded. Shunt flow was measured three times before dialysis (using duplex ultrasound), and a further three times during the first hour of dialysis (using ultrasound dilution). After a day free of dialysis, patients underwent a second session, during which measurement of shunt flow by ultrasound dilution was repeated. Duplex ultrasound measurements were carried out at a location between the sites of arterial and venous puncture, where the vessel cross section was as circular as possible. OUTCOME MEASURE The endpoint was hemodialysis shunt flow rate. RESULTS Mean age among the 24 patients studied (16 male; 8 female) was 59 boolean AND years. More patients had brachiocephalic fistulas (n=11) than radiocephalic fistulas (n=10) or grafts (n=3). Shunt flow assessed by ultrasound dilution averaged 1,138 ml/min in the first dialysis session, and did not differ significantly in the second session (mean 1,242 ml/min). The mean variation between sessions was 16.4%. The pooled variation coefficient for ultrasound dilution within sessions was lower than the between-session value-although not significantly (7.7% vs 14.2%; P=0.25)-and similar to that of the duplex measurements (11.6%). Adjustment for blood pressure did not significantly alter the pooled between-session variation coefficient for ultrasound dilution measurements. There was a trend toward greater between-session variation in patients whose needles were placed in different directions during the two dialysis sessions (n = 11), than in patients whose needles were oriented similarly during both sessions (n = 10; 21.3% vs 12.7%; P=0.08). This difference became significant when values were adjusted for blood pressure (26.0% vs 10.1 %; P<0.01). In patients whose needles were oriented similarly in both sessions, the 95% CI for the between-session variation in ultrasound dilution shunt flow measurements (adjusted for blood pressure) was 1.8-18.5%. When patients were analyzed separately according to fistula type (radiocephalic vs brachiocephalic), only the radiocephalic group showed significantly higher between-session variation coefficients when needles were oriented differently than when needle placement was consistent (23.5% vs 6.7%; P=0.02). CONCLUSION Changes of 20-25% in ultrasound dilution shunt flow measurements between hemodialysis sessions can be considered clinically meaningful.