Flow monitoring: performance characteristics of ultrasound dilution versus color Doppler ultrasound compared with fistulography

Abstract

Background: Measurement of access blood flow is the preferred noninvasive screening test for hemodialysis arteriovenous (AV) fistula stenosis. However, performance characteristics of the 2 most frequently used ultrasound techniques compared with fistulography remain elusive.

Methods: We evaluated 59 hemodialysis patients with native forearm AV fistulae who underwent all 3 measurements in a prospective order: the ultrasound dilution technique (UDT), color Doppler ultrasonography (CDUS), and fistulography. Patients with angiographically diagnosed access stenosis underwent angioplasty and were followed up by means of monthly UDT measurements until restenosis occurred within the first 6 months.

Results: Both ultrasound techniques predicted access stenosis (P < 0.01). Performance was similar between both techniques, evaluated by receiver operating characteristic curves. Areas under the curve averaged 0.79 (95% confidence interval [CI], 0.66 to 0.91) for UDT and 0.80 (95% CI, 0.65 to 0.94) for CDUS. Correlation between measured UDT and CDUS blood flow rates was 0.37 (Spearman's rho, rho = 0.004). The calculated optimal cutoff value for the prediction of stenosis was 465 mL/min for the UDT and 390 mL/min for the CDUS technique. Access stenosis was diagnosed in 41 patients who subsequently underwent percutaneous angioplasty (PTA), which was successful in 34 patients. Restenosis occurred in 13 patients within the first 6 months after PTA. UDT access blood flow after PTA was significantly lower in these 13 patients compared with the other 21 patients.

Conclusion: Our data suggest that blood flow monitoring of AV hemodialysis access by ultrasound provides a reasonable prediction of access stenosis and restenosis.