Hemodialysis vascular access flow measurements by AFM^® and Transonic^® devices: A bicentric comparative study

Clinical monitoring remains central to vascular access care, while surveillance of vascular access blood flow (Qa) provides information for detecting dysfunction. Ultrasound dilution (UD), such as Transonic^®, is widely used for Qa measurement but may be limited by workflow, personnel, and time constraints. Alternatively, conductivity-based approaches integrated into dialysis machines enable automated intradialytic estimation of Qa without interrupting treatment. This bicentric study evaluated agreement between an automated ionic dialysance-based method, AFM^®, and UD, and assessed its clinical performance across prespecified Qa ranges.

We conducted an observational method-comparison study in two Italian hemodialysis centers. Paired Qa measurements were obtained in 64 patients during routine hemodialysis sessions using AFM^® and Transonic^®. Agreement and association were assessed using Spearman correlation, Passing-Bablok regression, Bland-Altman analysis, Lin's concordance correlation coefficient, and weighted Cohen's kappa. Clinical performance was evaluated across prespecified Qa ranges: ⩽600, >600 to <2000, and ⩾2000 mL/min.

Median Qa was 1145 (IQR, 807-1536) mL/min with AFM and 1436 (IQR, 947-1980) mL/min with UD. The two methods showed strong correlation (ρ = 0.849, p < 0.001) and moderate-to-good concordance (CCC = 0.73). Bland-Altman analysis revealed a systematic underestimation at higher flows, with magnitude-dependent variability. Overall categorical agreement was 84% (weighted κ = 0.72). Performance was highest in the clinically relevant ranges of ⩽600 and >600 to <2000 mL/min, with 100% and 93% sensitivities, and F1-scores of 0.89 and 0.88, respectively. Agreement declined in the high-flow range (⩾2000 mL/min), where sensitivity dropped to 53%.

AFM^® showed strong correlation and substantial clinical agreement with UD, particularly in the low- and intermediate-flow ranges most relevant to routine vascular access surveillance. Its automated intradialytic implementation supports its use as a practical screening tool for Qa surveillance. However, high-flow measurements should be interpreted cautiously and further evaluated when clinically indicated.