Purpose: Arteriovenous fistula (AVF) is the preferred treatment for long-term hemodialysis patients to allow reliable vascular access. Arteriovenous graft (AVG) is monitored using Doppler sonography to check a vessel's condition and predict complications such as steal syndrome. In this study, we developed an analysis algorithm and method to quantify steal syndrome using Doppler sonography.
Methods: Doppler sonography was used to determine the pattern of anterograde and retrograde flow. The ratio of blood volumes was calculated with a vision analysis software. First, performance of the developed algorithm was validated by comparing it with commercial Doppler sonography data. Doppler sonography was performed for an artificial vessel to analyze the steal flow.
Results: A total of 58 patients with steal flow were enrolled in this study. Of these patients, 23 did not have a difference in fingertip temperature between both sides. The median difference in temperature of 35 patients was 0.8°C (range, 0.3-1.9°C). The ratio of retrograde flow volume/antegrade flow volume in patients with the presence of temperature difference was significantly higher compared to that in patients without the temperature difference (p < .001). The ROC curve for the difference in flow volume had an AUC of 0.770. The optimal cutoff of difference in the flow volume between the two groups was 0.24 (sensitivity of 91.4 % and specificity of 52.2%). The flow volume difference was significantly positively correlated to temperature difference (r = 0.487, p < .003).
Conclusion: Our algorithm could measure steal flow volume of a bidirectional waveform by antegrade arterial flow and retrograde reversal flow.
Keywords: Hemodialysis; arteriovenous fistula; doppler sonography; steal flow; vision analysis.