Accuracy of ultrasound in the detection of inflow stenosis of arteriovenous fistulae: results of a prospective study.

While vascular ultrasound (US) has been highlighted to detect vascular access stenosis, its accuracy in the identification of inflow stenosis (IS) (anastomosis and/or juxta-anastomotic area) compared with the gold standard (angiography) has not been evaluated. One hundred three consecutive fistulae referred for interventions were included in this study. Preprocedure US of inflow segment was performed. Angiography from the feeding artery to the right atrium was then conducted. US comparison to angiography in the detection of IS (anastomosis and/or juxta-anastomotic area) was evaluated. Additionally, comparison of US to angiography in the assessment of juxta-anastomotic and anastomotic stenosis was reported separately. Data from 103 patients were available for analysis. Overall, US was negative for IS in 52 cases. Of these, 47 did not show a lesion on angiography. Only five cases demonstrated a stenosis on angiography. Fifty-one cases had IS by US, 50 were confirmed by angiography while one case did not show a lesion on angiography. Consequently, US had a sensitivity of 91%, specificity of 98%, and positive and negative predictive values were 98% and 90%, respectively. The sensitivity, specificity, negative, and positive predictive values for juxta-anastomotic and anastomotic lesions evaluated separately were 92%, 98%, 92%, 98% and 79%, 100%, 95%, 100%, respectively. Linear regression analysis showed a significant positive correlation between US and angiography for anastomotic (r2=0.71, p<0.0001; slope=0.63+/-0.098 and intercept=24+/-6) and juxta-anastomotic stenosis (r2=0.71, p<0.0001; slope=0.68+/-0.060 and intercept=23+/-4). These results reveal that US has a high degree of accuracy in the detection of IS.