Distribution and severity of stenoses in functioning arteriovenous grafts: A duplex and angiographic study

Expanded polytetrafluoroethylene (ePTFE) bridge grafts for dialysis access are the most frequently implanted vascular grafts. There remains controversy regarding the distribution and frequency of stenoses that occur within functioning grafts. To resolve this issue, we studied all patients with a functioning graft in a single dialysis center using color duplex imaging (CDI). Angiography (arterial inflow, entire graft, and venous outflow) was performed in all patients with stenoses of greater than or equal to 50% by CDI. Of 150 patients within the dialysis center, 110 had ePTFE bridge grafts. Greater than 50% stenoses were identified in 41 patients. Of the 41 patients, 54 stenoses were identified by CDI. Angiography demonstrated a total of 64 stenoses in the same patient cohort. Per CDI, 29 (54%) of the lesions were located at the venous anastomosis, 31 (48%) at the venous anastomosis per angiography. The mean percent stenosis of these lesions was 60% per CDI and 67% per angiography. Two lesions were located at the arterial anastomoses by both modalities. Nineteen (35%) of the stenoses were midgraft per CDI and 17 (27%) per angiography; the mean percent stenosis of these lesions was 57% and 60%, respectively. Angiography demonstrated unsuspected, remote, venous outflow lesions in 10 (16%) cases, with an average stenosis of 70%. Technically unsatisfactory studies, using CDI, occurred in three (5%) patients. Sensitivity of CDI was 86%. There was excellent concordance of stenoses identified within the arterial and venous anastomoses. There was a high degree of concordance between the two imaging modalities regarding the severity of the stenoses. Our study has demonstrated that 37% of patients in a single dialysis center with functioning ePTFE arteriovenous grafts have graft stenoses of greater than 50%. The average number of stenoses per graft was 1.38 per CDI and 1.61 per angiography. CDI effectively detected stenoses within the graft, its arterial and venous anastomoses, but was ineffective in detecting central venous lesions.