Vascular access stenosis: comparison of arteriovenous grafts and fistulas.

Background: Vascular access stenosis is a frequent problem in hemodialysis patients. There is little published literature comparing the features of stenosis between arteriovenous fistulas and grafts, relative outcomes of elective angioplasty, and clinical factors predictive of access patency after angioplasty.

Methods: Prospective data were collected for all patients referred for a fistulogram during a 2-year period because of suspected access stenosis. Angioplasty was performed if there was greater than 50% stenosis. For each procedure, we recorded the number and location of stenotic lesions, degree of stenosis (on a scale of 1 to 4), and ratio of access to systemic systolic pressure. All subsequent access procedures were tracked prospectively to calculate intervention-free access survival. Multivariable analysis was used to evaluate clinical factors affecting access patency after angioplasty.

Results: Five hundred forty-three fistulograms were obtained: 358 in grafts and 185 in fistulas. The likelihood of finding a significant stenosis was substantially lower in fistulas than grafts (39.4% versus 68.7%; P < 0.001). Among patients with a significant stenosis, those with fistulas were less likely to have 2 or more stenotic lesions (12.5% versus 33.1%; P < 0.001). After angioplasty, degree of stenosis (1.35 +/- 0.70 versus 1.23 +/- 0.52) and access to systemic pressure ratio (0.34 +/- 0.15 versus 0.32 +/- 0.14) were similar between fistulas and grafts. Intervention-free survival was similar for fistulas and grafts (median survival, 7.5 versus 6.2 months; P = 0.36). Using multivariable stepwise proportional hazard regression analysis, only female sex, residual access stenosis, and postangioplasty access pressure ratio greater than 0.4 significantly predicted access survival (P = 0.0006).

Conclusion: The positive predictive value of clinical evaluation for access stenosis is substantially lower for fistulas than grafts. The technical success of angioplasty and subsequent primary patency are similar for fistulas and grafts. Finally, female sex, residual stenosis, and high postprocedure access pressure ratio are each predictive of shorter access patency after elective angioplasty.