Prospective assessment of hemodialysis access patency after percutaneous intervention: Cox proportional hazards analysis.

Vascular access failure is the greatest limitation of successful hemodialysis, but the factors associated with long-term patency have not been fully elucidated. Outcomes in a consecutive series of 294 thrombosed or failing accesses [128 fistulas (43.5%) and 166 grafts (56.5%) in 179 patients] were analyzed with life table and multivariable Cox proportional hazards analysis. Initial success was achieved in 275 of 294 accesses (95.6%). The median patency after intervention was 206 days (interquartile range, 79-457 days). Fistulas had longer median patency after intervention than grafts (286 vs. 170 days). Nonthrombosed accesses had longer median patency than thrombosed accesses (238 vs. 136 days), but thrombosed fistulas had similar median patency as thrombosed grafts (140 vs. 136 days). The selective use of stents as a bailout for failed balloon dilatation did not significantly reduce long-term patency (196 days for stented accesses vs. 210 days for unstented accesses). Long-term patency was inversely related to final access pressure, but access patency was not related to the presence of central venous occlusions, graft age, patient age, sex, or diabetes. Catheter-based intervention of thrombosed and failing dialysis accesses significantly prolongs patency and usefulness of dialysis accesses. The expanding use of fistulas, improved detection of early access failure, and selective use of bailout stents should enhance long-term access patency. Copyright 2005 Wiley-Liss, Inc.