Back to Annual Meeting Program

OBJECTIVE: This study aims to elucidate factors contributing to primary and cumulative patency of hemodialysis arteriovenous grafts (AVG). Currently, many NKF-KDOQI guidelines for AVG placement are based on clinical opinion. We seek to provide a resource for evidence-based algorithms.
METHODS: Retrospective cohort study including all upper extremity AVG using Boston Scientific or Gore® materials based on axillary or brachial arteries between 03/31/2007 and 01/01/2012 at Mt. Sinai Hospital (271 procedures). Cox proportional hazards regression and Kaplan-Meier curves were used to model primary and cumulative patency. For grafts that remained patent at least one year, ANOVA testing was used to determine differences between graft types in the number of reinterventions performed in the first year of use.
RESULTS: Sixty forearm (F-) and 211 upper arm (U-) grafts were placed during the study period. Three patient characteristics (age, sex, diabetes) and four graft properties (manufacturer, location, tapering, configuration) were considered for their impact on graft survival. No variable had significant effect on primary patency but location and tapering were found to be the most important determinants of cumulative patency. U-grafts had greater cumulative patency (1,324.6±63.9days) as compared to F-grafts (1,152.8±121.1days) [hazard ratio (HR): 0.548,p=0.015]. Straight (S-) grafts had less cumulative patency (1,257.5±63.7days) compared to tapered (T-) grafts (1,430.4±122.5days) [HR: 1.827,p=0.055]. Location and tapering were subsequently used to divide grafts into four groups (U-S, U-T, F-S, and F-T). Kaplan-Meier analysis revealed significant differences between groups in cumulative patency (p=0.031). F-S faired most poorly (990.2±151.3days). Analysis of variance indicated group as a significant factor in pseudoaneurysm reintervention in the first year (F=2.85,p=0.039). Post-hoc Tukey HSD analysis revealed F-T had significantly more pseudoaneurysm repairs than U-S (p=0.041). Otherwise, there were no significant differences between groups in prevalence of thrombectomies, revisions, stentings, or angioplasties over a one-year period. Furthermore, it was observed that subsequent to F-graft failure, U-fistulas were only placed in 2 of the 26 patients (7.7%), both of which were on the contralateral side.
CONCLUSIONS: Location in the upper arm appears to be the most significant factor contributing to cumulative patency of dialysis access grafts. Given this fact and the observed lack of U-fistula placement after F-graft failure, further detailed study of AVG outcomes is necessary to better inform the dialogue on specific decisions in dialysis access placement.