The effect of location and configuration on forearm and upper arm hemodialysis arteriovenous grafts Presented at the 2013 Vascular Annual Meeting for the Society of Vascular Surgery, San Francisco, Calif, May 30-June 1, 2013.

Objective The arteriovenous graft (AVG) is most often used in hemodialysis patients when an autogenous fistula is not feasible. The optimal location (forearm or upper arm) and configuration (loop or straight) of AVGs are not known. To evaluate relationships of AVG location and configuration with patency, we conducted a secondary analysis using data from a randomized, placebo-controlled trial of dipyridamole plus aspirin for newly placed AVG. Methods Participants of the Dialysis Access Consortium (DAC) Graft Study with newly placed upper extremity prosthetic grafts involving the brachial artery were studied. Multivariable analyses adjusting for trial treatment group, center, gender, race, body mass index, diabetes, current treatment with chronic dialysis, and prior arteriovenous vascular access or central venous catheter were performed to compare outcomes of forearm (fAVG) and upper arm (uAVG) grafts, including loss of primary unassisted patency (LPUP) and cumulative primary graft failure (CGF). Subgroup analyses of graft configuration and outflow vein used were also conducted. Results A total of 508 of the 649 participants (78%) enrolled in the trial had an upper extremity brachial artery graft placed, 255 with fAVG and 253 with uAVG. Participants with fAVG were less often male (33% vs 43%; P =.03), African American (62% vs 78%; P <.001), and receiving dialysis at the time of surgery (62% vs 80%; P <.001). Participants with fAVG had a higher mean body mass index (33 vs 29; P <.001). The LPUP (fAVG 70% vs uAVG 78%; P =.07) and CGF (33% vs 36%; P =.91) were similar between fAVG and uAVG at 1-year follow-up. In multivariable analysis, AVG location (uAVG vs fAVG) was not associated with LPUP (hazard ratio, 1.21; 95% confidence interval, 0.90-1.63; P =.20) or CGF (hazard ratio, 1.36; 95% confidence interval, 0.94-1.97; P =.10). LPUP did not differ significantly between fAVG and uAVG among subgroups based on AVG configuration (P = 1.00) or outflow vein used (P =.16). Conclusions Patency was comparable between fAVG and uAVG despite the larger caliber veins often encountered in the upper arm in carefully selected patients. Our findings support the traditional view that, in order to preserve a maximal number of access sites, the forearm location should be considered first before resorting to an upper arm graft.