This retrospective study compares outcomes of transposed brachiobasilic fistulae (avf), nontransposed avf, and grafts (avg). There were 161 (21%) transposed avf, 321 (42%) nontransposed avf, and 285 (37%) avg placed and followed over a 3-year period. Of the nontransposed avf, 203 (63%) were forearm radiocephalic avf and 118 (37%) were upper arm brachiocephalic avf. Grafts were able to be used in shorter time than avf and had a lower primary failure rate. However, avg had a significantly higher infection rate (0.23/patient/year vs. 0.07/patient/year) and required more procedures over the life of the access than avf (2.90/patient/year vs. 0.51/patient/year). Despite a higher primary failure rate, avf had better overall patency. Among nontransposed avf, brachiocephalic avf had significantly better primary patency than radiocephalic avf (p < 0.0001). Transposed avf had better cumulative patency than either nontransposed avf or grafts (p < 0.0001). Cumulative patency was best for transposed avf, intermediate for brachiocephalic avf, and worst with radiocephalic avf and avg. There was no difference in cumulative patency between radiocephalic avf and grafts (p = 0.5601). This analysis supports the current recommendation to prefer transposed avf over avg. In addition, the data suggest that both nontransposed brachiocephalic avf and transposed brachiobasilic avf outcomes may be superior to radiocephalic fistulae.

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.