Outcomes of surgical revision of stenosed and thrombosed forearm arteriovenous fistulae for haemodialysis.

BACKGROUND: Surgery is an established treatment for stenosed and thrombosed forearm arteriovenous fistulae (AVFs), but the literature on its outcome is limited. We report our experience of the surgical repair of stenosis in patent and thrombosed forearm AVFs and evaluate the outcome of two procedures, proximal neo-anastomosis (NEO) vs replacement of the stenosed segment with a polytetrafluoroethylene graft interposition (GI). METHODS: Sixty-four stenosed forearm AVFs underwent surgery, 32 pre-emptively and 32 post-thrombosis. End points of the study were initial success, restenosis and access loss rates. After treatment, AVFs were surveilled for restenosis by measuring access flow quarterly and performing at least one follow-up angiogram. RESULTS: Initial procedural success was 92%; 100% for patent and 84% for thrombosed AVFs. The restenosis rate was 0.189 events/AVF-year for both patent and thrombosed AVFs, while the access loss rate was 0.016 events/AVF-year in patent and 0.148 in thrombosed AVFs. Stenosis was corrected by NEO in 27 AVFs and by GI in 30. The restenosis and access loss rates were 0.151 vs 0.214 and 0.033 vs 0.019 events/AVF-year for NEO vs GI, respectively. At Cox's hazard analysis, no variable was significantly associated with restenosis, while the timing of intervention was the only significant determinant of access loss, repaired clotted accesses carrying an 8.0-fold relative risk of access loss compared with patent AVFs (P=0.048). CONCLUSION: Our study shows that surgery remains a valid option for the pre-emptive repair of stenosis and to salvage clotted forearm AVFs, offering an excellent initial success rate and low restenosis rate. It confirms that it is better to treat stenosis pre-emptively than post-thrombosis (though the restenosis rate appears to be uninfluenced by the timing of intervention) and suggests that GI compares favourably with conventional NEO