Vascular access is considered the Achilles heel of the dialysis patient. It constitutes the largest single cause of morbidity in the chronic hemodialysis population, accounting for over 25% of hospitalizations at an estimated cost in the US of at least one billion dollars annually. Currently, complication free survival of vascular access ranges between 30-50% a year and multiple investigative efforts in this area have been initiated and are directed at prolonging the functional life of vascular accesses.

It is not well established whether intervention prior to overt malfunction or thrombosis of the vascular access could reduce these complications and thereby improve the functional longevity of the access. Moreover, once accesses at potential risk are identified, it is not well established which method of intervention, Surgery vs. Angioplasty vs. Expectant Management, is superior in terms of clinical and financial outcome. The proposed study aims to determine whether early intervention of a vascular access determined to be at risk of malfunction and thrombosis improves the long term outcome and, specifically, which means of intervention is preferred.

• Procedure: angioplasty
• Procedure: surgery

• No Intervention: control group
• Active Comparator: angioplasty intervention
• Active Comparator: surgery intervention

Inclusion Criteria:

1. Subjects with End Stage Renal Disease on chronic maintenance hemodialysis 3 times per week
2. Have an arteriovenous (polytetrafluoroethylene) graft as vascular access
3. Have a venous stenosis between 30% and 70% as determined by angiogram

Exclusion Criteria:

1. Native arteriovenous fistula
2. Known previous vascular accesses complications, such as central vein stenosis, and multiple access surgeries >4
3. Unwilling to participate
4. Allergy to iodine
5. Absolute contraindication for surgery (e.g. medical condition precludes anesthesia and surgery)
6. Known arterial limb stenosis or long vessel length venous stenosis which are unamenable to surgical or angioplasty techniques, respectively, and therefore prohibit randomization
7. Known hypercoagulable state