Outcome and prognostic factors of restenosis after percutaneous treatment of native hemodialysis fistulas.

PURPOSE: To assess patency after percutaneous treatment of dysfunctional and thrombosed native arteriovenous fistulas and to examine predictors of patency after intervention.

MATERIALS AND METHODS: A cohort of 65 consecutive patients with dysfunctional (n = 53) or occluded (n = 12) native fistulas who underwent 96 percutaneous interventions over an 18-month period was retrospectively analyzed. Fistula locations were radiocephalic (n = 37), brachiocephalic (n = 10), or brachiobasilic (n = 18). Primary interventions consisted of angioplasty (n = 50), stent placement (n = 3), or percutaneous thrombolysis/thrombectomy (n = 12). Additional interventions during follow-up consisted of angioplasty (n = 22), stent placement (n = 6), or percutaneous thrombolysis/thrombectomy (n = 3). Duration of fistula function was assessed clinically and examined as a function of anatomic and clinical variables with use of Cox hazards models and the Kaplan-Meier method.

RESULTS: Clinical success with resumption of at least one session of normal dialysis occurred in 94% (90 of 96) of interventions. The 30-day morbidity rate was 2.1%; no procedure-related deaths occurred. Primary, assisted primary, and secondary patency rates (+/- SE) of dysfunctional fistulas after intervention at 12 months were 26% +/- 11%, 80% +/- 6%, and 82% +/- 6%. Occluded fistulas after intervention had 3-month primary, assisted primary, and secondary patency rates of 60% +/- 15%, 60% +/- 15%, and 80% +/- 13%. Lesions 2.0 cm or more in length were five times more likely to have loss of patency than lesions smaller than 2.0 cm. The presence of at least one comorbid factor--diabetes, coronary artery disease, or peripheral vascular disease--was associated with nearly twice the risk of patency loss after any intervention.

CONCLUSION: Despite modest primary patency rates in our experience, high assisted and secondary patency rates can be achieved with percutaneous intervention in native arteriovenous fistulas. These findings emphasize the need for close surveillance of native fistulas and a low threshold for diagnostic fistulography after initial intervention. The most detrimental determinant of outcome was lesion length > or =2 cm.