Effects of a More Selective Arteriovenous Fistula Strategy on Vascular Access Outcomes

Abstract

Significance Statement The optimal choice of vascular access for patients undergoing hemodialysis—arteriovenous fistula (AVF) or arteriovenous graft (AVG)—remains controversial. In a pragmatic observational study of 692 patients, the authors found that among patients who initiated hemodialysis with a central vein catheter (CVC), a strategy that maximized AVF placement resulted in a higher frequency of access procedures and greater access management costs for patients who initially received an AVF than an AVG. A more selective policy that avoided AVF placement if an AVF was predicted to be at high risk of failure resulted in a lower frequency of access procedures and access costs in patients receiving an AVF versus an AVG. These findings suggest that clinicians should be more selective in placing AVFs because this approach improves vascular access outcomes. Background The optimal choice of initial vascular access—arteriovenous fistula (AVF) or graft (AVG)—remains controversial, particularly in patients initiating hemodialysis with a central venous catheter (CVC). Methods In a pragmatic observational study of patients who initiated hemodialysis with a CVC and subsequently received an AVF or AVG, we compared a less selective vascular access strategy of maximizing AVF creation (period 1; 408 patients in 2004 through 2012) with a more selective policy of avoiding AVF creation if failure was likely (period 2; 284 patients in 2013 through 2019). Prespecified end points included frequency of vascular access procedures, access management costs, and duration of catheter dependence. We also compared access outcomes in all patients with an initial AVF or AVG in the two periods. Results An initial AVG placement was significantly more common in period 2 (41%) versus period 1 (28%). Frequency of all access procedures per 100 patient-years was significantly higher in patients with an initial AVF than an AVG in period 1 and lower in period 2. Median annual access management costs were significantly higher among patients with AVF ($10,642) versus patients with AVG ($6810) in period 1 but significantly lower in period 2 ($5481 versus $8253, respectively). Years of catheter dependence per 100 patient-years was three-fold higher in patients with AVF versus patients with AVG in period 1 (23.3 versus 8.1, respectively), but only 30% higher in period 2 (20.8 versus 16.0, respectively). When all patients were aggregated, the median annual access management cost was significantly lower in period 2 ($6757) than in period 1 ($9781). Conclusions A more selective approach to AVF placement reduces frequency of vascular access procedures and cost of access management.