Prospective validation of an algorithm to maximize native arteriovenous fistulae for chronic hemodialysis access

The purpose of this study was to evaluate an algorithm to maximize native arteriovenous fistulae (AVF) for hemodialysis access.

The prospective study design was set in an academic, tertiary care medical center. The study subjects were adults referred for permanent, upper extremity hemodialysis access between April 1999 and May 2001. Intervention included Doppler arterial pressures/waveforms and duplex imaging of the basilic, cephalic, and central veins. The optimal configuration for an AVF was determined (criteria: vein >3 mm, no arterial inflow stenosis, no venous outflow stenosis) on the basis of the noninvasive studies, and unilateral arteriography/venography was performed to confirm the choice. Permanent hemodialysis access was created on the basis of the imaging studies, and remedial imaging/intervention was performed if the AVF failed to mature. Outcome measures included impact of the noninvasive/invasive imaging, perioperative morbidity/mortality, incidence of successful AVF, time to cannulation, and predictors of AVF failure with multivariate analysis.

A total of 139 new access procedures was performed in 131 patients (age, 53 +/- 16 years; male, 51%; white, 60%; diabetic, 49%; actively undergoing dialysis, 50%; prior permanent access, 26%). The noninvasive imaging showed that 83% of the patients were candidates for AVF, with a mean of 2.7 +/- 2.1 possible configurations. Invasive imaging was abnormal in 38% (forearm arterial disease > central vein stenosis > inflow stenosis) and impacted the operative plan in 19%. AVF were performed in 90% of the cases (brachiobasilic > brachiocephalic > radiocephalic > radiobasilic), with prosthetic AVF performed primarily because of inadequate veins. Among the patients who underwent AVF, the 30-day mortality rate was 1%, the complication rate was 20% (wound, 10%; hand ischemia, 8%), and 24% needed a remedial procedure. The AVF matured sufficiently for cannulation in 84% of those with sufficient follow-up and was suitable for cannulation by 3.4 +/- 1.8 months. On the basis of an intention to treat approach, an AVF sufficient for cannulation developed in 71% of the 139 cases referred for access. The multivariate analysis predicted that female gender (odds ratio, 9.7; 95% CI, 2.2 to 43.5) and the radiocephalic configuration (odds ratio, 4.6; 95% CI, 1.1 to 18.6) were both independent predictors of failure of the fistula to mature.

With the aggressive algorithm, the construction of native AVF is possible in the overwhelming majority of patients presenting for new hemodialysis access.