A prospective cohort study of predictors of upper extremity arteriovenous fistula maturation

Background: A substantial proportion of arteriovenous fistulas fail to function adequately for hemodialysis. Existing studies on arteriovenous fistula failure prediction assess patency rather than the more clinically relevant outcome of arteriovenous fistula function. We hypothesized that preoperative demographic and ultrasound characteristics, and postoperative assessment by an experienced vascular access nurse would predict which arteriovenous fistulas will not function adequately for hemodialysis. Methods: Prospective cohort study of chronic kidney disease patients at a tertiary care center in Vancouver, Canada, with arteriovenous fistula creation between 2009 and 2013. Pre and postoperative clinical assessment and ultrasound blood vessel mapping were performed by an experienced vascular access nurse. The primary outcome was failure to achieve a fistula used successfully for hemodialysis (FUSH). Results: Outcomes were assessed in 200 patients; 123 (61.5%) arteriovenous fistulas were radiocephalic. Overall, 26.5% of arteriovenous fistulas had FUSH failure (34.1% of lower arm vs 14.3% of upper arm, p = 0.002). Univariate predictors of FUSH failure included older age (p = 0.03), female sex (p = 0.05), smaller arterial diameter (p ⩽ 0.001), lower artery volume flow (p = 0.04), and smaller vein diameter (p = 0.01). In multivariable analysis, artery diameter (odds ratio: 0.44, 95% confidence interval: 0.28–0.68) most significantly predicted FUSH failure. Vascular access nurse assessment 6 weeks postoperatively correctly predicted outcome in 83.8% of FUSH and 65.0% of FUSH failure. Conclusion: Smaller artery diameter most strongly predicted FUSH failure. Early postoperative nursing assessment was more useful to predict FUSH than FUSH failure, and as such was insufficient in determining which arteriovenous fistulas should be abandoned as many predicted to fail could be salvaged with further intervention.