The middle-arm fistula as a valuable surgical approach in patients with end-stage renal disease

American and European guidelines recommend the distal radial-cephalic fistula (dRCF) as the first and best hemodialysis access in patients with end-stage renal disease (ESRD). However, this kind of arteriovenous fistula (AVF) shows a limited primary unassisted patency and frequently needs surgical revisions or angiographic procedures, or both. When dRCF is not feasible, guidelines suggest a proximal brachiocephalic AVF. The middle-arm fistula (MAF), or autogenous forearm radial-median direct access, has been suggested as a possible alternative approach. This study evaluated MAF primary unassisted patency, the most frequent causes of MAF failure, and the possible related factors.

Data on patients with a MAF placed from January 1991 until June 2008 were retrospectively collected. The probability of MAF failure overall and by the main subgroups was estimated according to Kaplan-Meier with Greenwood standard error (SE). Comparison of failure among different subgroups was performed using the log rank test in univariate analyses. The Cox regression model was used to investigate factors that independently affected the overall hazard of failure and cause-specific hazard of thrombosis.

At the end of follow-up, 14.0% of MAF failed (11.6% thrombosis, 1.7% stenosis, 0.7% failed maturation), and 44.2% of MAF were still working. Cumulative probability of MAF unassisted primary patency after 4 years from the creation was 79%. Univariate analyses highlighted that women (P = .019), underweight patients (P = .010), and MAF implantation after starting hemodialysis (P < .001) had a higher risk of MAF failure for any cause than men, normal and overweight patients, and MAF implanted before starting hemodialysis. Results of the Cox multivariate analysis for overall MAF failure confirmed that only MAF implantation before starting hemodialysis is a protective factor against any failure (P = .003), whereas female gender (P = .016) was associated with an increase of the thrombosis hazard ratio to 2.04 (95% confidence interval, 1.14-3.63).

Our data demonstrate that MAF has a good unassisted primary patency and suggest that this kind of AVF could be a valuable alternative surgical approach when dRCF is not feasible in ESRD patients.