Keep Seeking the Holy Grail: Predictive Modeling of Arteriovenous Fistula Maturation and Survival

Fistula maturation in its most clinically applicable form is the state when dialysis unit staff can use an autogenous access exclusively and reliably for the prescribed dialysis sessions. This use should be possible in a way that maximizes access survival, such as allowing for cannulation along a sufficiently long segment of the access vein to prevent or minimize skin thinning and aneurysm formation.

Practical criteria that make an access “mature” in this way are (1) delivery of flow volume that supports dialysis, (2) depth that allows palpation of the access and reaching it with standard length dialysis needles, and (3) augmentation of the access during cannulation to safely position the needle tip and shaft in the access vein.

Many investigations have sought to determine the magic formula to predict arteriovenous (AV) fistula maturation and/or describe risk factors for AV fistula maturation failure, yet validated, standardized criteria that define fistula maturation are lacking. As a result, published recommendations defining AV fistula maturation have largely resulted from experts, and different, somewhat complex criteria are applied to clinical trial design, making it difficult to compare outcomes.

The Dialysis Access Consortium Study, the first large US randomized controlled trial of an intervention to reduce early AV fistula failure, was characterized by a notably high failure to mature rate of 66% for forearm and 54% for upper arm fistulas¹; however, it is important to note that the reported intervention rate was low. If procedures to correct depth and improve augmentation are included, most forearm accesses will undergo interventions before use.²

This study was followed by the National Institute of Diabetes and Digestive and Kidney Diseases Hemodialysis Fistula Maturation (HFM) observational cohort study, which comprised 602 patients with 47% upper arm brachial-cephalic, 27% single-stage upper arm transposed brachial-basilic, and 22% forearm radial-cephalic autogenous accesses (there were an additional 2% of other accesses in forearm and upper arm). The primary outcome was defined as unassisted clinical maturation: the use of two needles for ≥75% of dialysis sessions over a continuous 4-week period and either (1) four consecutive session with two needles and >300 blood pump speed or (2) Kt/V ≥1.4 or urea reduction ratio >70% during any session when two needles are used. This study hypothesized that AV fistula maturation is determined by (1) vascular anatomy, (2) vascular biology, (3) other clinical attributes, and (4) process of care. However, no such preoperative factor was identified as predictive, while there was a notable center effect. In general, early flow volume was correlated with later successful use of the access (“what works [early] works”). The follow-up period was up to 4 years, shorter than the mean survival of AV fistulae in an era of low interventions, and this study was not designed to assess long-term AV fistula survival.

The study by Ng and colleagues,³ published in this issue of CJASN, evaluates the performance of the primary HFM study and alternative measures in predicting access maturity and survival. Three clinically relevant outcomes were assessed with the various models on 535 HFM study participants: unassisted AV fistula maturation, overall AV fistula maturation, and overall AV fistula survival.

In comparing the various criteria, no model was found to be statistically stronger than the other in predicting unassisted patency or overall maturation at any time point, nor was there any difference between younger versus older patients. Not surprisingly, the authors found that unassisted AV fistula maturation and attainment of overall maturation were associated with lower risk of AV fistula abandonment. Moreover, their results showed that the addition of dialysis parameters, such as solute clearance or pump speed, was not important in predicting AV fistula maturation or overall AV fistula survival. In short, even the simplest criteria of successful cannulation, “use of fistula,” perform well in predicting abandonment of the access over the study period.

Our dialysis access care system is best described as fragmented: surgeons create the access, aftercare may be at the same institution but different specialty (interventional nephrology or radiology), outpatient access centers completely removed from the surgeon's office, and patients move around between dialysis centers. Be that as it may, once an AV fistula has become the dialysis access that is used, this system is able to maintain the access for up to 4 years with reasonable success. Although the HFM study was not designed to test whether accesses that required assistance to reach cannulation fared worse, data presented by Ng et al. suggest that there is little difference. This also counters the notion that AV fistulae that require early interventions are doomed for failure,⁴ and a more differentiated view likely has be to be developed.

What Does That Mean for Clinical Practice?

First, to achieve useability of AV fistulae, proper selection of the inflow artery is needed to yield sufficient dilatation of the AV fistula and the required flow volume to support dialysis. Even when this is achieved, widespread obesity in the United States and the high prevalence of diabetes and its associated body habitus limit palpability as well as augmentation of an access. There are no guidelines for systematic assessment of access depth, and there are no general recommendations to perform procedures such as superficialization and lipectomy to establish a longer use segment for cannulation. Moreover, the presence of multiple side branches may divert flow from cannulation target. Both factors combine to limit access vein augmentation such that cannulators lack biofeedback during needle insertion, with an increase in infiltrations.

Second, long-term fistula survival is limited by the loss of patency and failure to re-establish flow or loss of skin integrity overlying the cannulation segment that cannot be repaired or revised necessitating access ligation. The development of needle insertion aneurysm is a cofactor for both difficult to resolve access thrombosis and skin ulceration. Large observational studies before the use of routine angioplasty and stent/stent-graft placement revealed median survival times for different types of AV fistulae once in use: upper arm brachial-cephalic AV fistula approximately 3.6 years, upper arm transposed brachial-cephalic AV fistula approximately 5 years, and forearm radial-cephalic AV fistulae approximately 7 years.⁵ Access failure in upper arm accesses is typically associated with recurrent stenosis of the at-risk outflow segments, cephalic arch, and basilic swing point, respectively, where relatively high inflow is impeded with ensuing high intra-access pressures leading to enlarged needle insertion site aneurysm. By contrast, forearm fistulas usually fail in the context of juxta-anastomotic stenosis, a low flow–low pressure combination that complicates needle insertion and leads to lower clearance.

It is worth mentioning that dialysis access needs are often discussed in the context of short expected survival of patients with kidney failure. However, it is now clear that patients with kidney failure who survive the first year on dialysis only experience an average 15% per year mortality thereafter compared with 21% in 2004,⁶ and wait times for transplantation are frequently measured in the 5–8 years range. For these reasons, creating a dialysis access with an expected use period of 10 (rather than <5) years is meaningful.

The original HFM study did not deliver a silver bullet to inform which patient will successfully “mature” an AV fistula in a specific location. Its “center effect” suggests that some clinical access creation systems are more successful, as they have figured out processes of care that allow surgeons, interventionalists, and cannulators in the dialysis unit to take vessels and fashion them into a physiologically mature access that can be successfully cannulated, while other systems fail to do so with the same starting material. The HFM study also serves as a reminder that surgical training is a key to both AV fistula placement and survival. We know that regardless of surgical subspeciality, increasing surgical volume is associated with higher odds of successful AV fistula maturation,⁷ suggesting that outcomes may be improved by identifying optimal practices among high-performing surgeons.

Overall, future efforts of the research community should focus on identifying and sharing the factors contributing to successful surgical access creation, examine when and how balloon angioplasty and side branch ligations may be performed to enhance access maturation, and, importantly, studying the factors and techniques that allow cannulators to successfully use physiologically mature fistulas without complications. To this end, study groups should include access surgeons, clinical scientists, dialysis access interventionalists, and cannulators and patients.

Disclosures

D.M. Hentschel reports consultancy for Bard-BD, Biosurfaces, BluegrassVascular, Laminate, Medtronic, Meritmedical, MinnHealth, and SurModics; ownership interest in BluegrassVascular, Nephrodite, and Venova Medical; and advisory or leadership role as President of Vascular Access Society of the Americas and role on the Scientific Advisory Board of Nephrodite. H. Wasse reports consultancy for BLuegrass Technologies, enVVeno, Medtronic, Vascudyne, and Venova; research funding from Medtronic; honoraria from Humacyte, Medtronic, Nipro iMEP, and Vascudyne; and role on Advisory Board for Medtronic.

Funding

None.

Author Contributions

Writing – original draft: Dirk M. Hentschel, Haimanot Wasse.

Writing – review & editing: Dirk M. Hentschel, Haimanot Wasse.