The most recent national data show that at the end of 2006, nearly a third of a million Americans with end-stage renal disease were receiving renal replacement through chronic hemodialysis.1 In order to receive hemodialysis, a person must have some type of vascular access that permits a large volume of blood to be cycled from the patient through the dialyzer and back into the patient. For most Americans on hemodialysis, vascular access is achieved through a vascular circuit with either a direct arteriovenous anastomosis (arteriovenous fistula [AVF]) or an interposed conduit between the artery and vein (arteriovenous graft [AVG]).

Both AVFs and AVGs are considered permanent hemodialysis access, although it is widely recognized that both of these types of access circuits are prone to failure on the basis of stenosis and thrombosis, as well as aneurysmal and pseudoaneurysmal degeneration. In Huber's meta-analysis of AVFs and AVGs, the primary patency of more than 1,800 AVFs was 51% at 18 months, but that does not include AVFs that never matured for use, often due to stenosis or thrombosis. In fact, in a recent multicenter prospective study of 877 AVFs,2 60% of all AVFs did not mature for use within 4 months, and more than half were abandoned without expectation of future use. When one looks at primary patency of an AVF based on the intention to use it for hemodialysis, the combined effect of failed maturation and attrition due to late stenosis and thrombosis would probably result in fewer than half of the AVFs remaining primarily patent and functional at 1 year.

For AVGs, early failure and maturity problems are not that common. Nevertheless, primary patency of a useable AVG is inferior to that of a useable AVF. In their meta-analysis, Huber et al found that primary patency for more than 1,200 AVGs was only 33% at 18 months.3 Miller et al4 reported 12-month AVG patency that was only 23% in their study of 256 AVGs. Most often, AVGs fail due to the development of stenosis at the venous anastomosis as well as in the venous outflow central to the graft.

AV ACCESS ANGIOPLASTY
The point is that there is a high failure rate for both AVFs and AVGs at 1 year, and both surgical and percutaneous options are available for maintaining hemodialysis access. Percutaneous transluminal angioplasty (PTA) has been widely adopted as a first line of therapy for AV access stenosis largely because it can be readily scheduled and performed (compared to surgical revision) due to the proliferation of outpatient treatment facilities where it is available.

AV access PTA has an extremely high technical success rate. In the early 1990s, Beathard5 reported that PTA was 94% anatomically successful (536 procedures in 285 patients) with a complication rate of 3%. In a more updated review, Beathard et al6 reported a 97% anatomic success rate in 1,561 AVF PTA procedures and a 98% success rate in 3,560 AVG procedures. The major complication rates were only 0.19% and 0.11%, respectively, for AVF and AVG angioplasties. Today, using the National Kidney Foundation's Kidney Disease Outcomes Quality Initiative (KDOQI) definition of post-PTA, anatomic success in which a treated lesion should have less than 30% residual stenosis,7AV access PTA is widely seen as a terrific technique that effectively treats stenosis, maintains access function, and has few associated major complications.

Yet despite achieving an acceptable anatomic result in nearly all patients, the problem with AV access PTA is that it is not very durable. Recoil and neointimal proliferation at the PTA site frequently lead to recurrent stenosis within several months after PTA. Based on many reports, the 2000 KDOQI Vascular Access Clinical Practice Guideline 19 recommended a target goal of "50% unassisted patency at 6 months" after successful PTA of an AVG stenosis.6 Clearly, the expectation for PTA durability in AV access is much lower than that of PTA used almost anywhere else in the body, whereas failure of PTA in half the patients at 6 months would not be acceptable.

AV ACCESS STENTS AND PERCUTANEOUS CUTTING BALLOON ANGIOPLASTY
Once believed to hold promise in prolonging PTA patency, it is now accepted that stents offer no advantage over successful AV access angioplasty.8-11 This is largely due to neointimal proliferation leading to in-stent stenosis (Figure 1). Contemporary use of stents in AV access intervention is best summarized in the 2000 KDOQI Vascular Access Clinical Practice Guideline 19,6 which states that, "stents are useful in selected instances (eg, limited residual access sites, surgically inaccessible lesions, contraindication to surgery) when PTA fails." Simply stated, stents are used as a PTA bailout. Although stents play a small role in AV access intervention, this role is nevertheless important. It is therefore surprising that during the past decade, the only stent that has actually received FDA approval for use in AV access is the Wallstent (Boston Scientific Corporation, Natick, MA), and it has been approved for use in central veins only.

Hopes that the Peripheral Cutting Balloon (Boston Scientific Corporation) would improve conventional angioplasty patency were dashed by the Cutting Edge trial, one of the few prospective, randomized, multicenter clinical trials regarding maintenance of AV access circuits. This study demonstrated that there was no patency advantage of the Peripheral Cutting Balloon over standard balloon angioplasty for treating venous anastomotic stenosis in AVGs.12 The Cutting Edge trial not only found that the initial results and 6-month patency were similar for the Peripheral Cutting Balloon and conventional angioplasty groups but that there were more procedure-related complications in the Peripheral Cutting Balloon group.

COVERED STENTS FOR AVG STENOSIS
In 2004, a multicenter, randomized clinical trial of the Flair covered stent (Bard Peripheral Vascular, Inc., Tempe, AZ) was completed and reported in 2005.13 The Flair is a self-expanding nitinol stent embedded within expanded polytetrafluoroethylene (ePTFE) graft material. The Flair trial compared conventional PTA to PTA with immediate covered stent placement for treatment of stenosis at the venous anastomosis of AVGs (where most AVG stenoses develop). Unlike nearly all other AV access studies, the Flair trial required that both angiographic and clinical criteria be achieved in order for the AVG to be declared patent. In other words, if the access was functional without any clinical problem but a 50% stenosis was seen at the treatment site, patency was lost. Or if there was any report of AV access dysfunction (based upon KDOQI parameters and defined in the clinical protocol) but no stenosis was found anywhere in the AVG, then access circuit patency was lost. Finally, if the interventionist decided to treat a 30% stenosis in a fully functional access even though there were neither clinical nor angiographic criteria for failure, primary patency was lost.

Given the very high standard that each AVG had to pass and also noting that in both groups there were bona fide stenoses in dysfunctional AVGs, the patency rates for the entire study were much lower than studies in which only clinical outcomes are measured. Requiring both angiographic and clinical patency success at 6 months, primary patency for AVGs in patients who received the covered stent was nearly twice as good as for patients who were treated with PTA alone (38% vs 19.8%; P=.008), and patency at the treatment site was more than doubled (50.6% vs 23.3%; P<.001). Not only did the Flair group have much better treatment site and access circuit patency, but some of the 6-month angiographic follow-up studies of the Flair demonstrated that a covered stent could remain free from hemodynamically significant stenosis. In some cases, there was negligible tissue anywhere within the covered stent or at either end (Figure 2), which was very different from the experience with bare-metal stents.

Now, largely based on these clinical data, the Flair has received FDA approval for primary use when performing PTA of AVG venous anastomotic stenosis, even when PTA is technically successful. With release of the Flair in the US, a larger randomized clinical study (the Post Approval Study of the Flair Endovascular Stent Graft [RENOVA] trial) has been initiated to collect data in 270 patients with venous anastomotic stenosis. The RENOVA study is an FDA-required postapproval trial. Although there are many similarities to the pivotal Flair study completed in 2004, the RENOVA trial will go further, following patients to 12 months after treatment. It is a prospective, multicenter, randomized trial of PTA versus Flair that will characterize outcome not only on the basis of AVG patency following PTA or Flair, but also freedom from intervention and time between interventions for both groups. Concurrently, there is a 280-patient prospective, multicenter, randomized clinical trial of the Viabahn Endoprosthesis (W. L. Gore & Associates, Flagstaff, AZ) for treating AVG venous anastomotic stenosis. This trial, the Vascular Access Revision with Viabahn Endoprosthesis vs Percutaneous Transluminal Angioplasty (REVISE) study, announced enrollment of its first patient in September 2008. With both the RENOVA and REVISE clinical trials underway, it is hoped that we will get a comprehensive understanding of the role of ePTFE-covered stents in AVGs.

COVERED STENTS FOR AVF STENOSIS
Although clinical trials are being done to explore the role of covered stents in AVGs, today, there are fewer AVGs than AVFs largely due to the success of the Fistula First program.14 How do covered stents fare when used to treat stenoses in AVFs? There are very little clinical data. A recent retrospective, single-center report describes use of the Fluency ePTFE-covered stent (Bard Peripheral Vascular) to treat five stenotic AVFs with 80% 9-month patency.15 Shemesh et al described a series of AVF cephalic arch stenoses randomized to treatment with either angioplasty and bare stent (Luminexx, Bard Peripheral Vascular) versus angioplasty and covered stent (Fluency).16 The Fluency covered stent afforded superior primary patency, less angiographic restenosis, and fewer reinterventions compared to the bare stent, with a mean clinical follow-up of 13.7 months. Although covered stents may be useful adjuncts to angioplasty in AVFs, larger studies are needed before covered stents can be broadly advocated for maintenance of AVF patency.

OTHER APPLICATIONS OF COVERED STENTS IN AV ACCESS
A few other potential applications of covered-stent technology in hemodialysis access should be mentioned, such as treatment of PTA-induced rupture, pseudoaneurysm repair, and as an adjunct to PTA during treatment of central vein stenosis and occlusion.

At the 2008 Society of Interventional Radiology meeting, we reported use of the Fluency covered stent to treat immediate PTA-induced rupture in both AVGs and AVFs17 with excellent technical success, although 6-month patency of the AV access was similar to reports in which bare stents were used to treat rupture. It is not clear why patency was not improved, although these access circuits may fail for many reasons, often related to the development of new stenoses elsewhere in the circuit.

There may be a role for covered stents in the treatment of AV access aneurysms and pseudoaneurysms as well. The Viabahn covered stent has been successfully used to treat AVG pseudoaneurysms, as reported by Vesely.18 These pseudoaneurysms form at the cannulation sites of the AVG due to repeated puncture of the graft material with dialysis needles. Treatment of these pseudoaneurysms with the Viabahn necessitated its placement at a cannulation site where it would be repeatedly punctured, so it was not a great surprise that Viabahn stent fractures were seen over time.

Finally, both Fluency and Viabahn covered stents have been used to treat central venous stenosis in hemodialysis circuits with anecdotal success, but so far, the data are insufficient to support this practice. Neither the Fluency nor the Viabahn were specifically designed for use in central veins, where covered stent length, diameter, design, and delivery system requirements are very different from what is needed in the AV access circuit, tracheobronchial, or peripheral arterial systems. While these devices may work better than PTA in central veins (although we do not know that for sure), they have not been optimized for this application.

CONCLUSION
The Flair covered stent is the first FDA-approved, stent-based device for use in peripheral AV access, where it is indicated as a primary treatment, rather than a bailout. Its approval was based on better patency than PTA when treating AVG venous anastomotic stenoses. Both the RENOVA and REVISE clinical trials will likely add a great deal to our understanding of covered stent use in AVGs. Meanwhile, the challenges and opportunities that lie ahead include evaluation of covered stent usefulness in AVFs, AV access pseudoaneurysms, and central vein obstructions.

Bart Dolmatch, MD, is Professor of Radiology and Director of Vascular and Interventional Radiology at the University of Texas Southwestern Medical Center in Dallas, Texas. He has disclosed that he is a paid consultant and speaker for Bard Peripheral Vascular, Inc., and receives royalties from Bard Peripheral Vascular, Inc. He receives stock options for participating on the advisory board of the Endovascular Forum. He is co-course director of Controversies in Dialysis Access, for which he receives an honorarium. Dr. Dolmatch may be reached at bart.dolmatch@utsouthwestern.edu.