Novel Therapies for Hemodialysis Vascular Access Dysfunction: Myth or Reality?

Summary

Hemodialysis vascular access dysfunction is a major source of morbidity for patients with ESRD. Development of effective approaches to prevent and treat vascular access failure requires an understanding of the underlying mechanisms, suitable models for preclinical testing, systems for targeted delivery of interventions to maximize efficacy and minimize toxicity, and rigorous clinical trials that use appropriate outcome measures. This article reviews the substantial progress and ongoing challenges in developing novel treatments for arteriovenous vascular access failure and focuses on localized rather than systemic interventions.

Vascular access dysfunction is a major source of morbidity for patients treated with maintenance hemodialysis, and the need for novel treatment approaches for hemodialysis arteriovenous (AV) access failure is well recognized. Two recent randomized, placebo-controlled trials found some benefits on AV graft (AVG) patency of orally administered agents, dipryridamole plus aspirin (1) and fish oil (2), and an ongoing randomized clinical trial is evaluating short-term use of orally administered sirolimus on the patency of AVGs and AV fistulas (AVFs) after clinically indicated angioplasty (Sirolimus Use in Angioplasty for Vascular Access Extension (SAVE) NCT01595841). However, enthusiasm for systemic administration of agents to address vascular access dysfunction is somewhat limited because of off-target effects and toxicities. Recent advances that are setting the stage for the successful development of novel, locally administered interventions for vascular access dysfunction include elucidation of underlying mechanisms (see Lee et al. [3] and Remuzzi et al. [4] in this issue of CJASN), availability of animal models for evaluating treatment candidates, advances in targeted delivery of pharmacologic and biologic agents, progress in establishing outcome measures for clinical trials, and interest by the biopharmaceutical industry.

Animal Models of AV Access

Animal models have been developed to study the pathogenesis of vascular access failure and test therapeutic approaches (Table 1). These models provide information about integrated, whole-body responses to hemodynamic and other factors that occur after placement of an AVG or creation of an AVF. They also provide dose–response and safety information that can guide human studies of new therapeutic agents. However, animal models are expensive and have important limitations that are described below.

Table 1.

Animal models of hemodialysis access dysfunction

Animal/Access	Location/Configuration	Advantages	Limitations
Mouse
AVF	End of CA to side of JV	Transgenic animals available; numerous commercially available assay tools; inexpensivea	Blood flow parameters (wall shear stress and velocity) different than humana; geometry not used in humans (vein anastomosed to side of artery in human)
AVF	Side of aorta to side of vena cava	Easy to perform	Induces congestive heart failure
AVF	End of JV to side of CA	Similar geometry to humans	Technically challenging to perform
AVG	2-mm microvascular catheter anastomosed between CA and JV	Easy to perform	Catheter material not the same as graft material used in humans
Rat
AVF	End of CA to side of JV	Relatively inexpensive; relatively easy to perform	Geometry not used in humans (vein anastomosed to side of artery in humans)
AVF	End of EPV to side of FA	Similar geometry to humans	Technically challenging to perform
AVF	End of lateral vein to side of ventral artery in tail	Easy access to vessels for perivascular application of drug	Technically challenging to perform; scar tissue formation can compress AVF
AVF	End of FA to side of FV	Relatively easy to perform	Geometry not used in human
Dog
AVF	End of FV to side of FA	Similar geometry to humans; blood flow parameters similar to humans, particularly in the peripherally placed access locations	Slow development of NH; companion animals; measures may be needed to prevent animal’s access to wounda
AVG	Graft anastomosed between FA and FV
AVG	Graft anastomosed between CA and JV
Sheep
AVF	Side FA to side of FV	Geometry similar to that used in humans. Blood flow parameters similar to that in humans, particularly in the peripherally placed access locations.	Expensive; cumbersome to handle; genome not completely sequenced; limited assay tools
AVG	Graft anastomosed between JV and CA
Swine
AVF	End of FV to side of FA	Similar geometry to humans; blood flow parameters similar to humans, particularly in the peripherally placed access locations; animal cannot access wound.	Expensive; cumbersome to handle; fewer commercially available; limited assay tools; genome not completely sequenced; with FA-FV, edema in limb can occur and restrict mobility; with CA-JV, bilateral access can result in cerebral ischemia and stroke
AVF	End of JV to side of CA
AVF	Side of FV to side of FA
AVG	Graft anastomosed between JV and CA
Nonhuman primates
AVG	Side of aorta to side of vena cava	Blood flow parameters similar to humans, particularly in peripherally placed access locations; phylogenetically close to humans	Very expensive and cumbersome to handle; highly sentient
AVG	Axial artery to brachial vein

These advantages or limitations apply to all models using this animal. AVF, arteriovenous fistula; CA, carotid artery; JV, jugular vein; AVG, arteriovenous graft; EPV, epigastric vein; FA, femoral artery; FV, femoral vein; NH, neointimal hyperplasia.

AVFs have been created in mice by anastomosing the end of the carotid artery to the side of the jugular vein (5). This configuration differs from the configuration of human AVFs, which are usually created through anastomoses between the end of a vein and the side of an artery. Recently, a mouse AVF was created by anastomosing the end of the external jugular vein and the side of the common carotid artery (6). Although creating an AVF with this configuration is more technically challenging, the resulting hemodynamics more closely approximate the hemodynamics of AVFs used for dialysis. Advantages of using mouse models include low cost compared with large animals, the availability of assay reagents, and the ability to study the influence of gene expression on access function through the use of available or relatively easily-created transgenic animals (7–11). Additionally, because mouse models of kidney failure are well established, mouse AVFs provide the important opportunity to study AV access treatments in the presence of uremia. Indeed, AVFs created in uremic mice have more pronounced neointimal hyperplasia, inflammation (12), and endothelial barrier dysfunction compared with AVFs created in the setting of preserved kidney function (13). Because of the small size of mouse vessels, creation of an AVF requires substantial surgical skill, and wall shear stress and blood velocities can be orders of magnitude greater than in human AV accesses. These hemodynamic characteristics likely contribute to the accelerated pace of neointimal hyperplasia in this model (8). It is less technically challenging to create an AVF in a rat compared with a mouse, but as with the mouse, the AV access blood flow parameters differ from those parameters in humans (14–16).

To better replicate human hemodynamics, AV access models have been developed using larger animals, such as dogs, sheep, swine, and nonhuman primates. Some of these models use peripheral locations, but others use central vascular locations, where the suture sites are less likely to be disturbed by the animal. However, this benefit is counteracted by higher blood flow rates in the central arteries and the possibility of lower resistance in the central veins compared with the peripheral locations that are usually used in humans. The development of a neointima is typically slower in dogs compared with other animals (17), suggesting that the dog AV access provides a better representation of the human AV access; however, this finding also means that longer studies are required to evaluate effects of interventions that target intimal hyperplasia.

The vascular system of swine is similar in size and flow parameters to that of humans. As in humans, neointimal hyperplasia in the pig develops predominantly at the vein–graft anastomosis; however, the process is more rapid, typically evident within 4–6 weeks (18–23). Misra et al. (24–26) have developed a porcine model of renal insufficiency that provides a more relevant background for evaluating treatment effects and the pathophysiology of AVG and AVF, but use of this model is costly and technically challenging.

Baboons have been used to study effects of hemodynamics and antithrombotic therapy on neointima development in AVGs (27,28) and to evaluate bioengineered grafts (29). However, baboon purchase and care is expensive, and there is significant public opposition to the use of nonhuman primates in research.

Novel Therapies

This review focuses on therapies to prevent or treat AV access failure that are delivered locally through endovascular approaches or perivascular administration (Table 2).

Table 2.

Localized approaches for AV access dysfunction

Approach	Comments
Endovascular approaches
PTA	Restores patency of thrombosed AVGs and AVFs
	May enhance maturation of AVFs
	Role in preventing access thrombosis unclear
Stents: bare metal and drug-eluting	Newer metallic alloys, such as nitinol, may be more efficacious than older stainless steel stents; stents can fracture, move, and restrict later access revisions; efficacy of drug-eluting stents in the hemodialysis access setting has not been evaluated by a clinical trial
Drug-eluting balloon angioplasty (Table 3)	No placement of permanent scaffold
Stent grafts
Nonbiodegradable stent grafts	Appear to be efficacious; stent grafts can fracture or move and restrict later access revisions
Bioresorbable stent grafts	Can deliver pharmacologic agents and provide temporary intravascular support; have not been assessed in dialysis access
VIABAHN endoprosthesis stent graft coated with bioactive heparin	Heparin-induced thrombocytopenia is possible but unlikely; clinical trial underway to assess efficacy after PTA in AVG
Cutting balloon angioplasty	Sharp atherotomes embedded in the balloon score the stenotic lesion longitudinally to create more controlled disruption of the lesion
Cryotherapy
Brachytherapy
Gene therapy: intraluminal application, naked plasmid DNA vectors, adenoviral vectors	Possible immune rejection and tumorigenesis
Microinfusion catheter	Delivers therapeutic agent into the vessel wall and pervascular space through microneedles; avoids systemic exposure
Perivascular approaches
Transfection of the adventitial layer with genes encoding antiproliferative proteins	Avoids disturbance of the endothelium
Introduction of endothelial cells that provide beneficial vasoactive molecules using a gelatin-based cell-seeded sponge	Uses allogeneic cells
Application of autologous adipose tissue as drug depot for glitazone drug and biofactory for adiponectin	Uses autologous cells
Recombinant human type I pancreatic elastase
Application of drug-laden biodegradable polymer at time of access creation and/or by ultrasound-guided injection after creation	Paclitaxel, dipyridamole, and sirolimus have been delivered; polymers can induce foreign body reactions
Far-infrared treatment	Electromagnetic therapy that induces vasodilation and vasoprotective gene expression
Novel grafts or engineered vessels
Grafts
Heparin-bonded synthetic graft (Propaten graft, Gore) for AVG	Heparin-induced thrombocytopenia is possible but unlikely
Optiflow (Bioconnect Systems, Inc.) synthetic anastomotic implant	Configuration is an AVF but uses synthetic graft material for anastomosis creation Initial studies in 10 hemodialysis patients showed 100% technical success and 90% patency at 6 wks (96)
Tissue-engineered blood vessel
Lifeline (Cytograft Tissue Engineering, Inc.)	Engineered vessel constructed from autologous fibroblasts and endothelial cells isolated from dermal and superficial vein biopsies; 6- to 9-mo vessel production time before implant
Decellularized porcine carotid arteries populated with sheep endothelial cells and preconditioned for flow	Withstood repeat needling three times a week for 6 mo in a sheep AVG model; 5-wk vessel production time
Polymer scaffold seeded with human allogeneic cells collected from cadaver donors and then decellularized for off-the-shelf use	Clinical trials underway in Europe and the United States

PTA, percutaneous transluminal angioplasty.

Endovascular Approaches

Percutaneous Balloon Angioplasty.

The recommended approach to treating stenosis in either AVGs or AVFs is percutaneous transluminal angioplasty (PTA) (30). Although PTA can restore patency and function of accesses with thrombosis, the benefits are less clear for accesses that are stenosed but still patent (31–36). Balloon-induced injury may stimulate more aggressive neointimal hyperplasia by damaging endothelial cells and stimulating smooth muscle cell proliferation.

Drug-eluting balloons compress a stenotic lesion and target an antiproliferative drug to the lesion (Table 3) (37,38). Katsanos et al. (39) compared the effects of PTA with a paclitaxel-eluting balloon or a conventional balloon on stenotic lesions in AVGs and AVFs. At 6 months, the primary patency rate was 70% in the paclitaxel-eluting balloon group compared with 25% in the conventional PTA group. Drug-eluting balloons have a number of benefits over drug-eluting stents: (1) they provide uniform delivery of drug to the entire stenotic lesion, whereas delivery from stents is not homogeneous, (2) they do not require placement of a scaffold that could later preclude other endovascular interventions, and (3) they can be used to treat complex lesions and longer vessel segments. Drug-eluting balloons do not have regulatory approval in the United States.

Table 3.

Drug delivery balloon technology

Name	Principle	Manufacturer
Paccocath	Paclitaxel embedded in hydrophilic iopromide coating	Bayer (Bavaria Medizin Technologie, Oberpfaffenhofen, Germany)
SeQuentPlease	Paclitaxel embedded in hydrophilic iopromide coating	B. Braun Melsungen AG (Melsungen, Germany)
Coroflex DEBlue	Combination of a paclitaxel-eluting balloon and a bare metal stent	B. Braun Melsungen AG
DIOR PTCA	Paclitaxel-coated microporous balloon	Eurocor (Bonn, Germany)
MAGICAL	Paclitaxel-coated balloon combined with a bare metal stent	Eurocor
Elutex	Balloon is coated with a hydrogel; on expansion, paclitaxel is released from balloon to gel; gel is retained on vessel wall for prolonged drug delivery and decreased washout	Aachen Resonance (Aachen, Germany)
GENIE	Delivers antiproliferative drug (e.g., paclitaxel) in liquid form	Acrostak Corp. (Winterthur, Switzerland)
IN.PACT Amphirion	A wrapped balloon combines urea and paclitaxel molecules to increase paclitaxel solubility	Medtronic Endovascular (Minneapolis, MN)
IN.PACT Admirala	A folding balloon structure delivers paclitaxel	Medtronic Endovascular (Minneapolis, MN)
Advance PTX	Balloon coated with paclitaxel	Cook Medical (Bloomington, IN)
Lutonix DCB	Paclitaxel and proprietary carrier coated on balloon	Lutonix Inc (New Hope, MN)

None of these devices is currently approved for use in the United States. Modified from ref 37.

^aA clinical trial in Singapore (Prospective Randomized Trial Comparing DEB Versus Conventional PTA for the Treatment of Hemodialysis AVF or AVG Stenoses; NCT01544907) is underway to assess the efficacy of this drug-eluting balloon versus conventional percutaneous transluminal angioplasty for treatment of arteriovenous graft and arteriovenous fistula stenosis.

Stents.

Because early trials showed that the unassisted patency rate of AVGs did not differ with stenting than with PTA alone (40–42), stent placement was recommended only in the setting of elastic stenoses or PTA failure. However, a recent prospective study of 61 patients with thrombosed AVG reported significantly higher patency rates after placement of a nitinol stent compared with PTA alone (43). Also, improved patency of stenosed AVGs with placement of nitinol stents compared with PTA alone has been reported (44,45). Using a porcine AVG model, sirolimus-eluting stents showed greater patency and flow rates at 4 weeks compared with bare metal stents or unstented AVG (46).

The 12-month interim results of a 2-year study (Post-Approval Study for the FLAIR Endovascular Stent Graft; NCT00677235) assessing the primary AVG patency rates after treatment of stenotic lesions with stent grafts (covered stents) or PTA alone indicate that the stent graft is associated with significantly greater primary patency rates (24.1% versus 10.3%; P=0.005) (published in abstract form only). A previous 6-month trial comparison between stent grafts and PTA alone also reported better patency rates with the stent grafts (47). However, in the 6-month report, the graft thrombosis rates were not significantly different than the rates after PTA alone, and graft survival rates were not reported. A phase III randomized open-label clinical trial (GORE VIABAHN Endoprosthesis Versus Percutaneous Transluminal Angioplasty (PTA) to Revise AV Grafts in Hemodialysis; NCT00737672) is underway to test the safety and efficacy of a stent graft with a heparin bioactive surface (PROPATEN graft; Gore, Inc., Flagstaff, AZ) in AVG. The study will compare primary patency rates with the stent grafts with PTA alone. Another randomized clinical trial is testing the safety and efficacy of a stent graft to treat in-stent restenosis of bare metal stents in the venous outflow of AVFs and AVGs and will be concluded in 2014 (FLUENCY PLUS Endovascular Stent Graf for In-Stent Restenosis; NCT01257438). Unfortunately, stent grafts and stents can limit the ability to perform other access revisions, and they can fracture or move. Polymer or metallic alloy bioresorbable stents provide a temporary support for the vessel wall but degrade over time, mitigating such factors. In a clinical trial of single de novo coronary lesions treated with an everolimus-eluting bioresorbable polymer stent (48), the stents were shown to be safe; no restenoses or thromboses were reported, and the stented arteries were observed to retain vasomotor responses (49). Use of biodegradable stents for hemodialysis vascular access has not been reported.

Cutting Balloon Angioplasty.

The use of cutting balloons for treatment of high-pressure balloon-resistant stenotic lesions in AVGs and AVFs has been described in clinical reports and small studies (50–54). A multicenter randomized clinical trial of 340 patients with venous outflow stenoses reported equivalent 6-month patency rates between PTA alone and cutting balloon angioplasty, but the cutting balloon approach experienced a greater number of device-related complications, including venous rupture and dissections (55). Another randomized clinical trial is currently recruiting to assess the primary 1-year patency rate after treatment of vascular access stenotic lesions with cutting balloon angioplasty or PTA (Cutting Balloon Versus Non-Cutting Balloon for the Treatment of Venous Stenosis in the Fistulas of Hemodialyzed Patients; NCT01321866).

Cryotherapy.

Another endovascular approach is cryotherapy, in which a balloon catheter (PolarCath; Boston Scientific) inflated with nitrous oxide gas compresses the neointima and cools the vessel wall to −5 to −10°C. Huijbregts et al. (56) compared preventive cryoplasty with conventional PTA on the vein–graft anastomosis in a porcine bilateral AVG model. At 4 weeks, a nonsignificant decrease in the intimal/media ratio was observed in the cryoplasty-treated vein–graft anastomoses compared with the PTA-treated grafts. However, outcomes from clinical studies have been mixed. In one study, cryoplasty of neointimal hyperplasia in the AVGs of five patients increased the time to restenosis or thrombosis from 3 weeks to greater than 16 weeks (57). Another study tested cryoplasty in 18 patients with preexisting AVG neointimal hyperplasia and noted that, although the procedure was safe, there were low anatomic success rates; also, patients reported that the procedure was more painful than PTA (58).

Brachytherapy.

Endovascular brachytherapy is the delivery of ionizing radiation to the luminal surface of the vessel. This procedure is considered safe and technically feasible, but outcomes of animal and clinical studies have been varied. El Sharouni et al. (59) reported that brachytherapy (12 Gy) administered 48 hours after AVG placement in patients did not inhibit stenosis development. In a multicenter trial, 30 patients receiving AVG were randomized to receive brachytherapy. At 1 year, the stenosis rate in the brachytherapy group was 56% compared with 37% in the control group (60). A multicenter clinical trial (Beta Radiation for Treatment of Arterial-Venous Graft Outflow) recruited 25 patients with a single stenosis at the vein–graft anastomosis of AVG for randomization to either radiation (18.4 Gy) after PTA or sham intervention after PTA alone (61). At 6 months, target lesion primary patency was significantly improved in the brachytherapy group (41.6% versus 0%; PTA alone); but 1-year thrombosis and patency rates were the same.

Endovascular Gene Delivery.

Efficient gene transfection of the vein wall after intraluminal delivery of naked plasmid DNA in a rat model of AVF was recently reported (15). Reporter gene expression was detected in the endothelium, adventitia, and media of the fistula vein at least 21 days after injection, with no evidence of inflammation. The AVFs were injected with plasmid DNA coding for the gene endothelial nitric oxide synthase (pl-eNOS) with the expectation that transfected AVF would have less neointima development, because endothelial nitric oxide inhibits cell migration and induces vasodilation. However, neointimal hyperplasia was markedly increased in the pl-eNOS–transfected AVFs, possibly because extraordinarily high levels of eNOS expression resulted in the production of excessive amounts of reactive nitrogen species and oxidative stress. The gene for β-adrenergic receptor kinase C terminus has been delivered to the jugular vein segment of AVG in swine (62). Significant inhibition of neointimal hyperplasia at the vein–graft anastomosis at 4 weeks occurred compared with mock-transfected AVG. However, a 100% stenosis rate was observed in the mock-transfected AVG that is much higher than stenosis rates reported by others in that swine model (63,64).

Microinfusion Catheters.

Microinfusion catheters are a novel means to deliver drugs or proteins from the lumen to the vessel wall and perivascular space. These catheters (Mercator Medsystems, San Leandro, CA) are similar to an angioplasty balloon, except that, on inflation, a small needle punctures the vessel wall and delivers the therapeutic agent. This approach has been approved by the US Food and Drug Administration for the delivery of drugs that have been previously approved for vascular injection.

Perivascular Approaches

Application of biologic or pharmacological agents to the adventitial layer of the blood vessel is appealing given that (1) the adventitial and medial layers are thought to be the source of many of the cells that populate the neointima (20,65–67), (2) application of treatments to the perivascular region can be readily accomplished during vascular access creation and also, long after access surgery through image-guided subcutaneous injection, and (3) denudation of the endothelial layer is avoided.

Biodegradable Polymer Gels.

Drugs, genes, or cells can be delivered to the perivascular region by means of solutions or biodegradable films or gels. A single application of a biodegradable polymer gel containing dipyridamole to the vein–graft anastomosis in a porcine model of AVG achieved tissue drug concentrations in the therapeutic range but did not inhibit neointimal hyperplasia (68). A single application of a polymer gel incorporated with the antiproliferative agent, paclitaxel, did inhibit neointimal hyperplasia in AVG in a dog model (17). Repeat application of a sirolimus-laden polymer gel through ultrasound-guided injection to the vein–graft anastomosis significantly inhibited neointimal hyperplasia at 6 weeks in the swine AVG model (63,69). Perivascular application of expression vectors to alter in vivo vascular gene expression to promote better access outcomes has been reported. C-type natriurietic peptide (CNP) is an endothelial-derived protein that inhibits smooth muscle cell migration and inhibits thrombosis. A solution of recombinant adenovirus-encoding mouse CNP was painted directly onto the adventitial surface of AVG in a porcine model at the time of graft placement (22). Although no decrease in neointimal hyperplasia was observed at 2 weeks, a significant increase in lumen area at the venous anastomosis occurred in the CNP-transfected AVG compared with the bilateral AVG transfected with empty adenovirus vector. Thus, this treatment may promote outward remodeling, which would be anticipated to be beneficial.

Biodegradable Polymer Wraps.

Biodegradable polymer wraps have been used to deliver antiproliferative drugs to the perivascular region of hemodialysis access vessels in both animal models (70) and patients. In animal studies, perivascular wraps that elute paclitaxel were beneficial in inhibiting neointimal hyperplasia in AVGs (71). The paclitaxel wrap combination (Vascular Wrap; Angiotech) was shown to be beneficial in preventing neointimal hyperplasia in peripheral arterial bypass grafts in a European cohort. However, a clinical trial in the United States assessing the use of this approach to inhibit neointimal hyperplasia in AVGs was terminated early because of an increase in graft infections (NCT00448708). Collagen-based wraps have been used to deliver sirolimus in animal models (72) and replication-deficient adenoviral vector expressing the gene for vascular endothelial growth factor–D in humans (Adenovirus Vascular Endothelial Growth Factor Therapy in Vascular Access-Novel Trinam AGainst Control Evidence; NCT00895479).

Cell-Based Delivery.

Seeding endothelial cells to AVGs and AVFs through perivascular administration at the time of access creation has been developed based on the presumption that endothelial cells will provide beneficial vasoactive substances to the cells of the vessel wall (73–75). Nugent et al. (76) cultured porcine aortic allogeneic endothelial cells in a gelatin-based sponge that was then wrapped around the side-to-side anastomosis of femoral artery and femoral vein in pigs. At 2 months, the neointimal hyperplasia index for the endothelial cell-treated AVFs was 68% lower than in controls treated with the gelatin-based sponge alone. A randomized, double-blinded, phase I/II trial investigated the safety of this approach in the AVG and AVF of patients (77). Sponges seeded with allogeneic endothelial cells collected from cadaver aortas (Vascugel) were wrapped around the anastomoses and outflow of the newly created AVG or applied to the anastomosis region and vein outflow of AVF. At 24 weeks, there were also no differences in primary or assisted primary patency rates; however, post hoc analysis of the data suggested that Vascugel treatment was associated with enlargement of the fistula vein in the subgroup of diabetic patients (78). A potential problem with this approach is immunosensitization resulting from the use of allogeneic cells, which is a particular concern for patients who are planning to undergo kidney transplantation. A multicenter phase II clinical trial is currently recruiting patients to assess the efficacy and safety of Vascugel in AVF maturation (NCT01806545). Patients for whom organ transplantation is anticipated are excluded from participation because of the possibility of immunosensitization.

Laborious methods for ex vivo seeding of synthetic grafts with endothelial cells have been developed in an effort to create nonthrombogenic graft surfaces that diminish neointimal hyperplasia. Successful in vivo endothelialization of a synthetic AVG has been accomplished in a porcine model by coating the graft surface with anti-CD34 antibodies. However, an increase in neointimal hyperplasia was observed, and it was postulated that the captured progenitor cells developed into proliferating nonendothelial cells and/or that the captured cells released proliferative growth factors to stimulate cell proliferation/migration from the native cell wall (79,80).

Another cell-based approach is to apply autologous subcutaneous adipose tissue transplants to the perivascular surface of accesses at time of creation. The adipose tissue is mixed with insulin-sensitizing drugs (pioglitazone or rosiglitazone) before transplantation to the perivascular surface of the access veins. Thiazolidinediones induce adipose tissue production of adiponectin, a protein that has many vasculoprotective effects, including inhibition of smooth muscle cell proliferation and inflammation. The adipose/glitazone mixture is proposed to serve as a biofactory for the production of adiponectin, which will then interact with the cells of the vessel wall. The adipose tissue also provides controlled release of thiazolidinedione, which inhibits inflammation and smooth muscle cell proliferation and promotes the release of nitric oxide. The ability of this approach to inhibit AVG stenosis is being assessed in a porcine AVG model in the laboratory by author C.M.T. (Figures 1 and 2).

Figure 1.

Bilateral arteriovenous grafts placed between the common carotid arteries and external jugular veins in swine. Modified from ref 63, with permission.

Figure 2.

Autologous adipose tissue transplants for drug delivery and protein production. (A) External jugular vein (EJV) in swine before adipose tissue transplantation. (B) EJV directly after adipose tissue placement. (C) Adipose tissue grafted perivascular to EJV at 6 days post-transplantation.

Topical Administration of Recombinant Elastase.

Topical administration of recombinant human elastase is being studied in clinical trials of newly created AVFs and AVGs. The hypothesis is that degradation of elastin in the vessel wall will allow greater early dilation after access creation, resulting in favorable hemodynamics. Additionally, it is postulated that elastin fragments have chemotactic properties that could inhibit migration of myofibroblasts from the adventitial to the intima and thereby, inhibit neointimal hyperplasia. An early phase placebo-controlled, dose-escalation trial of human type I pancreatic elastase (NCT01305824) applied to the adventitial surface of the newly created AVFs found that the treatment was well tolerated (81). Larger trials of human type I pancreatic elastase for AVFs (NCT01305824) and AVGs (NCT01001351) have recently been completed, and the results are forthcoming.

Far-Infrared Electromagetic Radiation.

The use of far-infrared electromagnetic radiation is not, strictly speaking, a perivascular approach, because it is applied to the external surface of the arm containing an AV access. This treatment involves the application of electromagnetic radiation (wavelengths of 3–25 μm) to the skin surface overlying the vascular access. The electromagnetic radiation, which travels 2–3 cm below the skin surface, is hypothesized to be beneficial for the vasculature by inducing vasodilation, decreasing oxidative stress (82), inhibiting cell proliferation (83), and inducing protective enzymes, such as heme oxygenase-1 (84). The emitting device is positioned approximately 25 cm above the arm for 30–60 minutes per treatment. A randomized trial of far-infrared therapy administered three times per week for 1 year to patients with new AVFs was recently published (85). Although there were no differences in vein diameter between the intervention and control groups at any time period evaluated, volumetric flow rates were greater in the intervention group at 1, 3, and 12 months (P=0.001), and the intervention group experienced a higher rate of AVF clinical maturation (P=0.008), which was defined as the ability of the AVF to support dialysis within the 12-month follow-up period. Limitations of the trial include its single-center, nonblinded design and enrollment of patients of a single ethnicity.

Novel Grafts and Engineered Vessels

Significant efforts have been directed to developing better materials for AVGs. Conduits derived from native products, such as bovine carotid arteries or ureter, decellularized native vessels, and a patient’s own transplanted veins, have been used with varying outcomes (86). Bioengineering approaches have been used to create autologous grafts from dermal fibroblasts and endothelial cells obtained through tissue biopsy and expanded in tissue culture (87,88). Clinical trials are currently underway in Europe and South America. A drawback of this approach is a prolonged production time of 6–9 months. Another approach used harvested porcine carotids arteries that were depopulated and reseeded with sheep endothelial progenitor cells captured from peripheral blood (89). The repopulated conduits were placed in a sheep AVG model, and after 1 month, they were punctured three times per week. The conduits withstood repeat puncture, but like standard polytetrafluoroethylene grafts, they failed because of neointimal hyperplasia.

Dahl et al. (90) reported the development of a synthetic polymer scaffold incorporated with allogeneic human vascular cells from cadaver donors. After culturing for 8–10 weeks, the cells deposit an extracellular matrix onto the scaffold, and the grafts are decellularized to remove immunogenic proteins. The grafts were tested in a baboon model of arteriovenous access, with seven of eight grafts remaining patent during the observation periods of 1, 3, or 6 months. Very recently, a conduit created using this method was placed in a patient with ESRD as part of a United States–based phase I trial with 20 patients (91). Clinical trials are also underway in Europe.

Outcome Measures for Clinical Trials

Adequate evaluation of the efficacy of the novel interventions described above requires properly designed clinical trials with well-developed outcome measures. Establishing outcome measures for vascular access interventions is in the early stages but progressing. Recent efforts by groups comprised of surgeons, interventional radiologists, and nephrologists to generate standardized terminology and definitions for access function and complications should promote harmonization across studies that will be important for developing outcome measures as well as comparing results across trials (30,92,93). In the absence of well validated surrogate end points, regulatory approval of a new treatment usually requires demonstration of benefit on clinical outcomes. For vascular access interventions, clinical outcomes often relate to usability of the access. Although seemingly straightforward, establishing criteria for access use outcomes has challenges. For AVFs, clinical use outcomes for trials evaluating interventions designed to promote maturation might differ from outcomes for interventions designed to promote long-term access function. Because the time period between access creation and its initial use can vary substantially, especially if the vascular access is created before the initiation of maintenance dialysis, outcomes that incorporate time to access failure can be problematic, particularly if access use, through cannulation trauma or deleterious hemodynamics, contributes to access failure. Excluding from trial participation the subset of patients who are not yet receiving dialysis treatment at the time of access creation should partially mitigate this problem, but it also reduces generalizability of the findings and would likely substantially delay recruitment. An alternative is to begin the time-to-failure clock at initial use of the access; however, there is not an obvious method for assigning failure time for accesses for which use is never attempted.

For interventions targeting AVF maturation failure, it is not clear how to incorporate maturation-enhancing procedures into the access use outcome. The need for a procedure may suggest maturation failure; however, if the AVF ultimately was used for dialysis, an alternative perspective is that maturation was successful. Because performance of a maturation-enhancing procedure is often driven by subjective assessments by treating clinicians, classifying AVFs that had procedures as maturation failures might place excessive reliance on clinical decisions to define the outcome. These issues are also relevant to AVG trials because prophylactic repair procedures are driven by a variety of factors that can be difficult to standardize.

An additional challenge is that access use outcomes can be affected by multiple factors that are remote from the direct effects of the intervention. For example, an outcome based on ability to use a new AVF for dialysis could be affected by clinical decisions about when to initiate use of the AVF or the skill of the individuals performing AVF cannulation—variables that cannot be controlled in a clinical trial and do not reflect the effects of the intervention. If clinical use of an AVF is an outcome for maturation-enhancing interventions, what duration of use should be considered sufficient for defining the AVF as having successful maturation? Should a fistula be required to be usable within a specific time period after surgical creation to be categorized as having successful maturation? If so, what is the appropriate time period?

The cost of conducting clinical trials is an important barrier to developing new treatments for vascular access dysfunction. Surrogate outcomes, although typically not adequate for pivotal clinical trials, can facilitate shorter, less-expensive, early phase evaluations of interventions. Ultrasonographic characteristics, such as blood flow rate and vessel diameter within the first few months after AVF creation, have been proposed as surrogate outcomes for evaluating AVF maturating-enhancing interventions. Although clinical practice guidelines recommend the use of ultrasound for early clinical assessment of AVF maturation (94), the suitability of ultrasonographic parameters as surrogate outcomes for clinical trials is not known. In the National Institutes of Health–sponsored Hemodialysis Fistula Maturation (HFM) Study, a prospective, multicenter, observational study that is currently underway, serial ultrasounds are being performed at uniform time points after AVF creation (95). The rigorous evaluation by the HFM Study of the utility of ultrasounds for predicting AVF maturation should provide important information about the appropriateness of using ultrasonographic characteristics as surrogate outcomes in clinical trials.

Unique Advantages and Challenges for Trials in Vascular Access

From a clinical trial perspective, there are several advantages of studying interventions for vascular access failure. The frequency of access failure (i.e., high event rate) reduces the necessary sample size and/or duration of follow-up needed to detect an effect of the intervention. The ability to follow trial participants in outpatient dialysis settings can reduce the burden to participants of repeated study visits (at least for visits that do not require specialized evaluations, such as ultrasound or other imaging). The superficial location of AV access allows for relatively easy assessment of certain surrogate or intermediate outcomes, and it also allows for repeated administration of localized therapies after its activity diminishes. The potential applicability of therapies for vascular access failure to other conditions, such as peripheral arterial disease or coronary artery disease, provides efficiencies that are advantageous and of interest to the biopharmaceutical industry.

A challenge for vascular access trials is the technical expertise necessary for administration of many of the novel interventions currently under development. Additionally, the underappreciation by many patients and clinicians of the clinical significance of vascular access failure can make it difficult to enroll participants in trials of novel treatments that have uncertain risks. For example, it might be more difficult to enroll patients in first-in-man studies that evaluate cells or viral vectors to deliver the intervention if the target is vascular access failure rather than heart failure or cancer. Efforts to increase awareness of the clinical importance of vascular access failure might reduce this challenge.

Development of effective approaches to prevent and treat vascular access failure is clearly challenging. Agents that seem to work in animals often do not work in the clinical setting, and efficacy of interventions for coronary artery disease or peripheral arterial disease does not necessarily translate into benefit in the context of an artery–vein connection or conduit rather than an artery–artery conduit, or in the setting of uremia. Immunogenicity of cell-based therapies might interfere with future kidney transplantation, the use of stents or stent grafts can limit future surgical revisions, and the perception that vascular access failure is not a highly significant clinical event can limit acceptance of novel experimental treatments. However, the high interest level by academic investigators in multiple disciplines, the growing interest and commitments from the biotechnology and pharmaceutical industries, and the broad array of therapeutic agents and delivery systems that are currently in preclinical and clinical phases of investigation provide strong indications that effective novel interventions in the near future are a distinct reality.

Disclosures

L.M.D. is a member of the Clinical Advisory Board of Proteon Therapeutics and has received research funding from Proteon Therapeutics, the manufacturer of human type I pancreatic elastase, which is discussed in the manuscript.