Arteriovenous conduits for hemodialysis: how to better modulate the pathophysiological vascular response to optimize vascular access durability

Abstract

Vascular access is the lifeline for patients on hemodialysis. Arteriovenous fistulas (AVFs) are the preferred vascular access, but AVF maturation failure remains a significant clinical problem. Currently, there are no effective therapies available to prevent or treat AVF maturation failure. AVF maturation failure frequently results from venous stenosis at the AVF anastomosis, which is secondary to poor outward vascular remodeling and excessive venous intimal hyperplasia that narrows the AVF lumen. Arteriovenous grafts (AVGs) are the next preferred vascular access when an AVF creation is not possible. AVG failure is primarily the result of venous stenosis at the vein-graft anastomosis, which originates from intimal hyperplasia development. Although there has been advancement in our knowledge of the pathophysiology of AVF maturation and AVG failure, this has not translated into effective therapies for these two important clinical problems. Further work will be required to dissect out the mechanisms of AVF maturation failure and AVG failure to develop more specific therapies. This review highlights the major recent advancements in AVF and AVG biology, reviews major clinical trials, and discusses new areas for future research.

INTRODUCTION

The majority of patients with end-stage renal disease (ESRD) use hemodialysis as their primary form of renal replacement therapy. The dialysis vascular access is the conduit that connects the patient’s bloodstream to the hemodialysis machine for hemodialysis therapy. The dialysis vascular access is often referred to as the lifeline for the patient on hemodialysis, and it is arguably the single most important component of the hemodialysis procedure.

There are three types of vascular access: the arteriovenous (AV) fistula (AVF; Fig. 1A), AV graft (AVG; Fig. 1B), which is made of expanded polytetrafluoroethylene (ePTFE), and tunneled dialysis central venous catheter (CVC). CVCs are not intended for long-term use for hemodialysis, so this review primarily focuses on AVFs and AVGs. The preferred long-term vascular access is the AVF due to its increased longevity and fewer maintenance interventions required, compared with AVGs, only “if” the AVF successfully matures for hemodialysis (4). There are three main types of AVFs: radiocephalic, brachiocephalic, and basilic vein transpositions. Currently, the major problem in the field of vascular access is the high rate of AVF maturation failure. AVGs are the next preferred vascular access type when AVF placement is not possible (113a), but the rate of AVG failure is also high (4). Clinically, both AVF maturation failure and AVG failure are associated with venous stenosis (Fig. 2) (5, 7–9, 76), which is commonly characterized by intimal hyperplasia. Our poor understanding of the mechanisms that lead to AVF maturation failure and AVG failure has hindered our advancement to develop effective therapies for this important clinical problem.

Fig. 1.

Arteriovenous fistula and graft. Representative images show a forearm radiocephalic arteriovenous fistula (AVF; A) and an upper arm brachiobasilic arteriovenous graft (AVG; B). [Reprinted from the Atlas of Dialysis Vascular Access with permission from Tuschar Vachharajani.]

Fig. 2.

Angiography and histology of venous stenosis in arteriovenous fistulas (AVF) and arteriovenous grafts (AVG). Representative angiograms and histology of stenosis in a radiocephalic AVF (A and B) and AVG (C and D) are shown. A: venous stenosis is the common angiographic lesion seen in AVF maturation failure and AVG failure. Stenosis in AVF occurs most frequently in the juxta-anastomotic region of the AVF. In other types of AVFs, common areas of stenosis include the cephalic arch in brachiocephalic AVFs and proximal swing segment stenosis in transposed basilic vein transposition AVFs. B: intimal hyperplasia is the most common histological lesion present in AVF maturation failure. Stenosis in AVG occurs most commonly in the vein-graft anastomosis (C) with intimal hyperplasia being the most common histological legion present (D). NH, neointimal hyperplasia. [Adapted and reprinted from Ref. 62 with permission from American Society of Nephrology (62).]

This critical review presents the following: 1) an updated status of the clinical problem of vascular access dysfunction, 2) aberrant hemodynamic and wall mechanical stresses in AV access, 3) intimal hyperplasia development versus outward remodeling in AVFs, 4) exploring novel genes and pathways to treat AV access dysfunction, 5) major clinical trials and studies in AV access, and 6) future perspectives.

CLINICAL PROBLEM

Epidemiology of Vascular Access Dysfunction

While the AVF is the recommended vascular access for patients on hemodialysis (1, 111), vascular access use has significantly changed in the United States (90, 94) and varies significantly between countries (90, 94). From the Dialysis Outcomes and Practice Patterns Study (DOPPS), AVF use in the United States has increased from 20.7% (1997–1999, DOPPS I) to 26.5% (2002–2003, DOPPS II), respectively (94). DOPPS has also reported (DOPPS IV and V) that AVF use increased from 63% to 68% (90) in the United States from August 2010 to August 2013. A recent DOPPS publication [DOPPS IV and V (2009–2015)] compared the location and use of AV access and reported 49% upper arm AVF, 26% lower arm AVF, and 25% AVG in the United States (89).

Several recent observational studies have directly compared AVF and AVG outcomes. Voorzaat et al. (114) reported that in The Netherlands, the overall AVF maturation failure rate was 24% for radiocephalic AVFs and 11% for upper arm AVFs, whereas the AVG maturation failure rate was 6%. Lee et al. (64) recently reported from the United States Renal Data System that at 6 mo, AVF maturation failure in the United States was 51% compared with AVG maturation failure, which was 45%.

Gap in Knowledge-Pathophysiological Understanding of AVF and AVG Dysfunction

After AVF creation, the vein is immediately exposed to high blood flow, high shear, and a high-pressure environment (45, 60, 62, 96). Successful AVF maturation requires the vein to adapt to the increased blood flow and shear stress after arterialization. This occurs through outward remodeling, which involves a combination of vasodilation and structural changes to the AVF vein, including a sustained increase in its luminal diameter and likely also an increase in its wall thickness (45, 60, 62, 96). During this adaptive remodeling period, the surgically and hemodynamically induced injury to the vessel translates into cellular activation and signaling involving cells [e.g., endothelial cells (ECs), smooth muscle cells (SMCs), myofibroblasts, and fibroblasts] within the adventitia, media, and intima of the AVF (Fig. 3) (45, 60, 62, 96). The cellular changes that occur within the AVF result in the regulation of local extracellular matrix (ECM) production, inflammation, growth factor production, and oxidant stress, as part of the remodeling process (Fig. 3) (45, 60, 62, 96). AVF maturation failure is hypothesized to result from insufficient outward remodeling and early intimal hyperplasia development (62, 96, 100). Although there has been advancement of knowledge in recent years, elucidating the molecular mechanisms that lead to intimal hyperplasia development and poor outward remodeling, our fundamental understanding of the “venous” environment of the AVF still remains limited.

Fig. 3.

Mechanisms of arteriovenous fistula (AVF) maturation failure. Shown are the proposed cellular mechanisms associated with fistula nonmaturation. Because of the formation of the arteriovenous fistula, the already very high blood flow shifts from a laminar to turbulent system, exercising high wall shear stress (WSS) on the venous cells. Endothelial dysfunction, damage, and deendothelialization of the intimal layer due to WSS occurs, thus, exposing the subendothelial extracellular matrix (ECM) directly to flow. The exposed subendothelial layer results in thrombus formation as well as the adhesion and activation of platelets and monocytes. Monocytes infiltrate the vessel and differentiate to macrophages, releasing proinflammatory cytokines, such as TNF-α, monocyte chemoattractant protein (MCP)-1, and IL-8. Meanwhile, vascular smooth muscle cell (VSMC) migration and proliferation are promoted by platelet-derived growth factor (PDGF) and TNF-α. Driven by cytokines such as transforming growth factor (TGF)-β, TNF-α, and PDGF, fibroblasts differentiate to myofibroblasts, which proliferate and excrete ECM components.

The most common etiology of the venous-anastomotic stenosis seen in AVG failure is intimal hyperplasia (52, 99). Similar to AVFs, the molecular mechanisms that result in this lesion after AVG placement remain poorly understood. In recent years, the roles of hemodynamic shear stress (56), anatomic configuration (55), and cellular phenotypes (117) within the venous intimal hyperplasia have been evaluated. At present, there are few, if any, effective therapies to prevent AVG dysfunction.

Unmet Clinical Need and Lack of Effective Clinical Therapies

The majority of our therapies to treat vascular access dysfunction (e.g., balloon angioplasty, stent grafts, and drug-coated balloons) have focused on the time period after AVF and AVG failure. To date, there are presently no effective therapies to “prevent” AVF maturation failure or AVG stenosis. The landmark multicenter randomized controlled trials have primarily focused on systemic therapies and have largely shown only little or marginal benefit (Table 1). At present, a recent study from the United States has reported that 2.8 billion United States dollars, which is ~12% of the entire Medicare budget, are spent for dialysis vascular access services (110). Furthermore, in this study evaluating AVF costs, a substantial amount of costs in the first year after initiating dialysis is attributed to patients with AVF failure and loss of primary patency (110). Thus, there remains an urgent clinical need to develop therapies to enhance AVF maturation and improve short- and long-term durability of these vascular accesses.

Table 1.

Landmark clinical trials in AVFs and AVGs

Study	Therapy	Primary Outcome	P Value	Secondary Outcome	P Value
DAC graft (n = 649) (32)	Dipyridamole/ASA (n = 321) vs. placebo (n = 328)	Loss of unassisted graft patency: 80% vs. 84%	0.02	Cumulative graft failure: 50% vs. 53%	NS
DAC fistula (n = 877) (30)	Clopidogrel (n = 441) vs. placebo (n = 436)	Thrombosis at 6 wk: 12.2% vs. 19.5%	0.018	Fistula nonsuitability: 61.8% vs. 59.5%	NS
Fish oil in AVG (n = 201) (68)	Fish oil (n = 101) vs. placebo (n = 100)	Loss of patency within 12 mo: 48% vs. 62%	0.06	1. Primary unassisted patency at 12 mo: 48% vs. 32%	<0.045
				2. Corrective interventions: 2.89 vs. 4.92 per 1,000 access days	<0.001
				3. Thromboses: 1.71 vs. 3.41 per 1,000 access/days	<0.001
Fish oil (n = 567) and ASA (n = 406) in AVF (48)	Fish oil (n = 284) vs. placebo (n = 283) and ASA (n = 203) vs. placebo (n = 203)	AVF failure: fish oil (47%) vs. placebo (47%) ASA (45%) vs. placebo (43%)	P = 0.78 P = 0.68	1. AVF abandonment within 12 mo: fish oil (19%) vs. placebo (22%) ASA (24%) vs. placebo (18%)	P = 0.43 P = 0.17 P = 0.90 P = 0.70 P = 0.81 P = 0.92
				2. Thrombosis of AVF within 12 mo: fish oil (22%) vs. placebo (23%) ASA (20%) vs. placebo (18%)
				3. Cannulation failure within 12 mo: fish oil (40%) vs. placebo (39%) ASA (38%) vs. placebo (38%)

ASA, aspirin; AVG, arteriovenous grafts; AVF, arteriovenous fistulas; DAC, Dialysis Access Consortium.

PATHOPHYSIOLOGY OF AV ACCESS DYSFUNCTION: ROLE OF VASCULAR MECHANICS

Role of Blood Flow and Pressure on Arteries and Veins in Health and Diseases

Blood vessel walls are constantly under mechanical stress, including fluid-wall shear stress, which is the frictional drag between the flowing blood and the vessel wall surface, and circumferential wall stress resulting from the vessel wall deformation under pressure. Physiological fluid-wall shear stress is 10–70 dyn/cm² in large arteries, which are under 5–10% cyclic circumferential stretch and under wall stress on the order of 10⁶ dyn/cm². This is higher than the venous environment, where large veins are exposed to fluid-wall shear stress at 1–5 dyn/cm² and wall stress on the order of 10⁵ dyn/cm², with minimal stretch (36, 83).

In the arterial circulation, fluid-wall shear stress and circumferential wall stress are well recognized to be major driving forces during remodeling of arteries, including the formation of atherosclerotic lesions. In areas of unidirectional, laminar, and high fluid-wall shear stress (e.g., straight segments of the arterial tree), the activated mechanotransduction pathways lead to a quiescent endothelial phenotype, release of nitric oxide (NO), and downregulation of proinflammatory and proliferative genes in ECs (18, 23, 25–27, 85), with limited intimal hyperplasia or atherosclerotic lesion growth (40, 53, 75). In contrast, disturbed, oscillatory, and low fluid-wall shear stress often occurs at sites of vascular bifurcation in native arteries or the graft-arterial anastomosis in bypass grafts. In these regions, upregulation of certain mechanotransduction pathways leads to decreased NO release and upregulation of proinflammatory, proproliferative, and prothrombotic genes in ECs (23, 25–27, 42, 44, 85), with increased intimal hyperplasia and/or accelerated atherosclerosis (17, 40, 53, 56, 75). Pressure-induced stretch renders the wall under circumferential wall stress. At the tissue level, increased blood pressure (14, 46) and wall stretch (16, 46) have been shown to induce arterial intimal hyperplasia and wall thickening; at the cellular level, wall stretch has been shown to stimulate the proliferation of arterial SMCs (65, 91, 118). In contrast, there has been a relative paucity of information relating mechanical stress to venous remodeling in its native environment. However, when native veins are used as arterial bypass grafts, they are exposed to arterial pressure and develop intimal hyperplasia, which is likely, in part, a response to the mechanical stress change in their environment.

Blood Flow and Pressure in AV Access and Vascular Remodeling in This Severe Environment

As a result of the reduced downstream resistance and the change of the blood flow direction at the anastomosis, AV conduits are exposed to a significantly increased blood flow and pressure gradient as well as disturbed flow in their anastomotic regions that are characterized by spiral flow, flow separation (or flow split), stagnation, recirculation zones (or vortexes and eddies), and extremely high or low shear stress in different areas (35, 58, 59, 78, 113). In the AVF setting, the regional AVF flow can be up to 20-fold higher than normal arterial flow in patients (70, 121, 126). Elevated blood flow is desired and is expected to induce flow-mediated dilation (FMD) and preserve the quiescent phenotype of ECs; in the context of AVF, a FMD-induced increase in the lumen diameter is interpreted as an attempt to return fluid-wall shear stress to pre-AVF levels (34, 39), as illustrated by Corpateux et al. (22). In contrast, disturbed flow in the anastomotic regions may stimulate proinflammatory, proproliferative, and prothrombotic responses of ECs. Elevated pressure-induced stretch and wall stress may stimulate SMC proliferation and migration.

Recently, researchers in the field of vascular access have made significant progress in advancing our knowledge and understanding of the detailed blood flow, pressure, fluid-wall shear stress, and circumferential wall stress profiles in both AVFs and AVGs. These advancements use medical imaging-based biomechanical modeling to simulate and characterize hemodynamics and/or wall mechanics in patient-specific AV accesses (15, 29, 43, 77, 106) and in animal models of AV access (59, 71, 112). Figure 4A shows a representative MR image of a human AVF, and it has been found that the carotid-jugular porcine AVF model (87) displays arteriovenous anastomotic hemodynamics very similar to that in a human AVF (Fig. 4B) (43). The graft-venous anastomotic hemodynamics in a carotid-jugular porcine AVG model (Fig. 4F) (88) is also very similar to that in a human AVG (Fig. 4C) (104). Thus, these animal models of AV access recapitulate at least a part of the key signatures of AV hemodynamics.

Fig. 4.

MRI and computational fluid dynamics (CFD) from human and animal arteriovenous conduits. A and D: MRI of a human arteriovenous fistula (AVF; A) and a mouse AVF (D). B and E: CFD results of a human AVF (B) and a porcine AVF (E). C and F: CFD results of a human arteriovenous graft (AVG; C) and a porcine AVG (F). [A and B: reprinted from Ref. 43 with permission from Elsevier; D: reprinted from Ref. 87 according to the BioMed Central license agreement; C: presented at the 2016 American Society of Nephrology Kidney Week Meeting (104); E and F: presented at the 2015 Summer Biomechanics, Bioengineering, and Biotransport Conference (88).]

Experimental and clinical studies have demonstrated the associative and/or causative relationships between fluid-wall shear stress and venous intimal hyperplasia in AV access (6, 10, 24, 67, 74, 84, 86, 95). Specifically, studies in porcine AVF models have reported that fluid-wall shear stress is inversely correlated with an average venous intima-medial thickness ratio (54), and low fluid-wall shear stress and high oscillatory shear index at early time points correlated to maximum intima-media thickness ratio (93). Blood pressure and wall stress can increase at least 10-fold or even higher compared those in native veins. Fluid-structure interface simulations have shown that the intimal hyperplasia-susceptible sites in a murine AVF model (Fig. 4D) (63) and in a human AVG (Fig. 4C) (104) are exposed to larger stretch and larger wall stress than intimal hyperplasia-resistant sites, but further experimentation is needed to establish causative relationships.

PATHOPHYSIOLOGY OF AVF MATURATION FAILURE: INTIMAL HYPERPLASIA DEVELOPMENT AND OUTWARD REMODELING

The pathophysiology of intimal hyperplasia in arteriovenous conduits for hemodialysis access is complex, with many influencing factors such as turbulent flow, compliance mismatch between the involved artery and vein, and vascular injury due to creation surgery and repetitive cannulation (62, 98). It is important to recognize that the AVF environment is a unique, nonphysiological condition. Although at a histological level, AVF failure shares some characteristics with coronary and peripheral vein graft disease, there is an enormous increase in the flow velocity and pressure gradient in the perianastomotic segment, which are unique and important stimuli that are thought to contribute to the poor functionality of AVFs. From an evolutionary perspective, veins are not designed to cope with these excessive mechanical stimuli in AVFs.

Intimal Hyperplasia Versus Outward Remodeling in AVFs

With regard to nonmaturation of AVFs, vascular stenosis has been traditionally attributed to intimal hyperplasia, whereas the contribution of impaired outward remodeling has attracted more attention (96). Furthermore, in AVFs, there are many biological factors involved as responses to the vascular injury that impact AVF maturation failure, such as inflammation, vascular SMC proliferation, matrix turnover, and EC damage that can affect the balance between intimal hyperplasia development and outward remodeling (Fig. 3). With angiography, these two phenomena are difficult to distinguish. Several intriguing studies have been executed that challenge the current paradigm that preexisting venous pathology is predictive or at least partly responsible for AVF nonmaturation and failure. The Hemodialysis Fistula Maturation (HFM) Study found no association between preexisting venous intimal hyperplasia and postoperative AVF stenosis at 6 wk (21), which is consistent with other smaller and single-center studies (73, 109). Interestingly, the HFM Study found that flow- and nitroglycerin-mediated dilation, assessments of arterial dilatory capacity, were positively associated with postoperative AVF flow and diameter at 6 wk (1). These results suggest that 1) arterial function, not venous pathology, is predictive of AVF outcomes and 2) vasodilation, which results from relaxation of vascular SMCs, may be one of the mechanisms associated with outward remodeling.

The ECM: a Focus on Collagen and Elastin

As outward remodeling is critical for AVF maturation, detailed knowledge about the composition and structure of the ECM and its response to vascular injury is pivotal to design new therapies, aimed to steer the vascular responses after AVF creation surgery. The most abundant vascular ECM components are collagen and elastin. Their degradation depends on the activation of proteases, including matrix metalloproteinases (38), which are activated by various stimuli, including an increase in flow (107).

A recent study investigated the association between preexisting medial fibrosis and AVF maturation (103). While arterial medial fibrosis was associated with a favorable increase in AVF diameter and blood flow as well as a decreased risk of AVF maturation failure, preexisting venous medial fibrosis did not associate with either outcome (103). In line with these observations, Martinez et al. (73) showed that postoperative venous fibrosis increases in patients regardless of preexisting levels and that only postoperative venous medial fibrosis is determinant for the stenotic potential of intimal hyperplasia in AVF. These data suggest that preexisting arterial fibrosis and postoperative venous fibrosis are important factors for AVF maturation failure.

Elastin is another major ECM component of the vascular wall. Wong et al. (120) used a murine model of AVF failure to explore the role of elastin in AVF remodeling. Enhanced venous outward remodeling was observed in AVFs created in haplodeficient tropoelastin knockout mice (120). These data suggest that elastic fibers impair outward remodeling in AVFs and support the hypothesis that elastin degradation might improve AVF maturation.

Lineage Tracing and Dual Role of SMCs

To date, the origin of neointimal cells responsible for venous stenosis within AVFs has not been elucidated completely (97). A variety of cells have been designated as contributors to intimal hyperplasia in AVFs, including venous SMCs, venous fibroblasts that differentiate into myofibroblasts, circulating vascular smooth muscle progenitor cells, and arterial SMCs. A recent study in a murine AVF model by Liang et al. (66) suggested that arterial SMCs are responsible for approximately half of the cells present in stenotic lesions in AVFs. The role of SMCs in AVF maturation has not been studied extensively. To allow the vein to increase in size substantially, it seems logical that SMC proliferation is needed to keep pace with the vessel expansion, as has also been described in the arterial setting (28). Therefore, the complete inhibition of SMC proliferation after AVF surgery may not translate to better AVF maturation. The dual role of SMCs within the setting of AVFs has recently been studied elegantly by Zhao et al. (125), who observed that differentiated SMCs predominantly contribute to medial wall thickening during venous maturation, whereas dedifferentiated SMCs contribute to intimal hyperplasia.

EXPLORING NOVEL GENES AND PATHWAYS IN AVF MATURATION FAILURE

Novel Genes

Elastin.

As described above, enhanced venous outward remodeling was observed in AVFs created in haplodeficient tropoelastin knockout mice (120). These data suggest that elastic fibers impair outward remodeling in AVFs and support the hypothesis that elastin degradation might improve AVF maturation. A phase 3 multicenter randomized controlled trial in 313 participants, evaluating the efficacy of recombinant elastase (PATENCY-1) (NCT02110901) delivered perivascularly over the AVF artery and vein immediately after AVF creation to enhance the maturation of radiocephalic AVFs, was completed in 2016 (61). The main findings from this study were that perivascularly delivered recombinant elastase improved use of radiocephalic AVFs for hemodialysis and secondary patency without adverse events compared with placebo (61). However, recombinant elastase did not statistically improve the primary outcome, which was primary unassisted patency (61). A second phase 3 multicenter clinical trial is currently being evaluated, the PATENCY-2 study (NCT02414841) in 600 patients. The coprimary outcomes in this study are secondary patency and AVF use for hemodialysis. The results from this study are expected in 2019.

Ephrin B4.

Ephrin B4 is an upstream activator of the phosphatidylinositol 3-kinase-Akt pathway that promotes cell migration and proliferation (108) and regulates essential cellular functions in vascular remodeling. Ephrin B4 is predominantly expressed in venous ECs and SMCs, whereas arterial cells are characterized by ephrin B2 expression (37). In murine and human experiments, Protack et al. (92) explored the role of ephrin B4 in AVF remodeling. Successfully matured human AVFs demonstrated increased ephrin B4 expression relative to normal veins from the same patients. In addition, partial deletion of ephrin B4 resulted in increased intimal hyperplasia in murine AVFs, whereas stimulation of ephrin B4 resulted in reduced wall thickness and improved patency of murine AVFs. Future studies should evaluate whether interventions aimed to enhance ephrin B4 activity might improve AVF durability in patients on hemodialysis.

Relaxin.

There has been increased interest in the role of vascular inflammation in AVF maturation. Animal studies have shown that the early phase after AVF creation is characterized by marked infiltration of macrophages and lymphocytes in AVF vessels and upregulation of proinflammatory cytokines (80, 117). Furthermore, clinical studies have suggested that this inflammatory response is harmful for AVF maturation, as plasma levels of C-reactive protein are inversely correlated with successful AVF maturation (51). Previous studies revealed that the proinflammatory cytokine monocyte chemoattractant protein-1 negatively influences vascular remodeling in murine AVFs (50). Relaxin is a hormone that acts on the vasculature via interactions with its receptor RXFP1, which results in vasodilatation, ECM remodeling, and decreased inflammation (49). A recent paper by Bezhaeva et al. (11) found that deletion of RXFP1 resulted in impaired venous outward remodeling, which was associated with increased vascular inflammation and impaired elastin degradation.

Toll-like receptor-4 signaling.

Another inflammatory signaling pathway that has recently been studied for its role in AVF remodeling is RP105, an endogenous Toll-like receptor-4 agonist. RP105 appeared to be highly expressed in stenotic lesions obtained from the venous outflow tract of AVFs. In addition, deletion of RP105 resulted in reduced infiltration of inflammatory cells, reduced proliferation of venous SMCs, and yet impaired outward remodeling in murine AVFs (12). Apparently, the end result of anti-inflammatory interventions in AVFs depends on the exact target site in the inflammatory cascade, which underlines the complex interplay between inflammation, SMC proliferation, and vascular remodeling in AVFs.

Stem Cells and Posttranscriptional Regulation

Stem cells.

Mesenchymal stem cells have been considered a promising therapy for treatment of vascular injury. Presently, there are few studies that have evaluated the role of mesenchymal stem cells in AVF development. In a recent study by Yang et al. (124), adventitial transplantation of human adipose tissue-derived mesenchymal stem cells delivered to the AVF vein of a murine model of AVF resulted in a decreased monocyte chemoattractant protein-1 gene expression, reduced venous intimal hyperplasia and cell proliferation, and prolonged retention of mesenchymal stem cells at the adventitia. Currently, this therapy is being evaluated in phase 1 clinical trials in dialysis patients at the time of new AVF creation (ClinicalTrials.gov identifier: NCT02808208).

MicroRNAs.

MicroRNAs (MiRs) are small (~22 nt in length) noncoding RNAs that are crucial regulators of gene expression, cell functions, and diseases (82). Several miRs have been implicated in AVF remodeling. Shang et al. (102) found that the serum miR-92a level was significantly higher in patients with CKD than in controls and that the serum miR-92a level increased with increasing chronic kidney disease stages and was negatively associated with estimated glomerular filtration rate. This group’s continuing effort has found that in a mouse AVF model, intimal hyperplasia was reduced by whole body knockout of miR-92a and by nanoparticles that encapsulated miR-92a inhibitors (105). Wang et al. (116) found that stenotic AVF tissues collected from patients receiving AVF revision surgery contained high levels of miR-155 and that AVF intimal hyperplasia was reduced in miR-155 whole body knockout mice. Lv et al. (69) performed a microarray study to compare differences in miR expression profiles between stenotic AVF vein samples collected from patients receiving AVF revision surgery and control vein samples collected from predialytic patients receiving AVF creation surgery in a limited number of patients and identified 33 miRs with markedly different expression levels between these two groups. Their bioinformatics analysis suggested that the MAPK pathway may play a role in the pathogenesis of stenotic AVFs (69). Future studies that investigate different miR profiles between mature and fail-to-mature AVF vein samples in larger cohorts should shed new light into the pathogenesis of AVF maturation failure. Finally, in the context of AVF restenosis, Wu et al. (123) found that serum miR-21 significantly increased within 2 days after angioplasty of stenotic AVFs, and a postangioplasty increase of serum miR-21 was associated with restenosis.

MAJOR CLINICAL TRIALS AND STUDIES IN AV ACCESS

Previous and Ongoing Clinical Trials: Why We Have Failed?

Clinical trials in AV access.

Presently, there are few effective therapies to significantly prevent AVF or AVG dysfunction. The landmark multicenter randomized controlled trials have largely shown no therapeutic benefit or only marginal benefit from systemic therapies (Table 1). Two studies from almost a decade ago from the National Institutes of Health Dialysis Access Consortium (DAC) evaluated antiplatelet agents (30, 32). In the DAC AVF study, whereas clopidogrel significantly reduced thrombosis at 6 wk in new AVFs, the failure of AVFs to mature was similar (nearly 60%) in both clopidogrel- and placebo-treated groups (30). In the DAC AVG study, dipyridamole modestly improved loss of AVG unassisted patency, which was the primary outcome, with no significant improvement in cumulative AVG patency (32). There have also been two randomized clinical trials evaluating fish oil therapy, one trial in AVFs and and one trial in AVGs. Recently, Irish et al. (48) reported that fish oil did not improve the primary outcomes, which were a composite of AVF thrombosis and/or abandonment and/or cannulation failure at 12 mo, or secondary outcomes, which were the individual components of the primary outcome. In AVGs, Lok et al. (68) showed that fish oil trended toward improving loss of AVG patency at 12 mo (primary outcome) but significantly improved primary unassisted patency at 12 mo, corrective interventions, thrombosis, and reduced cardiovascular events (secondary outcomes).

There is an ongoing clinical trial evaluating intravenous delivery of prednisolone encapsulated in liposomes in patients with new AVF creation (NCT02495662). Wong et al. (119) assessed the feasibility and efficacy of prednisolone encapsulated in liposomes to target the inflamed perianastomotic region in murine AVFs and observed that upon intravenous injection, liposomes accumulate in the perianastomotic area of AVFs, where they inhibited the inflammatory vascular response, resulting in enhanced venous outward remodeling. On the basis of these results, the LIPMAT trial was designed. This is a multicenter, double-blinded, placebo-controlled randomized controlled trial in The Netherlands, in which the efficacy of liposomal prednisolone to enhance radiocephalic AVF maturation is evaluated (115). The results of this trial are expected in 2019.

Modifying Hemodynamics

In the AVF field, researchers have investigated the effect of anastomosis angles on the severity of disturbed flow as well as the effect of surgical techniques or devices that aim at improving anastomosis flow. Numerical simulations of flow in radiocephalic AVFs with various anastomotic angles between the feeding artery and the draining AVF vein found that smaller anastomotic angles led to smaller disturbed flow areas (33). A novel surgical technique in the vascular anastomosis, called the piggyback Straight Line Onlay Technique (pSLOT), was designed to reduce the flow disturbance near the anastomosis and, in a single-center study, had less juxta-anastomotic stenosis and lower AVF maturation failure compared with control patients that received the traditional technique (13). VasQ (Laminate Medical Technologies) (20) and Optiflow (Bioconnect Systems, Ambler, PA) (41, 72, 81, 101) are external and internal, respectively, devices that are designed to minimize flow disturbance at the anastomosis and have been used in patients. VasQ is a Nitinol implant, externally surrounding the vein and hugging the artery near the junction site without being in contact with the blood flow (20). In contrast, Optiflow is an anastomotic connector made from highly nonthrombogenic siliconized polyurethane; it is placed inside the anastomosis and, thus, contacts the blood flow (19). Both devices appear to be safe in creating AVFs.

Why Do We Still Not Have Any Effective Therapies?

Why have the large clinical trials evaluating systemic therapies to prevent AVF and AVG dysfunction been shown to be ineffective? One possibility is that the systemic therapies tested to date have not targeted the appropriate biological pathways that lead to AVF and AVG dysfunction. Another possibility is that systemic therapies are not able to reach a sufficiently high local drug concentration in AVFs or AVGs to achieve a substantial beneficial effect at the site of vascular response. The pathology of AVF and AVG dysfunction is largely a localized area of constant vascular injury, which is the perianastomotic region in AVFs and vein-graft anastomosis in AVGs. Thus, local therapies that deliver drugs directly to the site of vascular injury (perivascularly delivered) may be the most effective approach (60, 62). There have been several recent multicenter randomized controlled trials evaluating local perivascularly delivered therapies at the time of AVF and AVG creation (Table 2). Several of these clinical trials have been unsuccessful due to early termination of the studies following adverse events and poor recruitment. Two phase III randomized clinical trials are evaluating perivascularly delivered elastase (Proteon Therapeutics, Waltham, MA) in new radiocephalic AVFs, as discussed above. One phase III clinical trial is evaluating perivascular delivered sirolimus in new AVFs (Vascular Therapies, Cresskill, NJ). The final results from these randomized controlled trials are expected in 2019.

Table 2.

Clinical trials of local drug delivery approaches in AVFs and AVGs

Company	Clinical Trial	Therapy/Device	Outcome	Status
Angiotech Pharmaceuticals	NCT00448708	Vascular wrap paclitaxel-eluting mesh (AVG)	Time to loss of target site primary patency	2007–2009 (unsuccessful)
Shire	NCT01806545	Allogeneic human aortic endothelial cells cultured in a gelatin matrix (Gelfoam) (AVF/AVG)	Fistula maturation at 12 wk based on ultrasound	2013–2014 (unsuccessful)
Proteon Therapeutics	NCT01305824	Elastase (AVF)	1. Primary AVF patency	Completed
			2. Secondary AVF patency	Completed
Vascular Therapies	NCT02513303	Sirolimus-eluting collagen implant (AVF)	Fistula suitability at 6 mo	Ongoing phase III RCT in AVF
Proteon Therapeutics	NCT02110901	Elastase (AVF)	1. AVF primary unassisted patency	Ongoing phase III RCT in radiocephalic AVF
			2. AVF cumulative patency

AVG, arteriovenous grafts; AVF, arteriovenous fistulas; RCT, randomized control trial.

The HFM Consortium: What Have We Learned So Far?

The two primary objectives of the HFM Study were 1) identifying predictors of AVF maturation failure and 2) elucidating the mechanisms of AVF maturation failure (31). Currently, the HFM Study has the largest biorepository and database for investigating the natural history of AVF development.

To date, the most important biology knowledge from the HFM Study that could potentially be used for prediction is from the preoperative vascular function tests it performed. Specifically, FMD and nitroglycerin-mediated dilation (NMD) of brachial artery measured shortly before AVF creation were positively associated with AVF blood flow rate and diameter at 6 wk (1). These results highlight the importance of vasodilation in AVF development processes within 6 wk after AVF creation and suggest that the functionalities of ECs (based on FMD results) and vascular SMCs (based on NMD results) are both important. However, neither FMD or NMD was associated with clinical AVF maturation. This has been partly attributed to nonbiological factor-related processes of care, as clinical AVF maturation is affected not only by postoperative biological changes of the AVF, leading to changes in diameter and blood flow, but also by center-specific factors, such as AVF cannulation experience and practices and other processes of medical care (3).

CONCLUSIONS AND FUTURE PERSPECTIVES: A MORE PERSONALIZED APPROACH

While AVF use has increased and AVG use has decreased worldwide, on the basis of epidemiological data from DOPPS (90), AVF maturation failure and AVG stenosis remain a significant clinical problem for patients on hemodialysis. There still remain no effective therapies to prevent and treat both AVF maturation failure and AVG stenosis in patients on hemodialysis. Going forward, how do we approach and address this problem?

First, the most important decision for successful short- and long-term AVF and AVG use is the proper choice of the optimal vascular access for each patient on hemodialysis. Thus, it is crucial to consider each patient’s ESRD life plan, which will guide the individual vascular access “life plan” (122). Equally important, we also need to consider a more patient-centered approach [“Right access, for the right patient, at the right time, for the right reason” (47)] (Fig. 5)−not a one-size fits all approach (79)−that incorporates life expectancy and unnecessary surgical and interventional procedures and emphasizes overall improvement in patient quality of life.

Fig. 5.

Personalized approach to optimizing vascular access outcomes. A future approach to improving vascular access outcomes will require an individualized approach to vascular access selection and the selection of future biological therapies.

Second, there remain no effective therapies to treat or prevent AVF maturation failure and AVG stenosis. However, recently, there has been promising advancements in basic and clinical research to help better elucidate the pathophysiology of AVF maturation failure and AVG stenosis, which will hopefully lead to better therapeutics. Advancements in identifying biological and physiological risk factors that are associated with future AV access dysfunction may help guide providers with vascular access selection. However, many important questions remain to be answered, such as the role of vascular health in AVF/AVG dysfunction and the origin of cells in intimal hyperplasic lesions. Innovative advancements in technology in areas such as genomics, genetics, molecular biology, nanotechnology, and bioengineering will provide the technological tools to help answer these questions (Fig. 5).

In conclusion, while understanding the mechanistic aspects of AVF and AVG dysfunction is critical and will help advance potential therapeutic development, this knowledge alone is not sufficient to predict which access will mature and function properly for individual patients. There are also process-of-care factors, particularly, patient selection, that impact successful AVF and AVG use and durability, independent of biology. Thus, an ideal future strategy to optimize vascular access outcomes will be to incorporate a more personalized approach that combines processes care and patient selection [“Right access, for the right patient, at the right time, for the right reason” (47)] with future available biological therapies (Right therapy, for the right patient, at the right time, for the right reason) (Fig. 5).

GRANTS

T. Lee is supported by National Institutes of Health (NIH) Grants R44-DK-109789 and R01-HL-139692 as well as Department of Veterans Affairs Merit Award I01BX003387. Y. T. Shiu is supported by NIH Grant R01-DK-100505 and Department of Veterans Affairs Merit Award I01BX004133. J. Rotmans is supported by Dutch Organization of Scientific Research VIDI Grant 016.156.328.

DISCLOSURES

T. Lee is a consultant for Proteon Therapeutics, Merck, and Boston Scientific. J. Rotmans is supported by research grants from Proteon Therapeutics and Enceladus Pharmaceuticals.

AUTHOR CONTRIBUTIONS

Y.-T.S., J.I.R., W.J.G., D.B.P., and T.L. prepared figures; Y.-T.S., J.I.R., and T.L. drafted manuscript; Y.-T.S., J.I.R., W.J.G., and T.L. edited and revised manuscript; Y.-T.S., J.I.R., W.J.G., D.B.P., and T.L. approved final version of manuscript.