Graft Modification Strategies to Improve Patency of Prosthetic Arteriovenous Grafts for Hemodialysis

Introduction

Recent trends in hemodialysis (HD) include an increase in elderly patients requiring dialysis, increased prevalence of diabetes mellitus, increasing proportion of long-term dialysis patients and use of high efficiency dialysis. This changing demography of the dialysis population has led to an increase in the proportion of patients requiring more complex vascular access modalities for long-term HD. Consequently, the use of non-autogenous or prosthetic arteriovenous grafts (AVG), which are considered as secondary access modalities, has increased steadily over the last decade. However, AVGs remain to have an inferior primary and secondary patency compared to autogenous AVFs (1–4). Despite initial high blood flow rates following placement, AVGs are frequently faced with the problem of continuous flow rate decline because of neointimal hyperplasia (NIH) formation at the (graft-vein) anastomosis site leading to stenosis, compliance mismatch, procoagulant state and, eventually, occlusion.

For the past 40 years, the vast majority of AVGs have been constructed using expanded polytetrafluoroethylene (ePTFE). In addition to ePTFE, polyurethane is used for prosthetic AVGs. This material has the benefit of early postoperative cannulation; however, patency rates have been the same as for ePTFE grafts. Undoubtedly, there is a clear need to improve patency rates of prosthetic AVGs and different strategies have emerged to either improve the graft materials available or to produce grafts made of novel materials. This review aims to describe current strategies and future perspectives on the modification of available graft materials to improve AVG patency.

Graft Geometry and Flow Modulation

Wall shear stress (WSS) and turbulent flow with oscillating flow patterns are known pathogenic factors in the development of NIH in AVGs. Several methods have been developed to manipulate graft hemodynamics in order to establish a favorable laminar flow pattern, which in theory should reduce the development of NIH.

The Venaflo™ graft is a PTFE graft with a cuff at its venous anastomotic side. It has been thoroughly tested in computational fluid dynamics (CFD), animal and human studies. In a CFD study using a pulsatile circuit in a silastic model, the Venaflo™ graft created a significantly lower WSS and a more homogenous flow pattern at the graft-vein anastomosis compared to a straight graft (5). An animal study supported this potentially beneficial hemodynamic effect on the development of NIH; however, a high incidence of pseudo-intima formation was observed within the cuff leading to early graft dysfunction (6). Several retrospective analyses and randomized clinical trials (RCTs) comparing either the Venaflo™ graft or a PTFE graft with autologous venous cuff with a standard PTFE graft have been conducted; however, with modest results (7–12). Although some beneficial effects have been observed, the Venaflo™ graft did not seem to increase graft patency compared to control grafts. In one study, the Venaflo™ graft demonstrated a decrease in NIH formation (12). The lack of a beneficial effect on patency of the Venaflo™ graft can possibly be attributed to pseudo-intima formation in its cuff, which has been described in a Venaflo™ porcine model (6). It is hypothesized that a flow separation caused by the large mainstream-to-wall angle of the expanding cuff favors thrombus formation (6, 13). To solve the issue of the large mainstream-to-wall angle, the Bi-flow graft was developed, which is a diffuser graft with a cuff separated into two flow channels (3). In a CFD model comparing the Bi-flow graft to the Venaflo™ graft, the Bi-flow graft with a small cuff showed a favorable flow pattern with a decrease in WSS and little flow separation (13). Further animal studies have to reveal if the Bi-flow graft can prevent pseudo-intima formation as well as NIH to increase AVG patency.

Another approach of modifying graft geometry to create favorable hemodynamics is the tapering of an AVG at the graft-artery anastomosis. The aim of this approach is to increase the resistance and thereby decrease the flow within the graft and subsequently the WSS at the graft-vein anastomosis. This would putatively lead to a reduction in NIH formation. CFD studies comparing the hemodynamics of tapered grafts with that of straight grafts revealed favorable flow patterns with lower WSS and less disturbed flow at the graft-vein anastomosis and the outflow vein in the tapered grafts (14–16). A potential pitfall of tapered grafts is that the high resistance at the graft-artery anastomosis may cause severe hemodynamic stress at the graft-artery anastomosis eventually leading to partial hemolysis (16). Tapered grafts have also been thoroughly tested in a clinical setting with different results. An RCT evaluating 4-7 mm tapered grafts, for instance, showed no beneficial effect on graft patency over the standard graft (17). On the other hand, another RCT showed that the patency of a 6-8 mm tapered graft was superior to that of a non-tapered graft (18). Further in vivo and CFD studies comparing different tapering methods and sizing are necessary to evaluate which tapering is the most beneficial.

Instead of aiming for a reduction in flow, helical grafts have been developed to establish a more natural, laminar flow pattern by introducing a helical geometry resembling that of native arteries (19). Hypothetically, this flow pattern should reduce the areas exposed to high WSS and thereby impair the development of NIH (20). Caro et al first tested a helical graft configuration in a porcine model and showed that it can reduce pathological changes in the graft but the flow was not evaluated in their experiment (21). Nevertheless, CFD studies showed that the helical geometry may reduce WSS at the graft-vein anastomosis, reduce flow separation, and create a stable flow pattern (20, 22). All these factors seem to play a role in NIH development. Moreover, the pre-clinical evaluation of helical grafts is still inconclusive about its actual effect on NIH as few in vivo studies are available. Huijbregts et al have first tested a helical graft in a clinical pilot study (19). Here again, the patency was similar to that of a conventional ePTFE graft. Angiographic analysis of the implanted grafts revealed that the helical structure diminished after implantation, which could explain the unexpectedly low patency. A possible solution to this problem might be the C-flex connector, a short helical connector at the graft-vein anastomosis (22). However, pre-clinical evaluation of the C-flex connector is still necessary.

Compliance mismatch between the graft and the connected vein has long been postulated to have a causal link with NIH development at the graft-vein anastomosis. In vitro modeling studies testing grafts with a longitudinal stretch component demonstrated favorable hemodynamics, namely lower wave amplitudes and slower wave conduction (23, 24). Gessaroli et al studied a stretch graft, which, compared to the previously studied stretch grafts, not only had a longitudinal but also a radial stretch component thereby mimicking the properties of native vessels (25). In a porcine model, a significant decrease in NIH was seen in the stretch graft group if compared to the non-stretch control group. These results indicate that tackling the compliance mismatch between vessel and graft might still be a viable approach to decrease NIH in AVGs.

Recently, a modular anastomotic valve was designed to normalize venous flow between dialysis periods by isolating the graft from the circulation between dialysis periods (closed position) and enabling vascular access during dialysis (open position). This valve consequently decreases flow disturbances and the overall exposure of the graft-vein anastomosis to the high arterial flow, which theoretically could attenuate NIH progression (26). However, currently only CFD studies are available.

Graft Drug Coatings

Coating foreign bodies prior to implantation into the human body to increase their biocompatibility, decrease the infection rate, or increase their patency has been successfully applied in many different instances. Applying drug coatings on AVGs to reduce neointimal hyperplasia or thrombotic occlusion is the primary focus. New agents as well as new delivery methods are continuously being developed and tested by various research groups.

Heparin coating of vascular grafts to reduce their thrombogenicity has long been proposed and has frequently been tested in the clinical setting (27–29–31). However, pre-clinical literature studying heparin-coated AVG in animal models is scarce. In a canine arterial bypass model, heparin-coated vascular grafts improved graft patency and decreased its thrombogenicity (32). Although not tested in an AVG animal model, these results suggest the potential of heparin-coated grafts to increase graft patency by inhibiting intra-luminal thrombus formation. Two retrospective clinical analyses comparing primary and secondary patency of heparin-coated AVGs to those of standard AVGs, found that there was no difference in patency rates (27, 31). Heparin-coated AVGs even increased the need for thrombectomies (31). However, it has to be considered that a selection bias might have been introduced in these studies as patients receiving the heparin-coated grafts were usually more morbid than those receiving the control un-coated graft. Moreover, in one study half of the control group received tapered grafts whereas no tapered grafts were used in the heparin-coated group eventually introducing a bias. In contrast, a RCT conducted to compare heparin-coated grafts and non-coated grafts showed a positive trend on patency in favor of heparin-coated grafts, and a significantly lower occurrence of thrombus formation (29). Moreover, in another RCT of femoro-popliteal bypass grafts, heparin-coated grafts had a significantly higher primary and secondary patency, confirming the potential benefit of heparin coatings (28). Therefore, as current clinical results of heparin-coated AVGs are conflicting, a larger RCT with stricter statistical criteria should be performed to provide the definite evidence on the potential benefit of heparin coating.

Another agent being tested as a candidate for AVG coatings is paclitaxel. Paclitaxel is a chemotherapeutic agent used in the treatment of cancer. It elicits a cytostatic effect on proliferating cells. Different from heparin coatings, paclitaxel coatings aim to increase the AVG patency by inhibiting NIH. It has already been shown to successfully inhibit coronary stent restenosis after percutaneous coronary interventions with paclitaxel-eluting stents, justifying this rational (33, 34). Several pre-clinical studies evaluating the potential of paclitaxel-coated AVGs have been published. Lee et al tested a paclitaxel-coated AVG in a porcine model (35). To coat the grafts, a simple dipping method was used leading to a diffuse coating of the external and internal surface of the graft. In the paclitaxel-coated grafts a significant decrease in NIH was seen compared to the control group. However, the drug release pattern was characterized by an initial burst of paclitaxel entering the systemic circulation early after implantation and it continued to do so for months at lower concentrations. Paclitaxel is a toxic compound. The high systemic paclitaxel levels caused by the initial burst could cause paclitaxel-induced side effects. Additionally, the paclitaxel coating of the outer AVG surface inhibits myofibroblast infiltration, which may reduce graft adherence to the surrounding tissues and thereby interfere with graft ingrowth causing a delay in cannulation and therefore increase the risk of AVG dysfunction independent of NIH or thrombosis (36).

Numerous delivery methods have been developed that counteract the shortcomings of the dipping method and preserve the promising effect of paclitaxel on NIH. Baek et al simply modified the dipping method by restricting the coating to the terminal end of the AVG and reducing the applied dose to minimize systemic paclitaxel exposure (37). However, the results of this porcine study showed that the terminal coating suppressed NIH formation less effectively than completely coated grafts. Unfortunately no measurements of systemic paclitaxel levels were performed to analyze whether the terminal dipping method decreased the systemic levels of paclitaxel. Moreover, the terminal coating seemed to impair graft-vein anastomosis maturation supporting the hypothesis that outer surface paclitaxel-coating may impair graft ingrowth.

Another promising approach of paclitaxel drug delivery is the application of a drug-containing perivascular polymer wrap around the graft-vein anastomosis (38, 39). Masaki et al demonstrated in a dog model that a perivascular polymer wrap (ReGel [MacroMed, Inc.]) containing paclitaxel applied to the outer surface of the graft-vein anastomosis could prevent NIH without leading to detectable systemic paclitaxel levels (38). In a porcine model testing another polymer wrap loaded with paclitaxel applied to the graft-vein anastomosis, similar promising results were obtained (39). While the release pattern of paclitaxel from the polymer wrap is also characterized by an initial burst followed by a continuous release for several months, no systemic levels of paclitaxel were detected in either study. However, the problem of interference with graft-vein anastomosis maturation is also present here (39).

To eliminate the interference with graft-vein anastomosis maturation, paclitaxel coatings restricted to the luminal surface of AVGs were developed and evaluated (36, 40). However, as seen with the grafts processed with the dipping method, paclitaxel coating of the luminal surface again lead to a systemic paclitaxel exposure. Recently, an elegant solution to both these problems was proposed (40). To control the luminal release and thereby reduce systemic paclitaxel exposure, Lim et al developed paclitaxel-containing nanoparticles that they coated only on the luminal surface of an AVG. The in vitro release pattern was characterized by a lower initial burst and a longer continuous release of paclitaxel compared to the AVGs processed with the dipping method. Its restriction to the luminal surface and the favorable release pattern makes the nanoparticle-based delivery method a promising approach for paclitaxel drug delivery. However, further in vivo evaluation is necessary to evaluate its effect on NIH. Despite of the promising results shown in most of the preclinical studies with paclitaxel-coatings, several issues remain to be resolved before clinical application can be considered. More studies evaluating the doses of paclitaxel necessary to inhibit NIH are necessary, since current studies often use different doses thereby impairing comparability.

Sirolimus is another anti-proliferative agent, which has been proposed as a potential coating agent for AVGs. Both the dipping method as well as a perivascular polymer wrap containing sirolimus were evaluated and showed an inhibition of NIH in porcine models, though to a lower extent then paclitaxel (41, 42). Regrettably, the issue of impaired graft adhesion to the surrounding tissue observed in the paclitaxel-coated grafts was also observed in sirolimus-coated grafts. Recently, a first-in-human study was performed with a sirolimus-containing perivascular collagen polymer. Perivascular implantation of the polymer during graft surgery safely delivered sirolimus to the vascular wall and systemic sirolimus levels remained sub-therapeutic for immunosuppression, justifying further RCTs to test the efficacy of sirolimus-coating therapies (43).

Various other compounds are being tested as possible graft coatings to improve AVG patency. A phosphorylcholine (PC)-coating, a phospholipid found on the outer surface of red blood cells, potentially increases the biocompatibility of AVGs (44). In an AVG dog model, use of a PC-coating reduced formation of NIH and thrombogenicity of the graft. An in vitro study also showed that a PC-coating reduces thrombogenicity of artificial surfaces with blood contact (45). Despite these promising results, PC-coatings were not further investigated in the setting of AVGs.

In another interesting first-in-human study, application of PRT-201, a recombinant human type I pancreatic elastase, to the venous anastomosis of AVGs was evaluated (46). Pre-clinical animal studies have indicated that PRT-201 at high doses causes persistent vasodilation by fragmenting elastin in the adventitia of arteriovenous fistula. It was postulated that at lower doses PRT-201 could inhibit NIH by partial fragmentation of the adventitia and subsequent inhibition of myofibroblast migration. In this phase VII clinical trial the safety of low- and high-dose PRT-201 was demonstrated and a trend towards improved secondary patency was seen in the low-dose group (46). Larger randomized studies are necessary to test the efficiency of different doses of PRT-201.

Graft Surface Modification: The Future Perspective

While the above-described approaches to increase AVG patency are promising in theory, clinical results are often modest. Several issues remain to be solved in order to optimize them in a way that they can drastically increase AVG patency. An issue all these modified AVGs have in common is the lack of a luminal, functional endothelial lining. An endothelial lining has been shown to have several beneficial effects on graft hemocompatibility by showing vasoactive properties and decreasing thrombogenicity and NIH formation (47). Endothelial coverage does, to some extent occur in PTFE and polyurethane grafts; however, it is usually restricted to the first 1-2 cm of the graft distal and proximal to the anastomosis (48). The remaining luminal surface remains uncoated. Numerous exciting methods have recently been developed to facilitate endothelialization of the luminal surface of vascular grafts. The predominant focus currently lies on surface modifications using micro- or nanotopography or chemical surface modifications.

Plasma treatment, a promising chemical surface modification method, aims to modify the surface by subtle but homogenously distributed chemical reactions that increase the hydrophilic properties of the graft surface and thereby decrease its interaction with the circulating blood. Whereas regular coatings, as seen with the luminal paclitaxel coating, dissolve after prolonged exposure to arterial shear stress, modifications obtained by plasma treatment are integrated within the surface and thus resistant to shear stress. Interestingly, plasma treatment has been shown to significantly increase endothelialization of artificial grafts (49).

Nanotopography on the other hand has also been postulated to facilitate the adhesion of endothelial cells to the graft surface and thereby stabilize the endothelial layer after formation. Several methods of nanotopography are available, all having their advantages and disadvantages. Photolithography, electron-beam lithography, soft lithography and electrospinning are the most frequently used nanotopographic techniques. All aim to create a nanopattern on the surface of a graft that triggers cues in endothelial cells, which facilitates their adhesion (50).

Various clinical studies have already been carried out with electrospun polyurethane AVGs (AVflo™; Nicast Ltd, Israel). However, instead of aiming to increase endothelialization on the inside of the graft, the AVflo™ graft is electrospun on the outer surface to facilitate graft maturation and subsequently early cannulation. The first clinical studies show that early cannulation is feasible and patency comparable to other prosthetic grafts (51, 52).

Chong et al elegantly summarized the current advancement in micro- and nanopatterning in the field of vascular graft research (50). Although not specifically focusing on AVGs, their review gives a valuable insight in this fascinating field and its possibilities. The same research group also developed an intriguing approach of modifying a vascular graft with a combination of plasma treatment and nanotopography to induce endothelialization of the luminal graft surface. They were able to demonstrate that plasma treatment greatly enhanced endothelialization, whereas nanotopography appeared to increase the adhesive capacity of the endothelial layer making it more resistant to the shear stress of arterial flow (49).

Finally, the recent development of randomized libraries of surface topographies can become an essential tool to unravel the interplay between (endothelial) cells and graft surface topography and to find improved material surfaces (53).

Conclusion

There remains a subset of patients who are depending on prosthetic AVG for vascular access for long-term hemodialysis because an autologous AVF cannot be created or maintained owing to their age, comorbid diseases such as diabetes, cardiac or peripheral vascular disease or multiple previous failed accesses. To increase the poor patency of AVG, when compared with autologous AVF, several strategies have been explored to improve the functionality of available prosthetic graft materials.

In the area of optimization of flow hemodynamics, despite the theoretical benefit of creating a favorable flow pattern at the graft-vein anastomosis, the clinical evaluation of flow-modifying grafts has not shown great benefit above standard AVGs on graft patency. Several problems remain to be resolved to effectively modify hemodynamics in a way that substantially promotes graft patency.

Promising approaches to increase AVG patency by drug-coatings are being tested, but here also many issues remain to be resolved before most of them can be tested in a clinical setting. The major issues are the variation in dosages used for the coatings, the potential systemic side effects, and the impaired maturation or ingrowth of coated AVGs.

The captivating results of graft surface technology on endothelialization of graft materials show the large potential of this fascinating approach and its possible applications. More exciting developments in these techniques and their combinations, such as surface modification combining both topographic changes, using randomized libraries of surface topographies, and supramolecular chemistry, on both inner and outer sides of grafts, will definitely be introduced in the near future and have the potential to open a new chapter in the field of vascular graft research, including AVG development for hemodialysis vascular access.