Stent Graft Outcomes in Dialysis Circuits: A Review of Clinical Trials and Literature

Abstract

End-stage renal disease is a prevalent medical condition causing significant problems as the majority of patients are being treated with dialysis. While dialysis provides a means to compensate for the renal impairment, arteriovenous grafts and fistulas are often complicated by neointimal hyperplasia, loss of primary patency, and graft failure. Over the last decade, stent grafts have emerged as an alternative to balloon angioplasty and bare metal stents for revascularization in the event of arteriovenous graft occlusion or stenosis. Several randomized controlled trials have investigated the efficacy of stent grafts compared with balloon angioplasty for improving the function and durability of dialysis circuits. In this review, we present a comprehensive review of clinical trials. Stent grafts appear to offer increased primary patency over balloon angioplasty in the treatment of dialysis circuit stenosis. However, there is generally no significant difference between the two treatment modalities on other clinically relevant measures like complication rates and overall survival.

Background

End-stage renal disease (ESRD) is a challenging medical problem that currently affects over 786,000 people in the United States. ¹ ESRD encompasses kidney failure and is primarily caused by diabetes and hypertension in U.S. adults. In 2018 alone, more than 130,000 patients in the United States began treatment for ESRD. Treatment typically involves dialysis or kidney transplant. Approximately 71% of ESRD patients are currently being treated with dialysis, and approximately 29% are alive with a kidney transplant. While hemodialysis provides a means to compensate for kidney failure, it is often plagued with complications related to patency of the vascular circuit.

Hemodialysis Access Patency Rates

Synthetic arteriovenous grafts and arteriovenous fistulas are mainstays of hemodialysis access. However, grafts and fistulas are often complicated by neointimal hyperplasia. In patients with arteriovenous grafts, published patency rates are approximately 50% at 1 year. ^{2 3 4} Arteriovenous graft stenoses often occur in the venous anastomosis. Arteriovenous fistulas may be less prone to stenosis and circuit failure than grafts, showing primary patency rates of approximately 60% at 1 year with primary failure rates of 23% within 72 hours of fistula creation. ⁵

Maintaining Patency of the Hemodialysis Circuit

The current standard endovascular treatment for maintaining patency within a dialysis circuit is balloon angioplasty. ^{6 7} Although balloon angioplasty can effectively open a stenosed arteriovenous graft or fistula, the durability of the treatment effect and the need for multiple reinterventions has become a topic of discussion among experts in the field.

Over the last decade, stent grafts have emerged as an alternative therapy for arteriovenous graft occlusion or stenosis. Several randomized controlled trials have been completed investigating the efficacy of balloon angioplasty compared with stent graft to improve functionality of dialysis circuits. In this review, we present a comprehensive review of trials completed on this subject to date.

Clinical Trials

FLAIR Trial

In 2010, Haskal et al ⁸ published their landmark study comparing stent graft versus balloon angioplasty for failing dialysis access grafts. This prospective, multicenter, randomized controlled trial included 190 dialysis patients with a stenosed venous anastomosis, and investigated the efficacy of balloon angioplasty alone compared with balloon angioplasty plus placement of a FLAIR endovascular stent graft (Bard Peripheral Vascular, Arizona). ⁸ They found that the stent graft group had significantly greater incidence of patency of the treatment area at 6 months when compared with balloon angioplasty alone (51 vs. 23%, p  < 0.001). The stent graft group had significantly greater incidence of primary patency of the access circuit (38 vs. 20%, p  = 0.008) as well as significantly greater incidence of freedom from subsequent intervention at 6 months (32 vs. 16%, p  = 0.03). Haskal et al ⁸ also found that the incidence of binary restenosis at 6 months was greater with balloon angioplasty alone compared with balloon angioplasty plus stent graft (78 vs. 28%, p  < 0.001). Restenosis was the only adverse event that occurred significantly more frequently in the balloon-angioplasty group—adverse events at 6 months were otherwise equivalent between the two groups. The authors concluded that the addition of stent grafts to balloon angioplasty for venous anastomotic stenosis in hemodialysis circuits improved long-term patency and lowered incidence of reintervention when compared with balloon angioplasty alone. ⁸

Viabahn Stent Trial

After Haskal et al's work ⁸ was featured in the New England Journal of Medicine in 2010, the following three relevant trials were published in 2016. Vesely et al conducted a prospective, multicenter trial published in the Journal of Vascular Surgery . This study randomized 293 adult patients with evidence of upper extremity prosthetic hemodialysis graft thrombosis or dysfunction to receive either a GORE VIABAHN Endoprosthesis stent graft (W.L. Gore & Associates, Delaware) or balloon angioplasty. ⁹ Prior to enrollment in the study, any patient who had a qualifying secondary stenosis received angioplasty with the goal of less than 30% residual stenosis as criteria for enrollment. Patients with thrombosed grafts underwent percutaneous thrombectomy or thrombolysis prior to angioplasty. Vesley et al ⁹ found that, at 6 months, the stent graft group had significantly greater target lesion primary patency than the balloon angioplasty group (intent to treat, 51.6 vs. 34.2%, p  = 0.006). The stent graft group also had an increased median time from the index procedure to reintervention by an additional 95 days compared with the balloon angioplasty group (203 vs. 108 days). Patients with dysfunctional grafts had higher target lesion primary patency compared with patients with thrombosed grafts, regardless of treatment modality. Among patients with thrombosed grafts, treatment of venous anastomotic stenoses with stent grafts provided an increased 6-month target lesion primary patency when compared with angioplasty (36.1 vs. 23.5%). The authors concluded that placement of a stent graft was superior to balloon angioplasty with regard to target lesion primary patency 6 months after treatment for venous anastomotic stenoses of dysfunctional and thrombosed prosthetic hemodialysis grafts. ⁹

RENOVA Trial

In 2016, two clinical trials were published on the subject in the Journal of Vascular and Interventional Radiology . Haskal et al ⁸ followed up their initial work with the Prospective, Randomized, Concurrently-Controlled Post-Approval Study of the FLAIR Endovascular Stent Graft (RENOVA). This multicenter trial enrolled 270 patients at 28 sites with ≥50% anastomotic stenosis of their malfunctioning arteriovenous graft, and randomized them to receive either FLAIR-expanded polytetrafluoroethylene stent graft (Bard Peripheral Vascular, Arizona) or balloon angioplasty alone. ¹⁰ All stenoses were dilated prior to implantation of devices. Seventy-four patients died during this study with no significant difference between treatment groups. The investigators found that, at 12 months, the stent graft group achieved significantly greater treatment area primary patency (47.6 vs. 24.8%, p  < 0.001), access circuit primary patency (24 vs. 11%, p  = 0.007), and index of patency function (5.2 months/intervention ± 4.1 vs. 4.4 months/intervention ± 3.5, p  = 0.009) when compared with balloon angioplasty alone. At 24 months, the stent graft group maintained its significant improvements in treatment area primary patency (26.9 vs. 13.5%, p  < 0.001) and access circuit primary patency (9.5 vs. 5.5%, p  = 0.01) compared with balloon angioplasty alone. Similar to the initial FLAIR trial, ⁸ restenosis was the only adverse event that was significantly different between the two groups, with 82.6% of patients in the balloon angioplasty group requiring reintervention compared with 63% of patients in the stent graft group ( p  < 0.001). ¹⁰ Haskal et al ¹⁰ concluded that the stent graft provides a greater than twofold advantage in treatment area patency and overall access patency compared with balloon angioplasty, and this advantage is sustained at 2 years.

RESCUE Trial

Falk et al published their Randomized Study of the Fluency Plus Endovascular Stent Graft (Bard Peripheral Vascular, Arizona) in the Treatment of In-Stent Restenosis in the AV Access Venous Outflow Circuit (RESCUE) in the Journal of Vascular and Interventional Radiology in 2016. This multicenter trial enrolled 275 patients at 23 sites with the purpose of investigating the efficacy of stent graft following balloon predilation for the treatment of in-stent restenosis in hemodialysis access circuits compared with percutaneous transluminal angioplasty (PTA) alone. ¹¹ Enrollment criteria for the study included in-stent stenosis >50% in bare metal stents placed at the venous outflow of a mature fistula or arteriovenous (AV) access graft. Patients with concomitant thrombosis at the treatment site, or stenosis at the elbow, cannulation zone, cephalic arch, or superior vena cava, were excluded from this study. All additional stenosis >50% that were at >3 cm away from the treatment area were treated with a target residual stenosis of <30% prior to enrollment in the study. Falk et al ¹¹ found that, at 6 months, access circuit primary patency was significantly higher in the stent graft group compared with PTA alone (18.6 vs. 4.5%, p  < 0.001). The stent graft group also benefitted from significantly improved treatment area primary patency at 6 months when compared with PTA alone (66.4 vs. 12.3%, p  < 0.001). The stent graft group's improvement in access circuit primary patency was maintained at 12 months, and improvement in treatment area primary patency was maintained at 24 months. The authors concluded that treatment with stent graft yielded improvement in both access circuit primary patency and treatment area primary patency in dialysis patients presenting with in-stent restenosis within their AV grafts and fistulas when compared with balloon angioplasty alone. ¹¹

REVISE Trial

Finally, in 2019 Mohr et al published their REVISE randomized trial in the Journal of Vascular and Interventional Radiology , which investigated the clinical and economic benefits of stent grafts in dysfunctional and thrombosed hemodialysis access graft circuits. This multicenter trial enrolled 269 patients at 30 sites and followed up them for 24 months, with the goal of investigating the cumulative number of reinterventions, associated costs, and total costs for all patients. ¹² The number of reinterventions and total costs at 24 months were compared between the stent graft and PTA groups regardless of presenting pathology. Patients were also stratified into two groups based on the presentation of their dialysis circuit, separating patients with thrombosed grafts from patients with dysfunctional grafts. Each of these two groups were further subdivided to receive either GORE VIABAHN Endoprosthesis stent graft or balloon angioplasty alone. All patients received initial target lesion angioplasty prior to randomization and enrollment in the study. Mohr et al ¹² found that, when examining the total patient population including both thrombosed and dysfunctional dialysis grafts throughout the course of 24 months, the stent graft arm required a significantly lower average number of reinterventions within the dialysis circuit compared with the balloon angioplasty alone arm (3.7 vs. 5.1, p  = 0.005). In patients presenting with thrombosed hemodialysis access grafts, treatment with stent grafts led to a similar significant decrease in number of reinterventions at 24 months when compared with PTA alone (3.7 vs. 6.2, p  = 0.022). Among patients with dysfunctional grafts, no statistical difference was seen in the number of reinterventions between the stent graft and PTA groups. Total costs at 24 months were significantly lower in patients who presented with thrombosed hemodialysis access grafts and received treatment with stent grafts compared with PTA alone ($30,329 vs. $37,206, p  = 0.027). Total costs were otherwise statistically equivalent at 24 months between treatment groups among patients presenting with dysfunctional hemodialysis access grafts as well as the total patient population of the study when examined all together. The authors concluded that usage of stent grafts significantly reduced the number of reinterventions for all patients in this study at 24 months, driven by patients presenting with thrombosed grafts. Treatment with stent grafts resulted in a significantly lower total cost burden at 24 months in patients presenting with thrombosed grafts. ¹²

Patient Selection for Stent Graft Placement

Across all these clinical trials, stent grafts prolonged the primary patency of the treatment area when compared with balloon angioplasty alone. In most studies, stent grafts were also shown to significantly reduce reintervention rates in the dialysis circuit. The FLAIR ⁸ trial established that treatment by stent graft showed improved primary patency both at the intervention site and over the entire dialysis circuit without increased risk of complication at 6 months post-intervention. ⁸ In Vesley et al's Viabahn stent trial, ⁹ stent grafts delayed reintervention at the treatment site by 95 days compared with balloon angioplasty alone. ⁹ The RENOVA ¹⁰ study quantified that a greater than twofold improvement in primary access at the treatment area was sustained for at least 24 months with the use of stent grafts. ¹⁰ The RESCUE ¹¹ study concluded that these improvements in treatment area primary patency persisted for at least the next 24 months in both AV grafts and fistulas treated for in-stent restenosis. ¹¹ The REVISE ¹² trial determined that the reduction in additional interventional procedures offset the increased cost of deploying stent grafts in such patients. ¹² The reason why stent grafts may provide this benefit is unknown; however, many hypothesize that the stent graft provides an inert barrier to reduce the matrix factors responsible for restenosis from intimal hyperplasia. ¹¹ In the management of dialysis circuit stenosis, stent grafts are currently indicated for restenosis, rupture, and recoil (commonly seen in central vein stenosis). ¹³ However, the literature suggests that the success of stent grafts in dialysis circuits can vary heavily with the clinical indication for placement and nature of the circuit.

Additional Applications of Stent Grafts

In addition to the benefits seen in AV grafts, stent grafts have also demonstrated efficacy in the treatment of central venous disease in hemodialysis patients with functioning AV fistulas. In 2011, Jones et al published their retrospective analysis of 30 patients with functioning AV fistulas and central venous disease not responsive to balloon angioplasty. ¹³ After placement of the stent grafts, patients saw a primary patency rate of 45% at 2 years, and primary assisted patency rate of 75% at 2 years. Although there was no comparison group in this study, the patency rates seen in this cohort show an improvement over previously published rates of primary patency achieved by balloon angioplasty alone. ^{14 15 16} Unfortunately, stent grafts are limited by their available diameters and lengths in treating central vein stenosis. They have been associated with high rates of collateral coverage and venous confluence coverage, leading to their use in central veins only after other conservative options have been exhausted. ¹⁷

Some studies have indicated that stent grafts may be highly successful at managing “swing-point” stenoses, a frequent complication of many AV fistulas. ¹⁷ These stenoses occur at the segment of the vein mobilized for arterial anastomosis and are likely to rupture even at low pressure under balloon angioplasty. ¹⁸ They include stenosis at the basilic angle of transposition (BAT) in brachial-basilic fistulas and the cephalic arch in brachiocephalic fistulas. One study reported that stent grafts resulted in significantly higher secondary patency rates over 12 months than balloon angioplasty or bare metal stents alone. ¹⁹ Another found that, in the treatment of BAT stenoses, stent grafts showed a 12-month assisted patency rate of 80% compared with 40% by balloon angioplasty alone. BAT stenoses are especially difficult to manage surgically, making them an appealing target for endovascular therapies. ¹⁷ When considering the application of stent grafts in dialysis circuits, the presence of specific clinical indications, like BAT stenoses, in which stent grafts may perform particularly well should also be considered.

Case Presentation

At the authors' institution, a 63-year-old male with a left upper arm AV graft presented in November 2019 with abdominal pain, nausea, and vomiting. His AV graft had been placed a year earlier and underwent successful angioplasty of the venous anastomosis 6 months following its creation. A month earlier, an AV graft occlusion with 90% stenosis was identified when the patient was admitted for acute left cerebellar stroke, abdominal pain, and severe hypertension. After the patient's arm was prepped and draped, heparin 10,000 units was given IV. Ultrasound of the upper arm AV graft found no flow and collapse of part of the graft, and a subsequent graftogram demonstrated a stenosed AV graft ( Fig. 1 ). Following mechanical thrombectomy with a Trerotola thrombectomy device (Teleflex, Pennsylvania), a venogram showed a severe long stenosis in the brachial vein starting at the venous anastomosis and extending to the start of the axillary veins. A 7 mm × 100 mm Viabahn stent graft was deployed in the venous anastomosis and was extended peripheral to the brachial vein ( Fig. 2a, b ). Contrast injection confirmed good positioning, expansion, no residual stenosis of the brachial vein, and good flow through the axillary, subclavian, and innominate veins. Contrast injection through the brachial artery identified a filling defect, likely representing thrombus, causing greater than 50% stenosis 3 cm from the arterial anastomosis ( Fig. 3a ) treated by balloon angioplasty ( Fig. 3b ). A 7 mm × 15 mm Viabahn stent was then deployed at the arterial anastomosis, resolving the filling defect on arteriogram. The final graftogram showed an excellent angiographic result with no residual stenosis at the treatment sites ( Fig. 4 ). Five months following the procedure, color Doppler imaging was performed of the inflow artery, entire dialysis circuit, and outflow vein. These waveforms corresponded with high-flow rate. The graft was found to be patent with adequate volume, no significant intimal hyperplasia, and no focal increase in flow velocities ( Fig. 5 ).

Fig. 1.

Angiography demonstrating a thrombosed AV graft. Arrow—graft. Arrowhead—dominant outflow vein.

Fig. 2.

Following declotting of the AV graft, there was a greater than 50% stenosis within the venous outflow. ( a ) Positioning of a 7 mm × 100 mm Viabahn stent within AV graft (arrow). ( b ) Ballooning of covered stent (arrow).

Fig. 3.

( a ) Filling defect at arterial anastomosis most likely representing thrombus (arrow). ( b ) Attempted balloon angioplasty and removal of thrombus.

Fig. 4.

Residual stenosis at the arterial anastomosis, which was successfully treated with an additional Viabahn stent graft. Subsequent graftogram shows an excellent angiographic result.

Fig. 5.

( a ) Color Doppler image of patent arterial anastomosis at 5-month follow-up. ( b ) Color Doppler image of patent venous anastomosis at 5-month follow-up.

Complications and Considerations

With regard to adverse events, there was a significantly lower rate of restenosis in patients treated with stent grafts compared with balloon angioplasty alone, and this was seen in both the FLAIR ⁸ and RENOVA ¹⁰ trials. Similarly, the REVISE ¹² trial found that usage of stent grafts significantly reduced the number of reinterventions in patients presenting with thrombosed grafts. The FLAIR ⁸ study found that stent grafts significantly reduced restenosis (40 vs. 77%, p  < 0.001) but not thrombotic occlusion of the dialysis circuit (33 vs. 21%, p  = 0.10) at 6 months. Vesely et al explained that, despite significantly delaying restenosis and reducing the rate of treatment site reintervention, stent grafts significantly increased the number of procedures outside the treatment area that ended primary circuit patency within the AV graft. ⁹ Although the effects of stent graft placement on hemodialysis access lifetime are disputed, no studies have found stent grafts to do worse than balloon angioplasty in preserving dialysis access. ¹⁸

The data on stent graft outcomes in dialysis circuits are promising, but there are still some important factors to take into consideration when tailoring treatment plans for individual patients. Although the primary risk of balloon angioplasty is vessel rupture, stenting can result in stent graft device fracture, migration, strut protrusion, infection, and skin erosion stenosis. ^{18 20 21} Stent graft complications nonetheless were exceedingly rare in all trials discussed. In all studies, there was no significant difference between rates of infection in the stent graft versus balloon angioplasty groups. However, stent grafts may restrict the available area for dialysis access and repeated cannulation through the graft can cause migration and fracture. ¹⁸ Stent grafts can also cover significant portions of outflow veins, limiting sites for future hemodialysis access. ¹⁹

In the treatment of AV fistula pseudoaneurysms, stent grafts have a higher risk for device fracture and migration, along with a threefold increase in infection rate observed when treating pseudoaneurysms of AV fistulas with stent grafts compared with bare metal stents. ^{22 23}

Conclusion

Compared with balloon angioplasty alone, published clinical trial results have shown stent grafts to significantly improve primary patency of failing dialysis access grafts. Stent grafts have also been shown to significantly reduce the rates of reintervention compared with balloon angioplasty alone. Studies have shown the value of stent graft placement for treatment of in-stent restenosis within the dialysis circuit. For more specific indications, such as BAT stenoses, initial studies have shown that grafts outperform balloon angioplasty on primary and assisted patency of the circuit. Future research should investigate stent graft performance in dialysis circuits on narrower subsets of conditions and prioritize study designs that analyze the effects of stent grafts on lifetime hemodialysis access as a primary endpoint.