Original Contribution

Primary Patency With Stenting Versus Balloon Angioplasty for Arteriovenous Graft Failure: A Systematic Review and Meta-Analysis

Konstantinos Marmagkiolis, MD1*;  Cezar Iliescu, MD2*;  Mohan Mallikarjuna Rao Edupuganti, MD3;  Marwan Saad, MD, PhD3,4;  Konstantinos Dean Boudoulas, MD5;  Akash Gupta6;  Nikolaos Lontos, MD7;  Mehmet Cilingiroglu, MD8 

Konstantinos Marmagkiolis, MD1*;  Cezar Iliescu, MD2*;  Mohan Mallikarjuna Rao Edupuganti, MD3;  Marwan Saad, MD, PhD3,4;  Konstantinos Dean Boudoulas, MD5;  Akash Gupta6;  Nikolaos Lontos, MD7;  Mehmet Cilingiroglu, MD8 

Abstract: Objectives. To evaluate the efficacy of advanced stent technology in the management of failing arteriovenous grafts (AVGs). Background. End-stage renal disease rates and the need for hemodialysis are increasing worldwide. AVG remains a common dialysis access site. Several techniques have been previously suggested to restore and preserve AVG patency. A quantitative evaluation and synthesis of this information are essential in elucidating the role of newer stent platforms for the management of failing AVG. Methods. We performed a literature search using PubMed, Web of Science, and Embase from January 2006 to December 2017. Studies comparing the primary patency rates with stent placement vs balloon angioplasty alone in patients with failed AVGs were included. Results. Seven studies with a total of 1109 patients met the inclusion criteria. The mean graft age was 2.89 years in the stent group and 3.29 years in the balloon angioplasty group. Stent placement was associated with improved primary patency rates compared with balloon angioplasty alone at short-term (3-month) follow-up (73.2% vs 42.6%, respectively; risk ratio [RR], 0.55; 95% confidence interval [CI], 0.35-0.88; P=.01) and mid-term (6-month) follow-up (50.8% vs 18.4%, respectively; RR, 0.65; 95% CI, 0.51-0.82; P<.001). The primary patency rates remained favorable with stent placement at 12-month (40.3% vs 13.0%, respectively; RR, 0.69; 95% CI, 0.63-0.77; P<.001) and 24-month follow-up (20.5% vs 6.8%; RR, 0.86; 95% CI, 0.80-0.92; P<.001) compared with balloon angioplasty alone. Conclusions. Stent placement is associated with improved patency rates compared with balloon angioplasty alone.  

J INVASIVE CARDIOL 2019;31(12):E356-E361.

Key words: balloon angioplasty, renal disease, stent

Hemodialysis access-site patency becomes increasingly important with the continuous rise in the prevalence of chronic kidney disease and end-stage renal disease. While autogenous arteriovenous fistulas (AVFs) remain the most frequently used vascular access modality for hemodialysis (HD), arteriovenous grafts (AVGs) constitute an important alternative in patients with unfavorable venous anatomy.1 They are technically easier to create even in an outpatient setting and achieve functionality as early as 2 weeks after implantation. However, most synthetic polytetrafluoroethylene grafts fail within 18 months of implantation, mainly due to intimal hyperplasia at the venous anastomosis site.2,3

Several techniques have been previously suggested to restore AVG patency. In the 1980s, surgical interventions to extend the lifetimes of AVFs and AVGs were largely supplanted by percutaneous transluminal angioplasty (PTA).4 Primary patency rates with PTA, however, are suboptimal.5 Endovascular stents were first used to treat failed dialysis access grafts in 1988, and early studies failed to show improvement in patency rates.6-8 With new advances in stent technology, nitinol and shape memory alloy recoverable technology (SMART) stents became available; they offered better wall apposition, radial strength, and elasticity compared with first-generation bare-metal stents.9

More recently, covered stents were used in cases of venous anastomotic stenosis. These stents are made of materials similar to the AVG, and aim to reduce elastic recoil and tissue in-growth, thus improving long-term patency.10,11 The current meta-analysis aims to evaluate the efficacy of advanced stent technology in the management of failing AVGs compared with PTA alone.  


Search strategy. We conducted an electronic search of PubMed, Web of Science, and EMBASE from January 2006 to December 2017 using the Medical Subject Heading (MeSH) terms “dialysis,” “chronic kidney disease,” “end-stage renal disease,” “vascular access,” “arteriovenous graft,” “stent,” and “balloon angioplasty” separately and in combination. The search was conducted according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. We hand-searched the reference lists of all articles identified in the search. We also reviewed unpublished studies from the website www.clinicaltrials.gov. We decided, a priori, to exclude pediatric studies. Figure 1 shows our search strategy. 

Inclusion criteria. We included clinical studies that compared stent placement with PTA alone for the management of failed AVGs in dialysis patients and reported clinical outcomes of interest. Published studies with the following characteristics were included: (1) comparison of stenting vs PTA alone for the management of AV graft failure at the venous anastomosis site; (2) a minimum of 5 patients examined; and (3) manuscript published in English. 

Data extraction and outcomes. Two investigators (KM, MC) independently reviewed the studies and reported the results in a structured dataset. Studies were evaluated carefully for duplicates or overlapping data. Discrepancies between investigators regarding the inclusion of each trial were resolved by consensus with a third independent investigator. Prespecified data elements were extracted from each trial as follows: sample size; sex; age; hypertension; diabetes mellitus; coronary artery disease; type of stent; and AVG age. The primary endpoint of the current study was the primary patency of the treated area after intervention. We evaluated the primary patency rates at 3, 6, 12, and 24 months (when reported). The Vesely et al study12 reported 24-month follow-up data as Kaplan-Meier curves, so only 6-month follow-up rates were included. The number of events for clinical outcomes in both arms was tabulated. 

Statistical analysis. We performed random-effects summary risk ratios (RR) using a DerSimonian and Laird model, as recommended by the Cochrane Handbook for Systemic Reviews of Intervention.13 P-values were 2-tailed, and considered statistically significant when P<.05. Confidence intervals (CIs) were calculated at the 95% level for the overall estimate effect. Heterogeneity was evaluated using I2 statistic value, with values >50% indicating high heterogeneity.14 Random effects meta-regression analysis was performed to evaluate for any modification in the primary outcome with baseline characteristics, including age, male sex, hypertension, diabetes mellitus, and coronary artery disease (CAD). All analyses were performed using STATA software, version 14 (StataCorp).


Of the 175 citations found, seven studies (3 randomized and 4 observational) with a total of 1109 patients met the inclusion criteria.1,11,12,15–18 Characteristics of each study are listed in Table 1. The majority of patients with AVG failure received covered stents. The stent group comprised 533 patients and the PTA-alone group comprised 576 patients. There were no differences between the two groups with respect to baseline characteristics (Table 2). There were no significant differences between the stent group and the PTA-alone group with respect to baseline demographic variables, including age (62.0 years vs 61.6 years), male sex (47.6% vs 48.0%), hypertension (91.5% vs 91.5%), diabetes mellitus (59.9% vs 61.2%), and CAD (37.5% vs 34.5%). 

In all patients with AVG failure, stenosis was located to the venous anastomosis of the AVG. Some studies restricted themselves to venous thrombosis, while others studied patients with stenosis of the venous limb. Venous stenosis was defined as >50% narrowing of the lumen. Most studies achieved 100% procedural success, except for Haskal et al,10 who reported 94% procedural success. There was no difference in the complication rates of the two groups. The mean graft age was 2.89 years in the stent group and 3.29 year in the PTA-alone group. 

Stent placement was associated with improved primary patency rates compared with PTA alone at short-term (3-month) follow-up (73.2% vs 42.6%; RR, 0.55; 95% CI, 0.35-0.88; P=.01; I2=68.1; P=.01), and mid-term (6-month) follow-up (50.8% vs 18.4%; RR, 0.65; 95% CI, 0.51-0.82; P<.001; I2=82.6%; P<.001). The primary patency rates remained favorable with stent placement at 12-month follow-up (40.3% vs 13.0%; RR, 0.69; 95% CI, 0.63-0.77; P<.001; I2=0.0%; P<.78) and 24-month follow-up (20.5% vs 6.8%; RR, 0.86; 95% CI, 0.80-0.92; P<.001; I2=0.0%; P<.89) compared with PTA alone (Figures 2 and 3). Table 3 summarizes the primary patency rates at different follow-up durations.


To our knowledge, the current study is the most updated meta-analysis to evaluate the role of the new stent technologies in dialysis patients with failed AVGs compared with PTA alone. The current analysis demonstrates about 3-fold higher patency rates with stent placement vs PTA alone at long-term follow-up (number needed to treat [NNT] = 3). This beneficial effect persisted up to 24 months of follow-up.  

The most common cause of AVG failure is stenosis due to neointimal hyperplasia at the venous outflow segment, leading to thrombosis and graft malfunction. In the past, several methods have attempted to prolong the functionality of the AVG with antiplatelet or antithrombotic agents and serial ultrasound follow-up.19 PTA alone initially, followed by the use of different stent platforms and stent-grafts, contributed to the prolongation of good AVG function. In the 1990s, well-designed studies failed to show the superiority of stents over PTA alone.8,20 

Maintaining the patency of AVGs for HD is challenging. The evidence thus far has demonstrated no difference between PTA alone and stenting. With the development of newer-generation stents, studies suggest improvement in patency rates; however, these studies were small with limited follow-up, making it difficult to determine the magnitude of benefit. A meta-analysis published in 2015 showed that stents may provide a durable result and improve patency above and beyond that achieved by PTA alone.4 This study was characterized by important diversity in content and analysis, including: (1) patients with bare-metal stents; (2) both failed AVGs and AVGs; and (3) stenosis located in the venous limb of the circuit and more upstream in the central venous system. 

The current meta-analysis was designed to address previous limitations. In addition, whereas previous studies had a broader scope, we only selected studies with lesions in the venous limb of the AV circuit alone. The meta-analysis suggests that the mean AVG age requiring intervention for AVG failure is approximately 3 years. It clearly demonstrates the superiority of bare-metal stents and covered stents for the management of venous anastomotic stenoses compared with plain PTA alone. It should be noted that most data about long-term patency came from the covered stent population, with few long-term data available from bare-metal stents. There was a uniformly high degree of procedural success without increased risk of procedural complications. The NNT between 3 and 5 (although based on a small number of studies) strongly supports the use of stenting in the management of these lesions. 

A randomized controlled trial including newer technologies (newer stents, drug-coated balloons, atherectomy or thrombectomy platforms) with optimal procedural techniques (sizing and duration of balloon and stent implantation), intravascular imaging, and dual-antiplatelet therapy (DAPT) is needed to better understand and optimize the management of AVG failure. 

Study limitations. The main limitation of the current study is the significant heterogeneity that resulted from differences in inclusion criteria, stent types used (including bare-metal and covered), definition of technical success, and method of follow-up. In the RENOVA trial, patients with AVG stenosis >50% with clinical evidence of graft failure were included.10  The RESCUE trial included patients with AVG as well as AVFs with in-stent restenosis (>50%) in previously placed bare-metal stents.18 Technical success was defined as a target-lesion stenosis ≤30% in the studies done by Haskal et al10 and Kakisis et al.16 Carmona et al1 defined technical success as a residual stenosis of 10% or less. Maya et al17 defined technical success by measuring the intragraft and systemic blood pressures upon completion of the procedure.

This meta-analysis included three small observational studies. Despite their small sample size, we felt that their inclusion was necessary for the completeness of the manuscript, in the absence of larger studies. 

Study follow-up also varied among studies. In the RENOVA trial, follow-up was achieved by a combination of clinical and angiographic evaluation. Repeat intervention was not performed in asymptomatic clinically silent stenosis. It is unlikely that this may have affected the current meta-analysis results, as this approach was applied to both comparison groups. In the study by Kakisis et al, follow-up consisted of dynamic and static dialysis venous pressures using Duplex ultrasound imaging.16 Carmona et al defined graft patency simply by the functionality of the AVG. The RESCUE trial18 and Vesely et al12 defined graft patency as the interval following the index intervention until the next access thrombosis or need for repeated intervention. 

Balloon and stent sizing were not documented or compared and it may have influenced the periprocedural technical success and the primary patency both in the stent and the PTA-alone arms.   

The use and duration of DAPT were not clearly documented in the included studies. It is probable that DAPT may affect the rates of thrombosis in the stent and angioplasty arms. Randomized controlled trials are needed to clarify the value of DAPT in that patient subgroup.


Stent placement is associated with improved long-term patency rates compared with PTA alone in patients with failed AVGs. Large randomized trials are needed to confirm the results of the current analysis. 


 1. Carmona J, Rits Y, Jones B, Dowers L, Bednarski D, Rubin JR. Patency of the Viabahn stent graft for the treatment of outflow stenosis in hemodialysis grafts. Am J Surg. 2016;211:551-554. 

2. Pisoni RL, Young EW, Dykstra DM, et al. Vascular access use in Europe and the United States: results from the DOPPS. Kidney Int. 2002;61:305-316. 

3. Kanterman RY, Vesely TM, Pilgram TK, Guy BW, Windus DW, Picus D. Dialysis access grafts: anatomic location of venous stenosis and results of angioplasty. Radiology. 1995;195:135-139. 

4. Fu N, Joachim E, Yevzlin AS, Shin J-I, Astor BC, Chan MR. A meta-analysis of stent placement vs. angioplasty for dialysis vascular access stenosis. Semin Dial. 2015;28:311-317. 

5. Glanz S, Gordon DH, Butt KM, Hong J, Lipkowitz GS. The role of percutaneous angioplasty in the management of chronic hemodialysis fistulas. Ann Surg. 1987;206:777-781. 

6. Zollikofer CL, Antonucci F, Stuckmann G, Mattias P, Brühlmann WF, Salomonowitz EK. Use of the Wallstent in the venous system including hemodialysis-related stenoses. Cardiovasc Intervent Radiol. 1992;15:334-341. 

7. Günther RW, Vorwerk D, Klose KC, et al. Self-expanding stents for the treatment of a long venous stenosis in a dialysis shunt: case report. Cardiovasc Intervent Radiol. 1989;12:29-31. 

8. Beathard GA. Gianturco self-expanding stent in the treatment of stenosis in dialysis access grafts. Kidney Int. 1993;43:872-877. 

9. Clark TWI. Nitinol stents in hemodialysis access. J Vasc Interv Radiol. 2004;15:1037-1040. 

10. Haskal ZJ, Trerotola S, Dolmatch B, et al. Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med. 2010;362:494-503. 

11. Haskal ZJ, Saad TF, Hoggard JG, et al. Prospective, randomized, concurrently-controlled study of a stent graft versus balloon angioplasty for treatment of arteriovenous access graft stenosis: 2-year results of the RENOVA study. J Vasc Interv Radiol. 2016;27:1105.e3-1114.e3. 

12. Vesely T, DaVanzo W, Behrend T, Dwyer A, Aruny J. Balloon angioplasty versus Viabahn stent graft for treatment of failing or thrombosed prosthetic hemodialysis grafts. J Vasc Surg. 2016;64:1400.e1-1410.e1. 

13. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177-188. 

14. Higgins JPT. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557-560. 

15. Chan MR, Bedi S, Sanchez RJ, et al. Stent placement versus angioplasty improves patency of arteriovenous grafts and blood flow of arteriovenous fistulae. Clin J Am Soc Nephrol. 2008;3:699-705. 

16. Kakisis JD, Avgerinos E, Giannakopoulos T, Moulakakis K, Papapetrou A, Liapis CD. Balloon angioplasty vs nitinol stent placement in the treatment of venous anastomotic stenoses of hemodialysis grafts after surgical thrombectomy. J Vasc Surg. 2012;55:472-478. 

17. Maya ID, Allon M. Outcomes of thrombosed arteriovenous grafts: comparison of stents vs angioplasty. Kidney Int. 2006;69:934-937. 

18. Falk A, Maya ID, Yevzlin AS; RESCUE Investigators. A prospective, randomized study of an expanded polytetrafluoroethylene stent graft versus balloon angioplasty for in-stent restenosis in arteriovenous grafts and fistulae: two-year results of the RESCUE study. J Vasc Interv Radiol. 2016;27:1465-1476. 

19. Mousa AY, Patterson W, Abu-Halimah S, et al. Patency in arteriovenous grafts in hemodialysis patients.Vasc Endovasc Surg. 2013;47:438-443. Epub 2013 Jul 12.

20. Quinn SF, Schuman ES, Demlow TA, et al. Percutaneous transluminal angioplasty versus endovascular stent placement in the treatment of venous stenoses in patients undergoing hemodialysis: intermediate results. J Vasc Interv Radiol. 1995;6:851-855.

*Joint first authors.

From the 1Division of Cardiovascular Disease, Florida Hospital, Tampa, Florida; 2Division of Cardiology, Department of Medicine, MD Anderson Cancer Center, University of Texas at Houston, Houston, Texas; 3Division of Cardiovascular Disease, Baptist Hospital, Little Rock, Arkansas; 4Division of Cardiovascular Disease, University of Arkansas for Medical Sciences, Little Rock, Arkansas; 5Division of Cardiology, Ain Shams University, Cairo, Egypt; 6Division of Cardiovascular Disease, The Ohio State University, Columbus, Ohio; 7Baylor College of Medicine, Houston, Texas; and 8Bahcesehir University, School of Medicine, Istanbul, Turkey.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted February 28, 2019, provisional acceptance given June 24, 2019, final version accepted October 11, 2019.

Address for correspondence: Konstantinos Marmagkiolis, MD, MBA, FSCAI, Advent Health Pepin Heart Institute, 3100 East Fletcher Ave, Tampa, FL 33613. Email: c.marmagiolis@gmail.com