Stent Graft versus Balloon Angioplasty for Failing Dialysis-Access Grafts

Transcription

1 The new england journal of medicine original article Stent Graft versus Balloon Angioplasty for Failing Dialysis-Access Grafts Ziv J. Haskal, M.D., Scott Trerotola, M.D., Bart Dolmatch, M.D., Earl Schuman, M.D., Sanford Altman, M.D., Samuel Mietling, M.D., Scott Berman, M.D., Gordon McLennan, M.D., Clayton Trimmer, D.O., John Ross, M.D., and Thomas Vesely, M.D. Abstract From the University of Maryland Medical Center, Baltimore (Z.J.H.); the Hospital of the University of Pennsylvania, Philadelphia (S.T.); the University of Texas Southwestern Medical Center, Dallas (B.D., C.T.); Oregon Surgical Consultants, Portland (E.S.); Open Access Vascular Access Center, Miami (S.A.); Vascular Access Center, Augusta, GA (S.M.); Tucson Vascular Surgery, Tucson, AZ (S.B.); Indiana University School of Medicine, Indianapolis (G.M.); Bamberg County Hospital and Nursing Center, Bamberg, SC (J.R.); and the Vascular Access Center of Frontenac Grove, Frontenac, MO (T.V.). Address reprint requests to Dr. Haskal at the Division of Vascular and Interventional Radiology, University of Maryland Medical Center, 22 S. Greene St., GK214, Baltimore, MD 21201, or at N Engl J Med 2010;362: Copyright 2010 Massachusetts Medical Society. Background The leading cause of failure of a prosthetic arteriovenous hemodialysis-access graft is venous anastomotic stenosis. Balloon angioplasty, the first-line therapy, has a tendency to lead to subsequent recoil and restenosis; however, no other therapies have yet proved to be more effective. This study was designed to compare conventional balloon angioplasty with an expanded polytetrafluoroethylene endovascular stent graft for revision of venous anastomotic stenosis in failing hemodialysis grafts. Methods We conducted a prospective, multicenter trial, randomly assigning 190 patients who were undergoing hemodialysis and who had a venous anastomotic stenosis to undergo either balloon angioplasty alone or balloon angioplasty plus placement of the stent graft. Primary end points included patency of the treatment area and patency of the entire vascular access circuit. Results At 6 months, the incidence of patency of the treatment area was significantly greater in the stent-graft group than in the balloon-angioplasty group (51% vs. 23%, P<0.001), as was the incidence of patency of the access circuit (38% vs. 20%, P = 0.008). In addition, the incidence of freedom from subsequent interventions at 6 months was significantly greater in the stent-graft group than in the balloonangioplasty group (32% vs. 16%, P = 0.03 by the log-rank test and P = 0.04 by the Wilcoxon rank-sum test). The incidence of binary restenosis at 6 months was greater in the balloon-angioplasty group than in the stent-graft group (78% vs. 28%, P<0.001). The incidences of adverse events at 6 months were equivalent in the two treatment groups, with the exception of restenosis, which occurred more frequently in the balloon-angioplasty group (P<0.001). Conclusions In this study, percutaneous revision of venous anastomotic stenosis in patients with a prosthetic hemodialysis graft was improved with the use of a stent graft, which appears to provide longer-term and superior patency and freedom from repeat interventions than standard balloon angioplasty. (ClinicalTrials.gov number, NCT ) 494 n engl j med 362;6 nejm.org february 11, 2010

2 Stent Graft vs. Balloon Angioplasty for failing Dialysis-Access Grafts By 2008, more than 341,000 patients in the United States were undergoing hemodialysis for treatment of their end-stage renal disease. 1 The National Kidney Foundation Kidney Disease Outcomes Quality Initiative seeks to increase the use of autogenous fistulas, yet many patients continue to undergo hemodialysis with the use of prosthetic arteriovenous grafts. The reasons for this discrepancy between the recommendation and practice are multifactorial and continue to be debated. 2,3 The costs of maintaining vascular access are substantial; for example, the cost of treating a patient who has failure of a hemodialysis access graft is significantly higher ($62,000 per patient-year) than the cost of treating a patient who does not have access failure. 1,4 Many percutaneous techniques and endovascular tools have been used to treat the neointimal stenoses that develop at the site of venous anastomoses of arteriovenous grafts. At best, secondary patency of arteriovenous grafts (i.e., patency after an intervention) is 50% at 3 years after the creation of the vascular access; typically, multiple interventions are required to maintain patency. 5,6 No reported mechanical, endovascular, or pharmacologic approaches have improved the patency of arteriovenous grafts as compared with balloon angioplasty alone We hypothesized that revision of a venous anastomotic stenosis with a stent graft constructed with the same material as the graft would prevent elastic recoil and tissue ingrowth, thereby improving long-term patency as compared with that afforded by standard balloon angioplasty. Methods Study Design In our prospective, multicenter, randomized, controlled trial, patients were eligible if they had end-stage renal disease and were undergoing long-term hemodialysis with the use of failing, but nonthrombosed, prosthetic arteriovenous grafts. The study was designed to assess the safety and efficacy of an expanded polytetrafluoroethylene stent graft, as compared with balloon angioplasty, for the treatment of hemodynamically significant venous anastomotic stenosis in an arteriovenous graft. Inclusion and exclusion criteria, listed in Table 1, were developed in accordance with guidelines of the National Kidney Foundation Kidney Disease Outcomes Quality Initiative and the Society of Interventional Radiology The study was approved by each center s institutional review board and the Food and Drug Administration and was in compliance with Health Insurance Portability and Accountability Act regulations; all patients provided written informed consent. An independent clinical events committee at Harvard Clinical Research Institute (Boston) adjudicated the clinical data and the Angiographic Core Lab of the Cardiovascular Research Foundation (New York) analyzed the angiographic films. The principal investigator designed the study, with assistance from the sponsor (Bard Peripheral Vascular). The data were collected by on-site investigators under the auspices of the sponsor and principal investigator for analysis; Harvard Clinical Research Institute performed the statistical analyses. The principal investigator prepared the manuscript, which was reviewed by all authors, who vouch for the accuracy and completeness of the reported data. Study End Points The study objective was to demonstrate that treatment with a stent graft is not inferior to treatment with balloon angioplasty alone regarding the primary end point, the 6-month primary patency of a stenotic venous anastomosis in the treatment area. Secondary end points included safety variables, procedural success (successful percutaneous insertion of the stent graft), primary patency of the access circuit at 2 months and 6 months, the percent stenosis of the treatment area at 2 months and 6 months, and freedom from subsequent intervention. See the Supplementary Appendix (available with the full text of this article at NEJM.org) for definitions of treatment area and primary patency and measures of successful intervention. Stent Graft The investigational device consisted of a selfexpanding nitinol stent covered in carbon-impregnated expanded polytetrafluoroethylene (Flair Endovascular Stent Graft, Bard Peripheral Vascular). Two stent-graft configurations were used: tubular (straight) and flared. The flared configuration was used when the diameter of the outflow vein beyond the stenosis was larger than that of the arteriovenous graft. The stent graft n engl j med 362;6 nejm.org february 11,

3 The new england journal of medicine Table 1. Inclusion and Exclusion Criteria in the Study. Inclusion criteria Age of 18 to 90 yr and a hemodialysis access consisting of a synthetic arteriovenous access graft located in the arm. Angiographic evidence of 1 stenoses, 7 cm in length and 50%, at the graft-vein anastomosis of a synthetic arteriovenous access graft, with the entire lesion located within 7 cm of the anastomosis, such that approximately 1 cm of the stent graft will be extended into a nondiseased vein and approximately 1 cm, but no more than 2 cm, of the stent graft will be extended into a nondiseased arteriovenous graft. Clinical evidence of a hemodynamically significant stenosis. Percutaneous endovascular therapy thought by the investigator to have been the best treatment choice for the identified lesion. Patient s ability to provide written informed consent. Synthetic arteriovenous access grafts implanted >30 days before enrollment and 1 successful hemodialysis sessions performed. During primary balloon angioplasty, full expansion of an appropriately sized angioplasty balloon, in the operator s judgment. Exclusion criteria Concomitant disease (e.g., terminal cancer) or other medical condition likely to result in death within 6 mo after the time of implantation. Stenosis with a corresponding thrombosis treated within 7 days before enrollment. A second lesion in the access circuit (area from the arteriovenous access graft arterial anastomosis to the superior vena cava right atrial junction), 3 cm from the edges of the primary lesion, either treated within 30 days before enrollment or 30%. The presence of a second lesion in the access circuit >3 cm from the edges of the primary lesion that was 30%. Second lesions that were 30% must have been treated before patient inclusion to reduce the percent stenosis to <30%. Being unwilling or unable to return for follow-up visits or reason to believe that adherence to follow-up visits would be irregular. A stent placed at the target lesion site. A blood coagulation disorder or sepsis. Requirement that the stent graft would have to cross an angle (between the inflow vein and synthetic arteriovenous access graft) of >90 degrees. Requirement that the stent graft would have to be deployed fully across the elbow joint (which is identified radiographically as a combination of the humeroulnar joint and the humeroradial joint). A contraindication to the use of contrast medium. Infected arteriovenous access graft. Current or scheduled enrollment in other, conflicting studies. Procedural use of another investigational device. Pregnancy. was available in diameters of 6 to 9 mm and lengths of 30, 40, and 50 mm. It was deployed through a 9-French delivery catheter. Randomization and Intervention Once the enrollment criteria were met, angiography of the graft and treatment area was performed with the use of orthogonal magnified views, each 30 degrees or more apart, and a radiopaque 1-mm graduated ruler in the imaged field of view. The percent stenosis was calculated as follows: [1 (minimal lumen diameter nondiseased lumen diameter)] 100, where the minimal lumen diameter was the narrowest lumen diameter within the stenosed area and the nondiseased lumen diameter was the lumen diameter of the nondiseased arteriovenous graft or vein just upstream of the lesion. Remote secondary lesions more than 3 cm from the treatment area were required to have been treated until the stenosis was less than 30%, before randomization. Randomization was performed with the use of permuted blocks, identified in sealed envelopes that were sent to each site in advance. If the lesion met study criteria, then balloon angioplasty was performed with the use of an appropriately sized conventional (noncutting, noncompliant) angioplasty balloon. After angioplasty, patients were randomly assigned to undergo placement of a stent graft or to receive no other treatment. No patients were excluded before randomization because of resistant stenoses that 496 n engl j med 362;6 nejm.org february 11, 2010

4 Stent Graft vs. Balloon Angioplasty for failing Dialysis-Access Grafts prevented full expansion of the initial angioplasty balloon. If a patient was randomly assigned to the stent-graft group, a device of appropriate length, configuration, and diameter ( 1 mm greater than the arteriovenous graft, to avoid the use of an oversized device) was chosen. The stent graft was placed and was then dilated, with a balloon diameter equal to that used in the balloon-angioplasty group. Lesions were dilated to reduce the stenosis in the treated area to less than 30%. The angiograms were later sent to the Angiographic Core Lab for analysis. Clinical and Angiographic Follow-up Regimen Patients were treated and discharged according to each center s standard of care. Anticoagulation or antiplatelet agents were administered after the procedure at the physician s discretion. A single, intravenous dose of prophylactic antibiotic usually cefazolin sodium, in patients who did not have an allergy was administered in the stent-graft group. Mandatory clinical evaluations and magnified quantitative angiography were performed 2 and 6 months after the index procedure, with the use of similar imaging protocols. Medical events, hospitalizations, access interventions, and adverse events were recorded. Graft function was assessed by means of a clinical evaluation of the same clinical or hemodynamic indicator used to assess graft function during the initial angiographic evaluation (one of the graft-function indicators accepted by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative), as well as by means of angiographic evaluation. Catheter-based interventions were performed in patients who both met clinical criteria for graft dysfunction and had stenoses of more than 50%. In accordance with the National Kidney Foundation Kidney Disease Outcomes Quality Initiative guidelines, investigators were instructed not to intervene regarding asymptomatic, clinically silent (i.e., incidentally diagnosed) stenoses found in the treatment area during angiography at 2 and 6 months. Statistical Analysis We calculated the sample size needed to test the primary noninferiority hypothesis using the methods of Blackwelder. 19 The incidence of primary patency at 6 months was estimated as 60% in the stent-graft group and 50% in the balloon-angioplasty group. The two rates were considered clinically noninferior if the difference was 10 percentage points or less (with a significance threshold of P = 0.05 on a one-tailed test and 80% statistical power). On this basis, the number of patients required for each of the two treatment groups was calculated to be 76. The target number of patients enrolled in each group was set at 95, to account for a dropout rate of up to 20%. Thus, the total target sample size was 190 patients. Intention-to-treat analyses were performed to evaluate the 6-month primary patency. The times to the return of symptoms and patency were analyzed with the use of Kaplan Meier product limit survival estimates. The data for patients with missed 6-month visits were censored in the estimation of the percentage of patients with graft patency at 6 months. Continuous secondary variables (e.g., percent stenosis) were analyzed by means of parametric or nonparametric analysis of variance with covariate adjustment. Outcomes for nonprimary effectiveness variables (e.g., primary patency of the access circuit) were analyzed at 2 and 6 months after the procedure. Subgroup analyses were performed to evaluate the influence of concomitant variables. P values less than 0.05 were considered to indicate statistical significance. Results Patients A total of 190 patients at 13 study sites were enrolled: 97 were randomly assigned to undergo implantation of the investigational stent graft and 93 were randomly assigned to undergo the control procedure, balloon angioplasty, only. The study patients consisted of 69 men and 121 women (Table 2). Participating centers were academic, community-based, inpatient, or freestanding outpatient dialysis centers. There were no significant differences between the two treatment groups at baseline with respect to demographic characteristics, relevant medical history, or characteristics of the arteriovenous access graft, with the exception of a higher incidence of axillary venous anastomosis in the balloon-angioplasty group. Nor were there significant differences between the two groups with respect to the nature or prevalence of abnormalities leading to the intervention in the arteriovenous access graft. The n engl j med 362;6 nejm.org february 11,

5 The new england journal of medicine Table 2. Characteristics of the Study Patients and Access Grafts at Baseline, According to Treatment Group.* Characteristic Stent Graft (N = 97) Balloon Angioplasty (N = 93) P Value Age yr 61.8± ± Male sex no./total no. (%) 33/97 (34) 36/93 (39) 0.55 Hypertension no./total no. (%) 96/97 (99) 87/93 (94) 0.06 Coronary artery disease no./total no. (%) 33/90 (37) 34/88 (39) 0.88 Congestive heart failure no./total no. (%) 25/89 (28) 19/86 (22) 0.39 Diabetes no./total no. (%) 59/97 (61) 58/93 (62) 0.88 Chronic obstructive pulmonary disease no./total no. (%) 7/91 (8) 5/87 (6) 0.77 Hypercoagulability no./total no. (%) 1/91 (1) 0/ Glomerulonephritis no./total no. (%) 5/90 (6) 3/84 (4) 0.72 Anticoagulant agents no./total no. (%) 40/97 (41) 36/93 (39) 0.77 Antiplatelet agents no./total no. (%) 11/97 (11) 6/93 (6) 0.31 Age at placement of arteriovenous graft yr 2.19± ± Location of arteriovenous graft no./total no. (%) Right arm 23/97 (24) 22/93 (24) 0.99 Left arm 74/97 (76) 71/93 (76) Forearm 20/97 (21) 24/92 (26) 0.64 Upper arm 73/97 (75) 67/92 (73) Across elbow joint (forearm with jump graft) 2/97 (2) 1/92 (1) Forearm and elbow 2/97 (2) 0/92 Configuration of graft no./total no. (%) 0.62 Looped 42/97 (43) 37/93 (40) Straight 55/97 (57) 56/93 (60) Arterial anastomosis no./total no. (%) 0.28 Axillary 2/97 (2) 2/93 (2) Brachial 92/97 (95) 87/93 (94) Radial 1/97 (1) 4/93 (4) Ulnar 0/97 0/93 Other 2/97 (2) 0/93 Venous anastomosis no./total no. (%) Axillary 22/97 (23) 30/93 (32) Basilic 56/97 (58) 51/93 (55) Brachial 14/97 (14) 3/93 (3) Cephalic 3/97 (3) 9/93 (10) Other 2/97 (2) 0/93 Previous access no./total no. (%) Arteriovenous access graft 55/94 (59) 50/90 (56) 0.77 Venous anastomosis 64/94 (68) 60/89 (67) 1.00 Venous outflow tract 40/90 (44) 29/86 (34) 0.17 Type of arteriovenous access graft no./total no. (%) 0.34 Tapered 14/80 (18) 10/77 (13) 0.51 Straight 53/80 (66) 61/77 (79) 0.08 Stepped 8/80 (10) 5/77 (6) 0.57 Other 5/80 (6) 1/77 (1) n engl j med 362;6 nejm.org february 11, 2010

6 Stent Graft vs. Balloon Angioplasty for failing Dialysis-Access Grafts Table 2. (Continued.) Characteristic Stent Graft (N = 97) Balloon Angioplasty (N = 93) P Value Graft diameter no./total no. (%) mm 0/81 0/80 6 mm 54/81 (67) 55/80 (69) mm 5/81 (6) 7/80 (9) mm 0/81 2/80 (2) mm/7 mm 17/81 (21) 16/80 (20) mm/8 mm 1/81 (1) 0/ mm/6 mm 0/81 0/ mm/6.5 mm 1/81 (1) 0/ Other 3/81 (4) 0/ Clinical sign of graft dysfunction no./total no. (%) 31/97 (32) 39/93 (42) 0.17 Elevated venous pressure during dialysis no./total no. (%) 25/97 (26) 35/93 (38) 0.08 Detection of decreased blood flow no./total no. (%) 24/97 (25) 17/93 (18) 0.30 Abnormal flow on Doppler ultrasonography no./total no. (%) 16/97 (16) 15/93 (16) 1.00 Angiographic data Target lesion length 0.21 No. with data Mean (±SD) length mm 35.3± ±12.7 Interpolated reference-vessel diameter 0.07 No. with data Mean (±SD) diameter mm 8.3± ±1.7 Mean lesion diameter 0.66 No. with data Mean (±SD) diameter mm 2.37± ±0.80 Target lesion preprocedure stenosis 0.17 No. with data Mean (±SD) % stenosis 70.9± ±9.0 * Plus minus values are means ±SD. Patients could have had a graft at more than one location and more than one previous procedure. Graft diameters separated by a slash represent the diameter at each end of a tapered graft. Clinical signs of graft dysfunction were defined as prolonged bleeding after needle withdrawal or altered pulse or thrill in graft. three most common triggers for intervention were clinical variables, elevated venous pressure during dialysis, and detection of decreased blood flow. Baseline Angiographic Characteristics of the Lesion There were no significant differences between the two treatment groups with respect to angiographic characteristics of the target lesion: interpolated reference-vessel diameter, minimum lumen diameter, or percent stenosis (Table 2). The average balloon diameters during dilation were similar in the stent-graft group and the balloonangioplasty group. The percentage of patients with secondary lesions in the access circuit was similar in the two groups: 39% of patients receiving a stent graft and 41% undergoing balloon angioplasty. Implantation of Stent Grafts A total of 125 stent grafts were implanted in 97 patients; 67% of the stent grafts were flared, 16% were straight, and 17% consisted of overlapping straight and flared grafts. Seventy-three patients n engl j med 362;6 nejm.org february 11,

7 The new england journal of medicine Table 3. Treatment Success and Patency End Points in the Intention-to-Treat Population, According to Treatment Group.* End Point Stent Graft Balloon Angioplasty no. of patients/total no. (%) P Value Anatomical success 91/97 (94) 68/93 (73) <0.001 Hemodynamic success 97/97 (100) 93/93 (100) Clinical success 85/97 (88) 78/93 (84) 0.49 Procedural success 91/97 (94) 68/93 (73) <0.001 Primary patency of treatment area 2 mo 77/96 (80) 71/92 (77) mo 46/91 (51) 20/86 (23) <0.001 Primary patency of access circuit 2 mo 76/96 (79) 71/92 (77) mo 35/92 (38) 17/86 (20) * Definitions of treatment success are listed in the Supplementary Appendix. Table 4. Adverse Events at 6 Months, According to Treatment Group.* Adverse Event Stent Graft Balloon Angioplasty no. of patients/total no. (%) P Value Infection 6/95 (6) 2/90 (2) 0.28 Thrombotic occlusion 31/95 (33) 19/90 (21) 0.10 Restenosis 38/95 (40) 69/90 (77) <0.001 Pseudoaneurysm 5/95 (5) 2/90 (2) 0.45 Vessel rupture 3/95 (3) 1/90 (1) 0.62 Hemorrhage 0/95 0/90 Hematoma 2/95 (2) 0/ Edema of arm or hand 3/95 (3) 2/90 (2) 1.00 Vascular insufficiency from graft 2/95 (2) 1/90 (1) 1.00 Congestive heart failure 4/95 (4) 2/90 (2) 0.68 Cerebrovascular accident 2/95 (2) 3/90 (3) 0.68 Death 5/95 (5) 5/90 (6) 1.00 Embolism 0/95 0/90 Problems with device Kinking 0/95 NA Migration 4/95 (4) NA Permanent deformation 1/95 (1) NA * NA denotes not applicable. (75%) received a single device. Stent grafts were overlapped when lesion lengths (including a nondiseased 10-mm landing zone between the arteriovenous graft and outflow vein) exceeded the longest available stent-graft length (50 mm). Device deployment was successful in 96 of the 97 patients (99%). A 30-mm stent graft was placed in 28% of the patients, a 40-mm graft in 36%, and a 50-mm graft in 36%. Follow-up of the Patients There were no significant differences between the two treatment groups regarding the attendance of patients at follow-up examinations at any study interval. Thirteen patients (6 of the 97 patients [6%] receiving a stent graft and 7 of the 93 patients [8%] undergoing balloon angioplasty) missed either the 2-month or 6-month follow-up evaluation: 2 patients (1 in each group) missed the 2-month follow-up visit and 11 (5 in the stentgraft group and 6 in the balloon-angioplasty group) missed the 6-month follow-up visit. A total of 94% of patients in the stent-graft group and 93% of patients in the balloon-angioplasty group completed the 6-month follow-up visit. Study End Points On the basis of an intention-to-treat analysis, at 6 months, the incidence of primary patency of the treatment area was significantly greater in the stent-graft group (51%) than in the balloonangioplasty group (23%) (P<0.001), as was the incidence of primary patency of the access circuit (38% vs. 20%, P = 0.008) (Table 3). Forty-five patients in the stent-graft group and 66 patients in the balloon-angioplasty group had loss of primary patency of the treatment area, owing to one or more of the following events: reintervention in the treatment area, thrombotic occlusion, surgical intervention that excluded the treatment area from the access circuit, and abandonment of the graft because of an inability to treat the primary lesions (Table 4). At 6 months, the percentage of patients with freedom from loss of primary patency of the treatment area was significantly greater in the stent-graft group than in the balloon-angioplasty group (P = by the log-rank test and P = by the Wilcoxon rank-sum test), as was the percentage with freedom from loss of primary patency of the access circuit (P = 0.03 by the log-rank test and P = 0.04 by the Wilcoxon rank-sum test) (Fig. 1). At 210 days, the stentgraft group showed superior freedom from subsequent interventions as compared with the balloon-angioplasty group (P = 0.03 by the log-rank test and P = 0.04 by the Wilcoxon rank-sum test). 500 n engl j med 362;6 nejm.org february 11, 2010

8 Stent Graft vs. Balloon Angioplasty for failing Dialysis-Access Grafts These findings supported both the primary and secondary study hypotheses: the noninferiority to and superiority of the stent graft as compared with balloon angioplasty. Procedural Success and Restenosis The rate of procedural success was significantly higher in the stent-graft group than in the percutaneous-transluminal-angioplasty group, with success in 94% versus 73% of patients (P<0.001) (Table 4). At 6 months, the minimum lumen diameter of the treatment area was, on average (mean ±SD), significantly greater with the stent graft than with balloon angioplasty alone: 5.1±1.5 mm vs. 3.3±1.5 mm (P<0.001). The average percent stenosis was lower in the stent-graft group (32.1±19.8%) than in the balloonangioplasty group (59.2±19.6%) (P<0.001). At 6 months, the incidence of binary restenosis (stenosis of >50% diameter) was significantly greater in the balloon-angioplasty group (78%) than in the stent-graft group (28%) (P<0.001). Restenotic lesions were, on average, significantly shorter in the stent graft group (18.0±12.5 mm) than in the balloon-angioplasty group (32.1±14.3 mm) (P<0.001). At 6 months, the presence or absence of remote secondary lesions that were required to have been treated before study enrollment did not negate the patency advantage of the stent graft over balloon angioplasty. The incidence of patency of the treatment area was greater, among patients who had secondary lesions, in the stentgraft group (44%) than in the balloon-angioplasty group (17%) (P = 0.02), as well as among patients who did not have secondary lesions (54% vs. 28%) (P = 0.006). Logistic-regression analysis of clinical variables at 6 months failed to identify any distinguishing or significant criteria affecting primary patency including diabetes, age, sex, history of hypercoagulability or glomerulonephritis, graft site, hypertension, or use of anticoagulation or antiplatelet therapy. By means of multiple logistic-regression analysis, the only criterion that was associated with primary patency at 6 months of follow-up was assignment to the stent-graft group (P<0.001). Safety and Adverse Events There was no significant difference in the incidence of reported adverse events between the A B Primary Patency of Treatment Area (%) Primary Patency of Access Circuit (%) P=0.003 P=0.03 balloon-angioplasty control group and the stentgraft group, except for the incidence of restenosis, which was higher with balloon angioplasty (77%, vs. 40% with stent graft; P<0.001) (Table 4). Discussion Days after Initial Procedure Stent graft Balloon angioplasty Balloon angioplasty Days after Initial Procedure Stent graft Figure 1. Estimated Percentages of Patients with Primary Patency during the Study Period, According to Treatment Group. Panel A shows the percentages with primary patency of the treatment area, and Panel B, percentages with primary patency of the access circuit. The P values shown were calculated with the use of the log-rank test. Despite nearly universal agreement that native fistulas should be the hemodialysis access of first choice, prosthetic grafts continue to play an important role in the creation of permanent hemodialysis-access circuits in patients in the United States. Though the percentage of patients undergoing dialysis through prosthetic grafts may continue to fall, the total number of patients with end-stage renal disease continues to grow each year. 1,3 Thus, hemodialysis grafts are likely to remain important vascular accesses, as planned bridges to native fistulas or in patients in whom fistulas have failed or cannot be created. 17,20-22 n engl j med 362;6 nejm.org february 11,

9 The new england journal of medicine Balloon angioplasty for stenotic arteriovenous grafts has limitations, however: the long-term durability of balloon angioplasty is limited 6,9,10 and may necessitate repeated invasive procedures 9-11 with attendant complications and costs. From 16 to 25% of hospital admissions of patients with end-stage renal disease in the United States are necessitated by complications related to a vascular access; the associated costs have been estimated at nearly 1 billion dollars per year. 1,4 Furthermore, the outpatient costs for patients with graft failure more than doubled between 1991 and Multiple devices and approaches for treating arteriovenous graft related stenosis have been used and reported on in retrospective series and prospective, randomized trials. To date, none have shown any benefit over balloon angioplasty. These techniques have included angioplasty with cutting or with ultrahigh-pressure balloons, brachytherapy, cryoplasty, anticoagulation therapy, placement of bare-metal stents, modified surgical techniques and graft configurations, and other pharmacologic approaches. Although preliminary research in minimizing the hyperplastic process at the time of graft creation by means of pharmacologic, cellular, or gene therapies appears promising, such efforts are in the early stages of evaluation or development. 7-15,23-29 The polytetrafluoroethylene self-expanding stent graft is a less-invasive endovascular approach for revision of failing prosthetic arteriovenous grafts that is intended to mimic open surgical revision of a graft. Unlike surgery, a percutaneous approach optimally allows for immediate use of the graft, which might obviate the need for interim catheter dialysis and its associated costs, risks of bloodstream infection, and other complications. The stent graft used in the present trial was designed to prevent both the elastic recoil that occurs after balloon angioplasty, thus sustaining the short-term gain in luminal patency an effect similar to that of uncovered (bare-metal) stents used in the access circuit and the late loss in luminal patency due to trans-stent growth of neointimal tissue. 13,28,29 The present endovascular approach also essentially converts the initial surgical end-to-side venous anastomosis into an end-to-end anastomosis, providing more laminar in-line flow and thus potentially reducing the turbulence and shear stress that contribute to the development of venous outflow stenosis. 7,8,32-36 Our study involved detailed assessments and definitions of patency. All patients underwent formal angiography at 2 and 6 months, regardless of the clinical graft function, allowing for uniform assessment of stenosis at the Angiographic Core Lab. The angiographic findings were an integral part of the study definition of patency. In contrast, most published studies have used the absence of clinical dysfunction alone as the measure of treatment success. 7-15,25-29 The definitions used in our randomized, controlled study may provide a more accurate measure of actual arteriovenous graft patency than previous retrospective studies of arteriovenous grafts. Prospective reports of the patency of autogenous fistulas have noted similar findings: the incidence of patency is lower when prospectively assessed than when retrospectively ascertained. 20,37,38 In conclusion, our data indicate that the expanded polytetrafluoroethylene self-expanding stent graft used in the present study is superior to balloon angioplasty for the treatment of arteriovenous access grafts that have venous anastomotic stenosis. As compared with balloon angioplasty, the stent graft was associated with graft function for a longer period before subsequent intervention and a graft lumen that had a greater diameter and had patency for a longer period. Supported by Bard Peripheral Vascular. Dr. Haskal reports receiving consulting fees from W.L. Gore and Associates and lecture fees from Bard Peripheral Vascular and holding stock in AngioDynamics; Dr. Trerotola, receiving consulting fees from W.L. Gore and Associates and Bard Peripheral Vascular; Dr. Dolmatch, receiving consulting and lecture fees from Bard Peripheral Vascular and royalties for the Flair Endovascular Stent Graft and serving as an expert witness for testimony concerning the healing characteristics of the Flair Endovascular Stent Graft; Drs. Schuman and Berman, receiving consulting and lecture fees from Bard Peripheral Vascular; Dr. Altman, receiving consulting fees from Bard Peripheral Vascular; Dr. McLennan, receiving consulting fees from Medtronic, Cook, and Bard Peripheral Vascular and grant support from Cook, Boston Scientific, Guerbet, and Medical International Research; Dr. Ross, receiving lecture fees from W.L. Gore and Associates, Hemosphere, Bard Peripheral Vascular, and Medrad Interventional Possis; and Dr. Vesely, receiving consulting fees from AngioDynamics, Elcam Medical, Spire Biomedical, and W.L. Gore and Associates and lecture fees from AngioDynamics and W.L. Gore and Associates. No other potential conflict of interest relevant to this article was reported. 502 n engl j med 362;6 nejm.org february 11, 2010

10 Stent Graft vs. Balloon Angioplasty for failing Dialysis-Access Grafts References 1. U.S. Renal Data System. USRDS 2009 annual data report: atlas of chronic kidney disease and end-stage renal disease in the United States. Bethesda, MD: National Institute of Diabetes and Digestive and Kidney Diseases, Vascular Access Work Group. Clinical practice guidelines for vascular access. Am J Kidney Dis 2006;48:Suppl 1:S248- S Chan MR, Sanchez RJ, Young HN, Yevzlin AS. Vascular access outcomes in the elderly hemodialysis population: a USRDS study. Semin Dial 2007;20: The economic cost of ESRD, vascular access procedures, and Medicare spending for alternative modalities of treatment. Am J Kidney Dis 1997;30:Suppl 1: S160-S Schwab SJ, Harrington JT, Singh A, et al. Vascular access for hemodialysis. Kidney Int 1999;55: NKF-DOQI clinical practice guidelines for vascular access: National Kidney Foundation-Dialysis Outcomes Quality Initiative. Am J Kidney Dis 1997;30:Suppl 3: S150-S Fillinger MF, Reinitz ER, Schwartz RA, Resetarits DE, Paskanik AM, Bredenberg CE. Beneficial effects of banding on venous intimal-medial hyperplasia in arteriovenous loop grafts. Am J Surg 1989; 158: Fillinger MF, Reinitz ER, Schwartz RA, et al. Graft geometry and venous intimalmedial hyperplasia in arteriovenous loop grafts. J Vasc Surg 1990;11: 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: [Erratum, Radiology 1995;196:582.] 10. Lilly RZ, Carlton D, Barker J, et al. Predictors of arteriovenous graft patency after radiologic intervention in hemodialysis patients. Am J Kidney Dis 2001;37: Lumsden AB, MacDonald MJ, Kikeri D, Cotsonis GA, Harker LA, Martin LG. Prophylactic balloon angioplasty fails to prolong the patency of expanded polytetrafluoroethylene arteriovenous grafts: results of a prospective randomized study. J Vasc Surg 1997;26: Misra S, Bonan R, Pflederer T, Roy- Chaudhury P. BRAVO I: a pilot study of vascular brachytherapy in polytetrafluoroethylene dialysis access grafts. Kidney Int 2006;70: Patel RI, Peck SH, Cooper SG, et al. Patency of Wallstents placed across the venous anastomosis of hemodialysis grafts after percutaneous recanalization. Radiology 1998;209: Rifkin BS, Brewster UC, Aruny JE, Perazella MA. Percutaneous balloon cryoplasty: a new therapy for rapidly recurrent anastomotic venous stenoses of hemodialysis grafts? Am J Kidney Dis 2005; 45(2):e27-e Sun S, Beitler JJ, Ohki T, et al. Inhibitory effect of brachytherapy on intimal hyperplasia in arteriovenous fistula. J Surg Res 2003;115: Aruny JE, Lewis CA, Cardella JF, et al. Quality improvement guidelines for percutaneous management of the thrombosed or dysfunctional dialysis access. J Vasc Interv Radiol 2003;14:S247-S Clinical practice guidelines and for vascular access. Am J Kidney Dis 2006;48: Suppl 1:S176-S Gray RJ, Sacks D, Martin LG, Trerotola SO. Reporting standards for percutaneous interventions in dialysis access. J Vasc Interv Radiol 2003;14:S433-S Blackwelder WC. Proving the null hypothesis in clinical trials. Control Clin Trials 1982;3: Biuckians A, Scott EC, Meier GH, Panneton JM, Glickman MH. The natural history of autologous fistulas as first-time dialysis access in the KDOQI era. J Vasc Surg 2008;47: Stevenson KB, Hannah EL, Lowder CA, et al. Epidemiology of hemodialysis vascular access infections from longitudinal infection surveillance data: predicting the impact of NKF-DOQI clinical practice guidelines for vascular access. Am J Kidney Dis 2002;39: NKF-K/DOQI clinical practice guidelines for vascular access: update Am J Kidney Dis 2001;37:Suppl 1:S137- S Luo Z, Akita GY, Date T, et al. Adenovirus-mediated expression of beta-adrenergic receptor kinase C-terminus reduces intimal hyperplasia and luminal stenosis of arteriovenous polytetrafluoroethylene grafts in pigs. Circulation 2005;111: Kohler TR, Toleikis PM, Gravett DM, Avelar RL. Inhibition of neointimal hyperplasia in a sheep model of dialysis access failure with the bioabsorbable Vascular Wrap paclitaxel-eluting mesh. J Vasc Surg 2007;45: Konner K. Paclitaxel a tool to prevent stenosis in vascular access for haemodialysis? Blood Purif 2006;24: Yevzlin AS, Conley EL, Sanchez RJ, Young HN, Becker BN. Vascular access outcomes and medication use: a USRDS study. Semin Dial 2006;19: Rajan DK, Platzker T, Lok CE, et al. Ultrahigh-pressure versus high-pressure angioplasty for treatment of venous anastomotic stenosis in hemodialysis grafts: is there a difference in patency? J Vasc Interv Radiol 2007;18: Hoffer EK, Sultan S, Herskowitz MM, Daniels ID, Sclafani SJ. Prospective randomized trial of a metallic intravascular stent in hemodialysis graft maintenance. J Vasc Interv Radiol 1997;8: Vogel PM, Parise C. SMART stent for salvage of hemodialysis access grafts. J Vasc Interv Radiol 2004;15: Guiu B, Loffroy R, Ben Salem D, et al. Angioplasty of long venous stenoses in hemodialysis access: at last an indication for cutting balloon? J Vasc Interv Radiol 2007;18: Vesely TM, Siegel JB. Use of peripheral cutting balloon to treat hemodialysis-related stenoses. J Vasc Interv Radiol 2005; 16: Roy-Chaudhury P, Kelly BS, Miller MA, et al. Venous neointimal hyperplasia in polytetrafluoroethylene dialysis grafts. Kidney Int 2001;59: Sottiurai VS, Yao JS, Batson RC, Sue SL, Jones R, Nakamura YA. Distal anastomotic intimal hyperplasia: histopathologic character and biogenesis. Ann Vasc Surg 1989;3: Longest PW, Kleinstreuer C. Numerical simulation of wall shear stress conditions and platelet localization in realistic end-to-side arterial anastomoses. J Biomech Eng 2003;125: Krueger U, Zanow J, Scholz H. Computational fluid dynamics and vascular access. Artif Organs 2002;26: Fillinger MF, Kerns DB, Bruch D, Reinitz ER, Schwartz RA. Does the end-toend venous anastomosis offer a functional advantage over the end-to-side venous anastomosis in high-output arteriovenous grafts? J Vasc Surg 1990;12: Huijbregts HJ, Bots ML, Wittens CH, Schrama YC, Moll FL, Blankestijn PJ. Hemodialysis arteriovenous fistula patency revisited: results of a prospective, multicenter initiative. Clin J Am Soc Nephrol 2008;3: Dixon BS, Beck GJ, Vazquez MA, et al. Effect of dipyridamole plus aspirin on hemodialysis graft patency. N Engl J Med 2009;360: Copyright 2010 Massachusetts Medical Society. n engl j med 362;6 nejm.org february 11,