CLINICAL STUDY

Long-Term Results of Stent-Graft Placement to Treat Central Venous Stenosis and Occlusion in Hemodialysis Patients With Arteriovenous Fistulas
Robert G. Jones, MBChB, MRCP, FRCR, Andrew P. Willis, MBBS, MRCS, FRCR, Catherine Jones, MBBS, BSc, FRCR, Ian J. McCafferty, MBBS, MRCP, FRCR, and Peter L. Riley, MBChB, MRCP, FRCR

ABSTRACT Purpose: To determine the effectiveness of stent-grafts for the treatment of central venous disease in hemodialysis patients with functioning arteriovenous (AV) stulas. Materials and Methods: Between October 2004 and March 2010, 42 VIABAHN stent-grafts were deployed in central veins of 30 patients (16 men, 14 women; mean age 60 y) with functioning AV stulas and central venous disease that did not respond to percutaneous transluminal angioplasty (PTA). Eighteen patients had central vein stenosis and 12 had occlusion. Previous PTA and/or bare metal stent placement had been performed in 23 patients (77%). Surveillance was carried out at 3, 6, 9, 12, 18, and 24 months with diagnostic stulography. The mean follow-up was 705 days (range, 66 1,645 d). Statistical analysis included KaplanMeier and log-rank studies. Results: Technical success rate was 100%. Primary patency rates were 97%, 81%, 67%, and 45% at 3, 6, 12, and 24 months, respectively. Primary assisted patency rates were 100%, 100%, 80%, and 75% at 3, 6, 12, and 24 months, respectively. Patients without previous procedures had signicantly shorter times to repeat intervention (P .018) than those who had undergone PTA or bare metal stent placement previously. Patients with occlusive lesions had a signicantly shorter primary patency interval (P .05) than patients with stenoses. Occluded veins were more likely to require further stent-grafts (P .02). Twelve patients required further stent-grafts to maintain patency. There was one minor complication. Conclusions: Stent-graft placement to treat central venous disease in hemodialysis patients with autogenous AV stulas is safe and effective if PTA fails to maintain luminal patency.

ABBREVIATIONS
AV arteriovenous, PTA percutaneous transluminal angioplasty

Central venous disease is a commonly encountered problem in the hemodialysis population and results in signicant morbidity. It is often clinically apparent in those with functioning hemodialysis access, either as an autogenous arteriovenous (AV) stula or an AV graft. Incidences are typically quoted between 1.5% and 17% but have been as high as 40% (13). AV stula or graft thrombosis can occur
From the Radiology Department, Queen Elizabeth Hospital Birmingham, University Hospital Birmingham NHS Trust, Edgbaston, Birmingham B15 2TH, United Kingdom. Received September 21, 2010; nal revision received May 29, 2011; accepted June 7, 2011. Address correspondence to R.G.J.; E-mail: robert.jones@uhb.nhs.uk None of the authors have identied a conict of interest.

secondary to central venous disease, resulting in loss of dialysis access altogether or necessitate surgical ligation for symptomatic control. Central venous disease is thought to result from local trauma invariably as a consequence of repeated percutaneous hemodialysis catheter placement and altered ow dynamics (4,5). Endovascular management is well established
From the SIR 2007 Annual Meeting. SIR, 2011 J Vasc Interv Radiol 2011; xx:xxx DOI: 10.1016/j.jvir.2011.06.002

Stent-Grafts in Hemodialysis Recipients With AV Fistulas

Jones et al

JVIR

and accepted in this setting, and includes percutaneous transluminal angioplasty (PTA) and/or bare metal stent placement, with no clear advantage of one modality versus the other (6 11). Signicant recurrent stenosis rates are reported, and the condition is thought to result from neointimal hyperplasia. The use of stent-grafts in AV stula salvage has been described in several small series with encouraging results (1215). More recently, stent-graft placement has been described as the gold standard for the treatment of AV graft venous anastomosis stenosis (16). With these observations in mind, it seems feasible that the use of stent-grafts is applicable to the central venous scenario. The objective of the present study was to evaluate patency and performance of stent-grafts as an alternative to bare metal stents in central venous disease that does not respond to PTA in patients with upper-limb AV stulas.

MATERIALS AND METHODS Study Design

This study is a retrospective analysis of data collected between October 2004 and March 2010 pertaining to the use of VIABAHN (W.L Gore and Associates, Flagstaff, Arizona) stent-grafts as an alternative to bare metal stents to treat central venous disease that did not respond to PTA in patients with ipsilateral upper-limb AV stulas who were undergoing hemodialysis at a single institution. Included were those with recurrent stenosis previously treated successfully with PTA, in-stent stenosis in previously placed bare metal stents, and/or recurrent stenosis in association with stent-grafts (including 0.5 cm to either side of the device, ie, edge stenosis) placed during the review period. Patients were identied via our venographic surveillance program or if symptoms of upper-limb venous hypertension developed or dialysis venous pressures were increased. A normal coagulation prole was required at the time of the procedure, and any form of access infection or concurrent systemic infection (including bacteremia) was considered a contraindication to stent-graft insertion. Surveillance was carried out at 3, 6, 9, 12, 18, and 24 months with diagnostic stulography, and information regarding resolution of symptoms was also collected at these intervals. The study was compliant with institutional guidelines pertaining to the retrospective collection and analysis of data.

beyond the lesion and the wire was exchanged for a long stiff guide wire (Amplatz Super-stiff wire; Boston Scientic, Natick, Massachusetts). The vascular access sheath was then exchanged according to PTA balloon and subsequent stent-graft compatibility. If direct AV stula access was required, consideration was given to the application of a purse-string suture to prevent prolonged bleeding on removal of the access sheath. Periprocedural antibiotic prophylaxis was not used. In all cases, 2,000 U of heparin was administered intravenously before the intervention. PTA was carried out with a high-pressure balloon in all cases (Conquest or Atlas balloon; Bard, Covington, Georgia), and a stentgraft was subsequently placed if residual stenosis of 50% or more was observed. Venous pressure gradients were measured in cases of borderline residual stenosis at the discretion of the individual interventional radiologist. A gradient across the lesion or more than 10% systolic pressure was considered signicant, and in this situation, a stent-graft was placed. The threshold used in the present study has been reported to dene signicant central venous stenosis (17). A 10% oversizing versus the normal adjacent venous diameter measured on venography was used when sizing balloons and stent-grafts, and an equivalent diameter was used for deployment in an existing bare metal stent. Stent-grafts were not considered if deployment would involve covering patent ipsilateral internal jugular and contralateral brachiocephalic veins. Patients did not routinely undergo anticoagulation, and no additional specic medical therapy was instigated after stent-graft placement.

Denitions
Technical success of stent-graft placement was dened as exclusion of a stenosis or occlusion without signicant residual stenosis ( 30%). Clinical success was dened as resolution of signs and symptoms related to upperlimb venous hypertension. Technical failure of PTA was dened by the presence of residual stenosis of at least 50%. Primary patency was dened as stent-graft patency without the need for repeat intervention (balloon dilation or further stent-graft placement) within the existing stentgraft, including 0.5 cm to either side (ie, edge segment). The time point marking the end of primary patency was the date of the rst repeat intervention. Primary assisted patency was dened as stent-graft patency irrespective of the need for, or number of, further interventions within the stent-graft or edge segment. Secondary patency was dened as occlusion of the central vein.

Procedural Technique
Initial central venous assessment was carried out with digital subtraction venography via direct puncture of the venous arm of the upper-limb AV stula. The femoral route was chosen primarily for subsequent intervention. A multipurpose catheter and hydrophilic guide wire combination was used to probe occluded veins, and, when they had been crossed, a 4-F catheter was advanced

Statistical Analysis
Statistical analysis was performed by using SPSS statistical software (version 17; SPSS, Chicago, Illinois). Means and SDs were calculated for continuous variables. Survival

Volume xx

Number x

Month

2011

Table 1. Demographic Data on Patients, Venous Lesions, and Previous Interventions Characteristic Age (y) Mean Range Sex Male Female Central venous lesion Occlusion Stenosis Site of venous lesion Left brachiocephalic Left subclavian Right brachiocephalic Right subclavian Site of stent-graft Left brachiocephalic Left subclavian Right brachiocephalic Right subclavian Previous intervention None PTA Bare metal stent Note.Values in parentheses are percentages. PTA cutaneous transluminal angioplasty. Value 60 2788 16 14 12 (40) 18 (60) 24 (75) 3 (9) 4 (12) 1 (4) 31 (74) 3 (7) 7 (17) 1 (2) 7 21 12 per-

Table 2. Number of Stent-Grafts and Dimensions According to Target Vein Stent-Graft Dimensions (mm Site Left BCV Right BCV Left SCV Right SCV Note.BCV 13 5 12 4 1 11 5 12 2 3 10 1 10 10 6 5 cm) 9 10 1

brachiocephalic vein; SCV

subclavian vein.

curve analysis with KaplanMeier methodology was used to demonstrate event-free outcomes of primary and primary assisted patency. Log-rank statistics were used to compare event-free outcomes between subgroups of patients. Follow-up data were censored from KaplanMeier analysis if the patient was lost to follow-up for logistical reasons or died from unrelated causes, if no event had occurred at the time of data retrieval, or if renal transplantation occurred for reasons unrelated to stent failure. A P value of .05 or lower was considered to represent statistical signicance.

RESULTS
In the period between October 2004 and March 2010, 16 men and 14 women (mean age, 60 y; range, 27 88 y) underwent stent-graft placement (Table 1). The mean follow-up period was 705 days (range, 66 1,645 d). One patient was lost to follow-up.

Technical and Clinical Outcomes

The technical success rate was 100%. The mean time for each case was 102 minutes (range, 45 min to 2 h). Initially, 30 stent-grafts were placed in 30 patients. Twenty-one patients had previous PTA, 12 had bare

metal stents (with true in-stent stenosis), and seven had received a stent-graft at rst intervention after unsuccessful PTA. Eighteen of the 30 patients required further intervention to maintain patency during the study period. Twelve of these patients required further stent-graft placement (at a mean of 1,016 d) to maintain patency after failed repeat PTA, 10 had edge stenosis, and two had true intrastent-graft stenosis. Stent-graft deployment in all cases (initial and repeat) was achieved via the right common femoral vein in 31 patient episodes, the left common femoral vein in one, the AV stula in three, and via through-and-through access in seven patients with occlusive venous disease. To avoid prolonged bleeding and potential thrombotic complications that could be associated with the larger vascular sheaths required, the femoral approach for stent-graft deployment was favored versus direct AV stula access. The following sizes of VIABAHN stent-grafts were used: 13 mm 5 cm (n 17), 11 mm 5 cm (n 17), 10 mm 5 cm (n 6), 10 mm 10 cm (n 1), and 9 mm 10 cm (n 1; Table 2). Thirty-eight stent-grafts (90%) were postdilated per manufacturers instructions for use. No migration was observed in those that were not dilated, and postdeployment venography did not suggest luminal irregularity, although we recognize this can be overlooked on venography. No stent-graft fractures were observed in the study period. All patients who initially presented with arm swelling reported complete resolution within 7 days of stent-graft placement, and this was recorded at the rst surveillance stulography visit on direct questioning. There was a single complication in one patient in whom a stent-graft was deployed across a venous collateral vessel supporting head and neck drainage, but there were no signicant clinical sequelae. No stent-graftrelated infections were encountered. There were 10 deaths (33%) during the study period, but none were procedure-related or occurred within 30 days of stent-graft placement.

Patency
Primary patency rates for all initial stent-grafts (excluding those that required a further stent-graft) were 97%, 81%,

Stent-Grafts in Hemodialysis Recipients With AV Fistulas

Jones et al

JVIR

Figure 2. KaplanMeier primary patency after stent-graft placement in central venous disease in patients with and without previous central venous intervention.

Figure 1. Primary patency after stent-graft placement for the 30 original stent-graft placements, not including those that required a further stent-graft to maintain patency. (Available in color online at www.jvir.org.)

Table 3. Primary and Assisted Primary Patency After Stent Placement Follow-up (mo) 3 6 9 12 18 24 No. of Pts 29 25 20 19 13 12 Primary Patency (%) 97 (3.3) 81 (7.6) 77 (8.3) 67 (9.7) 51 (11) 45 (11) Primary Assisted Patency (%) 100 (NA) 100 (NA) 95 (4.4) 94 (NA) 92 (NA) 91 (NA)

Note.Values in parentheses represent SE. NA not applicable (because only one patient did not have primary assisted patency at 9-month follow-up).

Figure 3. KaplanMeier primary patency after stent-graft placement for occlusive and stenotic lesions.

67%, and 45% at 3, 6, 12, and 24 months (Fig 1). Primary assisted patency rates were 100%, 100%, 80%, and 75% at 3, 6, 12, and 24 months, respectively. Patency rates and event-free survival analysis is shown in Table 3. Log-rank analysis showed that patients undergoing stent-graft placement without previous procedures (n 7) had signicantly shorter time to repeat intervention (P .018) than those who had undergone previous PTA or bare metal stent insertion (n 23; Fig 2). Patients with occlusive lesions (n 12) also had a signicantly shorter primary patency interval (P .05) than patients with stenoses (n 18; Fig 3) and were more likely (P .02) to require a further stent-graft to maintain primary assisted patency. Primary assisted patency could not be maintained in only one patient, in whom patency was lost at 259 days after the initial

stent-graft procedure. A survival curve was not created for that reason.

DISCUSSION
PTA and bare metal stent placement are recognized and established methods of treatment of hemodialysis-associated central venous disease. PTA alone has been reported to be associated with primary patency rates at 3, 6, and 12 months as high as 58%, 23% 63%, and 12%53%, respectively (7,18,19). Central venous lesions that exhibit elastic recoil after PTA may show early failure and often require stent placement to maintain luminal patency, as this provides additional structural support. Stent placement is rec-

Volume xx

Number x

Month

2011

Figure 4. (a) Diagnostic central venogram in a patient with an existing stent-graft in the left brachiocephalic vein that had been placed 6 months earlier. Note the edge stenosis (arrow), which required a second stent-graft to maintain patency (b).

ommended in this scenario or when there is a recurrence of stenosis within 3 months (20,21). As with PTA, similar trends of disease recurrence with bare metal stents are reported, with patency rates at 3, 6, and 12 months ranging from 63% to 100%, 42% to 89%, and 14% to 73%, respectively (19). The application of stent-grafts in hemodialysis autogenous AV stula circuits has been reported in several series with encouraging results (1215,22), but these studies mainly address their function in the peripheral arm of the AV stula and in the setting of stula salvage following PTA-associated rupture or pseudoaneurysm formation. Application of stent-grafts in the central veins has been reported by Quinn et al (23), who used preexpanded polytetrauoroethylene sewn onto Palmaz stents (Johnson and Johnson, Warren, New Jersey). In this study (23), only six central venous stenoses were treated, with 2-, 6-, and 12-month primary patency rates of 40%, 32%, and 32%, respectively. In addition to the mechanical support offered by the metallic stent component of the stent-graft, the polytetrauoroethylene lining may provide a more compatible and less turbulent luminal environment for endothelialization than a bare metal stent. In animal models, it is suggested that stent-grafts maintain longer patency rates than bare metal stents (24). Most historical studies looking at PTA and bare metal stent outcomes (7,10) lack the interval 3-month venographic follow-up carried out in the present series. Although a more critical analysis would be required to make an accurate comparison, our stent-graft patency rates compare favorably with these outcomes, and, in particular, our primary assisted patency rates exceed previously reported results (9 11). As with PTA and bare metal stents, we observed recurrent stenosis in our stent-graft series, and this typically occurred within 0.5 cm of the edge of the stent-graft (Fig 4). In two cases of recurrent stenosis, further stent-graft placement was required for true intrastent-graft stenosis that did not

respond to repeat PTA. Similar observations of edge stenosis have been reported previously (25), which may have resulted from turbulent ow in this region. The central portion of the stent-graft is less likely to show intimal hyperplasia and may be a more favorable environment for endothelialization. In our sizing protocol, the diameter of the normal adjacent vein as measured on digital subtraction venography was used as a reference, with a 10% oversizing of the stent-graft diameter. In cases of in-stent stenosis, we used a stent-graft of the same diameter as the preexisting stent. A larger-diameter device would be constrained by the bare metal stent, whereas an equivalent diameter is further supported by an outer circumference of intimal hyperplasia. Length measurement included an additional 0.5 cm to either side of the lesion. Deployment across contralateral brachiocephalic or ipsilateral jugular veins was avoided except when these were already occluded as demonstrated by sonographic and venographic studies (Fig 5). Dominant venous collateral vessels were also considered in this fashion, and one complication was encountered in the present series in which a stent-graft compromised collateral drainage of the head and neck, which resulted in head swelling. Other smaller collateral vessels were present and developed further within 24 hours, and the swelling resolved spontaneously, obviating further intervention. The limitations of the present study include the cohort size and the lack of correlation with measurable dialysis function; however, all patients reported clinical improvement after stent-graft placement. Another limitation is the lack of comorbidity data, as this can be associated with increased risk of recurrent stenosis after central venous interventions (7,10). In conclusion, this retrospective review suggests that the application of stent-grafts to treat central venous disease unresponsive to PTA in hemodialysis patients with functioning AV stulas is safe and effective. However, larger

Stent-Grafts in Hemodialysis Recipients With AV Fistulas

Jones et al

JVIR

Figure 5. Central venography in a patient with a functioning right brachiocephalic AV stula and a swollen right arm. (a) Diagnostic venography demonstrates in-stent stenosis within a right brachiocephalic vein bare metal stent, which had been placed 8 months earlier. (b) No improvement was noted after PTA, and therefore a stent-graft was placed within the existing bare metal stent. A satisfactory result was obtained, followed by resolution of arm swelling within 7 days.

prospective studies are required to accurately determine their exact role.

ACKNOWLEDGEMENT
The authors acknowledge Dr. M. Dhillon for his input during the early phase of this work.

REFERENCES
1. Hernandez D, Diaz F, Runo M, et al. Subclavian vascular access stenosis in dialysis patients: natural history and risk factors. J Am Soc Nephrol 1998; 9:15071510. 2. Lumsden AB, MacDonald MJ, Isiklar H, et al. Central venous stenosis in the hemodialysis patient: incidence and efcacy of endovascular treatment. Cardiovasc Surg 1997; 5:504 509. 3. Schillinger F, Schillinger D, Montagnac R, Milcent T. Central venous stenosis in hemodialysis: comparative angiographic study of subclavian and internal jugular access. Nephrologie 1994; 15:129 131. 4. Morosetti M, Meloni C, Gardini R, et al. Late symptomatic venous stenosis in three hemodialysis patients without previous central venous catheters. Artif Organs 2000; 24:929 931. 5. Roy-Chaudhury P, Lee TC. Vascular stenosis: biology and interventions. Curr Opin Nephrol Hypertens 2007; 16:516 522. 6. Surowiec SM, Fegley AJ, Tanski WJ, et al. Endovascular management of central venous stenosis in the hemodialysis patient: results of percutaneous therapy. Vasc Endovasc Surg 2004; 38:349 354. 7. Bakken AM, Protack CD, Saad WE, Lee DE, Waldman DL, Davies MG. Long-term outcomes of primary angioplasty and primary stenting of central venous stenosis in hemodialysis patients. J Vasc Surg 2007; 45:776 783. 8. Quinn SF, Schuman ES, Demlow TA, et al. Percutaneous transluminal angioplasty versus endovascular stent placement in the treatment of venous stenosis in patients undergoing hemodialysis: intermediate results. J Vasc Interv Radiol 1995; 6:851 855. 9. Haage P, Vorwerk D, Piroth W, Schuermann K, Guenther RW. Treatment of hemodialysis related central venous stenosis or occlusion: results of primary Wallstent placement and follow up in 50 patients. Radiology 1999; 212:175180. 10. Nael K, Kee ST, Solomon H, Katz SG. Endovascular management of central thoracic veno-occlusive diseases in hemodialysis patients: a single institutional experience in 69 consecutive patients. J Vasc Interv Radiol 2009; 20:46 51.

11. Ozyer U, Harman A, Yildrim E, Aytekin C, Karakayali F, Boyvat F. Long-term results of angioplasty and stent placement for treatment of central venous obstruction in 126 hemodialysis patients: a 10-year single center experience. AJR Am J Roentgenol 2009; 193:1672 1679. 12. Donaldson J, Borzatta MA, Belville J. Covered stents for the treatment of venous-to-graft stenosis: one year later. J Vasc Nurs 2003 21:114 116. 13. Belville J, Borzatta MA. Covered stents for treatment of arteriovenous graft stenosis. vascular access for hemodialysis VIII. Access Medical Press 2002: 10:117121. 14. Ross JR. Non surgical management of vascular access stenosis, aneurysms, and extravasation with the Viabahn exible self-expanding stent-graft. Endovasc Today 2003; October:29 33. 15. Bent CL, Rajan DK, Tan K, et al. Effectiveness of stent-graft placement for salvage of dysfunctional arteriovenous hemodialysis stulas. J Vasc Interv Radiol 2010; 21:496 502. 16. 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. 17. Rajan DK, Chennepragada SM, Lok CE, et al. Patency of endovascular treatment for central venous stenosis :is there a difference between dialysis stulas and grafts ? J Vasc Interv Radiol 2007; 18:353359. 18. National Kidney Foundation. NKF-K/DOQI clinical practice guidelines for vascular access. Am J Kidney Dis 2001; 37(Suppl 1):S137S181. 19. Kundu S. Review of central venous disease in hemodialysis patients. J Vasc Interv Radiol 2010; 21:963968. 20. Gray RJ, Horton KM, Dolmatch BL, et al. Use of Wallstents for hemodialysis access-related venous stenoses and occlusions untreatable with balloon angioplasty. Radiology 1995; 195:479 484. 21. Aruny JE, Lewis CA, Cardella JF, et al. Quality improvement guidelines for percutaneous management of the thrombosed or dysfunctional dialysis access. Standards of Practice Committee of the Society of Cardiovascular and Interventional Radiology. J Vasc Interv Radiol 1999; 10:491 498. 22. Vesely TM. Use of stent-grafts to repair hemodialysis graft-related pseudoaneurysms. J Vasc Interv Radiol 2005; 16:13011307. 23. Quinn SF, Kim J, Sheley RC. Transluminally placed endovascular grafts for venous lesions in patients on hemodialysis. Cardiovasc Intervent Radiol 2003; 26:365369. 24. Dolmatch B, Dong YH, Heeter Z. Evaluation of three polytetrauoroethylene stent-grafts in a model of neointimal hyperplasia. J Vasc Interv Radiol 2007; 18:527534. 25. Tricht IV, Wachter DD, Tordoir J, Verdonck P. Hemodynamics and complications encountered with arteriovenous stulas and grafts as vascular access for hemodialysis: a review. Ann Biomech Eng 2005; 33:11421157.