Use of PTFE Stent Grafts for Hemodialysis-related Central Venous Occlusions: Intermediate-Term Results

Transcription

1 Cardiovasc Intervent Radiol (2011) 34: DOI /s CLINICAL INVESTIGATION Use of PTFE Stent Grafts for Hemodialysis-related Central Venous Occlusions: Intermediate-Term Results Sanjoy Kundu Milad Modabber John M. You Paul Tam Gordon Nagai Robert Ting Received: 12 August 2010 / Accepted: 5 October 2010 / Published online: 12 November 2010 Ó Springer Science+Business Media, LLC and the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) 2010 Abstract Purpose To assess the safety and effectiveness of a polytetrafluoroethylene (PTFE) encapsulated nitinol stents (Bard Peripheral Vascular, Tempe, AZ) for treatment of hemodialysis-related central venous occlusions. Materials and Methods Study design was a single-center nonrandomized retrospective cohort of patients from May 2004 to August 2009 for a total of 64 months. There were 14 patients (mean age 60 years, range years; 13 male, 1 female). All patients had autogenous fistulas. All 14 patients had central venous occlusions and presented with clinical symptoms of the following: extremity swelling (14%, 2 of 14), extremity and face swelling (72%, 10 of 14), and face swelling/edema (14%, 2 of 14). There was evidence of access dysfunction with decreased access flow in 36% (5 of 14) patients. There were prior interventions or previous line placement at the site of the central venous lesion in all 14 patients. Results were assessed by recurrence of clinical symptoms and function of the access circuit (National Kidney Foundation recommended criteria). S. Kundu (&) M. Modabber Department of Medical Imaging, Scarborough General Hospital General Division, Scarborough, ON, Canada sanjoy_kundu40@hotmail.com S. Kundu The Vein Institute of Toronto, Toronto, ON, Canada J. M. You Department of Vascular Surgery, Scarborough General Hospital General Division, Scarborough, ON, Canada P. Tam G. Nagai R. Ting Department of Nephrology, Scarborough General Hospital General Division, Scarborough, ON, Canada Results Sixteen consecutive straight stent grafts were implanted in 14 patients. Average treated lesion length was 5.0 cm (range, cm). All 14 patients had complete central venous occlusion (100% stenosis). The central venous occlusions were located as follows: right subclavian and brachiocephalic vein (21%, 3 of 14), right brachiocephalic vein (36%, 5 of 14), left brachiocephalic vein (36%, 5 of 14), and bilateral brachiocephalic vein (7%, 1 of 14). A total of 16 PTFE stent grafts were placed. Ten- or 12-mm-diameter PTFE stent grafts were placed. The average stent length was 6.1 cm (range, 4 8 cm). Technical (deployment), anatomic (\30% residual stenosis), clinical (resolution of symptoms), and hemodynamic (resolution of access dysfunction) success were 100%. At 3, 6, and 9 months, primary patency of the treated area and access circuit were 100% (14 of 14). Conclusions This PTFE encapsulated stent graft demonstrates encouraging intermediate-term patency results for central vein occlusions. Further prospective studies with long-term assessment and larger patient populations will be required. Keywords Hemodialysis Central venous disease Covered stents Abbreviations HD Hemodialysis PTA Percutaneous transluminal angioplasty Introduction Central venous stenosis and obstruction (central venous disease) is an important and prevalent problem in the

2 950 S. Kundu et al.: Stent Grafts for Central Venous Occlusions management of hemodialysis (HD) patients. Central venous disease compromises the integrity of the HD access circuit by causing venous hypertension with/without debilitating symptoms. This can result in loss of the access site as a result of access dysfunction or ligation for symptom relief. The incidence of central venous disease has been reported in the range of 30% in the literature [1 3]. There has been a strong association of central venous disease with previous placement of central venous catheters and pacemaker wires [2, 3]. Furthermore, there is a very high incidence of central venous disease in patients with a history of subclavian catheters [4 7]. A suggested mechanism for the development of central venous disease includes central venous catheter induced trauma to the venous endothelium and secondary inflammatory damage within the vessel wall at the time of insertion. There has been extensive literature on the treatment of this important and prevalent problem. Endovascular treatment options include percutaneous transluminal angioplasty (PTA) and bare metal stents (BMSs). Unfortunately, all the available treatment options have demonstrated variable rates of patency, thus requiring repeated intervention [8 12]. To date, no proven advantages have been observe in BMSs over PTA, leading to widely varied methodology for the treatment of central venous disease in HD patients [13 15]. In this study, we retrospectively assessed the intermediate-term outcomes of covered stent (CS) placement in HD patients over a 64-month period. The purpose of our study was to assess the safety and efficacy of a polytetrafluoroethylene (PTFE)-encapsulated nitinol stent (Bard Peripheral Vascular, Tempe, AZ) for the treatment of HD-related central venous occlusions in patients with clinical symptoms or access dysfunction. Materials and Methods Study Design Institutional review board approval was obtained for this retrospective study. Study design was a single-center nonrandomized retrospective cohort of HD patients from May 2004 to August 2009 for a total of 64 months in all consecutive patients who had CSs placed for central vein occlusions. The patient charts, digital database, and electronic imaging records were retrospectively reviewed to identify the type and dates of percutaneous intervention, access history, demographics, presenting symptoms and indications for intervention, existing comorbidities (diabetes mellitus, hypertension), and risk factors such as previous central venous catheter placement. There were 14 patients (mean age 60 years, range years; 13 male, 1 female), all of whom had autogenous fistulas. All 14 patients had central venous occlusions and presented with clinical symptoms of the following: extremity swelling (14%, 2 of 14), extremity and face swelling (71%, 10 of 14), and face swelling/edema (14%, 2 of 14). There was evidence of access dysfunction with decreased access flows measuring \650 ml/min when we used an ultrasound dilution flow method in 36% (5 of 14) patients as per our institutional protocol. There was previous line placement that bridged the site of the central venous lesion in all 14 patients. There were previous percutaneous intervention in 6 (43%) of the 14 patients at the site of the central venous lesion. Of the 6 patients with previous intervention, all had previous central vein angioplasties at the site of eventual occlusion. The types of dialysis access were radiocephalic fistula in 9 cases and brachiocephalic fistula in 5. The age of the HD access ranged from 1 to 7 years. The cause of renal failure was as follows: essential hypertension (n = 8), polycystic kidney disease (n = 2), glomerulonephritis (n = 3), and Wegner granulomatosis (n = 1). Six (43%) of the 14 patients had hypertension, 7 (50%) had diabetes mellitus, and 6 (43%) had significant coronary artery disease. Technique All procedures in this study were performed by one operator. All patients underwent an initial diagnostic fistulogram by ultrasound-guided percutaneous puncture using a micropuncture kit (Cook, Bloomington, IN) into the outflow vein, within 3 cm of the anastomosis. The access circuit was evaluated via power injector contrast digital subtraction venography from the arterial anastomosis to the right atrium. No peripheral venous outflow or arterial anastomotic stenoses or occlusions were identified in these patients outside of the central venous occlusions. Upon identification of a central vein occlusion, a second ultrasound-guided percutaneous micropuncture access was obtained in a upper arm vein approximately 7 cm below the axilla level for radiocephalic fistulas (brachial vein, n = 5; cephalic vein, n = 2; basilic vein, n = 2), and more proximally in the outflow vein in brachiocephalic fistulas (n = 5). Bilateral upper arm venous access was obtained in one patient in the basilic veins of both arms. Further central venograms in multiple projections with high magnification were performed to characterize the length, confirm the occlusion, and determine the diameter of the normal vein. In all patients, an initial 5F, 5-cm Terumo sheath (Terumo, Somerset, NJ) was placed in the upper arm venous access over a inch, 3-mm J tip, 145-cm-long Rosen wire (Meditech/Boston Scientific, Natick, MA). In 5 patients, the central venous occlusions were crossed with the above inch Rosen wire and a 5F, 40-cm-long Kumpe catheter (Cook). The time to cross the lesion for

3 S. Kundu et al.: Stent Grafts for Central Venous Occlusions 951 these 5 patients ranged from 5 to 12 min with a mean duration of 8 min. Once the lesion was crossed, the Rosen wire was exchanged for a inch Super Stiff Amplatz with a 3-mm J tip, 5-cm floppy distal length, and 260-cm length (Meditech/Boston Scientific). The tip of the Super Stiff Amplatz wire was placed into the inferior vena cava below the insertion of the renal veins with a 5F, 65-cm-long multipurpose angled catheter (MPA) (Cook). In a further 6 patients, the central vein occlusion could not be crossed with the above methodology. In such case, the sheath in the upper arm venous access site was exchanged for a 5F, 30-cm-long Abrahams sheath with an angled tip (Cook). The tip of this sheath was placed as close to the occlusion as possible. Through this sheath, a 5F angled 65-cm Glide catheter (Terumo) was placed. The Rosen wire was then exchanged for a inch angled or straight, 150-cm-long hydrophilic Glide wire (Terumo). The central venous occlusion was crossed using the above combination in 6 patients. Once the lesion was crossed, the Rosen wire was exchanged for a inch Super Stiff Amplatz with a 3-mm J tip, 5-cm floppy distal length, and 260-cm length (Meditech/Boston Scientific). The tip of the Super Stiff Amplatz wire was placed into the inferior vena cava, below the insertion of the renal veins with a 5F, 65-cm-long MPA catheter (Cook). The time to cross the lesion ranged from 18 to 94 min, with a mean of 48 min. In a further 2 patients, the methodologies stated above were unsuccessful. Therefore, a third venous access was obtained under ultrasound guidance using a micropuncture kit (Cook) into the right common femoral vein. A 5F, 5-cm Terumo sheath was then placed in the right common femoral venous access over a inch, 3-mm J tip, 145-cm-long Rosen wire. An 80-cm MPA catheter was then placed over the Rosen wire. Either a inch angled or straight, 150-cmlong hydrophilic Glide wire was placed through the MPA catheter. The MPA catheter and hydrophilic Glidewire combination via a right common femoral vein access were used to cross central venous occlusions in the 2 patients. Once the lesion was crossed, the Rosen wire was exchanged for a inch Super Stiff Amplatz with a 3-mm J tip, 5-cm floppy distal length, and 260-cm length (Meditech/Boston Scientific). The tip of the Super Stiff Amplatz wire was placed into the distal subclavian vein with a 5F, 100-cm-long MPA catheter. The time to cross the lesion ranged from 105 to 138 min, with a mean of 126 min. All the above techniques were unsuccessful in one patient. Therefore, a 11.5-mm occlusion balloon, 80 cm long (Meditech/Boston Scientific), was placed from the right common femoral vein access through an 8F, 5-cm Terumo sheath (Terumo) below the central vein occlusion. From the upper arm venous access through the Abraham sheath, a trocar cannula from a 6F drainage catheter (Cook) was inserted over a inch angled, 150-cm-long hydrophilic Glide wire. After insertion of the trocar needle into the trocar cannula, a sharp needle cannulation was used to cross the central vein occlusion using the occlusion balloon as a target. Once the lesion was crossed, the hydrophilic Glidewire was exchanged for a inch Super Stiff Amplatz with a 3-mm J tip, 5-cm floppy distal length, and 260-cm length. The tip of the Super Stiff Amplatz wire was placed into the inferior vena cava below the insertion of the renal veins with a 5F, 65-cm-long MPA catheter. The time to cross the lesion was 183 min in this single patient. The upper arm venous access sheath(s) was then exchanged for an 8F, 5-cm-long Terumo sheath in the 12 patients who were crossed from an upper arm access. In the 2 patients crossed from the right common femoral vein access, the venous access sheath was then exchanged for an 8F, 5-cm-long Terumo sheath. A noncompliant ultra-highpressure balloon (Conquest; Bard), with a diameter of 12 mm and length of 40 mm, was used to dilate the central vein occlusions with inflation pressures ranging from 6 to 12 atm. An ultra-high-pressure balloon was used as per our department protocol for HD interventions. At the time of treatment of the central venous lesions, no patients underwent any other intervention in the access circuit. The indications for CS placement were suboptimal PTA, which was defined as [ 30% narrowing with persistent filling of collateral vessels, despite effacement of the central venous occlusion with the ultra-high-pressure balloon. Throughout the study period, all consecutive patients received a CS when there was a central venous occlusion with suboptimal PTA, despite complete effacement of the inciting lesion. Before CS placement, patients underwent extensive power-injector central venous system venography at a rate of 5 ml/s for a total volume of 12 ml to determine the length of the occlusion, insertion sites of the innominate veins, internal jugular veins, and origin of the superior vena cava. PTFE-encapsulated nitinol CSs (Fluency Plus; Bard) ranging in diameter from 10 to 12 mm and in length from 40 to 80 mm were used. The use of CSs in this study were obtained as an off-label application through a special-access government medical programmer. Ten-millimeter-diameter stents were used for the first 6 cases as a result of a lack of availability of larger diameters from the manufacturer. Subsequently, 12-mm-diameter stents became available and were used for the remainder of the study period because larger-diameter stents were not available. The length of the stent was determined on the basis of the length of the lesion, as measured by use of a calibrated measuring pigtail catheter (Cook). Before insertion of the CS, 10F, 5-cm Terumo sheaths (Terumo) were placed at the site of the upper arm access (n = 12) and in the right common femoral vein access (n = 2) to allow passage of the stent delivery system. CSs were deployed with careful attention to the insertion of the

4 952 S. Kundu et al.: Stent Grafts for Central Venous Occlusions contralateral innominate vein and confluence of both innominate veins into the superior vena cava. It was intended to not cover the insertion of the contralateral innominate vein into the superior vena cava with the CS. If the occlusion extended to the insertion of the innominate vein into the superior vena cava, the CS was deployed with the intention of the edge of the CS extending 1 to 2 mm into the superior vena cava. Ipsilateral internal jugular veins or collateral veins were covered with the CS if necessary. Postdilatation of the CSs was performed with an ultra-high-pressure balloon of equal diameter to the CS (Conquest; Bard). Chest x-rays performed at 6 and 9 months after intervention were reviewed to assess for CS migration or fracture. As part of our institutional access surveillance program, all patients underwent access surveillance via ultrasound flow measurements along with 2D grayscale imaging and pulsed wave Doppler spectral analysis from the anastomosis to the ipsilateral axillary veins at 3-month intervals. All patients underwent close clinical follow-up composed of the following: examination by a dialysis coordinator (nurse practitioner) on a weekly basis with a focused history and physical examination for the presence of central venous symptom in addition to routine access follow-up. In addition, six surveillance digital subtraction contrast central venograms were performed by micropuncture access (Cook) into the outflow vein 3 cm beyond the anastomosis to confirm the clinical and Doppler ultrasound findings at 6 months after CS placement. Multiple projections with high magnification were performed of the treated central venous lesions to assess for any restenosis. Data Analysis Data were collected on the distribution of central venous disease, technical success rate, complication rate, and short- and intermediate-term patency. The central venous system was anatomically divided into four segments for the purpose of this study, including bilateral innominate veins and bilateral subclavian veins. Results and outcome definitions were based on Society of Interventional Radiology reporting standards [16]. Technical success was defined as a successful procedure with no complication or significant residual stenosis in the treated central veins. These patients continued to have functional dialysis access, resolution of access dysfunction (if present), and resolution of clinical symptoms as observed by decreased extremity and/or face swelling. Technical failure was defined as an inability to cross the target vein lesion and place a CS or the presence of [30% residual stenosis in the treated vein relative to the adjacent normal nondiseased vein after the procedure. A complication was defined as any event potentially related to the access intervention that occurred during the time of the study and required alteration in the normal pre-, intra-, or postprocedural care [17]. All patients continued their routine medications and were not placed on an anticoagulant, aspirin, or clopidogrel after CS placement as per institutional protocol. The primary patency rate of the central veins was defined as the postintervention interval until the next access thrombosis or a repeat intervention is required. Primary patency ended with treatment of a lesion anywhere within the access circuit from the arterial inflow to the superior vena cava right atrial junction [16]. The postintervention-assisted primary patency rate of the central veins was defined as the interval after intervention until access thrombosis or a surgical intervention that excludes the treated lesion from the access circuit took place. Endpoints to a functional access included (1) placement of a new access, (2) abandonment of the access site, (3) ligation of the access site, and/or (4) placement of a dialysis catheter for access. Follow-up studies were only intended to be performed as indicated by clinical recurrence of symptoms or access dysfunction. Statistical Analysis Measured values are reported as percentages. All digital subtraction images were analyzed from digital subtraction angiography images on the picture archiving and communication system with electronic calipers to perform vessel diameter and lesion length measurements. Results There were a total of 24 patients at our institution treated for benign HD-related central venous occlusion between 2004 and Of these 24 patients, 10 patients (42%) had successful angioplasty, and 14 patients required placement of a CS for failed angioplasty with [30% residual stenosis. Sixteen consecutive straight stent grafts were implanted in 14 patients. Average treated lesion length was 5.0 cm (range, cm). All 14 patients had complete central venous occlusion (100% stenosis). All 14 patients had chronic complete central vein occlusions with no flow through the lesions. Because the occlusions were chronic, no thrombolysis was performed. The central venous occlusions were located as follows: right subclavian and brachiocephalic vein 21% (3 of 14), right brachiocephalic vein 36% (5 of 14), left brachiocephalic vein 36% (5 of 14), and bilateral brachiocephalic vein 7% (1 of 14). Ten-millimeter-diameter (38%, 6 of 16) or 12-mm-diameter (62%; 10 of 16) PTFE stent grafts were placed in the study patients. Ten-millimeter CSs were initially used for the first 6 patients because this was the largest diameter available from the manufacturer. One year

5 S. Kundu et al.: Stent Grafts for Central Venous Occlusions 953 into the study period, larger-diameter 12-mm stents became available and were used for the remainder of the study period. The average stent length was 6.1 cm (range, 4 8 cm). All central venous occlusions were cannulated and crossed with a guide wire as described in Materials and Methods. Initial PTA was attempted and unsuccessful in all 14 patients with a residual stenosis of [50%. There was complete effacement of the central vein occlusions with dilation of the angioplasty balloons in all patients. However, there was significant elastic recoil with ongoing residual stenosis of [50% in all 14 patients. One patient required the placement of two CSs as a result of the length of the lesion involving the right subclavian and innominate vein. One patient required the placement of two CSs as a result of the presence of bilateral innominate vein occlusion. Technical success rate was 100%. In all patients, the stent delivery system (10F diameter) was successfully delivered to the site of the central venous occlusion and deployed. Post-CS placement balloon dilatation was successful in all patients with \30% residual stenosis at the site of the treated central vein lesion. No CSs extended [3 mm into the superior vena cava. The jugular vein insertion was covered by the CS in 6 patients. At the 3-, 6-, and 9-month clinical follow-up examinations, no recurrence of clinical symptoms, as evidenced by face or extremity swelling or evidence of access dysfunction, were observed in any of the 14 patients. The ultrasound evaluation with pulsed Doppler spectral tracing revealed normal polyphasic atrial waveforms in the axillary veins distal to the treated sites in all 14 patients. A polyphasic atrial waveform in the axillary vein has a 95% negative predictive value for ruling out hemodynamically significant ([80% stenosis) of the central veins. Chest radiographs performed at 6- and 9-month follow-up demonstrated no evidence of CS migration or fracture. The six central venograms demonstrated no evidence of any recurrent or in-stent stenosis. Figures 1, 2, 3 demonstrate a representative case, with a short segment central venous occlusion with failed angioplasty, followed by CS placement. There were no procedure-related or periprocedural deaths. There were no minor or major complications, and no alteration in postprocedure care protocols. The primary patency of the treated central venous lesion and access circuit at 3, 6, and 9 months was 100% (10 of 10). There was no recurrence of clinical symptoms or access dysfunction during the follow-up period. Discussion It is rare for central venous disease to occur in HD patients who do not have a history of previous central venous catheterization. Central venous catheter placements, Fig. 1 Short segment occlusion in left brachicephalic vein Fig. 2 Failed angioplasty, with [30% residual stenosis and flow within collateral vessels multiple placements of central venous catheters, and increased duration of catheter dwell times have been associated with a greater risk of central venous disease [3, 7, 18]. The location of the central venous catheter is also an important causative factor for central venous disease. Central venous catheters placed by a subclavian access have a particularly high risk, with a 42 50% incidence of central venous disease compared to the 10% rate observed with catheters placed via an internal jugular vein access [4 7]. There is also an increased predilection for central

6 954 S. Kundu et al.: Stent Grafts for Central Venous Occlusions Fig. 3 Venogram after placement of a 12-mm-diameter, 60-mm-long CS, with no residual stenosis venous disease to occur with left-sided access for catheter placement. This may be related to the more tortuous course catheters have to traverse from a left-sided access [6, 19 21]. Given the high incidence of central venous disease with HD catheters, the large caliber of these catheters may be a causative factor in central venous disease. Peripherally inserted central venous catheters, central venous port catheters, and pacemaker and defibrillator wires are also becoming an increasingly important risk factor for central venous disease. Most patients with central venous disease resulting from these devices are usually asymptomatic and present clinically after a hemodynamic challenge, such as placement of an ipsilateral atrioventricular access line [8, 22 26]. Endovascular intervention is the mainstay treatment in HD patients with central venous disease. The treatment options include PTA, placement of BMSs, and more recently, placement of CSs. The literature to date on the various treatment options has encompassed small retrospective studies with variable patency rates. This has led to a variety of endovascular treatment algorithms that are based on user preference rather than an evidence-based treatment approach. Surgery remains a treatment option of last resort, for patient s refractory to endovascular treatment options. The K/DOQI guidelines recommend PTA, with or without stent placement, as the preferred treatment approach for central venous disease [27]. PTA for central venous disease was first reported by Glanz et al. in 1984, with a 100% technical success rate [2]. PTA has demonstrated variable technical success rates ranging 70 90% [2, 8, 11 14]. A PTA study by Kovalik et al. in 1994 made some interesting observations, including a technical failure rate of 7%, with[50% improvement (nonelastic lesions) in 70% of patients with central venous disease, and\50% improvement (elastic lesions) in 23% of patients with central venous disease. The study concluded that there were two types of central venous lesions: nonelastic lesions, which responded well to PTA, and elastic lesions, which were unresponsive or poorly responsive to PTA. It was suggested that the histology of the two types of lesions was different, as based on intravascular ultrasound observations [11]. These results are not in keeping with our study: we noted complete effacement of the central venous occlusions during balloon angioplasty with [ 50% residual stenosis after PTA in all 14 patients. This suggests a higher degree of elastic recoil in our study than that of Kovalik et al. [11]. This is because we only placed CSs in patients in whom PTA failed. Patients who had no residual stenosis after PTA did not receive further intervention and were not included as part of this study. The PTA patency results for central venous disease demonstrate a wide range of variability. There is a 6-month primary patency range of 23 63% and a cumulative patency range of %. There is a 12-month primary patency range of 12 50% and a cumulative patency range of % [2, 8, 11 14]. One of the largest studies to date on PTA for central venous disease, by Bakken et al. in 2007, comprised 47 patients. This study demonstrated a technical success rate of 77%. There was a primary patency rate of 58% at 3 months, 45% at 6 months, and 29% at 12 months. There was a cumulative patency rate of 76% at 3 months, 62% at 6 months, and 53% at 12 months [15] (Table 1). In summarizing PTA, technical failures rates in the range of 10 30% are to be expected when treating central venous disease. There is clearly a subgroup of central venous disease patients with elastic lesions who are unresponsive to PTA. It is also evident that repeated interventions are required with PTA for central venous disease to maintain patency over the long term. There is a large body of literature to date on the placement of BMSs for central venous disease. BMSs provide mechanical support to a site of stenosis that is resistant or unresponsive to PTA and are potentially useful in central venous disease in the setting of kinked stenoses, collapsing or elastic stenoses after PTA, sealing dissections or circumscribed perforations after PTA, and establishing and maintaining patency of chronic central vein occlusions; and after PTA of highly resistant stenoses. However, there are significant limitations to BMSs. After deployment, BMSs may migrate, shorten, or fracture on a subacute or delayed basis. BMS placement may also preclude future endovascular procedures or surgical revision. Moreover, it is clearly evident that all BMSs incite intimal hyperplasia, leading to recurrent stenoses and multiple repeat interventions to

7 S. Kundu et al.: Stent Grafts for Central Venous Occlusions 955 Table 1 Primary patency rates for central venous intervention Study Procedure 3 months (%) 6 months (%) 9 months (%) 12 months (%) Bakken et al. [15] PTA Haage et al. [35] BMS Vogel et al. [31] BMS Bakken et al. [15] BMS Quinn et al. [47] CS This study CS PTA Percutaneous transluminal angioplasty, BMS Bare metal stent, CS Covered stent maintain patency. The Society of Interventional Radiology Quality Improvement Guidelines recommends that BMSs be reserved for central vein lesions in which PTA has failed or that recur within 3 months after an initially successful PTA, or rupture after PTA [28]. Similarly, the consensus guidelines of the National Kidney Foundation Dialysis Outcomes Quality Initiative recommend that the use of BMSs be reserved for surgically inaccessible stenoses in which PTA fails [11, 29, 30]. The results for BMSs in the setting of central venous disease demonstrate a very high technical success rate, in the range of 100%, with variable short- and long-term patencies. There is a 3-month primary patency range of % and a cumulative patency range of %. There is a 6-month primary patency range of 42 89% and a cumulative patency range of %. There is a 12-month primary patency range of 14 73% and a cumulative patency range of 31 91% [11 15, 31 38]. One of the largest retrospective studies to date on the treatment of central venous disease with BMSs, by Haage et al. using Wallstents published in 1999 with 50 patients, demonstrated a 3-month primary patency rate of 92% and a 6- and 12-month primary patency rate of 84 and 56%, respectively. There was a cumulative patency rate at 6 and 12 months of 97% [35] (Table 1). Unfortunately, these results have not been replicated elsewhere in the literature. A more recent retrospective study on nitinol BMSs for cardiovascular disease by Vogel et al. in 2004 with 16 patients demonstrated 3-, 6-, and 12-month primary patency rates of 81, 74, and 67%, respectively. Cumulative patencies were not reported in this study [31]. A retrospective study by Bakken et al. published in 2007 comparing PTA and BMSs for central venous disease demonstrated 3-, 6-, and 12-month primary patencies with PTA of 58, 25, and 29% in comparison with 3-, 6-, and 12-month primary patencies with BMSs of 65, 54, and 45%, respectively. There were 3-, 6-, and 12-month cumulative patencies with PTA of 76, 62, and 53% in comparison with 3-, 6-, and 12-month cumulative patencies with BMSs of 72, 55, and 46%, respectively. There was no significant difference in patency results between the PTA and the BMS group [15]. Another recent study published by Ozyer et al. in 2009 demonstrated a higher primary patency in the angioplasty group of 24.5 ± 1.7 months compared with the BMS group of 13.4 ± 2.0 months with equivalent assisted primary patency in both groups [39]. Given the additional costs of BMSs and aggressive intimal hyperplasia, central venous sent placement should only be considered selectively for refractory stenoses or occlusions. Given the variable patency results of the alternative treatments with the requirement of multiple repeat interventions, CSs also known as peripheral endografts have been proposed as a treatment option for central venous disease. The potential advantages of CSs include providing a relatively inert and stable intravascular matrix for endothelialization while providing the mechanical advantages of a BMS. This could potentially reduce the intimal hyperplastic response, causing restenosis after PTA or BMS placement. The disadvantages of CSs include covering collateral veins, which may provide critical outflow in recurrent stenoses or occlusions; inadvertent covering of adjacent veins such as the internal jugular or contralateral innominate vein; and high cost, which may be two to three times the cost of a BMS. There is little literature on CS usage in the HD access circuit. Most of the literature to date has been on the treatment of graft or outflow vein aneurysms and refractory venous outflow stenoses [40 47]. CSs for central venous disease have only been mentioned in two publications to date. Sapoval et al. in 1996 mentioned the use of a nitinol plus Dacron-covered stent (Craig Endopro, Mintec, LaCiotat, France) for an in-stent restenosis of a Wallstent, with asymptomatic recurrent restenosis after 6 months [46]. A study by Quinn et al. in 2003 placed 6 CSs for central venous disease and 11 CSs for peripheral venous outflow stenoses. There was a combined primary patency at 2, 6, and 12 months of 40, 32, and 32%; and secondary patency at 2, 6, and 12 months of 70, 55, and 39%, respectively. They used a Palmaz stent (P308, Johnson and Johnson, Warren, NJ) with an eptfe graft material manually sewn on [47] (Table 1). Our study demonstrates interesting short- and intermediate-term patency results for treating central vein occlusions using

8 956 S. Kundu et al.: Stent Grafts for Central Venous Occlusions newer PTFE-covered stent technology (Fluency Plus; Bard). Given the significant cost of CSs, our results will need to be assessed with larger patient series and longer term follow-up, in randomized, controlled trials if possible. Our study and results have a number of limitations including the nonrandomized retrospective methodology, small number of patients, the use of a single center, absence of long-term follow-up, performance of all procedures by one operator, the use of smaller diameter (10 mm) CSs for the initial 6 patients and use of 12-mmdiameter stents in the remainder of the patients in this study, uncovered nitinol stent (1 mm) at the edges of the CSs used in this study, absence of follow-up central thoracic venograms in all patients, and the lack of comparative data with alternative technologies. In reference to the uncovered nitinol stent at the edges of the CS in this study, these uncovered portions of stents may act to incite intimal hyperplasia and lead to edge restenosis over a period of time. In reference to the absence of follow-up central thoracic venograms, none of the referenced central venous disease studies have had follow-up central thoracic venograms. All follow-ups have been on a clinical basis, or to time of next intervention. We have also tried to address this limitation through close clinical surveillance, duplex ultrasound, and six random venograms to confirm our clinical and duplex ultrasound examinations as outlined in Materials and Methods. All of our study patients were also part of our institutional access surveillance program and underwent access surveillance via duplex ultrasound with ultrasound access flow measurements along with 2D grayscale imaging and pulsed-wave Doppler spectral analysis from the anastomosis to the ipsilateral axillary veins at 3-month intervals. The pulsed Doppler spectral tracing revealed normal polyphasic atrial waveforms in the axillary veins distal to the treated sites in all 14 patients. A polyphasic atrial waveform in the axillary vein has a 95% negative predictive value for ruling out hemodynamically significant ([80%) stenosis of the central veins [48 50]. We believe these limitations can be addressed in the future through larger multicenter randomized, controlled trials comparing all current treatment alternatives with longerterm follow-up. We conclude that CSs for central venous disease are a potentially interesting treatment alternative worthy of further investigation and research. Venous territory and capital are limited in HD patients, and any treatment alternative that prolongs the functional patency of the access circuit should be critically evaluated for scientific and clinical value with prospective studies. Conflict of interest of interest. The authors declare that they have no conflict References 1. Lumsden AB, MacDonald MJ, Isiklar H et al (1997) Central venous stenosis in the hemodialysis patient: incidence and efficacy of endovascular treatment. Cardiovasc Surg 5: Glanz S, Gordon DH, Lipkowitz GS et al (1988) Axillary and subclavian stenosis: percutaneous angioplasty. Radiology 168: Agarwal AK, Patel BM, Farhan NJ (2004) Central venous stenosis in hemodialysis patients is a common complication of ipsilateral central vein catheterization. J Am Soc Nephrol 15: 368A 369A 4. Barrett N, Spencer S, McIvor J et al (1988) Subclavian stenosis: a major complication of subclavian dialysis catheters. Nephrol Dial Transpl 3: Cimochowski GE, Worley E, Rutherford WE et al (1990) Superiority of the internal jugular vein over the subclavian access for temporary dialysis. Nephron 54: Schillinger F, Schillinger D, Montagnac R et al (1991) Post catheterization venous stenosis in hemodialysis: comparative angiographic study of 50 subclavian and 50 internal jugular accesses. Nephrol Dial Transpl 6: Vanherweghem JL, Yasine T, Goldman M et al (1986) Subclavian vein thrombosis: a frequent complication of subclavian cannulation for hemodialysis. Clin Nephrol 26: Beathard GA (1992) Percutaneous transvenous angioplasty in the treatment of vascular access stenosis. Kidney Int 42: Glanz S, Gordon D, Butt KMH et al (1984) Dialysis access fistulas: treatment of stenoses by transluminal angioplasty. Radiology 152: Trerotola SO, McLean GK, Burke DR, Meranze SG (1986) Treatment of subclavian venous stenoses by percutaneous transluminal angioplasty. J Vasc Interv Radiol 1: Kovalik EC, Newman GE, Suhocki P et al (1994) Correction of central venous stenoses: use of angioplasty and vascular Wallstents. Kidney Int 45: Quinn SF, Schuman ES, Demlow TA et al (1995) Percutaneous transluminal angioplasty versus endovascular stent placement in the treatment of venous stenoses in patients undergoing hemodialysis: intermediate results. J Vasc Interv Radiol 5: Dammers R, de Haan MW, Planken NR et al (2003) Central vein obstruction in hemodialysis patients: results of radiological and surgical intervention. Eur J Vasc Endovasc Surg 26: Surowiec SM, Fegley AJ, Tanski WJ et al (2004) Endovascular management of central venous stenoses in the hemodialysis patient: results of percutaneous therapy. Vasc Endovasc Surg 38: Bakken AM, Protack CD, Saad WE et al (2007) Long-term outcomes of primary angioplasty and primary stenting of central venous stenosis in hemodialysis patients. J Vasc Surg 45: Gray RJ, Sacks D, Martin LG, Trerotola SO (1995) Reporting standards for percutaneous interventions in dialysis access. J Vasc Interv Radiol 6: Sacks D, Marinelli DL, Martin LG et al (1997) Reporting standards for clinical evaluation of new peripheral arterial revascularization devices. J Vasc Interv Radiol 8: MacRae JM, Ahmed A, Johnson N et al (2005) Central vein stenosis: a common problem in patients on hemodialysis. ASAIO 51: Schon D, Whittman D (2003) Managing the complications of long-term tunneled dialysis catheters. Semin Dial 16: Moss AH, Vasilaksi C, Holley HL et al (1990) Use of a silicon dual-lumen catheter with a Dacron cuff as a long-term vascular access for hemodialysis patients. Am J Kidney Dis 16:

9 S. Kundu et al.: Stent Grafts for Central Venous Occlusions Salgado OJ, Urdaneta B, Comenares B et al (2004) Right versus left internal jugular vein catheterization for hemodialysis: complications and impact on ipsilateral access creation. Artif Organs 28: Gonsalves CF, Eschelman DJ, Sullivan KL et al (2003) Incidence of central venous stenosis and occlusion following upper extremity PICC and port placement. Cardiovasc Intervent Radiol 26: Trerotola SO, Kuhn-Fulton J, Johnson MS et al (2000) Tunneled infusion catheters: increased incidence of symptomatic venous thrombosis in subclavian versus internal jugular venous access. Radiology 217: Korzets A, Chagnac A, Ori Y et al (1991) Subclavian vein stenosis, permanent cardiac pacemakers and the hemodialysed patient. Nephron 58: Chuang C, Tarng D, Yang W et al (2001) An occult cause of arteriovenous access failure: central vein stenosis from permanent pacemaker wire. Am J Nephrol 21: Sticherling C, Chough SP, Baker RL et al (2001) Prevalence of central venous occlusion in patients with chronic defibrillator leads. Am Heart J 141: NKF-K/DOQUI (2001) NKF-K/DOQUI clinical practice guidelines for vascular access. Am J Kidney Dis 37:137s 181s 28. Aruny JE, Lewis CA, Cardella JF et al (1994) 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 5: NKF-DOQI (1997) NKF-DOQI clinical practice guidelines for vascular access. National Kidney Foundation Dialysis Outcomes Quality Initiative. Am J Kidney Dis 30(suppl 3):S150 S Clark TWI (2004) Nitinol stents in hemodialysis access. J Vasc Interv Radiol 15: Vogel PM, Parise CP (2004) SMART stent for salvage of hemodialysis access grafts. J Vasc Interv Radiol 15: Aytekin C, Boyvat F, Yagmurdur MC et al (2004) Endovascular stent placement in the treatment of upper extremity central venous obstruction in hemodialysis patients. Eur J Radiol 49: Chen CY, Liang HL, Pan HB et al (2003) Metallic stenting for treatment of central venous obstruction in hemodialysis patients. J Chin Med Assoc 66: Oderich GS, Treiman GS, Schneider P, Bhirang K (2000) Stent placement for treatment of central and peripheral venous obstruction: a long-term multi-institutional experience. J Vasc Surg 32: Haage P, Vorwerk D, Piroth W et al (1999) Treatment of hemodialysis-related central venous stenosis or occlusion: results of primary Wallstent placement and follow-up in 50 patients. Radiology 212: Vesely TM, Hovsepian DM, Pilgram TK et al (1997) Upper extremity central venous obstruction in hemodialysis patients: treatment with Wallstents. Radiology 204: Gray RJ, Horton KM, Dolmatch BL et al (1995) Use of Wallstents for hemodialysis access-related venous stenoses and occlusions untreatable with balloon angioplasty. Radiology 195: Vorwerk D, Guenther RW, Mann H et al (1995) Venous stenosis and occlusion in hemodialysis shunts: follow-up results of stent placement in 65 patients. Radiology 95: Ozyer U, Harman A, Yildrim E et al (2009) 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 193: Vesely TM (2005) Use of stent grafts to repair hemodialysis graft-related pseudoanuerysms. J Vasc Interv Radiol 16: Naoum JJ, Irwin C, Hunter GC (2006) The use of covered nitinol stents to salvage dialysis grafts after multiple failures. Vasc Endovasc Surg 40: Hausegger KA, Tiessenhausen K, Klimpfinger M et al (1998) Aneurysms of hemodialysis access grafts: treatment with covered stents: a report of three cases. Cardiovasc Intervent Radiol 21: Silas AM, Bettman MA (2003) Utility of covered stents for revision of aging failing synthetic hemodialysis grafts: a report of three cases. Cardiovasc Intervent Radiol 26: Shemesh D, Goldin I, Zaghal I et al (2008) Angioplasty with stent graft versus bare stent for recurrent cephalic arch stenosis in autogenous arteriovenous access for hemodialysis: a prospective randomized clinical trial. J Vasc Surg 48: Haskal Z, Trerotola S, Dolmatch BD (2010) Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med 362: Sapoval MR, Turmel-Rodrigues LA, Raynaud AC et al (1996) Cragg covered stents in hemodialysis access: initial and midterm results. J Vasc Interv Radiol 7: Quinn SF, Kim J, Sheley RC (2003) Transluminally placed endovascular grafts for venous lesions in patients on hemodialysis. Cardiovasc Intervent Radiol 26: Rose SC, Kinney TB, Bundens WP et al (1998) Importance of Doppler analysis of transmitted atrial waveforms prior to placement of central venous access catheters. J Vasc Interv Radiol 9: Passman MA, Criado E, Farber MA et al (1998) Efficacy of color flow duplex imaging for proximal upper extremity venous outflow obstruction in hemodialysis patients. J Vasc Surg 28: Weber TM, Lockhart ME, Robin ML (2007) Upper extremity venous Doppler ultrasound. Radiol Clin N Am 45: