- Email: [email protected]
Management of Central Vein Stenoses and Occlusions: The Critical Importance of the Costoclavicular Junction
Management of Central Vein Stenoses and Occlusions: The Critical Importance of the Costoclavicular Junction Karl A. Illig, MD The failure of an autogenous or prosthetic arteriovenous hemodialysis access is usually related to the failure of the venous outflow resulting from a stenosis somewhere in the venous system, commonly at the venous anastomosis for a prosthetic access or within the central veins. The National Kidney Foundation’s Kidney Disease Outcomes Quality Initiative guidelines state that percutaneous transluminal venoplasty with or without stenting is the preferred initial treatment for a central venous stenosis, but the results of these therapies have been have relatively disappointing when analyzed as a whole. Although endoluminal intervention works well (and is, indeed, the primary option) for treating areas of stenosis surrounded by soft tissue, we believe stenoses occurring at the costoclavicular junction are caused by extrinsic bony compression and, therefore, should be considered dialysis-associated venous thoracic outlet syndrome. The treatment of venous thoracic outlet syndrome, based on decades of experience, generally requires bony decompression for long-term patency. In the last 2 years, we have treated 12 patients with dialysisassociated venous thoracic outlet syndrome with surgical decompression of the thoracic outlet. Functional patency was achieved in 75% of patients at a mean follow-up of 8 months. We would contend that not all central vein stenoses are equivalent and that an individualized approach is most appropriate based on the extent and anatomic location of the lesion. Semin Vasc Surg 24:113-118 © 2011 Elsevier Inc. All rights reserved.
T
HERE ARE APPROXIMATELY 370,000 patients in the United States on chronic hemodialysis, with 77% of them dialyzing through an autogenous (AVF) or prosthetic (arteriovenous graft [AVG]) arteriovenous access.1 Unfortunately, these “permanent” accesses have a relatively short lifespan with the half-life for AVGs being approximately 1 year with that for an AVF only marginally better. When an AVF or AVG fails, it is most often due to the development of a stenosis at the venous anastomosis (ie, AVG) or in the venous outflow. In several studies, the prevalence of central vein stenosis was found to be between 15% and 20% for all patients undergoing dialysis, 30% in those with a history of prior catheter placement, and as high as 50% in those considered “symptomatic.”2,3 While central venous stenosis is most often attributed to subclavian catheter use, many paDivision of Vascular Surgery, University of Rochester Medical Center, Rochester, NY. Address reprint requests to Karl A. Illig, MD, Division of Vascular Surgery, University of Rochester Medical Center, Rochester, 601 Elmwood Avenue, Box 652, Rochester, NY 14642. E-mail: Karl_Illig@ urmc.rochester.edu
0895-7967/$-see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1053/j.semvascsurg.2011.05.008
tients with jugular catheters or no history of catheterization at all also develop these lesions.2 In our experience, the incidence of this problem has not decreased significantly, despite the more recent emphasis on a jugular-only approach for dialysis catheters (ie, avoidance of subclavian catheters). The pressure in the “venous” portion of an AVF or AVG (ie, region of access used for cannulation and/or venous outflow) is relatively low, typically in the range of 50 mm Hg at the point of cannulation and normal central venous pressure at the level of the axillary vein.4 This dramatic pressure drop occurs immediately distal to the arterial anastomosis, and is due to the fact that a normal vein is large, distensible, and very compliant (ie, capacitance vessel). Even though flow is extremely high, the vein can accommodate the slightly increased pressure without problems. If there is a stenosis or occlusion at some point in the circuit distal to the arterial anastomosis, however, the intraluminal pressure increases dramatically. This phenomenon accounts for the observed change from a thrill to a pulse when the outflow of an AVG or AVF is compressed. Unfortunately, some dialysis and emergency medicine personnel are not aware of this phenomenon 113
K.A. Illig
114 and inadvertently exacerbate post-cannulation bleeding at the access cannulation site by occluding the access central to the site, thereby increasing the intraluminal pressure. Chronically increased pressure within an AVG or AVG as a result of a venous outflow stenosis can result in several adverse events. Most visibly, the arm can swell. The outflow veins from the access can tolerate some increased pressure due to their capacitance, but these adaptations are overwhelmed at arterial pressures. If the intraluminal pressure is elevated high enough, swelling will ensue. In a manner analogous to chronic venous insufficiency in the lower extremities, this initially consists of acellular fluid alone, and produces reversible, pitting edema. However, as this process becomes more chronic, the fluid can include cells, cellular debris, proteins, and other macromolecules. The increased pressure will also produce a greater tendency to cause bleeding (both during and after cannulation) and increased recirculation as more and more “clean” blood is redirected back into the arterial needle to pass through the dialysis machine leading to inefficient dialysis. It is overly simplistic to lump all vascular access-associated venous stenoses or occlusions together and to treat them with a single interventional technique. Usually, stenoses in the arm or chest can be effectively treated by an endovascular means, while occlusions require an open surgical approach. We believe that the optimal treatment for stenoses that occur at the costoclavicular junction, however, must include surgical decompression of the subclavian vein.
Diagnosis and Evaluation Patients with venous stenoses will usually present with one of three problems: venous hypertension, prolonged bleeding, or inefficient dialysis. Venous hypertension should be fairly obvious and patients can present with swelling ranging from soft and pitting to brawny and indurated, depending on the chronicity of the problem. Although we have never seen a true venous ulceration, some patients have presented with increased pigmentation and lipodermatosclerosis. The swelling (and skin changes) can involve the whole arm or be limited to the forearm and the area overlying the fistula itself. The swelling can lead to patient discomfort and can complicate cannulation during dialysis. The increased bleeding can occur during dialysis or after de-cannulation and results from an increased pressure gradient across the puncture site (ie, lumen to outside world). This is often exacerbated by illadvised pressure held diffusely along the vein (rather than gentle point pressure on the site itself), which will increase the intraluminal pressure and further increase bleeding. The inefficient dialysis results from the recirculation of “clean” blood back into the dialysis machine from reflux of blood back into the arterial needle as a result of the elevated venous pressures. These elevated pressures can trigger alarms on the machine, but it is not the elevated pressures per se that are the problem, rather the inefficient dialysis. In the most extreme situation, an endless loop can occur in which the same blood is filtered through the machine with little net effect. All patients with symptoms related to presumed venous
stenoses or occlusions should undergo a fistulagram to identify the lesion and facilitate definitive treatment, most of which can be performed at the same setting. Although there is usually no sense of emergency, the fistulagram should be performed in a timely fashion because the access-related venous hypertension rarely improves without intervention. Furthermore, early intervention may be associated with improved results because it is easier to treat a stenosis than an occlusion. Although ultrasound may be helpful to image the access in the extremity, it cannot image the full extent of the venous outflow tract to the atrium, given the bony thoracic cavity.
Definitive Management: Anatomic-Based Approach The findings on the fistulagram or venogram will identify the anatomic problem and suggest the definitive treatment. There are three distinct anatomic locations where problems are commonly encountered and the vein can be stenotic or occluded at each of these sites. The modality chosen to treat an access-related central venous stenosis or occlusion is based on the location and nature of the venous lesion.
Central Veins Proximal to the Costoclavicular Junction This anatomic segment is comprised of the body of the access (ie, AVF), the peripheral outflow vein(s), and the brachial/ axillary veins up to the level of the costoclavicular junction (Fig 1). Although the access itself and the peripheral veins in the upper extremity are not usually considered along with the discussion of central venous lesions, they will be addressed for completeness. Venous stenoses, defined by the ability to cross the lesion with a wire, in these locations are almost always amenable to conventional endovascular techniques. Percutaneous access should be obtained peripheral to the lesion being treated (ie,
Figure 1 Fistulagram showing smooth stenosis in the axillary vein in the axilla (arrow). This responded well to balloon angioplasty.
Central vein stenoses and occlusions proximal on the access circuit). Lesions close to the arterial anastomosis can be approached retrograde through the access or through the inflow artery, although it may be difficult to establish a stable platform from the groin because of the anatomic distance. Conventional balloon angioplasty is usually the first step, and a reasonable amount of overdilation is safe in this low-pressure, superficial venous system. Balloon dilation of a stenotic lesion can be somewhat painful and necessitate additional sedation and/or pain medication. If there is significant recoil after angioplasty alone or the stenotic “waist” cannot be eliminated, a high-pressure or cutting balloon can be used as the next step in the treatment algorithm. Technical success can be assessed by the elimination of the stenosis, the disappearance of collateral channels and presence of a thrill in the access (or elimination of the pulsatile character). Intravascular ultrasound can be helpful in this setting to define the residual lumenal diameter. Consideration should be given to stenting the lesion if the stenosis persists. Stents, even in the segment of the vein used for cannulation, seem to be well-tolerated. Using these conventional techniques, a relatively good result can be expected although restenosis is common and an aggressive surveillance protocol is likely justified. Occluded outflow veins in the region peripheral to the costoclavicular junction require individualized treatment. An attempt can be made to cross the occlusion with a wire (and, if successful, treatment rendered as mentioned here), but in our experience these occlusions are usually chronic and cannot be crossed with a wire. In this situation, open surgical repair is usually required, with the goal being to establish adequate venous outflow. The surgical options are usually dictated by the extent and location of the occlusion and the status of the superficial and central veins. Potential solutions including mobilizing the distal segment of the AVF and reimplanting it on another vein with patent outflow (eg, distal cephalic vein-axillary vein), translocating the basilic or brachial vein (eg, distal cephalic vein-basilic vein), or using an interposition graft (autogenous or prosthetic) to bypass the lesion.
Central Veins Distal to the Costoclavicular Junction This zone is comprised of the innominate veins and the superior vena cava (SVC). Patients with stenoses at this location may exhibit signs and symptoms of SVC syndrome, which can be severe enough to threaten their airway. Innominate vein and SVC stenoses can usually be treated with conventional endovascular techniques, presuming the lesion can be crossed with a wire (Fig 2). Percutaneous access, including an introducer sheath, can be obtained through the AVF or AVG. It is important to “anchor” the distal end of the wire in the inferior vena cava (rather than the SVC or atrium) to assure that the platform is sufficient for successful treatment. We almost always use a self-expanding stent in this location, one that is sufficiently large in diameter to assure full wall apposition. Intravascular ultrasound is valuable for assessing the size of the normal, adjacent vessel.
115
Figure 2 Fistulagram of a patient with smooth stenosis of the superior vena cava (arrow). Note that his dialysis catheter, inserted from the left, does not seem to be involved with this lesion in any way. This responded well to placement of a 14-mm self-expanding stent that was angioplastied with a 12-mm balloon.
Although the balloon should be sized appropriately for the lumenal diameter, over dilation should be avoided because there have been anecdotal reports of fatal SVC rupture in this setting. At times it may be necessary to place a stent across the ipsilateral jugular or contralateral innominate vein. This is usually well-tolerated because the “caged” vein will remain patent via the blood flow through the interstices of the stent. Should subsequent intervention become necessary, the interstices of the stent covering the caged vein can usually be crossed, similar to stents placed across the orifices of the iliac veins for patients with iliocaval stenoses. Not infrequently, patients with an innominate or SVC stenosis will also have a dialysis catheter. If the central stenosis cannot be crossed with a wire (ie, between the wall of the vessel and the catheter), the wire can be passed through the catheter, the catheter removed, and the endovascular interventions performed. A new catheter can be placed at the completion of the procedure as necessary. Occlusions of the innominate vein and SVC are more problematic than stenoses. Aggressive attempts at crossing the occlusion with a wire are justified because the open surgical alternatives for salvaging the access and relieving the venous hypertension are associated with significant morbidity. Several techniques for blindly crossing central vein occlusions have been described, although these should not be attempted without significant endovascular experience.5 One technique involves the use of a retrograde approach to the SVC from the femoral vein. If the lesion cannot be crossed, the options include access ligation (and abandonment) or salvage with surgical revascularization or bypass through a sternotomy or thoracotomy.6,7 If the life expectancy of the patient is limited and alternative options for access placement exist, ligation and abandonment of the access is reasonable.
116
K.A. Illig
Central Veins at the Costoclavicular Junction Access-related central venous stenosis can develop in the segment of the subclavian vein adjacent to the costoclavicular junction, the point where the first rib and the clavicle come together at the sternum. The vein may be tethered and externally compressed by the subclavius muscle and tendon and the costoclavicular ligament at this location (Figs 3 and 4). In nondialysis patients, this anatomic narrowing may cause chronic injury to the vein, resulting in venous thrombosis (ie, venous thoracic outlet syndrome).8,9 It is our impression that the high venous flow and associated turbulence resulting from the arteriovenous access further exacerbates this chronic functional injury leading to stenoses and occlusions. Based on the collective experience from the treatment of venous thoracic outlet syndrome, it is apparent that subclavian vein stenoses adjacent to the costoclavicular junction respond poorly to treatment with simple balloon angioplasty.9 Furthermore, the use of stents in this location may actually be harmful because of the “nutcracker” effect of the clavicle and first rib, which can cause stent fracture and venous occlusion (Fig 5).10 We believe that the accepted treatment paradigms for venous thoracic outlet syndrome can be applied to the treatment of access-related venous stenosis at the costoclavicular junction and that surgical decompression (with concomitant endovascular intervention, if needed) provides the best hope for long-term success.
Figure 3 Basic anatomy of the thoracic outlet. The axillosubclavian vein passes anteriorly, passing by the junction of the first rib and clavicle. This “space” is open superiorly, but the vein is tethered in this location by surrounding tissue. The two bones and the subclavius muscle and tendon chronically and repetitively exert pressure on it. In patients with high flow (ie, with an ipsilateral arteriovenous fistula) this area can quickly become stenotic. (Reprinted from Illig KA, Doyle A: A comprehensive review of Paget-Schroetter syndrome. J Vasc Surg 51:1538-1547, 2010, with permission.9)
Figure 4 Fistulagram of a patient with “dialysis-associated venous thoracic outlet syndrome,” manifested clinically as massive arm swelling and fistula dysfunction. Note the rather long, complex stenosis beginning at the costoclavicular junction with fairly normal vein peripherally (solid arrow) and extensive collateralization that is pathognomonic for this lesion (open arrow).
We treat access-related subclavian vein stenosis at the costoclavicular junction by first performing a transaxillary first rib resection and mobilizing the subclavian vein to the jugular confluence. Often, there is a dense cicatrix surrounding the vein and it is important to perform an extensive venolysis to ensure complete venous decompression. Surprisingly, we have not noticed a difference in bleeding during the surgical decompression procedure in patients with access-related venous thoracic outlet compared to our “conventional” venous thoracic outlet patients. After the incision is closed, the patient is repositioned and a venous balloon angioplasty is performed as described here. Although some believe that surgical decompression of the vein permits the use of stents, we place stents across this vein segment only if absolutely necessary (Fig 6).10,11
Figure 5 A posterioanterior chest radiograph is shown for a patient with a right upper extremity autogenous access and an ipsilateral central vein occlusion. Note the compressed stent at the thoracic outlet (arrow). Interestingly, the patient also has a thoracic endograft.
Central vein stenoses and occlusions
Figure 6 The same patient shown in Figure 5 is shown again after transaxillary first rib excision with external venolysis and balloon angioplasty; note that the collateral channels are gone. The lesion within the innominate vein did not respond to venoplasty alone and required stenting. This vein remained patent and the fistula functional until the patient died of unrelated causes.
Treatment of subclavian venous occlusions at the costoclavicular junction generally requires direct reconstruction of the vein or a jugular transposition. Although the methods described for exposing this vein segment for direct reconstruction generally involve partial resection of the clavicle, a method of exposure that avoids claviculectomy also has been described.12,13 This latter method may be preferable in patients who are wheelchair-dependent because claviculectomy in such patients can be debilitating. These options may be considered to salvage the access and relieve symptoms in young patients who are fit and have a reasonable life expectancy. For patients whose life expectancy is short and alternative access sites exist, access ligation and abandonment may be more appropriate. An alternative approach that should be investigated in this setting includes extraanatomic bypass from the axillary/subclavian vein to the ipsilateral internal jugular or contralateral axillary. Although this is rarely an option given the widespread proliferation of tunneled dialysis catheters and the common finding that the jugular and axillary veins are occluded, it merits consideration. The axillary-jugular vein bypass is relatively straightforward. The axillary and jugular veins can be exposed through an infraclavicular and supraclavicular incision, respectively, and a large prosthetic graft (eg, 10 to 12 mm expanded polytetrafluoroethylene) can be easily tunneled between the incisions deep to the clavicle.
Results The literature reporting the outcomes for balloon angioplasty of access-related central vein stenosis is highly variable, with
117 6-month patency rates ranging from 30% to 60%.2,7,14,15 Unfortunately, the use of stents has not dramatically altered these marginal results. One potential reason for this variability is the fact that most reports combine outcomes for stenoses and occlusions and do not define the anatomic location of the lesion. The inclusion of patients with venous stenosis at the costoclavicular junction in these reports may account for many of these treatment failures. Two authors previously alluded to the notion that subclavian vein stenosis at the costoclavicular junction may respond poorly to endovascular intervention.5,7 However, there are no published series of access-related venous stenoses at the costoclavicular junction treated with concomitant surgical decompression and endovascular intervention. Interestingly, several of the figures contained in these articles documenting the outcomes for access-related central venous lesions actually involve lesions at the costoclavicular junction.2,5,7 Since November 2008, we have treated 12 patients with access-related central venous stenoses or occlusions with open surgical intervention.16 The procedures were performed to salvage an access in patients with severe venous hypertension and access dysfunction. All patients had multiple previous endovascular interventions (mean 2.3 interventions) at the costoclavicular vein segment. Eight patients underwent transaxillary first rib resection with venolysis and subsequent balloon angioplasty and/or stenting of the residual stenosis (Figs 4 and 6), while four patients underwent claviculectomy with direct venous repair or bypass. The overall mean follow-up in our series was 8 months. Salvage of the access occurred in eight cases and relief of the upper extremity swelling occurred in nine cases.
Conclusions Venous outflow stenosis is the most common cause for access failure. Guideline 20 of the Kidney Disease Outcomes Quality Initiative guidelines recommends percutaneous transluminal venoplasty with or without stenting as the preferred initial treatment for these lesions.17 However, not all of the venous outflow lesions are the same and treatment must be individualized by the nature and location of the lesion. Venous lesions in the arm or chest can usually be treated very effectively with conventional endovascular or surgical approaches. However, the lesions at the costoclavicular junction are intrinsically different because of extrinsic compression at thoracic outlet. As in venous thoracic outlet syndrome, the optimal treatment of access-related venous stenosis at this location must include surgical decompression of the vein with first rib resection and venolysis.
References 1. Arteriovenous Fistula First: Fistulat First data. Available at: http:// www.fistulafirst.org/AboutFistulaFirst/FFBIData.aspx. Accessed May 15, 2011 2. Agarwal AK: Central vein stenosis: current concepts. Adv Chronic Kidney Dis 16:360-370, 2009 3. Roy-Chaudhury P, Spergel LM, Besarab A, Asif A, Ravani P: Biology of arteriovenous fistula failure. J Nephrol 20:150-163, 2007
118 4. Illig KA, Surowiec S, Shortell CK, Davies MG, Rhodes JM, Green, RM: Hemodynamics of distal revascularization-interval ligation. Ann Vasc Surg 19:199-207, 2005 5. Kim YC, Won JY, Choi SY, et al: Percutaneous treatment of central venous stenosis in hemodialysis patients: long-term outcomes. Cardiovasc Intervent Radiol 32:271-278, 2009 6. Glass C, Maevsky V, Massey T, Illig KA: Subclavian vein to right atrial appendage bypass without sternotomy to maintain AV access in patients with complete central vein occlusion: a new approach. Ann Vasc Surg 23:465-468, 2009 7. Dosluoglu H, Harris LM: Hemodialysis access: nonthrombotic complications, in Cronenwett JL, Johnston KW (eds): Rutherford’s Vascular Surgery (ed 7). Philadelphia, Saunders Elsevier, 2010, pp 1137-1154, 8. Aziz S, Straehley CJ, Whelan TJ Jr: Effort-related axillosubclavian vein thrombosis: a new theory of pathogenesis and a plea for direct surgical intervention. Am J Surg 152:157-161, 1986 9. Illig KA, Doyle A: A comprehensive review of Paget-Schroetter syndrome. J Vasc Surg 51:1538-1547, 2010 10. Meier GH, Pollak JS, Rosenblatt M, Dickey KW, Gusberg RJ: Initial experience with venous stents in exertional axillary-subclavian vein thrombosis. J Vasc Surg 24:974-981; discussion 981-983, 1996
K.A. Illig 11. Kreienberg PB, Chang BB, Darling RC 3rd, et al: Long-term results in patients treated with thrombolysis, thoracic inlet decompression, and subclavian vein stenting for Paget-Schroetter syndrome. J Vasc Surg 33:S100-S1005, 2001 (suppl 2) 12. Hu SH, Dilley RB: Internal jugular vein turndown for subclavian vein occlusion. Oper Tech Gen Surg 10:149-153, 2008 13. Green RM, Waldman D, Ouriel K, Riggs P, DeWeese JA: Claviculectomy for subclavian venous repair: long-term functional results. J Vasc Surg 32:315-321, 2000 14. Beathard GA: Percutaneous transvenous angioplasty in the treatment of vascular access stenosis. Kidney Int 42:1390-1397, 1992 15. 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 45;776783, 2007 16. Glass C, Dugan M, Gillespie D, Doyle A, Illig KA: Costoclavicular venous decompression in patients with threatened arteriovenous hemodialysis access. Ann Vasc Surg 2011 Apr 20. [Epub ahead of print] 17. NKF-K/DOQI clinical practice guidelines for vascular access, 2000. Am J Kidney Dis 2001;37:S137–S181, 2001 (suppl 1)
T
HERE ARE APPROXIMATELY 370,000 patients in the United States on chronic hemodialysis, with 77% of them dialyzing through an autogenous (AVF) or prosthetic (arteriovenous graft [AVG]) arteriovenous access.1 Unfortunately, these “permanent” accesses have a relatively short lifespan with the half-life for AVGs being approximately 1 year with that for an AVF only marginally better. When an AVF or AVG fails, it is most often due to the development of a stenosis at the venous anastomosis (ie, AVG) or in the venous outflow. In several studies, the prevalence of central vein stenosis was found to be between 15% and 20% for all patients undergoing dialysis, 30% in those with a history of prior catheter placement, and as high as 50% in those considered “symptomatic.”2,3 While central venous stenosis is most often attributed to subclavian catheter use, many paDivision of Vascular Surgery, University of Rochester Medical Center, Rochester, NY. Address reprint requests to Karl A. Illig, MD, Division of Vascular Surgery, University of Rochester Medical Center, Rochester, 601 Elmwood Avenue, Box 652, Rochester, NY 14642. E-mail: Karl_Illig@ urmc.rochester.edu
0895-7967/$-see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1053/j.semvascsurg.2011.05.008
tients with jugular catheters or no history of catheterization at all also develop these lesions.2 In our experience, the incidence of this problem has not decreased significantly, despite the more recent emphasis on a jugular-only approach for dialysis catheters (ie, avoidance of subclavian catheters). The pressure in the “venous” portion of an AVF or AVG (ie, region of access used for cannulation and/or venous outflow) is relatively low, typically in the range of 50 mm Hg at the point of cannulation and normal central venous pressure at the level of the axillary vein.4 This dramatic pressure drop occurs immediately distal to the arterial anastomosis, and is due to the fact that a normal vein is large, distensible, and very compliant (ie, capacitance vessel). Even though flow is extremely high, the vein can accommodate the slightly increased pressure without problems. If there is a stenosis or occlusion at some point in the circuit distal to the arterial anastomosis, however, the intraluminal pressure increases dramatically. This phenomenon accounts for the observed change from a thrill to a pulse when the outflow of an AVG or AVF is compressed. Unfortunately, some dialysis and emergency medicine personnel are not aware of this phenomenon 113
K.A. Illig
114 and inadvertently exacerbate post-cannulation bleeding at the access cannulation site by occluding the access central to the site, thereby increasing the intraluminal pressure. Chronically increased pressure within an AVG or AVG as a result of a venous outflow stenosis can result in several adverse events. Most visibly, the arm can swell. The outflow veins from the access can tolerate some increased pressure due to their capacitance, but these adaptations are overwhelmed at arterial pressures. If the intraluminal pressure is elevated high enough, swelling will ensue. In a manner analogous to chronic venous insufficiency in the lower extremities, this initially consists of acellular fluid alone, and produces reversible, pitting edema. However, as this process becomes more chronic, the fluid can include cells, cellular debris, proteins, and other macromolecules. The increased pressure will also produce a greater tendency to cause bleeding (both during and after cannulation) and increased recirculation as more and more “clean” blood is redirected back into the arterial needle to pass through the dialysis machine leading to inefficient dialysis. It is overly simplistic to lump all vascular access-associated venous stenoses or occlusions together and to treat them with a single interventional technique. Usually, stenoses in the arm or chest can be effectively treated by an endovascular means, while occlusions require an open surgical approach. We believe that the optimal treatment for stenoses that occur at the costoclavicular junction, however, must include surgical decompression of the subclavian vein.
Diagnosis and Evaluation Patients with venous stenoses will usually present with one of three problems: venous hypertension, prolonged bleeding, or inefficient dialysis. Venous hypertension should be fairly obvious and patients can present with swelling ranging from soft and pitting to brawny and indurated, depending on the chronicity of the problem. Although we have never seen a true venous ulceration, some patients have presented with increased pigmentation and lipodermatosclerosis. The swelling (and skin changes) can involve the whole arm or be limited to the forearm and the area overlying the fistula itself. The swelling can lead to patient discomfort and can complicate cannulation during dialysis. The increased bleeding can occur during dialysis or after de-cannulation and results from an increased pressure gradient across the puncture site (ie, lumen to outside world). This is often exacerbated by illadvised pressure held diffusely along the vein (rather than gentle point pressure on the site itself), which will increase the intraluminal pressure and further increase bleeding. The inefficient dialysis results from the recirculation of “clean” blood back into the dialysis machine from reflux of blood back into the arterial needle as a result of the elevated venous pressures. These elevated pressures can trigger alarms on the machine, but it is not the elevated pressures per se that are the problem, rather the inefficient dialysis. In the most extreme situation, an endless loop can occur in which the same blood is filtered through the machine with little net effect. All patients with symptoms related to presumed venous
stenoses or occlusions should undergo a fistulagram to identify the lesion and facilitate definitive treatment, most of which can be performed at the same setting. Although there is usually no sense of emergency, the fistulagram should be performed in a timely fashion because the access-related venous hypertension rarely improves without intervention. Furthermore, early intervention may be associated with improved results because it is easier to treat a stenosis than an occlusion. Although ultrasound may be helpful to image the access in the extremity, it cannot image the full extent of the venous outflow tract to the atrium, given the bony thoracic cavity.
Definitive Management: Anatomic-Based Approach The findings on the fistulagram or venogram will identify the anatomic problem and suggest the definitive treatment. There are three distinct anatomic locations where problems are commonly encountered and the vein can be stenotic or occluded at each of these sites. The modality chosen to treat an access-related central venous stenosis or occlusion is based on the location and nature of the venous lesion.
Central Veins Proximal to the Costoclavicular Junction This anatomic segment is comprised of the body of the access (ie, AVF), the peripheral outflow vein(s), and the brachial/ axillary veins up to the level of the costoclavicular junction (Fig 1). Although the access itself and the peripheral veins in the upper extremity are not usually considered along with the discussion of central venous lesions, they will be addressed for completeness. Venous stenoses, defined by the ability to cross the lesion with a wire, in these locations are almost always amenable to conventional endovascular techniques. Percutaneous access should be obtained peripheral to the lesion being treated (ie,
Figure 1 Fistulagram showing smooth stenosis in the axillary vein in the axilla (arrow). This responded well to balloon angioplasty.
Central vein stenoses and occlusions proximal on the access circuit). Lesions close to the arterial anastomosis can be approached retrograde through the access or through the inflow artery, although it may be difficult to establish a stable platform from the groin because of the anatomic distance. Conventional balloon angioplasty is usually the first step, and a reasonable amount of overdilation is safe in this low-pressure, superficial venous system. Balloon dilation of a stenotic lesion can be somewhat painful and necessitate additional sedation and/or pain medication. If there is significant recoil after angioplasty alone or the stenotic “waist” cannot be eliminated, a high-pressure or cutting balloon can be used as the next step in the treatment algorithm. Technical success can be assessed by the elimination of the stenosis, the disappearance of collateral channels and presence of a thrill in the access (or elimination of the pulsatile character). Intravascular ultrasound can be helpful in this setting to define the residual lumenal diameter. Consideration should be given to stenting the lesion if the stenosis persists. Stents, even in the segment of the vein used for cannulation, seem to be well-tolerated. Using these conventional techniques, a relatively good result can be expected although restenosis is common and an aggressive surveillance protocol is likely justified. Occluded outflow veins in the region peripheral to the costoclavicular junction require individualized treatment. An attempt can be made to cross the occlusion with a wire (and, if successful, treatment rendered as mentioned here), but in our experience these occlusions are usually chronic and cannot be crossed with a wire. In this situation, open surgical repair is usually required, with the goal being to establish adequate venous outflow. The surgical options are usually dictated by the extent and location of the occlusion and the status of the superficial and central veins. Potential solutions including mobilizing the distal segment of the AVF and reimplanting it on another vein with patent outflow (eg, distal cephalic vein-axillary vein), translocating the basilic or brachial vein (eg, distal cephalic vein-basilic vein), or using an interposition graft (autogenous or prosthetic) to bypass the lesion.
Central Veins Distal to the Costoclavicular Junction This zone is comprised of the innominate veins and the superior vena cava (SVC). Patients with stenoses at this location may exhibit signs and symptoms of SVC syndrome, which can be severe enough to threaten their airway. Innominate vein and SVC stenoses can usually be treated with conventional endovascular techniques, presuming the lesion can be crossed with a wire (Fig 2). Percutaneous access, including an introducer sheath, can be obtained through the AVF or AVG. It is important to “anchor” the distal end of the wire in the inferior vena cava (rather than the SVC or atrium) to assure that the platform is sufficient for successful treatment. We almost always use a self-expanding stent in this location, one that is sufficiently large in diameter to assure full wall apposition. Intravascular ultrasound is valuable for assessing the size of the normal, adjacent vessel.
115
Figure 2 Fistulagram of a patient with smooth stenosis of the superior vena cava (arrow). Note that his dialysis catheter, inserted from the left, does not seem to be involved with this lesion in any way. This responded well to placement of a 14-mm self-expanding stent that was angioplastied with a 12-mm balloon.
Although the balloon should be sized appropriately for the lumenal diameter, over dilation should be avoided because there have been anecdotal reports of fatal SVC rupture in this setting. At times it may be necessary to place a stent across the ipsilateral jugular or contralateral innominate vein. This is usually well-tolerated because the “caged” vein will remain patent via the blood flow through the interstices of the stent. Should subsequent intervention become necessary, the interstices of the stent covering the caged vein can usually be crossed, similar to stents placed across the orifices of the iliac veins for patients with iliocaval stenoses. Not infrequently, patients with an innominate or SVC stenosis will also have a dialysis catheter. If the central stenosis cannot be crossed with a wire (ie, between the wall of the vessel and the catheter), the wire can be passed through the catheter, the catheter removed, and the endovascular interventions performed. A new catheter can be placed at the completion of the procedure as necessary. Occlusions of the innominate vein and SVC are more problematic than stenoses. Aggressive attempts at crossing the occlusion with a wire are justified because the open surgical alternatives for salvaging the access and relieving the venous hypertension are associated with significant morbidity. Several techniques for blindly crossing central vein occlusions have been described, although these should not be attempted without significant endovascular experience.5 One technique involves the use of a retrograde approach to the SVC from the femoral vein. If the lesion cannot be crossed, the options include access ligation (and abandonment) or salvage with surgical revascularization or bypass through a sternotomy or thoracotomy.6,7 If the life expectancy of the patient is limited and alternative options for access placement exist, ligation and abandonment of the access is reasonable.
116
K.A. Illig
Central Veins at the Costoclavicular Junction Access-related central venous stenosis can develop in the segment of the subclavian vein adjacent to the costoclavicular junction, the point where the first rib and the clavicle come together at the sternum. The vein may be tethered and externally compressed by the subclavius muscle and tendon and the costoclavicular ligament at this location (Figs 3 and 4). In nondialysis patients, this anatomic narrowing may cause chronic injury to the vein, resulting in venous thrombosis (ie, venous thoracic outlet syndrome).8,9 It is our impression that the high venous flow and associated turbulence resulting from the arteriovenous access further exacerbates this chronic functional injury leading to stenoses and occlusions. Based on the collective experience from the treatment of venous thoracic outlet syndrome, it is apparent that subclavian vein stenoses adjacent to the costoclavicular junction respond poorly to treatment with simple balloon angioplasty.9 Furthermore, the use of stents in this location may actually be harmful because of the “nutcracker” effect of the clavicle and first rib, which can cause stent fracture and venous occlusion (Fig 5).10 We believe that the accepted treatment paradigms for venous thoracic outlet syndrome can be applied to the treatment of access-related venous stenosis at the costoclavicular junction and that surgical decompression (with concomitant endovascular intervention, if needed) provides the best hope for long-term success.
Figure 3 Basic anatomy of the thoracic outlet. The axillosubclavian vein passes anteriorly, passing by the junction of the first rib and clavicle. This “space” is open superiorly, but the vein is tethered in this location by surrounding tissue. The two bones and the subclavius muscle and tendon chronically and repetitively exert pressure on it. In patients with high flow (ie, with an ipsilateral arteriovenous fistula) this area can quickly become stenotic. (Reprinted from Illig KA, Doyle A: A comprehensive review of Paget-Schroetter syndrome. J Vasc Surg 51:1538-1547, 2010, with permission.9)
Figure 4 Fistulagram of a patient with “dialysis-associated venous thoracic outlet syndrome,” manifested clinically as massive arm swelling and fistula dysfunction. Note the rather long, complex stenosis beginning at the costoclavicular junction with fairly normal vein peripherally (solid arrow) and extensive collateralization that is pathognomonic for this lesion (open arrow).
We treat access-related subclavian vein stenosis at the costoclavicular junction by first performing a transaxillary first rib resection and mobilizing the subclavian vein to the jugular confluence. Often, there is a dense cicatrix surrounding the vein and it is important to perform an extensive venolysis to ensure complete venous decompression. Surprisingly, we have not noticed a difference in bleeding during the surgical decompression procedure in patients with access-related venous thoracic outlet compared to our “conventional” venous thoracic outlet patients. After the incision is closed, the patient is repositioned and a venous balloon angioplasty is performed as described here. Although some believe that surgical decompression of the vein permits the use of stents, we place stents across this vein segment only if absolutely necessary (Fig 6).10,11
Figure 5 A posterioanterior chest radiograph is shown for a patient with a right upper extremity autogenous access and an ipsilateral central vein occlusion. Note the compressed stent at the thoracic outlet (arrow). Interestingly, the patient also has a thoracic endograft.
Central vein stenoses and occlusions
Figure 6 The same patient shown in Figure 5 is shown again after transaxillary first rib excision with external venolysis and balloon angioplasty; note that the collateral channels are gone. The lesion within the innominate vein did not respond to venoplasty alone and required stenting. This vein remained patent and the fistula functional until the patient died of unrelated causes.
Treatment of subclavian venous occlusions at the costoclavicular junction generally requires direct reconstruction of the vein or a jugular transposition. Although the methods described for exposing this vein segment for direct reconstruction generally involve partial resection of the clavicle, a method of exposure that avoids claviculectomy also has been described.12,13 This latter method may be preferable in patients who are wheelchair-dependent because claviculectomy in such patients can be debilitating. These options may be considered to salvage the access and relieve symptoms in young patients who are fit and have a reasonable life expectancy. For patients whose life expectancy is short and alternative access sites exist, access ligation and abandonment may be more appropriate. An alternative approach that should be investigated in this setting includes extraanatomic bypass from the axillary/subclavian vein to the ipsilateral internal jugular or contralateral axillary. Although this is rarely an option given the widespread proliferation of tunneled dialysis catheters and the common finding that the jugular and axillary veins are occluded, it merits consideration. The axillary-jugular vein bypass is relatively straightforward. The axillary and jugular veins can be exposed through an infraclavicular and supraclavicular incision, respectively, and a large prosthetic graft (eg, 10 to 12 mm expanded polytetrafluoroethylene) can be easily tunneled between the incisions deep to the clavicle.
Results The literature reporting the outcomes for balloon angioplasty of access-related central vein stenosis is highly variable, with
117 6-month patency rates ranging from 30% to 60%.2,7,14,15 Unfortunately, the use of stents has not dramatically altered these marginal results. One potential reason for this variability is the fact that most reports combine outcomes for stenoses and occlusions and do not define the anatomic location of the lesion. The inclusion of patients with venous stenosis at the costoclavicular junction in these reports may account for many of these treatment failures. Two authors previously alluded to the notion that subclavian vein stenosis at the costoclavicular junction may respond poorly to endovascular intervention.5,7 However, there are no published series of access-related venous stenoses at the costoclavicular junction treated with concomitant surgical decompression and endovascular intervention. Interestingly, several of the figures contained in these articles documenting the outcomes for access-related central venous lesions actually involve lesions at the costoclavicular junction.2,5,7 Since November 2008, we have treated 12 patients with access-related central venous stenoses or occlusions with open surgical intervention.16 The procedures were performed to salvage an access in patients with severe venous hypertension and access dysfunction. All patients had multiple previous endovascular interventions (mean 2.3 interventions) at the costoclavicular vein segment. Eight patients underwent transaxillary first rib resection with venolysis and subsequent balloon angioplasty and/or stenting of the residual stenosis (Figs 4 and 6), while four patients underwent claviculectomy with direct venous repair or bypass. The overall mean follow-up in our series was 8 months. Salvage of the access occurred in eight cases and relief of the upper extremity swelling occurred in nine cases.
Conclusions Venous outflow stenosis is the most common cause for access failure. Guideline 20 of the Kidney Disease Outcomes Quality Initiative guidelines recommends percutaneous transluminal venoplasty with or without stenting as the preferred initial treatment for these lesions.17 However, not all of the venous outflow lesions are the same and treatment must be individualized by the nature and location of the lesion. Venous lesions in the arm or chest can usually be treated very effectively with conventional endovascular or surgical approaches. However, the lesions at the costoclavicular junction are intrinsically different because of extrinsic compression at thoracic outlet. As in venous thoracic outlet syndrome, the optimal treatment of access-related venous stenosis at this location must include surgical decompression of the vein with first rib resection and venolysis.
References 1. Arteriovenous Fistula First: Fistulat First data. Available at: http:// www.fistulafirst.org/AboutFistulaFirst/FFBIData.aspx. Accessed May 15, 2011 2. Agarwal AK: Central vein stenosis: current concepts. Adv Chronic Kidney Dis 16:360-370, 2009 3. Roy-Chaudhury P, Spergel LM, Besarab A, Asif A, Ravani P: Biology of arteriovenous fistula failure. J Nephrol 20:150-163, 2007
118 4. Illig KA, Surowiec S, Shortell CK, Davies MG, Rhodes JM, Green, RM: Hemodynamics of distal revascularization-interval ligation. Ann Vasc Surg 19:199-207, 2005 5. Kim YC, Won JY, Choi SY, et al: Percutaneous treatment of central venous stenosis in hemodialysis patients: long-term outcomes. Cardiovasc Intervent Radiol 32:271-278, 2009 6. Glass C, Maevsky V, Massey T, Illig KA: Subclavian vein to right atrial appendage bypass without sternotomy to maintain AV access in patients with complete central vein occlusion: a new approach. Ann Vasc Surg 23:465-468, 2009 7. Dosluoglu H, Harris LM: Hemodialysis access: nonthrombotic complications, in Cronenwett JL, Johnston KW (eds): Rutherford’s Vascular Surgery (ed 7). Philadelphia, Saunders Elsevier, 2010, pp 1137-1154, 8. Aziz S, Straehley CJ, Whelan TJ Jr: Effort-related axillosubclavian vein thrombosis: a new theory of pathogenesis and a plea for direct surgical intervention. Am J Surg 152:157-161, 1986 9. Illig KA, Doyle A: A comprehensive review of Paget-Schroetter syndrome. J Vasc Surg 51:1538-1547, 2010 10. Meier GH, Pollak JS, Rosenblatt M, Dickey KW, Gusberg RJ: Initial experience with venous stents in exertional axillary-subclavian vein thrombosis. J Vasc Surg 24:974-981; discussion 981-983, 1996
K.A. Illig 11. Kreienberg PB, Chang BB, Darling RC 3rd, et al: Long-term results in patients treated with thrombolysis, thoracic inlet decompression, and subclavian vein stenting for Paget-Schroetter syndrome. J Vasc Surg 33:S100-S1005, 2001 (suppl 2) 12. Hu SH, Dilley RB: Internal jugular vein turndown for subclavian vein occlusion. Oper Tech Gen Surg 10:149-153, 2008 13. Green RM, Waldman D, Ouriel K, Riggs P, DeWeese JA: Claviculectomy for subclavian venous repair: long-term functional results. J Vasc Surg 32:315-321, 2000 14. Beathard GA: Percutaneous transvenous angioplasty in the treatment of vascular access stenosis. Kidney Int 42:1390-1397, 1992 15. 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 45;776783, 2007 16. Glass C, Dugan M, Gillespie D, Doyle A, Illig KA: Costoclavicular venous decompression in patients with threatened arteriovenous hemodialysis access. Ann Vasc Surg 2011 Apr 20. [Epub ahead of print] 17. NKF-K/DOQI clinical practice guidelines for vascular access, 2000. Am J Kidney Dis 2001;37:S137–S181, 2001 (suppl 1)