Chapter 57. Interventional Nephrology: Endovascular Procedures

Over the past decade, there has been a resurgence of interest by nephrologists in the management of hemodialysis vascular access. The early days of dialysis were marked by advances in vascular access conceived and developed by visionary nephrologists, including the Scribner shunt and the Brescia-Cimino arteriovenous (AV) fistula. Without these means of obtaining reliable repeated blood access, the delivery of chronic hemodialysis would not have been possible. Some nephrologists have maintained this primary role in the creation and maintenance of vascular access, particularly in Europe. One successful example reported the construction of a series of 748 consecutive native AV fistulas, with 2 year assisted access survival rates in diabetics and nondiabetics ranging from 75% to 96%. During the 1970s and 1980s, at least in the United States, interest and involvement in vascular access largely faded. This may have been due to exciting progress in what were perceived to be more scientifically rewarding areas of study, as opposed to the relatively mundane “plumbing” problems of vascular access. Certainly neither technical proficiency nor rigorous academic attention to vascular access was emphasized in most nephrology training centers in the United States during that time. In many programs and practices management of vascular access was left exclusively to the surgeons. At the same time, particularly in the United States, there was increased promotion and utilization of synthetic polytetrafluoroethylene (PTFE) grafts in favor of native AV fistulas. This shift may have been driven by marketing and reimbursement practices, poor long-term venous access catheters available for use as “bridges” to native fistulae, and increasing emphasis on short, high efficiency dialysis treatments. The result for the United States nephrology community was a large hemodialysis patient population with a high prevalence of PTFE grafts, a low usage of AV fistulas, and perhaps incidentally, the highest dialysis patient mortality of all industrialized nations. In 1999, 49% of hemodialysis patients in the United States were dialyzing with AV grafts, 28% with native fistulas, and 23% with venous catheters.

During this period of a rapidly growing hemodialysis patient population, increasing PTFE graft utilization, and decreased involvement of nephrologists in the management of vascular access problems, there was a predictable crisis in the access-related medical care of these patients. Management of access dysfunction and thrombosis was largely “reactive” and primarily utilized open surgical techniques. The role of venous stenosis in contributing to AV graft thrombosis and failure was underappreciated. In the late 1980s, interventional radiologists began to recognize these problems and applied their tools and techniques to treating access dysfunction. A method for declotting AV hemodialysis grafts using pharmacomechanical thrombolysis and angioplasty was reported in 1991. Numerous other reports and variations on this method followed, with increasing acceptance of percutaneous interventions in the management of hemodialysis access dysfunction. Largely, however, nephrologists remained on the periphery, as vascular access continued to be the province of the vascular surgeon and more recently interventional radiologists. This collaboration of expert subspecialties might have been all that was needed to provide timely, high-quality hemodialysis access care. Undeniably, in some settings, this was the case. However, while access dysfunction was of critical and immediate importance to the patient, dialysis unit, and nephrologist, for many programs this could not be the first priority for the surgeons or radiologists, creating a service void and an opportunity for improvement in care.

The central role of vascular access in the care of patients on hemodialysis cannot be overemphasized. Comprehensive medical care of the hemodialysis patient includes management of uremia, hypertension, sodium and water balance, anemia, mineral metabolism, metabolic bone disease, and nutritional status. This care cannot be properly delivered unless there is reliable, efficient blood access for dialysis. Leaving this critical aspect of care entirely in the hands of others puts the patient and the nephrologist at a significant disadvantage. Under ideal circumstances, when the skills and priorities of the multispecialty access team come together, patient care may be very well served. Conversely, if the appropriate surgical or interventional services cannot be delivered in a timely fashion, the patient may suffer in terms of delayed dialysis, temporary venous hemodialysis access, unnecessary hospitalization, or other avoidable morbidity. This also may result in a significantly greater financial burden to the health care system.

In the early 1990s, this problem was recognized and confronted by Dr. Gerald Beathard in Austin, Texas. He acquired the necessary training, adapted the reported interventional radiology techniques, and developed a nephrology-run service for percutaneous management of vascular access. He then liberally shared this expertise, training many nephrologists from various practices and backgrounds, this author included. It is a remarkable fact that all interventional nephrologists practicing in the United States can trace their roots directly back to Dr. Beathard. As these nephrologists brought these techniques to their practices, the field of “interventional nephrology” was effectively born. When nephrologists began to perform these access-related percutaneous interventions, the first priorities were to master the techniques, establish suitable facilities in which to work, and then deal with the multitude of day-to-day access failures, largely centered on the PTFE graft. While this led to an immediate and dramatic improvement in care, it was very evident that the poor performance of PTFE grafts compared to native AV fistulas was contributing to an excessively high rate of access failure and hence a large volume of percutaneous interventions. This was good business, but very bad medicine. This realization led to the next phase in the evolution of interventional nephrology, which was to take on comprehensive vascular access management for the patient on hemodialysis. To improve vascular access outcomes, it would be folly to address only the technical aspects of percutaneous interventions. To achieve optimal vascular access, the following aspects of care take on equal or greater importance.

• Preservation of peripheral veins for native AV fistulas.
• Avoidance of peripherally inserted central venous catheters and subclavian catheters.
• Judicious use of internal jugular vein tunneled hemodialysis catheters for temporary hemodialysis access.
• Early referral to a surgeon for construction of a native AV fistula.
• Education and selection of surgical colleagues willing and able to master the techniques for the creation of a native AV fistula.
• Preoperative imaging of veins and arteries, including ultrasound vein mapping and venography.
• Evaluation and treatment of a poorly functioning or immature AV fistula.
• Conversion from a failing PTFE graft to a “secondary” native AV fistula.
• Maintenance of a complete and accurate clinical database with regular quality analysis including rates of usage of AV fistulas and catheters success rates of technical procedures, complications, and patient satisfaction.

Another essential principle must be recognized in order to provide optimal hemodialysis access care: For each scenario of access dysfunction, there is a “best solution.” There has been a tendency for vascular access solutions to follow “the path of least resistance” or, worse yet, “the path of greatest reimbursement,” both of which may be very different from the optimal pathway. From the perspective of the interventionist, there is the temptation to view all problems as best solved using percutaneous means. However, there are clearly situations in which surgical solutions are preferable and should be employed. This of course works both ways. It would be equally poor care to perform frequent, repeated percutaneous thrombectomies and angioplasties on a failing PTFE graft as it would be to attempt an open surgical declot of a fistula that thrombosed due to central venous stenosis. Knowing the anatomy and history of each patient is the critical element that allows these judgments to be made correctly. In this regard, keeping a detailed clinical database is essential. This allows the operator to fully assess the vascular access problem at hand, make the most appropriate management decision based upon the history and known anatomic factors, and perform the necessary procedure with the lowest risk and best possibility of a successful outcome.

When a patient presents with access dysfunction, a timely solution is required. This may represent an urgent problem such as a thrombosed AV fistula that requires immediate intervention to salvage the access and provide dialysis. Although arguably not a true “medical emergency,” restoration of fistula or graft function is of paramount importance to the care of the patient on hemodialysis, with potentially grave implications for both short-term and long-term morbidity and mortality. Other access problems may be less immediate, such as prolonged bleeding from needle puncture sites related to venous outflow stenosis; this may not prevent dialysis, but puts the patient at risk, may lead to access thrombosis, and should be dealt with before it becomes an urgent problem. Other access problems may be relatively elective in nature, such as the evaluation of limb swelling associated with central venous stenosis or a slowly enlarging pseudoaneurysm. In all cases, the goal of an interventional program should be to address each problem in an appropriate timely fashion, minimizing disruption of the dialysis schedule of patients. Nephrologists are in an ideal position to provide these services when equipped with the necessary skills, allowing for seamless delivery of care and management of dialysis-related problems as required before, during, and after an interventional procedure. Alternatively, there are many practices in which excellent care and service are provided by interventional radiologists or vascular surgeons. The title of the individual responsible for vascular access interventions is not as important as his or her knowledge, skill, availability, and willingness to work with surgeons, nephrologists, and the dialysis staff as part of a multidisciplinary access management team.

The American Society of Diagnostic and Interventional Nephrology (ASDIN) was founded in 2000 to promote the proper application of procedures in the practice of nephrology. These include diagnostic ultrasound imaging and peritoneal dialysis catheter placement in addition to the full complement of percutaneous interventions required for management of hemodialysis vascular access. A central goal of this society was to develop standards for training, certification, and accreditation relative to these diagnostic and interventional disciplines. These were published in 2003 and remain the only standards specific to these procedures. Previously there were no specific standards and there was tremendous variation in the training and credentialing requirements of health care facilities. It is expected that with increasing awareness and acceptance of these criteria in the United States and elsewhere, interventional training will become more uniform and rigorous, ultimately improving the quality of care offered as well as patient outcome. Over the past several years, interventional nephrology training programs have become established at some U.S. academic centers. As these centers develop, they should be expected to advance the standard for training, quality, and clinical research related to vascular access.

The core procedures of interventional nephrology include the placement and management of venous hemodialysis catheters, diagnostic imaging of AV accesses and native veins, percutaneous angioplasty, and percutaneous thrombectomy of occluded AV access. Other related procedures include the placement of venous stents, the ligation or embolization of native fistula accessory branches, and the implantation of subcutaneous venous hemodialysis access ports.

Venous Hemodialysis Catheters

Venous catheters are the least desirable method of hemodialysis access. In a sense, every placement of a venous catheter represents a failure to prepare a native fistula in advance of initiating dialysis as well as a failure to detect a dysfunctional AV access and to intervene preemptively either to maintain its function or to create a new alternative access. Nevertheless, venous catheters are an unavoidable necessity for many patients who do not have functional AV access. For most interventional nephrologists, insertion of tunneled dialysis catheters is the first and most basic procedure acquired, building on the common skill of temporary hemodialysis catheter placement. Nevertheless, the risks of this procedure should not be understated, with the potential for severe injury to major central vessels from large-bore catheters and dilators. The use of real-time ultrasound guidance is widely considered to be essential for safe and efficient venipuncture based on higher procedural success rates and fewer complications compared to the use of landmarks only. The low posterior approach to the right internal jugular vein is ideal; this keeps the catheter low on the neck, with minimal patient discomfort and a good cosmetic result, and creates a smooth bend of the catheter that avoids kinking. Fluoroscopy is also recommended for proper catheter tip positioning, although there is limited evidence to support this requirement. There is general agreement that the performance of chronic dialysis catheters is optimized when the catheter tips are placed in the right atrium, and Disease Outcomes Quality Initiative (DOQI) guidelines support this approach. Nevertheless, this remains controversial, and achieving the desired tip position is difficult, even with fluoroscopy and careful attention to anatomic landmarks. In any case, to achieve the best possible outcomes, the operator must adhere to a meticulous sterile surgical technique, utilize ultrasound guidance for venipuncture, and pay careful attention to tip positioning using fluoroscopic and/or landmark guidance. Figure 57–1A shows a poorly placed left internal jugular vein tunneled dialysis catheter, with a high vein puncture from the anterior approach, a tight bend with the catheter kinked, and its split tips extending only into the superior vena cava. This catheter did not function for dialysis. Figure 57–1B shows the catheter replaced with a new Ash-Split catheter (Medcomp, Harleysville, PA), using the right internal jugular vein puncture from a low posterior approach with a smooth bend in the neck. In Figure 57–1C, the new catheter tips are shown extending into the high right atrium, yielding excellent catheter function as required for delivery of hemodialysis.

Over the past several years, two totally implantable venous hemodialysis access devices have been developed: the LifeSite hemodialysis valve (Vasca Inc., Tewksbury, MA) (Figure 57–2), and the Dialock Access System (Biolink Inc., Norwell, MA). In the United States, only the LifeSite device became available. The LifeSite system consisted of two independently implanted subcutaneous valves, each connected to a single-lumen catheter placed into the right atrium. These were accessed using 14-gauge needles through a “buttonhole” tract. Disinfection with isopropyl alcohol before and after each dialysis needle access was essential to reduce serious device infections. The LifeSite system provided improved device survival and lower rates of infection than conventional tunneled dialysis catheters. Interventional nephrologists have been instrumental in the LifeSite clinical trials and the application of the LifeSite device in practice. However, considerably greater time, skill, risk, and expense were associated with the implantation of these devices, factors that had to be weighed when considering the use of these devices versus the use of a conventional venous catheter. There is also a significant “learning curve” for both the implanting physician and the dialysis nursing staff, requiring a number of patients to be maintained in a program with the LifeSite device in order to achieve consistent successful results.

Because the product did not gain wide acceptance in most nephrology communities, the manufacturing company, Vasca Inc., eventually decided to withdraw LifeSite from the market in 2005.

Arteriovenous Access Dysfunction

Stenosis associated with AV grafts due to neointimal hyperplasia typically occurs at or near the venous anastomosis. This may be triggered by turbulent blood flow, the shear force on the vessel wall, and material compliance mismatch between the synthetic graft and vein. Proliferation of smooth muscle cells, endothelial cells, and fibroblasts leads to stenosis, as seen in Figure 57–3. The pathophysiology of stenosis affecting native AV fistulas is not as well studied or understood and may involve different factors in the absence of a surgically constructed venous anastomosis or synthetic material. Trauma to the vein during surgical manipulation or from intravenous catheters may be a contributing factor. Stenosis may also affect other sites in the AV access system, whether native fistula or synthetic graft. These include the arterial anastomosis (Figure 57–4), the fistula or draining peripheral vein (Figures 57–5 and 57–6), the central veins (Figure 57–7), or in the case of synthetic grafts, the body of the graft itself. Left untreated, progressive stenosis will lead to a reduction in access blood flow, ineffective hemodialysis, and ultimately access thrombosis.

Numerous studies have evaluated the role of screening and preemptive intervention for stenosis associated with synthetic AV hemodialysis grafts in maintaining function and preventing thrombosis. These studies are based on the premise that graft dysfunction and thrombosis are commonly associated with stenosis of the venous outflow, typically at the venous anastomosis. Outflow obstruction will result in diminished AV access flow, with a large pressure gradient across the stenosis, causing pressures within the access to rise toward arterial pressure. The utility of a good history and physical examination of the access should not be underestimated in detecting stenosis. Prolonged needle site bleeding, or an excessively pulsatile quality of the access, without a prominent “thrill” may indicate outflow stenosis or poor flow that may predispose the patient to thrombosis. Early studies suggested that elevated pressures measured through the venous dialysis needle during hemodialysis were predictive of PTFE graft thrombosis, and that preemptive angioplasty of these lesions was effective in reducing the rate of graft thrombosis. This relatively crude technique may be enhanced by a reduction of dialysis circuit flows, by more sophisticated analysis of dynamic venous pressure screening, or by measurement of static pressures within the graft in the absence of dialysis circuit flow. A number of studies have shown that the periodic measurement of access flow rates is useful in detecting graft stenosis, with intervention for low or reduced flow leading to reduced graft thrombosis. However, this remains somewhat controversial, and it is not clear that preemptive intervention results in longer overall access survival.

It should be noted that there are no comparable published studies evaluating screening and intervention for native AV fistulas despite the fact that this is the preferred form of hemodialysis access. As the nephrology community in the United States successfully transitions to a patient population dominated by native fistulas rather than synthetic grafts, questions concerning surveillance and intervention for failing PTFE grafts will become less relevant.

Percutaneous Transluminal Angioplasty

Angioplasty of a vascular stenosis is an essential tool in the management of hemodialysis access dysfunction. The principles and techniques of percutaneous angioplasty are well established, and these must be mastered by the interventional nephrologist, radiologist, or surgeon performing the procedure. There is considerable controversy as to whether angioplasty is an “effective therapy.” While it is almost always possible to dilate a stenosis and achieve an anatomically improved appearance, it is not clear that this will result in a sustained resolution of stenosis or improved long-term access patency. Elastic recoil may occur immediately following an apparently successful angioplasty. The angioplasty procedure itself is injurious to the vessel and may promote rapid restenosis. There are clearly lesions that will respond more favorably to surgical correction with improved duration of patency compared to percutaneous treatment.

It should be noted that some venous stenoses, especially those related to native AV fistulas, are quite resistant to expansion but may be successfully dilated using specialized cutting balloons or ultrahigh-pressure angioplasty. Conventional angioplasty balloons are rated with “burst pressures” of 12–15 atm and are typically operated at pressures up to 50% greater than this rating. Even this pressure may not be sufficient to dilate certain resistant lesions, which may require the use of ultrahigh-pressure balloons, such as the Conquest balloon (Bard Peripheral Vascular Inc., Tempe, AZ), with burst pressure ratings of 30 atm. Figure 57–5A shows a right upper arm cephalic vein fistula referred for evaluation of excessive dialysis needle-site bleeding. On physical examination the fistula was firm and markedly pulsatile, with a poor palpable thrill. Severe stenosis was demonstrated in the mid portion of the upper-arm cephalic vein, correlating with the history and physical findings. The lesion was extremely resistant to angioplasty (Figure 57–5B), ultimately responding with sustained inflation at ultrahigh pressure, >35 atm (Figure 57–5C). Postangioplasty there was improved flow, with restoration of a palpable thrill; a fistulogram demonstrated improved stenosis along with a ragged vessel lumen likely representing an intimal tear and disruption (Figure 57–5D). This indicates that the angioplasty procedure itself is injurious to the vessel and may incite local factors leading to further neointimal hyperplasia, contributing to eventual restenosis.

Percutaneous Thrombectomy

Prior to the 1990s, management of AV access thrombosis was almost exclusively surgical. Diagnosis of associated venous stenosis was limited, and there was very little preventive intervention. Since the first publications of percutaneous methods for declotting AV grafts, many devices and techniques for declotting hemodialysis access have been reported. For the most part, no one method or device has been shown to have any advantage over others in terms of procedural success, complications, or duration of access patency. The principal determinant of outcome is the correction of underlying stenoses or other conditions responsible for the thrombosis. The most notable progress has been made in the management of native AV fistula thrombosis. Until very recently, salvage of a thrombosed native fistula was not considered possible. Several studies have now shown that results of percutaneous native fistula thrombectomy can be excellent, ranging from 76% to 94%. Most of these studies, and our own technique of thrombus aspiration, involve a combination of pharmacologic thrombolysis and/or mechanical clot removal or maceration, with no proven difference in efficacy or safety between different methods. Long-term secondary fistula patency after successful declotting is also quite favorable, with 50–86% remaining patent at 24 months. It should be emphasized that to achieve this secondary or “assisted” patency and to avoid repeated thrombosis, other percutaneous interventions may be required. In the United States, as we continue to improve our utilization of native fistulas, and in other countries in which this has already been achieved, there clearly will be more emphasis on the application of percutaneous interventional techniques for the maintenance of native fistula function. Furthermore, as we attempt to create functional fistulas in more patients with poor quality veins, it is likely that the rate of native fistula dysfunction and thrombosis will be higher in these “marginal” fistulas than in those previously reported.

Venous Stents

The role of endovascular stents in the management of stenosis associated with AV hemodialysis access is poorly defined and rather controversial. This remains in sharp contradistinction to coronary and renal artery stenosis, where primary stent placement is widely accepted to be the appropriate standard for most lesions. No such clear data exist to demonstrate improved outcomes of stents versus simple angioplasty in hemodialysis vascular access applications. In general, a stent should not be placed at a site where a definitive surgical revision would be technically feasible, given the opportunity for prolonged access function following surgical revision. Stent placement should be considered only after the failure of conventional angioplasty, due to either immediate recoil, rupture, or repeated rapid restenosis, or in the management of complete vessel occlusions.

There are a variety of self-expanding metallic stents suitable for this purpose, although only the stainless-steel Wallstent endoprosthesis (Boston Scientific, Natick, MA) has been approved in the United States for a central venous indication. Other stents have biliary, tracheobronchial, or arterial indications, but are commonly utilized in the venous system “off-label.” These are primarily nitinol (nickel-titanium alloy) stents, which have improved radial force, flexibility, and other material property advantages compared to stainless-steel stents. A recent uncontrolled, retrospective study utilized the nitinol SMART Stent (Cordis/Johnson & Johnson, Warren, NJ) for treatment of central or peripheral vein stenosis and demonstrated a primary patency of 14.9 and 8.9 months, respectively. This study also suggested that the SMART Stent conferred longer periods of intervention-free access patency than angioplasty alone. Clearly further study will be needed to determine the appropriate indications for venous stents. It is possible that drug-eluting stents used for treatment of AV hemodialysis access stenosis may result in less rapid or severe restenosis, by inhibiting local proliferative responses. One of the challenges for the interventional nephrology community is to take the lead on such studies and advance the clinical scientific basis for vascular access care.

Potential sites for stent placement include the peripheral draining veins related to a native fistula (Figure 57–6) or synthetic graft, central veins (Figures 57–7 and 57–8), graft venous anastomosis, or intragraft for the treatment of stenosis. Covered stent grafts may be particularly useful for the treatment of graft or fistula pseudoaneurysms in situations in which open surgical repair is not practical. Figure 57–9A shows a left upper arm PTFE graft segment with multiple large pseudoaneurysms resulting from difficult needle cannulation and poor closure of needle puncture sites. An 8 mm × 80 mm Fluency tracheobronchial stent graft (Bard Peripheral Vascular Inc., Tempe, AZ) was placed with complete exclusion of the pseudoaneurysms (Figure 57–9B).

Fistula Accessory Branch Ligation

To be usable for hemodialysis access, a native fistula must have sufficient flow through an accessible vein or veins. In selected cases, accessory veins may divert flow from the dominant fistula vein into branches that are not suitable for needle access. This may leave insufficient flow through the dominant vein for dialysis, especially when there is relatively poor blood delivery due to small or diseased peripheral arteries, low blood pressure, or poor cardiac output. In these circumstances, in addition to correcting any flow-limiting stenosis, ligation or embolization of accessory veins may result in improved fistula performance. This can be done using minimally invasive techniques, with immediate confirmation of successful branch occlusion. Figure 57–10A shows an upper-arm brachial artery-to-cephalic vein fistula with a single accessory vein diverting flow from the preferred vessel; this fistula was not functioning well for hemodialysis due to inadequate maturation of the vessel and ongoing difficulty with needle access. Ligation of the accessory vein was performed, with elimination of flow into this vessel on repeat fistulogram (Figure 57–10B). When selecting accessory veins for occlusion, great care must be taken not to sacrifice veins that might eventually be suitable for needle cannulation or useful for a surgical revision. Furthermore, collateral veins that are providing outflow due to occlusion or stenosis of the dominant vein should never be sacrificed. In these cases, the outflow stenosis should be corrected before any vein branches are considered expendable.

Interventional nephrology is an exciting and rapidly evolving field. It is extremely important to address and resolve the critical vascular access problems that arise in patients on hemodialysis. Our ability to provide timely, efficient, high-quality, and cost-effective service is central to the care of these patients and to the operation of our dialysis units and practices. Our role as nephrologists also provides multiple opportunities for improvements in vascular access, including early vein preservation, the creation of AV access in advance of the need for dialysis, the minimization of venous catheter use, monitoring and screening for access dysfunction, and planning for the creation of secondary native fistulas in anticipation of impending access failure. Working closely with surgeons, radiologists, nurses, and dialysis technicians, our challenge is to continue to improve the quality of access care that is provided to patients on dialysis. Improved hemodialysis access care and enhanced outcomes will result in higher native fistula rates, lower patient morbidity and mortality, reduced requirements for inpatient care, and reduced costs to the health care system.