Management of hemodialysis catheters

Ziv J. Haskel, MD, FSCVIR and Michael C. Cohn, MD

Dr. Haskel is Associate Professor of Radiology and Director of the Interventional Radiologic Research Laboratory, and Dr. Cohn is Clinical Instructor in Radiology at the Hospital of the University of Pennsylvania in Philadelphia, PA.

In 1991, approximately 200,000 Americans were treated for end-stage renal disease 60% with hemodialysis (HD). ¹ In 1993, over 160,000 patients with chronic renal failure underwent HD. ² The yearly incidence of patients with end-stage renal disease has nearly doubled during the last decade. Although renal transplantation is the preferred treatment for most renal failure patients, the availability of donor organs continues to limit growth in the number of transplants. Thus, most patients are maintained with hemodialysis. As the average age of the U.S. population continues to rise, both the prevalence of dialysis and associated comorbid conditions increase.

Creation and maintenance of vascular access for hemodialysis continues to present a complex challenge for nephrologists, surgeons, and interventional radiologists. Present methods of dialysis access include surgically created arteriovenous fistulae (e.g., Brescia Cimino fis tulae ³ , prosthetic arteriovenous grafts, and dual lumen central venous catheters. Many physicians consider the autogenous radiocephalic fistula to be the most desired option because of its high long term patency and low rate of associated complications. Indeed, the Dialysis Outcomes Quality Initiative (DOQI) guidelines recommend that physicians try to establish working native fistulae in as many dialysis patients as feasible. ⁴

Despite the relatively trouble free nature of mature arteriovenous fistulae, the majority of patients in the United States are dialyzed through prosthetic grafts. Results from the Medicare End-Stage Renal Disease program revealed that 83% of patients who were dialyzed for more than 3 months were treated with prosthetic grafts. ⁵ These grafts carry their own set of complications, including pseudoaneurysms, infection, ipsilateral extremity edema due to venous hypertension and, above all, graft thrombosis. Graft failure most commonly is the result of smooth muscle cell proliferation at and just beyond the venous anastomosis, and often it requires repeated angioplasty, stenting, or surgical revisions to preserve patency.

Central venous catheter placement for HD has replaced the original Scribner external arteriovenous shunt. ⁶ The procedure may be indicated for patients with acute renal failure requiring temporary HD, and for patients awaiting renal transplantation or maturation of surgical prosthetic grafts or fistulae. In some cases, these catheters must provide life long HD access because all potential fistula sites have been depleted. In 1979, Uldall et al emphasized the advantages and convenience of transvenous subclavian access using a double lumen catheter. ⁷ Since then, many centers have reported their results using similar catheters for temporary and permanent HD access. ^7-33

Temporary hemodialysis catheters

Temporary hemodialysis catheters are noncuffed, semistiff, tapered, dual lumen devices designed for short term HD, plasmapheresis, or photopheresis. They are available in a variety of lengths and are placed using standard percutaneous techniques. The majority of temporary HD catheters are constructed of polyurethane, which has the advantage of tensile strength, allowing construction of large bore catheters with thin walls. It is also stiff enough for percutaneous insertion without a sheath, and yet it softens in response to body temperature. ³⁴

Typically, a 15-cm long catheter is suitable for right internal or external jugular catheters, while 20-cm devices are necessary for left-sided catheters. These catheters can be bonded with antiseptic substances, such as silver sulfadiazine and chlorohexidine, which may decrease bacterial colonization of the catheter. ³⁵ While some investigators have reported satisfactory long-term success using these catheters, ^32,33 most practitioners agree that tunneled access catheters provide improved catheter stability and lower infection rates when used for long-term access. DOQI guidelines recommend use of a tunneled catheter if the length of catheterization is expected to exceed 3 weeks. ^4,36 In our experience, the steps required to tunnel and place a cuffed long-term catheter add little additional time to catheter insertion. Accordingly, we now place tunneled catheters in all patients in whom we anticipate the need for more than 2 weeks of HD. This policy has reduced catheter infection rates and the incidence of accidental catheter withdrawal, and has improved patient acceptance of central catheters. The exceptions to this policy include patients with bacteremia and those with uncorrectable coagulopathies. Ideally, prothrombin time (PT) should be corrected to less than 16 seconds, international normalization ratio (INR) less than 1.3, and platelets to more than 50,000/mm ³ before tunneled catheter placement; for translumbar catheter placement, the PT is corrected to normal if possible. In patients with bacteremia, placement of a tunneled catheter is deferred until blood cultures are proven to be negative and the patient is afebrile for at least 24 hours. ^37,38

Long-term/permanent HD catheters

In the past 5 years, percutaneous placement of long-term or permanent hemodialysis catheters by radiologists, nephrologists, and surgeons has dramatically increased. The advantages of radiological placement include accurate and safe venipuncture using sonographic guidance, precise final catheter positioning, the ability to provide same day service, the use of light conscious sedation instead of general anesthesia (and resultant shorter recovery times), and the potential for reducing costs (figure 1).

FIGURE 1. A properly positioned tunneled, left internal jugular hemodialysis catheter. Note the smooth curve of the catheter in the subcutaneous tunnel (arrows). The catheter tip is in the right atrium.

The technical skill in catheter placement under fluoroscopic and sonographic guidance that is intrinsic to interventional radiology makes primary catheter misplacement and malfunction rare. This has driven corporate development toward catheters and introduction kits more suited to nonsurgical techniques. A number of commercial catheter sets have been developed to approach these goals. They are largely distinguished by the outer shapes of the catheters (round vs oval), the shapes of the lumina (semilunar vs round), and whether they are a single catheterdouble lumen design, or a twin cathetersingle shaft design. Most of these catheters have a Dacron cuff that is positioned within the subcutaneous tunnel at least 2 cm from the skin exit site. The Dacron cuff eventually becomes ingrown with connective tissue and provides a secure anchor and mechanical barrier to infection. ³⁹ Placing the catheter tip in theright atrium results in fewer catheter malfunctions when compared to placing the tip in the superior vena cava. ^40,41

The results of catheter placement can be assessed in terms of technical success, complication rates, vascular thrombosis, catheter patency, and infection rates. Unfortunately, much of the relevant published literature on these topics consists of retrospective reviews of single site experiences. Prospective comparative trials of different catheter designs and placement approaches are forthcoming.

Technical success

Technical success can be defined as the ability to introduce a tunneled catheter into a venous location to allow immediate and effective HD. In one of the largest series of radiologist placed catheters, Lund et al retrospectively reviewed the outcome of 237 subclavian tunneled silastic catheters. Immediate technical success was not noted. ¹⁸ Nevertheless, operators should expect a very high technical success rate once they become familiar with central venous punctures and catheter placement techniques.

Relatively caudal jugular catheterization between the two heads of the sternocleidomastoid muscle generally is preferable to the anterior approach because of the resultant gentler curve and less kinking of the catheter as it turns caudally into the subcutaneous tunnel. The external jugular vein, often prominent in HD patients because of their high intravascular fluid states, also can be used for vascular access. Particular attention must be paid to the course of the catheter as it passes from the external jugular venotomy to the subcutaneous tunnel. In our experience, catheters placed in this vessel may be more prone to kinking because of the highly superficial position of the catheter and the sharp turn it must make into the tunnel (figures 2,3,4).

FIGURE 2. Demonstration of use of the external jugular vein for tunneled catheter placement. The superficial entry into the vein has led to kinking of the catheter (arrow).

FIGURE 3. Unsatisfactory catheter function due to poor catheter positioning and fibrin sheath formation. A tunneled catheter was introduced through a high internal jugular approach (arrow). The catheter tip lay against the wall of the superior vena cava (curved arrow). A fibrin sheath encases the catheter shaft within the inominate vein (arrowheads).

FIGURE 4. Inadvertent catheter retraction. An initially well positioned leftsided catheter (A) pulled back several centimeters after the patient sat up (B). The operator did not take into account the effect of the thick soft tissues overlying this patient's chest.

In the last few years, small transducer sonographic units have become available for vascular guidance (e.g., Site Rite, Bard Access Systems, Salt Lake City, UT). These relatively inexpensive units can easily be moved from room to room and sterilely draped. Sonographic guidance of jugular punctures has proved invaluable in choosing venous access sites (particularly in patients with multiple prior catheters) and in minimizing puncture site complications. In over 400 HD catheter placements and reinterventions, our technical success rate has exceeded 97%. Procedurerelated pneumothoraces have proved extremely rare, to the point that routine chest radiographs after sonographically guided venipunctures are no longer obtained. While serious complications have been reported with catheter placement, the incidence is typically less than 1% (table 1).

Vascular thrombosis

For many years, both temporary and long-term HD catheters have preferentially been placed into the subclavian veins. Prospective venographic studies have since documented catheter-related subclavian vein stenoses and thromboses in 42 to 50% of patients. ^42-46 Left-sided catheters also are more prone to cause problems, perhaps due to constant motion of the catheter against the wall of the inominate vein. Venous obstructions can develop remarkably quickly. In a study of 52 patients with temporary subclavian and internal jugular HD catheters, Cimochowski et al found that 50% of subclavian catheter patients developed marked venous stenoses (mean catheter dwell time of 11.5 days) compared with internal jugular catheter patients (no incidence; mean catheter dwell time 15 days). ⁴²

Agraharkar et al sonographically evaluated 96 patients who underwent 144 separate internal jugular vein HD catheter placements (116 veins). Only two jugular thromboses related to percutaneous catheter placement were found (1.7%), compared with five (4.3%) found in patients who underwent surgical cannulation. ⁸ A catheter tip positioned within the cephalic portion of the superior vena cava has also been linked to a higher risk of central venous thrombosis (figure 5). ^41,47

FIGURE 5. Superior vena cava syndrome caused by the tip of a dialysis catheter positioned within the distal superior vena cava (arrow). The cavagram, performed from a right internal jugular approach, demonstrates retrograde flow within the azygous (open arrow) and hemiazygous (curved arrow) veins.

Subclavian vein stenosis may lead to transient ipsilateral arm swelling until venous collaterals sufficiently enlarge. In non-HD patients, the thrombosis may prove of little clinical importance. Yet, in HD patients it can render that arm useless for surgical creation of a permanent arteriovenous access because of the resultant venous hypertension. While central lesions can often be treated with thrombolysis, angioplasty, and stent placement, most surgeons choose not to create an arteriovenous fistula or graft peripheral to a previously stenotic or thrombosed subclavian or inominate vein which has been angioplastied or stented. However, we have recanalized and stented occluded brachiocephalic veins in order to provide a route for HD catheter placement. In general, it must be emphasized that subclavian catheters should not be placed in patients who have or may need permanent upper extremity HD access.

Use of the right internal jugular vein (IJV) obviates many of the risks associated with subclavian vein access. The right IJV approach also makes placement easier because the large peelaway sheaths through which the catheters are advanced are less likely to kink, given the straight line of access. Accordingly, right IJV catheter approaches should be the primary choice for HD catheter placement. If possible, occluded IJVs should be recanalized in order to avoid use of secondary access sites (figure 6).

FIGURE 6. A 55-year-old woman with failed extremity accesses now requires permanent dialysis catheters. Both left and right catheter-related inominate and subclavian occlusions developed, requiring venous stent placement. (A) A right arm venogram. The right subclavian and inominate vein stent is occluded (arrow). Collateral veins supply the superior vena cava. The left inominate vein stent was also occluded (not shown). (B) After long-term failure of multiple transhepatic dialysis catheters, transhepatic access was used to assist in recanalization of the occluded right inominate vein stent. The stent was punctured directly, and a guidewire was threaded through the occlusion, snared, and exteriorized through the transhepatic approach (arrows). (C) An angioplasty balloon was threaded through the mesh of the stent and used to create a channel within the occluded stents. (D) Finally, a tunneled hemodialysis catheter was inserted through the stent mesh. (Images courtesy of Michael C. Soulen, MD, Hospital of the University of Pennsylvania, Philadelphia, PA.)

When IJV access is no longer feasible, translumbar catheters provide a viable alternative route. ^48-51 Although common femoral vein catheter placement is straightforward, femoral catheters possess a higher rate of infection, a risk of pelvic vein thrombosis, and the theoretic possibility of excessive mechanical stress upon a catheter due to motion at the hip joint. ^37,52,53 Despite these caveats, in our opinion, alternate approaches should be exhausted before resorting to subclavian catheter placement. In patients with extensive central venous and caval thrombosis, successful transhepatic and intercostal collateral vein placement of catheters has proven useful (figure 7). ^54-56

FIGURE 7. A transhepatic dialysis catheter. The leading tip of the catheter has been directed into the caudal inferior vena cava. In this case, thrombus has formed around the catheter tip (arrow); the clot was removed using a combination of lytic and mechanical means (not shown).

Catheter patency

Poor catheter flow may be caused by catheter malpositioning, migration, thrombosis, or fibrin sheath formation. Fibrin sheath formation around indwelling catheters is a common biologic response to all existing venous catheters. ^57-60 In a 1971 autopsy study, Hoshal et al found that fibrin enveloped most catheters within 5 to 7 days of insertion. ⁶¹ In that series, fibrin deposition began at the venous entry site and at the leading end of the catheter where the tip contacted the vein wall. Fibrin deposition spreads onto the entire catheter from these two sites of intimal injury. This process, in large part, accounts for non-infection-related loss of catheter function. The angiographic signs of fibrin sheath formation include a filling defect around the catheter tip, occlusion of contrast flow through the end hole of the catheter, or backflow of contrast along the side of the catheter within the sheath (figure 3). ³⁷ Many companies have touted improved catheter function and durability based upon their unique designs (figure 8). As an example, the Tesio Twin Cath (Medcomp, Harleysville, PA) is said to provide a longer interval to failure because each of its single lumen catheters has multiple side holes and a larger effective lumen. Intuitively, larger bore catheters will deliver higher initial flow rates, allowing shorter HD sessions. Whether such catheters provide durable high flow must still be addressed.

FIGURE 8. Several paired lumina or split-tip design catheters have been touted as providing higher dialysis flow rates.

In fact, it is difficult to compare patencies of different catheter designs because of widely varying definitions of loss of catheter function. In some series this is simply not defined, while in others it is defined as catheter occlusion or inability to sustain HD flow rates above 200 ml/min. ^{11,13-15,18-20,22,25,31} In other reports, loss of patency was defined as catheters that did not respond to urokinase therapy and required removal. These reporting differences account, in large part, for the wide range reported (1.4% to 50%) of thrombotic complications. ^{12,18,21,24,26}

Ideally, catheter function is best described using a clear definition of function and life-table or Kaplan-Meier analysis. Using the aforementioned definition of low flow rates and/or occlusion, Lund et al reported an overall cumulative patency of 44% at 6 months and 25% at 12 months. ¹⁸ In our review of 114 tunneled HD catheters, 6-month patency was 55% and 12-month patency was 50%. In our series, loss of patency was defined as catheters from which blood could not be aspirated or injected through one or both ports, or catheters in which consistently adequate flow rates (greater than 250 ml/min) could not be achieved despite repeated use of transcatheter urokinase. ⁶² Trerotola et al randomly placed conventional 13.5-F silicone HD catheters or high flow splittip cathetersin 24 patients requiring tunneled access. These researchers found that both catheters delivered acceptable flow rates (within DOQI recommendations), although the split tip catheter did provide higher effective flow rates. ⁶³ These results emphasize the need for larger scale, prospective, comparative trials of different catheter designs, as well as research aimed at blunting the thrombotic response to all such indwelling devices.

Catheter infection

Infection is the second most common cause of catheter malfunction. Site infections can range from mild erythema to purulent exit site and tunnel infections. Staphylococcus aureus and Staphylococcus epidermidis typically cause these infections. Bacteremia and septicemia also can be caused by these pathogens, although more virulent gramnegative rods (e.g., Enterobacter , Proteus , Pseudomonas , Escheria coli , and Serratia species) often are implicated. A number of techniques can be used to decrease catheter related infections:

1) Use of surgical hand scrubs, hats, and masks worn during insertion of catheters, as well as use of barriers such as sterile drapes, gowns, and gloves.

2) Skin preparation with chlorohexidine at the time of catheter insertion. This has proven more effective than povidone iodine or alcohol in the prevention of infection during insertion of percutaneous intravascular devices, although additional studies are recommended for verification. ⁶⁴

3) The use of antiseptic or silver ion-impregnated cuffs. This may reduce risks of catheter infections by inhibiting migration of bacteria into the puncture site. ^65-68

4) Coating catheters with heparin. This has been found to reduce the incidence of catheter infections in some studies. ⁶⁹

5) Coating catheters with antimicrobial agents. This has reduced the rate of catheter-related infections. However, the role of such catheters in HD patients is unclear, as their bacteriocidal activity is measured in weeks. ³⁵

6) Use of prophylactic antibiotics before catheter placement has not been shown to decrease the rate of catheter-related infection. ^70-72

7) The use of permeable dressings such as gauze and tape dressings has a lower risk of insertion site colonization and, therefore, catheter related infections when compared with transparent, relatively impermeable adhesive dressings. ⁷³ Impermeable transparent dressings can actually increase the rate of catheter infection by creating a warm, moist, contained environment that promotes growth of microbial organisms on the skin. ⁷⁴

Other risk factors for catheter infection include fever or bacteremia at the time of insertion, and the age of the patient (children have a higher rate of infection when compared to adults). ^75,76 HIV disease has not, in and of itself, been associated with an increased risk of infection. ⁷⁷

In addition, the duration of catheterization has been strongly linked to increasing the risk of infection. In a review of 786 non-tunneled HD catheter placements, Vanholder et al found a significantly increased incidence of bacteremia in catheters left in place for more than 10 days. ³³ The bacteremia rate in the "chronic" group was 7.2% (mean catheter dwell time 23 days) versus 3.4% in short-term patients (mean catheter dwell time 6.1 days). Infection rates with tunneled catheters appear to be lower than their non-tunneled equivalents, although a risk remains, in part, related to the duration of catheter placement. Schwab et al reported one episode of bacteremia in a prospective study of 80 PermCaths (Quinton, Bothell, WA) (median catheter use 8 weeks, range 3 to 22 weeks). Twenty-three site infections were treated successfully with antibiotics. ²⁴ Moss et al reported one episode of bacteremia in 54 patients (0.49 episodes per 100 patient-months), and nine site infections (5.3 episodes per 100 patient-months). ²¹ A study-to-study comparison of infection rates is difficult due to inconsistencies in data reporting. Therefore, catheter infection rates are probably best described in a manner that reflects the duration of catheter placement. Lund et al reviewed several recent studies with this approach, calculating infection rates using life-table analysis, number of events per 100 days at risk, and time to first catheter infection (table 3). In their study, there were 32 episodes of infection (18.5%), seven of which were treated conservatively with antibiotics. The infection rate was 0.20 per 100 days at risk. The probability of infection-free survival was 0.71 and 0.56 at 6 and 12 months, respectively. Notably, only one site infection developed within 2 weeks of catheter placement. Our infection rate in placing catheters under similar conditions is 15%, with a frequency of 0.15 per 100 days at risk. ⁶² The above data strongly support the contentionthat HD catheters can be safely placed under sterile conditions outside of the operating room without an increase in catheter infection rates.

Strategies for maintaining catheter function patency/thrombosis

The initial accumulation of thrombus or fibrin around the catheter tip of the arterial (red) port impairs the ability to withdraw blood at a rate sufficient to allow adequate HD. However, this process may not yet affect the distal venous (blue) port, so that reversal of the port connections to the dialyzer may allow the treatment to proceed. This provides a temporizing effect and may still yield unsatisfactory HD, as higher recirculation may occur (figure 9).

FIGURE 9. Venographic signs of fibrin sheath formation. This patient had undergone prior left inominate vein stent placement for treatment of an elastic, flow-limiting stenosis (arrow). (A) Contrast injection into the venous (distal) port demonstrates a subtle filling defect around the catheter tip (arrow). (B) Contrast injection into the arterial (proximal) port demonstrates a clear-cut filling defect around the catheter tip (arrow). There is retrograde filling of the azygous vein (curved arrow), reflecting the partial occlusion of the superior vena cava.

Initial medical management of failing catheters is instillation of a thrombolytic agent into the catheter lumina. This has proven effective at restoring function in most cases. A number of different regimens can be used, ranging from a single per port infusion of urokinase (Open-Cath, Abbott Laboratories, Abbott Park, IL, 5000 IU/cc) to high dose infusions of higher concentrations of urokinase (50,000 IU/cc). ^78-80 The Open-Cath solution typically is allowed to dwell for 20 to 30 minutes, with repeated periodic attempts at catheter aspiration. This process can be repeated with longer dwell times (30 to 60 minutes) before HD is reattempted. Six- to 12-hour infusions of urokinase also have been successful. ⁸¹ Low dose warfarin (1 mg/day) appears to reduce the incidence of catheter related thrombosis as well. ^82,83 Additionally, in a study of bone marrow transplant patients, a reduction of catheter thrombosis from 20% to 3.2% was found in patients receiving 325 mg aspirin perday. ⁸⁴

Earlier this year, urokinase was removed from the U.S. market by the Food and Drug Administration (FDA). While there is currently no specific recombinant tissue plasminogen activator (rt-PA) formulation for catheter clearance, it is expected that one will become available in the near future (Genentech, Personal communication, May 1999).

When thrombolysis fails to restore patency, standard therapy consists of catheter removal and insertion of a new catheter at a different access site. The disadvantage of this approach is the surrendering of a venous access site in patients with limited potential access sites. Crain et al reported success using percutaneous fibrin sheath stripping to prolong the functional patency of failing catheters (blood flow rates <200 mL/min). ^85,86 In this technique, a nitinol loop snare was used to tightly encircle the catheter tip and strip the encasing sheath from the catheter. Forty procedures were performed on 24 catheters in 23 patients. Immediate technical success was 100%, with a calculated median added patency of 2.8 months. The procedure was repeated in several patients to restore and prolong catheter function.

In contrast, our experience led us to abandon the above technique because it appeared to provide no durable benefit in improving catheter function. ⁸⁷ Twenty-four procedures were performed on 22 patients with occluded or failing catheters (blood flow rates <250 mL/min). Although occluded catheters were reopened and failing catheters demonstrated marked improvement in flow rates, these changes disappeared within five HD treatments. ⁸⁷

Currently, we manage failing catheters by exchanging existing catheters over guidewires for new catheters through the existing subcutaneous tunnels. This is rapidly performed on an out-patient basis using local anesthesia (lidocaine with epinephrine) at the catheter entry site and along the subcutaneous tract. The catheter cuff is dissected free using a hemostat introduced at the skin exit site. A single or pair of hydrophilic guidewires is introduced through the catheter (one wire per lumen) and used to exchange the old catheter for the new catheter. In many cases, the new catheter is positioned slightly deeper within the atrium by using a longer catheter or pushing the new catheter deeper into the tract.

In a retrospective comparison of 74 de novo catheter placements and 40 catheter exchanges, catheter patency was equivalent: 67% and 60% at 3 months, 55% and 50% at 6 months, and 40% and 41% at 12 months. ⁶² Notably, no significant difference between infection rates was found in the two groups. We have since used this technique in over 180 cases with similar results. In a minority of cases, we have introduced a vascular sheath over the exchange guidewire and used an angioplasty balloon to disrupt the cavoatrial fibrin sleeve prior to catheter replacement. It is unclear whether this provides any additional benefit over simple catheter exchange, a process that in and of itself typically disrupts the fibrin sheath.

Occasionally, there can be excessive oozing at the puncture site after catheter exchanges. Numerous hematologic abnormalities accompany renal dysfunction, including increased bleeding time, defective platelet aggregation, adhesiveness, and factor 3 availability. While the resulting coagulopathy is best treated with hemodialysis, intravenous desmopressin acetate (DDAVP) infusion can provide a rapid and effective treatment for temporary reversal of uremic hemorrhage in patients undergoing catheter placement or other invasive procedures. ^88,89

Catheter infection

Management of catheter infections continues to evolve. Typically, mild site infections can be treated with antibiotic therapy without catheter removal. Grossly purulent exit site infections, tunnel infections, and bacteremic or septic patients should be treated with catheter removal and antibiotics. Catheter tips should be sent for bacterial and yeast cultures upon removal.

Our practice has been to delay placing a tunneled access at a new site in a patient with bacteremia for several days after initiation of antibiotic therapy and until blood cultures no longer indicate the presence of circulating pathogens. In the interim, we place a temporary jugular hemodialysis catheter, or the patient undergoes one or two treatments with transient femoral catheterization, performed by a nephrologist. There have, however, been several reports of successful eradication of septicemia complicating tunneled HD and non-HD catheters with antibiotics and catheter exchange over a guidewire. ^90-93 Carlisle et al treated 21 episodes of sepsis in HD patients with this approach. ⁹³ Most infections were due to staphylococcal species and gramnegative rods. Shaffer reported similar results in 13 consecutive cases of sepsis. ⁹² In both series, catheters were not replaced in situ in patients with frankly purulent site infections. This somewhat controversial approach has the advantage of preserving an important access site, limiting the cost and number of procedures, and limiting hospital days required for management.

Conclusion

In summary, the role of interventional radiologists in the management of dialysis patients continues to expand, encompassing both placement and maintenance of many forms of venous access. It is clear that we have both the abilities and sophisticated techniques to percutaneously place all forms of HD catheters safely and to maintain and preserve their function. The effective use of these out-patient techniques can provide more rapid patient therapy and lower hospital costs compared with standard operative approaches.     AR