Worldwide the population of patients with chronic kidney disease (CKD) is expanding, leading to increasing numbers of patients hitting end-stage renal failure (ESRF) and requiring renal replacement therapy (RRT). In the UK the incidence of new dialysis starters increased from 115 per million population (pmp) in 2014 to 120 pmp in 2015, resulting in 7814 new patients initiating RRT. The median age of these new dialysis starters was 64.4 years, highlighting the fact that a significant proportion of the UK dialysis population are elderly patients, many of whom have multiple comorbid conditions, such as diabetes, hypertension, ischemic heart disease, and obesity. There has been a paradigm shift in recent years and the leading cause of kidney disease is now diabetes, which accounts for 30% of all cases. This is important to recognize, because many of these patients are high-risk surgical candidates and are particularly challenging with regard to creating dialysis access because of small, diseased vessels, a paucity of good-caliber veins, and higher infection rates.
Long-term access for RRT through the creation of an arteriovenous fistula (AVF), placement of an arteriovenous graft (AVG), or peritoneal dialysis (PD) catheter requires the input of multiple dedicated access specialists. Over the past decade this has expanded to include nephrologists, surgeons, radiologists, and nurse specialists who work closely as a multidisciplinary team to provide timely access creation and ongoing access maintenance for each individual renal patient. There is a growing recognition that the approach to access creation should be increasingly bespoke for each individual patient. The historical tradition of starting dialysis via a tunneled dialysis catheter while waiting for a wrist fistula to mature has evolved to a “fistula first,” “line last” strategy. This is employed with the aim of preserving venous real estate and ensuring patients start dialysis using an access that is patient-specific based on comorbidities and needs.
Technologic advancements of products specifically for dialysis access have provided an array of options for patients on RRT, particularly when access options have been exhausted. The improved long-term outcomes with newly developed grafts along with increasing armamentarium such as drug-eluting balloons and stents for the maintenance of access or technologies to recanalize occluded central veins are providing renal failure patients with ongoing reliable access and improved life expectancy rates on dialysis. This chapter will explore much of the current modern-day practices associated with establishing and maintaining RRT for patients with ESRF.
Vascular Access Catheters
Temporary Dialysis Catheters
Central venous catheters (CVCs) are an option in patients with renal failure to provide dialysis. Catheters are either cuffed or uncuffed. Uncuffed CVCs are used as a temporizing measure in patients who require immediate or emergent hemodialysis. About 40% of patients with renal failure present acutely and require short-term vascular access and dialysis. These patients include the “acute presenter” and the “crash lander.” Acute presenters require short-term RRT to allow renal recovery, and crash landers, those who present suddenly with new ESRF, require short-term access during planning for a long-term option. In both instances a temporary uncuffed CVC may be placed to facilitate dialysis. Temporary catheters can be inserted at the bedside and can remain in situ for up to 3 weeks depending on the site of insertion.
The three most common locations for temporary vascular access are the internal jugular vein (IJV), subclavian vein (SCV), and common femoral vein (CFV). Upper limb catheters can remain in situ for up to 3 weeks, patients can remain fully ambulatory, infection rates are lowest, and they are therefore preferred to groin CVCs. Currently the favored site for temporary venous access is the right IJV, although some patients find the visibility of the catheter in the neck above the collar unsightly. The advantages of IJV placement are lower risk of infection, higher patency rates, and lower risk of insertion-related complications.
Permanent Dialysis Catheters
Patients requiring long-term dialysis and in whom there has not been an opportunity to create a fistula, AVG, or peritoneal dialysis catheter in a timely fashion will require a permanent dialysis catheter. In the US the incidence of patients starting dialysis on a tunneled catheter remains high at around 70%. The UK incidence is not much better with 48% of new starters initiating hemodialysis via a CVC. The current indications for central venous catheter insertion are displayed in Box 5.1 . Permanent catheters are always tunneled, dual-lumen, large-diameter (14 to 16 French) polyurethane catheters that have an annular fibrous cuff that holds that catheter in place within the tunnel tract. CVCs are composed of biologically neutral material that should not induce catheter lumen thrombosis or a perivascular reaction and subsequent venous thrombosis. The catheter should be soft and compliant, easy to insert, durable, and should be coated with an agent that reduces bacterial proliferation and biofilm formation. Furthermore it should be inexpensive and permit blood flow of >350 mL/min to facilitate efficient dialysis. Numerous catheters are available for use, and they can be differentiated by the design of the distal tip and the presence of side holes and flow dividers. The hemodynamic effects of these design variations are regularly debated and remain largely unknown and are influenced by numerous factors such as insertion technique, site of insertion, and position of the catheter tip. Although interesting, catheter design and insertion principles are less important than providing efficient dialysis. This includes the ability of the catheter to be able to process at least 50 L of fluid during a standard 4-hour dialysis session with favorable arterial and venous pump pressures and minimal recirculation.
During maturation of autogenous arteriovenous fistula
During maturation of peritoneal dialysis catheter
Patients awaiting a living donor transplant
Dialysis bridge after failure of current access to permit planning and imaging for long-term access
Permanent access—all other sites exhausted, severe cardiac dysfunction or patient choice.
The right IJV is the most favored site of placement of a permanent catheter, which is then tunneled to an exit site on the anterior chest wall. Care should always be taken to ensure that the cuff is at least 4 cm from the exit site to minimize the risk of cuff extrusion and minimize infection rates. Internal jugular vein catheter placement can be performed surgically using a cut down at the medial border of sternocleidomastoid. However, the Seldinger technique using ultrasound to guide the needle into the vessel is more frequently performed. It is essential to place catheters under fluoroscopic control, to ensure that the tip of the catheter is placed in the superior vena cava (SVC). The malposition rate of catheters when placed without fluoroscopic control has been shown to be as high as 29%. The tip of the catheter should be at the junction of the SVC and right atrium as this affords the most optimal blood flow through the catheter. A list of complications associated with central venous catheter insertion is displayed in Box 5.2 .
Tunneled catheters can provide access for months and even years, particularly in those patients in whom all native venous conduits for AVF or AVG formation have been exhausted or where arteriovenous strategies are likely to induce risks of limb loss from steal syndrome or precipitating heart failure from the high volume venous return. Indeed, in such patients who become dependent on a tunneled catheter, placement is becoming more adventurous with reports of transhepatic, translumbar and even transmediastinal approaches being used as last resort procedures.
Complications of Hemodialysis Catheters
Catheter dysfunction was historically defined as “failure to attain and maintain adequate extracorporeal blood flow sufficient to perform dialysis in a timely fashion.” A recent collaboration of transplant surgeons and physicians, the North American Vascular Access Consortium (NAVAC), introduced a new definition in an attempt to create a uniform standard. Dysfunction was defined as the first occurrence of either (1) peak blood flow of 200 mL ⁄ min or less for 30 minutes during a dialysis treatment, (2) mean blood flow of 250 mL ⁄ min or less during two consecutive dialysis treatments, or (3) inability to initiate dialysis owing to inadequate blood flow, after attempts to restore patency have been attempted.
Dysfunction accounts for 17% to 33% of all catheter removals and can be either early or late. Early dysfunction is due to technical error such as kinking of the catheter in the subcutaneous tunnel or catheter malpositioning. Late dysfunction is due to central vein occlusion, catheter thrombosis, and fibrin sheath formation ( Fig. 5.1 ). Catheter thrombosis has an estimated frequency of 0.5 to 3 episodes per 1000 days and an incidence of 46% and is the major cause of catheter dysfunction. Antifibrinolytic therapy introduced via the catheter is the treatment of choice and first-line therapy is usually 5000 IU/mL of urokinase of sufficient volume to fill the lumen. This can be repeated immediately and if bolus dosing fails can be followed by a systemic infusion of 20,000 IU/mL/hr over 6 hours or a continuous infusion during dialysis of 250,000 IU. An RCT of tenecteplase versus placebo in patients with catheter dysfunction demonstrated that patients treated with tenecteplase had significantly increased flow rates (>300 mL/min). Patients within the study with <1 day of catheter dysfunction had the greatest benefit with 35% of tenecteplase-treated patients gaining adequate dialysis flow rates compared with 8% in the placebo group. If thrombolysis fails then imaging of the catheter to obtain further information is indicated. Direct injection of contrast via the arterial port can help elucidate the presence of a fibrin sheath. Treatment of a fibrin sheath usually involves replacement of the catheter at a new location although there is some evidence to support mechanical stripping, which is carried out using a snare around the catheter.
Catheter-Related Central Vein Stenosis
The long-term presence of any device within the central venous system is associated with the development of venous stenosis. The exact mechanism by which this happens is unclear; however, catheter-related thrombosis or infection along with progressive trauma to the endothelium is the most likely cause. The incidence of CVC-induced stenosis is reportedly as high as 50% although there are values as low as 18% in the literature. The rate of stenosis after subclavian vein cannulation is higher than that seen when the catheter is placed in the IJV, and catheters placed on the left side of the neck have a higher reported incidence of associated stenosis compared with those placed on the right. Fig. 5.2 shows an example of a right arm venogram in a patient with a central venous stenosis as a result of a CVC.
Central Vein Occlusion
The presence of occlusive thrombus in the central veins, in particular the SVC, is asymptomatic in about 30% of cases. Although it can manifest as catheter dysfunction, the presence of arm and facial edema or prominent superficial vessels should raise suspicion of central venous occlusion or stenosis. Thrombus is usually detected by angiography; however, transesophageal Doppler and contrast-enhanced cross sectional imaging are also useful diagnostic tools. The treatment of central vein thrombus depends on the chronicity of the thrombus. Acute thrombus may respond to fibrinolytic therapy. However, fibrinolytic therapy will not have any effect on older, organized thrombus. Percutaneous angioplasty and stenting have an increasing role in maintaining central venous catheter patency and recanalization of stenosed or occluded vessels. In addition, the advent of novel technology such as the Surfacer Inside-Out device (Merit Medical) has provided a potential means to persist with upper limb access in the presence of SVC occlusion. The details of the device are covered in the Advances in Access section later in the chapter.
Septic complications from CVCs are well documented, and the longevity of modern cuffed and tunneled catheters is still limited by infection. Catheter-related bacteremia rates for cuffed, tunneled catheters ranges from 1.5 to 5.5 per 1000 days. Catheter-related infections include exit site infection, tunnel infections, and catheter-associated bacteremia. Sepsis is the most significant cause of catheter-associated morbidity and mortality ( Fig. 5.3 ). Catheter infection is responsible for between 6% and 28% of catheter failure. Gram-positive skin dwelling bacteria, in particular Staphylococcus species, account for 40% to 80% of infections with gram-negative organisms accounting for 20% to 40%. Polymicrobial (10%) and rare causes such as fungal infection (<5%) make up the remainder. The route of infection is twofold, either tracking along the external surface of the catheter or via the lumen. Exit-site infections present with localized erythema with or without discharge and often respond to topical or oral antibiotics. Unresponsive infections or those with discharge and no signs of systemic sepsis and no bacterial growth on blood culture should be treated with parenteral antibiotic therapy. The Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines recommend treatment for 3 weeks. However, if the patient is hemodynamically unstable or fails to improve after 36 hours of parenteral antibiotics, with bactericidal serum levels of an appropriate antibiotic, the catheter should be removed. Serious complications from catheter-associated infections occur in 3% to 44% of cases and include endocarditis, osteomyelitis, thrombophlebitis, septic arthritis, spinal epidural abscess, and large atrial thrombi. Mortality from catheter-associated infections is greatest when Staphylococcus aureus is the culprit with a reported rate of up to 30% in some units. Methicillin-resistant Staphylococcus aureus (MRSA) has three to five times higher mortality compared with methicillin-sensitive strains and is considerably more costly to treat and manage.
Fistulas and Synthetic Grafts
In the past decade there has been a concerted effort to improve the outcomes for patients requiring arteriovenous access for dialysis. This has led to an expansion in the volume and quality of evidence in vascular access. There has been an increase in the number and variety of randomized controlled trials (RCTs), prospective cohort studies, and guidelines to inform practice. This has coincided with the concept of a personalized access plan for each patient and has challenged clinicians to consider all possible options, including AVGs as an initial access procedure. “Fistula first” should still be considered the foundation for the majority of patients; however, decision making on access should be on an individual basis.
AVFs using native veins are associated with the highest long-term patency and lowest complications rates, although they suffer from high initial failure rates. In patients with enough time for fistula planning, creation, and maturation, they should be the vascular access of choice for patients requiring long-term hemodialysis. Synthetic grafts have lower rates of primary failure but can be used for dialysis earlier; indeed many synthetic grafts can be needled immediately. However, the complication and intervention rates are higher, and the use of an AVG may reduce the amount of native vein available for future use. Furthermore the National Kidney Foundation-Kidney Disease Outcomes and Quality Initiatives (NKF-KDOQI) guidelines state that all suitable patients should have a native AVF for dialysis access.
Planning and Timing of Vascular Access
Construction of an AVF involves the conversion of an accessible vein to a high-flow vessel from which blood can be rapidly removed and returned via two needles during a dialysis session. Veins are low-pressured, low-resistance capacitance vessels. Their relative lack of elastic recoil compared with that seen in arteries allows them to progressively increase in size if high-pressure flow passes through them. After AVF formation, high-pressure arterial flow passes into the low-pressure capacitance vein. The subsequent increased turbulent flow gives rise to a palpable thrill and an audible bruit within the AVF. In addition, the fistula vein increases in size over time. It is this second feature that is crucial for needling and subsequent hemodialysis to take place. As a general rule, the fistula vein needs to be a minimum of 6 mm diameter 6 weeks after creation of the fistula if needling is to be attempted and hemodialysis successful.
Given that there is a minimum of 6 weeks maturation time of a fistula from the point of its creation, the timing of fistula creation in relation to the predicted start of dialysis is critical. Permanent vascular access in the form of a fistula should be created well in advance to allow not only for maturation but also for further procedures should there be failure of the initial access procedure. Many guidelines exist regarding the timing of vascular access. In the US the NKF-KDOQI guidelines recommend:
Patients with glomerular filtration rate (GFR) <30 mL/min should be educated on all modalities of RRT including transplantation.
A fistula should be placed at least 6 months before the anticipated start of dialysis.
A new AVF should be allowed to mature for at least 1 month and preferably 3 to 4 months before cannulation.
Prosthetic grafts should be placed 3 to 6 weeks before anticipated need for hemodialysis.
Never insert hemodialysis catheters until actually needed.
In the UK the Renal Association and the joint working party of the Renal Association, the Vascular Society of Great Britain and Ireland, and the British Society of Interventional Radiology have produced the following standards :
65% of all patients presenting within 3 months of dialysis should start hemodialysis on a useable AVF.
85% of all prevalent hemodialysis patients should be dialyzed using a native AVF.
No patient requiring dialysis should wait more than 4 weeks for fistula construction.
Permanent access should be constructed 16 weeks and preferably 6 months before anticipated need for dialysis.
Preservation of arm veins is critical and unnecessary venipuncture must be avoided.
The Vascular Access Society recommendations are similar, and as with those of NKF-KDOQI and the Renal Association, state that arm veins must be preserved and that the planning of access should commence when a patient’s GFR reaches 20 to 25 mL/min. Patients should see an access surgeon when the GFR is 15 mL/min so that access can be created 6 to 12 months in advance of dialysis (or even sooner if progression is rapid).
Emerging data suggest that AVF outcomes in terms of patency are significantly improved if access is constructed earlier rather than when the patient is on the brink of dialysis. It is unclear why this may be but it has been suggested that worsening uremia affects vascular biology especially at the level of the AVF anastomosis.
The historical focus in vascular access has been AVF location-specific, rather than patient-specific. Vascular access has become much more patient centered and bespoke even if that conflicts the view of wrist fistula first. However, the importance of carefully planning access procedures to protect venous real estate still remains. There are a few general principles that are important in the planning of vascular access. The upper limb should be used in preference to the lower limb and distal sites should be used before proximal sites. This permits the maximum use from a particular vessel and preserves proximal sites for future use. It is also preferential to use the nondominant arm ahead of the dominant arm, and this is particularly important in patients who needle their own fistula at home. The formation of preemptive AVF requires liaison between nephrologist and surgeon and flexibility in theater list planning to ensure patients with rapidly deteriorating renal function can be accommodated before reaching end-stage disease. However, patients who are referred late, who are already undergoing RRT, or who present acutely with an emergent need to dialyze, should have planning of vascular access adapted from the wrist fistula first model, to suit the patient. Such examples include the use of early cannulation AVG (ecAVG), which can be needled immediately and used in preference to a tunneled dialysis catheter. A recent RCT randomized 121 patients who needed dialysis within 48 hours to ecAVG or tunelled central venous catheter (TCVC). They demonstrated a reduction in bacteremia rates at 3 months in the ecAVG group and lower mortality. The higher upfront cost of ecAVG was offset at 6 months by the increase in complications of the TCVC, and those in the ecAVG had a shorter hospital stay and reduction in use of temporary CVCs.
The three requirements for a successful AVF are as follows:
Good arterial supply—inflow
Patent central veins—outflow
Adequacy of the inflow to the arm can usually be assessed clinically with palpation of the brachial, ulnar, and radial pulses and Allen’s test can be used to determine patency of the palmar arch and the dominance of radial or ulnar artery in supplying the hand. Suitability of arm veins can usually be determined clinically using a tourniquet placed proximally on the arm. However, vein diameter and patency can be confirmed by duplex ultrasound, which is cheap, noninvasive, and improves patency rates in those patients with veins that are difficult to assess clinically. It is generally accepted that to stand a good chance of success the vein needs to be greater than 2 to 2.5 mm in diameter although the evidence is conflicting. Assessment of the basilic vein is usually performed using duplex ultrasound. In patients with previous CVCs or suspected central venous pathology, venography should be performed before surgery.
The majority of upper limb AVF can be performed under local anesthetic (LA), in particular wrist and elbow fistulas. For more extensive procedures such as transposition of the basilic vein and forearm loop grafts, a regional block with local infiltration may suffice. Regional anesthesia (RA) has the added advantage of blocking sympathetic nerves as well as sensory nerves, which causes vasodilation. In prolonged operations or if the patient is intolerant of the procedure, the anesthetist can administer a short-acting sedative to avoid a general anesthetic (GA). This is advantageous because many patients with renal failure have significant cardiovascular disease and other comorbid conditions making a GA high risk. Surgical procedures in the axilla and lower limb usually require GA. Comparison of LA and RA on patency rates of AVF has been the subject of a recent RCT and meta-analysis. An observer blinded RCT of 126 patients allocated to brachial plexus block (BPB) or LA demonstrated greater patency in the BPB group at 3 months. However, functional patency rates, which are a more clinically relevant outcome, did not differ between the groups. A meta-analysis of four RCTs comparing AVF failure rates in those performed using RA with LA concluded that failure rates were lower in the RA group. Follow-up in the studies varied from 40 to 100 days, and the definition of failure was slightly different in each paper; however, a sensitivity analysis still demonstrated that failure rates were lower in RA. There are no longer-term data on patency rates of AVFs formed under RA or LA, which would be interesting, because this information would better inform future practice.
Autogenous Arteriovenous Fistulas
The initial procedure of choice, in patients with suitable vessels, is a radiocephalic AVF at the wrist. The operation is performed under LA as a day case procedure. The original Brescia-Cimino AVF involved a side-to-side radial artery to cephalic vein anastomosis although more recently an end of cephalic vein to side of radial artery has come into favor. The original Brescia-Cimino AVF often led to venous hypertension in the hand, whereas the end-to-side variation does not. The disadvantages of radiocephalic AVFs are the high primary failure rate and failure to mature. Mean primary patency rates at 1 year are 55% (46%–63%) and secondary patency rates are 71% (65%–77%). At 1 year, 23% (17%–29%) of fistulas will have been abandoned without use.
The surgical technique for formation of radiocephalic fistulas follows. The cephalic vein and radial artery are exposed via a longitudinal, oblique, or S -shaped incision depending on the proximity of the vessels and the preference of the surgeon. The cephalic vein is mobilized for 3 cm from beneath the lateral skin flap ensuring preservation of the sensory dorsal branch of the radial nerve. The artery is located lateral to the tendon of the flexor carpi radialis and lies underneath the fibers of the deep fascia, which must be divided. Between 2 to 3 cm of artery should be mobilized and branches of the vessel can be ligated and divided. The vein and artery are controlled proximally and distally with vascular slings and microvascular clamps. In an end-to-side anastomosis the cephalic vein is ligated distally and divided obliquely to leave a spatulated end for anastomosis. An arteriotomy is then performed on the anterolateral surface of the radial artery. In a side-to-side anastomosis an arteriotomy (1–1.5 cm) is performed on the lateral aspect of the vessel, and afterwards a venotomy of equal length is performed on the medial side of the cephalic vein. A 6-0 or 7-0 nonabsorbable, monofilament suture with two needles is then used to perform a continuous anastomosis. Once the anastomosis has been completed and in the presence of an acceptable thrill within the vein, the distal cephalic artery can be ligated and divided. In a successful procedure there should be a thrill present once the clamps have been released and the slings loosened. In the absence of a thrill one must ensure the patient is not hypotensive, there are no adventitial bands constricting the vein, and there are no errors, such as an intimal flap, twisting of the vein, or presence of thrombus.
In the presence of failed radiocephalic fistulas or inadequate forearm vessels, an autogenous elbow fistula should be the next procedure of choice. A brachiocephalic AVF is usually the first procedure of choice at the elbow and is a straightforward procedure performed under LA.
Brachiocephalic Arteriovenous Fistulas
The original brachiocephalic AVF (BCAVF) was described as an end-to-side anastomosis of the cephalic vein to the brachial artery ( Fig. 5.4 ). As with radiocephalic AVF, if the elbow vessels are small then a side-to-side configuration can also be used. The venous anatomy in the cubital fossa is variable, and it is important to establish how the veins are related before deciding on which vessel to use; the cephalic vein at the elbow is often used for venipuncture and can be sclerosed and unsuitable. The median cubital vein often drains into the cephalic vein and can also be used for anastomosis to the brachial artery. The surgical technique for the brachiocephalic fistula is the same in principle as for the radiocephalic at the wrist. Patency rates of BCAVFs are similar to radiocephalic AVFs at 1 year, with mean primary patency of 52% (41%–61%) and mean secondary patency rates of 74% (63%–83%). They also have excellent longer-term patency of up to 70% at 3 years. Unlike with radiocephalic fistulas, there can be a significant increase in flow rates in BCAVFs, which can lead to high-output cardiac failure and steal syndrome. Such complications can be minimized by ensuring the arteriotomy does not exceed 75% of the diameter of the artery.
Brachiobasilic Arteriovenous Fistula
The basilic vein originates on the medial aspect of the forearm at the wrist and runs superficially in the forearm. It only remains superficial for a short distance in the arm before coursing beneath the deep fascia to run up the medial aspect of the arm alongside the medial cutaneous nerve of the forearm. Brachiobasilic fistula formation can be split into one-stage and two-stage procedures. A one-stage procedure is usually performed under general anesthesia and requires an extensive incision from the cubital fossa running longitudinally along the medial aspect of the arm toward the axilla. The basilic vein is mobilized from beneath the deep fascia and all tributaries are ligated and divided. Once an appropriate length has been exposed the vein is divided and placed superficial to the medial cutaneous nerve of the forearm. The vein can then either be transposed via a subcutaneous tunnel, or superficialized, where the lateral skin flap can be undermined and the vein placed in a suitable position for needling, above the deep fascia ( Fig. 5.5 ). The advantages of a one-stage procedure are that it only requires one operation and a single hospital stay, and the fistula can be used more quickly. The disadvantage is that if the fistula fails the patient has undergone a significant procedure, with a substantial incision, usually under general anesthesia. A two-stage procedure comprises a first stage, during which the basilic vein is anastomosed to the brachial artery under local anesthetic, through a small cubital fossa incision. The fistula is then assessed for maturation 4 to 6 weeks after formation, and if deemed adequate, the second stage can be performed. Innovative minimally invasive techniques have also been described using videoscopic assistance to mobilize the vein and ligate the tributaries although such techniques are not in widespread use. The mean 1-year primary patency rate is 55% (47%–63%) and mean secondary patency rate is 75% (67%–82%) with one randomized study yielding patency rates of 70% at 3 years. A recent systematic review and meta-analysis concluded that there is no statistical difference in outcomes between one- and two-stage procedures although individual studies tended to favor the two-stage procedure.
Historically, in the UK and Europe, prosthetic grafts were reserved for patients who had exhausted native venous access, and although that is still the most common indication, practice has changed. This is in an attempt to avoid the complications associated with the use of temporary “bridging” central venous catheters. The three categories of patients who require urgent access for dialysis in whom this is relevant are as follows:
Patients with an acute presentation of chronic renal failure
Patients who have lost their current hemodialysis access
Patients requiring a long-term modality switch from peritoneal dialysis to hemodialysis
The first group includes patients who require dialysis urgently and those who will need dialysis before an AVF can be performed and will mature. As a result there has been an increase in the use of prosthetic grafts, in particular, ecAVGs. Conventional AVGs are composed of expanded polytetrafluoroethylene (ePTFE), which is durable, easy to handle, and produces reproducible results. It requires about 14 days to become incorporated into the surrounding tissue before use, which limits its use as a rescue procedure. Early cannulation grafts such as Flixene (Maquet) or Acuseal (W.L. Gore) can be used within 24 hours of formation and can therefore avoid bridging dialysis catheters. This reduces the infection risk associated with catheters and minimizes damage to central veins thereby reducing risk of central venous stenosis.
Data on patency rates and complications in AVFs and AVGs generally favor AVFs. In a systematic review of more than 200 studies with 875,269 access events, 2-year pooled primary patency rate for AVFs was 0.55 (95% confidence interval [CI], 0.52–0.58) and secondary patency was 0.63 (95% CI, 0.59–0.67). In AVGs the values were inferior with primary patency of 0.40 (95% CI, 0.35–0.44) and secondary patency of 0.60 (95% CI, 0.55–0.65). In another review primary patency rates ranged from 0.18 to 0.70 at 1 year for AVGs compared with 0.53 to 0.74 for AVFs. Secondary patency rates at 2 and 3 years were 0.47 to 0.73 and 0.39 to 0.70 for AVGs compared with 0.26 to 0.94 and 0.64 to 0.84 for AVGs.
Complication rates are higher in AVGs than AVFs although AVFs had greater rates of stenosis (51.4% vs. 40.6%, P = 0.0182), whereas AVG had greater thrombosis rates (14.6% vs. 31.9%, P < 0.001) and overall increased risk of death compared with an AVF (relative risk = 1.18, 95% CI = 1.09–1.27). AVGs also have a higher rate of interventions to retain patency with data in the range of 1.7 to 3.5 per graft per year.
A prosthetic graft can be anastomosed to any suitable artery and vein, in an end-to-side disposition and then tunneled subcutaneously. It can be performed under LA but usually requires general anesthesia. Brachioaxillary grafts in a straight configuration are an excellent option in patients with unsuitable cephalic and basilic veins at the elbow. Grafts can also be configured as a loop and they are commonly located in the forearm, arm, and thigh ( Fig. 5.6 ). Forearm loop grafts are an excellent option in obese patients with deep upper arm veins because the procedure can be performed using RA or LA. It also preserves upper arm veins for future use. US data demonstrates that forearm loop grafts have similar patency rates to upper arm straight grafts. However, this may not be transferable to a European population because up to 90% of the patients had a graft as their initial access procedure. AVGs can be constructed in many anatomic locations once conventional access has been lost and such examples include forearm, arm, on the chest wall, in the thigh, and as a necklace—from one axilla to the other. More elaborate procedures have also been undertaken including brachiojugular grafts, and a straight graft has successfully been fashioned from the axillary artery to the external iliac vein and even the subclavian artery to the right atrial appendage. Once venous access is lost, arterioarterial grafts have been used with impressive success. A systematic review showed primary patency rates ranged from 67% to 94.5% at 6 months to 54% to 61% at 36 months, with secondary patency rates from 83% to 93% at 6 months to 72% to 87% at 36 months.
Complications of Arteriovenous Fistulas and Grafts
Bleeding can be categorized as early, which is within the first 24 hours, or late, which is anytime thereafter. Early bleeding may either be due to technical error within the anastomosis, slipping of a ligature, or due to generalized oozing as a result of uremic platelet dysfunction. Clinicians should have a low threshold to reexplore the fistula to resolve any technical cause of bleeding. Late hemorrhage from an AVF is usually from a needling site immediately after dialysis and often occurs in patients who are anticoagulated. Direct pressure on the site of venipuncture will usually stop the bleeding and reversal of anticoagulation is not usually necessary. Bleeding from aneurysms or at needling sites as a result of infection can be catastrophic and frightening for the patient. Although direct pressure may be sufficient to control hemorrhage, surgical exploration is often required and may result in ligation of the fistula, for which the patient must be consented. Exploration of a late bleed from a fistula should be performed under general anesthesia because an extensive incision is often required to gain proximal control of the vessels and it can be extremely unsettling for the patient.
The most common cause of AVF and AVG thrombosis is due to the presence of an underlying stenosis. In native AVFs stenosis can be present at any point from the arterial anastomosis to the central veins. The common locations are juxta-anastomotic, at the swing point, at needling sites and the cephalic arch (BCAVF). In AVGs the most common location is at the venous anastomosis. The pathophysiology of stenosis is as a result of neointimal hyperplasia, thought to be due to turbulent blood flow and shear stress, although a proportion of patients have intimal hyperplasia evident before AVF formation. There is conflicting evidence on the effect this has on AVF maturation. Early detection of stenosis increases the number of fistulas that mature and prolongs fistula patency. Fistula and graft surveillance is discussed in more detail later.
Immediate thrombosis in the presence of adequate quality and appropriate sized vessels may be due to technical error or a platelet plug and merits reexploration. Thrombectomy and refashioning of the anastomosis should salvage the fistula. Early thrombosis, that occurs after 24 hours but before the fistula maturing, may be as a result of patient factors such as hypotension, either as a result of fluid depletion after dialysis or cardiac failure, or it may be due to inadequate vessel size and/or quality. Attempted salvage of an early thrombosis is often unsuccessful and a pragmatic approach should be taken to avoid unnecessary and costly interventions that are often futile. Late thrombosis is usually due to the presence of a gradually progressive stenosis secondary to neointimal hyperplasia, and these account for about 85% of all stenoses. Treatment of fistula thrombosis can be radiologic or surgical and is dependent on local expertise. Radiologic intervention has the combined advantage of thrombectomy followed by venography and treatment of any underlying stenosis with balloon angioplasty. Surgical intervention can also provide definitive treatment of stenoses in the form of patch plasty ( Fig. 5.7 ) or interposition graft, as long as the stenosis is accessible to the surgeon. There is some evidence that stenoses that are treated surgically have better long-term patency. A recent comprehensive review of percutaneous intervention for thrombosed vascular access demonstrated an 80% success rate in retaining patency and avoiding temporary access catheters. Fistulas that are unsuitable for percutaneous intervention often require revision because of anastomotic or perianastomotic strictures that cannot be treated by angioplasty. Standard Fogarty vascular thrombectomy catheters are usually sufficient to clear early, soft thrombus but are not designed to penetrate mature thrombus. Fistulas that require surgical revision must retain enough length to accommodate two dialysis needles after revision.
Vascular access procedures are “clean surgery” although infection rates in renal patients are higher because of the relative immunocompromised state uremia produces. Uremia affects the immune system by inhibiting the bactericidal, phagocytic, and chemotactic action of neutrophils and by suppressing both B cell and T cell responses. Furthermore renal patients are more readily colonized by S. aureus compared with the general population, with an incidence of up to 70% upper respiratory tract colonization compared with 15%, respectively. S. aureus is the leading culprit in infective complications of vascular access conduits and is often resistant to first-line antibiotics. Routine use of antibiotics for autologous AVF formation is not universal. However, any procedure in which prosthetic material is used, a second-line intravenous antibiotic such as vancomycin or teicoplanin should be given. Wound infection rates are as low as 2% to 3% after autologous AVF formation although they are higher in AVGs, ranging between 5% and 35%. Drainage of any collections either by liberal suture removal or formal evacuation and irrigation under anesthesia alongside antibiotic therapy resolves the majority of infections. If there is concern about the severity of the infection or if MRSA is isolated and the infection is not responding to antibiotic therapy, then surgical debridement and inspection of the anastomosis should be undertaken. There is a small but potentially life-threatening risk of hemorrhage from an infected fistula, and ligation of the fistula may be required. Superficial wound infection in a patient with underlying prosthetic material should always be treated seriously. Aggressive, early antibiotic therapy should be employed to treat the infection and reduce the risk of graft infection. Superficial infection can often be treated successfully, but if there is a purulent infection or the graft is proven to be infected it must be removed.
Vascular access conduits can develop true and pseudoaneurysms. Unlike a true aneurysm, a pseudoaneurysm does not contain all layers of the blood vessel wall. It is a hematoma in direct communication with the lumen ( Fig. 5.8 ). The incidence of pseudoaneurysm is about 10% in prosthetic grafts and 2% in autologous AVF. Duplex ultrasound is usually diagnostic but venography or cross sectional imaging may be warranted if central venous pathology is suspected. Treatment can be radiologic or surgical, with the latter being the conventional method. Surgical repair usually involves oversewing the defect or removing the damaged section of graft and restoring the AVF either by direct end-to-end anastomosis or by using an interposition graft. Increasingly interventional radiologic methods are employed first; direct injection of thrombin into the defect is usually sufficient if the defect is small, and percutaneous deployment of a covered stent can also be used to exclude the aneurysm from the circulation. A true aneurysm can be defined as a threefold or greater increase in diameter compared with the rest of the access. True aneurysms are relatively commonly in upper limb fistulas, and older fistulas are more likely to become aneurysmal ( Fig. 5.9 ). Aside from the fistula being unsightly, most aneurysms are uncomplicated and at extremely low risk of rupture. However, aneurysms that are rapidly increasing in size, have thin skin, or infection present should be surgically corrected or ligated. Investigation of the rest of the access is essential before intervention, because up to 50% will have a clinically relevant stenosis. Many different surgical procedures have been detailed, and the choice is surgeon-dependent.