In clinical practice, a detailed history, physical examination, and duplex ultrasonography are commonly used in preparation for AV access creation. Detailed physical examination should assess the peripheral pulse and blood pressure in both extremities, examine the arm for edema, evaluate signs for heart failure, and identify marks of old lines and catheters, previous surgeries, and cardiac devices.

Preoperative vein mapping using vessel duplex ultrasonography was recommended in previous KDOQI 2006 guidelines solely on the basis of expert opinion to assess vein anatomy and diameter. However, several disadvantages are associated with duplex ultrasonography including operator errors, cost, patient inconvenience, and effect of temperature on vessel size. As such, KDOQI 2019 suggests the use of preoperative ultrasonography in a selected population, such as the elderly, females, those with multiple comorbidities (coronary and peripheral arterial disease), those with previous peripheral or central lines, and morbidly obese patients. Venography can be used when central venous stenosis is suspected. ²⁴ As discussed later, however, vascular ultrasonography can impact decision making and the success of AV access creation and is recommended if the operators are skilled in this technique.

Allon and colleagues have suggested a minimal vein size of >2.5 mm for AVF creation and >4 mm for AVG creation and an artery size of >2 mm. When these thresholds were used, the proportion of patients having an AV fistula placed increased from 34% to 64% as compared with historical controls, and the proportion of patients dialyzing with an AV fistula doubled from 16% to 34%.

Dageforde and colleagues retrospectively evaluated 158 patients who had brachiocephalic or brachiobasilic AVF and found that a vein size of >3.4 mm had a higher AVF maturation rate compared with those with vein size of <3.4 mm.

Most surgeons obtain intraoperative vascular ultrasonography. A prospective pilot study compared the surgeon’s decision about access placement in 70 consecutive patients with CKD, before and after the results of preoperative vascular mapping were provided to the surgeon. In almost one third of the patients, the surgeon changed his or her mind about the intended access procedure after receiving the mapping results. In most of these cases, the surgeon decided to place an AV fistula rather than an AV graft or used a different location for the fistula creation. Similar increases in AV fistula placement following the introduction of preoperative vascular mapping have been documented by other investigators ( Table 67.2 ), although a reduction in primary AV fistula failure has not been a consistent finding in all studies. ^, Two randomized clinical trials have compared outcomes of new AV fistulas placed after preoperative ultrasound vascular mapping, as compared with those placed after clinical evaluation alone. In a Turkish study of 70 patients, the two treatment groups had similar primary and secondary AV fistula survival. In a British study of 218 patients, the cohort that underwent a preoperative ultrasound had a lower incidence of immediate AV fistula failure (4% vs. 11%, P = 0.03). Primary (intervention-free) AV fistula survival was not different between the two groups (65% vs. 56%, P = 0.08), but the assisted primary (thrombosis-free) AV fistula survival was superior in the group undergoing preoperative ultrasound mapping (80% vs. 65%, P = 0.01).

About 80% of AV graft failures are due to thrombosis, whereas 20% are due to infection. ^, Thus improving AV graft longevity requires implementing measures to reduce the frequency of AV graft thrombosis. Reports from the U.S. Renal Data System showed that the number of angioplasties and thrombectomies (“declots”) for AV grafts had substantially increased from 0.49 to 1.10 (rate per patient year) and 0.15 to 0.48, respectively, from 1998 to 2007. Among AV grafts referred for thrombectomy, a significant proportion have underlying stenosis, most commonly at the venous anastomosis, recipient/draining vein, or central veins. This observation suggests that prophylactic angioplasty of hemodynamically significant AV graft stenosis may reduce the frequency of AV graft thrombosis and thereby increase cumulative AV graft survival. An early study suggested that measurement of dynamic venous pressures in AV grafts during consecutive hemodialysis sessions under carefully standardized conditions may be useful. Persistent elevation in the venous pressure measured at a low dialysis blood flow (200 mL/min) was associated with hemodynamically significant stenosis, and referral for prophylactic angioplasty in such cases led to a reduced frequency of AV graft thrombosis from approximately 0.6 to 0.2 events per year. This finding stimulated a large volume of subsequent studies aiming to answer two fundamental questions: 1. identifying noninvasive methods to screen for AV graft stenosis and 2. evaluating whether stenosis surveillance and prophylactic angioplasty improved AV graft outcomes.

A variety of methods have been used for detection of hemodynamically significant AV graft stenosis ( Table 67.3 ). Clinical monitoring consists of using information that is readily available from physical examination of the AV graft, abnormalities experienced during the dialysis sessions (difficult cannulation or prolonged bleeding from the needle puncture sites), or unexplained decreases in the dose of dialysis, urea reduction ratio, or volume-indexed clearance-time product (Kt/V). ^{, ,} The positive predictive value of clinical monitoring has been reported to be between 69% and 93% ( Table 67.4 ). ^{, ,} AV graft surveillance uses noninvasive tests requiring specialized equipment or technician training that is not obtained as part of the routine dialysis treatment. These include measurement of static dialysis venous pressure (normalized for the systemic pressure), measurement of the access blood flow, or duplex ultrasound to directly evaluate stenosis. Each of these monitoring or surveillance tools has been reported to have a positive predictive value for AV graft stenosis of between 70% and 100% (see Table 67.4 ). The negative predictive value has not been studied systematically, as it would require obtaining routine angiography in patients whose screening test is negative. However, the negative predictive value can be inferred from the proportion of AV graft thromboses not preceded by abnormalities of AV graft surveillance, which is around 25%.

In contrast, the predictive value of surveillance methods for AV graft thrombosis is much less impressive. Thus when AV grafts with abnormal monitoring criteria suggestive of stenosis are observed without preemptive angioplasty, only about 40% of the AV grafts clot over the following 3 months. ^, In practice, this means that in any program of AV graft monitoring, about half of preemptive angioplasties performed may be unnecessary. Unfortunately, there are no reliable tests to distinguish the subset of AV grafts with stenosis that will progress to thrombosis from those that will remain patent without any intervention.

Several observational studies have documented that introduction of a monitoring or surveillance program for AV graft stenosis with preemptive angioplasty lowered the frequency of AV graft thrombosis by 40% to 80%, as compared with the historical control period during which there was no monitoring program ( Table 67.5 ). ^{, , , ,} The promising findings from these observational studies led to the implementation of programs of graft surveillance and preemptive angioplasty aiming to decrease the frequency of AV graft thrombosis. However, over the past few years, the value of AV graft stenosis surveillance has been subjected to rigorous testing in randomized clinical trials ( Table 67.6 ). ^{, ,} Only one of the six randomized trials has demonstrated a benefit of ultrasound AV graft surveillance ; the other five showed no benefit of surveillance, despite a substantial increase in the frequency of preemptive angioplasty in the surveillance group. ^{, ,} For example, one study randomized patients with AV grafts to standard clinical monitoring alone or to a combination of clinical monitoring and ultrasound surveillance for stenosis. The patients in the ultrasound group had a 66% higher frequency of undergoing preemptive angioplasty, yet there was no difference between the two randomized groups in terms of frequency of AV graft thrombosis, time to first thrombosis, or likelihood of AV graft failure ( Fig. 67.5A and 67.5B ). Further, a meta-analysis of the randomized studies suggested that the benefit of surveillance with preemptive angioplasty, if present, is likely to be small, with a reduction of AV graft thrombosis not exceeding 23% and a reduction in AV graft failure no greater than 17%. Therefore a large-scale, multicenter study trial, which takes into account multiple factors involved in access thrombosis, would be required to provide a definitive answer to this controversial question. Accordingly, KDOQI 2019 does not recommend routine AVG surveillance using blood flow measurements, pressure monitoring, or access imaging due to the low quality of evidence.

(A) Comparison of cumulative graft survival between randomized patients with clinical monitoring versus clinical monitoring plus regular ultrasound surveillance of grafts.

P = 0.93 by the log rank test. (B) Comparison of thrombosis-free graft survival between randomized patients with clinical monitoring versus clinical monitoring plus regular ultrasound surveillance of grafts. P = 0.33 by the log rank test.

A and B, Reproduced from Robbin ML, Oser RF, Lee JY, et al. Randomized comparison of ultrasound surveillance and clinical monitoring on arteriovenous graft outcomes. Kidney Int. 2006;69[4]:730–735.

The question remains: If the underlying AV graft stenosis is an important predictor of AV graft thrombosis, why is preemptive angioplasty not more successful in reducing AV graft thrombosis? The fundamental problem appears to be the short-lived efficacy of angioplasty to relieve AV graft stenosis. When serial access blood flows have been used as a surrogate marker of successful angioplasty, 20% of AV grafts develop recurrent stenosis within 1 week of angioplasty and 40% within 1 month of angioplasty. ^, In another study, the mean vascular access blood flow following angioplasty increased from 596 to 922 mL/min and then decreased to 672 mL/min within 3 months after angioplasty. In addition, the angioplasty-induced injury may actually accelerate neointimal hyperplasia, thereby resulting in recurrent stenosis. Not surprisingly, patients undergoing angioplasty for AV graft stenosis require frequent re-interventions due to recurrent stenosis. The median intervention-free patency following AV graft stenosis is only about 6 months. ^, Finally, AV graft surveillance is likely to be most useful in new AV grafts (within 3 months of the creation). A large proportion of AV grafts fail before there has been adequate opportunity for surveillance to be performed and for hemodynamically significant stenosis to be detected and treated.

The goal of elective angioplasty is to correct the stenotic lesion that impairs optimal delivery of dialysis, and hopefully delay AV graft thrombosis. Measurements of the degree of stenosis are not always feasible, nor are they entirely objective. Caliper measurement or “eyeballing” are to various degrees, subjective. Even electronic quantitative analysis is not reliable in detecting the various densities and edges of the normal vessels and requires operator adjustment of detected vascular edges (a subjective intervention). Arteriovenous graft loops originating from the brachial artery and those that have anastomosis at a 90-degree angle with the basilic vein are hardest to image in profile, which is necessary in order to get an accurate measurement of stenosis. In practice, the degree of stenosis is estimated subjectively by evaluating the angiogram images before and after angioplasty. In hospital-based centers, intra-graft pressure measurement before and after angioplasty is usually obtained.

The most common location of the stenosis localized by angiography is the venous anastomosis, followed by the peripheral draining vein, the central veins, and intragraft ( Table 67.7 ). Inflow (arterial anastomosis) stenosis has been rare (<5%) in most series. A number of published series have documented the short-lived primary patency (time to next radiologic or surgical intervention) of AV grafts following elective angioplasty ( Table 67.8 ), ^{, , ,} with only 50% to 60% patency at 6 months and 30% to 40% at 1 year. This means that, on average, each AV graft requires two angioplasties per year. The primary patency is shorter after angioplasty of central vein stenosis, as compared with other stenotic locations. In one study, the primary patency at 6 months was only 29% for central vein stenosis, as compared with 67% for stenosis at the venous anastomosis. Most studies have documented progressively shorter patency after each consecutive angioplasty. Notably, the primary patency of AV grafts following elective angioplasty is not affected by patient age, race/ethnicity, diabetes, or peripheral vascular diseases. However, the patency tends to be shorter in women than in men. The primary patency after angioplasty is also not influenced by the location of the AV graft or the number of concurrent stenotic lesions found. The technical success of an angioplasty procedure may be assessed in several ways. The first is visual inspection of the angiogram before and after the procedure to determine whether the magnitude of the stenosis (percent stenosis relative to the normal vessel diameter) has been reduced. The degree of stenosis of each lesion can be quantified with calipers, electronic quantitative analysis or graded semi-quantitatively using the following scale: grade 1, no (<10%) stenosis; grade 2, mild (10%–49%) stenosis; grade 3, moderate (50% to 69%) stenosis; and grade 4, severe (70%–99%) stenosis. ^, A second approach is to measure the intragraft pressure before and after the procedure and normalize it for the systemic blood pressure. A third approach is to measure the change in access blood flow before and after the procedure. Each of these measures has been shown to be predictive of the primary patency of AV grafts following elective angioplasty. In one large series, following elective angioplasty, the median intervention-free survival of AV graft with no residual stenosis was 6.9 months, as compared with 4.6 months with any degree of residual stenosis. Similarly, the median primary patency of AV grafts following angioplasty was inversely related to the intragraft-to–systemic pressure ratio, being 7.6, 6.9, and 5.6 months, when this ratio was <0.4, 0.4 to 0.6, and >0.6, respectively. Finally, a failure to significantly increase the access blood flow after angioplasty is observed in 20% of AV grafts at 1 week and in 40% by 1 month, ^, confirming the short-lived benefit of this intervention.

Technical procedure: percutaneous graft angioplasty

A digitally subtracted antegrade angiogram is performed to visualize the entire access circuit. The presence or absence of stenotic lesions and their number and location are assessed, and arterial anastomosis, intragraft, venous anastomosis, and draining vein and central vein lesions are documented. The degree of stenosis of each lesion is quantified with calipers or graded semiquantitatively. ^, Lesions with at least 50% stenosis are considered to be hemodynamically significant and undergo angioplasty using an appropriately sized balloon ( Fig. 67.6A–67.6D and see Fig. 67.7A–67.7C ). The majority of anastomotic lesions require higher pressure than those required for peripheral arterial angioplasty. Therefore high pressure balloons with minimal burst pressure >25 atmospheres are routinely used. If a residual (>30%) stenosis is found, prolonged angioplasty (2- to 3-minute inflation), higher-pressure balloons (up to 30 atmospheres), and occasionally covered stent deployment may be required to treat these lesions. The main complication of this procedure is vessel extravasation and/or rupture of the vessel during angioplasty. Most of these extravasations are controlled with balloon inflation for prolonged time and occasionally require covered stent deployment. Surgical repair is rarely indicated and usually reserved for ruptures that are not controlled with stent placement and associated with major local hematoma.

(A) Left upper arm arteriovenous (AV) graft angiogram showing a severe (95%) stenotic lesion at the level of the venous anastomosis.

(B) Left upper arm AV graft stenotic lesion at the venous anastomosis with the angioplasty balloon partially inflated. (C) Left upper arm AV graft stenotic lesion at the venous anastomosis with the angioplasty balloon fully inflated. (D) Final postangioplasty left upper arm AV graft angiogram showing a treated lesion with minimal residual stenosis.

(A) Digital subtraction angiography (DSA) of a left upper arm AV graft showing a moderate stenotic lesion at the level of the arterial limb of the graft.

(B) Spot film showing the stenotic lesion being angioplastied. (C) Final postangioplasty DSA showing excellent results.

The pathophysiology of AV graft stenosis involves proliferation of vascular smooth muscle cells (neointimal hyperplasia), with progressive encroachment of the lesion into the AV graft lumen. There has been ongoing interest in pharmacologic approaches to prevention of neointimal hyperplasia. Two small, single-center, randomized clinical trials have documented a beneficial effect of dipyridamole and fish oil in preventing AV graft thrombosis. ^, A multicenter, randomized, double-blind study compared clopidogrel plus aspirin with placebo for prevention of AV graft thrombosis. The study was terminated early due to an excess of bleeding complications in the intervention group; there was no difference in the rate of AV graft thrombosis between the two randomized groups. Similarly, a randomized clinical trial found that low-intensity warfarin posed a substantial risk of major hemorrhagic complications without reducing the frequency of AV graft thrombosis.

The Dialysis Access Consortium was a larger multicenter, randomized, double-blinded clinical trial comparing long-acting dipyridamole + low-dose aspirin (Aggrenox) with placebo in prevention of AV graft failure. There was a modest but statistically significant improvement in primary unassisted AV graft survival in patients receiving dipyridamole-aspirin, as compared with those treated with placebo (28% vs. 23%, at 1 year P = 0.03). Cumulative AV graft survival did not differ between the two treatment arms. ^, Another double-blind, randomized, controlled study evaluating fish oil on PTFE AV graft patency and cardiovascular outcomes reported that fish oil did not decrease loss of native patency (time from creation to first intervention) of AV grafts at 12 months but did show improvement in overall AV graft patency and cardiovascular outcomes. In summary, based on available evidence, the use of combination of aspirin and dipyridamole to improve the graft primary patency should be individualized on the basis of patient risk-benefit analysis. Additionally, the use of oral fish oil may reduce the incidence of thrombosis in patients with newly created AV grafts. ^, Finally, to improve the patency of AV grafts following angioplasty, stent deployment was investigated. It was postulated that the rigid scaffold of the stent could potentially keep the vascular lumen open. The use of stents in dialysis access is discussed later.

Compared with AV grafts, AV fistulas require a much lower frequency of intervention (angioplasty or thrombectomy) to maintain their long-term patency for dialysis. However, AV fistulas have a substantially higher primary failure rate (AV fistulas that are never usable for dialysis). Similar to AV grafts, the rate of angioplasty utilization increased from 0.16 to 0.47 (rate per patient year) in U.S. Renal Data System data, likely due to the increasing frequency of treating primary AV fistula failures. The proportion of AV fistulas with primary failure has ranged from 20% to 60% in multiple series, even when routine preoperative vascular mapping has been employed. ^{, ,} Primary AV fistula failures fall into two major categories: early thrombosis and failure to mature. Early thrombosis refers to AV fistulas that thrombose within 3 months of the creation, before they have been used for dialysis. Failure to mature refers to AV fistulas that never develop adequately to be cannulated reproducibly for dialysis. Nonmaturation is less common with upper arm than forearm AV fistulas. Among upper arm AV fistulas, nonmaturation is less likely with transposed brachiobasilic than brachiocephalic AV fistulas.

In some patients, maturation of the AV fistula can be assessed easily by clinical evaluation by the nephrologist, surgeon, or an experienced dialysis nurse. In less straightforward cases, duplex ultrasound may be useful in predicting whether a new AV fistula can be used successfully for hemodialysis. One pilot study used a combination of two simple sonographic criteria to assess AV fistula maturation: AV fistula diameter and access blood flow. When the ultrasound documented a draining vein diameter ≥4 mm and an access blood flow ≥500 mL/min, 95% of the AV fistulas were subsequently usable for dialysis. In contrast, when no criterion was met, only 33% of the AV fistulas achieved adequacy for dialysis. The likelihood of AV fistula adequacy for dialysis was intermediate (∼70%) when only one of the two criteria was met.

Failure of an AV fistula to mature can be related to one of several anatomic defects, which can be identified either by sonography or angiography. The main causes of fistula failure to mature are stenosis at the arterial anastomosis/juxta-anastomotic segment or in the draining vein and the presence of large accessory veins that affect fistula blood flow. In most cases, these anatomic problems can be corrected percutaneously or surgically. Stenotic lesions can be treated by angioplasty or surgical revision. Superficial side branches can be ligated by a suture through the skin; deeper branches can be embolized.

It is worth mentioning that in obese patients, the AV fistula may have adequate caliber and blood flow but is too deep to be cannulated. This may require surgical superficialization to facilitate AV fistula cannulation.

In immature AV fistulas with one or more of the anatomic lesions described, specific interventions to correct the underlying lesion may promote subsequent AV fistula maturation. Several published series have evaluated the ability to salvage immature AV fistulas such that they are subsequently usable for dialysis. A number of studies using only radiographic procedures (angioplasty of stenotic lesions or obliteration of side branches) in immature AV fistulas have shown a high success rate ( Table 67.9 ). An initial salvage (ability to use the AV fistula for dialysis) was accomplished in 80% to 90% of patients, with a subsequent 1-year primary patency of 39% to 75%. A study using a combination of radiologic and surgical salvage procedures in an unselected dialysis population reported a more modest salvage rate of 44%. The frequency of salvage procedures for immature AV fistulas in this study was twice as high in women than men. Retrospective analysis of immature AV fistulas with an anatomic abnormality that were corrected with angioplasty or surgically found that these were more likely to be usable dialysis, compared with similar fistulas that did not undergo a salvage procedure. Balloon-assisted maturation may enhance the maturity rate in these AV fistulas. The premise of balloon-assisted maturation is to perform repeated long-segmental sequential angioplasty of the whole access to promote maturity. Further, balloon-assisted maturation was used intraoperatively during fistula creation in patients with small arteries and veins (artery, <2 mm; vein, <2.5 mm). ^, Overall, these techniques were not compared with a control group and fistulas generally required further angioplasty after maturity to maintain patency. Importantly, immature AV fistulas requiring interventions to achieve maturity had significantly shorter cumulative patency and required more interventions to maintain patency, compared with AV fistulas achieving maturation without interventions ^, ( Fig. 67.8 ). Randomized studies comparing different types of interventions or different timing of interventions in terms of their effect on AV fistula maturation are generally lacking. In practice, proactive percutaneous or surgical interventions are attempted to salvage these immature accesses.

Interventions to promote arteriovenous fistula maturation.

Cumulative survival, defined from the time of access cannulation to permanent failure, is shorter in patients receiving two or more interventions before arteriovenous fistula maturation compared with those with zero interventions (hazard ratio, 2.07; 95% confidence interval, 1.21 to 2.94; P = 0.0001).

Reproduced from Lee T, Ullah A, Allon M et al. Decreased cumulative access survival in arteriovenous fistulas requiring interventions to promote maturation. Clin J Am Soc Nephrol. 6:575–581, 2011 with permission from the American Society of Nephrology.

There has also been interest in pharmacologic interventions to promote AV fistula maturation. A randomized study of 877 U.S. patients undergoing creation of a new AV fistula allocated patients to receive clopidogrel or placebo for the initial 6 weeks. While clopidogrel significantly decreased the frequency of AV fistula thrombosis within 6 weeks, there was no effect on access maturation. Subsequently, another randomized study from Australia and New Zealand of 567 patients reported that neither fish oil nor low-dose aspirin prevented new AV fistula failure. There is some evidence to support the role of far-infrared therapy in improving the primary patency of AVF. The use of this modality was evaluated in four studies in Taiwan ( N = 763) and reported higher primary patency rate at 1 year of follow-up. The use of fish oil or aspirin did not show to prevent AV fistula flow dysfunction. There is inadequate evidence to support the use of simvastatin, ezetimibe, or clopidogrel-prostacyclin to reduce fistula thrombosis or improve primary failure.

Technical aspects of salvaging immature arteriovenous fistulas

Percutaneous transluminal angioplasty of arteriovenous fistulas

The stenosis at the juxta anastomotic segment and within the venous outflow can be treated with sequential balloon dilatations over one or more angioplasty sessions. Generally, the fistula is cannulated in a retrograde manner using ultrasonography guidance and the feeding artery is cannulated. Subtracted digital angiograms of the dialysis circuit usually delineates the culprit lesion(s) that are treated with appropriately sized balloons ( Fig. 67.9A and 67.9B ). In practice, these patients are brought back in 2 to 4 weeks for physical examination and repeat angiogram as clinically warranted. Once the fistula is deemed suitable for cannulation, it is highly recommended that an experienced nurse cannulates the new access using one short needle and then advance to two needles as feasible.

(A) Fistula salvage.

Digital subtraction angiography (DSA) of a radiocephalic arteriovenous fistula showing a severe stenotic lesion at the level of the juxta arterial anastomosis. (B) Fistula salvage: Postangioplasty DSA of a radiocephalic fistula showing radiologic improvement of the stenotic lesion at the level of the juxta arterial anastomosis.

Ligation of accessory veins

Accessory veins that affect the fistula maturity can be treated by either surgical ligation or endovascular coil deployment. Accessory veins are treated depending on size, location, and number ( Fig. 67.10 ). While superficial and deep large accessory veins are usually ligated surgically, vessel embolization using coil deployment is used to treat small deep accessory veins ( Fig. 67.11 ) To accomplish this, the fistula is accessed and a selective cannulation and an angiogram are performed on the intended veins, followed by deploying appropriately sized coils. A final angiogram of the AV fistula is performed to ascertain proper coil deployment and occlusion of all collateral veins.

Digital subtraction angiography of a left radiocephalic arteriovenous fistula showing multiple collaterals.

There is a metallic plate from a prior open reduction and internal fixation of a radius bone fracture.

Coil deployed in a collateral vein of an upper arm arteriovenous fistula.

Although the frequency of interventions is several-fold lower in AV fistulas than grafts, AVFs are also susceptible to developing stenosis and thrombosis. Most studies have documented a comparable primary patency of fistulas and grafts following elective PTA ( Table 67.10 ), ^{, , ,} although one study observed a higher primary patency in fistulas. As is the case with angioplasty of grafts, the primary patency of fistulas following PTA is inversely related to the magnitude of postangioplasty stenosis, as well as the magnitude of the postangioplasty intraaccess to systemic pressure ratio.

Table 67.10

Primary Patency After Elective Angioplasty: Fistulas Versus Grafts

Reference	Primary Access Patency at 6 Months
	Grafts	Fistulas
Safa, 1996	43%	47%
Turmel-Rodrigues, 2000	53%	67%
McCarley, 2001	37%	34%
Van der Linden, 2002	25%	50%
Maya, 2004	52%	55%

After performing the initial digitally subtracted angiogram of the access circuit, PTA is done on the hemodynamically significant lesions (>50% stenosis). A final digitally subtracted angiogram is performed to assess for residual stenosis and requirement for further treatment of the stenotic lesion ( Fig. 67.12 ). In some centers, intrafistula pressure and systemic pressure are measured before and immediately after the intervention to assess the efficacy of angioplasty; a reduction in the intrafistula to systemic pressure ratio is used to confirm hemodynamic improvement. The major complications of this procedure are vessel extravasations and rupture of the vessel after the angioplasty. As in AV grafts, most of these complications are controlled with prolonged balloon tamponade at the extravasation site. Stent deployment is indicated occasionally, and surgical referral is rarely needed ( Fig. 67.13 ).

(A) Digital subtraction angiography (DSA) of a left radiocephalic arteriovenous (AV) fistula showing a long severe segment of stenosis distal to the arterial anastomosis followed by a pseudoaneurysm of the fistula.

(B) Spot film of a left radiocephalic AV fistula showing the segment of stenosis being angioplastied. (C) Postangioplasty DSA shows a successful treatment, and the pseudoaneurysm is unchanged.

(A) Angioplasty complication.

Digital subtraction angiography (DSA) showing a rupture of the left cephalic vein after aggressive percutaneous transluminal angioplasty. There is a coexisting stenosis of the left subclavian. (B) Angioplasty complication: DSA showing salvage and correction of the complication by deploying a covered stent (wall-graft).

The diagnosis of dialysis access thrombosis is usually made clinically; the absence of thrill bruit or pulse during physical examination indicates access thrombosis. Access ultrasonography is rarely needed for diagnosis. Once the diagnosis is made, access thrombectomy is recommended as soon as possible to improve access survival and avoid CVC use. Localized pain and tenderness can be encountered over the thrombosed AVF site due to local inflammatory thrombophlebitis. If these findings are seen in a thrombosed AVG, it usually denotes graft infection.

There are few contraindications to perform endovascular thrombectomy in a timely fashion. These contraindications are classified into absolute, relative, and temporary. Access infection, pulmonary hypertension, recent access creation, severe ipsilateral steal syndrome, and right-to-left cardiac shunts are considered absolute contraindications. Performing endovascular thrombectomy in a recently created vascular access carries a risk of rupture of the fresh surgical anastomotic site. As such, these procedures should be done by experienced interventionalists or vascular surgeons. Fluid overload, hyperkalemia (K>6mEq/L), and hemodynamic instability are regarded as temporary contraindications. In these cases, a temporary dialysis catheter will be required to perform dialysis therapy before access thrombectomy. Recent intracranial hemorrhage, mega fistula, contrast allergy, and cardiac arrest are considered relative contraindications and should be left to the operator discretion and consideration of the risk of the required anticoagulation.

Currently, endovascular thrombectomy is considered the procedure of choice to treat dialysis access thrombosis. The time between vascular access thrombosis and reestablishment of access flow is an independent predictor of vascular access survival; therefore it is imperative to perform thrombectomy as soon as possible. When performing access thrombectomy, four fundamental principles require special attention from the operator. Firstly, the location of the thrombus and the clot burden can vary between fistulas and grafts. Whereas the thrombus generally occupies the entire length of the synthetic graft, in the thrombosed AVF, the thrombus is usually localized at the juxta-anastomotic segment area. Secondly, thrombectomy of AVFs is more challenging compared with AVGs. The reason for this difference is that AVF thrombosis can trigger an inflammatory response that leads to the adherence of the thrombus to the vessel wall. Consequently, the success rate for AVF thrombectomy declines significantly if thrombectomy is delayed by more than 2 to 3 days. However, AVG thrombectomy can be performed successfully up to 2 weeks following its thrombosis. Thirdly, thrombosis of the dialysis access within the first month of creation requires surgical thrombectomy due to the inherent risk of rupturing the suture line with balloon inflation. Finally, recurrent AVGs thrombosis (>3 episodes of thrombosis within 3 months) and AVF thrombosis soon after its creation often require surgical revision.

Thrombectomy of dialysis access is a multistep procedure that aims to restore the blood flow within the access circuit by utilizing a combination of thrombolysis and mechanical thrombectomy using different devices, balloons, and stents as clinically indicated.

The fundamental principles of access thrombectomy center around 1. removal of the thrombus from the access conduit, and 2. correcting the underlying culprit stenosis of the access ( Fig. 67.19 ).

Principles of performing dialysis access thrombectomy.

From Almehmi A, Al-Balas A. The clotted access. In: Illig KA, Scher LA, Ross JR, eds. Principles of Dialysis Access . Springer; 2024:419–434.

After obtaining informed consent, an intravenous line should be established to administer agents such as midazolam and/or fentanyl for conscious sedation. The endovascular thrombectomy procedure is performed in the interventional suite, where the patient’s blood pressure, heart rate, oxygen saturation, and cardiac rhythm are continuously monitored. Local anesthesia is achieved with 1% to 2% lidocaine and 2000 to 5000IU of heparin is given systematically. Two antegrade and retrograde access sheaths (6 or 7 Fr) are inserted facing each other without overlapping. Occasionally, ultrasound guidance is used to access the AVF/AVG. Next, a pullback angiogram is performed to assess the outflow veins and the level of stenotic lesions. Most interventionalists currently use a combination of thrombolysis and mechanical thrombectomy to remove the clot burden ^{, ,} ( Fig. 67.20 ).

Dialysis access thrombectomy using Fogarty balloon.

(A) Two 7-French sheaths inserted in the arterial and venous sides of the graft (retrograde and antegrade, respectively). (B) Pullback contrast injection to determine the level of stenosis. (C) Removal of arterial plug using Fogarty balloon (arrow). (D) The arterial limb is patent ( arrow —blood flowing from the sheath sidearm). (E) Angioplasty of the venous anastomosis ( arrow —angioplasty balloon). (F) Final angiogram of the dialysis access showing patent flow.

From Almehmi A, Sheta M, Abaza M, et al. Endovascular management of thrombosed dialysis vascular circuits. Semin Intervent Radiol . 2022;39[1]:14–22.

Before using thrombolytic agents, it is crucial to assess the patient for the presence of any contraindications to thrombolytic therapy, such as recent surgery, bleeding disorder, recent episode of bleeding, and the presence of severe hypertension.

Several thrombolytic agents are available including urokinase and streptokinase, as well as tissue plasminogen activator (tPA), which is the most used agent in the United States. The technical success of thrombolysis in reestablishing the blood flow in a thrombosed dialysis circuit is modest ranging between 33% and 80%. For this reason, mechanical thrombectomy is often combined with pharmacologic thrombolysis to improve the access patency rate.

Mechanical thrombectomy uses different devices in combination with thrombolytics to reestablish blood flow. In essence, devices of mechanical thrombectomy are divided into two categories: direct wall contact devices and hydrodynamic devices (see Fig. 67.19 ).

• 1.

  Direct Wall Contact Devices: Using a balloon, such as Fogarty balloons (Edwards Laboratories, Santa Ana, CA) or a rotational device, such as Arrow-Trerotola (Teleflex, Wayne, PA), and Cleaner XT Rotational Thrombectomy System (Argon Medical Devices, Plano, TX), the luminal clot burden is removed by pullback approach. The rotational device works by fragmenting the clot within the access circuit via generating a hydrodynamic vortex created by the high speed (800–5000rpm) rotating impella. ^, These devices are inexpensive and disposable. Some of them can be used over the wire. Almehmi and colleagues described the use of Fogarty in combination with a no-wait lysis approach in performing AVG thrombectomy. In this study, the Fogarty balloon was used to remove the arterial plug and establish the blood flow from the feeding artery. Once the arterial flow was established, tPA and heparin were infused and the clot burden was macerated using noncompliant angioplasty balloon. This approach was associated with a high technical success rate of thrombectomy and shorter radiation and procedure time.

• 2.

  Hydrodynamic (rheolytic, non–wall contact) Devices: These devices use high-speed rotation to generate a hydrodynamic vortex that disrupts the clot and creates a negative pressure gradient. Subsequently, this negative gradient leads to the removal of the clot microfragments into a collection bag without vessel wall contact. The hydrodynamic vortex results from the high-speed fluid jets that produce the Venturi (rheolytic) effect. ^, Several devices are used, such as AngioJet (Boston Scientific, Marlborough, MA), Oasis catheter (Boston Scientific, Natick, MA), Hydrolyzer (Cordis, Miami, FL), Amplatz thrombectomy device (Microvena, White Bear Lake, MN), and Straub Rotarex catheter (Straub Medical AG, Wangs, Switzerland). For instance, an AngioJet catheter is a wire-guided catheter that comes in different sizes, 4- to 6-Fr systems. After passing the catheter over a wire beyond the distal end of the thrombus, the catheter is activated and drawn back into the clot. At this point, the catheter generates high-speed saline jets in a retrograde direction that produces a Venturi suction gradient that breaks down the clot burden. Finally, these generated fragments are evacuated into the device. Overall, the success rate of the above devices in reestablishing blood flow in thrombosed dialysis access is more than 90%. ^,