Heidi R. Umphrey, Lauren F. Alexander, Michelle L. Robbin and Mark E. Lockhart

Arteriovenous Fistulas and Synthetic Arteriovenous Grafts

At the end of 2009, there were 571,414 patients being treated for end-stage renal disease in the United States with 116,395 new cases in that year¹; approximately 65% of these patients received haemodialysis.¹ Complications associated with vascular access procedures are a major cause of morbidity and increasing healthcare costs in patients undergoing haemodialysis.² For patients with end-stage real disease (ESRD) requiring haemodialysis, there are two options for vascular access, either the surgical creation of an arteriovenous fistula (AVF) or implantation of a synthetic arteriovenous graft (graft). Due to the lower rates of infection and thrombosis, mature AVFs are the preferred access when possible.^3,4 Preoperative ultrasound evaluation of the upper extremity arteries and veins has been shown by several studies to increase the success rate of AVF creation by influencing surgical planning.^5–7 While sonographic postoperative haemodialysis access evaluation may be beneficial in assessing AVF maturation,^8,9 the role of postoperative ultrasound monitoring for early detection of access pathology allowing prompt intervention to increase the longevity of an access is still evolving.^10–15

When clinically feasible and anatomically possible, the surgical creation of an AVF is preferred over a graft. Access placement in the non-dominant arm allows activities of daily living to continue as the non-dominant arm recovers; however, it is preferential to place a dominant arm AVF as opposed to a non-dominant arm graft in most patients. Potential haemodialysis access sites in decreasing order of preference are as follows: (1) forearm AVF (radiocephalic AVF or transposed forearm basilic vein to radial artery AVF); (2) upper arm brachiocephalic AVF or transposed brachiobasilic AVF; (3) forearm loop graft; (4) upper arm straight graft (brachial artery to upper basilic/axillary vein); (5) upper arm axillary artery to axillary vein loop graft; and (6) thigh graft. Other less common access configurations may also be utilised based on surgical experience.¹⁶ Cephalic vein use is preferred over a basilic vein transposition for fistula formation because the cephalic vein procedure involves less dissection and venous handling.

Preoperative Evaluation

Ultrasound exam optimisation has been previously reported.9,17–19 For evaluation of upper extremity arteries and veins, the patient should preferably be sitting upright. Following arterial assessment, a tourniquet is placed for evaluation of the veins²⁰ (Fig. 7-1). The tourniquet should be repositioned on the arm above each level of examination, so those veins are maximally distended. Assessment of the internal jugular and subclavian veins should be performed with the patient supine for improved distension.

The vascular structures are generally superficial and a high-frequency linear array transducer, at 12–15 MHz or higher, can provide good spatial resolution with adequate penetration. A light touch and plenty of coupling gel are recommended, so as to not deform the circular shape of the vessels. Preoperative criteria suggested by the literature include a minimum intraluminal arterial diameter of 2.0 mm (Fig. 7-2) and a minimal intraluminal venous diameter of 2.5 mm (Fig. 7-3) to allow successful AVF creation.^5,9 Suggested criteria for graft creation include a minimum intraluminal venous diameter of 4.0 mm and a minimum arterial diameter threshold of 2.0 mm.⁵ In addition to diameter measurements, the caudal third of the brachial and radial arteries are evaluated for intimal thickening, calcification, stenosis or occlusion. Severity of arterial calcification may be categorised, depending on surgeon preference (Fig. 7-4). Evaluation for a normal triphasic or biphasic arterial waveform is performed, and peak systolic velocity (PSV) is measured in the region of artery planned for use (Fig. 7-5). If two arteries with paired veins are seen in the upper arm, a high radial artery take-off from the brachial artery should be suspected (Fig. 7-6). These arteries should be followed into the forearm to confirm high radial artery take-off and to exclude a prominent arterial branch supplying the elbow.

Venous assessment is performed with a tourniquet in place, and each vein should be visually inspected and compressed along the entire venous length to exclude thrombus. Internal transverse diameters of the cephalic vein are measured in the forearm at multiple levels. Internal diameters of the cephalic and basilic veins are measured in the upper arm at multiple levels in a similar fashion. The axillary vein, subclavian vein, and internal jugular vein should be assessed for compressibility and/or normal waveforms (Fig. 7-7).

Postoperative Ultrasound Evaluation

A surgically created dialysis access can have a variety of complications, and most of these can be identified and evaluated by ultrasound. Prior to the sonographic evaluation of a haemodialysis access, pertinent patient history and surgical notes are reviewed. An initial cursory ultrasound scan is performed to obtain an overview of the haemodialysis access. Once the general layout is understood, sonographic assessment is performed with greyscale and Doppler. The caudal third of the feeding artery is assessed for possible stenosis, and the intraluminal diameter is measured in the transverse plane. Colour and spectral Doppler assessment of the feeding artery is performed in the longitudinal plane to confirm the presence of normal low-resistance flow (Fig. 7-8). The anastomosis(es) are evaluated with both colour and spectral Doppler to assess for stenosis. The draining vein of the AVF or graft is inspected for wall thickening, stenosis, and thrombosis along its entire length. For a fistula, the intraluminal draining vein diameter and the vein depth from skin surface are measured at several points cranial to the arteriovenous anastomosis. A depth greater than 5 mm may suggest the need for superficialisation to improve accessibility. The draining vein is evaluated for accessory vein branches (Fig. 7-9). Internal diameter and distance from anastomosis are recorded for each identified accessory vein branch within 10–15 cm of the anastomosis. Flow volume measurements are obtained within the mid draining vein of an AVF or within the graft. This measurement is obtained in an area with parallel vessel walls, minimal vessel tortuosity and no stenosis (Fig. 7-10).

If intraluminal thrombus is identified within an access or draining vein, a longitudinal greyscale image is obtained for documentation. Colour and spectral Doppler should be performed when access occlusion is suspected to confirm and document absent flow. If there is any uncertainty, power Doppler should be utilised since it is more sensitive for detection of residual slow flow. If fluid collections are present, they should be measured and inspected for echogenic foci or gas. Subcutaneous air within a collection or adjacent to the access can be the result of recent access cannulation. Echogenic foci associated with a fluid collection, however, can represent a developing infection or an abscess.

The artery downstream from the arterial anastomosis is routinely evaluated for flow reversal, which may be asymptomatic. If there are clinical symptoms suspicious for arterial steal, such as hand pain, the arterial flow to the hand should be carefully evaluated. In severe cases, revision or sacrifice of the access may be necessary.

In the setting of upper extremity swelling, the subclavian and internal jugular veins should be included in the postoperative sonographic evaluation. Doppler imaging of the internal jugular and subclavian veins provides indirect assessment of the brachiocephalic veins. In a patient with an AVF, spectral Doppler evaluation of normal central veins shows respiratory phasicity and/or transmitted cardiac periodicity (Fig. 7-7). Monophasic subclavian and internal jugular venous waveforms are suggestive of central venous stenosis or occlusion, especially if these veins are distended (Fig. 7-11). MR venography may be useful to exclude critical stenosis in this setting.

A normal AVF has antegrade flow through both the arterial and venous limbs without visible narrowing, thrombus, or flow disturbance. Potential signs of stenosis include areas of visible narrowing, aliasing, or focal increased velocity, and are further described in the following pathology section. A functioning AVF has flow volume of at least 300 to 800 mL/min (Fig. 7-10).^8,21 When an AVF has a minimum draining vein diameter of > 4 mm or a blood flow rate of > 500 mL/min, approximately 70% can be successfully used for haemodialysis.⁸ When both criteria are present, the likelihood of fistula maturation is 95%.⁸ If neither of these are met, only 33% of fistulas are adequate for haemodialysis.⁸ The National Kidney Foundation Kidney Disease Outcomes Quality Initiative published sonographic criteria suggestive of maturation that included the following: draining vein greater than 6 mm diameter, blood flow rate greater than 600 cc/min, and less than 6 mm skin depth.³ However, these criteria exclude many fistulas that can still be successfully used for haemodialysis.

A normal haemodialysis graft is seen as two echogenic lines that represent strong specular reflection from the polytetrafluoroethylene (PTFE) material. Flow within the graft should be antegrade demonstrating low resistance, arterialised flow on spectral Doppler. Findings suggestive of stenosis include visible narrowing, focal turbulence, or focally elevated velocity. Antegrade flow without focal turbulence or aliasing is seen in the draining vein of a normal graft. Because of the larger blood flow volume within grafts, the lack of phasicity or periodicity in the central veins is less specific for central occlusion in patients with a graft than in patients with an AVF.

Diagnosis of Pathology

Pathology associated with AVFs includes stenosis, thrombosis/occlusion, pseudoaneurysm, peri-fistula fluid collections, and arterial steal. Two features characterise AVF stenosis; visual narrowing as assessed on greyscale imaging and an increased ratio of the PSV at the stenosis as compared to the PSV measured 2 cm upstream from the stenosis. The diagnostic criteria vary in accordance with the location of the stenosis. A juxta-anastomotic stenosis is defined by the following criteria: (1) location within 2 cm of the anastomosis, encompassing the feeding artery and the draining vein;22–23 (2) visible narrowing; and (3) a PSV ratio of greater than or equal to 3:1 (Fig. 7-12). A draining vein stenosis is defined by visible narrowing and by a PSV ratio of 2:1. Stenoses of the feeding artery are rare but can occur and are characterised by visible narrowing and a PSV ratio of 2:1. Poststenotic waveforms with a delayed systolic upstroke (tardus parvus) may suggest proximal arterial stenosis and prompt further direct evaluation.

Stenoses associated with an AVF are listed in order of decreasing frequency by location as follows: the juxta-anastomotic region, the draining vein, the feeding artery, and central veins. Stenoses are clinically relevant since they can be associated with subsequent thrombosis. If thrombosis occurs, it is usually visualised within the vessel lumen on greyscale imaging (Fig. 7-13) and confirmed when no colour or spectral Doppler flow is identified within the affected portion of the AVF or draining vein. If the thrombus is nonocclusive, slow flow or a trickle of flow may be identified with power Doppler.

Pseudoaneurysms may develop anywhere along the course of a fistula secondary to suboptimal compression after cannulation. Colour Doppler of a pseudoaneurysm shows a typical ‘yin/yang’ circular flow pattern (Fig. 7-14). There may be ‘to-and-fro’ flow in the pseudoaneurysm neck. If the pseudoaneurysm thromboses, there may be clot partially filling the space or total lack of flow on Doppler. Avascular, hypoechoic lesions usually represent post access or post procedure haematomas and can be seen associated with fistulas or grafts (Fig. 7-15). Fluid collections complicated by echogenic foci with ‘dirty’ shadowing suspicious for gas may represent abscess in the appropriate clinical setting.

Graft abnormalities are similar to those described above for AVFs, with slight variations and differences in the diagnostic thresholds and criteria. Studies have shown that grafts with decreased blood flow are at increased risk for thrombosis.24–26 The four sonographic criteria utilised to characterise graft stenosis are as follows: (1) visual luminal narrowing on greyscale imaging; (2) a high-velocity jet on colour Doppler; (3) a PSV ratio greater than 2:1 for the venous anastomosis¹⁷ or draining vein; and (4) a PSV ratio greater than 3:1 for the arterial anastomosis. The most common sites of graft stenosis are in decreasing order of frequency as follows: the venous anastomosis, draining vein, intragraft, arterial anastomosis, and the central veins.²⁷ Graft thrombosis is characterised by echogenic material identified within the graft lumen and by lack of flow within the graft (Fig. 7-16); power Doppler should be utilised to exclude slow flow.

The formation of pseudoaneurysms along grafts is relatively common with typical ‘yin/yang’ flow pattern seen with colour Doppler within the cavity or ‘to-and-fro’ flow at the neck. Degeneration of graft synthetic material is unique to grafts. At ultrasound, irregularity of the graft wall can be seen and may be associated with numerous pseudoaneurysms along the length of the graft wall (Fig. 7-17). Arterial steal is defined as flow reversal in the native artery caudal to the anastomosis and may be associated with hand pain and paraesthesias that may worsen during dialysis. If severe, ischaemia or tissue necrosis of the fingers can occur. However, not all flow reversal is symptomatic. Sonographic evaluation demonstrates reversal of flow in the distal artery and rarely shows an arterial occlusion (Fig. 7-18). Brief manual compression of the graft will usually demonstrate a change in the reversed arterial flow, yielding an antegrade high-resistance flow pattern toward the hand.²⁸

Conclusions

The sonographic evaluation of the haemodialysis access has become a mainstream diagnostic technique, and it continues to gain importance as the number of patients with renal failure progressively increases. Doppler can aid in the selection of the optimal access type and location for fistula or graft creation in these patients. Once an access is surgically created, sonography is useful to determine maturation of the access for successful use in haemodialysis. Furthermore, it can identify the common complications associated with these accesses.

REFERENCES