Central Venous Stenosis Associated With Arteriovenous Access

Introduction

Arteriovenous (AV) access for dialysis is established by connecting an artery and a vein together to create the superficial, high-flow conduit necessary for easy access and efficient dialysis. This is ideally done by connecting an anatomically or surgically superficial vein to an artery, but can also be done by interposing a prosthetic conduit between the in-flow artery and out-flow vein. For success, there must be sufficient in-flow from the artery to provide enough flow (usually at least 400–600 cc/min) and unobstructed venous out-flow to the level of the atrium. Categorized according to treatment strategy, out-flow stenosis can develop in three general areas: peripherally, defined as distal to the cannulation site, venous anastomosis in a graft, and brachial and axillary veins lateral to the thoracic outlet; centrally, veins central to the thoracic outlet; and in the thoracic outlet itself ( Fig. 40.1 ). Endovascular interventions are the first line of therapy for both central and peripheral stenoses or occlusions as they are surrounded by soft tissue and relatively fixed bony structures. Alternatively, the veins in the thoracic outlet run in close proximity to bony structure and cross joints, and thus, surgical or hybrid techniques are a required part of the armamentarium. Stenoses can develop for several reasons. In general, the end result histologically is intimal hyperplasia which can result from intimal injury due to several factors: turbulence or high flow, a stent, indwelling catheter or pacemaker, extrinsic compression at the costoclavicular junction, thoracic outlet, or between sternum and aortic arch vessels, and/or compliance mismatch at the venous anastomosis.

The subclavian vein enters the neck and thorax at the venous thoracic outlet. Anatomy is critical. The vein is anterior, passing by the costoclavicular junction (CCJ). The CCJ is formed by the first rib inferiorly, and clavicle, critically involving the subclavius muscle and tendon and costoclavicular ligament superiorly and anteriorly ( Fig. 40.2 ). The anterior scalene muscle and associated phrenic nerve abut the vein posteriorly, but these do not insert on the clavicle and a true, bounded space is not present. The subclavian vein is vulnerable at this point, leading to injury and thrombosis, especially in those with muscular development and/or vocation or avocation with arms overhead. This situation, independently described in 1875 by Paget and in 1884 by Von Schroetter and officially coined Paget-Schroetter syndrome in 1934 (also termed effort thrombosis based on typical risk factors), is the venous form of thoracic outlet syndrome (VTOS).

Although the issues surrounding VTOS have been understood for decades, it is only in the past 10 years or so that it has been recognized that those with AV access are susceptible to the same problem. The true cause of this is unknown, but it seems logical to assume that the vein is vulnerable in this location in essentially everyone. Even if only a small amount of narrowing occurs, adding the very high flow rates of liters per minute across this area would seem to predispose toward intimal hyperplasia; once this situation begins, a positive feedback loop is created ( Fig. 40.3 ). Why does this matter? Decades of experience with conventional VTOS have definitively shown that angioplasty and stenting do not work in this situation, and that lasting relief will not be obtained without formal decompression of the external bony compression in this area, usually by means of first rib resection. It seems reasonable to extend this to patients with AV access-associated out-flow stenosis at the CCJ; for lasting relief, the bony compression must be removed.

This chapter will address complications during endovascular or hybrid treatment of central venous stenoses, defined as those occurring at the bony costoclavicular junction and centrally in the innominate vein to atrium.

Presentation, Diagnosis, and Treatment

Any clinically relevant obstruction to out-flow will, by definition, create increased venous pressure in the arm. This increased pressure will do several things. First, it will cause swelling. AV access creation is often associated with edema of the involved extremity for the first few weeks after access creation because of the inflammatory response from surgery alone. This is usually minor. Severe or persistent swelling lasting longer than 2 to 3 weeks should raise concern for underlying central venous stenosis. If extreme, brachiocephalic or superior vena cava stenosis can be associated with facial edema, shortness of breath, difficulty swallowing, hoarseness, and headaches. Collateral veins at the shoulder and anterior chest wall are suggestive of underlying central venous stenosis. In addition, the increase in pressure can cause excessive or prolonged bleeding after decannulation, and if out-flow restriction is enough to decrease overall flow, increase recirculation as the “clean” blood “recirculates” back into the machine through the out-flow needle. In this situation, on physical examination, the access may feel pulsatile.

With experience, the diagnosis of out-flow stenosis should be extremely straightforward. A patient who presents with late, gradual arm swelling, or immediate severe swelling, has out-flow stenosis in all but extremely unusual cases. Even if swelling is not present, complaints of post-decannulation bleeding, high recirculation rates and/or inefficient dialysis, or physical findings of a pulsatile fistula and/or visible shoulder or chest wall collaterals are likely to present this problem. Out-flow stenosis in the periphery can sometimes be distinguished from more central stenosis by physical examination. “Popeye arm,” massive swelling in the entire arm ( Fig. 40.4 ), shoulder, and chest wall collaterals, and any evidence of superior vena cava syndrome suggest that the problem is more than at the venous anastomosis and probably located more centrally. Interestingly, we have seen many patients with CCJ stenoses who complain of chest wall pain, often worse with dialysis, which is perhaps a specific, although not sensitive, marker of this problem.

The diagnosis of central venous stenosis is often clear based on physical examination alone. Ultrasound may show decreased volume flow or blunted venous waveforms but is not an accurate identifier of more central stenosis. Cross-sectional magnetic resonance imaging (MRI) or computerized tomography (CT), even with modern techniques, is less useful because of artifact created by the bones in this area. In addition, it is difficult to obtain images with the arm abducted. Although CT venography can diagnose central venous stenosis, it is limited by high contrast loads and expense, in addition to the problems previously mentioned. In contrast, conventional venography with fistulagram offers the opportunity to diagnose and potentially to treat an area of stenosis. We find it useful to image this area with full-strength contrast and breath-hold. If no lesion is immediately seen, repeat imaging is performed with the arm abducted. The presence of collaterals is essentially always pathognomonic ( Fig. 40.5 ). IVUS has been shown to identify areas of stenoses or residual stenosis after venoplasty or stenting that are not evident with conventional venography, although this technique may over-estimate stenoses that are probably not clinically significant.

We feel that central venous stenosis causing impending access failure, significant disability, or rapidly progressive symptoms should always warrant open or endovascular intervention. Although no data exist to support this statement, a randomized trial would not seem to be cost-effective. We believe stenosis in this area, especially if causing symptoms, predicts fistula failure and, even when failure does not immediately occur, leads to significant discomfort, at times massive cosmetic problems, and functional disability.

Treatment of central venous out-flow stenosis should be determined based on the location of the problem, specifically with regard to whether the CCJ is involved or not. In cases of venous stenosis central to the CCJ, balloon venoplasty is the first option and is associated with 70%–100% initial success rates and 1-year primary assisted patency rates of 86%. While often successful, recoil is effectively and easily treated with stenting. If the problem is at the CCJ, however, decades of experience with VTOS shows that lasting relief will not be achieved without decompression of this area via removal of one of the bones constricting the vein, most often the first rib, and often combined with hybrid endovascular approaches to restore venous patency (see Fig. 40.2 ).

Complications

Based on the arguments mentioned and knowledge from decades of outcomes from treatment of venous TOS, we treat patients with symptomatic stenosis at the thoracic outlet with a combined surgical and endovascular approach: decompression of the bony thoracic outlet plus fistulagram and endovascular intervention.

Bony decompression can be performed in several ways. Most commonly, the first rib is removed. The anterior location of this compression must be kept in mind, along with the substantial role that the subclavian muscle and tendon undoubtedly play. An infraclavicular approach allows for full vascular exposure and mobilization of the subclavian vein, with resection of the muscle, and is performed with the patient supine, allowing a fistulagram and intervention at the same time. Thompson et al. have described the addition of a supraclavicular incision to allow resection of the entire vein in patients with VTOS (paraclavicular approach), but we do not feel this is necessary in the average case. The transaxillary approach is cosmetically pleasing but does not allow safe vascular reconstruction and precludes concomitant endovascular intervention. The Molina approach is a medial extension of the infraclavicular incision to include a partial sternotomy involving the first interspace for full venous exposure and more central access and is very useful if surgical intervention is required ( Fig. 40.6A,B ). If necessary for increased exposure, claviculectomy can also be performed with surprisingly little long-term morbidity.

The second step in treating these patients is endovascular intervention. What can go wrong?