It is very important to be clear and methodical about the route and safety precautions to adopt once a significant stenosis is identified on angiogram and indication for dilation rationalized. On rare occasions, dilation can result in complications like severe vessel rupture. A cautious approach ensures that the right tools, measures, and skills are available to handle serious complications when they do happen. This entails the interventionist being able to easily deploy an occluding balloon in an afferent artery to arrest flow into a ruptured vessel segment and leaving enough length between the introducer-sheath tip and stenosis to allow placement of a stent. The arterialized vein of an AVF is preferably accessed in a segment that can be most easily punctured by palpation and as far away from the stenosis of interest as possible. Dilation of peri-anastomotic stenoses involves placing a guidewire in the proximal artery, which is not always an easy maneuver in forearm AVFs.

The vessel cannulated for diagnostic angiography may not always be the best route for deploying the materials required for dilation particularly when this happens to be the brachial artery. In the case where the vein or graft is punctured, it is necessary to subcutaneously tunnel the needle 1–2 cm away from the wall of the vessel before entering it. This avoids direct access to venipuncture that can potentially lead to secondary pseudoaneurysm formation. It also makes hemostasis of the puncture site by compression at the end of the procedure easy without risking access lumen occlusion. Flat or low flow and therefore not easily palpable AVFs are cannulated after placing a tourniquet at the mid-upper arm or under echographic guidance.

An 18-G cannula and 0.035-in. angled-tip or “J” hydrophilic guidewire make up the primary tools required to cannulate the majority of accesses and allow placement of an introducer-sheath (usually 5–8 F depending on the size of dilating balloon to be used).

Angled-tip hydrophilic guidewire can be steered across most stenoses but should be avoided in heterogeneous segmental stenoses. Its passage through long occlusions and stenoses adjacent to aneurysmal dilatations is also largely challenging if not impossible.

Heterogeneous stenoses, both long and short, are encountered only in arterialized veins which have never undergone dilation. They are due to pre-dialysis venipuncture-induced endoluminal trauma, transient thrombosis, and partial spontaneous recanalization (Fig. 10.5a–d). A hydrophilic guidewire in this situation is more likely to false track, create a false lumen, and cause irreversible vessel wall dissection. Dissections antegrade to blood flow are usually more sinister. A more steerable and atraumatic guidewire, in the like of the 0.014-in. prototypes used in renal artery stenosis or coronary angioplasties (Spartacore, BMW, Whisper, etc.), supported by a 5-F vertebral diagnostic catheter stands a greater chance of passing through such stenoses.

The technique is to place the vertebral catheter mounted on a hydrophilic guidewire in contact with the stenosis and then to maneuver through it the 0.014-in. guidewire several times until its tip is in the axis of stenosis and traverses it (Figs. 10.6a–c and 10.7a–e). The 0.014-in. guidewire has a very soft and flexible tip which can penetrate stenotic segment mm by mm before hitting against a new area of resistance without causing trauma. The 5-F vertebral catheter should follow the trajectory of the guidewire tip and be maneuvered by subtle changes of angulation so that the guidewire can be gently steered through the entire stenosis at alternating angles along the axis of the access lumen. Some interventionists prefer railroading a 3-F microcatheter across the 5-F catheter to provide better support for the guidewire.

Passage of the guidewire through the stenosis can be facilitated by a simple technique of stretching the overlying skin and vein between two fingers in order to modify its morphology and angulation of the stenosis. Once the guidewire has completely crossed the stenosis, the catheter is advanced beyond the stenosis, and the initial 0.014-in. guidewire is replaced by an angled-tip 0.035-in. metallic “safety” guidewire which is sturdier, more secure, and therefore unlikely to slip out. Failed attempts at crossing the stenosis with a 0.014-in. guidewire require a change of strategy which should include attacking the stenosis from a different direction or location (femoral vein in the case of central vein stenosis or brachial or distal radial artery). As a last resort, a 0.035-in. straight-tip hydrophilic guidewire can be used to attack the stenosis, preferably in a direction retrograde to blood flow. A vessel wall dissection in the opposite direction to blood flow is less likely to result in loss of the access by thrombosis and may thus leave the option open for surgical intervention.

The maneuvering of an angled-tip hydrophilic guidewire through a stenosis adjacent to an aneurysm or pseudoaneurysm can be challenging. The guidewire tends to hit a dead end and coils up inside the aneurysmal formation. A straight-tip guidewire particularly has an advantage and is preferred as it can be more easily maneuvered into the neck of the aneurysm and aligned in the axis of the stenosis. Compression and stretching of the aneurysm are also useful techniques that often help get the guidewire across the stenosis (Fig. 10.8a, b). A rather exceptional alternative but workable technique is to advance the vertebral catheter without a leading guidewire across stenosis found between two aneurysms. These stenoses are usually soft and can accommodate the tip of a 5-F catheter. The guidewire should follow once the catheter is in the lumen of the stenosis. It must be cautioned that this is perhaps the only instance when the golden rule that a catheter should never be advanced into any lesion without a leading guidewire is breached.

An undulating or tortuous arterialized vein segment can be traversed by an angled-tip hydrophilic guidewire with a bit of effort. The guidewire initially assumes the curvature of the vein as it is advanced but usually straightens the vessel undulations once it reaches the straight normal segment further downstream.

Segmental occlusions should be attacked with a vertebral catheter/straight-tip hydrophilic 0.035-in. guidewire pair, preferably against the direction of blood flow whenever this is possible. The trick is to prod at the occlusion and find the true lumen either upstream or downstream depending on the direction of attack in relation to blood flow. The risk of failure and creating a false guidewire track in the process is in the order of 20 % in chronic occlusions, whereas technical failure in passing stenoses is rare in well-experienced hands.

Once the guidewire has crossed the stenosis, its tip should be placed either in the brachiocephalic trunk, superior vena cava, contralateral internal jugular vein, inferior vena cava, or hepatic vein. The guidewire tip should never be left in the right atrium or ventricle during interventions on outflow vein stenoses where it is likely to induce arrhythmias or cause potentially fatal endocardial and myocardial trauma during forceful manipulations and torquing. A number of such complications have previously been reported.

All these tips and tricks on maneuvering around stenoses should however not be a substitute for steadfast and adequate hands-on training on real-life cases and supervision by well-experienced seniors.

The next phase after steering a guidewire through a stenosis is balloon dilation. Dilation is potentially a very painful procedure, and it is essential to administer adequate analgesia as patients exhibit wide variations in pain threshold. Whenever possible, the stenosis and its surrounding area should be infiltrated subcutaneously with 1–2 % lidocaine. Administration of local anesthetics is not technically possible during dilation of deeply seated stenoses found on the medial aspect of the upper arm, cephalic arch, axillary, and upper chest areas.

The next consideration is the choice and sizing of the dilation balloon (diameter and length). If the aim of dilation is a perfect result, then no residual stenosis should remain such that from the completion angiogram it is difficult to tell where along the access the initial stenosis was located (Fig. 10.9a–c). In real practice, complete absence of residual stenosis is more ideal as some stenoses tend to reform immediately once the balloon is removed, a phenomenon known as elastic recoil. There exist three basic rules when it comes to selecting the appropriate dilation balloon size: the balloon diameter should be 1–2 mm larger than that of the reference vessel adjacent to the stenosis, high pressure balloons need to be readily available to replace conventional balloons when results are not satisfactory with the latter, and covered stents of different sizes should be kept in stock in anticipation of any major vessel rupture.

Dilation of venous anastomotic stenoses in prosthetic grafts (Goretex®, Flixene®, Propaten®, etc…) is perhaps the simplest form of dilation and shall be used as an example for discussion below. Prosthetic grafts are generally 6 mm in diameter. Venous anastomotic stenoses are usually dilated at a minimum with a 7 mm or more commonly with 8 mm dilation balloons. Rarely, a 10-mm balloon may be considered when the downstream vein segment is large enough to accommodate it. The standard balloon length is 4 cm. Shorter versions tend to slip out of the stenosis grip during inflation. The correct positioning of the balloon with respect to the stenosis needs to be verified on fluoroscopy, and the right placement of the guidewire in the lumen of the main vein rather than a smaller tributary or vessel wall should be verified by angiography. The balloon is connected to a manual inflation device or manometer which is filled with iodinated contrast diluted to a minimum 75 % by saline. The manometer has a needle gauge which indicates the pressure generated in the balloon lumen on inflation. Dilation is always performed under fluoroscopic guidance until there is complete effacement (elimination) of the balloon waist, indicating complete dilation. Balloon inflation transmits an incremental radial force around the stenosis tearing away the fibrotic tissue and potentially causing profound pain. The balloon rated burst pressure (RBP), which is the inflation pressure beyond which the balloon bursts in vitro, must be considered beforehand so that change to a higher pressure balloon can be made when full effacement is not obtained at the RBP of conventional balloon.

Up to 2012, there are on the international market three types of the so-called ultrahigh pressure (UHP) balloons that can take in inflation pressures exceeding 25–30 atm although their official RBP is slightly lower: Mustang® (Boston Scientific) and Dorado® (Bard) can usually be inflated to 25 atm and the Conquest® (Bard) to 30 atm. Blue Max® (Boston Scientific) and Powerflex Extreme® (Cordis) are relatively older UHP prototypes and used to be favored in a not too distant past being the first balloons that broke the 20-atm UHP defining barrier. Blue Max® is still used in the dilation of anastomotic and juxta-anastomotic stenoses in forearm AVFs. It has a shorter and more compliant tip-to-end-of-balloon length (balloon shoulder) than Mustang or Conquest and hence ideal for anchorage in the first centimeter of the venous post-anastomotic segment of radial–cephalic AVFs.

Globally, based on published literature, about 25 % of AVF stenoses require UHP balloons for full effacement [7]. Whether an ultrahigh pressure balloon is used for primary dilation of stenoses or only after failure of conventional balloons (RBP less than 15 atm) is at the interventionist’s discretion. First-line use of UHP balloons saves time but does not necessarily provide better outcome when compared to conventional balloons [8]. They are more expensive, less compliant, harder to deflate, and require sometimes larger introducer-sheaths.

The usual ultimate aim of dilation is 100 % immediate success as denoted by full effacement of the waisting on the balloon. According to a randomized control trial, good angiographic result is more achievable with prolonged balloon inflation (3 min) than standard immediate balloon deflation [9]. However, beyond immediate cosmetic results, prolonged balloon inflation does not confer any overall long-term survival advantage to dialysis accesses. Standard inflation and immediate deflation are on the contrary very suited for a heparin-free protocol as prolonged balloon inflation without heparinization (usually 3,000 units of heparin in one bolus) is associated with a higher rate of thrombus formation though this can be mitigated by frequent saline flushing (1–2 mL every minute) of the outflow vein.

It is sound practice to compress the arteriovenous anastomosis during balloon deflation to prevent the sudden transmission of high pressure flow around the weakened dilated zone and hence minimize the risk of venous rupture. The skin overlying the dilated segment should be examined for swelling or hematoma formation, suggestive of venous rupture, which is usually accompanied by painful sensation.

A post-dilation angiogram is obtained after balloon withdrawal with guidewire kept in situ. In the case of forearm AVFs with peri-anastomotic stenoses, contrast is injected either through the brachial artery if it was punctured for diagnostic angiography or through a catheter advanced over the guidewire into the proximal radial artery.

Four possible scenarios may play out post dilation:

Vascular rupture at arteriovenous anastomosis is the rarest but probably the most disastrous complication of balloon dilation. The narrow anastomotic chamber can accommodate for balloon tamponade but does not favor effective covered stent placement when the former measure fails. Access loss at this stage is inevitable. For the greater majority of such cases, bleeding control rather than access salvage should be the main aim of therapy. Hemostasis can be achieved by coil embolizing both the proximal and distal arteries of radial–cephalic AVFs. In the case of accesses fed by the brachial artery, the balloon across the ruptured anastomosis should be reinflated and the patient transferred urgently to theater for surgical repair of the artery.

In the cases where there is sufficient length of post-anastomotic brachial artery stump, this can be cannulated retrogradely and a covered stent deployed in the brachial artery to straddle the ruptured anastomotic segment.

Dilation-associated vascular rupture can occur at any anatomical location within the access, but is more frequently seen in the radial artery, juxta-anastomotic and transposed segments of veins, and at the cephalic arch [12]. It has never been described at the level of the subclavian or brachiocephalic veins.

Guidewires may induce venous spasm, just like a dilation balloon inflated across a point of vein bifurcation can do the same, even on the uncatheterized venous branch. It is not always easy to differentiate spasm from occlusive wall dissection though the former always responds to low-pressure (2 atm) balloon dilation or propagates further downstream or upstream into the proximal artery. The introducer-sheath is known to cause spasm at its point of entry into the vein. Such spasms are more frequently seen in newly created AVFs, and they can become occlusive. They sometimes develop on a relatively long venous segment and may cause pain. They may be eventually the only explanation to recurrent difficulties in cannulation for dialysis.

The AVF in this case may need to be cannulated from a different location to allow the introduction of a balloon to recanalize the occlusion and reestablish flow since local injection of vasodilators like nitrates is usually ineffective. Older AVFs, usually more than 1 year, rarely develop such spasms as their walls are thicker and more rigid from arterialization.

As a routine, the completion angiogram should be read with meticulous attention from anastomosis to central veins before removal of the introducer-sheath to ensure no subtle anomaly that may compromise the function of the access is missed and left unrectified. The presence on clinical examination of the access of a faint thrill or abnormal hyperpulsatility should raise the suspicion of an anomaly and prompt even more detailed review of the completion angiogram for corroborative lesions. The interventionists should put themselves in the place of dialysis nurses who will have to cannulate the access the same or ­following day.

Once the introducer-sheath is removed, the cannulation tunnel, but not the wall of the access, is compressed to stop bleeding. Manual compression rarely lasts longer than 10 min in cases where introducer-sheaths smaller than 8 F are used. Punctures made by larger introducer-sheaths, in patients on anticoagulants and during all thromboaspiration procedures, bleed longer and should be closed using a U-shaped suture technique subtending a cut end of the introducer-sheath dilator [13] (Fig. 12.6a–f). Any swelling that forms immediately around the dilated segment should be carefully examined to make sure it is due to a transient hematoma arising from the dilation and not an ongoing vascular leak causing the skin to harden.

A few patients do complain of diffuse or localized pain once all materials have been removed. In the majority of cases, the pain is not related to any rupture and can be managed with topical, oral analgesia, or even local anesthetics. Such cases should be discussed as to the course of analgesia and monitoring and passed on to the attending nephrologists.

The use of heparin in routine dilation is usually not necessary to prevent thrombosis after interruption of access flow by the inflated balloon, but it remains interventionist and lesion dependent. All it takes in some patients is 30 s of balloon occlusion for thrombi to form downstream. These thrombi are very fresh and can thus be easily aspirated through a 6-F catheter or if small in size can be discretely pushed into the pulmonary circulation. The simplest and safest way to prevent access thrombosis during dilation is to flush boluses of saline (in the order of a few milliliters) at 1-min intervals through the introducer-sheath side port. It is even more imperative to flush the immediate outflow vein when the balloon is left inflated for a few minutes like during balloon tamponade. Some interventionists more concerned about access thrombosis independent of the complexity of the case prefer administering 2,000–3,000 IU of heparin particularly when the patients are not on any oral anticoagulant. Great caution should be exercised with the adjunct use of heparin during transbrachial artery puncture or procedure.

The use of oral anticoagulants or antiplatelets post-dilation for better outcomes to date is not supported by any clinical study. Randomized control trials however have shown that there is a role for antiplatelets like clopidogrel and dipyridamole in maintaining early primary patency of newly created AVFs and prosthetic grafts [14, 15].

These venous stenoses are characterized by their central position and slit-like opacification surrounded by the concave lumps on each side which are the hypertrophic valves, in contrast to the full opacification seen beyond the valves (Fig. 10.14). Passing a guidewire across such stenosis is as challenging as getting it through stenoses associated with aneurysms. The trick is to torque the stiff end of a reversed hydrophilic guidewire through the slit. There is always a small risk of subintimal dissection and false guidewire tracking.

Retrograde cannulation of either the cephalic or basilic vein in the lower third of the upper arm is usually performed to access venous stenoses located in the forearm using the catheter–guidewire pair. It often does produce some surprises. The venous anatomy at elbow consists of both superficial and deep veins connected by a perforating vein. The guidewire may be wrongly steered into a deep vein, from where it becomes impossible to find the way to the stenosis of interest in the superficial vein. The trick is to opacify the stump of the cephalic vein that leads to the stenosis by performing local test angiographies with a tourniquet applied near the shoulder. The cephalic vein segment that needs to be catheterized may sometimes be located more laterally or higher than the interepicondylar line. Test angiographies from the brachial artery (in the case where it has already been retrogradely punctured) may help give a better definition of elbow venous anatomy and guide the direction to take. A combination of 5-F vertebral diagnostic catheter and hydrophilic guidewire guarantees the best chance of reaching and crossing the stenosis.

The general rule for dilating accesses with multiple significant stenoses is to start with the most central or downstream (outflow) stenosis and end with the stenosis closest to the anastomosis (inflow). Arterial stenoses should be dilated after venous stenoses. The reason is that any dilation can result in venous rupture that is usually controlled successfully by prolonged balloon inflation. If an upstream stenosis is then dilated (i.e., a stenosis located between the arteriovenous anastomosis and the initially dilated venous stenosis), inflating the balloon stops flow considerably, decreases downstream pressure and helps seal the leak. On the contrary, while dilation of a downstream stenosis (i.e., a stenosis located between the initially dilated venous stenosis and central veins) also stops fistula flow, it nonetheless considerably increases pressure in the upstream segment and facilitates reopening of the previous breach.

At least 50 % of venous stenoses fully dilate at a balloon inflation pressure of 15 atm according to one published report, which implies that any type of balloon, even the most basic, can be used for primary dilation of venous stenoses [7]. The same study also showed that 20 % and 10 % of stenoses in AVFs and prosthetic grafts, respectively, require a balloon pressure above 20 atm to fully dilate. It is therefore indispensable to also have in the intervention armamentarium ultrahigh pressure balloons. The Conquest® balloon (Bard) can withstand inflation pressure greater than 30 atm, but higher pressure generation is limited by the manometer designs which cannot generate pressures above 30 atm. There is also a theoretical risk of balloon rupture (rare but always possible). Trerotola found a way round the limits of current manometers by substituting them with a 1-mL polycarbonate syringe which can manually generate pressures above 30 atm in Conquest balloons [16]. Conquest balloons, however, have three shortcomings on top of their high price: they are rigid and poorly flexible, take a longer time to deflate, and their long shoulder makes them unsuitable for dilating anastomotic stenoses in forearm fistulas since the tip often encroaches into the distal radial artery. They only come in 5- to 12-mm diameter sizes. Their larger 14- to 16-mm diameter size balloons are called Atlas® and the smaller 4-mm version Dorado® from the same manufacturer. These latter two Conquest-derived balloons, unlike Conquest, are less able to take on ultrahigh pressure. A new UHP balloon, Mustang®, from Boston Scientific, gives as good a result as Conquest in the 3- to 7-mm diameter category. Mustang however cannot tolerate inflation pressures above 30 atm for balloons sized above 7 mm. Its advantage over Conquest is that it is extremely flexible and hence the dilation balloon of choice for stenoses located at acute angles like the arteriovenous anastomosis.

The size (diameter) of the balloon selected for a particular dilation is based on both the location of the stenosis and the access vintage. Arterial stenoses in the subclavian artery require 7- to 9-mm balloons, while those in the axillary artery 7–8 mm, brachial artery 6–7 mm, upper forearm radial artery 5–6 mm, lower forearm radial artery 4–5 mm, and ulnar artery 3–4 mm.

Venous stenoses in forearm cephalic and basilic veins need by standard 6-mm balloons near arteriovenous anastomoses of new AVFs and up to 9 mm in older AVFs and repeat cases. At cannulation and aneurysmal segments of forearm, AVFs balloons as large as 12 mm may be used.

Cephalic and basilic vein stenoses near the anastomoses of upper arm AVFs, as a precaution, are dilated with 6-mm balloons on first dilation. At cannulation and transposed segments of upper arm AVFs, the starting balloon size is 7 mm, and this can go up to 12 mm in older AVFs. Cephalic arch stenoses require 7-mm balloons in recent accesses though this can be increased to a maximum of 12 mm depending on the age of the accesses.

Subclavian vein stenoses are dilated with 10- to 14-mm balloons, while brachiocephalic vein stenoses require 12- to16-mm balloons.

Balloon rupture can occur at inflation pressures lower than the stated RBP as a result of manufacturing defects or when the balloon is trapped and damaged by intraluminal calcification seen more commonly in arteries than veins. It becomes a more frequent event once inflation pressure surpasses the RBP. In practice, this index is systematically underestimated for legal reasons as most balloons can tolerate pressures which are 50 % above their RBP (27 atm against a stated RBP of 18 atm for instance) though this goes totally against the manufacturers’ recommendation.

On first notice of balloon rupture, all efforts are made to deflate the balloon to the maximum (blood comes into the manometer), so it can be removed with ease over a guidewire through the introducer-sheath. If any difficulty is encountered with pulling out the balloon, the whole ensemble of balloon and introducer-sheath should be removed over a guidewire. A new introducer-sheath is then inserted over the wire. The ruptured balloon should be carefully inspected to make sure the fragments have been removed in toto and no pieces remain inside the vessel. There are reported cases whereby the distal piece of the balloon dislodged and embolized. The balloon tip is radiolucent and is therefore difficult to locate on angiography or fluoroscopy unless it is still lodged onto the guidewire from where attempts can be made to remove it with a goose neck snare (Fig. 10.15a–d). It is however futile and risky trying to retrieve it if it has already embolized into the pulmonary arteries where most venous emboli end and are unlikely to be of a major clinical risk, that is, pulmonary infarction or infection.

Balloon rupture can cause a small puncture injury in the vein wall which can convert to a frank venous rupture. This type of venous rupture should be managed the usual way with prolonged balloon tamponade and occasionally stenting.

Deliberate balloon rupture is rather uncommonly induced when it becomes impossible to deflate the balloon. A fine needle is used to gently poke the balloon through the skin under fluoroscopic guidance.