INTRODUCTION

The major limitations of balloon angioplasty involved acute problems (failure to achieve a satisfactory dilatation of a lesion as well as acute complications, including dissection of the vessel) and short- to medium-term restenosis. Failure to achieve a reasonable dilatation occurred in very fibrotic and calcified lesions, and in other vessels related to the bulk of focal disease and elastic recoil of the vessel. Certain anatomic situations were also difficult to treat, including aorto-ostial lesions and bifurcations. With the latter, plaque shift from one arm of the bifurcation to another could only partially be overcome using kissing balloon techniques. Dissections occurred particularly in calcified lesions, eccentric bulky lesions, lesions on bends or in tortuous segments of the vessel and in patients with very diffuse disease.

Although stenting has become the default method of PCI, the understanding in the early and mid-1980s of the components of acute failure and restenosis led to two major lines of development – atherectomy and ablative methods as well as methods of scaffolding the vessel. The new technologies that emerged included the directional atherectomy device and two methods of ablation, namely laser techniques and rotablation. Although some of the scenarios when these devices were used can now be treated with stents, certain clinical and anatomic scenarios exist where the use of ablative methods are essential to allow optimal delivery and expansion of stents and in some cases they may be the only means of achieving percutaneous dilatation of a lesion. One of the worst clinical situations to occur in the current era favouring direct stenting is the delivery of a stent to a previously unrecognised undilatable lesion. Only experience, recognition of a number of clinical clues that this situation might be met, and the ability to use ablative technologies can prevent this. This chapter will review the use of atherectomy devices and coronary rotablation.

DIRECTIONAL ATHERECTOMY

This is now largely a historical review, as directional coronary atherectomy (DCA) is now used very infrequently. Many of the devices previously available are no longer manufactured but Guidant (Guidant, Santa Rosa, CA) continues to manufacture the Flexi-Cut™ device and the FoxHollow SilverHawk™ device (FoxHollow Technologies, Redwood City, CA, USA) is also available.

Directional coronary atherectomy was developed by John Simpson to enable debulking of lesions. It was argued that excision of atheromatous material would overcome many of the limitations of balloon dilatation alone, allowing for better dilatation of the lesion with less dependence on stretch (and thus barotrauma), more controlled dilatation (and thus fewer dissections), and that there would, as a result, be better acute clinical results with less restenosis. The initial registries appeared favourable, but the results of several randomised trials were disappointing. The major studies of DCA are listed in Table 14.1. In particular, the use of DCA resulted in a relatively high peri-procedural release of cardiac enzymes, and clinical results were either only marginally better or in some cases appeared worse.^13–15 There are several reasons why early registry experience is not always translated into clinical benefit in a multi-centre randomised trial. These include the performance of a trial before the technique has been adequately developed, designing a trial that partially reflects the learning curve of operators with only limited experience of the technique and, perhaps most importantly, the fact that early registry experience is obtained in carefully selected patients by operators who have the greatest experience of the technique. Moreover, the original concepts of why DCA might be favourable might have been countered by the stimulus to restenosis caused by cutting into the plaque.

TABLE 14.1 MAJOR STUDIES OF THE ROLE OF DCA

Although the CAVEAT and CCAT trials did not establish a clear benefit from DCA, possibly because of inadequate debulking and a higher complication rate compared to balloon angioplasty alone, the OARS, BOAT, ABACAS and GUIDE II studies suggested that optimal debulking, with attempts to reduce residual plaque volume to a minimum (aided if necessary by IVUS guidance), would result in lower restenosis rates. A residual angiographic percent diameter stenosis after DCA (with or without adjunctive balloon angioplasty) of <10–15% was aimed for. However, the higher CK release during procedures remains an issue and given that DCA is more technically demanding, the emergence of stenting as an easier-to-use technique with equal or better results has resulted in a major reduction in the number of DCA procedures being performed worldwide. However, when only BMS were available, the Japanese START study suggested that DCA could still be considered as an alternative to stenting as it achieved a lower rate of restenosis with no differences in clinical events.⁸

Several generations of devices were produced by Devices for Vascular Intervention (acquired initially by Guidant Ltd and subsequently by FoxHollow). Following the original AtheroCath®, sequential improvements were seen in the SCA-1™, SCA-EX™, Short-Cutter™, the GTO®, the Bantam™ and the Flexi-Cut™ devices, with improved debulking capability and use even in partially calcified lesions. The original devices needing 10F sheaths and guide catheters gave way to 8F-compatible devices. Currently, only the Flexi-Cut™ device is available (Fig. 14.1). The technique requires stiffer than average guide catheters such as the Viking XT™ range (Guidant) with stiff wires such as the Iron Man™ (Guidant).

Figure 14.1 Directional atherectomy.

The device is inserted into the vessel, rotated until the cutting window is lined up with the side of the vessel containing the bulk of the lesion, the cutter is then drawn back and the balloon on the side of the device opposite to the window is then inflated at low pressure. The balloon stabilises the device and pushes the cutting window against the plaque. The device is then activated and the rotating cup-shaped cutter is advanced, slicing off the atheroma and pushing it forward into the collection nosecone. The balloon is then deflated, the device rotated slightly and further excisions made until adequate atherectomy has been performed. Different sized catheters are available for different vessels and lesions. In smaller vessels, the development of ischaemia during the procedure sometimes requires withdrawal of the device back into the guiding catheter between cuts. This is less of a problem with the Flexi-Cut™ device.

More recently, the FoxHollow SilverHawk™ Plaque Excision System has been introduced (Fig. 14.2), which has the advantage of being 6 F compatible and it does not have a contra-lateral balloon opposite the cutting window but instead a unique hinge design.¹⁶ This potentially reduces barotrauma. Three tip lengths are available with different collection chambers, allowing for a number of clinical scenarios in coronary cases, and the device can be used in vessels ranging from 1.5 to 3.5 mm in diameter. Catheters designed for peripheral vascular cases are also available. The carbide blade allows for excision of calcified plaque. Like the earlier DCA devices, the shaved plaque is stored in the nosecone of the device, allowing its retrieval. These atherectomy devices are the only available tools that allow for retrieval of human atheromatous tissue for research purposes. Another device is the Arrow-Fischell Pullback Atherectomy Device (Arrow Medical Devices Co, Reading, Pennsylvania, US).^17,18

Figure 14.2 The FoxHollow SilverHawk™ device.

Debulking with atherectomy, especially when aggressive, has been associated with a higher rate of vessel perforation (0.5–1.3%) than with ballooning alone, and there is also a small risk of aneurysmal formation during follow up, although this has not been associated with adverse clinical events. Whether this could be a problem if DCA were used before implantation of a DES (some of which are associated with negative vessel remodelling during follow up) is not known. Groin complications were slightly higher when using larger guide catheters.

Although stenting has emerged as the dominant method of PCI, there are still some anatomical situations where stenting is technically difficult or may not achieve optimal results, and stenting itself has produced the problem of how to deal with in-stent restenosis. The latter is due entirely to neointimal proliferation within the stent, but inadequate stent expansion (which thereby reduces the size of the lumen that can accommodate the proliferative tissue) has been recognised to be one of the major determinants of sub-optimal clinical results. DCA enthusiasts have, therefore, continued to use this technique in difficult anatomy, and have also explored the role of debulking prior to stent implantation. Again, early registry evidence was favourable but the randomised multi-centre AMIGO trial could find no benefit from this technique, although benefit was seen in the subset of patients where aggressive debulking (<20% residual stenosis) was achieved prior to stenting. Although certain groups have suggested that the benefits only emerge with aggressive debulking prior to stent implantation, this has become a somewhat academic exercise with the emergence of DES.

The lesions for which some operators continue to use DCA rather than stenting include eccentric lesions, ostial stenoses, bifurcations, in-stent restenoses, and some continue to use the technique of aggressive debulking prior to stenting (especially for left main stem lesions). Its use in these lesions has been reviewed recently.¹⁹ Although the use of DES has seen a major reduction in the use of DCA, these lesions remain difficult to treat and results with DES are less favourable than with more straightforward lesions.

ROTATIONAL ATHERECTOMY (ROTABLATION)

Rotablation is probably a more appropriate name for this technique as it results in virtual emulsification of the atheromatous plaque rather than being a method of excising atheroma. It utilises a high-speed rotating elliptically shaped burr, the front face of which is coated with 2000 to 3000 diamond microchips (30–50 μm in diameter). As the burr meets the plaque, the latter is fragmented into particles 5–10 μm in diameter (smaller than red cells) which are washed downstream, pass through the coronary capillary bed and then enter the venous circulation and are removed from the circulation by the reticulo-endothelial system (Fig. 14.3).

Figure 14.3 The Rotablator: A The Rotablator console and foot-pedal B The Rotablator device – burr catheter and advancer C The Rotablator burr.

There are two major principles of action of this device. The first is through orthogonal displacement of friction, whereby high-speed rotation (at speeds of 140 000 to 150 000 rpm) reduces the longitudinal friction between the burr and the guide wire and eases passage down the vessel. The second is referred to as ‘differential cutting’ with elastic or soft tissue deflecting away from the burr whereas inelastic tissue cannot deflect and is therefore ablated. The technique minimises wall stretch and thus barotrauma and it was hoped that this would result in a lower restenosis rate. However, this has not been found in clinical practice and the deleterious effects of the heat it generates probably counter any advantage the rotablator might have in this regard. Although rotablation is not, then, a means of reducing restenosis, it allows for successful treatment of lesions that cannot be adequately treated by balloon angioplasty. It is not a technique for removal or fragmentation of soft thrombus.

Since its introduction, the technique has changed somewhat. Modifications include the use of a verapamil/nitrate flush solution used down the guide catheter whilst rotablation is performed, a Rotaglide™ lubricant that reduces friction and disperses heat, shorter runs with only gentle advancement into the plaque using a pecking technique that allows for as much coronary perfusion during each run, lower burr speeds, avoidance of drops in speed of >5000 rpm, and up-front use of a glycoprotein IIb/IIIa inhibitor.^20,21 Many operators do not use the latter though until after rotablation in particularly difficult, angulated and heavily calcified vessels because of the small risk of perforation of the vessel, but instead consider its use during the remainder of the procedure once the rotablation is completed. For lesions in dominant right or circumflex coronary arteries, significant bradycardia or heart block can occur. The patients should be well hydrated, atropine pre-treatment should be considered and in some cases it is wise to insert a pacing electrode. Right ventricular pacing electrodes (especially if a glycoprotein IIb/IIIa inhibitor is used) can result in cardiac tamponade if an apical electrode perforates the ventricle, and it is probably wiser to select either a septal position for the pacing electrode or use a flotation pacing catheter.

The burrs come in sizes between 1.25 and 2.5 mm in diameter. They pass along a specialised stainless steel stiff guide wire with a floppy end. The guide wires are harder to torque than other angioplasty wires, and if there is some difficulty placing the wire, a conventional floppier wire can be used with a low profile over-the-wire (OTW) balloon catheter. Having crossed the lesion with the latter, the initial wire can be exchanged for the longer exchange-length rotablator wire. The OTW balloon can then be removed and a burr advanced.

The technique is demanding and must be performed carefully to achieve procedural success without complications. Rotablating at too high a speed and with too much forward force can result in larger particles of atheroma being released and activation of platelets (both of which can contribute to a slow-flow phenomenon and downstream myocardial damage). In addition, excess heat may be generated (a stimulus to restenosis).

MAJOR STUDIES OF ROTABLATION

The major studies of rotablation are listed in Table 14.2.^20–31 The earlier studies compared balloon angioplasty with rotablation (and laser technology in the ERBAC trial) in relatively complex lesion subsets and small vessels. In ERBAC, 685 patients with complex lesions were randomised between rotablation, excimer laser angioplasty or balloon angioplasty. Procedural success was highest in the rotablator arm but there was no difference overall in complications. Further revascularisation at six months was required least frequently with balloon angioplasty. In the COBRA study, procedural success rates were similar in both arms of the study, and restenosis rates were similar although there was more late loss in the rotablator arm. These studies, performed in the pre- and early stenting era, did not provide compelling evidence to use rotablation in the majority of complex lesions (B2 and C category). A number of nonrandomized studies have looked at the role of the rotablator in lesions such as chronic total occlusions, ostial lesions, and bifurcation lesions. In general, there is no evidence to suggest any benefit, although it may be necessary to use the device in individual resistant, fibrotic or calcified lesions. There is no particular advantage of using the device in small vessels (as shown in DART).

TABLE 14.2 MAJOR STUDIES IN ROTABLATION

Although these trials did not suggest routine use of rotablation in such lesions, it is generally accepted that the device has an important role in heavily calcified vessels. Identification of calcium on fluoroscopy does not reveal the nature and site of the calcification in the vessel wall. Studies using IVUS, however, have revealed significant variations, with calcification in the medial and adventitial walls rarely causing a major problem with dilatation of a stenosis whereas superficial calcium with an arc >270 degrees, and especially with a thick concentric band of calcification (‘napkin-ring’ calcification) can be totally resistant to balloon dilatation even at very high pressures. Less extensive calcification can sometimes be dilated but it can be extremely difficult to pass and fully deploy stents in these circumstances. Some operators have attempted to use the Cutting Balloon® (Boston-Scientific, San Diego, CA) in these circumstances although this technique was not designed for this purpose. Others use the FX miniRAIL balloon® (Guidant) that uses the guide wire and an additional wire running alongside the balloon to act as fulcrum points to incise the lesion (allowing dilatation in some vessels at lower pressure) or to act as a pressure point to try and crack calcified segments and thereby achieve dilatation. Whether extensive or relatively localised, superficial calcification is a risk factor for vessel dissection with balloon dilatation and there is a small risk of vessel perforation. This has to be taken into account when considering the small risk of perforation with the rotablator device, although this risk is probably falling with the change of strategy whereby the procedure is most commonly performed with small burrs (see below). There are no randomised trials of patients with heavily calcified lesions undergoing rotablation versus other techniques, mainly because it is impossible to design a trial when the operators know that treatment, with balloon angioplasty, for example, will either fail or given only a modest dilation at best. Prior to the emergence of stenting, many interventionists did not treat these lesions and, instead, patients were referred for coronary surgery. A number of non-randomized series have demonstrated good results with acceptable complication rates.

The issue of the optimal method to use rotablation was the subject of the STRATAS and CARAT studies. Two techniques were considered – a simpler technique utilising sufficient plaque ablation to allow successful dilatation with ballooning at reasonably low pressure (‘facilitated angioplasty’) versus more aggressive debulking using sequential burrs (final burr:artery ratio >0.7) and then finishing with either no ballooning or ballooning at very low pressure (to avoid deep tissue injury and thereby minimise the stimulus to restenosis). These trials did not show the hoped-for lower restenosis rate with the more aggressive technique, which was associated with a higher release of CK and other complications. Nowadays, rotablation using small burrs (burr:artery ratio <0.6) is most often used to remove difficult calcification and to alter vessel compliance prior to ballooning (and nowadays stenting). Use of the larger burrs has reduced significantly and most operators use a single 1.5 or 1.75 mm burr (reserving the 1.25 mm burr for those with a very tight or heavily calcified lesion). There are, however, individual cases that have such deep napkin-ring calcification that sequential burrs are necessary to achieve a satisfactory lumen. If there is any doubt about whether a lesion might be heavily calcified, it is advisable to inspect it first with IVUS. One tip is that if the IVUS catheter cannot cross the lesion, it needs rotablating.

Although stenting can now be performed in some calcified vessels, it may be extremely difficult to pass the stent down to and across the lesion (often requiring support catheters, stiff guide wires and the use of a ‘buddy wire’), and even if stent expansion is possible, full stent expansion may not be. Cases with inadequate stent expansion have a higher restenosis rate than cases with full stent expansion. For this reason, those able to use the rotablator often use the device to partially change the compliance of the vessel prior to stenting thus allowing optimal stent expansion (‘rotastenting’). It is argued by some that inadequate expansion of the stents in the era of DES is less important given that they are designed to minimise neointimal proliferation, but the counter arguments include a concern that difficult deployment of a DES may damage the stent coating and thus reduce its anti-proliferative efficacy. In addition, imperfect apposition of the stent to the vessel wall may be one factor that predisposes to late thrombosis. Although there is no compelling evidence to use debulking devices in general prior to stenting, better results will probably be seen with rotastenting in calcified lesions compared to stenting alone. Further research is necessary but, currently, interventionists fall into those who believe in pre-stent preparation of the vessel with the rotablator in such cases, and those who are inexperienced in the technique who use atherotomy devices (such as the Cutting Balloon®) and very high pressure dilatation of the stent after it is deployed.

Because of sub-optimal results of balloon angioplasty for the treatment of diffuse in-stent restenosis, a number of centres investigated the role of debulking with atherectomy or ablative devices prior to ballooning. Although early registry data was reasonably encouraging (as with the BARASTER registry), and the single centre ROSTER randomised trial suggested benefit with the rotablator with relatively aggressive debulking, the larger ARTIST randomised trial failed to demonstrate any benefit from this technique and there were more complications in the rotablator arm. Although this was in part related to inadequate debulking, most centres do not use this technique routinely for this purpose.

Complications with rotablation are lower with the less aggressive debulking technique but there is an incidence of Q wave infarction (1–1.3%), CK-MB >3 times the upper limit of normal in 4–6%, dissection (10–13%), abrupt closure (1.8–11.2%), the slow flow phenomenon (1.2–7.6%), perforation (0–1.5%) and severe spasm (1.6–6.6%). Many of these can be avoided by correct technique using smaller burrs. If they occur, they can be managed with post-rotablation ballooning and stenting (including the use of covered stents in the case of perforation), and generous administration of intra-coronary nitrates, verapamil or adenosine. If tamponade occurs, emergency pericardiocentesis and sometimes surgical bailout is necessary. If severe slow flow or spasm occurs, it is wise to abandon the rotablator and to use balloon angioplasty and stenting as appropriate, as continued rotablation in these circumstances greatly increases the chances of perforation. Particular care with the rotablator must be used in very angulated or heavily calcified lesions. In these cases it is sensible to start with the smallest 1.25 mm burr and to concentrate on optimal technique. In cases with poor left ventricular function it might be necessary to use an intra-aortic balloon pump prior to starting. The device is not necessary in very soft lesions, it is not designed as a thrombus removal device, and it should not be used in the body of vein grafts with friable disease.

The rotablator device continues to have a niche role but it is not required in the majority of cases. Because of this, training in the technique is difficult, and there are many operators who do not feel comfortable using it on an infrequent basis. For this reason, interventionalists should be encouraged to refer difficult cases requiring the rotablator to those with more experience. If an inexperienced operator is faced with a totally resistant lesion then a more experienced operator who is happy to use the rotablator, if available, should become involved with the case. If an operator experienced in rotablation is not available, it is better in many cases to stop the procedure rather than proceeding with very aggressive high-pressure ballooning (which can have unfortunate consequences). The patient can then be referred on to an interventionist who has the necessary rotablator skills.