Percutaneous coronary intervention (PCI) of chronically occluded coronary arteries represents the greatest technical challenge for the interventional cardiologist [1, 2]. Compared to intervention of nonoccluded stenoses, recanalization of chronic total occlusions (CTOs) requires more operator skill and procedural time, increases radiation exposure to the patient, physician, and catheterization laboratory staff, and consumes significantly greater resources. In no other lesion subset does the procedural success rate vary as much from physician to physician as with intervention in CTOs, and a definite learning curve exists in which technical success rates for individual operators continue to improve after the performance of hundreds of CTO procedures [3]. Moreover, until recently, the clinical benefits of CTO recanalization had not been clearly demonstrated, which, in concert with the technical complexity of the procedure, resulted in many patients with CTOs being either treated medically or referred to bypass graft surgery. There is now an increasing body of published evidence demonstrating that successful percutaneous recanalization of occluded coronary vessels subtending viable myocardium not only reduces angina and improves quality of life, but also improves left ventricular function and is strongly associated with enhanced survival [4-6]. Moreover, long-term patency and freedom from restenosis after successful recanalization of CTOs may be greatly enhanced by implantation of drug-eluting stents. As a result, tremendous interest has recently emerged among interventional cardiologists to learn the principles and develop the advanced skills required to maximize procedural outcomes in chronic coronary occlusions. In the first decade after the introduction of balloon angioplasty, technical success rates for PCI in true CTOs were achieved in only 40–57% of cases, reflecting the reliance on suboptimal guidewires, catheters, and dilatation equipment [7-10]. In the last 20 years, the procedural outcomes of intervention for CTOs have increased significantly as a function of improved guidewires and devices, as well as operator technique and experience, such that successful recanalization of true CTOs may now be achieved in approximately 80% of lesions [4, 5]. In order to achieve consensus on the techniques, methods, and devices required to maximize the likelihood of successful recanalization, an international group of 47 physicians from nine countries was convened in New York City in January 2004, representing many of the world's leading operators and thought leaders in the subspecialty of CTO intervention. Over a 2-day period, collective experiences were evaluated in detail and procedural outcomes were examined though a series of didactic lectures, roundtable discussions, breakout focus groups, and the performance of 14 live case demonstrations of complex angioplasty in occluded coronary arteries. Summarizing the lessons and guidelines generated from this summit, the present article will review the basic principles and advanced techniques required to develop expertise in CTO angioplasty and highlight the major innovations responsible for the current progress being achieved in the percutaneous revascularization of chronic total coronary occlusions. Appropriate patient selection is dependent on the experience of the operator in relation to the clinical and anatomic complexity of the CTO. Traditional predictors of lesion success have included shorter duration of the occlusion (< 3 months), functional total occlusions (those with TIMI 1 flow, which are no longer considered true CTOs), shorter lesion length, presence of a tapered or funneled stump, absence of a side branch at the occlusion site, minimal vessel and lesion tortuosity, absence of calcification, nonostial occlusion location, and a well-formed ipsilateral or contralateral collateral supply clearly delineating the vessel course distal to the occlusion [4, 11-16]. Conversely, the presence of bridging collaterals has traditionally been the strongest correlate of procedural failure [3, 4, 11, 12, 16]. However, with improvements in guidewire technology and procedural technique and experience, even bridging collaterals no longer represent the harbinger of procedural failure as once believed [2, 4, 10]. As a general rule, operators should begin with relatively straight-forward cases (e.g., short occlusions of limited duration, tapered stump present, and absence of bridging collaterals) and advance to more complex cases as they gain experience. For the most experienced CTO operators, the only near absolute lesion-specific contraindication is the absence of a visible distal vessel. Advance planning is necessary to optimize success rates of CTO angioplasty. The operator should review multiple projections of the occluded vessel in orthogonal projections frame by frame, as well as collateral vessels using contralateral injections when required to gain a complete understanding of the anatomy, the entry and exit points, the vessel course and side branches, as well as calcification (which serves as an important marker of vessel boundaries). If the CTO is old and bridging collaterals have formed, it is essential to differentiate the true lumen of a functional occlusion (with intracoronary microchannels, which may be crossed and dilated) from perivascular vasa vasorum or intravascular channels of a bridging collateral, which can easily be dissected or perforated with excessive wire manipulation. For each particular patient, the maximum allowable amount of radiographic contrast media should be defined prior to starting the procedure. To reduce the amount of dye, contralateral injections may be performed through an end-hole catheter inserted distally into the artery to perform superselective angiography. This technique can restrict the amount of contrast to less than 1 cc per injection. All patients undergoing PCI should be treated with at least 300 mg of oral aspirin at least 2 hr prior to the procedure. Because CTO angioplasty is rarely an emergent procedure, pretreatment with a loading dose of clopidogrel at least 6 hr before the procedure in anticipation of stent implantation during the procedure is recommended [17]. Procedural anticoagulation considerations are similar for PCI of nonoccluded stenoses, except that direct antithrombins and glycoprotein IIb/IIIa inhibitors are usually avoided because of the increased procedural risk of perforation. Glycoprotein IIb/IIIa inhibitors may be administered before angioplasty once the guidewire has successfully crossed the lesion and confirmed to be intraluminal. Similarly, the initial heparin bolus may be reduced to achieve an activated clotting time of approximately 200 sec until the guidewire has successfully crossed the occlusion, after which additional heparin should be administered before dilatation to achieve an activated clotting time of 250–350 sec (if a IIb/IIIa inhibitor is not utilized). Postprocedural heparin should rarely be administered following angioplasty, and the vascular sheath(s) removed as soon as the activated clotting time falls below 170 sec (or immediately postprocedure if a closure device is used). Aspirin (81 mg per day) should be prescribed after successful PCI of a CTO, and thienopyridine recommendations are the same as following PCI of a nonoccluded vessel (clopidogrel 75 mg per day for a minimum of 4 weeks after a bare metal stent, 3 months after a sirolimus-eluting stent, and 6 months after a paclitaxel-eluting stent) [18, 19]. Vascular access planning requires consideration of guide catheter support and the likelihood of requiring debulking devices or bifurcation stents, venous sheaths, or an intra-aortic balloon pump. Once the lesion is wired, difficulty passing a balloon across a CTO should be anticipated, necessitating excellent guide catheter support. Femoral artery access is preferred for CTO angioplasty by most operators, with utilization of 7–8 Fr guides for passive support, though 6 Fr catheters may be considered for short occlusions or by operators skilled with active guide manipulation. Larger guide catheters, however, provide the versatility to pass covered stent grafts more easily should a perforation occur, a complication that must be anticipated with PCI of CTOs [20, 21]. If a second angiographic catheter is necessary for contralateral injections, a 4 or 5 Fr catheter can be inserted into the contralateral femoral artery or either radial artery, though 4 Fr access from the ipsilateral groin may be an acceptable alternative by puncturing 1 cm medially and distally to the previously placed sheath [22]. Finally, the radial artery may be an acceptable alternative for CTO angioplasty by experienced operators in selected cases, especially when a guiding catheter no larger than 6 Fr is required, when the distal vessel is visible from ipsilateral collateral flow, when the location of the occlusion is mid or distal, and in the presence of otherwise favorable anatomy [23]. For the left coronary system, extra backup (EBU)-type guiding catheters (Voda left, extra backup, geometric left, left support) are preferable. Judkins-type guiding catheters, which typically preclude deep intubation, are associated with reduced success with hard fibrocalcific occlusions. For the right coronary artery (RCA), left Amplatz 0.75–2 shapes (which in general provide the maximal support) are preferred (especially with a superior or shepherd's crook takeoff), though hockey-stick shapes may be considered for the RCA with transverse or slightly superior takeoffs, or Judkins shapes for inferiorly oriented vessels. Typically, RCA guiding catheters should have side holes to allow perfusion of the sinus node and conus branches during tight seating of the guide. Aggressive manipulation of the guide catheter, or inadvertent deep intubation (which not infrequently occurs with the Amplatz shape) may dissect the ostial right coronary ostium (often requiring stenting), a complication that should be anticipated and recognized before guidewire removal. Guidewire crossing of the CTO is the most technically exacting phase of the procedure and the point at which success or failure is typically determined. There are three steps to crossing a CTO: penetrating the proximal fibrous cap, traversing the body of the CTO to reach the distal fibrous cap, and penetrating the distal fibrous cap. Optimal guidewire selection and technique are critical if procedural success rates are to be maximized. A vast array of guidewires are available, which may be used for CTO intervention, the most popular of which are listed in Table I. Wires designed for treating CTOs can be largely divided into two main groups: polymer-coated (hydrophilic or lubricious) guidewires and noncoated (nonlubricious) coil guidewires. Both groups may be further subdivided into those with nontapered tips (0.014″) and those with tapered tips (0.009″ to 0.010″). Each wire subgroup has inherent advantages and disadvantages when used for crossing CTOs, and each has its proponents recommending routine frontline use. We therefore present the consensus risk/benefit considerations when choosing each wire type. It is important for operators to become familiar with all wire types, but then to choose one or at most two wires for routine use. This is true whether one primarily uses conventional or hydrophilic wires. Most operators use nonhydrophilic-coated guidewires for CTO intervention. Conventional (nonhydrophilic-coated) wires are more controllable (and therefore less likely to dissect) and provide better tactile feel compared to hydrophilic wires. Noncoated coil wires tend to encounter more resistance inside the lumen than polymer-coated wires, but select coil wires (especially the Asahi Intec Miracle Brothers line; 0.014″ tip, available in "strengths" of 3, 4.5, 6, and, outside the United States, 12 g) and the Confianza (also known as the Conquest; 0.009″ tapered tip with 9 or 12 g force) have exceptional torque response even within a fibrocalcific CTO. The greater tactile feel of nonlubricious wires is especially important when attempting to penetrate the distal fibrous cap of a CTO and not create a false lumen. Notably, as the wire tip becomes stiffer, torque response increases, but less tip resistance is transmitted to the operator, making it easier to enter a false channel. Thus, lower-force wires are generally used initially (e.g., Miracle Brothers, 3 g), with progressive use of stiffer, more powerful wires if resistance to penetration is encountered. To minimize tip resistance and select small vascular microchannels within the CTO, some operators prefer tapered tip wires (e.g., the Guidant Cross-It series, which tapers to 0.010″ and comes in progressively greater strengths from 100 to 400, with the 400 corresponding to ∼ 6 g of force, and the Confianza). The technical success of these wires stems from their ability to engage and traverse through luminal microchannels within the occluded segment. In a contemporary PCI registry of 214 CTO revascularization attempts with tapered guidewires, overall technical success was achieved in 76% of patients; in the presence of a visible microchannel, however, the success rates ranged from 81% (incomplete microchannels) to 100% (microchannels with distal filling) [24]. Despite their benefit in penetrating resistant lesions, however, these needle-like tips can also easily dissect and perforate the vessel wall and thus should be used principally by experienced operators. The extreme lubricity of hydrophilic wires underlies both their strengths and weaknesses. Hydrophilic wires typically advance with minimal resistance and tactile feel, even down minute branches and false channels. These wires offer good maneuverability in tortuous vessels and, compared to coil wires, may be steered more easily in a true lumen immediately after a sharp bend. Conversely, in a true CTO, they are more likely to penetrate beneath plaque and dissect than noncoated wires, do not maintain their tip shape as well, and do not offer optimal tip control. Given the lack of tactile feedback, once in a false lumen a hydrophilic wire may be passed for long distances without resistance. This leads to a greater tendency to create large false channels that preclude success. Hydrophilic wires also tend to select small branches and perforate more frequently than noncoated wires. Such end capillary perforations may be difficult to manage, requiring coil embolization to seal [25]. Lubricious wires also more easily enter paralleling bridging collaterals and fragile vasa vasorum, the dilatation of which may result in perforation and tamponade. Finally, if the operator fails to recognize that a hydrophilic wire has entered a false passage or tiny branch and subsequently inflates a balloon, massive perforation, cardiac tamponade, and death can ensue. Therefore, it is imperative to visualize the distal course of hydrophilic wires in at least two orthogonal views and never inflate a balloon distally unless certain the balloon and wire are in the true lumen. Despite these caveats, a significant minority of operators select hydrophilic-coated wires as their wire of choice for CTOs because the reduced resistance to wire passage substantially accelerates the procedure (whether success or failure). The most commonly used hydrophilic wires currently are the Guidant Whisper (the floppiest lubricious wire), the Guidant Pilot (ranging in support from the 50 to the 200 version, with increasing stiffness), the Boston Scientific Choice PT and P2 and the stiffer PT and P2 Graphix, and the Cordis Shinobi (the stiffest hydrophilic wire, but also the one with the best torque response). As with noncoated wires, if hydrophilic wires are selected, initial attempts should begin with floppy wires, progressing to increasingly stiff wires if necessary. Operators who routinely use nonlubricious guidewires often find utility in hydrophilic guidewires when tortuous and/or fibrocalcific anatomy proximal to the occlusion is present. In such cases, the proximal passage may be simpler to negotiate with a hydrophilic wire, which can then be exchanged for a conventional wire for CTO crossing after passing an over-the-wire balloon angioplasty catheter to the point of the occlusion. Soft-tipped lubricious wires such as the Whisper (the least traumatic hydrophilic wire) may also be preferred when a faint channel is visible, consistent with an intracoronary microchannel that may allow easy access to the distal lumen. Care must be employed in this circumstance, however, not to create a false lumen, converting a simple case into a failure. Finally, the Confianza (Conquest) Pro is a hybrid 0.014″ wire that tapers to 0.009″ and is hydrophilic-coated except at the tip, thus reducing the friction as the wire shaft passes down the vessel and through the body of the occlusion while theoretically retaining tactile response at the distal end. Because of its combined stiffness, hydrophilic coating, and tapered tip, this powerful wire (which is available in 9 and 12 g versions) should be reserved for experienced CTO interventionalists. Once a stiff guidewire (whether noncoated or hydrophilic) has crossed the occlusion and has been passed into the distal vessel and the lesion crossed with an over-the-wire balloon dilatation catheter, it should be immediately withdrawn and a floppy-tipped noncoated wire placed distally to minimize the risk of distal wire perforation or dissection. The most common approach to CTO intervention is the use of nonhydrophilic-coated wires introduced through a 1.5–2.0 mm diameter over-the-wire balloon angioplasty catheter (with 6–10 mm balloon length) or an end-hole exchange catheter (e.g., the Cordis Transit, Spectranetics Probing catheter, or Boston Scientific Excelsior). The over-the-wire technique improves wire maneuverability and simplifies frequent wire exchanges or tip reshaping. In general, CTOs should be first approached with the same floppy guidewires used for PCI in nonoccluded lesions. As many as 20% of CTOs that are believed to be rigid or hard may be crossed with floppy wires (often by slightly stiffening the wire tip by bringing the balloon catheter closer to the tip), thus minimizing the risk of dissection. Penetrating the proximal fibrous cap, however, may be difficult when the occlusion is old, requiring progressively stiffer wires. Coaxial wire alignment along the course of the vessel and a shallow (∼ 15–20°) angled curve of the wire tip are recommended to pierce the proximal fibrous cap. A secondary bend may be placed in the wire ∼ 5–8 mm proximal to the primary bend to extend the reach of the guidewire if necessary without increasing the primary wire tip angle. Once the wire breaks through the proximal fibrous cap, it may be exchanged if desired for a softer wire with a slightly greater primary tip bend. After the proximal cap is penetrated, the wire must be passed through the body of the CTO to the distal fibrous cap, which typically requires experience and reliance on knowledge of the natural course of the vessel, as well as collateral visualization of the distal vessel. Lesion calcification or occluded stents serve as guides to the vessel course. When negotiating an angulated segment in the vessel, the wire should be steered toward the inner curve to avoid extraluminal passage (Fig. 1). a: When steering through an occlusion on a bend (typically the proximal right coronary artery), directing the wire to the outer part of the curve typically results in a subintimal dissection (b). The wire should be redirecting toward the inner curve of the vessel (c). Old occlusions (> 3 years) typically taper at the end to form a convex structure, making penetration of the distal fibrous cap problematic. The optimal point for the penetration of convex distal fibrous cap is its center, although the newly created proximal channel often leads laterally. In curved vessels, the optimal point to attempt to perforate the distal fibrous cap is usually on the myocardial mural side, and the majority of these lesions require the parallel-wire technique to succeed. Blunt (nontapered) occlusions can be difficult to cross, especially when a side branch arises at the site of the CTO. The knowledge that plaque typically accumulates opposite the side branch may help the operator identify the true vessel course. Typically, however, the guidewire will often repeatedly deflect into the side branch, and it may be difficult to penetrate the proximal fibrous cap. As a last resort, dilatation of the artery with a 1.5 mm angioplasty balloon in the side branch may modify the anatomy and disrupt tissue planes to allow the wire to be redirected into the main vessel and afford the purchase necessary to enter the true lumen. If the wire enters a false lumen, vigorous attempts to redirect it into the true lumen will often create an extramural hematoma that can cause compression of the true lumen and collapse of the proximal fibrous cap, making lumen reentry impossible. Moreover, vigorous and repeated wire manipulations widen the subintimal space, which reduces wire resistance and may create a false impression that the true lumen has been found. Once the wire has entered a false lumen, the parallel-wire technique is the best method to locate the true lumen while minimizing the risk of extensive dissection and perforation. In the parallel-wire technique, when a wire is appreciated to have entered a false lumen, it is left in place to mark the dissection channel, and a second (typically stiffer and often tapered) wire with a slightly different primary and secondary bend supported by an over-the-wire balloon catheter is passed along the same path parallel to the first wire, with care taken to avoid wire twisting. Leaving the first wire in place also offers the advantage of straightening bends in the vessel and provides a marker to guide the second wire around the inner curvature of a bend (Fig. 2). Three or more wires may occasionally be used. In a variation termed the seesaw technique, both wires are supported by over-the-wire angioplasty catheters and are alternatively used to probe the occlusion to find the true lumen. The parallel-wire technique. Left: Diagrammatic representation. A Miracle 12 g wire enters a subintimal space at the crux of the distal right coronary artery. The wire is left in place as a marker and a second wire (Conquest in this example) with a slightly different primary and secondary curve is passed in parallel and then steered toward the suspected correct path. Middle: Angiogram of an occluded right coronary artery. The first wire has entered a false lumen. Right: Using the parallel-wire technique, a second wire (supported by an over-the-wire balloon catheter) is able to enter the true lumen. In rare cases, the CTO cannot be successfully crossed from an antegrade approach, but the true lumen can be penetrated from a retrograde direction via a large collateral vessel (e.g., a septal branch) or a bypass graft that connects with the distal vessel. The wire is introduced directly into the distal vessel beyond the occlusion and then steered proximally to approach the end of the convex distal fibrous cap, which is often more easily punctured intraluminally from below. The retrograde wire can either serve as a marker or create a channel that can then be used to facilitate passage of a second wire in an antegrade direction. Rarely the CTO can actually be dilated through the collateral channel or retrograde via a graft. Recent IVUS studies have demonstrated that passage of a guidewire into a subintimal space is a strong factor contributing to unsuccessful recanalization of CTOs [26, 27]. IVUS can differentiate a true lumen from a false lumen by identifying side branches (which arise only from the true lumen) and intima and media (which surround the true lumen, but not the false lumen). Similarly, IVUS can confirm when the guidewire has reentered the true lumen from a false lumen. A typical case with a subintimal wire entry indicated by IVUS is shown in Figure 3. IVUS studies have also revealed that the major reason that it is difficult to penetrate the distal cap into the true lumen is that the guidewire tends to deflect into a false channel, not because of extensive calcification or fibrosis (Fig. 4), in contrast to the proximal entry site that is often densely fibrocalcific. IVUS findings in a chronic total occlusion of the left anterior descending coronary artery. A: Preintervention. B: The short guidewire was used first and was passed into a false lumen. The second wire (longer wire to the left) is in the true lumen. IVUS was performed over this wire. C and D: IVUS images of the proximal and middle part of occlusion, demonstrating that the wire is in the true lumen. E and F: The false lumen made by first wire is seen along the circumference of media (white arrows). J: The IVUS catheter is verified to be in the true lumen as evidenced by a side branch (septal perforator) arising adjacently. Differences between the proximal and distal fibrous caps of a chronic total occlusion. Longitudinal and cross-sectional IVUS views at both ends of a chronic total occlusion > 15 years of age at the ostium of the left anterior descending artery after postdilation with a 1.5 mm balloon. A dense fibrous cap is seen at the proximal end of the occlusion. Although calcification and a thick fibrous membrane were absent at the distal cap, it was difficult to penetrate into the true lumen at the distal end of occlusion as the guidewire frequently deflected off the distal cap, creating a false lumen. IVUS can also be useful in identifying the origin of a CTO when the ostium is occluded or the entry site is otherwise angiographically indistinct. In such cases, the IVUS catheter can be inserted into a proximal side branch to identify the location of the occlusion and guide passage of the wire. An example of this use of IVUS to recanalize a CTO is shown in Figure 5. Example of the utility of IVUS in finding the origin of an ostial occlusion. A: Angiogram showing occlusion of the origin of an obtuse marginal branch of the circumflex. The entry point is indistinct, with several possibilities. B: IVUS at a superior point demonstrating the absence of a side branch. C: IVUS slightly lower demonstrating the occluded branch (marked by an asterisk). D: With the wire containing the IVUS catheter placed in the branch in the left atrioventricular groove, a second wire is used to probe the occlusion. E: IVUS is used to guide the shadow of the wire (arrowhead) to the central portion of the occlusion. F: Under IVUS guidance, the wire is passed distally through the body of the occlusion. The first attempt to cross a CTO should be with an over-the-wire, 1.5–2.0 mm diameter, low-profile balloon dilatation catheter with a tapered tip and a lubricious coating to facilitate crossing even long occlusions. The catheter should transmit force efficiently and have an excellent transition from balloon to tip. Some operators use an end-hole exchange catheter instead, which affords slightly better flexibility, but may not cross as easily, and of course sacrifices the dilatation function of a balloon angioplasty catheter. If, despite orthogonal projections and contralateral injections, it is uncertain whether the distal wire is in the true lumen, the balloon may be passed across the occlusion to Dotter the lesion, then withdrawn and dye-injected. If the distal vessel is still not opacified, the balloon may be passed again across the CTO, the guidewire removed, and a small amount of dye carefully injected distally via the guidewire lumen. Such distal dye injections should be reserved as a procedure of last choice, however, as they will either demonstrate intraluminal position of the catheter or worsen a subintimal dissection, typically ending the case. After the occlusion is crossed and dilated with the over-the-wire catheter, the true dimensions of the CTO may be appreciated and subsequent balloon angioplasty and stent implantation performed with appropriate-sized devices. It should be noted, however, that myogenic spasm in the distal vessel is common after CTO recanalization, often necessitating large and repeated doses of intracoronary nitroglycerin or other vasodilators so the true reference vessel diameter is not underestimated. Crossing and/or dilating the CTO with a balloon angioplasty catheter may be difficult and, in up to 5% of cases, impossible, typically due to the presence of extensive fibrocalcific plaque, especially when guide support is suboptimal. Methods that may be considered for such difficult situations include use of larger and more supportive guiding catheters; deep guide catheter intubation; introduction of a second buddy wire into a branch proximal to the occlusion to increase the support of the guiding catheter, or in the true lumen adjacent to the first wire to increase the dimension of the wire channel (after which it is removed); inflation of an angioplasty balloon either in the main vessel or in a side branch to stabilize the guide catheter; and use of debulking devices. Though most studies have not demonstrated a role for debulking of CTOs to reduce restenosis [28], excimer laser or high-speed rotational atherectomy may allow balloon passage or expansion of otherwise nondilatable CTOs [29, 30]. Lasing is best accomplished over the 0.014″ guidewire with the Spectranetics low-profile X-80 catheter using high fluence (80 mjoules/mm2) and an 80 Hz pulse repetition rate; saline flush is not required as a photoacoustic effect is desirable in this situation. High-speed rotational atherectomy is an excellent alternative if a laser is not available. Use of this device, however, requires recrossing the occlusion with a dedicated 0.009″ stainless steel guidewire; typi
