Bifurcation lesions have been reported to constitute 15% to 20% of all percutaneous coronary interventions (PCI).¹ Since they are associated with an increased risk of procedure-related complication, especially side branch (SB) occlusion after stent implantation, treating these distinct lesion subsets has been a significant challenge for interventional cardiologists. Rapid advancements in novel techniques, devices, and adjunctive pharmacotherapies have considerably reduced the risk of acute complications, restenosis, and stent thrombosis (ST), and ultimately have led to the extension of PCI’s clinical application for various complex bifurcation lesions. Moreover, interventional cardiologists have learned lessons from extensive clinical experiences in that many important anatomic features, including relative plaque distribution to the bifurcation, degree of SB angulation, and severity or length of SB lesion, should be taken into account for technical success and favorable clinical outcome.² Considering that currently there are no clear guidelines to address the use of particular interventional techniques with regard to the specific anatomy of a given bifurcation lesion, every effort should be made to obtain understanding of the technical, clinical, and fundamental aspects of the management of bifurcation disease.

Definition and Classification of Bifurcation Lesions

According to the consensus of the European Bifurcation Club, bifurcation coronary lesion can be defined as “a coronary artery narrowing occurring at, or adjacent to, a significant division of a major epicardial coronary artery.” A “significant SB” is a branch that the operator does not want to lose in a global context of a particular patient (eg, symptoms, comorbidity, diameter and length of SB, size of the myocardial mass supplied by the SB, location of ischemia, viability of the supplied myocardium, collateralizing vessel, left ventricular function, results of functional tests).³

Although several classification systems of bifurcation lesions have been proposed to facilitate PCI, planning the treatment approach according to the angulation of the bifurcation and the degree of plaque burden seems to be most practical (Fig. 38-1). The Y-shaped lesion is defined when the SB and the main vessel (MV) angulation is less than 70%, while that of the T-shaped lesion is greater than 70%. In general, the risk of atheromatous plaque shifting (snow-plow effect) and deterioration of SB flow during stent implantation is higher in Y-shaped lesions. Furthermore, it is technically difficult to achieve the whole coverage of the SB ostium using popular T-stenting in these lesions. Although the risk of SB occlusion is relatively low in T-shaped lesions, wire access can be problematic when large plaque burden is present at a significant SB ostium. The location of stenosis in each of the 3 segments that constitute the bifurcation is also crucial for the treatment strategy, and 7 lesion types, with each type assumed to be associated with a specific treatment technique, have been described in the Medina classification, the simplest and most widely used classification system⁴ (Fig. 38-2).

One-Stent Simple Crossover Technique

The most frequently used approach to treat a bifurcation lesion is to place a stent in the MV while covering the ostium of the SB with the stent (Fig. 38-3). Although this treatment strategy may be very useful in lesions involving a normal SB ostium, it can be used after the predilation in lesions with SB ostial stenosis. In case of significant flow limitation to the SB after MV stenting, further treatment of the SB with a balloon or additional stent (provisional stenting) may be performed. In contrast, even if the diameter stenosis of the SB is up to 70% to 80% after MV stenting, the long-term patency and clinical outcome are usually excellent if the stenosis is confined to the ostium and the flow is not impaired. Therefore, treating such lesions with stent implantation is not recommended.

Since SB compromise occurs to some extent unpredictably, a decision to protect the SB by placing a wire to be left in until the procedure on the MV has been completed, including high-pressure postdilation, should be made carefully. In general, an SB <2.0 mm in diameter and supplying a small amount of viable myocardium is rarely considered for protection, whereas relatively large SBs (≥2.5 mm in diameter) with an ostial diameter stenosis ≥50% need to be considered for protection in case of further specific treatment. The jailed wire in the SB can act as a marker of the occluded SB and straighten the angle between the SB and the MV to facilitate further access. In the provisional technique, rewiring through the distal strut (close to the carina) following the MV stenting is strongly recommended since it creates better SB scaffolding compared with the proximal wire crossing.

One- and Two-Stent Techniques: Clinical Trials in the Drug-Eluting Stent Era

The sirolimus-eluting stent bifurcation study was the first randomized trial that was conducted to assess the feasibility and safety of drug-eluting stent (DES) implantation for bifurcation lesions.⁵ In this study, 22 of the 43 patients who were randomly assigned to the strategy of single stent implantation in the MV with balloon angioplasty for the SB (group B) were crossed over to the strategy of T-stenting in both branches with 2 stents (group A) due to flow impairment or >50% residual diameter stenosis in the SB after stent implantation into the MV. Overall, 63 patients were treated with 2 stents, and 22 patients were treated with 1 stent. Although the high crossover rate (51.2%) made it difficult to directly compare the 2 groups, the in-segment restenosis rate at 6 months did not differ significantly between the 2 treatment groups (28.0% in group A vs 18.7% in group B; P = .53). The researchers in this study postulated that the relatively high restenosis rate at the SB in the 2-stent technique group was caused by the incomplete coverage of the ostium of the SB since T-stenting may leave a gap between the 2 stents at the bifurcation. To address this problem, other bifurcation stenting techniques have been introduced. In the Nordic Bifurcation Trial (NORDIC), the investigators randomized 413 patients with complex lesions to MV-only stenting (n = 207) versus MV and SB stenting using crush, culotte, Y, or other techniques (n = 206) with sirolimus-coated stents. In the MV-only treatment arm, the threshold for crossover to SB stenting was high and was allowed only if the Thrombolysis in Myocardial Infarction (TIMI) flow grade was 0 following SB dilation.⁶ Overall, only 4.4% of the patients in the MV-only stenting group received a stent in the SB. The combined rate of MV restenosis (stenosis >50% in diameter) and SB occlusion after 8 months was low and similar in both the provisional and routine 2-stented groups (5.3% vs 5.1%; P = not significant). Long-term results of the Nordic Bifurcation Study were recently published and showed comparable incidences of all-cause mortality (5.9% vs 10.4%; P = .16), non–procedure-related myocardial infarction (MI; 4% vs 7.9%; P = .09), target vessel revascularization (TVR; 13.4% vs 18.3%; P = .14), and definite ST (3% vs 1.5%; P = .32) over 5 years between the 1-stent (n = 202) and 2-stent (n = 202) approaches.⁷ The investigators therefore concluded that excellent clinical and angiographic results appeared to be obtained with the percutaneous treatment of de novo coronary artery bifurcation lesions, independent of the stenting strategy that was used. However, the 1-stent strategy was associated with reduced procedure and fluoroscopy times and lower rates of procedure-related biomarker elevation.

The BBC ONE (British Bifurcation Coronary Study: Old, New, and Evolving Strategies) study further provided evidence for the preference of 1-stent technique for bifurcations.⁸ A total of 500 patients were randomized either to a simple or a complex stenting strategy (crush or culotte) using paclitaxel-eluting stents. At 9 months of clinical follow-up, there was a significant difference between the 2 groups in terms of the primary end point (a composite of death, MI, and target vessel failure; 8.0% in the simple group vs 15.2% in the complex group; hazard ratio [HR], 2.02; 95% confidence interval [CI], 1.17-3.47; P = .009), MI (3.6% vs 11.2%; P = .001), and in-hospital major adverse cardiovascular events (MACE; 2.0% vs 8.0%; P = .002). A meta-analysis of previous randomized studies demonstrated that a provisional 1-stent approach was comparable to a 2-stent approach in terms of mortality, repeat revascularization, and quality of life.^9,10 However, the 1-stent technique was superior in regard to the risk of periprocedural MI and ST.¹¹ Based on these studies, the 1-stent technique with provisional SB stenting is currently recommended as the primary approach for most bifurcation coronary lesions.³ However, it should be noted that the results of the randomized studies are limited in generalizability due to selective inclusion criteria. Because the aforementioned randomized studies could not include all patients with bifurcations, patients with complex anatomies who may have potentially benefited from the 2-stent approach might have been excluded in enrollment. Moreover, a lack of enough experience in 2-stent techniques across operators might have inflated the risk of adverse events for patients.

The 2-stent approach in provisional or planned situations is still a viable option for a minority of patients who have complex true bifurcations. In 1170 true bifurcation lesions from the Coronary Bifurcation Stenting (COBIS) registry, 17.3% of lesions were treated with 2-stent approaches.¹² The types of 2-stent approaches were T-stenting (47.8%), crush technique (34.1%), V-stenting (14.7%), and culotte stenting (3.4%). There were no differences in the rates of death, MI, target lesion revascularization (TLR), and their composite of MACE. Even in the randomized studies, for patients assigned to a 1-stent approach, a 2-stent technique was eventually performed in 3.5% of patients from the NORDIC and BBC ONE studies and 31% from the CACTUS (Coronary Bifurcations: Application of the Crushing Technique Using Sirolimus-Eluting Stents) study by the operators’ decision. In the recently updated randomized Optimal Stenting Strategy for True Bifurcation Lesions (PERFECT) trial, 28.2% of patients with true bifurcation lesions who were randomized to 1-stent approach eventually received SB stenting.¹³

The 2-stent approach can be attempted for bifurcation lesions if the operator is concerned about the acute complications, including hemodynamic compromise or periprocedural MI in the circumstances of SB loss. In the COBIS registry, SB occlusion occurred in 8.4% of 2227 bifurcation lesions after MV stenting during 1-stent treatment.¹⁴ Preprocedural percent diameter stenosis of the SB (≥50%) and SB lesion length were independent predictors of SB occlusion, with odds ratios of 2.34 and 1.03, respectively. Moreover, SB occlusion was associated with an increase in the adjusted HR of 2.34 (95% CI, 1.15-4.77; P = .02) for the composite of cardiac death or MI and 6.19 (95% CI, 2.00-19.13; P = .002) for ST. In the PERFECT randomized trial, patients treated with provisional SB stenting experienced greater procedural time and contrast amount than the planned 2-stent approach.¹⁵ These findings indicate that bifurcations with complex SB stenosis, which are expected to be compromised or occluded during the 1-stent approach based on angiography with or without intravascular ultrasound (IVUS), may be an indication for a planned 2-stent approach.^16,17 In general, if the SB is large enough (reference diameter ≥2.5 mm), has significant stenosis extending beyond the ostium (10-20 mm or more), supplies sufficient size of the territory to justify stent implantation, and/or has an unfavorable angle for re-crossing after MV stenting, the 2-stent technique can be considered.¹⁸

Two-Stent Techniques

Current 2-stent techniques using DES include culotte, crush and its variants, T-stenting, and V-stenting or simultaneous kissing stent (SKS) techniques. Classic crush is now rarely performed, and variants of the crush technique, such as mini-crush, step crush, and double kissing (DK) crush techniques, are preferred. The potential advantages and disadvantages of each technique are summarized in Table 38-1. Because of lack of studies on the comparative outcomes of diverse 2-stent techniques, selection of proper stenting technique should be dependent on the patient’s clinical manifestation, bifurcation morphology, and operator’s preference. In the same context, since the favorable outcome is more related to the successful procedure itself, not with the type of 2-stent technique, a careful angiographic evaluation is required to identify disease severity, vessel size, and the angle of both branches before the treatment of bifurcation lesions with 2-stent technique.

T-Stenting Technique

The technique of T-stenting is better suited to treat the SBs that originate with an angle close to 90°.^19,20 The deployment of 2 stents in the SB and the MV can be performed in different steps. Since this method is most frequently used for provisional SB stenting, the majority of T-stenting is performed with MV stenting first. However, with the intention to treat with a 2-stent technique at the beginning, initial deployment of the SB stent is preferred to avoid re-crossing the stent strut in the MV (Fig. 38-4). The technique is safe and eliminates the difficulty of advancing the second stent. However, this strategy may carry the risk of incomplete coverage of the SB ostium. This small gap between the stent implanted in the MV and the one implanted in the SB may result in an uneven distribution of the drug, which leads to late restenosis at the SB ostium (Fig. 38-5). The T-stenting and small protrusion (TAP) technique is a modification of the T-stenting technique to overcome this limitation and is based on an intentional minimal protrusion (1-2 mm) of the SB stent within the MV.²¹

Culotte Technique

The culotte technique can be a suitable approach when there is small discrepancy in vessel size between the proximal MV segment and SB (Fig. 38-6). This technique provides full coverage of the bifurcation at the expense of an excess of metal component in the proximal MV.²² The culotte technique is very useful for all angles of bifurcation lesions and can be applied either in the provisional SB stenting strategy or planned 2-stent technique. The more angulated branch is usually stented first. However, if predilation results in a dissection or occlusion in 1 branch, this branch should be stented first in case of the difficulty of rewiring through the stent struts. It is advisable to make minimal overlap of stents in proximal MV segment whenever possible.

Crush Technique

Crush technique is the most widely used 2-stent technique.²³ Despite the complex procedural steps, this technique can be applied to any true bifurcations requiring a 2-stent technique with complete lesion coverage for SB ostium (Figs. 38-7 and 38-8). Because of some limitations of the classic crush technique, its variants, such as mini-crush, step crush, or DK crush, are preferred.²⁴ The mini-crush technique is recommended over classic crush to avoid the large area of 3 strut layers in the proximal MV.²⁵ Using the step crush technique, in which the SB stent strut is crushed by MV balloon instead of stent, separate manipulation of 2 stents is possible and allows precise placement of each stent using 6-Fr guiding catheter²⁶ (Fig. 38-9). The DK crush technique may aid rewiring into the SB after MV stenting and also increases the expanded stent cell area in front of the SB detectable at follow-up.²⁷ In any crush techniques, final kissing balloon inflation (FKI) is mandatory because it allows better strut contact against the ostium of the SB and better drug delivery.²⁸ FKI may also correct the stent deformation and ensure optimal stent scaffolding. As compared with the culotte technique, the mini-crush technique showed similar incidence of MACE (composite of death, MI, TVR, or ST) at 6 months (4.3% for crush vs 3.7% for culotte; P = .87) in the NORDIC II randomized study.²⁹ However, a recent DKCRUSH-III randomized study showed that DK crush (n = 210) was superior to the culotte technique (n = 209) in terms of 1-year MACE including death, MI, and TVR for unprotected distal left main (LM) bifurcation stenosis (6.2% vs 16.3%; P = .001).³⁰

V-Stenting and SKS Technique

The V-stenting or SKS technique can be preferred for bifurcations with a large proximal MV such as LM stenosis.^31,32 V-stenting is preferable for selected bifurcation lesions where the lesions are limited to the distal portion of the bifurcation and have a narrow angle (Fig. 38-10). If there is a proximal stenosis in the MV, the SKS technique is preferred (Fig. 38-11). Since the SKS technique has advantages of fast procedural time and no need of strut reopening, it may have potential benefit for hemodynamically unstable patients with large LM bifurcation stenosis. However, the risk of restenosis or ST may be relatively high compared with other techniques due to the formation of long metal carina; hence, the SKS technique is not recommended for routine 2-stent treatment.

As a result of the long-term benefit of coronary artery bypass graft (CABG) surgery compared with medical therapy, CABG has been regarded as the standard therapy for unprotected left main coronary artery (LMCA) stenosis.^33,34 However, because of easy anatomic accessibility and a relatively large vessel caliber, PCI for LMCA disease has become an attractive option for interventional cardiologists. In addition, technical advances in PCI and devices have emboldened physicians to test the feasibility of LMCA intervention and, coupled with the widespread availability of DESs, have led to a reevaluation of the role of PCI as a viable alternative treatment for unprotected LMCA disease.³⁵ As a result, over the last decade, the prevalence of LMCA stenting has significantly increased worldwide.^36,37 In addition, several recent large registries and randomized controlled trials have demonstrated that LMCA stenting yields comparable mortality and morbidity rates compared with CABG.^38-50 This chapter will provide an overview of current techniques and contemporary outcomes of PCI with either bare metal stents (BMSs) or DESs.

Assessment of Lesion Severity

Clinically, significant LMCA disease has been found in 3% to 5% of all patients who undergo coronary angiography and in 10% to 30% of patients who undergo bypass surgery.⁵¹ Traditionally, angiographic diameter stenosis of 50% has been considered a cutoff for significant LMCA stenosis. However, conventional coronary angiogram is only a luminogram and has critical limitations in assessing lesion morphology and plaque characteristics. The LM trunk is often short in length and lacks a normal segment for comparison. In addition, contrast material in the aortic cusp sometimes obscures the ostium, and “streaming” of contrast may result in a false impression of luminal narrowing. As a result, marked discrepancy in interpretation of degree of stenosis of LMCA narrowing has been documented in several studies.^52,53 Therefore, fractional flow reserve (FFR) and IVUS are often used to assess the severity of LMCA stenosis.

An FFR value of 0.75 to 0.80 or greater has been suggested to be a strong predictor of excellent survival and low event rates in patients with intermediate LMCA disease, making it a useful cutoff value to determine significant LMCA stenosis. In a study involving 213 patients with intermediate LMCA stenosis, 5-year survival rates of 138 patients treated medically with an FFR ≥0.80 and 75 patients treated surgically with an FFR <0.80 were 89.8% and 85.4%, respectively (P = .48).⁵⁴ Several other studies using FFR cutoff values of 0.75 to 0.80 as a surrogate for revascularization showed similar outcomes, and as a result, FFR-guided decision making for the treatment of LMCA stenosis is generally accepted.^55-58 Practically, the FFR of LMCA stenosis should be interpreted with caution because isolated LMCA disease is rare, with most stenosis associated with disease in the left anterior descending artery (LAD) and/or left circumflex artery (LCX), both of which tend to increase the FFR value measured across the LMCA stenosis. Therefore, in these situations, the functional significance of intermediate LMCA stenosis should be reassessed after the correction of distal coronary artery stenosis.⁵⁹

The IVUS-derived minimal lumen area (MLA) has frequently been used to determine the functional significance of intermediate LMCA stenosis, and traditionally, an MLA cutoff value of 6.0 mm² has been considered to represent functionally significant LMCA stenosis. This value was derived primarily from the Murray law, with an MLA of 4.0 mm² considered to represent the ischemic threshold of the LAD or LCX, and was supported by a clinical study conducted by Jasti et al.⁵⁸ However, recent studies reported that the IVUS MLA value corresponding to ischemia-producing lesions of non-LM epicardial coronary arteries to be <3 mm², and application of the Murray law to these values suggests that the IVUS MLA of a stenotic LMCA should be <5.0 mm².^60-62 Park et al⁶³ attempted to determine the IVUS-derived MLA criteria corresponding to an FFR <0.80 in 112 patients with isolated intermediate LMCA stenosis who underwent preinterventional IVUS and FFR measurements. They found that the IVUS-derived MLA value within the LMCA that best predicted FFR <0.80 was <4.5 mm² (77% sensitivity, 82% specificity, 84% positive predictive value, and 75% negative predictive value; area under the curve, 0.83; 95% CI, 0.76-0.6; P < .001), which was similar to the theoretical cutoff of <5.0 mm². It is interesting to note that the positive predictive value of IVUS-measured MLA <4.5 mm² was acceptably high and the anatomic parameter provided by IVUS appeared to correlate well with functional significance of LMCA stenosis. Thus, in cases when FFR measurement is not feasible, IVUS-derived MLA criteria could possibly be used as a surrogate of functional significance of LMCA disease.

Outcomes of Unprotected Left Main Intervention

LMCA stenosis might be considered to be an attractive target for percutaneous intervention because of its large vessel size, short lesion length, and lack of tortuosity. Consequentially, LMCA intervention has been shown to be feasible and to have acceptable short- and mid-term outcomes in the BMS era. The ULTIMA registry enrolled 279 patients with unprotected LMCA stenosis who were treated with BMSs; 46% of the patients were inoperable or at high surgical risk.⁶⁴ The in-hospital mortality rate was 13.7%, and the 1-year incidence of all-cause mortality was 24.2%. However, among the 32% of patients at relatively low risk (age <65 years, left ventricular ejection fraction >30%, and no shock), there were no periprocedural deaths, with a 1-year mortality rate of only 3.4%. Park and colleagues⁶⁵ found that elective BMS stenting for LMCA bifurcation (n = 63) in highly selected patients with normal left ventricular ejection fraction and a large reference vessel might be safe and effective. In this report, the procedure was successful in all patients, and major in-hospital events did not occur in any patients. Moreover, no significant differences were noted in the rates of 2-year freedom from TLR among LMCA ostium, shaft, and bifurcation stenting (82%, 86%, and 85%, respectively). Several other studies also showed favorable short- or mid-term outcomes (in-hospital mortality, 0%-4.3%; mortality at 6-12 months, 2.5%-10.8%) in low-risk patients undergoing elective PCI using BMS for unprotected LMCA disease.^66-70 However, considerable risk of restenosis (18%-31%) and repeat revascularization (7.3%-33.6%), which may lead to worst-case outcomes such as sudden death and acute MI, have limited the durability of LMCA stenting with BMSs. Therefore, LMCA stenting had to be reserved for selected patients who were not candidates for CABG or refused to receive CABG in the BMS era.

After the introduction of DESs, with a remarkable reduction of restenosis and repeat revascularization, PCI with DES has been widely performed for more complex clinical and anatomic subsets of LMCA disease. Although limited by their nonrandomized nature, small numbers of patients, and short follow-up periods, several early observational studies have shown promising outcomes after PCI using early-generation DES compared with BMS.^45,71,72 Even after safety concerns regarding very late ST associated with early-generation DES, physicians’ threshold for performing PCI at the LMCA became less restrictive, and the worldwide frequency of LMCA stenting started to sharply increase.⁷³ In a subsequent meta-analysis comparing outcomes for DES and BMS after LMCA stenting, a total of 44 studies including 10,342 patients who received a DES or BMS were analyzed.⁷⁴ The respective (DES vs BMS) cumulative event rates at 3 years were 8.8% and 12.7% for death, 4.0% and 3.4% for MI, 8.0% and 16.4% for TVR/TLR, and 21.4% and 31.6% for MACE. Adjusted outcomes at 3 years favored DES, including the reduction of mortality.