Due to anatomical reasons and the distributive function of the coronary tree, bifurcation sites are prone to the development of atherosclerotic lesions as a result of flow turbulence generating pro-atherogenous low wall shear stress (WSS). Over the past few years, coronary bifurcation lesions have been the subject of intense therapeutic discussions fuelled by new definitions, classifications (Medina) and measurement and imaging methods as well as various clinical series, registries or randomised studies and meta-analyses. Numerous new findings with a direct impact on therapeutic strategies have emerged from experimentation with various stent implantation techniques in bench tests, which will be soon replaced by digital simulation techniques. The purpose of this article is to examine the current strategies implemented in the treatment of coronary bifurcation disease and to highlight the most debated issues.
In the Drug-eluting Stent Era, Provisional Side Branch Stenting Is the Gold Standard Strategy for Most Bifurcation Lesions
The use of drug-eluting stents (DES) has resulted in a sharp decrease in the rate of repeat intervention in stented bifurcation lesions. In instances where complex strategies such as culotte stenting1 were implemented with bare-metal stents (BMS), the rate was reported to be as high as 45%.
The relatively low rate of repeat intervention in patients undergoing a provisional side branch (SB) stenting strategy2 suggests that, although only a small number of complex strategies have been carried out with BMS, the use of DES is highly recommended for bifurcation stenting, especially when the lesion is complex and treated with a multiple-stent strategy.3
Six randomised studies4–9 of varying quality were undertaken in order to compare the provisional SB stenting strategy with one or more dual-stent techniques. The results of these studies and of relevant prospective registries were pooled in meta-analyses.10–15
These randomised studies were performed using the CYPHER® stent, except for the British Bifurcation Coronary study Old, New and Evolving Strategies (BBC One), the largest study, in which the TAXUS™ stent was used. The complex strategies used were the crush technique (used in most cases), the culotte technique and T-stenting (starting with the SB or the main branch [MB] in the Bifurcations Bad Krozingen [BBK] study and the study by Pan).
Overall, the meta-analyses did not reveal any differences at six- to seven-month follow-up in terms of mortality, major adverse cardiac events (MACE) and target-lesion revascularisation (TLR), although there was a significant difference in favour of the single-stent group in the BBC One study. However, a significant difference in the rate of myocardial infarction (MI) (<30 days early MI) was observed in favour of the single-stent strategy. A tendency towards a higher incidence of stent thrombosis was also noted in the complex strategy group.
One of the meta-analyses focused on the angiographic results, showing no differences in the MB and better early angiographic outcome in the SB treated with a stent (minimum lumen diameter [MLD] compared with balloon angioplasty alone, although this was no longer observed during follow-up (higher acute gain and higher late loss with a DES in the SB).15 The obvious conclusion to be drawn from this is that when a good angiographic result is obtained in the SB with balloon angioplasty alone, it is pointless, unnecessarily complex and costly (procedural time, X-ray exposure) to implant a stent into the SB. This also favours the use of a stenting strategy starting with the MB.
These results were achieved in all types of lesions, except for Medina 0,0,1 lesions (isolated SB stenosis), which were usually excluded, and diffuse SB lesions in the study by Pan. In the Crushing Technique Application Using SES in Coronary Bifurcations (CACTUS) study, all lesions were true bifurcation lesions (Medina 1,1,1, 0,1,1 or 1,0,1), and the rate was 83% in the BBC One study.
Stenting techniques cannot currently be indexed according to lesion type, as the provisional strategy has been successfully tested in all lesion types.
The impact of final kissing balloon inflation on dual-stent strategies is not discussed. However, in patients undergoing provisional SB stenting without SB stenting, final kissing was reported in a large registry16 as a predictor of events in the SB, resulting in an increased rate of SB stenting. Most importantly, the conclusion of two randomised studies was that its systematic use was not associated with any advantage or disadvantage apart from higher procedural complexity with kissing.17,18 Nevertheless, the implementation of the final kissing balloon strategy was shown to reduce the risk of ischaemia and residual symptoms (Burzotta, personal communication).
In Which Complex Lesions Can the Role of Complex Strategies Be Discussed?
The advocates of complex strategies report the following indications: SB lesions with difficult access due to an ‘A’ angle configuration (access between the proximal segment and the SB); however, the angulation can only be assessed after wiring the SB. When angle A is <120°, it increases by an average of 33° after guidewire placement.19
In the CACTUS study involving only true bifurcation lesions, the length of the SB lesion was between 5.1±4.4mm in the provisional group and 5.9±4.7mm in the crush group. The presence of relatively long lesions in both groups did not result in any differences in outcome during follow-up.
The presence of a lesion in a very large SB may call for stent placement in order to treat significant ischaemia or symptoms. The presence of very long lesions in the SB may call for primary SB stenting in order to anticipate the impossibility of stenting the SB if it becomes necessary.
The risk of haemodynamic deterioration associated with potential SB occlusion (poor left ventricular ejection fraction [LVEF], distal left main stenting) may prompt the operator to stent the SB first for safety reasons.
Which Strategies Should Be Implemented in Complex Bifurcation Lesions?
Many bifurcation stenting techniques using multiple stents have been described (see Figure 1).20 These strategies do not always start with the stenting of the SB (T-stenting, modified T-stenting and crush stenting variations). In some cases, the proximal segment of the bifurcation is stented first,21 or both branches simultaneously (V or touching stenting, simultaneous kissing stent [SKS]), or the main branch across the SB (T-stenting, culotte, internal crush, T and protrusion).
The advantages and disadvantages of each technique can be assessed using various criteria: easy handling, stent-to-wall apposition for drug delivery, change in flow dynamics inside the bifurcation and compliance with bifurcation branching laws,22–26 which are closely related to the distributive function of the epicardial coronary arteries.
Murray’s law27 describes the relationship between the diameters of the three segments of a bifurcation, each having its own reference. Finet’s formula28 is a simplification of Murray’s law: PM diameter = (DM diameter + SB diameter) x 0.678.
This formula is reproducible (self-replication) in all bifurcations governed by a fractal geometry principle, according to which an artery has a reference diameter decreasing by successive stages and not gradually, the diameter remaining constant between two bifurcations. The ramification law of the epicardial coronary tree governs its distributive function via linear relations: diameter/ flow, length/flow or diameter/flow/myocardial mass. The anatomy of the coronary tree is directly related to the efficacy of its distributive function.
T-stenting starting with the SB29,30 is theoretically the ideal strategy for true bifurcations with an A angle close to 90°. However, in most cases this angle disappears when the SB is wired. When angle A is open, performing T-stenting in the SB first may be complicated by two potential issues: protrusion into the MB and formation of gaps. These may result from a stent being implanted too distally in the SB and also from plaque shifting from the MB to the SB over the SB stent during secondary stenting of the MB. Protrusion of the SB stent into the MB may induce flow turbulences and stent thrombosis.
The crush stenting technique,32 which is relatively easy to carry out without kissing, was designed to prevent the gap phenomenon at the SB ostium. A stent is implanted in the SB with struts protruding into the MB, in which a wire, a balloon or even an undeployed stent are present. After removal of the wire and balloon from the SB, the proximal segment of the SB stent is crushed by means of a balloon (balloon crush)32 or a stent in the main branch (classic crush). The accumulation of three layers of stents, often inadequately apposed by a balloon of insufficient size, has been shown to result in flow turbulences and blood stagnation, thus neutralising drug efficacy or hindering the endothelialisation process with the resulting risk of stent thrombosis. Furthermore, intravascular ultrasound (IVUS) examination has revealed that following crush stenting the minimal surface area of the MB is within the crushed area in the proximal MB and not in the distal MB;33 this contradicts the ramification law governing coronary bifurcations and results from the impossibility of deploying the MB stent adequately when a stent has already been placed in the SB.
A decrease in MACE-free survival rate was reported in cases where the angle between the two distal branches was >50°. This was probably related to the extreme difficulty of deploying and apposing the SB stent at the carina adequately in this type of angulation. Therefore, it is preferable to avoid the crush technique in open angles.36 Moreover, it was shown that the risk of TLR was three-fold higher when final kissing balloon inflation was not performed.37 Although indispensable, kissing balloon inflation is not always feasible (91.1% in the CACTUS study).
Enhanced versions of the crush technique may improve the technical outcome and probably also the clinical outcome. High-pressure inflation of a balloon in the SB prior to kissing inflation may enable the operator to position the stent struts on the wall at the SB ostium. The double kissing crush technique (DK crush)38 consists of performing a kissing balloon inflation after successive implantation of each stent, thus allowing better stent deployment in the proximal segment and in the SB. However, it is recommended that the segment of crushed stent be as short as possible (minicrush).39
The culotte strategy is not an easy technique to carry out. As with the crush technique, it poses the same problems of stagnation, drug delivery and endothelialisation in the proximal segment. Furthermore, expansion of the SB and MB stents depends on the MB and SB stent strut surface area, which may result in incorrect stent apposition in certain areas, flow turbulence and inadequate drug delivery. The culotte technique has also been associated with less than optimal results in open-angle bifurcations (B angle, between the two distal branches).40
Considered as simple to implement, the SKS technique41 has not been thoroughly assessed from a clinical point of view or with respect to its impact on intracoronary flow and drug delivery. It is probable that coronary flow is significantly disrupted by the new carina created in the proximal segment and also by the presence of two additional circulating channels at the junction of the two stents implanted in this segment.
Apposition of the two stents is seldom optimal as they are likely to become entwined.42 Some studies report high rates of stent thrombosis, especially when short-duration antiplatelet treatment has been prescribed. Restenosis with the added difficulty of a membrane on the neo-carina is difficult to treat.
Why Not Start with the Main Branch?
The provisional SB stenting strategy is appropriate in instances where stent deployment in the SB definitely seems to be required. The advantage of this strategy is that the need for an SB stent is confirmed or invalidated during the course of the procedure. This strategy enables the operator to appose a layer of stent correctly and to optimise stent deployment in the main branch with appropriate proximal and distal diameters (proximal optimisation technique [POT]) (see Figure 2).
The difference in diameter between the proximal and distal stent segment brings about a change in the SB ostium plan, allowing guidewire exchange and balloon or stent insertion (see Figure 3). A kissing balloon inflation is performed when the SB is large, stenosed or occluded, with slow flow, chest pain or electrocardiogram (EKG) changes. Kissing balloon inflation results in the MB stent struts being pushed into the SB ostium, thus allowing stent deployment in the SB ostium without leaving a gap. Coronary flow is disrupted by only one layer of struts; the drug is delivered homogeneously and efficiently (including in areas where it is not necessary, such as the carina) and adequate diameter/flow/myocardial relations are ensured.
The implementation of this technical strategy is as follows (see Figure 4). Systematic wiring of each of the two branches takes place, starting with the least accessible branch in order to limit the risk of wire twisting. It is preferable to avoid wires with a polymeric sheath, which might peel off during wire removal. Pre-dilatation of the MB should be performed according to conventional techniques. It confers the advantages of direct stenting in cases of unstable coronary syndromes. The absence of SB dilatation eliminates the risk of dissection but also precludes subsequent passage into the SB through several cells. Passing through the most distal strut is the only strategy, allowing the apposition of struts against the wall of the proximal SB. Selection of the MB DES is made according to the diameter of the main distal segment in order to avoid displacement of the carina, which may lead to SB occlusion. The stent should be implanted across the SB after placement of a jailed wire. Several studies have shown a better follow-up outcome using a Cypher stent instead of a Taxus,42,43 except in the distal left main trunk.44 Optimisation of stent deployment in the proximal segment of the MB should be performed with a short balloon (6mm) in order to avoid the occurrence of geographic miss.
This POT45 enables the operator to appose the proximal stent onto the vessel wall for efficient drug delivery and also to prevent the wire from slipping outside the proximal segment of the MB stent during guidewire exchange. The POT is especially useful when the SB is large and where the difference between the proximal and distal segments of the artery is maximal. Subsequent guidewire exchange is facilitated. The operator starts the manoeuvre by withdrawing the guidewire from the distal MB while attempting to penetrate the most distal cell towards the SB; the SB wire is then unjailed and advanced into the distal MB.
Kissing balloon inflation is required when the SB is occluded or severely stenosed with slow flow, chest pain or EKG change or in the presence of a significant stenosis in a large SB requiring treatment, when inadequate treatment of the SB could result in the occurrence of ischaemia or residual symptoms. Koo46 clearly showed the discordance between an angiographically tight stenosis in the SB ostium after MB stenting and the absence of functional consequences as analysed by fractional flow reserve (FFR). Kissing balloon inflation enables the operator to stent the SB ostium while avoiding stent distortion on the opposite artery wall in anticipation of potential SB stenting. As the objective is no longer to optimise deployment of the proximal segment of the MB stent, the use of very short balloons selected according to the distal diameter of both vessels is recommended. Similarly, the use of non-compliant balloons prevents the occurrence of dissections in the SB ostium.
Accurate T implantation of a second stent in the SB is a simple manoeuvre in most cases. The markers/stent relations on the stent delivery balloons, which vary according to the stent used, should be clearly identified in order to achieve optimal stent deployment. Final kissing balloon inflation starting with the MB is performed to complete the double stenting procedure.
Many dedicated stents were designed for provisional SB stenting in order to allow reproducible strut projection in the SB ostium and permanent access to both branches without guidewire exchange, but their use has not proved to be as simple as expected.
Conclusions
Treatment of coronary bifurcation lesions is still a polemical issue. However, in the vast majority of cases, provisional SB stenting results in satisfactory angiographic outcome associated with excellent mid-term clinical outcome in recipients of DES.
Situations in which primary placement of two stents is required have not been accurately described. The various dual-stent techniques currently available have been evolving in compliance with the following principles: optimal stent deployment and wall apposition for achievement of a biological effect; limitation of stent overlapping areas in order to allow endothelialisation; and reduced strut exposure to intracoronary flow in order to avoid thrombus-generating turbulences.
In instances where double stenting seems to be necessary, the SB is stented first in anticipation of potential difficulties in implanting a stent in the SB through the struts of the MB stent. Double stenting can also be carried out with the MB being stented first provided that stent deployment does not alter the initial anatomical configuration of the MB. This allows accurate placement of a single layer of DES in the whole bifurcation without hindering stent implantation in the MB.