Drug-eluting stents (DESs) have contributed to a significant lowering of the incidence of restenosis and target vessel revascularisation (TVR) in bifurcations.1–4 A randomised study of bifurcation lesions using sirolimus-eluting stents revealed restenosis rates of only 4 % in the main branch (MB) and a TVR rate as low as 8.2 % at six-month follow-up,2 a marked improvement over that in historical controls using bare metal stents. Results on the side branch (SB) remained poor, with restenosis rates as high as 20 %, despite the use of DESs and irrespective of whether the SB was stented or the procedure was completed only with final kissing balloon inflation (FKI).
The total restenosis rate at six months was 25.7 % overall, 28.0 % in the double-stenting group and 18.7 % in the provisional SB-stenting group, respectively.2
In another study using DESs for bifurcation treatment, Ge et al. reported target lesion revascularisation (TLR) rates of 8.9 % (two-stent group) and 5.4 % (one-stent group) at nine-month follow-up with sirolimus-eluting stents.5 The incidence of in-hospital major adverse cardiac events (MACEs) was similar to the experience with bare metal stents, with 8.8 % in the group with one stent and 10.3 % in the group with two stents. No specific conclusion could be drawn concerning the advantage of stenting one branch versus two branches, because a two-stent approach was more often preferred for complex bifurcational lesions.
Randomised trials have demonstrated that, for a lesion localised mainly in the main vessel (MV), stenting of both the MV and SB with two stents does not produce any clinical benefit compared with stenting the MV only.2,6,7 Currently, bifurcation percutaneous coronary intervention (PCI) with two stents is performed mainly as a cross-over from a provisional strategy in the case of a suboptimal result in a large-sized SB (abrupt closure, flow-limiting dissection, >75 % stenosis, thrombolysis in myocardial infarction [TIMI] flow <3). Bifurcations involving large-sized SBs with extensive disease beyond the ostium or having a steep angle are much less likely to be treated optimally with a one-stent technique. It is estimated that approximately 30 % of true non-left main bifurcation lesions encountered in everyday practice warrant treatment with a two-stent technique.8
When planning a ‘true’ complex bifurcation treatment strategy, we have to think in advance of a two-stent approach, because stenting of the SB is an early step in several two-stent techniques (crush, conventional culotte and some dedicated stents). A two-stent strategy as intention to treat might be needed in true bifurcations (Medina 1.1.1, 1.0.1 and 0.1.1) when a significant SB is involved (>2.5 mm, large amount of myocardium subtended, disease extending >5 mm from the ostium). A provisional approach is preferred on the other hand in true bifurcations with no severe SB disease, extending <5 mm from the ostium, and all ‘non-true’ bifurcations.
In a recent meta-analysis of randomised trials including patients with coronary bifurcation lesions who were randomly selected to undergo PCI by either double or single stenting, Katritsis et al. found an increased risk of myocardial infarction (MI) with double stenting (risk ratio 1.75, p=0.001), as well as an increase in the risk of stent thrombosis (risk ratio 1.85, p=0.19). Point estimates suggested approximately a 3 % excess risk of MI and 1 % excess risk of stent thrombosis with the double-stent versus single-stent approach. No differences were found in TLR (risk ratio 1.09, p=0.67).9
In summary, a provisional T-stenting strategy, with stenting of the MV and elective stenting of the SB, appears to have equivalent efficacy to, with less risk of MI than, a planned two-stent strategy in most bifurcation lesions.4,7,10,11 It is estimated that approximately 70 % of true non-left main bifurcations are currently being tackled with a provisional T-stent approach.12
Comparison of single versus complex strategies is often limited because complete randomisation is difficult and a high cross-over rate is generally observed between groups. Such bias is present in almost all bifurcation studies, with operators selecting a two-stent approach in more complex lesions while treating simple lesions with one stent.
Limitations of the Provisional Strategy
In clinical practice, however, true bifurcation lesions with extensive lesions in the SB are unlikely to be successfully treated by a provisional approach and deployment of one stent only in the MB in a bifurcation lesion involving disease in the SB is limited by the risk of restenosis, particularly at the SB ostium, compromising SB patency.2
The goal of bifurcation stenting treatment remains to maintain patency of both branches and it is estimated that, in 30 % of true non-left main bifurcations, stent coverage is required for both the MV and the SB, and these lesions should be treated with a two-stent technique.13 True complex bifurcations with large SB diameter (>2.5 mm) are still an area of uncertainty and questions for interventionalists. This is particularly the case for true left main bifurcations, which remain nowadays one of the last taboos in PCI.14
Conventional Two-stent Strategies
Culotte, Crush and Simultaneous Kissing Stents
Culotte and crush techniques have been designed to provide complete scaffolding of the SB and MB. These techniques are usually employed for large SB (>2.5 mm) with relatively low take-off angle. In the culotte technique, a first stent is deployed in one of the branches, usually across the SB to the MB, and the MB is then rewired through the strut of the first stent and dilated with a non-compliant balloon. The second stent is implanted followed by FKI.
In the crush technique, a stent is first positioned in the SB and retracted to protrude into the MB (>5 mm in the classical crush or 1–2 mm in mini-crush). The protruding portion of the SB stent is then crushed against the wall by deployment of the MV stent or dilatation with a NC balloon. The procedure needs to be completed with the rewiring of the SB and FKI post-dilatation (see Figures 1 and 2).
In the Nordic stent technique study (Randomized comparison of coronary bifurcation stenting with the crush versus the culotte technique using sirolimus eluting stents), a total of 424 patients with bifurcation lesions were randomised to crush (n=209) and culotte (n=215) stenting. At six months there were no significant differences in MACE rates between the groups. Both crush and culotte bifurcation stenting techniques were associated with similar MACE rates: 4.3 % for crush and 3.7 % for culotte (p=0.87). Angiographic in-stent restenosis of MV and/or SB after eight months was found in 12.1 versus 6.6 % (p=0.10), and in 10.5 versus 4.5 % (p=0.046), in the crush and culotte groups, respectively.15
In the recent Coronary bifurcations: application of the crush technique using sirolimus-eluting stents (CACTUS) study, a randomised prospective trial comparing crush stenting with provisional T-stenting in true bifurcation, Colombo et al. found that angiographic restenosis rates were not different between the crush and provisional groups, with 4.6 and 6.7 % restenosis in the MB (p=not specified [ns]) and 13.2 and 14.7 % restenosis in the SB, respectively.16 MACE rates were also similar, with 15.8 % in the crush group versus 15.0 % in the provisional stenting group (p=ns).
FKI was performed in the majority of patients (91.1 %), and absence of FKI was associated in both groups with a significantly higher rate of in-hospital and follow-up MI (29.0 versus 7.5 % with FKI, p<0.001), higher stent thrombosis (6.5 versus 0.9 %, p=0.06), higher incidence of TLR (12.9 versus 6.3 %, p=0.25), as well as higher angiographic restenosis rate in the MB (16.0 versus 4.7 %, p=0.03) and in the SB (36.0 versus 11.9 %, p=0.001).16
Recent studies suggest that a crush technique with a double kissing post-dilatation (DK crush) may provide better results than provisional stenting in true bifurcation lesions.17 In the DK crush strategy, a first kissing balloon (KB) post-dilatation is performed after the implantation of the SB stent and re-crossing. A second KB dilatation is performed after MV stent implantation, similarly to the conventional crush technique. Excellent results with the DK crush technique have been reported, with a similar MACE rate and inferior TLR rate compared with controls treated with a provisional strategy. This study underlines the importance of post-dilatation in complex strategies to achieve satisfactory clinical outcomes.
Another technique for complex bifurcation treatment is the simultaneous kissing stent (SKS) technique. Both the SB and MB stents are deployed either simultaneously or sequentially to form a double-barrel stent with a neocarina in the MB. Such a technique has the advantage of not requiring re-crossing of the stent, but the SKS technique raises concerns over the potential risk of stent thrombosis induced by the long double neocarina in the MB.
T and T with Protrusion
In classical T-stenting, a stent is implanted in the SB up to the SB ostium. A second stent is then deployed in the MB and the procedure is completed by FKI. It is a popular approach for high-angle (>70 degrees) T-shaped bifurcation, because it provides coverage of the SB with minimal stent protrusion into the MB.
In a randomised study with sirolimus-eluting stents in 101 patients assigned either to provisional T-stenting or routine T-stenting, Ferenc reported binary restenosis rates in the MB of 7.3 and 3.1 % (p=0.32), respectively, at nine-month follow-up, with TVR at one year of 10.9 % after provisional and 8.9 % after routine T-stenting. SB restenosis remained high in both groups, with 23.0 ± 20.2 % after provisional T-stenting and 27.7 ± 24.8 % (p=0.15) after routine T-stenting. They conclude that routine T-stenting is not significantly different in terms of angiographic outcome as compared with provisional SB stenting.
For angles <70 degrees, the T-stenting technique is limited by the potential gap in scaffolding at the SB ostium. A popular adaptation of the T-stenting that offers both good strut apposition and coverage of the ostium is the T with protrusion (TAP). The SB stent is advanced and left with a minimal protrusion (1–2 mm) into the MB. The aim of the TAP technique is to provide full continuity of scaffolding between the MB stent and the SB stent. Good clinical results have been observed with the TAP technique and long-term TLR rates as low as 6.8 % have been reported.13 The main limitation of TAP remains the risk of misplacing the SB stent and producing a neocarina, extensively protruding into the MB lumen (see Figure 2).
Two Stents or One Stent – The Scaffolding versus Apposition Dilemma
Whereas a simple provisional approach ensures good stent strut apposition and warrants preservation of the MV patency, the technique often fails to protect the SB ostium with the risk of a late focal re-narrowing. On the other hand, complex techniques such as the crush, culotte or SKS techniques have been designed to provide continuity in the stent coverage between the SB and the MB to ensure protection of the SB. Gaps in stent scaffolding and focal ostial restenosis remain major limitations of the T-stenting technique, particularly for angles <50 degrees.8 Complex techniques are, however, limited by the significantly higher rate of struts remaining unapposed.18,19 This metallic presence in the middle of the lumen may compensate for the benefits of protecting the SB ostium with an additional stent.
The major limitation of current bifurcation strategies is summed up in a dilemma for the interventionalist, who must choose between lesion scaffolding and strut apposition.
With the culotte and crush techniques, another issue lies in the risk of compromising one of the branches in the case of failure to rewire or re-cross a balloon through the stent strut to complete the procedure and perform a mandatory FKI. Absence of FKI in the CACTUS trial was associated with a dramatic increase in the rate of in-hospital and follow-up MI (29.0 versus 7.5 % with FKI, p<0.001) and stent thrombosis (6.5 versus 0.9 %, p=0.06).16
Dedicated bifurcation designs have emerged, with the aim of overcoming these limitations and providing continuous scaffolding of both the SB and MB without detrimental strut malapposition throughout the bifurcation (see Figures 3, 4 and 5).20,21
Potential Impact of Strut Malapposition in Two-stent Strategies
High rates of malapposition are found in the complex technique, due to the superposition of two layers (culotte) or three layers (crush) of stent struts in the MV.22 Multiple layers of stent struts at the carina and in the proximal MB are a cause of concern for stent thrombosis, particularly with DESs.23In vivo optical coherence tomography (OCT) analysis and bench models have shown that complete apposition of overlapping layers of struts can be hard to achieve despite use of high-pressure non-compliant balloons and FKI.19,22
Stent malapposition raises particular concerns about the risk of stent thrombosis. In vitro models have shown that shear rate, which represents the gradient in velocity within the blood flow, can activate platelets at rates >1,000 s-1 in a dose-dependent manner through von Willebrand factor binding to glycoprotein (GP) Ib and GP IIb/IIIa receptors.24 Unapposed stent struts in the lumen create a situation of facing steps for blood flow and result in flow separation and disturbance with high shear gradient, thereby increasing the risk of platelet adhesion and stent thrombosis.25–27 Several clinical studies have confirmed the correlation between incomplete stent deployment and a higher risk of stent thrombosis.3,28–31
Strut coverage with endothelium provides an antithrombotic barrier between the blood and the implanted stent, which may be critical for preventing the risk of stent thrombosis.32–34 In a DES, antiproliferative drugs and polymer coating damage delay strut coverage by endothelium and DESs typically present a lower rate of re-endothelialisation than bare metal stents (BMSs).32 Uncovered malapposed struts may thereby persist for years after implantation, necessitating dual anti-platelet treatment and leaving a risk of potential late stent thrombosis.32,34,35
Neocarinas after SKS and crush techniques, in particular, can cause severe flow disturbances by the juxtaposition of a large number of struts, grossly unapposed, with the high velocity component of the bloodstream, which is particularly concerning for the risk of stent thrombosis (see Figures 1, 6 and 7).
This correlation is a concern in bifurcations, considering the high rate of strut malapposition observed in vivo.19,25 As shown here, complete dilatation of the bifurcation ostium and apposition of the struts remains almost impossible with current DESs, even when following recommendations on the treatment of bifurcations and FKI. In vivo rates in excess of 40 % of malapposed struts have been reported after bifurcation stenting towards the ostium, despite systematic final KB dilatation.19 This might partially explain why such anatomical subsets are more vulnerable to high rates of stent thrombosis and MACE after PCI.3,6,7,16
Conclusions
At present, the following conclusions can be drawn from the data and various factors discussed in this article:
- Stent strut malapposition, which is suggested to increase the risk of stent thrombosis, is frequent in bifurcation even after recommended FKI.
- A trade-off exists with current bifurcation strategies between stent coverage at the SB ostium and stent strut apposition.
- Two-stent complex techniques lead to an increased rate of strut malapposition, particularly with the crush technique.
- Introduction of dedicated designs represents an interesting potential for overcoming the limitations of conventional devices in the treatment of bifurcation lesions. Randomised trials are needed to validate their benefits and popularise their routine use.