Bifurcation Stenting - Have We Made Progress in the Last 10 Years

Permissions× For commercial reprint enquiries please contact Springer Healthcare:

For permissions and non-commercial reprint enquiries, please visit to start a request.

For author reprints, please email
Average (ratings)
No ratings
Your rating


Bifurcation stenting is one of the unsolved challenges for interventional cardiologists. Patients with bifurcation lesions tend to have more advanced disease and multiple co-morbidities. Over the last 10 years we have come to understand the importance and made progress in bifurcation imaging for planning the procedure and for assessment of procedural success, from 3D angiography and multislice computed tomography (MSCT) to fractional flow reserve (FFR) and optical coherence tomography (OCT). Recent clinical trials have shown improved results with selective use of a two-stent strategy and drug-eluting stents. Rates of major adverse cardiac events (MACE) and peri-procedural myocardial infarction (MI) for bifurcation lesions still remain high. Many challenges such as side-branch access, wire-trapping, incomplete side-branch coverage and restenosis still remain. No single dedicated bifurcation stent design thus far has been able to solve them all. More long-term, prospective, efficacious studies of these novel stent designs with concomitant imaging are needed for the field to progress.

Disclosure:The authors have no conflicts of interest to declare.



Correspondence Details:Willem J van der Giessen, Interventional Cardiology, Thoraxcentre Erasmus MC, 's-Gravendijkwal 230, 3015CE Rotterdam, The Netherlands. E:

Copyright Statement:

The copyright in this work belongs to Radcliffe Medical Media. Only articles clearly marked with the CC BY-NC logo are published with the Creative Commons by Attribution Licence. The CC BY-NC option was not available for Radcliffe journals before 1 January 2019. Articles marked ‘Open Access’ but not marked ‘CC BY-NC’ are made freely accessible at the time of publication but are subject to standard copyright law regarding reproduction and distribution. Permission is required for reuse of this content.

Approach to the Bifurcation Patient

Historically, bifurcation lesions have been associated with lower procedural success and a poorer clinical outcome than non-bifurcation lesions. This may be due to the technical difficulty of the procedure but also due to the complexity of the patients. According to the recently presented LEADERS trial sub-study,1 patients who have bifurcation disease are significantly more likely to have had a previous myocardial infarction (MI) (37.2 versus 30.4%; p=0.007) and multivessel disease (36.0 versus 17.0%; p<0.001). The latter was reflected in a much higher Syntax score (17 versus 12; p<0.001), indicating that next to the target bifurcation there is, on average, at least one additional lesion. In addition, patients with bifurcation disease tended to have smaller-diameter vessels. The rate of peri-procedural MI in LEADERS was significantly higher in patients with bifurcation disease (6.6 versus 3.6%, p=0.01). Similarly, in the ARTS-II bifurcation sub-study,2 the bifurcation group had more extensive and complex disease (i.e. diffuse lesions, type C lesions). The procedure itself was also more complex, with a higher number and length of stents implanted and longer procedural time. Thus, the first important step in planning the bifurcation procedure is to carefully evaluate the patient and minimise his or her peri-procedural complication risk. This may include routine use of 2β3α inhibitors or newer antithrombotic agents on top of the usual dual antiplatelet therapy. In addition, pre-hydration in patients with renal disease and careful assessment of anatomy using state-of-the-art imaging techniques is paramount. Based on this careful assessment of clinical and procedural risk, one may even positively decide to change the revascularisation strategy from percutaneous to surgical.1

Imaging Techniques

The treatment of bifurcation lesions requires careful assessment of the bifurcation disease anatomy, including the size of the proximal versus the distal vessel, the extent of the disease in the side branch, the degree of calcification and the side-branch angle. Accurate assessment of these bifurcation features is critical in choosing the strategy and tools that would provide the best side-branch coverage with minimal overlap of struts and good stent apposition, thereby minimising the complication rates. However, angiographic 2D assessment of bifurcations is often incomplete because of vessel overlap, foreshortening and poor reconstruction of angles. 2D-angiography-based Medina classification of bifurcation lesions is the most accepted, and easiest to use, classification system.3 It divides the bifurcation into main proximal, main distal and side-branch segments and assigns binary (0 and 1) values to describe plaque distribution.
Angiographic assessment of bifurcation lesions was recently summarised by Lansky et al. in 2009.4 One has to be aware of the inherent inaccuracies in measuring QCA in bifurcation lesions. QCA algorithms have the assumption that the vessel tapers minimally along its course. This assumption is not true for bifurcations where Murray’s law must be applied and standard QCA results in the underestimation of the proximal vessel size. Dedicated bifurcation software is now available from Medis and Pie Medical (The Netherlands). The two software programs are described by Lansky et al.4 and Ramcharitar et al.,5 respectively. The consensus is that bifurcation lesions should be systematically described according to the Medina classification,3 at least three projections are required and quantitative analysis performed in two views with minimal overlap and foreshortening. 2D angulations between the proximal main branch and side branch should be measured. This software has been developed with trial reporting in mind. Single restenosis for the side as well as main branch should be reported and overall restenosis for the entire bifurcation. It is also recommended to perform a more detailed segmental analysis of restenosis, which may give insight into its mechanism.

3D Angiography

Given the limitations of 2D angiography, we currently recommend the use of 3D reconstruction of the bifurcation lesions with specialised software such as Paeion Medical (Israel). This system uses two angiographic orthogonal images to reconstruct 3D anatomy and allows the measurement of the bifurcation angle, defined as the angle where the centre lines of the lumen of the main and side-branch cross (see Figure 1A). The mean 3D bifurcation angle and the systolic and diastolic angle are then analysed. Careful assessment of angulation not only allows one to choose the proper stenting technique, such as T-stenting for 90º angles and other techniques for lower angulations to avoid side-branch gap, but is also a powerful predictor of clinical outcomes including mortality.6

Multislice Computed Tomography

For similar reasons, multislice computed tomography (MSCT) with volume-rendered reconstructions of the coronary tree may facilitate assessment of bifurcations, with respect not only to the angulation but also to the distribution of plaque and calcifications. Knowledge of the plaque distribution and calcifications allows one to better choose the wire, navigate the calcifications and make a decision about the need for pre-dilation, in addition to choosing the best stenting strategy. MSCT 3D reconstruction is less user-dependent than 2D angiography and allows for more accurate angle measurements.7 As recently shown, steep angles >72º between the left main and left circumflex artery are predictive of incomplete apposition and distortion of the side-branch stent with the crush technique, which, in turn, is associated with a higher mortality.8 Careful review of MSCT data prior to the planned intervention is therefore highly advisable (see Figure 1B). In addition, a prior MSCT study may also decrease contrast use during the actual procedure.

Intravascular Ultrasound

Although intravascular ultrasound (IVUS) is invasive and available intra-operatively, it can offer invaluable insights into the anatomy (size and length) of the bifurcation pre- and post-stenting and guide stent optimisation. Obtaining IVUS data is particularly important, but poor correlation exists between angiography and IVUS measurements of the mean bifurcation angle.9 It can be particularly valuable in cases where MSCT is not available or the degree of calcification obscures the ability to assess the lesion severity.

With regard to future approaches to bifurcation treatment, pre-procedural imaging may help us to solve two classic bifurcation problems: wire-crossing and/or twisting and re-cross. Importing the imaging data set onto a wire navigation system might facilitate the engagement of even the most difficult side branches. In our current practice, we are actively exploring the use of both MSCT and 3D-angiography with magnet navigation and GPS-systems (RJ van Geuns, personal communication).

Post-stenting Fractional Flow Reserve

The decision to dilate the side branch after stenting of the main branch, and whether a side-branch stent is necessary, often cannot be reliably made based on angiographic appearance. A shift in the carina angle after placement of the stent often creates an illusion of significant ‘pinching’ of the ostium. Fractional flow reserve (FFR) is a physiological measurement of maximal myocardial flow that can be maintained in the presence of a given stenosis. It is calculated from the ratio of the distal coronary to proximal aortic pressure at maximal hyperaemia. FFR <0.75 denotes significant stenosis. The seminal paper by Koo’s group10 has shown that there was a negative correlation between the per cent diameter of side-branch stenosis and FFR measurement. In addition, fewer than 30% of the lesions with angiographically significant stenosis had flow limitation shown by FFR measurement. FFR has not been evaluated in more complex diffuse side-branch lesions or in cases of elective two-stent strategy.

Post-stenting Optical Coherence Tomography for Optimisation of Results

Optical coherence tomography (OCT) is emerging as the preferred technique for post-stent assessment in terms of apposition, strut distortion and protrusion and overlap or presence of gaps (incomplete coverage). At 10–20μ resolution, when performed in both limbs, OCT allows for bifurcation reconstruction at the level of detail similar to that of the ‘Ormistogram’ model or ‘microCT’. Combined assessment of bifurcations with IVUS-VH and OCT can identify plaque composition as well as plaque burden and the presence of a necrotic core with ‘thin cap’ (fibroatheroma), which is most often present in the proximal bifurcation rim.11 At follow-up stent-strut coverage and neo-intima thickness can be assessed with great detail. We believe that OCT may be the technology that will give us the best insight into the mechanisms of in-stent restenosis in bifurcation stenting and allow for the evaluation of both two-stent techniques in conventional stenting and dedicated stents.

When the full potential of 3D OCT reconstruction is realised, it may be used for high-resolution hydrodynamic modelling of the stented bifurcation. We are currently participating in first-in-man studies and registries of dedicated bifurcation stents, using OCT extensively (see Figures 1D and 1E).

Review of Recent Trials and Evidence

Despite the complexity of bifurcation stenting and the high restenosis rates, particularly in the side branch, several advancements such as the introduction of drug-eluting stents and the more selective use of two-stent strategies – as described in the MADS classification – have reduced major adverse cardiac events. This reduction has been reflected in recent clinical trials.
Over the last few years, Colombo et al.,12 Pan et al.,13 Steigen et al.,14 Ferenc et al.15 and Tsuchida et al.2 have shown equivalent major adverse cardiac event (MACE) rates for one- and two-stent strategies, ranging from 3.4 to 19% (see Table 1). The meta-analysis of these trials including BBC-One (TCT 2008) has recently been published, and showed that there are no differences in mortality and TLR between one- and two-stent strategies.16 However, there is a 43% increase in peri-procedural myocardial infarction (MI) rate with two-stent strategy. This early difference (11.2 versus 3.6%) was particularly apparent and drove the overall MACE rate in the BBC-One trial of elective versus provisional T-stenting (15.2 versus 8%, hazard ratio [HR] 2.0; p=0.0009). The Cactus trial12 showed equivalent outcomes of crush-stenting compared with the provisional approach. One-third of the patients in the one-stent/provisional approach arm have crossed over to the two-stent arm due to residual stenosis in the side branch (72% of cases), poor flow (1.9%) or significant dissection in the side branch (39%). The Nordic 2 trial showed equivalent event rates with culotte versus crush-stenting. In addition, Colombo and his group have demonstrated that over the last several years as the use of kissing-balloon inflation increased, after crush-stenting the rate of restenosis and TLR has decreased by half.17 The rate of in-stent restenosis in the side branch remains at 13.2% with crush-stenting, even when kissing-balloon inflation is used.

Based on these studies, involving a total of 1,600 patients, the current recommendations for the approach to bifurcation stenting are as follows:
• protect the side branch with a guidewire that facilitates access and marks the ostium; • perform provisional rather than elective side-branch stenting only, with possible final kissing-inflation; and • two-stent treatment should be limited to flow-limiting side-branch dissection, side-branch disease extending well beyond the ostium or an unfavourable angle after main-branch stent implantation.
We have recently shown, in the ‘all-comers’ LEADERS study, that in the major European centres this strategy for true bifurcation treatment is followed in the majority of patients.1 Two-thirds of the patients have both branches wired and one-third of those are treated with a two-stent approach. In 88% of the lesions treated with a two-stent strategy, kissing-balloon inflation is performed. With these techniques and the use of drug-eluting stents the MACE rate for bifurcation lesions was 16% at 12 months. Even when following these guidelines, bifurcation stenting remains a technical challenge, the major issues being:
• difficulty in maintaining the side-branch access due to jailing of the wire and difficulty in re-crossing through double struts, especially in stent designs that have small cells;18 • distortion of the main-branch struts during side-branch dilatation (seen on micro-CT in vitro and OCT in vivo), which may be one of the reasons for in-stent restenosis and thrombosis; and • inability to fully cover the side-branch ostium with some angulations (e.g. shallow angles with T-stenting strategy).
Given these challenges, dedicated bifurcation stents that address some of these problems have been developed and, hopefully, will not only facilitate the procedure but also improve the long-term outcomes.

Dedicated Bifurcation Stents – What Progress Have We Made in the Last 10 Years?

Our group was one of the first to report the acute and long-term outcomes of the implantation of dedicated bifurcation stents such as the Jostent, NIR Side Royal and BARD stents.19,20 Since the cells in the first two stents were larger at the carina, the side branch had incomplete coverage. In particular, the true bifurcation stents were bulky and required a 9Fr guide catheter for delivery.

The alignment, especially with the ‘trapped-wire’ approach stents, was also an issue (see Figure 2). The design of the dedicated bifurcation stents has improved and most of them can be delivered via a 7Fr sheath for self-expanding designs and 5 or 6Fr sheath for the balloon-expandable designs. However, none of the current designs has solved all of the challenges.
Most of the currently available bifurcation stents are non-drug-eluting except for the Devax stent, which is coated with biolimus and delivered from a biodegradable polymer and the new BiPax (Nile Croco) paclitaxel-eluting stent. The bifurcation stents are meant to facilitate the following:
• main proximal first – Skirt in the MADS classification – ensures access to both branches and obviated the need for any re-crossing with the wires. The best-studied stent in this category is the Axxess Plus (Devax, Irvine, CA); • side-branch access after main branch-stenting – these stents are designed for provisional side-branch stenting21 (see Figure 3) – one of the examples of main branch across the side branch is the NILE CROCO stent (Minvasys, Genevilliers, France); and • stents for the side-branch first approach in the MADS classification – severely diseased large side branches may require this approach, which mimics classic culotte-stenting; however, without layers of overlapping struts, an example discussed here is the Tryton stent (Tryton Medical, MA, US).
Table 2 summarises dedicated bifurcation stent studies published to date according to the MADS classification.

Main Proximal Branch First

The best-studied example in this category of dedicated bifurcation stents is Axxess Plus (Devax), with a 302 patient prospective DIVERGE registry published recently in JACC.22 The stent is a 7Frcompatible, self-expandable, nitinol, single-wire system and elutes biolimus from a biodegradable polymer. It has a modular design, in that it can be tailored based on the bifurcation lesion anatomy, and provides access to both branches but may require three stents in total to complete coverage. DIVERGE enrolled 64% true bifurcation medina (1,1,1) lesions and 64% of the patients had both branches stented.

In-segment late loss in the side branch was only 0.17mm at nine-month angiographic follow-up. This translated into a 4.3% target lesion revasularisation (TLR) rate and 7.7% MACE rate at 12 months, which compares favourably to the results of a biolimus stent with a conventional approach to bifurcations in the LEADERS study, where the TLR rate was 11.1% and overall MACE rate was 16%. In addition, the MI rate was 4.3%, which is considerably lower than the 10% MI rate in the recent BBC-One trial.

Main Branch Across

Dedicated stents in this category are designed to provide easier access to the side branch for provisional stenting. All of the stents in this class have only been studied in small first-in-man studies or registries. One of the examples in this category is the Nile Croco stent, which is a 6Fr-compatible, balloon-expandable, cobalt– chromium (73μ strut thickness) stent. It uses a double-balloon, dual rapid-exchange system with two independent catheters that track over two wires. It is designed for ‘provisional’ side-branch stenting and provides partial side-branch ostium coverage/scaffolding (with a couple of struts). The main balloon has three markers, with the central marker indicating the position of the side-branch ostium. After stent deployment in the main branch, a side-branch balloon is advanced and a simultaneous kissing inflation can be performed. While theoretically this system should obviate the need for re-crossing, the wire twist often observed at the initial attempt of stent delivery necessitates the withdrawal of the wire from the side branch and re-crossing. Although there is no need to re-cross through the struts, this is often done after initial pre-dilation in the presence of potential dissection in the side-branch. The Nile Croco registry of 75 patients by Lefevre et al.23 demonstrated a MACE rate of 10.7% at six months. This included a 2.7% MI rate and a 6.7% TLR rate. Similarly, the Spanish registry24 showed excellent procedural success rates and a 12% MACE rate at six months. The NilePax paclitaxel-eluting stent is now available and is undergoing clinical testing in the Bipax study.
The Multi-link Frontier (Abbott Vascular, Santa Clara, C, US) was one of the first stents in this category and was studied in a 105-patient registry.25 It is a balloon-expandable 316L stainless steel stent pre-mounted on a dedicated delivery system with two balloons (monorail for the main branch and over-the-wire for the side branch) sharing a single inflation port. This stent was 7Fr-compatible and difficult to deliver in calcified vessels. While the procedural success rate was 93%, the MACE rate at six months was 17.1%, with a 44.8% overall restenosis rate (29% for both main branch and side branch). The new- generation Pathfinder, which incorporates the Xience V platform, is expected to offer better results.

Side Branch First

The best example in this category is the Tryton (Tryton Medical, MA, US), which is a 5 or 6Fr-compatible balloon-expandable cobalt– chromium slotted-tube bare-metal stent. It uses a single balloon and single rapid-exchange system. The stent consists of three zones: distal side branch, transition zone at the carina and main branch zone. The central transition zone has a specific geometry to provide the best scaffolding and is made of three elements, which can be independently deformed. The proximal main branch zone (the so-called collar) has three fronds and a minimal amount of metal, and allows for the delivery of a standard work-horse stent such as the drug-eluting Xience V stent (most often used in our institution). This design minimises the amount of overlapping and protruding struts while still providing adequate coverage and scaffolding. Tryton stent use commits users to the two-stent strategy, similar to culotte.
The first-in-man study of 30 patients with six-month follow-up showed excellent results, with an overall MACE rate of 11.2%. This was composed of an MI rate of 6%, a TLR rate of 7.5% and a stent thrombosis rate of 2.2%.26 More interestingly, the six-month angiographic follow-up in this study showed an extremely low late loss of 0.17mm in all three bifurcation segments, including the side branch. This is despite the fact that the side-branch portion is covered by a bare-metal stent only. Long-term follow-up in a larger population will be needed to confirm the results. We have been able to successfully use Tryton in very complex cases including left main bifurcations and chronic total occlusions at bifurcation sites with good procedural success and we are performing a careful OCT evaluation of the stent post-procedure in this ‘all-comers’ population (see Figure 1E).
The Side-guard ostium protection device (Capella, MA, US) is a self-expanding trumpet-shaped nitinol device with a low profile that allows for T-stenting and is designed for bifurcation angles between 45 and 135º. The six-month results in the first 20 patients showed a MACE rate of 12.5%.

As illustrated, none of the dedicated bifurcation stents has been studied in large randomised trials with long-term follow-up. The data are limited to small first-in-man and registry studies. While it appears that dedicated drug-eluting bifurcation devices may halve the MACE rate compared with historical controls (MACE of 7.7% in DIVERGE versus 16% in LEADERS), these promising preliminary data must be confirmed.


Bifurcation stenting is an unsolved challenge for interventional cardiologists. Patients with bifurcation lesions tend to have more advanced disease and multiple co-morbidities. Over the last 10 years we have understood the importance and made progress in bifurcation imaging for planning of procedures and in procedural success (from 3D angiography, IVUS and MSCT to FFR and OCT). Recent clinical trials show an improvement in results with more selective use of a two-stent strategy and drug-eluting stents. Rates of MACE and peri-procedural MI for bifurcation lesions still remain higher than in non-bifurcation lesions and patients. Many challenges, such as side-branch access, wire trapping and incomplete side-branch coverage, still remain and have only partially been addressed by the new dedicated bifurcation stent designs since their development in the 1990s. More long-term prospective studies of these novel stent designs are needed for the field to progress with concomitant imaging. We believe that in addition to randomised trials of the dedicated bifurcation stents, an ‘all-comers’ registry of these devices, with long-term follow-up, would be invaluable in the assessment of their safety and efficacy compared with conventional approaches. Table 3 summarises our current practical approach to bifurcation treatment. Ôûá


  1. Garg S, Wykrzykowska J, de Vries T, et al., EuroPCR presentation. Available at: (accessed 21 May 2009).
  2. Tsuchida K, Colombo A, Lefevre T, et al., Euro Heart J, 2007;28(4):433–42.
    Crossref | PubMed
  3. Medina A, Suarez de Lezo J, Pan M, Revista espanola de cardiologia, 2006;59(2):183.
    Crossref | PubMed
  4. Lansky A, Tuinenburg J, Costa M, et al., Catheter Cardiovasc Interv, 2009;73(2):258–66.
    Crossref | PubMed
  5. Ramcharitar S, Onuma Y, Aben JP, et al., Eurointervention, 2008;3:553–7.
    Crossref | PubMed
  6. Girasis C, EuroPCR 2009, Barcelona, Spain.
  7. Van Mieghem CA, Thury A, Meijboom WB, et al., Euro Heart J, 2007;28(16):1968–76.
    Crossref | PubMed
  8. Girasis C, Onuma Y, Kukreja N, et al., I2 Summit, ACC 28–31 March 2009, Orlando, FL, US, J Am Coll Cardiol, 2009;53(10):A77 (Suppl. A).
  9. van der Waal EC, Mintz GS, Garcia-Garcia HM, et al., Catheter Cardiovasc Interv, 2009;73(7):910–16.
    Crossref | PubMed
  10. Koo BK, Kang HJ, Youn TJ, et al., J Am Coll Cardiol, 2005;46(4):633–7.
    Crossref | PubMed
  11. Gonzalo N, Garcia-Garcia HM, Regar E, et al., JACC, 2009;2(4):473–82.
    Crossref | PubMed
  12. Colombo A, Bramucci E, Sacca S, et al., Circulation, 2009;119(1):71–8.
    Crossref | PubMed
  13. Pan M, de Lezo JS, Medina A, et al., Am Heart J, 2004;148(5):857–64.
    Crossref | PubMed
  14. Steigen TK, Maeng M, Wiseth R, et al., Circulation, 2006;114(18):1955–61.
    Crossref | PubMed
  15. Ferenc M, Gick M, Kienzle RP, et al., Euro Heart J, 2008;29(23):2859–67.
    Crossref | PubMed
  16. Brar S, Gray W, Dangas G, et al., Eurointervention, 2009; in press.
  17. Latib A, Colombo A, Sangiorgi GM, Heart (British Cardiac ociety), 2009;95(6):495–504.
    Crossref | PubMed
  18. Mortimer P, van Loo D, De Beule M, et al., Eurointervention, 2008;4(3):391–6.
    Crossref | PubMed
  19. Cervinka P, Foley DP, Sabate M, et al., Catheter Cardiovasc Interv, 2000;49(1):105–11.
    Crossref | PubMed
  20. Carlier SG, van der Giessen WJ, Foley DP, et al., Catheter Cardiovasc Interv, 1999;47(3):361–96.
    Crossref | PubMed
  21. Louvard Y, Thomas M, Dzavik V, et al., Catheter Cardiovasc Interv, 2008;71(2):175–83.
    Crossref | PubMed
  22. Verheye S, Agostoni P, Dubois CL,et al., J Am Coll Cardiol, 2009;53(12):1031–9.
    Crossref | PubMed
  23. Lefevre T, Pavlides G, on behalf of the Nile Registry investigators. Available at: medtool/webmedtool/icpstool01/stud0112/html/Nile/fra me.htm (accessed 2 December 2007).
  24. Garcia Del Blanco B, Brenot P, Royer T, et al., Interventional Cardiology, 2008;46–50.
  25. Lefevre T, Ormiston J, Guagliumi G, et al., J Am Coll Cardiol, 2005;46(4):592–8.
    Crossref | PubMed
  26. Onuma Y, Müller R, Ramcharitar S, et al., Eurointervention, 2008;3(5):546–52.
    Crossref | PubMed