Simultaneous kissing stents to treat unprotected left main stem coronary artery bifurcation disease; stent expansion, vessel injury, hemodynamics, tissue healing, restenosis, and repeat revascularization

Objectives To perform detailed analysis of stent expansion, vessel wall stress, hemodynamics, re‐endothelialization, restenosis, and repeat PCI in the simultaneous kissing stents (SKS) technique of bifurcation left main stem (LMS) stenting. Background The SKS technique is useful to treat patients with true bifurcation disease of the LMS but remains controversial. Methods and Results Computational structural analysis of SKS expansion demonstrated undistorted and evenly expanded stents. Computational fluid dynamics modelling revealed largely undisturbed blood flow. 239 PCI procedures were performed on 217 patients with unprotected bifurcation LMS disease with SKS using DES (2004‐2017). We electively studied 13 stable patients from baseline to 10 years post‐SKS with repeat angiography and optical coherence tomography, and demonstrated tissue coverage of the stent struts at the carina, with no evidence of lacunae behind the stents. We studied all patients with symptomatic recurrence. Target lesion revascularization rate was 3.2% at 1 year and 4.6% at 2 years. Of all 20 patients with restenosis, the site was the LMS‐Cx stent in 7, the LMS‐LAD stent in 2 and both in 11. Two‐year recurrence rate was 7/32 (5.3%) for first, and 4/108 (3.7%) for second generation DES. Treatment with repeat kissing techniques was undertaken in 19/20, with sustained clinical results with re‐SKS. Conclusion The SKS technique for treating unprotected LMS bifurcation disease does not distort the stents, is associated with favorable hemodynamics, tissue coverage of the exposed struts, and a low restenosis rate when performed with contemporary stents. Re‐PCI with repeat SKS appears feasible, safe, and durable.


| I N TR ODU C TI ON
With modern drug-eluting stents, intravascular imaging, physiological guidance, improved patient selection, and insight from large contemporary trials such as EXCEL and NOBLE, percutaneous coronary intervention (PCI) is assuming an increasing role in the treatment of left main stem (LMS) disease [1,2].
Approximately 70% of lesions affecting the LMS involve the bifurcation. Percutaneous coronary intervention (PCI) of such lesions is both more challenging and associated with higher risk than for those of the body or ostium of the LMS [3]. A single stent 'provisional' technique is the most widely used, especially if the circumflex (Cx) is large and minimally diseased, or small and nonsignificant. The results of this approach are excellent, provided crossover to a two-stent technique is avoided [4]. Unlike other bifurcations, however, the implications of losing flow to the Cx, especially when it is large and diseased, are substantial, and 'bailout' two-stent techniques are often less than satisfactory. A twostent strategy with simultaneous kissing stents (SKS) for true bifurcation disease of the LMS has gained some favor, particularly in the emergency setting. The first cases of SKS were reported by Colombo [5], Teirstein [6], and Sharma [7]. We conducted preclinical and early clinical assessment in a series of 30 patients in the first-generation DES era, and confirmed its simplicity, high acute success rate, procedural speed and safety, with systematic angiographic follow-up [8]. We also provided the overall outcomes of our first 150 unselected cases [9].
The SKS procedure has previously been described in detail [7], but it remains controversial chiefly because of perceived problems related to the artificially created carina and the feasibility and approach to re-intervention.
We therefore now present a technical analysis of SKS to address those concerns, and focus on the mechanics of stent expansion, the hemodynamics, the extent of re-endothelialization of the carina, the rate and pattern of restenosis, and the feasibility and safety of repeat PCI. other coronary bifurcation stenting techniques [10][11][12]. Two Resolute Integrity TM stents (Medtronic, USA) were deployed in SKS pattern in an idealized geometry comprising a 608 bifurcation measuring 4.0 mm at the LMS, 3.0 mm at the LAD and 2.5 mm at the Cx, a configuration obeying Finet's law ( Figure 1). The stent sizes were 3 3 15 mm for the LAD and 2.5 3 14 mm for the Cx. The balloon model of each stent was calibrated to comply with the pressure-diameter curve given by the manufacturer. The structural analysis focused on the geometry of the deployed stents, deformation, and areas of malapposition. Furthermore, the vessel wall stress induced by the SKS and the stent stress after SKS were evaluated.

| Computational fluid dynamics (CFD) model of SKS
We constructed a CFD model based upon the structural simulations to analyze the velocity field, regions of disturbed flow and wall shear stress immediately after SKS procedure. A pulsatile CFD simulation was performed following a previously described procedure [11,13]. A typical human LMS flow waveform was applied at the inlet [14]. A flow-split of 0.6:0.4, calculated using a diameter-based scaling law [15], was imposed between the LAD and Cx. The lumen and the stents were considered rigid with a no-slip wall-boundary condition. Blood was modeled as a non-Newtonian fluid.  Recruits had to be willing to return for re-catheterization, with no contraindications; namely frailty, peripheral vascular disease, renal impairment, anticoagulation, a bleeding tendency or inability to tolerate lying down. Local Ethics Committee approval was obtained (STH16312; UK REC 12/YH/0010). The patients underwent repeat coronary angiography with an 8F guide catheter via the femoral artery or a 7.5F sheathless catheter via the radial artery. Angiographic images were recorded.

| Patients and setting
With dual antiplatelet drug cover, and after a standard dose of heparin, a guidewire was passed down each branch of the SKS and an optical coherence tomography (OCT) catheter (Optis TM , St. Jude Medical, St. Paul, MN) advanced through the SKS wherever possible. Standard pullback imaging was then recorded through the LMS stents. The images were recorded and analyzed offline with particular attention to tissue coverage of the exposed metal carina.

| Restenosis rate and angiographic pattern
All patients with recurrent symptoms underwent repeat coronary angiography with a view to repeat PCI during the same procedure. We

| Percutaneous treatment of SKS restenosis
A guidewire was passed down each 'limb' of the SKS, predilatation was performed, and definitive treatment given with simultaneous expansion of balloons (drug eluting when available) or stents, aiming to increase the size of the SKS by 0.5 mm in each stent. The details of revascularization and the acute success rate were recorded. The long-term results of patients treated with re-PCI were also documented.

| Statistical analysis
Descriptive data are presented as mean [standard deviation (SD)], median [interquartile range (IQR)], or as n (%), as appropriate. Betweengroup differences were analyzed using Student's t tests for continuous data and Chi-square tests for categorical data. Statistical significance was accepted at the 95% level.

| Computational structural model of SKS
The results of the structural analysis of SKS are shown in Figure 1

| CFD model of SKS
The velocity field is shown in Figure 2. The presence of the 'virtual' metallic carina caused flow separation in the proximal LMS with the creation of two inner channels and an increase of the flow velocity.
These two channels behaved as an extension of the two daughter   Figure 2. Low TAWSS were located close to the stent struts and at the metallic carina. The percentage lumen area exposed to TAWSS lower than 0.4 Pa was 20.3% in the stented region. The median TAWSS was 0.58 Pa.

| Patients
During the period of the registry, 601 ULMS PCI procedures were performed. Thirty of these were in the BMS era and were excluded, 571 LMS procedures were performed using DES. Of these, 239 (42%) were with SKS. The remaining 332 cases comprised 289 single/provisional stents, 24 T-stents, 14 POBA, one culotte, two mini-crush and two unsuccessful procedures. Details of the 217 patients undergoing SKS for ULMS disease are shown in Table 1, and this includes 20 patients re-treated for SKS restenosis. Median follow up was 8.3 years. The mortality rate was 3.3% at 1 year and 5.2% at 2, 64% of these having originally been urgent or emergency cases.

| Symptomatic recurrence
Of the 217 patients who underwent SKS, the running total of clinical recurrence was 7 by 1 year (3.2%), 10 by 2 years (4.6%), and 16 by 5 years (7.4%). A further four patients re-presented between years 5 and 8. The timelines of recurrence are displayed in Figure 5 and the clinical and procedural details are presented in Table 3. For all 20 patients with recurrence, their status at the original procedure (left side of Table 3) was as follows: mean age 64.1 years, 13 male (65%), and 2 diabetic, 9 elective, and 11 nonelective. The median (IQR) estimate of in-patient mortality by the logistic EuroSCORE-1 was 2.83% (0.88-11.1%). The  Table 3. The mean time to presentation was 30.8 (SD 32) months. Presentation was with chronic stable symptoms in 9, urgent in 7, and as an emergency procedure in 4. The recurrence rate at 2 years for the first generation DES was 7/132 (5.3%) and for the second gen-

| Repeat revascularization
Of the 20 patients with clinical recurrence, one was treated with CABG because of occlusive in-stent restenosis in the right coronary artery

| D ISC USSION
This is the first analysis of SKS that addresses the detailed technical issues related to the sustainability of the SKS technique. Computational modelling revealed even expansion of the stents to their nominal diameter, without distortion, uneven cell opening or large gaps. This contrasts with the pattern seen in stents used with other bifurcation techniques. However, the 'virtual' artery was undiseased and, in the 'real world', with eccentric calcific plaque formation, it is common to see differential expansion at moderate pressure with any technique.
Each stent in the LMS take on an oval cross-section, seen in our OCT images. We also noted that the two inner channels exhibited increased blood flow velocity in the proximal LMS. Largely undisturbed blood flow was found through the stents, with only a small recirculation zone at the ostium of the Cx [16,17]. This reflects the smooth curve between LMS and branch arteries, and contrasts with two-stent strategies with more angulation, such as 'T' [18,19]. The gaps in the LMS between the stents and the LMS wall matched our findings of SKS deployment in undiseased porcine arteries in our first publication [6], but have not been seen clinically, and none of our OCT images showed this. This might happen clinically in the rare case of a highly localized tight bifurcation stenosis with a long and completely normal LMS; but in such a case SKS would not be recommended.
The OCT study of re-endothelialization and healing was driven by concern about the exposed metal carina. We observed that all of the stent struts at the carina were endothelialized from 4 months onwards.
Furthermore, of the 13 patients studied, seven had a carina that was completely covered with tissue without visible fenestrations. Differences in coverage of SKS may also depend on the stent type, particularly older generation DES [20]. Having previously recommended lifelong dual antiplatelet therapy for patients with SKS, with the advent of thinstrut second generation DES we now recommend a duration of 3 years.
The clinical recurrence rate in our patients was 3.2% at 1 year and 4.6% at 2 years. This compares favorably with other 'real world' Abbreviations: Cx, circumflex; ISR, in-stent restenosis; LAD, left anterior descending; LMS, left main stem; Mixed c , a complete, unfenestrated diaphragm at the carina but with fenestrations proximally towards the LMS ostia. a Case 14 was studied ten years after re-do SKS. b Case 15 was studied twelve months after re-do SKS.   Table 3) angiography was only performed for recurrent symptoms, so subclinical, mild, angiographic restenosis could not be documented.

| C ONC LUSI ON
SKS is an effective PCI treatment for patients with disease at the LMS bifurcation who require a two-stent approach, and is particularly useful in the emergency situation. The stents are deployed with minimal distortion, the blood flow pattern is favorable, the struts of the neocarina become covered with tissue either individually or in a continuous sheet, the clinical restenosis rate is low, and re-PCI is safe and feasible with a repeat SKS approach.

ACKNOWLEDGMENTS
We are grateful to the staff of the Cardiac Catheter Suite, Northern