Interventions for Left Main Coronary Artery Disease

Registry Data

In the large, multinational, “all-comers” IRIS-MAIN (Interventional Research Incorporation Society–Left MAIN Revascularization) registry,7 over the past two decades increasingly complex patients with worsening risk-factor profiles have been recruited with improvements in concomitant medical therapy. Over time, the proportion of PCI treatment has progressively increased to approximately 61%. Risk-adjusted survival, composite outcomes, and repeat revascularization have significantly improved for PCI over time but have relatively remained stable for CABG narrowing the gap in the treatment effect between PCI and CABG from the BMS (bare metal stent) period to the early DES period and then to the late DES period. The risks of mortality and the composite of death, myocardial infarction (MI), or stroke (9.1 vs. 9.7%; hazard ratio [HR] 0.97; 95% confidence interval [CI]: 0.81–1.17; p = 0.75) were equivalent between PCI and CABG, but the risks of repeat revascularization and major adverse cardiac and cerebrovascular events (MACCE) were higher in the PCI group than in the CABG group.

The results of DELTA-2 (DrugEluting Stent for Left Main Coronary Artery) registry8—an international, all-comers, multicenter registry—were compared with those from the historical DELTA-1 CABG cohort using propensity score stratification. The primary endpoint was the composite of death, MI, or stroke at the median time of follow-up. The distal LMCA was involved in 84.6% of the lesions. The primary endpoint of death, MI, or cerebrovascular accident was lower in the PCI DELTA-2 group compared with the historical DELTA-1 CABG cohort (10.3% vs. 11.6%; p = 0.03) at 17 months of follow up. A favorable result was observed for PCI with respect to cerebrovascular accident whereas CABG was more advantageous with respect to target vessel revascularization.

Randomized Trials

The SYNTAX trial9 included a subset of 705 patients with LM disease. There was no significant difference in MACCE (death, MI, stroke, and repeat revascularization) at 5 years (36.9% in PCI patients and 31% in CABG patients) (HR 1.23; 95% CI: 0.95–1.59; p = 0.12). Mortality was 12.8 and 14.6% in PCI and CABG patients, respectively (HR 0.88 [0.58, 1.32]; p = 0.53). Stroke was significantly increased in the CABG group and repeat revascularization in the PCI arm. MACCE was similar between arms in patients with low/intermediate SYNTAX scores but significantly increased in PCI patients with high scores (≥ 33).

In the EXCEL (Evaluation of XIENCE versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization) trial,10 1,905 patients—with ULMCA disease with ≥ 70% diameter stenosis, or ≥ 50% or ≤ 70% with either (1) noninvasive evidence of LM ischemia, (2) intravascular ultrasound-derived minimal lumen area (IVUS-MLA) ≤ 6.0 mm2, or (3) fractional flow reserve (FFR) ≤ 0.80, SYNTAX score ≤ 32, and clinical and anatomic eligibility for both PCI and CABG as agreed to by the local heart team—were randomized to PCI with fluoropolymer-based cobalt-chromium everolimus-eluting stents or CABG. The primary endpoint, the composite of death from any cause, stroke, or MI at 3 years occurred in 15.4% of patients who underwent PCI group versus 14.7% of the patients who underwent CABG group meeting the noninferiority criterion. The secondary endpoint event of death, stroke, or MI at 30 days was lower in the PCI group (4.9% in the PCI group vs. 7.9% in the CABG group). The secondary endpoint event of death, stroke, MI, or ischemia-driven revascularization at 3 years occurred in fewer patients in PCI group. CABG patients had fewer primary endpoint events than the PCI group patients between 30 days and 3 years after the procedure. Revascularization rates at 3 years were higher with PCI than with CABG (12.6 vs. 7.5%) whereas the rates of early MI and major adverse events were lower with PCI than with CABG (8.1 vs. 23.0%).

The EXCEL trial enrolled predominantly males (76%). Most patients were clinically at low risk (60% had stable ischemic heart disease and most patients had normal ejection fraction). Thirty percent were diabetic. According to the sites, the SYNTAX was low (≤ 22) in 60.5% of the patients and intermediate (23–32) in 39.5% of the patients. The core laboratory assigned higher scores, low in 35.8% of the patients, intermediate in 40.0%, and high (≥ 33) in 24.2%, the group that met the exclusion criteria. Most of the patients (80.5%) had distal left main bifurcation or trifurcation disease, and two- or three-vessel coronary artery disease was present in 51.3% of the patients. Average 1.9 lesions were treated per patient. A mean of 2.4 stents with a mean total stent length of 49.1 mm were implanted per patient. Femoral access was used in 73% patients, and IVUS guidance was used in 77% of patients. Off-pump CABG was performed in 29%. Internal mammary artery (IMA) was used in 98.8% with bilateral IMA in 28.8% patients.

NOBLE (Nordic-Baltic-British Left Main Revascularization Study) trial11 randomized patients with significant LM lesion visually assessed stenosis diameter ≥ 50% or FFR ≤ 0.80 and no more than three additional noncomplex lesions (additional non-LM complex lesions were chronic total occlusions, bifurcation lesions requiring two-stent strategies, calcified or tortuous lesions). A total of 1,201 patients were randomly assigned to PCI or CABG. Kaplan-Meier 5-year estimates of MACCE (death from any cause, nonprocedural MI, repeat revascularization, or stroke) were 29% for PCI and 19% for CABG (HR 1.48; 95% CI: 1.11–1.96) with CABG being significantly better than PCI. Though all-cause mortality was not different, the rates of MI, revascularization, and stroke were higher in PCI patients compared with CABG patients. Repeat revascularization was higher due to de novo lesion and target non-LMCA lesion revascularization. Notably 1-year MACCE was same in both the groups. Surprisingly there was no association with SYNTAX score and MACCE.

Twenty percent of the population was female with 15% diabetics. Eighty percent had stable angina and most had preserved ejection fraction. Eighty-eight percent had distal LM lesion. Elective double stenting was used in 35% with culotte being preferred technique. 11% of PCI cases received a first-generation DES in the LM. Kissing balloon inflation (KBI) was done in 55%, and any ostial circumflex post-dilatation was done in 79%. Complete revascularization was achieved in 92%. IVUS was done pre-PCI in 47% and post-PCI in 74%. CABG was done with the on-pump technique in 84%, with 93% patients receiving left internal mammary artery. The number of grafts per patient was one in 4%, two in 52%, three in 39%, four in 4%, and five in < 1% of patients.

Conflicting Results of EXCEL and NOBLE

Both were well-conducted robust trials. Both studies included similar patients with similar SYNTAX scores, and distal LM was treated in the majority. The conflicting results could be due to:

1. Differences between studies in patient assessment, risk profiles, trial process, or procedural characteristics.

2. Differences in the stent used: EXCEL used a thin strut fluropolymer-based cobalt chromium everolimus-eluting stents whereas NOBLE used first-generation sirolimus-eluting Cypher stent (11%) or the biolimus-eluting Biomatrix Flex stent (89%), resulting in a substantial difference in stent thrombosis (0.7 vs. 3%). As a result, the rate of stent thrombosis or symptomatic graft occlusion was much higher after CABG than after PCI (5.4 vs. 0.7%) in EXCEL but similar (4 vs. 3%) in NOBLE.

3. Strategy of stenting: Both the trials used provisional stenting approach. In approximately 75% of patients IVUS was used. Data were not available from EXCEL regarding two-stent strategy. Thirty-five percent needed a two-stent strategy with culotte technique being used in the majority in NOBLE.

4. Differences in primary endpoint used: EXCEL used death, MI, and stroke. Investigators included periprocedural MI. The EXCEL trial did not include repeat revascularization. NOBLE trial, on the other hand, used only nonprocedural MI in the endpoint, and repeat revascularization was part of primary composite endpoint. In the NOBLE trial, there was not much difference in target LM artery revascularization but more than twofold increase in de novo lesion revascularization in the PCI group with most patients undergoing repeat PCI (96%).

5. Difference in trial duration: EXCEL reported data at 3 years whereas NOBLE reported data at 5 years. Remarkably in NOBLE, the 5-year mortality was not different in PCI and CABG groups whereas death, stroke, and repeat revascularization were higher in PCI group. In fact, data from EXCEL may look similar to NOBLE if repeat revascularization is included with evaluation done at 5 years.

6. An unexplained higher risk of stroke after PCI in NOBLE, which was more than twice after PCI than CABG.

Overall, 4,594 patients from five trials (EXCEL, NOBLE, SYNTAX, PRECOMBAT, Boudriot et al) were included in the meta-analysis by Nerlekar et al,12 with 50% of patients in the metanalysis undergoing PCI using DES. The prevalence of isolated ULMCA stenosis ranged from 10 to 29%, with between 55 and 80% of patients having a distal bifurcation ULMCA lesion. The primary endpoint of death, MI, and stroke was similar between PCI and CABG (odds ratio [OR] 0.97; 95% CI: 0.79–1.17; p = 0.73). As expected, CABG was associated with a significant reduction in the risk of repeat revascularization.

Access and Guiding Catheter

Radial access with large lumen 6F guiding catheters can be used. In case of an aorto-ostial ULMCA lesion, a 6F guide catheter is suitable for most of cases, except when debulking is planned and a larger guiding catheter and a stronger support are needed. For very tight aorto-ostial ULMCA lesions or severe damping of the guiding catheter, it is safer to place a guidewire in the guiding catheter, to be able to wire the LMCA LAD or LMCA LCX and promptly disengage the guiding catheter from the LMCA ostium to avoid catheter damping. It is preferable to use a 7F guiding catheter in distal LM when elective double stenting is planned or when > 3.5-mm balloons are needed for kissing. The advantages of a 7F catheter are better visualization, decreased friction during advancement of the stents, accommodation of all sizes of burr if rotablation is required, and allowing for IVUS to be performed in the presence of multiple guidewires.

Ostial and Shaft Lesions

The operator must choose the best angiographic view (AP cranial or LAO-cranial) to visualize the ostium of LM and properly position the stent. The implantation of the stent for ostial LM stenosis must be done with a small protrusion into the aorta. After implantation, it is important to post-dilate the stent with properly sized balloons to facilitate good stent apposition at the ostium and allow reengagement of the ostium with the catheter (Fig. 2). The results of LM ostial and shaft lesions are gratifying according to DELTA registry22 though first-generation DES were used. At 3 years, the rate of major adverse cardiac events (MACE) was 19.1% versus 28.5% for distal LM, and the rate of TLR was 4.5% as against 12.6% for distal LM.

Close

Fig. 2 (A) Left main ostial stenosis in 45-year-old woman with unstable angina. (B) Positioning of stent in left main in AP cranial view. (C) Stent deployment. (D) Stent boost. (E) Post-dilatation of stent (F) Final result.

Export to PPT

Fig. 2 (A) Left main ostial stenosis in 45-year-old woman with unstable angina. (B) Positioning of stent in left main in AP cranial view. (C) Stent deployment. (D) Stent boost. (E) Post-dilatation of stent (F) Final result.

Wiring

LAD and LCX should be wired with the most difficult branch wired first. Jailing of the wire in the SB during main vessel stenting is important as it keeps the SB open and is the only marker for rewiring when SB is occluded. The SB wire changes bifurcation angle that facilitates rewiring. Polymer-coated wires were more resistant to retrieval damage and were more efficient in crossing the SB ostium than nonpolymer-coated wires,23 but another study reported that jailed polymer-jacketed guidewires were associated with procedural MI.24

Lesion Preparation

Lesion is predilated with noncompliant and scoring balloons. Rotational atherectomy is needed for severely calcified LM lesions.

Stenting Strategy—One versus Two Stents

Several studies have reported that the provisional one-stent approach for distal LM bifurcation was associated with more favorable outcomes compared with two-stent techniques due to lower risk of death, MI, and target vessel revascularization,25 26 27 including a lower risk of stent thrombosis.27 A meta-analysis of seven observational studies involving 2,438 patients with mean follow-up of 32 months showed decreased risk of MACE with provisional versus two-stent strategy (20.4 vs. 32.8%; OR: 0.51; 95% CI: 0.35–0.73), and also decreased target vessel/target lesion revascularization (10.1 vs. 24.3%; OR: 0.35; 95% CI: 0.25–0.49) favoring provisional stenting as against two-stent strategy.28 In a nonrandomized, retrospective study including all consecutive patients with 50% stenosis of the LM treated with DES in nine European centers between 2002 and 2004, 62.5% underwent provisional stenting whereas 37.5% were in the two-stent group. After 10 years, no differences in TLR or MACE were detected.29

Single-Stent Strategy

European Bifurcation Club (EBC) 12th consensus30 recommends a provisional SB stenting approach for LM treatment in most cases (Fig. 3). It was the strategy used in EXCEL and NOBLE trials with 35% patients requiring a two-stent strategy in NOBLE trial. A simple lesion defined as SB stenosis < 70% and lesion length < 10 mm based on criteria described in the DEFINITION study21 favors a provisional approach. Other factors such as a small LCX < 2.5 mm in diameter or diminutive LCX in a right dominant coronary system and a wide angle between LAD and LCX favor provisional approach.

Close

Fig. 3 (A) CAG revealing LM bifurcation lesion involving proximal LAD and LCX beyond ostium (Medina 1, 1, 0). (B) FFR in LCX to decide provisional versus two-stent technique. (C) FFR in LCX 0.97. (D) Mid LAD stenting. (E) Left main to LAD stent. (F) POT balloon marker at carina. (G) Proximal optimization. (H, I) Final result showing good result with no stenosis of ostial LCX.

Export to PPT

Fig. 3 (A) CAG revealing LM bifurcation lesion involving proximal LAD and LCX beyond ostium (Medina 1, 1, 0). (B) FFR in LCX to decide provisional versus two-stent technique. (C) FFR in LCX 0.97. (D) Mid LAD stenting. (E) Left main to LAD stent. (F) POT balloon marker at carina. (G) Proximal optimization. (H, I) Final result showing good result with no stenosis of ostial LCX.

Second-generation DES should be the first choice. The stent platform should be selected keeping in mind the maximum expansion capability of the stent. The stent size should be selected according to the distal reference diameter to avoid carinal shift. Single-stent crossover from LM into the LAD is the most common approach (Fig. 4). If the predominant lesion is in the LCX and LAD ostium is spared, provisional stenting from LM toward the LCX can be done (inverted provisional). While planning to stent toward the LAD or LCX, it is important to remember that stenting toward the tightest vessel rather than the largest vessel has better outcomes and reduces the number of stents used.31

Close

Fig. 4 Provisional stenting approach.

Export to PPT

Fig. 4 Provisional stenting approach.

Proximal Optimization

Proximal optimization (POT) is performed after stenting by inflating a short balloon just proximal to the carina, to change the tubular stent to a tapered device fitting the LM and distal MB either LAD or LCX, respecting the anatomy of the bifurcation core segment.30 Care must be taken so that at least 6 to 10 mm of stent length is proximal to the carina (the smallest length of commonly available balloons). Careful positioning of the balloon for POT is crucial and may influence the final result: If too distal, it increases the risk of SB occlusion; if too proximal, it has no effect on the stent strut toward the SB. Ideally, the distal shoulder of the balloon should be positioned just proximal to the carina while the proximal part is still in the stent to avoid geographical miss. The main problem is that the positioning of the distal marker compared with the distal shoulder varies among the different balloons currently available. Compliant or noncompliant balloons in a 1:1 relation to proximal reference diameter of the LM can be used. POT apposes the stent to LM, reduces ellipticity of stented segment, and prevents accidental abluminal wiring. POT allows strut protrusion into the SB with larger strut opening and minimizes carinal shifting for easier guidewire exchange.

After POT there are three options for the SB (Fig. 4):

1. Leaving the SB alone if the result is acceptable.

2. If intervention is required to the SB, SB is rewired through the most distal strut (closest to carina) with either a new wire or by pulling back the MB wire. The jailed wire should be withdrawn carefully to avoid deep intubation of the guiding catheter and consequent damage to the stent. SB is opened with a short noncompliant balloon, and re-POT is done to restore stent distortion opposite the SB. (POT-side-POT [PSP])

3. After wire exchange, KBI is performed with two short, preferably noncompliant balloons sized according to the distal diameters. Re-POT is performed (POT-KBI-POT).

The clinical outcomes of recent studies comparing KBI versus non-KBI failed to show its advantage over non-KBI treatment in terms of the rate of MACEs, despite lowering SB diameter stenosis at 6- to 12-month follow-up.32

According to the Korean Coronary Bifurcation Stenting (COBIS) registry, KBI was associated with a higher MACE rate due to higher TLR rather than death or MI.33 The COBIS II registry demonstrated the opposite result with reduced MACE and TLR with KBI. Occlusion of the SB after MV stenting was associated with more frequent cardiac death and MI, which suggested the importance of KBI to secure SB patency.34 KBI procedure can be optimized by final POT to correct proximal malapposition, proper guidewire cross in the distal cell to optimize SB strut opening and selecting the appropriately sized balloons for optimizing the geometry.32 To reduce proximal deformation, a “modified KBI approach” was recently proposed, using asymmetric inflation pressures: the SB is first inflated to 12 atm, then partly deflated back to 4 atm with simultaneous inflation of the MB balloon at 12 atm. The modified FKBI procedure reduces elliptical stent deformation and optimizes SB access.35

Two-Stent Techniques

A complex LMB lesion according to DEFINITION criteria21 has SB diameter stenosis > 70%, and lesion length > 10 mm may necessitate a two-stent technique. This corresponds to Medina classification 1, 1, 1 or 1, 0, 1 or 0, 1, 1. Large LCX ≥ 2.5 mm in diameter or diseased left dominant coronary system also may necessitate two-stent techniques.

T-stenting offers good SB scaffolding with angles > 70 degrees but could leave a potential gap at SB ostium or protrusion of SB stent into the MB (in the case of TAP). Variation of the T-stenting, the mini-crush technique was developed involving minimal (usually 1–2 mm) retraction of the SB stent into the MB before crushing with a stent or with a balloon followed by stent. It is relatively simple, leaves the SB open, and provides good coverage of the SB though it leaves multiple layers of strut. However, due to the low success rate of final KBI and suboptimal long-term outcome in the absence of KBI, this procedure has been modified by Chen et al as the double kissing (DK)-crush technique.38

Double Kissing Crush Technique

This technique consists of stenting the SB, completely crushing the SB stent with main vessel balloon sized 1:1 to the proximal vessel diameter, proximal SB recross, first KBI, MV stenting, second SB recross and second KBI, followed by final POT (Figs. 6, 7). Another key step in the procedure is the alternative inflation with a noncompliant balloon at high pressure (≥ 16 atm) for the SB before each kissing. First kissing can optimize the distorted SB stent, enlarge the cell of the SB stent, and leave only one layer of struts at the ostial SB, which probably facilitates the second kissing after stenting the MV. FKBI was successfully performed in 100% of cases by DK crush.

Close

Fig. 6 Double-kiss (DK) crush technique. (A) Baseline CAG revealing significant LM bifurcation lesion involving proximal LAD and LCX (Medina 1, 1, 1) (B) Significant calcification in LM, LAD, and LCX and IVUS catheter did not cross. (C) Rotablation of LCX. (D) Rotablation followed by balloon dilatation of LAD. (E) After adequate lesion preparation. (F) Stenting of LM-LCX. (G) Crushing of LCX stent with noncompliant balloon placed in LM-LAD. (H) First kissing balloon inflation after LCX recross. (I) Deployment of stent in LM-LAD. (J) Second kissing balloon inflation after high-pressure sequential dilatation of LAD and LCX. (K) POT in LM with short noncompliant balloon. (L) Result.

Export to PPT

Fig. 6 Double-kiss (DK) crush technique. (A) Baseline CAG revealing significant LM bifurcation lesion involving proximal LAD and LCX (Medina 1, 1, 1) (B) Significant calcification in LM, LAD, and LCX and IVUS catheter did not cross. (C) Rotablation of LCX. (D) Rotablation followed by balloon dilatation of LAD. (E) After adequate lesion preparation. (F) Stenting of LM-LCX. (G) Crushing of LCX stent with noncompliant balloon placed in LM-LAD. (H) First kissing balloon inflation after LCX recross. (I) Deployment of stent in LM-LAD. (J) Second kissing balloon inflation after high-pressure sequential dilatation of LAD and LCX. (K) POT in LM with short noncompliant balloon. (L) Result.

Close

Fig. 7 Double-kiss (DK) crush technique. (A) Baseline CAG revealing significant LM bifurcation lesion involving proximal LAD and LCX (Medina 1, 1, 1), (B) Significant calcification in LM, LAD, and LCX. (C) Stenting of LM-LCX. (D) Crushing of LCX stent with noncompliant balloon placed in LM-LAD. (E) First kissing balloon inflation after LCX recross. (F) Deployment of stent in LM-LAD. (G) POT in LM with short noncompliant balloon. (H) Second kissing balloon inflation after high-pressure sequential dilatation of LAD and LCX. (I) Stent boost. (J) Re-POT. (K) Final result.

Export to PPT

Fig. 7 Double-kiss (DK) crush technique. (A) Baseline CAG revealing significant LM bifurcation lesion involving proximal LAD and LCX (Medina 1, 1, 1), (B) Significant calcification in LM, LAD, and LCX. (C) Stenting of LM-LCX. (D) Crushing of LCX stent with noncompliant balloon placed in LM-LAD. (E) First kissing balloon inflation after LCX recross. (F) Deployment of stent in LM-LAD. (G) POT in LM with short noncompliant balloon. (H) Second kissing balloon inflation after high-pressure sequential dilatation of LAD and LCX. (I) Stent boost. (J) Re-POT. (K) Final result.