Percutaneous Coronary Intervention for Unprotected Left Main Disease

Introduction

Left main (LM) coronary artery disease, defined by a visually estimated diameter stenosis of at least 50%, is present in approximately 3% to 5% of patients undergoing coronary angiography. Without revascularization, mortality is as high as 50% within 3 years of follow-up. Current treatment guidelines recommend coronary artery bypass grafting (CABG) for most patients with unprotected LM disease ; however, significant advances in stent technology, revascularization techniques, and antithrombotic therapies have made percutaneous coronary intervention (PCI) a safe, feasible, and effective method of revascularization in many patients with LM disease. The present chapter will review the current evidence related to LM PCI and describe PCI techniques to manage diverse LM disease anatomy.

Randomized Controlled Trials

To date, there have been four randomized controlled trials (RCTs) comparing PCI with CABG for LM intervention ( Table 8-1 ). The Study of Unprotected Left Main Stenting versus Bypass Surgery (LE MANS) was the first RCT to compare the safety and efficacy of PCI with CABG. LE MANS enrolled 105 symptomatic patients with >50% diameter stenosis in the LM with or without concomitant multivessel coronary artery disease (CAD) to revascularization with either PCI or CABG. The primary endpoint was the change in left ventricular ejection fraction (LVEF) at 12 months as measured by two-dimensional echocardiography. The secondary endpoint was major adverse cardiac and cerebrovascular events (MACCE), the composite of all-cause death, myocardial infarction (MI), stroke, or target vessel revascularization (TVR). At 30 days, PCI was associated with a significantly lower rate of MACCE (2% vs. 13%; p = 0.03) and shorter duration of hospitalization (6.8 ± 3.7 vs.12.0 ± 9.6 days; p = 0.0007). At 1 year, LVEF improved to a significantly greater degree with PCI than with CABG (3.3 ± 6.7% vs. 0.5 ± 0.8%; p = 0.047), resulting in greater LVEF in the PCI group (58.0 ± 6.8% vs. 54.1 ± 8.9%; p = 0.01). Angina status and treadmill exercise performance were comparable in the two groups. There was no difference in all-cause mortality, and 1-year MACCE rates were also similar in both groups (relative risk [RR], 1.09; 95% confidence interval [CI], 0.85 to 1.38). At 28 ± 9.9 months of follow-up, MACCE-free survival was comparable in both groups, but there was a trend of fewer deaths in the PCI group (p = 0.08). The major limitations of this trial were its small size, a relatively low rate (72%) of internal mammary artery graft use in the surgical cohort, and a high rate of bare-metal stent use (65%), with mainly first-generation drug-eluting stents (DES) in the remainder.

The Synergy Between PCI with Taxus and Cardiac Surgery (SYNTAX) trial is the largest RCT to date to evaluate optimal revascularization strategy (PCI vs. CABG) in patients with LM disease. SYNTAX randomized 1800 patients with three-vessel or LM disease to PCI with first-generation TAXUS paclitaxel-eluting stents (PES) or CABG, stratified by the presence of LM disease and medically treated diabetes. The primary endpoint was 1-year rate of MACCE (death from any cause, stroke, MI, or repeat revascularization), and patients were followed up to 5 years. The study used a hierarchical statistical approach that prespecified statistical testing on the LM subgroup only after the demonstration of noninferiority between PCI and CABG in the overall population for the primary endpoint of 1-year MACCE. Since PCI was not shown to be noninferior to CABG in the trial (1-year MACCE: 17.8% for PCI vs. 12.4% for CABG; p = 0.002), the findings of the LM subgroup are considered hypothesis-generating.

A total of 705 patients were enrolled with LM disease. The mean EuroSCORE of the LM cohort was 3.8, and the mean SYNTAX score was 30, consistent with complex and extensive CAD. The 1-year rate of MACCE in the LM subgroup was similar in the two groups (15.8% PCI vs. 13.7% for CABG; p = 0.44). The 1-year rate of TVR was significantly greater in the PCI group (11.8% vs. 6.5%; p = 0.02); however, the rate of stroke was significantly greater in the CABG group (2.7% vs. 0.3%; p = 0.009). There were no differences in the rates of all-cause death or MI between the two groups. Comparing PCI to CABG, the 1-year MACCE rate was greater in the CABG group in patients with isolated LM or LM + one-vessel CAD (7.5% vs. 13.2%, respectively) but was greater in the PCI group in patients with LM + two-vessel or LM + three-vessel CAD (19.8% vs. 14.4% and 19.3% vs. 15.4%, respectively). When stratified by SYNTAX score tertiles (0-22, 23-32, >32), the MACCE rates in the lower two tertiles were similar between PCI and CABG; however, the highest tertile had a significantly greater MACCE rate with PCI compared with CABG.

The final 5-year results of the SYNTAX trial have been recently published. The MACCE rate did not differ significantly between the two treatment groups (36.9% for PCI vs. 31.0% for CABG; p = 0.12). When stratified by SYNTAX tertiles (0-22, 22-32, >32), the MACCE rates were similar between PCI and CABG in the lower two tertiles; however, in the highest tertile, PCI had a significantly higher rate of MACCE compared with CABG ( Figure 8-1 ). While these data suggest that PCI with TAXUS stents may be a reasonable alternative to CABG in patients with low and intermediate SYNTAX scores, given the small sample sizes of these subsets, prospective trials are required for confirmation before such an approach is widely adopted.

The Premier of Randomized Comparison of Bypass Surgery versus Angioplasty Using Sirolimus-Eluting Stent in Patients with Left Main Coronary Artery Disease (PRECOMBAT) randomized 600 patients with LM disease to first-generation Cypher sirolimus-eluting stents (SES) versus CABG. The primary endpoint was MACCE (the composite of death from any cause, MI, stroke, or ischemia-driven TVR) at 1 year, powered for noninferiority. The mean EuroScore was 2.7, and the mean SYNTAX score was 25. At 1 year, PCI was shown to be noninferior to CABG (8.7% for PCI vs. 6.7% for CABG; p = 0.01 for noninferiority) for MACCE. Additionally, there were no differences at 1 year in any of the individual components of the composite endpoint. At 2 years the rate of MACCE was also not significantly different between the groups (12.2% for PCI vs. 8.1% for CABG; p = 0.12), although ischemia-driven TVR was significantly more common in the PCI group (9.0% vs. 4.2%; p = 0.02). The main limitations of this trial were the modest number of patients enrolled, the wide margin for noninferiority, and the low event rates in both groups, making these results hypothesis-generating rather than conclusive.

The fourth RCT comparing PCI with CABG in unprotected LM disease is a small study from Boudriot et al. This trial enrolled 201 patients with documented ischemia and LM diameter stenosis >50% with or without multivessel disease. The primary clinical endpoint was the composite rate of major adverse cardiac events (MACE) at 1 year, including cardiac death, MI, or TVR, powered for noninferiority. PCI failed to reach noninferiority compared with CABG for 1-year MACE (19% PCI vs. 13.9% CABG; p = 0.19 for noninferiority), driven by greater TVR after PCI. The overall results remained unchanged at extended 3-year follow-up. Major limitations of this study included a small sample size, the noninclusion of stroke as an endpoint, and use of first-generation DES.

Metaanalyses

Several metaanalyses of observational registries and RCTs have been conducted comparing PCI and CABG for LM revascularization. In a recent metaanalysis, Athappan and colleagues compared the long-term outcomes between PCI using first-generation DES and CABG in unprotected LM patients. This analysis included 21 observational studies and three RCTs, for a total of 14,203 patients. At 5 years, there were no significant differences between PCI and CABG in all-cause death (odds ratio [OR] = .0.79; 95% CI, 0.57 to 1.08), cardiac death (OR = 0.95; 95% CI, 0.36 to 2.50), or nonfatal MI (OR = 1.38; 95% CI, 0.71 to 2.70). PCI was associated with significantly lower 5-year rates of stroke (OR = 0.27; 95% CI, 0.13 to 0.55) and cumulative MACCE (OR = 0.64; 95% CI, 0.51 to 0.80) but a greater rate of TVR (OR = 3.77; 95% CI, 2.43 to 5.87) compared with CABG. When stratifying patients by complexity and extent of baseline coronary disease by SYNTAX score assessment (reported in only three studies), PCI and CABG had nonsignificantly different rates of all-cause mortality, MI, and MACCE in the lower two tertiles of the SYNTAX score (<32); however, with more complex disease (SYNTAX score >32), CABG was associated with significantly better outcomes at 3 years. The combination of randomized trials and registries (which dominated the numbers of patients) limits the interpretation of this metaanalysis.

Capodanno and colleagues performed a metaanalysis of the four RCTs of PCI with first-generation DES versus CABG for unprotected LM disease (total 1611 patients). The 1-year rate of MACCE was nonsignificantly different between PCI and CABG (14.5% vs. 11.8%; OR = 1.28; 95% CI, 0.95 to 1.72; p = 0.11), although TVR was greater after PCI (11.4% vs. 5.4%; OR = 2.25; 95% CI, 1.54 to 3.29; p = 0.001). Conversely, stroke occurred less frequently with PCI (0.1% vs. 1.7%; OR = 0.15; 95% CI, 0.03 to 0.67; p = 0.013). There were no significant differences in the rates of all-cause death (3.0% vs. 4.1%; OR = 0.74; 95% CI, 0.43 to 1.29; p = 0.29) or MI (2.8% vs. 2.9%; OR = 0.98; 95% CI, 0.54 to 1.78; p = 0.95) between the two groups. Long-term outcomes were not reported.

Ongoing Trials

On the basis of these data, two large-scale trials are ongoing which will importantly inform the relative risks and benefits of PCI with contemporary DES versus CABG. Both trials are based on the observations that in the trials completed to date, PCI with first-generation DES appeared comparable to CABG in patients with low or intermediate anatomic disease complexity, but not in those with concomitant extensive multivessel CAD. Both trials are also utilizing second-generation DES for which outcomes have been shown to be improved compared with first-generation devices used in the earlier studies. The Evaluation of XIENCE Prime or XIENCE V versus CABG for Effectiveness of Left Main Revascularization (EXCEL) trial (NCT01205776) is a prospective, randomized multicenter trial in which 1905 patients with significant LM CAD and a SYNTAX score of ≤32 have been enrolled at 148 international sites and randomized 1 : 1 to either PCI using XIENCE everolimus-eluting stents or CABG. The primary endpoint, powered for sequential noninferiority and superiority testing, is the composite rate of all-cause death, MI, or stroke at median 3-year follow-up. Follow-up for all randomized subjects is planned for at least 5 years with an option for additional follow-up to 10 years. A second RCT is enrolling at 126 European sites and is comparing PCI using the bioabsorbable polymer biolimus-eluting BIOMATRIX stent (n = 600) with CABG (n = 600) in approximately 1200 patients with LM CAD with ≤3 additional noncomplex lesions (excluding lesions with length >25 mm, chronic total occlusions, bifurcation lesions requiring two stents, and calcified or tortuous vessels) (NCT01496651). The primary endpoint is the composite endpoint of death, stroke, nonindex treatment related MI or repeat revascularization, with follow-up for 5 years.

Current Guidelines and Appropriate use Criteria

Pending the results of EXCEL and Nordic-Baltic-British Left Main Revascularization (NOBLE), the evidence from the randomized trials of first-generation DES versus CABG has been reflected in the current guidelines ( Tables 8-2 and 8-3 ). In the American College of Cardiology/American Heart Association/Society of Cardiovascular Angiography and Interventions guidelines (ACC/AHA/SCAI), LM PCI has been upgraded from a Class III indication in 2006 to a Class IIb indication in 2009 and to a Class IIa indication in 2011 for selected patients in whom the coronary anatomy is deemed suitable for PCI and surgical risk is high ( Table 8-2 ). To guide selection of the most appropriate revascularization modality for significant unprotected LM disease, these guidelines recommend the following:

• 1.

  A heart team approach to decision making involving at least one interventional cardiologist and one cardiothoracic surgeon (Class 1, level of evidence [LOE] C)

• 2.

  Utilization of angiographic risk stratification by the SYNTAX score and clinical risk stratification by the Society of Thoracic Surgeons (STS) score (Class IIa, level of evidence B)

PCI is considered a reasonable alternative to CABG in patients with stable ischemic heart disease having a SYNTAX score <22 and ostial/shaft LM disease with STS predicted operative mortality >5% (Class IIa, LOE B) and in patients with SYNTAX score <32 and LM bifurcation disease with STS predicted operative mortality >2% (Class IIb, LOE B). It is important to note that most of the RCTs comparing PCI and CABG enrolled patients who could undergo equivalent revascularization with either modality.

The Appropriate Use Criteria (AUC) recommendations have also evolved to incorporate the long-term evidence from randomized clinical trial data. Table 8-3 summarizes their recommendations.

The European Society of Cardiology and the European Association for Cardio-Thoracic Surgery (ESC/EACTS) 2014 guidelines for myocardial revascularization have recently upgraded the status of PCI for LM revascularization. LM with SYNTAX score ≤22 has been given a Class I recommendation (LOE B), while LM with SYNTAX score 22-32 was given a Class IIa recommendation (LOE B). LM with SYNTAX score >32 was given a Class III recommendation (LOE B). The ESC/EACTS also focused on the central importance of the multidisciplinary heart team discussion in all scenarios involving complex multivessel CAD and unprotected LM disease.

Scoring Algorithms

Scoring algorithms to assess clinical risk and anatomic disease complexity bring objectivity to risk stratification and, paired with clinical judgment, can guide the decision-making process to select the most appropriate revascularization strategy for a given patient. ESC/EACTS recommendations for various scores are shown in Table 8-4 . The clinical STS and EuroSCORE II are recommended to assess suitability for CABG (Class I, LOE B, and Class IIa, LOE B, respectively), whereas the anatomic SYNTAX score is useful to discriminate relative outcomes between PCI and CABG (Class I, LOE B).

The SYNTAX score, first developed and evaluated in the SYNTAX trial, was subsequently validated and studied in many unprotected LM studies. In most of these studies, rates of composite ischemic endpoints (death, MI, target lesion revascularization [TLR], or TVR) were significantly greater after PCI in the highest tertile of SYNTAX score than in the lower two tertiles. Capodanno et al. demonstrated that a SYNTAX score >34 was associated with significantly higher rates of ischemic events in patients with LM CAD undergoing PCI than in those undergoing CABG. Of note, only the baseline SYNTAX score had prognostic value; the lesion location in the LM (ostial, shaft, or bifurcation) and number of stents implanted were not predictive of clinical outcomes. Similarly, analysis from the ISAR-Left Main trial also showed a clear association between the overall burden and complexity of CAD, as assessed by SYNTAX score, and the rate of TLR and MACE at 3 years.

By combining clinical variables and the angiographic SYNTAX score, several other scores have been created and studied in an attempt to improve the predictive capability of the solely anatomic SYNTAX score. Of note, the Global Risk Classification, which is a combination of the SYNTAX score and EuroSCORE, showed the best calibration and discrimination in the prediction of adverse events such as mortality after LM PCI. Additionally, the Global Risk Classification identified a low-risk cohort of patients (low SYNTAX score) who could be safely treated with PCI.

The recently developed SYNTAX Score II combines the SYNTAX score with anatomical and clinical variables that were shown to alter the threshold value at which equipoise was achieved between CABG and PCI for long-term mortality. The variables included are age, creatinine clearance, LVEF, presence of unprotected LM disease (vs. three-vessel CAD), peripheral vascular disease, female sex, and chronic obstructive pulmonary disease. This score may be used to identify some patients considered low-risk by the anatomic SYNTAX score who might preferentially benefit by CABG, and some patients considered high-risk by the anatomic SYNTAX score who might preferentially benefit by PCI ( Figure 8-2 ). Further validation of the SYNTAX Score II (including assessment of endpoints other than mortality) is required prior to its widespread adoption.

Impact of Stent Choice

Given the fact that ≥70% of the myocardial mass is typically perfused by the LM, stent failure, such as restenosis and stent thrombosis, may have potentially catastrophic consequences. Despite its large diameter, disease involving the LM frequently involves the distal bifurcation and side branches (SBs). Table 8-5 presents data from observational registries and trials comparing different stent types for LM PCI. A metaanalysis of observational studies and RCTs involving 10,342 patients demonstrated lower event rates for DES than BMS for mortality, repeat revascularization, and MACE at 6 to 12 months, 2 years, and 3 years. Adjusted analyses among 5081 patients demonstrated a significantly lower risk of mortality with DES at 2-year (OR = 0.42; 95% CI, 0.28 to 0.62; p < 0.001) and 3-year (OR = 0.70; 95% CI, 0.53 to 0.92; p < 0.01) follow-up. Although these analyses should be seen as hypothesis-generating, given their retrospective nature, these data support a default strategy of DES for LM PCI in most cases unless long-term dual antiplatelet therapy is contraindicated.

To date, only two RCTs have compared different DES in LM PCI. In ISAR-Left Main, no significant difference in the 1-year composite endpoint of death, MI, or TLR was present between Cypher SES and Taxus PES (both first-generation DES). Similarly, no difference in the same endpoint was observed in ISAR-Left Main II between slow-release resolute zotarolimus-eluting stents (ZES) and everolimus-eluting stents (EES) (both second-generation DES). The modest number of patients included in these studies (~600), however, precludes definitive conclusions. Recently, a retrospective study demonstrated better mid-term (2-year) outcomes when comparing a first-generation DES (PES) to a second-generation DES (EES) in terms of target vessel failure (a composite of cardiac death, target vessel MI, or TLR). Once again, given the small number of patients included in this study (n = 344) and the retrospective nature of the analysis, these results are hypothesis-generating. That being said, in light of the important burden of evidence supporting the improved safety and efficacy of second-generation DES compared with both first-generation DES and bare-metal stent (BMS), second-generation DES should be the default platform when performing unprotected LM PCI, assuming the absence of contraindication to their use.

In-Stent Restenosis

No dedicated RCTs examining treatment options of LM in-stent restenosis (ISR) have been conducted so far, and most of the current evidence has originated from registry data and observational series. In the CORPAL (Córdoba and Las Palmas) registry, 7% of patients who underwent PCI with DES in unprotected LM disease developed ISR at a median follow-up of 9 months. The location of restenosis was divided equally among the main vessel (MV) (LM/left anterior descending artery [LAD]) or isolated to the ostium of circumflex or both arteries. Angiographically, ISR lesions were divided equally between focal (47%) and diffuse (51%). Intravascular ultrasound (IVUS) was conducted in 79% of ISR patients and demonstrated that stent under expansion was present in 14% of the cases. All patients except four were treated with repeat PCI with DES—58% with provisional stenting and 42% with a two-stent approach. During a follow-up period of 4 years, the overall recurrent MACE rate was 22%, with the provisional approach having a significantly higher MACE-free survival compared with the two-stent approach (85% vs. 53%; p < 0.05). Similarly, patients with ISR involving only one bifurcation segment had better MACE-free survival compared with patients with more than one segment involved (84% vs. 47%; p < 0.05) at 4 years.

In the MITO (Milan and New-Tokyo) registry, 92 out of 474 (19%) unprotected LM patients undergoing PCI with DES developed ISR, and 84 (19%) were treated with repeat PCI (43 with balloon angioplasty alone and 41 with further DES implantation). Of note, patients with focal left circumflex (LCX) stenosis were frequently asymptomatic and were discovered only on angiographic follow-up. During a follow-up period of 2 years, the patients undergoing POBA had significantly greater MACE rates compared with repeat PCI with DES (higher risk [HR], 2.75; 95% CI, 1.26 to 5.98; p = 0.01).

One of the important independent predictors of LM ISR is the final minimum stent area as determined by IVUS. Post-PCI optimal values for the MV, both branches, and the polygon of confluence (POC) have been established and were shown to be associated with significantly improved outcomes ( Figure 8-3 ).

Disease Location: Ostial/mid-Shaft versus Distal LM

PCI of the distal LM bifurcation has been associated with higher follow-up MACE rates than PCI limited to the LM ostium or mid-shaft in most, but not all, studies. The Drug-Eluting Stent for Left Main Coronary Artery Disease (DELTA) registry is the largest study to examine this issue. A total of 1612 LM PCI patients were included in the study, including 1130 patients with distal bifurcation lesions. During a median follow-up of 3.4 years, patients with distal bifurcation disease had a greater rate of MACE (a composite endpoint of death, MI, or TVR) compared with ostial/mid-shaft lesions (propensity-score adjusted HR, 1.48; 95% CI, 1.16 to 1.89; p = 0.001), primarily driven by a higher rate of TVR (propensity-score adjusted HR, 1.68; 95% CI, 1.19 to 2.38; p = 0.003). Similar findings have been reported by other groups.

In another registry analysis of 1111 patients with 2 years of follow-up, Palmerini et al. showed that distal bifurcation lesions had a lower freedom from MACE compared with ostial/mid-shaft lesions (72% vs. 80%; p = 0.03). However, whereas patients with distal bifurcation lesions treated with a two-stent strategy had worse outcomes, patients with distal bifurcation lesions treated with a one-stent (provisional) approach had clinical outcomes similar to those with ostial and mid-shaft LM lesions. Given the difference in outcomes after PCI of distal bifurcation LM and nonbifurcation LM lesions, the ACC/AHA guidelines recommend PCI as a Class IIb recommendation for distal bifurcation lesions and Class IIa for ostial/mid-shaft lesions.

One-Stent versus Two-Stent Approach for LM Bifurcation Disease

Bifurcation lesions represent >50% of all LM PCIs. True distal bifurcation lesions may be treated by either a single-stent or a two-stent strategy. Choice of strategy is based on vessel and lesion characteristics (plaque distribution, diameter of the branches and the angle between them, anatomy of the SB) in addition to operator experience and expertise.

Few dedicated RCTs have compared stenting strategies in LM bifurcation lesions. Most bifurcation studies have been performed in populations that include both non-LM and LM bifurcation, with the latter usually representing only 2% to 10% of all bifurcation lesions. Nevertheless, outcomes from these studies may be extrapolated to LM patients, although the results of adverse outcomes can be more severe in the LM cohort. Nearly all RCTs in non-LM bifurcations have demonstrated that a routine two-stent approach (stenting both the MV and SB) offers no angiographic or clinical benefits compared with a provisional stenting strategy (stenting the MV only, with stenting of the SB reserved for suboptimal balloon angioplasty results). One trial did report lower rates of angiographic and clinical restenosis with a routine two-stent approach, whereas another trial demonstrated an increase in adverse ischemic events, mainly periprocedural MI, with the two-stent approach. All of these studies were, however, limited by modest sample size, heterogeneity in SB lesion severity and length, and diversity in the two-stent technique.

From a registry report of 782 LM bifurcation patients undergoing PCI with DES, provisional stenting resulted in significantly higher MACE-free survival at 2-year follow-up compared with a routine two-stent approach (75% vs. 67%; p = 0.02). The lower MACE in the provisional group was primarily driven by greater TLR-free survival (87% vs. 73%; p = 0.00001). Thus, most experts agree that the provisional approach should be the default LM bifurcation treatment strategy for most cases. Exceptions include severe disease in both the LAD and LCX, and/or a sharply angulated SB entry.