A critical stenosis in left main coronary artery (LMCA) is observed in 3–5% of all patients who undergo coronary angiography and in 10–30% of patients who undergo coronary artery bypass grafting (CABG).1 Significant LMCA disease is a high-risk condition because of the extent of jeopardised myocardium.
Current practice guidelines recommend CABG as the standard therapy for patients with unprotected LMCA (ULMCA) disease,2–4 primarily because the long-term outcomes of surgical revascularisation are superior to those of medical treatment.5,6
Historical studies evaluating percutaneous coronary intervention (PCI) for ULMCA stenosis with balloon angioplasty or bare-metal stents (BMS) have reported poor mid-term results and high rates of restenosis-related events such as repeated revascularisation or even cardiac death. The introduction of drug-eluting stents (DES), which led to a significant reduction in restenosis and target lesion revascularisation (TLR) rates compared with BMS, has improved the immediate and mid-term outcome of PCI for ULMCA.7–32
Current guidelines recommend the indication with an uncertain benefit (Class IIb) for PCI only in cases of very high surgical risk.2 However, there are still only limited data regarding the long-term safety and efficacy of PCI for ULMCA disease and no adequately powered randomised trials have compared PCI versus CABG in this subset of patients.
The purpose of this article is to examine recent advances and the current status of PCI with DES in ULMCA disease treatment, as well as to determine concerns that should be resolved before the eventual redefinition of standards for ULMCA revascularisation.
Different Anatomical Complexities of the Left Main Coronary Artery Lesion – Are the Outcomes Equal?
The LMCA can be divided into three anatomical regions: ostium, mid-shaft and distal (which involves the bifurcation between the left anterior descending [LAD] and left circumflex [LCx] arteries). Histologically, the ostial portion resembles the aorta (rich in smooth-muscle cells and elastic fibres), while the distal bifurcation is the part of the LMCA most susceptible to developing an atherosclerotic lesion because of low-shear flow disturbance.33 Clinical outcomes may vary according to ULMCA complexity and lesion location. In approximately 40% of patients undergoing PCI for ULMCA disease, the atherosclerotic plaque is confined to the ostium or mid-shaft segment.
PCI of non-distal lesions is associated with favourable clinical and angiographic outcomes. In a multicentre study of 147 patients with non-distal ULMCA lesions treated with DES implantation, the success rate was 99% and the restenosis rate was 0.9%. At 2.5 years of follow-up, cardiac death and TLR rates were 2.7 and 0.7%, respectively.18 Conversely, the ULMCA lesion is located distally in 60–90% of all patients; PCI on such lesions is technically more challenging because it can be complicated by plaque shift in 4.5–26% of patients.34 On this basis it is not surprising that clinical outcomes following distal LMCA treatment have been less encouraging than the results of non-distal ULMCA PCI.16,35
A meta-analysis of 17 trials involving PCI for ULMCA identified a distal lesion as the most significant predictor of repeated revascularisation and overall major adverse cardiac event (MACE).22 However, different outcomes have been described following distal-bifurcation LMCA lesions depending on lesion complexity (simple versus complex distal lesions with extensive involvement of both branches) and consequent stenting strategies (one- versus two-stent approach). Current evidence suggests that the results in cases of simple bifurcation lesions treated with a one-stent approach are more favourable compared with complex bifurcation lesions treated with a two-stent approach.12,36 The TLR rate is relatively low (<5%) with a one-stent approach, reaching nearly equivalent results to those obtained with DES for ostial or mid-ULMCA lesions.8,10
Conversely, because of the extensive plaque burden, patients with distal ULMCA disease approached with two-stent techniques show a TLR rate as high as 25%, with restenosis confined mainly to the LCx ostium. Although preliminary results for bifurcation ULMCA PCI showed better results with a one-stent approach compared with a two-stent approach, the latter was historically adopted in approximately 40% of cases.37,38 There is little consensus and few data on the optimal two-stent strategy (crush, culotte, V- or T-stenting) to approach a distal ULMCA lesion, since any existing comparison is challenged by operator and institutional preferences. Because restenosis or stent thrombosis (ST) could have tremendous consequences following PCI for ULMCA, all possible measures to achieve the optimal final result should be considered. In addition, because of the measurable risk of restenosis and revascularisation after complex stenting, the use of ULMCA-dedicated stents is currently under investigation.39
How to Stratify the Risk of Procedural and Long-term Outcomes Following Unprotected Left Main Coronary Artery Stenting
Most of the clinical risk-scoring systems currently used for ULMCA revascularisation have been extrapolated from patients treated with CABG. Recently, the application of a coronary anatomical risk score based on lesion severity and extent (SYNTAX score)40 has provided insight into both patient selection and safety and efficacy outcomes. In the left main subgroup of the Synergy Between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery (SYNTAX) trial, isolated ULMCA stenosis was identified in 13% of patients and the increasing number of additional vessels treated was identified as the single independent procedural determinant of one-year major adverse cardiac and cerebrovascular events (MACCE).41,42 For patients with ULMCA disease identified in the upper tertile of lesion complexity, represented by a SYNTAX score ≥33, PCI was associated with a higher MACCE rate up to three years of follow-up (PCI 37.1% versus CABG 21.2%; p=0.003).43 Conversely, no significant differences regarding three-year MACCE rate were observed between treatment strategies in patients with a low SYNTAX score (0–22: PCI 18% versus CABG 23%; p=0.33) and those with an intermediate SYNTAX score (23–32: PCI 23.4% versus CABG 23.4%; p=0.90).43
A recent study demonstrated that combining the SYNTAX and the EuroSCORE into a common risk model (Global Risk Classification) was correlated with a significant improvement in predicting cardiac mortality in patients undergoing PCI for ULMCA.44 Another novel score, the New Risk Stratification Score (NERS), which consists of 54 variables (17 clinical, four procedural and 33 angiographic) was recently described and compared with the SYNTAX score in patients undergoing PCI for ULMCA.45 At an average clinical follow-up of 1,764 days, NERS ≥25 demonstrated significantly higher sensitivity and specificity (MACE as stated variable) than an intermediate- or high-risk SYNTAX score. Moreover, NERS ≥25 (hazard ratio [HR] 1.13, 95% confidence interval [CI] 1.11–1.6; p<0.01) was the only independent predictor of cumulative MACE and ST at follow-up (odds ratio [OR] 31.04, 95% CI 19.36–67.07; p<0.001).45 The incorporation of clinical risk factors into the future estimation scoring system may refine its prognostic ability and may reduce the gap between the art and science of decision-making when choosing the most appropriate coronary revascularisation strategy for patients with severe coronary artery disease (CAD).
Current Evidence for Percutaneous Coronary Intervention in Unprotected Left Main Coronary Artery Disease
The efficacy of DES in LMCA PCI has been evaluated in observational studies with limited clinical follow-up.7–10,12,15,24,26,30,32,35,37,46–54 More recent clinical trials and comparative studies evaluating PCI with DES versus CABG for ULMCA disease have shown important results regarding efficacy and safety outcomes with DES treatment (see Table 1).11,13,14,19,21,25,55–60
The Revascularization for unprotected left main coronary artery stenosis: comparison of percutaneous coronary angioplasty versus surgical revascularization (MAIN-COMPARE) registry was the first multicentre non-randomised study comparing long-term outcome following PCI with stenting versus CABG in ULMCA disease.25 This registry involved 2,240 patients with ULMCA stenosis who underwent stenting (DES n=784; BMS n=318) or CABG (n=1138). Patients in the PCI cohort were less likely to have diabetes or multivessel coronary artery disease; however, a propensity scoring model found that treatment with DES was associated with a significantly lower rate of freedom from repeat revascularisation versus CABG at three-year follow-up (90.7 versus 98.4%; p<0.001). Of note, no significant differences in safety end-points including freedom from death (DES 91% versus CABG 93.1%; p=0.26) or the composite end-point of death, myocardial infarction (MI) or stroke were identified.25
Similar results were found even when follow-up was extended to five years.61 The five-year risk of death (HR 1.13, 95% CI 0.88–1.44; p=0.35) and the combined risk of death, Q-wave MI or stroke (HR 1.07, 95% CI 0.84–1.37; p=0.59) were not significantly different between the PCI and the CABG group, whereas the target vessel revascularisation (TVR) rate was significantly higher in the PCI group (HR 5.11, 95% CI 3.52–7.42; p<0.001).61
Consistent with this finding, a non-randomised comparison from our group showed no difference between PCI with DES implantation (107 patients) and CABG (142 patients) in the occurrence of cardiac death (adjusted OR 0.502, 95% CI 0.162–1.461; p=0.24), although the PCI group showed a trend towards a lower occurrence of the composite end-point of cardiac death and MI (adjusted OR 0.408, 95% CI 0.146–1.061; p=0.06). However, CABG was correlated with a lower TVR rate (adjusted OR 4.411, 95% CI 1.825–11.371; p=0.0004), and no difference was detected in the occurrence of MACCE (adjusted OR 1.578, 95% CI 0.825–3.054; p=0.18) at five-year clinical follow-up.62
The randomised SYNTAX trial compared PCI with paclitaxel-eluting stent (PES) implantation versus CABG for left main/multivessel CAD.41 Recently, the three-year results of the ULMCA subgroup from the SYNTAX trial (PCI arm 348 patients versus CABG arm 357 patients) were presented by Serruys at Transcatheter Cardiovascular Therapeutics 2010. PCI with PES implantation (348 patients) resulted in equivalent three-year overall MACCE to CABG (22.3% CABG versus 26.8% PCI; p=0.20). Notably, the MACCE rate was similar between the groups for patients with low (23% CABG versus 18% PCI; p=0.33) and intermediate (23.4% CABG versus 23.4% PCI; p=0.90) SYNTAX score, but was significantly higher in the PCI arm in the high SYNTAX score group (21.2% CABG versus 37.3% PCI; p=0.003). Even the overall safety outcomes (death/cerebrovascular event [CVE]/MI) were similar between the two groups (14.3% CABG versus 13% PCI; p=NS).
As reported at one-year follow-up,42 there was a higher revascularisation rate in the PCI group (11.7% CABG versus 20% PCI, p=0.004) and a higher rate of CVE in the CABG group (4% CABG versus 1.2% PCI, p=0.2) even at three-year follow-up (see Table 2) Moreover, favourable results of PCI for ULMCA disease treatment were also reported in a gender subanalysis.43 Although the encouraging results for PCI, especially in cases with anatomically simple LMCA disease (isolated LMCA ostial or mid-shaft disease or non-distal LMCA plus single-vessel disease), because of the hypothesis-generating nature of subgroup analysis, results from adequately powered trials for patients with ULMCA disease are needed. Table 3 shows the variables influencing revascularisation strategy in clinical practice.
Acute Unprotected Left Main Coronary Artery Occlusion in the Setting of ST-elevation Myocardial Infarction – Is Percutaneous Coronary Intervention the Preferred Revascularisation Approach?
Acute occlusion involving the ULMCA accounts for 0.8% of patients who undergo primary PCI.63 This event is often associated with catastrophic events such as cardiogenic shock, lethal arrhythmias and sudden death. Uncertainty surrounds the optimal revascularisation strategy for ST-elevation MI (STEMI) patients due to acute ULMCA occlusion. The revised 2004 American College of Cardiology (ACC)/American Heart Association (AHA) STEMI guidelines64 indicate that PCI and CABG have a class Ia indication in patients with cardiogenic shock, but do not provide specific treatment recommendations for ULMCA disease.
Data from individual trials with small sample sizes show a high mortality rate for patients with STEMI and ULMCA acute occlusion regardless of the revascularisation strategy chosen. The in-hospital mortality rate of STEMI patients with ULMCA occlusion treated with primary PCI ranges between 35 and 44%,65–67 while a rate of 46% was reported after emergency CABG in the same clinical setting.68 Although data on long-term follow-up are limited, patients who survive to discharge following ULMCA PCI have a favourable prognosis in terms of subsequent death and MI.65–67 However, considering the high clinical risk profile of this subset of patients, a treatment bias favouring PCI over CABG prohibits direct comparison between the two revascularisation modalities. Moreover, it is unlikely that a randomised controlled trial with sufficient size for this indication will be conducted given the logistical complexity of such a study and the treatment biases that favour one therapy over another. Nevertheless, there is an opportunity through studies to further refine our understanding of ULMCA PCI in STEMI.
PCI may be performed more expeditiously than CABG to promptly reperfuse the infarcted artery, potentially reversing arrhythmic and haemodynamic instability. Delays to reperfusion with CABG, which may take one hour or longer during off-peak hours to establish cardiopulmonary bypass, can be catastrophic in this situation. Hence, emergency PCI performed in a timely fashion by experienced operators should be considered as a preferred alternative to CABG in the following situations:64 ULMCA occlusion with a Thrombolysis In Myocardial Infarction (TIMI) flow grade less than 3; cardiogenic shock and/or life-threatening arrhythmias; or co-morbidities that pose an excessive risk of CABG-related complications such as chronic obstructive pulmonary disease, cerebrovascular disease.69
Is Intravascular Ultrasound Guidance Needed for Unprotected Left Main Coronary Artery Percutaneous Coronary Intervention?
Evaluation of ULMCA disease may be difficult because of anatomical and haemodynamic factors such as large size, overlapping major vessels, aortic cusp opacification and varied angulations. There is a need for proper evaluation of both the distribution and disease severity in cases of ULMCA lesions. Therefore, haemodynamic and intravascular ultrasound (IVUS) assessment of ULMCA disease has associated the functional and/or anatomical relevance of stenosis with the need for treatment and clinical outcome.70,71
Although a small retrospective study reported that IVUS-guided ULMCA PCI with DES did not have a significant clinical long-term benefit compared with angiography guided PCI,46 the usefulness of IVUS-guided stenting may not be hampered by this underpowered retrospective study. Angiography has a clear limitation in assessing the true luminal size of ULMCA, because the left main artery is often short and lacks a normal segment for comparison. In addition to the assessment of an ULMCA lesion before the procedure, IVUS guidance is helpful to obtain an adequate expansion of DES, prevent stent malapposition and achieve full lesion coverage with DES. A subgroup analysis from the MAIN-COMPARE registry reported that IVUS guidance was associated with improved three-year mortality rates compared with conventional angiography-guided procedures after adjustment with propensity-score matching (6.3% IVUS versus 13.6% angiography, log rank p=0.063, HR 0.54, 95% CI 0.28–1.03).27 In particular, for patients receiving DES, IVUS-guided PCI was associated with a significantly lower three-year incidence of mortality compared with angiographically guided PCI (4.7% IVUS versus 16% angiography, log rank p=0.048, HR 0.39, 95% CI 0.15–1.02).27 Despite the non-randomised design of the study, these results indicate that IVUS-guided PCI for ULMCA disease may play a significant role in reducing ST and long-term mortality. IVUS assessment of stent underexpansion, incomplete lesion coverage, suboptimal stent area and incomplete stent apposition have been found to predict early and late/very late ST after DES implantation.72–76 Therefore, IVUS-guided PCI with DES for ULMCA stenosis should be strongly recommended. Finally, optimal coherence tomography (OCT) to assess vascular response to LMCA stenting has recently been reported.77
Are Clinical Safety and Efficacy Equivalent for the Various Drug-eluting Stents in Unprotected Left Main Coronary Artery Percutaneous Coronary Intervention?
To date, data are mostly limited to sirolimus-eluting stents (SES) and PES. Several small observational studies have compared outcomes following SES or PES implantation for ULMCA stenting, showing similar results at mid-term follow-up.17,47
In the Intracoronary stenting and angiographic results: drug-eluting stents for unprotected left main lesions (ISAR-LEFT MAIN) randomised trial,37 607 patients were assigned to ULMCA PCI with SES or PES to compare the one-year composite outcome of death, MI and TLR. At 12- month follow-up, no significant differences were reported in the one-year primary end-point (PES 13.6% versus SES 15.8%, relative risk 0.85, 95% CI 0.56–1.29), angiographic restenosis (six- to nine-month follow-up: PES 16% versus SES 19.4%; p=0.30) and two-year left main specific revascularisation rates (9.2% PES versus 10.7% SES; p=0.47). The incidence of definite (0.7% PES versus 0.3% SES) and probable (0.3% PES versus 0% SES) ST was also similar in the two groups at two-year follow-up.37
Although promising results have been demonstrated with new-generation DES in selected and unselected subsets of patients and lesions, only limited data are as yet available regarding their safety and efficacy in ULMCA PCI. In the Left main Taxus and left main Xience (LEMAX) non-randomised registry, 173 patients with ULMCA disease treated with an everolimus-eluting stent (EES) (Xience V; Abbott Vascular, Santa Clara, CA, US) were compared with a historical cohort of 291 patients treated with PES (Taxus, Boston Scientific, Natick, MA, US) for ULMCA stenosis. At 12-month clinical follow-up, the EES was associated with a lower rate of target lesion failure (a composite of cardiac death, target vessel MI and TLR), MACE and overall ARC ST compared with PES.78 The ongoing ISAR-LEFT MAIN 2 randomised trial, which is evaluating the safety and efficacy of the EES versus a zotarolimus-eluting stent (Endeavor Resolute, Medtronic Vascular, Santa Rosa, CA, US) in ULMCA PCI will provide more information about the performance of new-generation DES platforms in this complex subset of lesions.
Are Percutaneous Intervention and Drug-eluting Stent Implantation for Unprotected Left Main Coronary Artery Stenosis Safe in Terms of Long-term Clinical Outcomes?
DES implantation in ULMCA revascularisation represents a narrow margin between the need for a strong antirestenotic effect balanced by the risk of ST related to delayed healing. Since ST at any time is associated with MI and mortality rates of approximately 30%,79 its occurrence following ULMCA intervention could be associated with catastrophic consequences. Thus, any measure to reduce the occurrence of ST would have a potent clinical impact and should be considered when moving forward with a ULMCA PCI strategy recommendation.
However, recent data attenuate concerns about the safety of PCI with DES in the treatment of ULMCA stenosis.23,25,32,80 Data from a multicentre international registry of 731 patients undergoing elective ULMCA stenting with DES have reported four episodes of definite or probable ST, with a cumulative incidence of 0.95% at 30 months of follow-up. Of these four episodes, two were acute, one subacute and one late ST.23 A report from the Drug eluting stent for left main (DELFT) registry, which included 358 patients undergoing ULMCA stenting with DES, demonstrated that the incidence of definite, probable and possible ST was 0.6, 1.1 and 4.4% respectively, at a minimum of three-year follow-up.32 In a large real-world registry (ASAN-MAIN), of 176 patients treated with DES implantation for ULMCA disease, the cumulative incidence of definite or probable ST was 1.8% at five-year follow-up.80 Among the definite or probable ST events, one was acute, one subacute and one very late (4.3 years after the procedure). The acute and subacute ST occurred on dual antiplatelet therapy (DAT), while the very late case arose two months after DAT discontinuation.80 In the ULMCA subset of the SYNTAX trial, the three-year rate of any ST following PES implantation was 4.1% compared with a 3.7% rate of clinically manifest bypass graft occlusion (p=0.80).43 Notably, in the PCI cohort of the SYNTAX trial, ST outcomes were reported as any event and were not specified to the ULMCA target lesion. All these results indicate that DES implantation in patients with ULMCA disease results in relatively lower rates of ST than when DES are implanted in other patients and lesion subsets (see Table 4).81
For How Long Should Dual Antiplatelet Therapy Be Continued After Unprotected Left Main Coronary Artery Stenting with a Drug-eluting Stent?
The optimal duration of DAT in patients with ULMCA disease treated with DES remains to be defined. The long-term benefits of DAT use beyond six or 12 months are unclear in non-ULMCA cohorts.82–84 Current ACC/AHA/Society for Coronary Angiography and Interventions and European Society of Cardiology/European Association for Cardio-Thoracic Surgery guidelines offer limited assistance about this practical issue.2,4
Current guidelines support long-term aspirin administration and at least six to 12 months of therapy in patients receiving a DES (Class I, level of evidence B); however, this is not specific for ULMCA stenting. Moreover, the support for long-term (over one year) DAT after DES implantation is even less robust (Class IIb, level of evidence C). Although the risk–benefit ratio of long-term DAT is not well defined, many clinicians prolong DAT for up to several years or indefinitely after ULMCA stenting with DES. The utility of functional assays to provide the most appropriate antiplatelet therapy is also not clearly defined. Migliorini et al.85 reported the outcomes of 215 patients treated with DES for ULMCA who had prospective platelet reactivity assessment by light transmittance aggregometry after a loading dose of 600mg of clopidogrel. The incidence of high residual platelet reactivity (HRPR) after clopidogrel loading was 18.6%. The three-year cardiac mortality rate was 8.0±3.1% in the low residual platelet reactivity (LRPR) group and 28.3±10.4% in the HRPR group (p=0.005). The three-year ST rate was 4.2±1.8% in the LRPR group and 16.0±7.3% in the HRPR group (p=0.021). By forward stepwise regression analysis, HRPR after clopidogrel loading was the only independent predictor of cardiac death (HR 3.82, 95% CI 1.38–10.54; p=0.010) and ST (HR 3.69, 95% CI 1.12–12.09; p=0.031).85 Current American guidelines suggest performing platelet aggregation tests in patients undergoing DES implantation for ULMCA disease and empirically advocate increasing the clopidogrel dose to 150mg daily if <50% platelet aggregation is demonstrated, whereas European guidelines suggest a 150mg daily maintenance dose of clopidogrel (without mentioning platelet response to the drug) following PCI with DES implantation for complex coronary lesions.2,4 Additional studies are strongly required to resolve these issues and to determine the optimal duration of DAT administration after DES placement in ULMCA disease.
Based on This Evidence, Should We Redefine The Therapeutic Standards?
CABG is still the standard of care for significant left main disease in patients eligible for surgery, and the Coronary artery surgery study (CASS) registry reported a median survival advantage of seven years in 912 patients treated with CABG rather than medically.86 However, non-randomised comparative studies of PCI versus CABG in ULMCA disease are consistent in reporting similar hard clinical end-point event rates following risk adjustment, up to five-year (DES) and 10-year (BMS) clinical follow-up.61,62
Moreover, data obtained from pre-specified subgroup analysis from randomised clinical trials21,41 comparing PCI versus CABG in ULMCA disease are encouraging and in accordance with previously reported experiences involving PCI with DES in ULMCA cohorts.
Support for the potential indication of PCI at least for low-risk ULMCA disease comes from a meta-analysis of 3,773 patients with ULMCA stenosis. The authors reported no differences in mortality or MACCE rates for up to three years between PCI and CABG. However, in the PCI group there was a significantly higher revascularisation rate.87
This emerging evidence suggests that PCI provides at least equivalent results to CABG for lower-severity ULMCA lesions up to three years of follow-up.43 However, it is vital to confirm that these results remain with longer-term follow-up. The currently available evidence was substantive enough for reclassification of European guideline recommendations for ULMCA revascularisation with PCI to class IIa B status in non-bifurcated lesions that are isolated or associated with single-vessel disease, although the indication for more complex ULMCA lesions (bifurcated lesions or ostial-shaft ULMCA lesions plus two- or three-vessel disease) remains class IIb B4 (see Table 5). There remains a large number of patients with intermediate anatomical and clinical complexity for whom there is still uncertainty about the optimum treatment. It is hoped that we will receive more information on these patients from the upcoming Evaluation of Xience V Prime versus coronary artery bypass surgery for effectiveness of left main revascularization (EXCEL) trial, which will help to move forward the process of redefining treatment standards.