Left main coronary disease is a common problem faced by interventional cardiologists. Percutaneous left main intervention has gained popularity mainly as a result of recently published encouraging results of trials examining the use of drug-eluting stents. However, important questions remain unanswered in the interpretation, implications and treatment of left main stenosis. Nevertheless, use of experience gathered over the last four decades may facilitate the optimisation of clinical decision-making.
In this article, important aspects that may help to determine the best management in patients with a left main stenosis are reviewed, with a focus on left main assessment and percutaneous treatment.
Scope of the Problem
A left main stenosis greater than 50% can be found in 4–10% of coronary angiograms,1–6 most frequently in combination with significant disease in at least one other coronary artery.1 Up to 9% of patients referred for bypass surgery and 7% of patients with an acute myocardial infarction have a significant left main lesion.7 The occurrence of an isolated left main stenosis is rare: it is encountered in fewer than 1% of patients undergoing angiography.8–13 Isolated left main lesions are more common at the ostium8 and in females.9 When seen in the presence of coronary disease in other epicardial vessels, it is more commonly located at the shaft or left main bifurcation.
The presence of left main disease has been shown to be associated with a poor outcome.1,14,15 Survival data collected by Conley et al. in the 1970s, when medical management for coronary disease was virtually non-existent, demonstrated that in the absence of surgical revascularisation, three-year survival in patients with a 50–70% and >70% left main stenosis was 66% and 41%, respectively.1
Left main stenosis is rarely an isolated finding. Most frequently it occurs in the presence of multivessel disease. For example, in the aforementioned study, concomitant three-vessel coronary disease was documented in 82% of patients. It is therefore impossible to separate the impact of the left main lesion itself from the consequences of atherosclerosis in the remaining coronaries. Nevertheless, the following information may be useful. First, data from a small observational study suggest a prognostic impact of angiographically silent left main atherosclerosis detected by intravascular ultrasound performed during percutaneous intervention to a coronary other than the left main artery.16 Second, in patients with three-vessel disease and concomitant left main stenosis >70%, survival was inferior compared with those with three-vessel disease and a normal left main artery.1 Moreover, taking clinical characteristics (left ventricular systolic function, baseline electrocardiogram, history of previous congestive heart failure, presence or absence of chest pain, left ventricular end-diastolic pressure, arteriovenous oxygen difference and size of the heart on chest X-ray) into account, patients with left main stenosis >70% and three-vessel coronary disease can be risk-stratified into a low- or high-risk category.1 While in the high-risk subgroup defined by the presence of two or more of the above-described risk factors the three-year survival was dismal (25%), it was 74% in the low-risk group.1
Importantly, the reason for poorer survival in patients with left main stenosis remains unclear. Is plaque rupture with subsequent left main thrombosis or occlusion the mechanism of death, or is it the haemodynamic impact causing profound myocardial supply–demand ischaemia followed by a lethal arrhythmia? Currently available data regarding coronary disease in general would suggest that plaque rupture is the most frequent cause of cardiac death in patients with coronary disease and that the haemodynamic impact of a chronic lesion plays a secondary role.17,18 It is not clear how well this concept translates to the left main artery specifically.
Although a dedicated randomised controlled trial to address the role of surgical revascularisation in patients with left main disease does not exist, there appears to be general agreement reflected in various current guidelines that surgical intervention in patients with significant left main stenosis improves survival compared with medical management and is therefore the treatment of choice.19,20 This notion is based mainly on the results of a meta-analysis of seven small to moderate-sized trials (including a total of only 150 patients) demonstrating a relative risk reduction in mortality of 68% at five years and an absolute survival benefit of 19 months.21
Subgroup analyses of two randomised trials – the VA Cooperative Study22 and the European Cooperative Study23 – as well as the large Coronary Artery Surgery Study (CASS) registry,24 all performed in the 1970s, supplied most of the data for this meta-analysis. Two-year survival in the VA Cooperative Study was 93% for surgically treated patients and 71% for medically managed patients. The VA Cooperative Study and CASS registry unequivocally demonstrated a significantly longer survival after surgical revascularisation in patients with a left main stenosis >50% compared with medical management and the European Cooperative Study showed an equally improved but non-significant survival (likely to be related to the small number of patients: total n=59) with surgery. Despite these results, a number of important findings deserve attention as they may be helpful when making treatment decisions.
First, in the VA Cooperative Study (total n=91), at 42 months there was no significant survival benefit in patients with a 50–75% stenosis (n=47) and/or normal left ventricular systolic function (n=23).25,26
Second, in the CASS registry (total n=1,492) at three and/or 15 years there was no survival benefit in patients with a 50–69% left main stenosis and preserved or mildly reduced left ventricular systolic function, with left main stenosis in the absence of a significant right coronary lesion or with left main disease with a right coronary lesion but preserved left ventricular systolic function.24,27 To put these findings into better perspective, in patients with normal left ventricular systolic function the 15-year survival was 51% in the medically treated group versus 42% in the surgically treated group.
Third, the two randomised trials were too small and therefore underpowered to demonstrate small differences in outcome in subsets.
Fourth, it is unclear how well the data are applicable to the present as surgical treatment has become safer with more durable results, and medical management has considerably improved the natural course of coronary disease. For example, in the VA Cooperative Study and CASS trial, the left internal mammary artery was used in fewer than 10% of patients assigned to surgery. Medical management in the study was virtually non-existent, as statins and angiotensin- conversion enzyme inhibitors were not available and aspirin was used in fewer than one-quarter of patients.
Fifth, the number of patients >75 years was very limited and recommendations based on these trials should be extrapolated into the elderly population with caution.
Finally, the merit of surgical revascularisation may have been underestimated as the patients enrolled would be considered at low risk on clinical grounds and the analysis was performed on an intention-to- treat basis, with a 40% cross-over rate to surgical revascularisation. Nonetheless, it is unlikely that any similar randomised trial will be repeated in the foreseeable future and it is therefore necessary to put the available data to clinical use as best as possible.
Angiographic, Ultrasound and Physiological Assessment
Angiographic assessment can be challenging due to vessel tortuosity, overlap, eccentricity and the frequent absence of a healthy reference segment. Hence, it is not surprising that large interobserver variability has been reported, particularly in angiographically intermediate left main disease.28–30 In fact, angiographic assessment of the left main artery is probably the least reproducible of any coronary segment.29,31,32 Furthermore, significant discrepancy has been found when comparing necropsy measurements with angiographic measurements.30 Angiographic assessment more frequently underestimates the degree of stenosis found at necropsy.30 To facilitate clinical decision-making in the presence of an angiographically intermediate left main stenosis, two tools have most frequently been used: intravascular ultrasound (IVUS) and fractional flow reserve (FFR) measurement.
Unfortunately, given the paucity of data, there is no consensus on which IVUS parameter is the most important and at which measurement revascularisation for a left main stenosis should be considered. For example, the most frequently discussed IVUS cut-off parameter – the minimal luminal area – below which a left main lesion is considered significant, has been suggested to be as high as 9mm2 33 and as low as 6.0mm2.34 Where do these numbers come from? The more conservative cut-off of <6.0mm2 has been determined from a single-centre study of 55 patients by comparing the minimal luminal area and minimal luminal diameter with FFR measurements.34 While there was no correlation between minimal luminal diameter determined by quantitative angiography and FFR measurements, left main minimal luminal areas of <6.0mm2 and diameters of <2.9mm by IVUS correlated very well with an FFR measurement <0.75.
In a different study, minimal luminal areas of angiographically normal left main coronary arteries were assessed in 121 patients.35 It was proposed that a minimal luminal area of <7.5mm2, which represented two standard deviations from the mean, may indicate significance. To validate this number, 214 patients with angiographically indeterminate left main lesions underwent IVUS examination. Patients with a minimal luminal area of ≥7.5mm2 were treated conservatively and did as well as patients who underwent surgical revascularisation for a minimal luminal diameter of <7.5mm2. Perhaps this allows safe deferral of left main revascularisation with luminal diameters equal to or greater than 7.5mm2. However, it does not clarify whether all patients with luminal diameters <7.5mm2 require revascularisation. Hence, it would be premature to conclude that revascularisation is invariably indicated in patients with a minimal luminal area of <7.5mm2.
An important potential problem with the use of IVUS to determine the significance of a left main lesion is the variable size of a normal left main artery depending on the patient’s gender and body surface area.36,37 These differences have not been taken into account in the currently available studies comparing IVUS parameters with FFR. For instance, it is likely that a left main stenosis with a minimal luminal diameter of 6.2mm2 has a different haemodynamic impact in a female with a body surface area of 1.0m2 than in a male with a body surface area of 2.5m2. For IVUS assessment of the left main artery, coaxial catheter and coronary wire alignment to the vessel is important. In the presence of a distal, bifurcational lesion, measurements during pullback from the left anterior descending (LAD) and left circumflex (LCX) arteries are also important. However, despite a meticulous technique, it is sometimes impossible to achieve coaxial alignment. Measurements under these circumstances should be interpreted with caution.
Fractional Flow Reserve
FFR has been used to distinguish which patients require left main artery revascularisation. In one study by Suemaru et al., 15 patients with angiographically equivocal left main disease underwent FFR measurements.38 Patients with an FFR >0.75 were managed conservatively, whereas patients with an FFR <0.75 underwent bypass surgery. After approximately three years, there was no difference in survival. Importantly, all of the medically treated patients had FFR >0.8: in fact, it was 0.85 or greater in seven out of the eight patients treated medically. Therefore, although it may be reasonable to defer left main artery revascularisation in patients with an FFR >0.80, concluding that a left main stenosis with an FFR of 0.75–0.80 may safely be left alone is risky.
In a similar, larger study by Bech et al., 54 consecutive patients with angiographically indeterminate left main disease underwent FFR.39 Patients with an FFR <0.75 were treated surgically and patients with an FFR >0.75 were treated conservatively. There was no difference in survival at three years, but the number of patients in the conservatively treated group with an FFR of 0.75–0.80 was not reported.
Finally, in a study by Jasti et al.,34 55 patients with intermediate left main lesion underwent FFR assessment. Patients with an FFR >0.75 (total=41) were treated medically or, if there was an additional lesion, in either the LCX or LAD artery (13 patients), they underwent percutaneous coronary intervention (PCI). Four patients with stenoses in the mid- to distal LAD and LCX arteries underwent coronary artery bypass grafting (CABG) and were excluded from follow-up. None of the patients with a left main FFR >0.75 who were treated medically or with PCI of the LAD or LCX arteries (total=37) died due to cardiac aetiology at a mean follow-up of 38 months. However, three non-cardiac deaths did occur. Two patients eventually underwent CABG and two PCI of a vessel other than the left main artery. At the end of follow-up, 10 patients had Canadian Cardiovascular Society (CCS) class 1–2 angina. The remainder were symptom-free. Once again, similar to the study performed by Suemaru et al., in the conservatively treated group all FFRs were equal to or greater than 0.80 and none were between 0.75 and 0.80.
The application of FFR for left main artery assessment has potential drawbacks. First, in the presence of concomitant lesions in both the LAD and LCX arteries, without repairing the downstream lesions the FFR may underestimate the true significance of the left main lesion. Furthermore, in ostial left main lesions with catheter-induced damping upon engagement, uncomfortable manoeuvres are sometimes repetitively required to engage, inject intracoronary adenosine and disengage the guide. Under these circumstances, the administration of intravenous adenosine should be considered.
Above all, if one is most concerned that plaque rupture is the reason for coronary events in patients with left main stenosis, the plaque composition should be more important than the luminal area or diameter or the haemodynamic impact of the lesion. While several technologies are currently under investigation, no tool has been proved to reliably measure plaque vulnerability.
Individualisation of Management Decisions
Given the above-described limitations and uncertainties in the assessment of a left main stenosis, one aspect becomes clear: although desirable, there is currently no single agreed parameter to allow a clinician to consistently make the right decision. Apart from critical lesions that clearly need further surgical or percutaneous intervention, or mild lesions that can be managed conservatively, treatment needs to be individualised according to the clinical scenario. Clinical (e.g. age, symptoms, electrocardiogram findings) and angiographic parameters (particularly the concomitant presence of right coronary disease and assessment of left ventricular systolic function in addition to physiological assessment and/or IVUS examination) need consideration when deciding whether a left main will require revascularisation or not. For example, it is probably reasonable to proceed with conservative management in a sedentary elderly female patient with non-limiting stable angina, a 60% left main lesion on angiography with no significant right coronary lesion, normal left ventricular systolic function, FFR of 0.78 and IVUS minimal luminal diameter of 5.8mm2. On the other hand, it may be more prudent to consider revascularisation in an active middle-aged male with a body surface area >2m2, stable angina, a 50% left main stenosis on angiography, a 70% mid-right coronary lesion, normal left ventricular systolic function, FFR of 0.78 and a minimal luminal area of 6.0mm2.
What Is the Role of Percutaneous Intervention and How Does It Compare to Surgical Revascularisation?
Balloon Angioplasty and Stand-alone Debulking
The first trials to evaluate the safety and efficacy of percutaneous left main intervention demonstrated variable results. Plain old balloon angioplasty has been associated with unsatisfactory outcomes related to high procedural mortality as well as restenosis. A higher incidence of restenosis may in part be related to the high proportion of elastic tissue and smooth-muscle cells causing significant recoil,40 a unique feature of the left main artery. For example, O’Keefe et al. reported results of elective balloon angioplasty in 33 patients with high procedural (9%) and late mortality (65% at 20 months). Repeat revascularisation was required in 42% of these patients.41 Likewise, Hartzler et al. reported a procedural mortality rate of 4% after balloon angioplasty in 103 patients with unprotected left main disease.42 More promising results were published by Stertzer et al.43 Nineteen patients underwent balloon angioplasty of the left main artery with a procedural success rate of 95% and emergency surgery in 5% (one patient). There were no procedural deaths; however, 37% ultimately required surgical revascularisation. Elective percutaneous balloon angioplasty of the left main artery was more or less abandoned44 until the advent of routine coronary stenting.
Of note, debulking (e.g. with rotational atherectomy) with or without balloon angioplasty in the absence of subsequent stenting has been reported in a handful of cases.45–48 It may be very useful in select patients as an adjunct treatment followed by stenting. For example, directional atherectomy at the left main bifurcation may allow single stenting, avoidance of more complex bifurcational stenting49 and, perhaps, restenosis.50 Likewise, rotational atherectomy may facilitate stent delivery and expansion in heavily calcified lesions.51–53 However, as a stand-alone strategy, rotational or directional atherectomy can no longer be recommended.
The use of stents has reduced the rate of acute complications as well as the incidence of restenosis.54,55 It thereby rekindled interest in left main intervention in the mid to late 1990s. As a result, numerous case series and small observational studies using bare-metal stents have been published including more than 800 patients. A detailed description of the individual trials would surpass the limit of this review.
Without taking all of these observational studies into account, the results were variable, with procedural or in-hospital mortality rates ranging between 0 and 20%.46,56–62 In addition, follow-up was almost universally short (most lasting for 12 months or less).
Many of these early trials were hampered by one major shortcoming: left main intervention was frequently reserved for patients deemed to be at very high surgical risk. For example, in the frequently quoted Unprotected Left Main Trunk Intervention Multicenter Assessment (ULTIMA) registry, the majority of patients were considered at high or prohibitive risk for surgical revascularisation. Incidentally, in this trial stents were used in only approximately 50% of patients. It is therefore not surprising that the results appeared to be less satisfactory than those reported for surgical revascularisation, yet these trials represented the basis for the most recent guidelines classifying percutaneous treatment for left main stenoses as class III (harmful).
It has become clear that the use of drug-eluting stents for the treatment of left main artery disease significantly reduces the rate of in-stent restenosis and the need for target vessel revascularisation compared with bare-metal stenting.63,64 In the only randomised trial comparing bare-metal versus drug-eluting stents, a total of 103 patients were randomised to receive a bare-metal or a drug (paclitaxel)-eluting stent.64 There was no difference in death, myocardial infarction or stent thrombosis at six months. However, the bare-metal stents were associated with a significantly higher incidence of in-stent restenosis and target vessel revascularisation (16 versus 2%).
Two randomised clinical trials comparing surgical with percutaneous revascularisation, as well as numerous reports of observational studies and case series, have been published. In the LE MANS trial,65 105 patients were randomly assigned to CABG or PCI. In the PCI group, the left main lesion was located distally in 56% of patients and the Syntax and Euro- scores were 25±8 and 3.3±2.3, respectively. The primary end-point was the change in left ventricular systolic function one year after revascularisation, which improved significantly in the PCI group but remained unchanged in the CABG group. The secondary end-point – major adverse cardiac and cerebrovascular events at one year – was similar in the two groups (Major adverse cardiac and cerebrovascular event [MACCE]-free one-year survival of 71 and 76% in the PCI and CABG groups, respectively). In this context, it is worth mentioning that only 35% of patients in the PCI group received a drug- eluting stent. An arterial conduit to the LAD artery was used in only 81% of the CABG patients (72% left internal mammary and 9% radial artery).
In the more recently published SYNTAX trial,66 the subgroup of 702 patients with left main disease were randomised to CABG versus stenting with paclitaxel-eluting stents.67 At one-year follow-up there was no difference in the primary end-point – major adverse cardiac or cerebrovascular events (15.8 and 13.7% for PCI and CABG, respectively). There was no difference in mortality or myocardial infarction; however, the stroke rate was significantly lower and the need for repeat revascularisation higher in the PCI group. The higher rate of repeat revascularisation in the PCI group was seen only in patients with concomitant two- and three-vessel disease. Importantly, while the MACCE rate was similar in patients with a low and intermediate SYNTAX score, it was higher in the PCI subgroup with a high SYNTAX score (>33), related mainly to a higher rate of repeat revascularisation. A unique feature of this trial is the high enrolment rate of screened patients (>70%), which has been <10% in most previous trials comparing percutaneous intervention with CABG or medical therapy. Therefore, the patient population and selection are more representative of a real-world population.
Although the Comparison of Bypass Surgery versus Angioplasty using Sirolimus-eluting Stents in Patients with Left Main Coronary Artery Disease (CoMBAT) trial comparing percutaneous intervention versus sirolimus-eluting stents with CABG will not be completed, the planning of a randomised multicentre trial, the Evaluation of Xience Prime versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial, was announced at a symposium of the Cardiovascular Research Foundation in March 2010. This trial will be designed to specifically examine percutaneous versus surgical left main revascularisation. The enrolment of 2,500 patients is anticipated.
In addition, numerous non-randomised series and registry data have been published, most with encouraging results. Four of these studies, including a total of more than 1,000 patients, have compared percutaneous intervention with drug-eluting stents versus surgical revascularisation in a non-randomised fashion with comparable early and intermediate (one-year) outcomes. There was a procedural mortality rate in the percutaneous group of 0–5% and a one-year mortality rate of 3–14%.68–71 Importantly, the majority of these lesions involved the left main bifurcation. As expected, the target revascularisation rate was higher after percutaneous intervention, ranging between 5 and 25%. The variability in the need for repeat intervention is related to lesion location (body or ostial versus distal/bifurcational) and the need for concomitant intervention in another coronary artery. The risk of target revascularisation is highest (up to 38% in one study)72 in patients with distal left main lesions and is very low for the rather uncommon ostial or mid-shaft lesions.73
The MAIN-COMPARE registry included 2,240 patients, 1,102 of whom were treated with unprotected left main stenting (784 with drug-eluting stents) and the remainder surgically.74 The unique feature of this study is the longer follow-up of three years. It demonstrated equivalent event rates (combined death, Q-wave myocardial infarction or stroke) but higher rates of target vessel revascularisation after percutaneous intervention (5% after drug-eluting stenting and 11% after bare-metal stenting).
Finally, several meta-analyses with conflicting results, depending on the design, have been published. One included data from 16 observational studies with a total of 1,278 patients treated with drug-eluting stenting for an unprotected left main stenosis.75 In-hospital mortality was low at 2.3%, as was the intermediate-term mortality rate (5.5% at 10-month follow-up). The MACCE rate was more favourable in the percutaneously treated patients than in the CABG patients. By contrast, another systematic review suggested a lower early and late mortality with CABG, regardless of whether bare-metal or drug-eluting stents were used.76 Although the rate of immediate procedural and in-hospital complications with left main stenting has generally been low, two important aspects that potentially affect the long-term results of left main stenting with drug-eluting stents deserve further discussion: stent thrombosis and restenosis.
Stent thrombosis, the Achilles’ heel of drug-eluting stenting, has critical clinical implications when it occurs in the left main artery as this would, in most instances, have immediate lethal consequences. Limited data are available to precisely define this risk; however, registry data suggest a thrombosis rate no higher than in other coronary arteries, and perhaps even lower. For example, in a multicentre registry including 731 patients treated with unprotected drug-eluting left main stenting, the incidence of probable or definite stent thrombosis at 30 months was 0.95%.77 Likewise, in the ISAR-LEFT MAIN trial including 607 patients treated with drug-eluting left main stenting, the two-year rate of definite or probable stent thrombosis was 0.5–1.0%.78 Incidentally, this is also the only randomised trial comparing the results of paclitaxel- versus sirolimus-eluting stents, with results demonstrating the equal safety and efficacy of both regarding the rate of stent thrombosis and restenosis. The three-year risk of definite stent thrombosis in the MAIN- COMPARE registry was 0.6%.74 Finally, in the DELFT registry including 358 patients who underwent left main stenting with a drug-eluting stent, the risk of definite, probable and possible stent thromboses at three- year follow-up were 0.6, 1.1 and 4.4%, respectively.79
Despite the overall low thrombosis rates, it is clear that the location of the stenosis at the bifurcation is associated with a higher thrombosis rate.72 In this context, the duration of dual antiplatelet therapy has been widely debated and remains to be determined. However, most would agree that for drug-eluting stents antiplatelet therapy should be taken for no less than 12 months, preferably longer and (some would argue) indefinitely in patients with complex stenting.80 Evidence aside, late and very late stent thromboses are seen after coronary stenting. All clinicians have been consulted at one time or another regarding the need for premature discontinuation of double antiplatelet therapy. When the stent or stents are placed in the left main artery, many will likely admit to an uncomfortable feeling when double antiplatelet therapy interruption or discontinuation is necessary for various clinical reasons.
Similar to the incidence of stent thrombosis, the target revascularisation rate is highly dependent on stenosis location and is highest for bifurcational lesions.68,69,72 For example, in a single-centre study of 70 patients with in-stent restenosis of drug- eluting left main stenting, 49 (70%) were located at the bifurcation.81 Similarly, Valgimigli et al. demonstrated a significantly higher major adverse cardiac event rate in 130 patients treated with sirolimus- or paclitaxel-eluting stents (after a little over one-year follow-up) when the lesion was located distally versus in the shaft or at the ostium (30 versus 11%).82 This was related primarily to a higher rate of target vessel revascularisation. Likewise, Price et al. reported a high rate of in-segment restenosis (42%) in 47 patients with distal left main stenosis treated with a sirolimus-eluting stent.72 Taking the available data into account, it is reasonable to state that the procedural, short- and intermediate-term event rate with drug-eluting left main stenting is low. Furthermore, stent thrombosis is likely to be low. Location of the lesion is key. Unfortunately, treatment of the most common (bifurcation) lesions, particularly with two stents, appears to be associated with a high rate of target lesion revascularisation and, perhaps, a higher stent thrombosis rate. Therefore, in bifurcational lesions, provided it is technically feasible, a provisional approach of single stent deployment across the bifurcation (most commonly into the LAD artery) with additional side-branch stenting only if the result after side-branch balloon angioplasty is suboptimal currently appears to be the most favourable option.83–87
In conclusion, the importance of left main coronary disease has been recognised and characterised in the past four decades. The importance of surgical or percutaneous intervention at the extremes of the spectrum is understood, but the management of intermediate lesions remains a challenge that requires many more factors to be taken into account than the angiographic degree of stenosis. Once committed to treatment, a dogmatic approach to the treatment modality is likely to lead to unfavourable outcomes. At this point, it is reasonable to offer percutaneous treatment to patients with prohibitive surgical risk. However, for patients suitable for both surgical and percutaneous revascularisation, although stenting is very promising, until more data from randomised trials dedicated to left main revascularisation are available, the surgical approach should remain the primary consideration in complex anatomy (particularly in challenging bifurcation disease). Percutaneous intervention should be reserved for select patients with very favourable clinical and anatomical (e.g. ostial or shaft lesions) characteristics after a thorough discussion with the patient. This should ideally include both a surgeon and interventional cardiologist, emphasising the as-yet unknown long-term outcomes.
Important further research directions include the exploration of dedicated bifurcation stents that will hopefully reduce the rates of in-stent restenosis and/or thrombosis in the most common subset of patients with distal (bifurcation) disease. Perhaps more widely available percutaneous assist devices (e.g. Impella) will permit safer performance of left main interventions in patients with complex disease and/or severely reduced left ventricular systolic function. In addition, further studies will need to clarify the best type and duration of antiplatelet therapy. Finally, further research is needed to examine the fundamental mechanism of poor outcome in patients with left main artery disease and to determine which plaque characteristics are associated with clinical events.