Stenting the Unprotected Left Main: Finding Answers With the EXCEL Trial

EXCEL is a global, prospective, multi-center, randomized trial to assess the safety and efficacy of the Xience Prime/Xience V everolimus-eluting stent (Abbott Vascular, Santa Clara, California) compared to coronary artery bypass grafting (CABG) in select patients with unprotected left main coronary artery disease. Dr. Stone is a principal investigator.

What do we know about the percutaneous treatment of left main disease?

There is a long history of angioplasty in the left main, starting with balloon angioplasty back in the late 1970’s, then bare metal stents, and now drug-eluting stents. While we are incrementally getting closer and closer to surgery in terms of outcomes, we are still not there, although we think we might be. The most recent major study, the SYNTAX trial, randomized 705 patients with unprotected left main disease to first-generation paclitaxel-eluting stents versus bypass surgery. Overall, the unprotected left main cohort did fairly well with intervention, with similar rates of death and less stroke with stenting, although there was a higher repeat revascularization rate with percutaneous coronary intervention (PCI) compared to surgery. However, when you look to the outcomes according to the SYNTAX score, a measure of the anatomic complexity for PCI, the lower 2 tertiles with left main disease — approximately two-thirds of patients — seemed to fare very favorably, even with the first-generation paclitaxel stent, when compared to surgery. That is why we have now planned the large-scale, definitive EXCEL trial, which will include 2,600 randomized patients and compare surgery to stenting, but with the next-generation, everolimus-eluting Xience Prime stent. We are hoping that with this study, which is restricted to the lower two-thirds of SYNTAX tertiles, we will be able to show for the first time that stenting is at least as good, if not better than, surgery for selected patients with unprotected left main disease.

What are the study endpoints?

The primary endpoint is death, myocardial infarction (MI) or stroke, measured at three years after the procedure. There are three important things to note about the primary endpoint. First, it is a composite of the three hard endpoints that matter the most: death, myocardial infarction and stroke. Prior studies have also added repeat revascularization, which while being an efficacy indicator of the two procedures, is certainly much less important than death, MI or stroke. Most people feel that repeat revascularization can be placed in a long list of other adverse events such as atrial fibrillation, renal failure, and many other events that could have been included as a primary endpoint. But everybody agrees that death, MI and stroke are the most important — and these occur frequently in left main patients not revascularized. Second, the primary endpoint is measured at three years, which is a departure from most other studies that have had a primary endpoint at one year. We believe that three years is more appropriate, because it certainly allows all the restenotic episodes to be included. There is data to suggest that some events might continue to accrue in the PCI arm, but perhaps not in the bypass surgery arm, at least up to that time period. After seven years, more events may occur in the bypass surgery arm, and so in this trial, we are going to follow patients for at least five years and possibly ten years. Third, the trial is powered for sequential non-inferiority and superiority testing, and by that we mean that the main goal of the trial is to show that PCI is non-inferior to surgery, which means at least as good as or better. If that bar is met, then we will test whether unprotected left main stenting with the everolimus-eluting stent is actually superior to surgery.

Can you describe the planned patient population?

For the most part, patients enrolled will be elective, but the trial will also include acute coronary syndromes. Patients within the acute throes of an ST-segment elevation myocardial infarction (STEMI) will be excluded, but we can include unstable or recent non-STEMI. The main selection criteria are a left main stenosis of at least 70% diameter stenosis (or greater than 50% diameter stenosis with associated ischemia), plus a SYNTAX score of ≤ 32. Very high-risk patients with multiple chronic total occlusions and diffuse triple-vessel coronary disease, those with a SYNTAX score of ≥ 33, should, as a rule, be treated with surgery. For those with a SYNTAX score of ≤ 32, there is clinical equipoise based on all current data that angioplasty with a very good drug-eluting stent may be as good or better than surgery in these patients, so this is the population that will be randomized.

Why is the left main so difficult to stent?

The left main is a high-risk segment of the coronary artery tree because it supplies up to 60-70% of the myocardium of the heart. If a stenosis develops in that one segment, it is like developing a stenosis at the trunk of a tree. If a tiny branch of the tree is lost, the tree survives. If the whole trunk is lost, the tree dies. That is a pretty apt analogy for the amount of myocardium supplied by the left main coronary artery. It is arguably the most important inch of artery in the human body. On the other hand, the left main is actually a favorable target for angioplasty because it is a large vessel, proximal, and contains relatively short lesions. With the technically advanced devices and techniques we have developed over the years, it is relatively easy to obtain a good angioplasty result in this location, and restenosis and stent thrombosis are quite low in the left main coronary segment. In the old days, when we just had balloon angioplasty, technically, left main PCI was a very difficult procedure. When we had bare metal stents, the procedure became easier, but restenosis was still significant. Restenosis in an area supplying so much myocardium would often result in sudden death, heart failure or shock. But now, with drug-eluting stents in the left main segment, restenosis and stent thrombosis rates are extremely low. The bigger issue is that patients usually do not just have isolated left main disease; they also have other lesions in one, two or three coronary arteries. The real benefit of the second-generation stent is in treating those additional lesions.

Should stenting a protected versus unprotected left main be considered two completely different procedures?

Yes, it really is very, very different. A protected left main is more like treating a single coronary artery. By definition, you are not treating such a large myocardial bed. You are only treating part of the myocardium that the left main supplies, and you can for the most part ignore the other major branches that are protected by bypass grafts.

What about possible complications for unprotected left main stenting?

Technique does have to be meticulous, given that the left main supplies so much myocardium. If an occlusion or a complication in that segment does occur, while fortunately rare, the left main supplies such a large extent of myocardium that the patient will become hypotensive and/or fibrillate and/or die much more frequently than if there was an occlusion or complication in a smaller arterial segment. It requires definite operator expertise to handle the left main segment, especially in the two-thirds of cases where the distal bifurcation of the left main is involved. That is where the left main coronary bifurcates into the left anterior descending artery and the left circumflex artery. In fact, sometimes it even trifurcates, and there are a variety of different ways, depending on the anatomy, to handle those situations, but it can be technically challenging. You need very good, very experienced operators that are used to dealing with large amounts of myocardium, complex bifurcations, calcified disease, and avoiding and managing complications.

What about the use of intravascular ultrasound (IVUS) or a left ventricular assist device (LVAD)?

We strongly recommend IVUS for left main intervention. It is very important to confirm an optimal stent result in the left main segment with IVUS. Other devices such as intra-aortic balloon pumps or LVADs such as the Impella are used selectively. They are not required in most cases of left main intervention, but if you have a patient with a high grade left main disease and either an occluded right coronary artery, hypotension, or left ventricular dysfunction, then often we will perform those cases with either a balloon pump or an Impella. But I would say those are ≤ 10% of patients.

Has it historically been difficult to choose appropriate patients?

Choosing appropriate patients for any major procedure is always a critical issue. The study handles it by first prequalifying operators so we know we are only getting good operators and good sites, both for surgery as well as for intervention, because left main surgery is also higher risk than non-left main surgery. There are very strict inclusion and exclusion criteria. There must be clinical equipoise for an individual patient to be enrolled, meaning both physicians believe that PCI is a feasible option as well as surgery. We are also very proscriptive in describing what techniques are used in different situations: how IVUS should be used, how fractional flow reserve (FFR) should be used, when you might consider using a balloon pump or Impella, staging a procedure, how to manage a distal bifurcation. All of that is spelled out in extreme detail in the protocol.

Are there aspects of the Xience stent that make it ideal for this type of trial?

Xience is the one stent that seems to have the lowest stent thrombosis rate in all settings, and that is presumably attributable to many factors, but mainly to the fluoropolymer that is being used as a drug-release vehicle on the stent. Xience also has a very low rate of restenosis. It is very deliverable, so it is very good in complex disease, and the outcomes in general have been about 40% better in non-selected patients compared to the Taxus stent. I do not know if it will offer any advantages in the left main segment itself; some studies suggest that the drug-eluting stent choice is less critical in that segment, because it is a large vessel, usually with a short lesion. However, in all other coronary lesions, the Xience stent will significantly reduce adverse events. This is important because very few patients enrolled in EXCEL will have isolated left main disease.

What about the interaction between interventionalists and surgeons?

Right now, the standard of care to treat unprotected left main disease is surgery. That is how most patients continue to be treated, even in the United States, where I would estimate 99% of elective left main patients undergo surgery. In the older guidelines, elective left main stenting used to be a class III recommendation, which means don’t do it, it does not lead to benefit and it may be harmful. By the current guidelines after SYNTAX, it is now considered a class IIb recommendation, which means it may be considered in selected cases when unprotected left main stenting is low risk and/or complex surgery is high-risk. But while class IIb means it may be considered and may be reasonable, this is not a ringing endorsement, and so that is why we are doing the current EXCEL trial. That being said, all cases of left main disease require the collaborative decision-making of a general cardiologist, interventional cardiologist, and the cardiac surgeon, the so-called heart team. That should happen in real life and it is also happening in the confines of the EXCEL trial.

How frequently have you been stenting both types of left main disease?

We have been stenting protected left mains for many, many years. That is a post bypass patient where one or more of the grafts are diseased or closed, but there is at least one patent graft to either the left anterior descending or left circumflex artery, a not uncommon occurrence. We have also for decades been treating high-risk patients with unprotected left main disease who are not candidates for surgery. These are patients who absolutely refuse surgery, have a porcine aorta, prior mediastinal radiation, or other situations that make them too high-risk for surgery, like chronic obstructive lung disease. But now I must say that with the more recent class IIb recommendation for elective left main stenting, at Columbia we are treating more elective patients outside of the study with unprotected left main disease, although the majority of patients still go to surgery. As the EXCEL trial begins at Columbia University Medical Center, all of those patients who otherwise meet study eligibility criteria and agree to participate will be randomized to surgery versus PCI.

Dr. Stone can be contacted at gs2184@columbia.edu.

Disclosure: Dr. Stone reports he is a consultant to Abbott Vascular, Boston Scientific and Medtronic.