Our panel of experts dives deep into the ISCHEMIA trial, focusing on trial design, questions answered, need for future study, and take-home points.
Issam Moussa: Thank you for joining us for the inaugural session of the Trial Examiner, a video-based series to discuss important landmark trials in the field of interventional cardiology in which we will have conversations with experts and dig into the details. With ISCHEMIA, we want to discuss what it means for clinical practice and understand its future implications.
It is a pleasure to welcome Drs. Waxman and Bangalore who are obviously very well known experts to discuss the ISCHEMIA trial. Dr. Waxman is going to moderate the session and Dr. Bangalore and myself will be the recipients of his questions and engage in discussion.
Ron Waksman: Thank you very much. My name is Ron Waxman and I am an interventional cardiologist at the Medstar Washington Hospital Center in Washington DC.
I will serve as a moderator and our two discussants are Dr. Sripal Bangalore, who is an interventional cardiologists and director of research at the cardiac catheterization at NYC New York and he also was very instrumental in the ISCHEMIA trial.
We also have Dr. Issam Moussa, who is professor and head of the department of clinical sciences at the College of Medicine in University of Illinois.
To set the stage, ISCHEMIA is a study that has been going on for a while and it has been almost a year since it was presented at the American Heart Association annual meeting. Dr. Bangalore, please give us the highlights. What was the main question of the ISCHEMIA trial and how was ISCHEMIA designed to answer the question?
Sripal Bangalore: Thanks, Ron. The ISCHEMIA trial was primarily designed because of the questions that were left over after the COURAGE and BARI 2D trials. The questions that many of us were asking is if you were to take a higher risk group of patients who have stable ischemic heart disease, say, those who have moderate or severe ischemia, would a strategy of routine revascularization be associated with significant reduction in cardiovascular events. And this was important because there was some criticism of the prior trials given the fact that they enrolled patients after diagnostic cath. Maybe many of the investigators if they saw a significant tight proximal LAD, they would have not enrolled those patients. So, we just don't know the answer for those.
And, also around that time Rory and others have shown from nuclear SPECT data that, if you look at observational data, if you have 10% or more ischemia, revascularization at least in an observational data set was associated with a reduction in death or myocardial infarction (MI).
So, we wanted to address the question if you specifically chose a higher-risk group of patients. Is there a benefit of a routine invasive strategy in reducing cardiovascular events. That was our primary outcome and secondarily, if they are symptomatic, would there be a benefit at reducing symptoms and improving quality of life?
One of the big criticisms for prior trials was could there be a selection bias. If you do a diagnostic cath, could people pick and choose who they want to enroll in clinical trials versus saying, I know to manage this really complex patient. I don't want to put them in a clinical trial.
Because of this the trial was designed to enroll patients upstream of the cath lab so patients needed to have moderate or severe ischemia and the enrollment was before they could get coronary angiography.
Ron Waksman: It's a very interesting question. I kind of want to challenge the question and maybe ask Dr.Moussa to tell us what he thinks. Is this question legitimate, how do you measure the ischemia and where do you put the cutoff to address the question of this study?
Issam Moussa: I think, as Dr. Bangalore mentioned, the concept that we don't have sufficient information about management of patients with severe ischemic heart disease is absolutely needed. Now, obviously we have the benefit of hindsight in terms of patient the trial included and what it means.
So, the question I would ask is whether the population in the ISCHEMIA trial was actually high risk, as intended and how we make that decision. I would say that, obviously now we know the that the stress testing on functional testing didn't correlate whether you help mild or moderate or severe but I would go back a little bit more. If you look at the mortality in the medical group in the ISCHEMIA trial, which was I think was approximately 8.2%, it's exactly similar to the COURAGE trial, which I think the follow-up was 4.6%. So, if the mortality in the medical group between ISCHEMIA and COURAGE is exactly the same, does that mean that the population of ISCHEMIA is similar, essentially to the population of COURAGE, especially with the view that only 40% of multivessel disease and 36% had proximal LAD.
The question that I would like to bring out is whether the population of the ISCHEMIA trial is high risk or is it really a more COURAGE-like population? In BARI 2D, the mortality in the medical group was 12%.
If you take the STITCH trial, which is bypass heart failure, the mortality is 42%. So, do we believe that the ISCHEMIA trial population is high risk or is it really a COURAGE-like population?
Ron Waksman: Maybe before you answer, Dr. Bangalore, I think the question arises also from the difficulties in the actual study to enroll only patients at high risk of ischemia and it was blended with some patients who had a little bit less ischemia. Maybe you can address the challenges of enrolling into those subsets of patients and what was the differentiation between this kind of study and to COURAGE, which was not associated and didn't measure the degree of ischemia.
Sripal Bangalore: These are definitely critically important questions. We started with having a higher threshold for entry; it has to be moderate or severe ischemia determined by site but confirmed by a core lab. The core lab would work with the sites and suppose the site gets it wrong. The core lab would have a session with them to educate them and make sure that they read it right to make sure that we are enrolling patients with moderate or severe ischemia.
I’ve been on this trial for now 10 years and it feels like it's been a long time. When I first started, I was a bit skeptical. I thought that this was going to be a trial of moderate ischemia. Anytime you design a trial and you say you're going to enroll the highest-risk group of patients and you say, moderate, or severe ischemia, you would think that sites would do the bare minimum to get in. But we were really surprised—at the end of the trial, 54% of patients had severe ischemia. If you look at the absolute numbers, there was more severe ischemia than moderate, CCTA-based multivessel disease was approximately 77%, and LAD disease that was already present, and proximal disease in nearly half of the patients based on CCTA.
If you purely look at ischemia anatomy, and I’ve talked to Gregg Stone, MD, and Philippe Genereux, MD, and Hamovitz and they say that they've never seen this degree of coronary artery disease and this degree of atherosclerosis in any other trial.
Based purely on those metrics, I think we succeeded in enrolling the higher-risk group of patients that we envisioned we would. And I think that Dr. Moussa’s point is absolutely right. If you look at the mortality, it's not that different from COURAGE. And the question is twofold.
First, maybe it's because of modern medical therapy. COURAGE was over a decade ago and things have changed substantially and maybe it's because of that the mortality rates are, even with a moderate, severe ischemia higher-risk group of patients in stable risk coronary disease the mortality rate is very comparable.
Second, I think we really need to start questioning the importance of ischemia itself. I mean, is it a good stratification, or is anatomy a better stratification than ischemia alone.
Ron Waksman: That's a very good question because in your introduction you mentioned the study of Hamovitz, which is almost 20-30 years ago. That's the only study that was actually available. So, how do you design a study based on such a very old study that has never been replicated in a meaningful way? How much do we really need to rely on that study to design such a robust study that costs millions of dollars, across 370 sites, etc?
Sripal Bangalore: I must say that the Hamovitz study, even though it was just one study, it was what we practiced. I can tell you that we would get phone calls from the stress lab, saying, ‘Oh, we have a patient with severe ischemia. Can you take him to the cath lab?’ It became an urgent referral. We were practicing that. I think that was one of the impetus to say COURAGE has been criticized because of this assumption that there is a lower-risk patient set.
So, why don't we set the threshold higher and challenge what we were doing, which was if you have moderate or severe, and especially severe ischemia, people are rushing patients to the cath lab.
Issam Moussa: I think it's important to define what sort of population this really is. For example, yesterday I was discussing about a patient with ostial LAD who was admitted with heart failure and ejection fraction over 40% (not 35%).
The question in ISCHEMIA shows there's no difference in this ostial LAD. So, we go back and evaluate angiographically and ischemia was 36% proximal LAD. I don't know how many are ostial and how many are actually heart failure that were excluded that you have 35%.
I think this particular point is important, especially in the view that those with severe angina were excluded (approximately 35% had no angina).
If somebody has truly severe multivessel disease with moderate to severe ischemia, how could it be that 35% have no angina and the rest really have morbid class 2 or 3?
The purpose of this discussion is to put out an interpretation; do we believe at the end of the day, knowing what we know now, incorporating CT angiography medical arm outcome, do we feel that the ISCHEMIA population was truly a high-risk population compared to say COURAGE or BARI 2D? Let's not talk about STITCH, which I think was very high-risk bypass low EF. But between COURAGE and BARI 2D, do we feel it's in the middle?
BARI 2D mortality was 12% in the revascularization I think, not the total population. So, do we feel that the ISCHEMIA population risk is closer to COURAGE, to BARI 2D, or in the middle?
Sripal Bangalore: Those are all great questions. You know, more than what patients were enrolled in the trial, we always have to absolutely make sure to know what patients were not enrolled in the trial. I think your point is absolutely right.
Some things, for example, in a patient with NYHA class III or IV? That was an exclusion. If the patient had severe angina, that was an exclusion. If the patient had unacceptable angina despite medical therapy. For example, they were patients who were already maxed out on their medical therapy and there was nothing else to do. We just felt that, based on the COURAGE experience, if you were to enroll those patients, they're going to cross over. That’s not the question we were trying to answer.
The question that we really were looking for is the ischemia hypothesis, which was that if you have a high degree of ischemia, relieving that ischemia by invasive revascularization is actually better and that's one of the reasons we do not want a very high symptom group of patients.
If we go back to take a look at all of the trials, they've been consistent in showing that revascularization works for symptoms. I don't think any of us were questioning that.
So, we excluded patients with heart failure, who have less than 35%, ACS within 2 months. So, the example that you were giving of ostial LAD and heart failure would likely have been excluded. And, I think your point is absolutely right, people have to know what was not included in the trial.
I think we should all ask ourselves since we are all interventional cardiologists, is ISCHEMIA a positive trial or a negative trial?
I think ISCHEMIA is as positive as it can get for revascularization. What did we show in ISCHEMIA? If the patient is symptomatic, there is a significant benefit in the number needed to treat (3 patients) to relieve symptoms, and that symptomatic relief is long lasting at 4 years.
That has not been shown in other trials. We showed that revascularization, PCI or CABG, relieve symptoms. We also showed no difference in death or MI, but there was a reduction in spontaneous MI, an increase in procedural MI, and a reduction in unstable angina.
I think that makes for a nice discussion with the patient because there is this misconception that, I mean if the overwhelming goal is to say revascularization in stable coronary artery disease improves survival, that's where the problems comes in, in my opinion.
That's exactly what we showed in ISCHEMIA; there's no survival benefit. There may be benefit for other things, but that's what we need to discuss with patients.
Ron Waksman: Well said. I think we have discussed a lot about the population and the degree of ischemia and should move on to, not only the results, but also to the endpoints that have been discussed. I'd like to spend a few minutes discussing the composite of the endpoints.
So, we have revascularization, we have death, and we have an MI. I don't think there is much dispute about those are all hard endpoints, but within MI there was a differentiation between spontaneous MI and periprocedural MI.
I'd like to ask both of you to give me your sense. Why do you have to include periprocedural MI in those questions because we actually don't know the relevance of periprocedural MI. It's becoming such a controversial issue. When we look at spontaneous MI there is actually no dispute.
Sripal, because you are one of the study leadership, can you tell us why and how did you tackle the whole issue of periprocedural MI?
Sripal Bangalore: I can tell you that periprocedural MI was the biggest thing that we discussed and debated for a long time. And you can imagine them in this steering committee, as surgeons, interventionists, and also cardiologists, we had to have a balance.
If we were to say we’re not including procedural MI at all, we would be accused of saying that this is a free ride for doing what you want to do for an invasive strategy. There will be a lot of people criticizing us.
So, we came up with a very stringent definition because we did not want to just look at biomarkers with the universal MI definition, which is fairly relaxed. I mean, it's five times the elevation in troponins and we all know if you were to do a complex lesion you may reach that threshold and there may not be any clinical consequence of such a thing. So, we debated and discussed for a long time and came up with this definition, in which the thresholds were very high.
You not only needed CKMB threshold, which are much higher, you also needed something else in addition to that just as pure biomarker release. So, our goal was to see if we could come up with a clinically meaningful definition that is just not super sensitive that you're just going to focus on this biomarker release but potentially could have implications in the longer term. It's not a simple biomarker release but a more stringent definition.
Ron Waksman: Dr. Moussa, do you want to rebuttal this?
Issam Moussa: I think I would go back to the initial endpoint of death and MI, before the sort of adjustment of the endpoint because there weren't enough primary endpoints. So, really there are two important points to discuss.
First, that as the trial went through and the primary endpoints were being collected it was recognized that the trial will have the power to show a difference in death and MI.
Think about the population in their early 60s and with follow-up of only 3.5 years. As you know, the STITCH trial looked a very high risk and took 5 years to show a minor difference in cardiac mortality and 10 years to show a difference in total mortality in a population that's extremely high risk.
My major issue with the primary endpoint in ISCHEMIA is that it did not have the power to answer the question about a reduction in death and MI. Also, the follow-up is really too short for somebody who is 62 and whose anatomic risk seems to be intermediate.
I would say an interpretation of the ISCHEMIA trial might be that in somebody who is 62 and who has this level of ischemia maybe needs 10 years to show a difference in mortality.
We would say at 3 or 3.5 years this strategy doesn't improve mortality.
Ron Waksman: You're deviating from the question.
Issam Moussa: Really?
Ron Waksman: This is not about the power to the mortality right now, that's a separate issue by itself. You know, they are somewhat connected because, in all honesty, why do people introduce periprocedural MI?
Issam Moussa: I am with you. The question though, and I'm going to put myself in the other shoe, which could be a cardiac surgeon, could be a general cardiologist to say if the periprocedural MI affects the long-term outcome, then I want to know about it. Right? And I think if the periprocedural MI is troponin 3 and there's nothing, then I agree with you Ron.
So, really, to me the question is does the ISCHEMIA definition of periprocedural MI correlate with long-term outcome. Obviously, this is a different discussion. I think it's good that was chosen to be a little bit of a higher threshold. But I think that is a different discussion and we're going to see and work, hopefully, to understand whether the periprocedural MI ischemia by other trials correlates with long-term outcomes.
Ron Waksman: I will submit to you that it's really not relevant, other than power. I mean, you can put it as a secondary endpoint to address the question of whether periprocedural MI predicts long-term outcome, but for a study like that you need it just for power. It's a little artificial in terms of the interpretation of the study, especially in this specific study when the curves have been convergent. And then they have been flipped. Its just very, very confusing. I'm going to ask Sripal in a second, should you have not included periprocedural MI, would you able to do that study from the power perspective and what would be the outcome of this study?
Because I agree with you, if periprocedural MI is still important, then we should be seeing that in the long-term outcomes. We don't have to measure it as a biomarker, to include that biomarker in the outcome of the study.
As a matter of fact, this has haunted us again and again. I mean, consider just the last controversy regarding EXCEL. I know you're going to have it in the Trial Examiner series so we're not going to talk about it right now. But I'd like to throw out something provocative: that periprocedural MI should not be combining any primary endpoint, for any study, because this is just basically an easy ticket for power and it could be distorted because we really don't know. I don't know if it's 10 times more or 20 times more. And if it does, let it be. So, we’ll see that the patients are not going to do that well over time.
Issam Moussa: I really never thought about that. I think, in that sense, the argument that we’re going to capture long-term adverse events from periprocedural MI as a justification for not including it, I think it makes sense. The question is, what would be the long term? You would imagine higher mortality, maybe heart failure. I don't know. Sripal, what are your thoughts? I think Ron’s point sounds reasonable.
Sripal Bangalore: Yeah, it’s definitely reasonable. But this is a debate we've been having for almost 2 decades, or even more. The issue is, we really need to draw a line. The routine procedureal MIs, i agree. It's not prognostically important, the biomarkers. But really, don't we want to know if somebody does a bad job of revascularization, for example closes off a large branch and that causes a procedural MI. We need to know; that's an upfront risk for the patient.
In fact, our definition actually included a higher by market threshold, they should have either and geographic evidence of closing off a major side branch or EKG changes, etc.
This is an interesting discussion. We are looking more into this data on procedural MI. We’re coming up with a paper which will be published soon on whether procedural Mis have the same prognostic implications.
Issam, to your point I’m pleased to announce that we have been approved for a long-term follow-up of the ISCHEMIA trial. So, we should know whether at 10 years, if there is a difference in mortality between the two treatment arms. NDLBI is funding it so we will have that data.
Issam Moussa: Excellent. In a future trial, maybe it can be included as a safety measure. If we're thinking it is important and it is necessary communicate to patients, and again this is going back to looking at long-term spontaneous endpoints and whether the periprocedural MI . . . again discussion going to go on, I agree. But I think it's an important point.
Ron Waksman: Before we, before we leave the topic, I would say that most centers right now do not collect the enzymes from the procedure the patient is discharged the same day to home. So, it's really not a measure that is relevant in my view.
I'd like to hear from you Sripal, what would have happened if you didn't have the periprocedural MI. Would the study results have been changed with respect to the outcome?
Sripal Bangalore: Good question. This is one of the reasons why we not only had the composite endpoints, but also individual endpoints. One thing that I must say is, forget procedural MIs. Let's look at death curves. We’d all agree that whether procedural MI is important or not, you should have an impact on some longer-term outcomes. Take 4-year death curves superimposed—it's right on top of each other. I mean, whether you trust procedural MI or not, in terms of survival what we're seeing is no different.
So, for some endpoints like that I don't think we would have seen a difference. And the fact that there is a difference in spontaneous MI would have persisted, even if you did not include procedural MI.
Issam Moussa: I want to hear your perspective on the issue of a patient who's in their early 60s with a normal ejection fraction. Knowing what we know about mortality from other trials, why would we even expect to have mortality within 3-4 years?
Sripal Bangalore: Absolutely. I think that's the key take home because as you were saying, there was this notion that if you have severe ischemia, they need cath now. And the reason that people say that they would get referrals directly from the stress lab is people just felt like these are a really high-risk group of patients and we cannot wait a week.
And now we're saying at 4 years there’s no mortality difference and we kind of expected that. I think that allays some of the concerns that people had about the severe ischemia.
We've shown that you don't need to rush and in fact in a meta-analysis of 14 randomized trials we recently published showed that even when you combine all of the trials there just isn't a mortality difference.
I think we as interventionalists need to move beyond mortality in patients with stable ischemic heart disease. We are not going to find it.
Ron Waksman: We’ve already moved beyond that, I think. We never claimed that absolutely.
Sripal Bangalore: Yeah.
Ron Waksman: For patients the PCI for ischemic/nonischemic impact on mortality, but I think it was mainly on quality of life. I think you have equipoise here between the therapy and between PCI for quality of life.
You have shown that the study was positive for PCI, it was a big win for the quality of life of those patients. So, the question is, what is the interpretation and what do you take from that? How you differentiate the results of the ISCHEMIA trial from the COURAGE trial?
The reason I'm asking that because after the COURAGE trial there was a huge impact on PCI volume, it dropped by nearly 15%.
It actually enforces into the guidelines; into the appropriate use criteria. It has had a lot of impact, even though the study itself was controversial in terms of the interpretation because the quality of life was not there all the way.
And the bottom line, there was a question about patient selection and there was no degree of ischemia and there was a nuclear study that actually was in favor. But again, it was not primary endpoint. So, that actually left a lot of questions unanswered. My question to both of you is do you think that are wiser today after the ISCHEMIA trial than we were after the COURAGE trial? And, do you expect that the ISCHEMIA trial will penalize the PCI volume by another 15%, or 5%, or not at all?
Sripal Bangalore: Tough questions. What I take away from ISCHEMIA is that it provides a great platform for informed decision-making. Prior to this, guidelines have historically said patients should be on two or three antianginals and fail, and then you consider revascularization
We didn't do that in ISCHEMIA. If you were randomized to an invasive strategy you could be on zero medications and you got cath and revascularization and that strategy was superior in terms of quality of life.
I think this provides a nice platform for informed decision-making. So, if you have a patient who is symptomatic, I would tell them you can choose two options; one is medical therapy and one is cath and revasc. If you choose cath and revasc, you likely will have faster relief of angina and it is likely going to be durable longer term. But, if you choose to try medical therapy first, it's not going to change your survival. You can try it and if it doesn't work, you can always do cath and revasc.
My take on ISCHEMIA is not that you should always try to maximize your medical therapy before you offer the choice of cath and revac; I think that's the distinction that I draw.
Ron Waksman: Do you believe that we should make any guidelines comments on that part because, you know, COURAGE came out very categorically that we have to try medical therapy first and only then, if you fail, are you allowed to do PCI, even if it was a type A proximal LAD. Do you think the results of ISCHEMIA should change anything or should we basically leave things as they were with COURAGE?
Sripal Bangalore: I must disclose that I'm on the revasculatization guidelines writing committee, so I won't be able to give you the answer to that, but I think these points need to be discussed because it's the trial conduct and the application of the trial and my take on this is both are reasonable options you present to the patient and let the patient decide and not force patients to say, okay, you need to make sure you follow a conservative strategy first and then only do the invasive strategy. That’s not what ISCHEMIA did.
Issam Moussa: I think there are two points left like to bring about.
First, thallium or stress testing in general really does not prove or disprove ischemia. It's a functional physiological test. And I think there's now a lot of doubts that the accuracy of thallium stress testing, and potentially other stress testing, is not where we thought it should be.
Obviously, the FFR studies, even ISCHEMIA, when you have 20% of patients with severe ischemia and you have no severe disease, I would say there's a lot less confidence in what an ischemic stress test means based on this data, and otherwise other data.
Second, it is important for the community to put up front what population we're talking about. Even though the publication clearly states upfront the studied population to the community, physicians aren't looking in a granular fashion about who are the patients were presented in this trial. I would say if somebody who's in their early 60s, with a normal ejection fraction, and who has little or no angina then I'm absolutely willing to go, Okay, I'm kind of feeling well. If you're feeling well and you're not on four medicines and you know have to be a two-vessel disease and we’ll leave you alone.
If somebody with clear angina, on one drug, a type C proximal LAD and double vessel disease, the question is do I believe a stress tests or do I do CT FFR, or what else? So, I think clearly, this study guides us in the intermediate-risk population. And I think it opens a lot more questions about the validity of stress testing in general.
Ron Waksman: You know, I'm confused because we have a lot of studies that shows that complete revascularizations are better than partial revascularizations. So, you ask yourself what is complete revascularization? Should you basically obliterate any ischemia and why are they better than partial revascularization?
I'm not sure that the answer is really clear and maybe it's to blame the type of the tests that we are doing. Maybe we should not rely so much on thallium stress tests, or any other stress test, but that's the best that we have right now. I don't know if you have any comments on that part, but it is very confusing.
I just want to go to one other imaging test that some claim actually was the winner of the study, which was the CT. Dr. Bangalore, is there anything about the CT that resulted in a change of practice because of ISCHEMIA? Is that something that we should now screen all our patients and run CT and find out that we should or should not treat them?
Sripal Bangalore: I think this debate will continue, whether it should be imaging first or CT. I think even among the leadership here, we have different takes. I do believe from the totality of data that anatomy does provide important prognostic information. I prefer some form of an atomic evaluation, so I'm kind of slightly biased toward saying CT.
That being said, if you just step back and say I have a patient in front of me and the patient says he is having symptoms, the question is, do you want to try medicines or do you want to try an invasive strategy. If they choose an invasive strategy, forget CT, forget stress testing. You do the cath and revasc. If they say they want to try a conservative strategy, then I think the ISCHEMIA trial pathway starts because you just want to make sure, if it is a high-risk patient, make sure that there is no left main disease.
One of the things that I commonly tell my fellows is if you're saying that you're going to treat patients medically without knowing the anatomy, you’d rather be sure they’re epicardial CAD and that is what you are treating. There have been many patients who have angina or atypical symptoms; you keep throwing medications on them and then a year later you do an angiogram and they have nonobstructed CAD. Something different and you've been throwing all of these medications at them thinking that it's obstructive symptoms. So, the way I see imaging is really dependent on what pathway the patient wants to pursue.
Issam Moussa: If I had a patient with unequivocal stage three angina, and I'm convinced there’s CAD, and the patient is on whatever medical therapy or not, certainly cath is the way to go.
I think, in anybody else with an intermediate score or lower, as you know there's two randomized trials of CTA versus stress testing and there's more trials coming, I think increasingly I would go to CTA if there's any doubt and the diagnosis of CAD is not clear. If it's clear with symptoms you need to define the anatomy and make a decision later. And I think that would be a way to go.
If really it's not clear that the patient has higher-risk CAD I think CTA, with or without FFR is a different discussion that we will come to later, I feel strongly that that should begin to adopt that strategy because there's reasonable evidence that ischemia by stress testing, as determined clearly by the ISCHEMIA trial and other studies, that it’s not a reliable indicator of action.
Ron Waksman: Clinical judgment is always important. I think one of the issues with all those studies is that it's not one size fits all because we're dealing with different patients and different skills from the practitioner.
I mean, some can smell out those with severe CAD just by looking at the patient, whereas others need to ask more questions, and some need to send them to CPA. But if you're fully convinced that this patient is going to need an intervention, why to do the two steps? You can take them to the cath lab, you can do a radial approach, it's saving contrast, etc.
From an interventionalist perspective, I still like to go directly to the cath lab. But if I do have conviction, and the story is very compelling that this patient is symptomatic and has angina, and it's a very classical angina, that patient would not fit into the ISCHEMIA trial. The main question is all of those patients that you don't know for sure.
Let’s move to another topic, which was discussed a lot at the AHA meeting and continues to be discusses since then—the conversion of the curves.
The curves are now turning. If you start to do a landmark analysis and that's going to be the trend, then that does actually support intervention more than medical therapy.
Is that a fair interpretation, Dr. Bangalore?
Sripal Bangalore: That is a fair interpretation but that's going by an assumption that upfront risk of procedural MI is . . . that you can just throw it away. I'm sure people who are noninterventionalists would feel strongly against that.
I think this is one of those things that will continue to be debated. What do we do with that upfront risk of procedural MI? Is it prognostically important? Of course, we'll get more insight into it.
But, you're absolutely right. I mean, if you take away the upfront risk, what we are seeing is there is divergence and there is benefit. But, I don't think everybody would agree that you can take that upfront risk away and say you're going to close your eyes and not look at that.
Ron Waksman: It will continue to be a debate. I just . . . trying to trade a biomarker with a spontaneous MI, I don't think they are the same.
Spontaneous MI would require an intervention. Hospitalization usually will have some impact on your functionality of the heart, maybe also on the periprocedural MI. But if that's the case, usually we know in our patients if they have a residual risk for a thrombotic event or ischemic event, it's going to pan out at some point.
In the ISCHEMIA trial, if you really put so much on the periprocedural MI, you would expect that actually those patients who had periprocedural MI will do worse over time. It's not going to flip at anytime point; so maybe they did worse, we just don't know because we never heard what happened to those specific patients who had periprocedural MI. It would be nice to differentiate those from those who never had periprocedural MI, did they really do worse over time.
Sripal Bangalore: You will see the results very soon.
Issam Moussa: I think this as important as SYNTAX and other trials. But I think the problem with a lot of these things is that the number of patients with periprocedural MI is really too small. When you put a curve and you compare these patients to the other population you get very artificial curve. So, they need that at some point maybe a large meta-analysis or something. But the problem is there's different definitions. So, this question will remain open for a while and it's an important I think point to keep in mind.
Ron Waksman: What is still missing from the ISCHEMIA trial? What questions need to be addressed and is there still equipoise in your opinions that needs further research?
Sripal Bangalore: I think that's a great question. If you go back to the guidelines, and of course the revascularization guidelines are pretty old and we are in the process of coming up with new guidelines, but 10 years ago the guidelines would say there are high-risk group of patients, those with triple-vessel disease or two-vessel disease and a proximal LAD and for those the guidelines recommend CABG to improve survival with a class I indication. So, the question, potentially, is if you have high-risk and atomic subsets would there be a mortality benefit?
We didn't see that in ISCHEMIA, but of course we were not powered to look at this individual subset of patients.
But what I want to also ask is if that really is an important question. Are we still trying to look for a holy grail of survival benefit in stable ischemic heart disease by revascularization?
I think all of us, at least on this panel, feel that we don't trust that there is going to be mortality benefit that is more benefit beyond that. So, I am not convinced that that's actually as important of a question.
Issam Moussa: I agree very strongly, especially based on the premise that those guidelines are based on registries that are 30 years old. I really think it is time to incorporate the contemporary evidence, including STITCH, BARI 2D, and ISCHEMIA to show that patients with a normal ejection fraction and non left main disease at the mortality in point, especially at less than 5 years, certainly there is no evidence to support it.
The future would be, as you said, going back to the stress testing. Are we going to continue to do this, despite the evidence that we have that's not reliable. Because in whatever trial we do, if another trial uses stress testing, I think that would be a problem because I think people should not ignore the fact that stress testing is clearly unreliable to determine prognosis as it was taught.
And then the question becomes, if you don't use stress testing what do you use; CT, CT FFR, and what are the endpoints?
I mean, periprocedural MI, you could capture it, maybe some would argue, as a safety standpoint that you could follow it up if we're going to look at long-term outcome. Now, these are all controversial, but I would say somewhat novel, somewhat new that would move the field beyond the last 20-30 years where we considered restenosis equal to mortality. So, I really think it is important to incorporate these in your thoughts to move on to a different level of investigation.
Ron Waksman: And the problem is that a lot of individuals take the position that it's either medical therapy or prevention. And while more people appreciate the progress in medical therapy with lowering lipid production without a platelet therapy, etc etc, anti-inflammatory, it is accepted but, I think, still very hard to accept for many individuals who are non interventionalists that PCI has some value.
We have to think more about the combination therapy rather than two separate therapies. It's not either or, it is the combination therapy that potentially makes the difference for our patients.
I would like to move to the last segment, which is what’s next? What can you share with us about the duration of the follow up that's going to be done now that you got the grant for the continuation of this study? What should we expect to see more of from the current ISCHEMIA trial?
Sripal Bangalore: Of course, there's plenty of data to answer many different questions about procedural MI, spontaneous MI, does completeness of revasc matter, what about PCI CABG, and all of those things that we were actively working on. But, the longer term follow-up is mainly designed to answer the mortality question. Yes, you have an upfront risk of procedural MI, but does it impact long term mortality? Sure, you have the benefit of spontaneous MI, but that does that benefit of spontaneous MI reduction translate to long-term survival benefit?
In that sense, I think we will be able to understand better. But, one of the things that we have to recognize is that the ISCHEMIA sample size is much more than COURAGE and BARI 2D combined.
I think Dr. Moussa absolutely was spot on to say, even with that the mortality was not very high. It just goes to show that I think we've come a long way in treating our patients.
Ron Waksman: Dr. Moussa?
Issam Moussa: I think in ISCHEMIA obviously, the long-term follow-up is key, up to 10 years at least could define what is the prognosis of this population. I think, it's the future that matters in terms of what studies we design and what primary endpoints we choose. But I would hope that we'll see more vigorous discussion of not continuing to do the same thing that we've done for 30 years in terms of testing, primary endpoints, and duration of follow-up to make the trials that we do more impactful.
Ron Waksman: Yeah, I think it's all depends on the patient. I'm not that thrilled about the mortality endpoint in general for these kinds of studies because I think that if I have a patient today who has angina or who has ischemia, they want to take care of the problem today. They’re going to worry about being alive or dead with therapy 10 or 15 years from now. We've seen that with other technology that we have, even with TAVR, they want to get their relief now. So as long as you don't tell them this therapy is going to kill them, they say, “okay, I'm going to go with the therapy. I don't care that he doesn't improve my survival rate.” And I'm not sure we can, we have so many confounders that enter into the follow-up. So, at the end of the day, guess what—we’re all going to die. So, if you take a 65-year-old patient and follow them for 15 years, they’re going to be 80. Hopefully, with medical therapy they will live longer. But at some point it's becoming moot. I don't think we have to put so much on mortality.
So, I’d like to congratulate you and the NIH for extending the follow-up because I think there was so much put into that study that it would it be a loss if we wouldn't do that effort to understand more because we may find so many other things about progression of disease, impact of duration, compliance, further intervention down the road.
There's tons of information and it's nothing that can substitute for good research. In that respect this is very important information. How to change the guidelines based on that.? I'm not sure what I can advise you other than to look at each patient individually because at the end of the day there is no one size fits all. And we learned that, I mean there is such a heterogeneity of the patient, the therapies, the lesions, the degree of the ischemia, which we cannot settle.
So, I would like to give both of you the last word. This has been a very interesting, in-depth discussion. It looks like most of the stuff we now agree, we are much more mature and after 1 year you leave all the things and you kind of converge with your ideas and the conclusions.
Issam Moussa: The ISCHEMIA trial is a great trial and that did provide new information. I think the new information specifically is that there is accumulating evidence that stress testing is unreliable to prognosticate patients, that any other anatomy is important, that if you have normal ejection fraction and limited symptoms a medical therapy would work as well as PCI.
But if you do have symptoms, clearly improve your prognosis in the short term and the long term is yet to be determined rather than say there's no difference. That would be my interpretation.
Sripal Bangalore: I think ISCHEMIA, in being the largest treatment strategy trial, has given us more ammunition as to how best to discuss with the patient so that we can come up with a shared decision-making to inform them on these are what each strategy does, let us know what you want. I think this discussion was really important because this is not a trial of PCI or revascularization versus medical therapy. This was a strategy trial so if the patient says they want to start medical therapy, that's fine. They can try it for some time. If their symptoms are not getting better, it's a strategy of changing treatment as you see what is happening to the patient.
It really emphasizes the fact that it should be a patient-centric approach: you discuss the pros and cons and you do what is best for the patient.