CRT Clinical Trials: What We Know And Why
William Abraham, MD
Ohio State University
Columbus, OH
But the good news is you all know the data, and one of the things that Cecelia and the meeting organizers asked me in particular to talk about is how we measure response to CRT. So we will talk a little bit about what we know, who CRT is currently indicated for, and we are going to look at possible definitions of nonresponse or inadequate response, and I believe that is one of the objectives is to move to some consensus in terms of that definition.
So let’s start with what we know. We know that cardiac resynchronization therapy is an effective form of therapy for patients with New York Heart Association class III and class IV heart failure. The weight of evidence supporting current practice and our current guideline recommendations is summarized on this slide. More than 4000 patients have been evaluated in landmark randomized controlled trials; in fact, the leadership for most of those trials is with us in this room. These trials have demonstrated consistent improvements in quality of life, functional status, and exercise capacity. They have provided strong evidence for reverse remodeling, the ventricle gets smaller and stronger, and functional mitral regurgitation is diminished, and there are reductions in both heart failure and all cause morbidity and mortality.
So I will just show you two or three slides to provide some data to support what was summarized on the prior slide. This slide depicts the effects of CRT on quality of life and functional status from four of our landmark trials, which includes the MUSTIC trial, Drs. Daubert, Linde, and others in this room, MIRACLE, CONTAK-CD, and MIRACLE ICD, and you can see in terms of quality of life and New York Heart Association functional class ranking, very consistent improvements.
Similarly, these trials have demonstrated consistent improvement on exercise capacity, whether that be measured by the six-minute hall walk distance, or the change in peak oxygen consumption. We have data from MIRACLE and now from many, many other studies that demonstrate the effects of CRT to produce reverse remodeling, improvements in the ejection fraction, reductions in mitral regurgitation, reductions in LV volumes including both LV and diastolic, and then systolic volumes and dimensions.
And then finally, from the very large CARE-HF study from our colleagues in Europe, the first demonstration that CRT alone without a defibrillator prolongs life, prolongs survival, morbidity and mortality is markedly reduced in patients randomized to CRT plus optimal medical therapy versus CRT alone.
So it is this, these types of data, that have led to the following guideline recommendations for heart failure. Here are the patients who currently should receive CRT on the basis of those studies. A class I indication supported by multiple randomized controlled trials, so a level of evidence graded as A, that patients with LV ejection fractions less than or equal to 35% in sinus rhythm and with New York Heart Association functional class III or ambulatory class IV symptoms, despite recommended optimal medical therapy and who have cardiac dyssynchrony currently defined as a QRS duration greater than 120 msec should receive CRT with or without an ICD, unless contraindicated. So that covers the patients with sinus rhythm.
There is now also a class IIA recommendation; the level of evidence here is graded as B because it comes from either small randomized controlled trials or from observational studies supporting the use of cardiac resynchronization therapy in patients with atrial fibrillation. And I won’t read this guideline recommendation to you because it is identical to the one shown on the previous slide with the exception of specifying atrial fibrillation rather than normal sinus rhythm.
And then there is additionally a class IIA recommendation in the guidelines for patients who have a frequent dependence on, or an indication for ventricular pacing. Here, the level of evidence is C; this really means that the recommendation is based on consensus and some of the randomized controlled trials to actually support this recommendation are still ongoing.
Now, in addition, we know that we are on the verge of an expansion of the clinical indication for cardiac resynchronization therapy into a group of patients with either asymptomatic or only mildly symptomatic heart failure, and this slide summarizes the weight of evidence supporting an indication for those patients. More than 3,000 patients evaluated in landmark randomized controlled trials, most of those patients to date studied in either the REVERSE or the MADIT-CRT trial. These studies have also provided strong evidence for reverse remodeling and demonstrated improvements in functional status and outcomes as measured by either a clinical composite endpoint or the combined endpoint of heart failure or cardiovascular morbidity and all cause mortality.
Just very briefly because you already know the data, we will take a look at one slide from REVERSE and a couple of slides from MADIT-CRT. Here are the REVERSE data at 24 months from Drs. Linde, Daubert and others that demonstrate a statistically significant improvement in the percent of patients who are worsened over 24 months randomized to CRT versus no CRT. So you see that a greater percentage of patients are improved or unchanged, a lesser percent of patients are worsened, compared to those patients who are randomized to no cardiac resynchronization therapy.
The MADIT-CRT trial, which randomized just over 1,800 patients, confirmed the effects of resynchronization therapy in this class I and class II population, demonstrating a reduction in heart failure–free, or an improvement in heart failure–free survival. Overall, there was about a 30% improvement in survival in these patients randomized to cardiac resynchronization therapy.
I am certain that you are all aware of the subgroup analysis. This was part of the primary paper, the primary report, of the MADIT-CRT trial that demonstrated a couple of curious findings. One was that the benefits of CRT in this population appear to be most evident, most prominent in the class II patients, a little bit tougher to show that the confidence intervals are broad for class I and cross the line of unity, and there appear to be no beneficial effect in patients with QRS durations less than 150 msec. It appeared that the entire benefit was seen in patients with QRS durations greater than 150 msec.
And while the US guidelines have not yet caught up with these data, the European guidelines have, and so, based on the ESC heart failure guideline, specifically, the update on the use of devices for heart failure just released a couple of months ago, Ken Dickstein and others, there is now a class I indication with a level of evidence graded to be A, supporting the expansion of the indication for CRT into mildly symptomatic heart failure patients.
And you will see that, and this is part of what we have wrestled with and may become part of our discussion today, based on the MADIT-CRT and REVERSE data, which patients should actually get CRT? Well, the European Guideline Committee decided that this should be recommended to patients who have a reduced ejection fraction, less than or equal to 35%. Interesting cutoff; it agrees with the current guideline recommendation for class III and class IV patients, but remember that MADIT-CRT enrolled patients with LVEFs less than 30%, and REVERSE less than 40%. So this represents a compromise of the two, it falls right in-between, and it provides consistency with our current guideline for class III and class IV heart failure.
They also, and I am not sure where the II dropped out, but that should say New York Heart Association functional class II, so the European Guideline Writing Committee did not extend this indication to class I patients on the basis of currently available clinical trials. So they specified New York Heart Associational functional class II patients, and also limited this to patients with QRS durations greater than or equal to 150 msec. So that is the consensus of the European Guideline Writing Group or Committee, on the basis of currently available data.
Well, now let’s talk a little bit about how we define response to CRT in our patients. If we look at clinical trials, while the clinical trials really run the gamut from assessing quality of life to New York Heart Association functional class ranking, exercise capacity including six-minute hall walk and metabolic exercise testing, cardiac remodeling, morbidity and mortality, when we look at studies that have emerged either from single centers or large multicenter, and, in some instances, multinational studies, that have tried to evaluate this issue of CRT responsiveness, the most common endpoints used for assessing response in CRT clinical trials have been the heart failure clinical composite response, which was first described, the methodology published in the Journal of Cardiac Failure in 2001 from Milton Packer, and remodeling, particularly changes in left ventricular end-systolic volume or LVESV index.
For those of you who are not fully familiar with the clinical composite response, its definition is shown on this slide. It incorporates measures of both patient well-being or functional status, the New York Heart Association functional class ranking and the patient global assessment, as well as heart failure disease–specific morbidity, heart failure hospitalization, and all-cause mortality.
So patients are judged to be improved if they have an improvement in New York Heart Association functional class ranking, or an improvement in patient global assessment; they are worse if they die, if they experience worsening heart failure leading to hospitalization or permanent withdrawal of the therapy, or if they experience worsening of New York Heart Association functional class ranking or patient global assessment; and if they don’t meet either of those two definitions, they are judged to be unchanged.
So let’s look at how this clinical composite endpoint performs in clinical trials of cardiac resynchronization therapy. And in fact, this was the first trial to prospectively evaluate the clinical composite response to cardiac resynchronization therapy, and it has been widely cited as forming the basis for the assertion that about a third of patients don’t respond to cardiac resynchronization therapy. The data come from the MIRACLE trial, and this demonstrated in the CRT arm that about 65% of patients were improved, according to this clinical composite response definition; about 18% of patients were unchanged; and 17% were worsened.
So that was our first experience with this clinical composite response in heart failure. But remarkably, in now many studies that have reevaluated this over the past decade, the results have been strikingly similar.
So let me just show you one example. Here are the data from the PROSPECT trial, which evaluated the clinical composite response as one of its primary endpoints; the second was LV end-systolic volume. And you will see that the data, shown on the left-hand figure, are almost identical to those seen in the MIRACLE trial. And we have seen this at least in the class III and IV heart failure population, we have seen this emerge time and time and time again. So by this definition, at six months follow-up, you get about 65%-70% of patients demonstrating improvement with cardiac resynchronization therapy.
So that is what we do, or that is what our experience is, in clinical trials, and those are two definitions of response to CRT which have been widely used in investigations.
What about in the real world? And I am just going to conclude with a brief summary of this survey that was done of participating sites in the FREEDOM trial. The FREEDOM trial was a very large multinational randomized study which evaluated optimization strategies in CRT. But at the onset or outset of the FREEDOM study, a survey was conducted at the participating centers, essentially to ask them, how do you define responsiveness to CRT?
So there were 117 centers from 16 countries which participated in, or provided answers to, the survey. This represented nearly 80% of the 150 participating freedom centers. So we got a really good response rate to these surveys. And again, questions in the survey focused on the period of time when CRT responds was assessed. We wanted to know when did people sort of make that judgment, was the patient improved or unchanged or worsened? We also wanted to know what criteria they used to form that judgment. Were they using echocardiograms or New York Heart Association class or some subjective impression of how the patient was doing, their symptoms and so on and so forth?
And we also asked them the percent of patients responding to CRT and then ultimately when and how CRT is optimized. And I am going to tell you the answer to the third bullet point because I forgot to include the slide in the presentation. I meant to because it turned out that no matter how, no matter what definition that the center was using, whether it be a single criteria or a multiple criteria, whether it be subjective or an echo, no matter how the center defined responsiveness to CRT, when you asked them what percentage of your patients respond to CRT, regardless of the criteria used, the number was 70%, turned out to look just like what we see in the clinical trials. So our real world clinical experience or impression of responsiveness to CRT is virtually identical to that which has been seen in clinical trials.
Well, let’s take a look at what people are using in the real world as we enter into our own discussion of how to define responsiveness to CRT, when and how do people, clinicians assess the response to CRT? So in answer to the period of time when CRT responds is assessed, while there may have been device follow-up, and certainly was in most patients device follow-up which occurred before this period, the first follow-up which involved a clinical assessment of how the patient was doing, generally occurred between two and six months after CRT implant.