And now I think we decided that the order goes on with Doug Pecker, who's going to tell us what Cabana is about when we put all the evidence together. Doug Pecker, we're delighted to hear your talk. Thank you. So I appreciate this opportunity. And I'm going to give you everything about Cabana. And we're going to do that in 10 minutes. My name is really John Sapp, because I'm second in the program. But some of my friends call me Doug. And that's OK. But if we put all the evidence together, we first have to put the evidence of disclosures. And those are they. None of them have anything to do with Cabana. But the purpose of Cabana was really to look at ablation versus state-of-the-art drug therapy for patients with new, unfit, or undertreated atrial fibrillation. The primary endpoint was all-cause mortality. Disabling stroke, serious bleeding, and cardiac arrest, and major secondary endpoints were all-cause mortality. Death or cardiovascular hospitalization. Halfway through, the primary and the secondary endpoints were flipped. And we did that at the recommendation of DSMB. And I think it's important that people know that. We did that because a slow enrollment rate was really the major factor. And the event rates were lessened. And so that, in any review of Cabana, needs to be in there. Now, if you look at the consort diagram, there are 2,200 patients. There are basically 1,100 that went to each one of the arms. And 90% of patients were ablated. 215 underwent repeat ablation. 102 were not ablated for a variety of different reasons. And there were 100 or 1,000 of them that completed the process over the course of about 48 months. Drug therapy is almost identical, except for 301 crossed over from drug therapy to ablative therapy because of inefficacy of the drug, largely in some side effects. Now, if you look at some of the demographics, then they're pretty well aligned. There are not much in the way of differences there that we need to review. About 43% of them had paroxysmal atrial fibrillation. And 57% had persistent or longstanding persistent atrial fibrillation. And they're all fairly highly symptomatic. Now, if you look at the outcomes here, and we're comparing ablation versus drug therapy, and let's just look at the composite, it's 8% versus 9.2%. And that's not different. If you look at death, it's 5.2% versus 6.1%. You get down in here to death or hospitalization, then there is a significant difference. 52% versus 58% in all-cause mortality was not different. And that's what we showed in the very first KM curve of the primary endpoint that you can see, that there was not a difference between ablation and drug therapy. You know, it was reduced by 14%, but that simply wasn't enough to make a substantial difference. If you look at the primary endpoints and the subgroup analyses, these mostly huddle right around null. There are some interesting issues that we'll bring up in a second about patients who are under the age of 65, patients who are minorities, and patients who had heart failure class two or three that tended to lean towards the side of ablation being better in intention to treat. And all the information I've given you is intention to treat. Now, one of the problems with intention to treat is it loses precision. If you have a strategic trial, you're looking at interventional approaches. And if you're looking at places where people don't stay in the arm, to which they were randomized, there's 300 that moved in one direction, there were 100 that moved in the other direction, so your difference in looking at patients is really 900 versus 700. Now, how much of a difference is that? Well, it's enough of a difference that we had pre-specified looking at treatment received and per-protocol analyses. And if you look at these, the primary outcome here was 7% primary outcome event rates in ablation, and then the drug therapy group was about 11%. You see where the 95% interval is with the estimate of .67, and that was quite significant. If you look at all-cause mortality, going all in the same direction, and death and hospitalization really was what it was before. If you look at per-protocol, then there is a difference between drug therapy and ablation. Patients who were ablated had fewer event rates than patients who were treated with drugs. And the reduction here is about 27%. Again, a point estimate of about .73. I think you learn more about these numbers right here about a trial and a KM curve. If you look at the endpoint analyses in per-protocol, then you start to see this age issue becoming more of an issue. You find out that minority may be more of an issue, and you find out that the heart failure may be more of an issue. That's some of the things that we've unpacked for this meeting. Now, AF recurrence has demonstrated what you would expect it would, whether it's intention to treat or other analyses. Ablation did substantially better in eliminating first recurrences of atrial fibrillation than did those that were treated with drug therapy. Here, the reduction's about 47%, and you can see in the split here that it's substantially different. Is there something in Cabana that's wildly positive? Yes. Are there things that are indeterminate? Well, those primary endpoints that I showed you first. And in the words of Clyde Yancy yesterday, then we've rescued Cabana from a negative trial. It's not a negative trial. There are parts of the trial that are indeterminate, and most of it is quite positive, and this is one of the areas where it is. Now, this is work that was done by Gene Poole. You look at paroxysmals, or you look at persistence, and you look at the burden on Holter analysis, and you can see that both groups had a lot of misery at the beginning. Patients who were ablated did significantly better than did those who were treated with drug therapy, whether it was paroxysmal or persistent, but as you'd expect, there are more persistence that had recurrences or had a higher burden as we tracked along the course of 60 months. Dan Mark looked at quality of life, and he looked at it in a couple of different ways. He looked at effect overall scores, and these are unadjusted, and if you look at this, you have to keep track of which way these are tracking, but at the end of the day, ablation patients did better in terms of improvement in quality of life than did those who were treated with drugs, and so our green turns to a little bit of red there. If you look at the MAFSI scores, and this is intention to treat analysis, and what he did is he looked at adjusted mean differences of ablation minus drug therapy, then if you're looking at three months versus looking at 60 months, all the way along the line, then ablation patients fare better in this measure of quality of life. Now, I think everybody in the room can probably name 45 other inventories that can be used for this. If you look at age, and this is some work that Tris Bonson did, then there's a difference in patients under the age of 65 versus 65 to 75 and those that are greater than 75. If you look at the younger age groups, and you see the hazard ratios, and you see the more important bits of information along the way, and you see the overall look at all three groups, but ablation patients did better with lower event rates than, and this is in the primary endpoint than did those who were treated with drug. Not a lot of difference in the middlings, and as I look out here, then I see a lot of people who are, well, probably not 65 to 75 years of age, but there's a problem with the 75 and older. So the patients who were ablated had higher event rates than did those that were treated with drug therapy. Now, it's important here to make a point. It doesn't mean you cannot ablate a patient for symptoms and for quality of life because they still did very, very well. How do I know that? Because that's the next slide, and if you look at the freedom from recurrence, then you can see whether you're 65 or 65 to 75 or 75, then there's a fairly substantial difference. However, if your mother-in-law is 75 years of age or older, you cannot look to Cabana to solve the problems of having her live longer. The same thing must be said about fathers-in-law who are equally obnoxious, and I think that that's another point that needs to be brought out of Cabana. Now, for heart failure, we put this together, and this is heart failure. We're looking at the primary endpoint of mortality, disabling stroke, serious bleeding, cardiac arrest, and this is intention to treat an impact of heart failure, and the thing that you see is that drug therapy patients did not do as well as did ablation patients. You can see what the hazard or the estimate is 0.64, and you can see 0.99, so we've cleared everything out from null, and you can see what the p-value is. I think p-values in a plot like this are right up there next to worthless. I think that the 595, particularly the 95, I think is very important, and the hazard ratio tells us a lot. If you look at mortality, the mortality outcomes were basically the same, so for patients with heart failure, there weren't the areas of confusion seen with all comers, and so the question at hand is, are we going to be using this to a greater extent in patients with heart failure? Are we gonna do CRTs in all of them? So this is yet another positive finding of Cabana that I think will be coming out. Three of the papers are already out, and there's about five more that are underway, and so if you look at the cumulative risk of AF here with intention to treat, again, the patients had less in the way of AF recurrence than they did if they were treated with drugs. This is kind of the forest plot, and it basically says everything that I said. There's one more thing that we need to talk about. Miriam Rettman is sitting right down here, and she did a paper that looked at reverse remodeling, so these are patients where we did scans before and at about six months afterwards, and the thing that we found was that in terms of left atrial volume index, and the same thing was true, certainly with right-sided pulmonary veins and then left-sided pulmonary veins, is that the absolute change was greater in ablation than it was in drug therapy, and this makes a comparison with that at the time of the first recurrence. Now, these are all going to be on the positive side of null, but the thing that we found is that those patients who had the greatest change in left atrial volume index, which was ablation patients, they did significantly better in terms of recurrence. We already know about the adverse events in cabana. They were low, and I will simply close by saying what I said before. I think the things that we can say is ablation did not produce a significant reduction in primary endpoint null cause mortality if you look in tension to treat, and I think that that was harmed by crossovers and by some lower-than-expected event rates. However, this is not an observational trial. It's a trial that looks at strategy, and it looks at interventions where sometimes you have troubles keeping patients in their arm of randomization, and so it makes a difference, and it can drop the precision of your estimates. The other thing is if you look at per protocol, you still have the benefit of randomization, and it's incredibly important that we have that benefit of randomization, but as we discussed in great detail yesterday in a prolonged seminar on this, there are issues that come up with intention to treat as well. Ablation significantly reduced mortality rates in cardiovascular hospitalization by 17%, 47% reduction in recurrence of atrial fibrillation, 39% reduction in primary endpoint, and 40% mortality if you look at treatment received, and ablation was acceptably low risk, and if the heart size changed, the outcome of ablation was better, and if you have heart failure, then there's a substantial difference, and age makes a difference, and there are about seven other things that we're in the process of unpacking, and I'm just not gonna tell you because these guys have told me my time is up, so thank you very much. Thank you, Doc, great presentation, and we are anxiously looking forward to the next things you will be unpacking from Cabana. It's gonna be great, so are there any questions from the audience? So the thing that struck me most was obviously heart failure, an important point that you made, and of course the fact that ablation did go well here, and not very much of complications that were seen. When we look at this as a sub-analysis for the heart failure group of patients, can you shed some light in terms of what the statistical significance of the power of this sub-group analysis means, and do you think it's well in line with all the data that we've seen from Castle Evidence supporting that, and can we draw conclusions from these sub-studies in Cabana? I think that whether it's Cabana or whether it's Castle, there's some problems in looking at power, and just exactly what that looks like and what it does in a clinical trial. We think that the power is as good in the heart failure patients as it is in the rest of the trial, but there's a fewer, smaller number of patients. Some of the issues that they had in Castle we did not have in Cabana, so we think that the information from Cabana certainly supports what Castle said, and we think that it actually supports it to a greater degree. We need to unpack the rest of it, though, and so I think probably it's not fair to say more than what I can say, but what I think is it's gonna change guidelines. One other question in terms of the, please. Thank you. Dr. Parker, thank you for your presentation. I would like to ask two things, two simple things. First one, you described in 2009 in the editorial the Cabana trial. If the all-cause mortality was analyzed like that, and the crossover was not permitted, the study was a positive study or not before the changes that you introduced. The second thing, you answer 25.7% of crossover in direct treatment in those patients in the supplement three of your publication as only 2% of death and disabled stroke against 5% of the ablation group that have done ablation and 9% of the direct treatment that have not done crossover. I would like to have two comments on this. Okay. I think you must have been one of the people that wrote a letter to the editor, to JAMA that we're gonna answer in greater detail, but I think that in 2009, you know, we had total mortality as the first, and then in 2013, we switched it. We did not have any knowledge whatsoever of outcomes at that point. We didn't know what the outcomes were going to be, so we made that decision independently. And your second question? Second question is the very good results of your crossover group, because half of the deaths is in stroke than the group of ablation, and a quarter of the deaths and strokes of the patients that have direct treatment without crossover. Yeah, so... This means an explanation. All of that information is already published, so you can get that in JAMA in the supplement, and I'm not quite sure that we would agree with your assertions, but we do agree with what we put into the subgroup. It's difficult to analyze your study, and I finalize immediately, without knowing the patients that are lost for follow-up in this study. I think... Because if we know that, we can verify if, in fact, there are a significant difference due to the fact only, if you have a very, very good result in the crossover group. I think you've just given us a lesson in Statistics 101. It's difficult to deal with patients who drop out of a study. And I agree with you. If we had information from those, it could either change the outcomes massively or not at all, and unfortunately, we don't know that. I mean, it's a good point, and it'd be something that I'd be delighted to know. The problem we had is whether we were going to break all ethics and try to get into these people's houses and call them and find out what happened to them. We won't do that. Okay. For the sake of time, we have to move on. Okay, thank you. Thank you very much. Thank you. And the next...

Cabana trial

ablation versus drug therapy

atrial fibrillation

all-cause mortality

heart failure