Dr Mandrola discusses recent trail data with Dr. Thomas Nero including the results from Artesia, NOAH Afnet 6, Frail AF, Azalea and Oceanic-AF. They also discuss Left Atrial Appendage Occlusion and utility of wearable devices. Finally, a shout out to Dr. Andre Foy and his work on multi-morbidity and its role in clinical trial interpretation.
Dr Mandrola discusses recent trail data with Dr. Thomas Nero including the results from Artesia, NOAH Afnet 6, Frail AF, Azalea and Oceanic-AF. They also discuss Left Atrial Appendage Occlusion and utility of wearable devices. Finally, a shout out to Dr. Andre Foy and his work on multi-morbidity and its role in clinical trial interpretation.
Dr John Mandrola on Atrial Fibrillation
Tom: . Welcome John.
John: Thank you. Thanks for having me.
Tom: I've been listening to your podcast , every week. And, it's sort of like, I already know you because I've heard your voice so often.
I think that a lot of people who are listening to this podcast have probably already heard you. , we know that you're a medical conservative and I think that actually does you a little bit , of injustice, , because a medical conservative, it's almost as if your curmudgeon and it's not that at all, it's in the great continuation , of evidence based medicine.
, for those who haven't already listened to John's podcast , this week in cardiology on Medscape, and he also does another podcast on sensible medicine . And then of course your sub stack, uh, which I am a paid subscriber
John: Oh, thank you.
Tom: love, love reading every week.
So, , if you haven't already, , been, , introduced to John's work, you should, , dive into that. , so Dr. Mandrola, you are one of the national leaders in, in the way that we're discussing clinical trials. And. Pushing back on just accepting the top of the line results.
, when I initially, , started discussions with you about the podcast, it was, it was mostly on the heels of. The dual information that we were getting from Azalea and then the oceanic trials. So I figured that we'd start there and then a leading into a discussion about atrial fibrillation then we'll, , let it go from there.
So, uh, the Azalea trial, it had a decreased risk of bleeding, uh, by about 70%, and, it did not demonstrate a significant increased risk of stroke.
So, , now that we're three, four weeks out after Azalea's results, have you, , changed your, opinion on how to view that trial?
John: , the Factor XI inhibitors, you know, I've been going to meetings, and I've been seeing the excitement, and , I'm not a hematologist, but, , it seems to have these properties of sort of the perfect anticoagulant because it can decrease bleeding, but also decreased thrombosis, .
, so the Azalea trial did show that bleeding was dramatically reduced, and I think this led to termination of the trial, so there's a lot of excitement. Avalacemabs are injectable, , and then, , we, , I think you were as well, but our group, , which is a private practice group in Louisville, Kentucky, we were actually, , recruiting for Oceanic, the Ascenduxian, which is the oral Factor XI.
And it was, it was actually quite easy to, , recruit because anybody who had trouble paying for their Apixaban, we could just propose a trial. And so , we were recruiting and having no trouble. And then boom, like. Like, I think it was at 11 months or one year into the trial, , the company announced that it was, , being stopped for, , increased harm, which you assume are thrombotic events.
So that is a real, , let's call it a headwind, right? Uh, but now I'm hearing voices about, well, maybe the dose wasn't right. And there are other factor 11 inhibitors. So I think that the opposing team has scored a few runs on the factor 11 inhibitors, but the game probably isn't over.
But, , it's not looking as great , as it was. And it also speaks to the fact that , Apixaban is pretty good. So, we'll have to, we'll have to see, um, it really was going to play into the, the left atrial appendage closure discussion that we were going to have. Right?
Because. If factor 11 inhibitors were a better anticoagulant, a safer anticoagulant, that might have been a headwind for the appendage closure folks. But , it looks like we need to really wait on more data, but maybe not as exciting as it was a year ago.
Tom: , we're involved. In the Oceanic as well as the Librexia AF trials. And the Librexia AF , is, , still ongoing, uh, which is positive. , and there is some question about whether or not the factor 11a inhibitors are the same as the factor 11 inhibitors.
So, uh, for azalea, it's a, it's a different mechanism. And I think that, um, you know, obviously when you look at the genetic questions behind this, that when you have patients who have genetic factor 11 deficiency, they don't have thrombosis. And so that's really where this all came from. And it's quite an interesting sort of concept.
That we can get all this great information from, uh, the UK biobank, et cetera. And now that they've now released the entire genetic, uh, data behind the UK biobank, I think we're going to get even more information out of it, but that translating from the genetics and what we think is going on into a therapeutic that actually makes a difference clinically is a very, very large step.
John: It's really hard. I, it's really sobering and I. I'm just a clinician, so I, I just use the data. I don't create the data, but, um, uh, it's sobering how hard it is to bring a drug to market, a new drug or new mechanism, mechanistic, drug to market.
Tom: and, and we'll see, , they're continuing on the oceanic, uh, CVA trial, uh, that's ongoing and it did not stop. They did have not seen harm yet, although they're recruiting a little bit more slowly.
Um, but you know, the other possibilities that what we're going to see is different effects by. Different diagnostic groups and , you were discussing, uh, I want to say on sensible medicine, you know, it's always nice when you have. A robust result that transcends a specific diagnosis, right?
One thrombosis risk, you know, one thrombosis diagnosis, you know, dbt leads to an improvement in a fib, uh, thrombosis leads to an improvement in ACS thrombosis, et cetera, but maybe it's not going to be quite that simple and that we're going to have to have more nuanced therapies for these.
John: Yeah. Uh, exactly. Uh, it's, it's early. . It's exciting. Uh, there's still more to be said and who knows? Maybe, you know, a, a year or two from now, it's, it is, it's hard to be too pessimistic, but it, it wasn't like a hit it out of the park, the first try thing.
Tom: , I agree with you with it, with the discussion that I had with, uh, Dr. Gibson on the factor 11 inhibitors. , we were musing about all the potentials that that could be coming down the road with it and clearly we're going to have to, uh, hold off on all of that excitement.
The other trial that you had highlighted and that I was really happy that you focused in on was frail AF. ,
John: well, I love Frail AF because I'm a clinician and this was , a group of primary care clinicians. From the Netherlands. And I'm part Dutch, so I'm totally biased that I, I have a little bit of emotion in, in, in why I like this. But the idea was, um, uh, and I'll, and I'll probably pronounce his name wrong, but Gert John ing is this sort of.
Uh, primary care physician, academic and takes care of older patients, frail, older patients. And, um, many of his colleagues were saying that we needed to switch these frail, elderly folks with a fib from vitamin K antagonist and in the U. S. it's warfarin, but there, there are others, but basically warfarin like drugs to, uh, DOACs.
And he's like, ah, you know, um, all the DOAC warfarin trials randomized really healthy patients that were ambulatory outpatient clinics, very different patients. They weren't a lot of frail patients, and so they actually did a randomized trial, , where they randomized older, frailer patients, um, and, uh, which is really cool, right?
Because. I don't know about you, but , in Louisville, , I take care of a lot of older Frail patients. , and , it's where the rubber hits the road. And, um, they found actually results that were pretty shocking that, that actually it was not a good thing to, switch.
And, , I'm looking at the trial now just to remind myself the mean age of these patients were 83 years old. And, , they actually had to stop the trial because of, , increased bleeding. And so they basically found that we shouldn't be switching older, frailer patients who are doing well on warfarin like drugs.
We shouldn't switch them. And I think the lessons were so great because, , not only does it inform, uh, that patient population of 83 year old patients with high frailty scores, but it also really globally teaches us that the translation of clinical trials to the clinic is really something that we should do carefully because what we show in ambulatory outpatients who are good enough to be in a trial, , may not be true in a, , older patient population with more comorbidity.
Tom: It also , begs that question about once you've, once you've had someone who has been stable on a medication, it doesn't mean people will be stable on other medications. Right. You, you've given them their vitamin K stress test and they passed their vitamin K stress test. So it's fine for them is going to be the other group passed their DOAC stress test, and it is like you said, it's, it's a question of translation afterward and so you really have to re-look at all of this data in that different light.
John: And I think a really important message, I think it's important not to say that for older, frailer patients that vitamin K antagonists are the better drug. This was, this was, this was really a strategy trial where patients who were doing well on those drugs, should we just switch them because the, uh, the apixaban data looked, or the DOAC data looked better.
And it's not, it's a switching trial, not so much initiation trial, but , it hazard a ratio for the primary outcome, which was, which was bleeding, , was 1. 69. and so 69 percent increase in, in bleeding. , and so it's just remarkable how such a surprise and, , it was presented as a late breaker at ESC.
, it didn't get into New England Journal, but it really deserved, I think, to be at the highest level journal , , the investigators said that. We don't know the right answer. And so we're going to do a randomized trial 2600 patients and they recruited a bunch of centers and they did it.
And I just, it's so exemplary for so many reasons , I just, it's one of my favorites,
Tom: , what the trial says , was definitely beyond just the top of the line outcome. . And what it says about how we do trials is, is probably even more important.
. And I think that, uh, this is 1 of the reasons why we need to put a megaphone to what you are saying and what other people are saying because. This is not a trial that was going to make a lot of people a lot of money. This was going to be a trial that was going to purely about helping patients and solving a clinical question
, but just to dive into it for a second is, we, we have all this information and all these databases ,? All of the information that we've given them is because we've had a bias on how we're treating our patients before.
And, you know, when you talk to the people who are developing, uh, all these AI tools, Um, and they believe that if you have make a large enough database that you're going to be able to get important clinical information. And , when you push back saying, you know, there's bias, I said, no, we'll be able to, we'll be able to identify that bias and, re stratify it.
And I'm, I'm hesitant. .
John: Yeah, that's a huge topic that you just delved into. , what they did in this, what they did in this trial is they randomized patients. And so when you randomize patients, you balance, you balance a known and unknown baseline characteristics and confounders. But. You know, the idea was, could you, could you do this with an observational database looking back at all of the older, frailer patients?
And I really don't think you can, because it's very difficult to control for factors that are not in those, , databases that, , big data can look at. And so, the way they answered this question is, is, is is awesome. And people say, you know, Vinay Prasad talks about this all the time. And Vinay says, you know, we can't do a clinical trial.
And here, , in the North of the Netherlands, they basically got into an argument about what they should do with frail older patients. And they said, well, we can do it. We can do clinical trials and, and, and they did it. .
Tom: Yeah, , we'll dive into that again in just a second. I wanted to talk quickly about, , ARTESIA and Noah AFNET 6. Uh, again, another really sort of interesting trial, uh, looking at patients with short duration atrial fibrillation. With Artesia that it wasn't a benign placebo, um, I don't think that it was the right placebo. I don't think that we should have included aspirin, uh, as, as the placebo there.
John: ,
well, I think this is one of the most important topics of the year, actually, Tom. The deal is that we Uh, in the electrophysiology department, we get a ton of alerts, , almost daily about a patient who has a pacemaker or a defibrillator, and they have two hours of AFib or three hours of AFib.
They don't feel this AFib, but when you look at the recordings, it's definitely AFib , but it's not the same as the AFib of old, right? So the AFib that established anticoagulation as beneficial, , were patients who had enough AFib to present to a doctor's office to have a 12 lead EKG. Usually they had symptoms or they had the AFib long enough to have it recorded on a 12 lead and so that, so-called clinical AFib.
I think what, what these two trials have shown, , NOA Edoxaban versus placebo and Artesia apixaban versus aspirin, what they have shown is that subclinical AFib, the short duration AFib episodes is a different species than clinical AFib. I think it's the most important lesson. And, , both trials, I think, showed consistent results.
NOAA AFib was a smaller trial, 2, 500 patients, , shorter duration, , stop for futility, basically a almost 20 percent reduction , in stroke , and systemic embolism and CV death, 31 percent increase in bleeding. So Decreased stroke, increased bleeding, and artesia with aspirin versus apixaban, apixaban versus aspirin, about the same thing.
Uh, but, but artesia had more patients and longer follow up, and so they had more events, and so the, the hazard ratio for stroke came out positive. It's a, it's a 37 percent reduction, and, and it was statistically significant, but 80 percent increase in bleeding. I think the, the commonalities of the trials is that it's number one, it's going the same way strokes are going down with DOACs uh, bleeding's going up that net benefit is problematic because you have a very low stroke rate.
When you look at both trials, I think this is one of the most important messages that the stroke rates, the yearly stroke rates are about less than 1%. So we typically say Chads VASc is 2.2% per year, and that's where the, for sure, a net benefit occurs. But at a stroke rate less than 1%, how do you get a net benefit with a DOAC?
, the reason why the stroke rates are probably that low is the median duration of episodes was very short. I think in the hour, one or two hours. So, . What I'm looking for, and I think they'll do, is they'll put these trials together and they'll do an individual patient level meta analysis.
And they'll draw a graph where we can look at duration of AFib and we can look at net benefit or stroke reduction and bleeding increase. And there's probably going to be, there's probably going to be a point where 6 hours, 8 hours, 10 hours. Uh, where it's a 12 hour episode of a fib is going to be different than a 1 hour episode of a fib, and it's not going to be perfect.
And so we're still, unfortunately, despite . Nearly 7, 000 patients randomized. In really good trials, it's very difficult to have an exact answer of , when a colleague asked me, John, do we need to put this patient on anticoagulation for this three hour episode of AFib?
And I don't know that , these trials don't allow for an algorithm or a best practice advisory. There, you're still going to have to use clinical judgment and we're going to get more data about duration of AFib and I think that's going to be a big player. But I think that the, the biggest take home for me is that subclinical short duration AFib is not the same species as true clinical AFib.
And I, and I'll tell you where it really plays into, another place it plays into is post stroke, , loop recorders. Because , when we do loop recorders in post stroke patients, , we're actually going to find a lot of these one or two hour episodes. And it's not clear to me That anticoagulation , is going to be beneficial in those patients.
Tom: , I couldn't agree more. And , when you looked at , , the stroke events here were half of what you would expect , and when you look back at it, the other AFib trials, they were pretty much spot on with what the Chad Vest risk would say pretty close.
Um, and this was less than half. And so it tells you that Chad's best risk probably is correct in clinical AF. And it also says that, yeah, the risk does increase, but it is not the same and , then you'll have to make that decision of timing. Um, I'm always intrigued about the patients who had, we even had.
A day or two of a fib during hospitalization, we start them on anticoagulation, uh, because of that. And then we don't see it again for a year or two. Uh, and they're actually not involved in the clinical trials, right? For the clinical trials for these medications, they have to have had a fatigue of population.
You're recorded more than 24 hours. At least once within the past 12 months. And so that patient was in the hospital who had the atrial fibrillation two years ago and you haven't seen it since. I don't know whether they're actually getting the benefit and we're, we're certainly putting them at a bleeding risk.
You know, that we can say for certain, right? There's, you know, there's no question as long as they're taking the medication.
John: Right. , but I also think that, I also think that these were amazing trials, and , really, um, it, I wish they gave us a clear answer, but they're still amazing trials and, , and kudos to the investigators for these things. They were , really good.
Tom: Yeah. , there was, , some online discussion, criticizing clinical trials and, and how we're doing it. And I don't think that any of the things that we're saying , when we're being critical, , and we're dissecting them and discussing them and trying to make them better.
, , there are very few people who are going into this work trying to do something harmful or bad, right? We're, we're all trying, we're all looking for the same, a good positive outcomes and good results, , but, , you want to make it better, right? How do we get these things better in the future?
Um, and actually we'll jump forward. A question , is . , this idea of trying to do sham procedures when we're doing clinical trials, making sure that we have appropriate placebo controls, making sure that they're that they're answering clinical questions that we really need to have answered or that that they're looking in a way that is going to give us.
information that's important to us clinically. And, uh, you've pointed out in a number of your reviews of some of the trials that, you know, having these composite endpoints that have five or six different pieces to them, yeah, it may help us to do, to say, yes. There's a statistical benefit when you combine all these things together, doesn't really matter to us when, when, when does it, you know, when, when do we say, yeah, that's going to make me use this drug or not.
John: Yeah, thanks for saying that. I mean, all of all of the critical appraisal that I do is, it's never personal about, uh, it's never an affair. I don't think any of this is nefarious. It's just, you know, we're, we're, we're criticizing and critically appraising the trials or appraising the trials just because as consumers of medical evidence, , we have to use them and, , and it's complicated and, , so that's, that's the point.
It's not to, , bust up trials or, , Say this is bad or this is good. It's just to say, you know, , here's what they did. Here's the patients they recruited. Here's what they found. And here's how we might use it in the clinic.
Tom: So now let's go into, uh, into many people's personal bet noirs, which is a left atrial appendage occlusion, , devices. , it's interesting because it's, , it's certainly, it is a popular intervention. People, , like the idea, they want it.
, I think that there is a, there's certainly selection bias involved. Uh, and I think that there is, uh, if not outcome bias, there's certainly going to be almost a, a placebo bias to the physician for doing it. Yeah. Okay.
John: my emotion about this because I've lost this debate on 5 continents. I've. debated against, I've debated against appendage closure, uh, seriously on five continents and you can see how popular it is. So obviously, uh, my opinion is a minority opinion, but, but in a nutshell, in a nutshell, uh, since you have me on, I will say my piece.
Um, I didn't, it wasn't even my fault. It was my, my editors at heart. org asked me to look into this many, many years ago. And I started looking into it and I was like, Oh my gosh. And, um, and Andrew Foy at Penn State and I have published a lot on it, but basically the problem, you can put it on one slide. The problem is young doctors don't.
They're busy learning procedures, and it's a complicated procedure, it takes a lot of skill. Um, I always encourage young doctors to go back and look at the seminal data that underpinned the approval. And basically, there was the two clinical trials, PROTECT AF. PROTECT AF was a Watchman versus Warfarin, and it did meet non inferiority.
And if you read JACC, or JAMA, where it's published, both. Uh, you'll find that it was a positive trial. The problem is that it didn't pass FDA muster. FDA evaluators said, ah, there's too many problems. There were basically internal validity problems with the trial. And so they said, do another trial. And so they, they did prevail and they borrowed some patients from protect, but they did prevail.
And they And actually Prevail, the main primary endpoint, the co primary, the stroke systemic embolism and CV death did not meet non inferiority in Prevail. Now it met non inferiority in the secondary co primary endpoint, which excluded the seven days after the procedure. So, Prevail actually did not meet non inferiority and non inferiority margin was 1.
75, so it could be 75 percent worse and it didn't meet that. And then people say, okay, well, it decreases bleeding, John. You can take people off anticoagulants. But when you look at the Holmes meta analysis, , in JACC, , when you include the procedure, actually, major bleeding is not any different between
And then they say, well, we have a low complication rate. They have watchman flex. They have iterations. But when you really look at, , real world data, , mandatory data, not these voluntary registries from Europe, , patients start this preventive gamble. It's a prevention gamble with a, at least a 5 percent major complication rate, uh, in, in real world data.
And then of course, I think one of the most important studies to consider with appendage closure is Averroes. And you remember Averroes was Apixaban versus aspirin in patients felt to be ineligible for anticoagulation. So the exact kind of patient that we would consider Watchman and in Averroes, uh, a pixaban was much better than aspirin for, for stroke prevention, but the.
The real important slide is that in major bleeding, at least in the intention to treat, there was no difference between apixaban and aspirin. And so I, for me, the, the seminal data is the problem. And then, , I think the newest study that came out this year was the one year data from Swiss Aperol.
And Swiss Aperol was this humble little study from Switzerland that looked at Amulet versus Watchman, two different companies, , closure device. And they basically found that they were similar, but that they had a 50 percent patency rate. So that means that, uh, when you look at these things with CT scan, only 50 percent of the devices had totally sealed the appendage.
And I think that explains the higher ischemic stroke rates in, in a seminal trial. So I am, I am open minded the nanosecond that somebody comes out with a study that shows that there's a net benefit in a. group of patients, any group, I would refer for it. The one group of patients, the one positive thing I'll say is that when we, when I accidentally electrically isolate the left atrial appendage during complex ablation, usually the second or third or fourth LA ablation, when the appendage is sitting there and it's not electrically activated, I think it's reasonable to consider, , appendage closure , in those patients, especially if they're, have stroke or if they have difficulty taking anticoagulants, because that's one of the highest risk thromboembolic states, , akin to a mechanical valve.
So I'm just, I'm open minded and I use new stuff all the time. Conduction system pacing, you know, I'm open to new things. It's just, , I'm not convinced.
Tom: , I was approached whatever it was, 15, 16 years ago to be involved , the first watchman trial. And when we were looking at it, , we said, wait a minute, this is the wrong comparator group, right?
You're trying to talk about it in patients who you do not think should be on anticoagulation. And yet you have to have people who are on anticoagulation. And you're going to take them off, um, without clearly demonstrating ahead of time that this is going to be something that's going to work. And so, you know, it's almost like you're threading the needle here.
You want to have the perfect patient who can be on anticoagulation for some period of time that you're not going to then have to stop, or if you have to continue on for a longer period of time that they don't have the risk and you have to have them have it for long enough so that your initial problem that you have with the procedure.
Your procedural risk gets outplayed by the decreased bleeding risk over time. Yeah. patients who clearly are going to live for three years and then they're going to start accrue. Yeah. And then they may accrue, accrue benefit. Yeah. So I know it's funny in my family, my, my son has four grandparents, four grandfathers.
And so Uh, that's, that's a longer story, but, um, lots of divorces and remarriages. So four grandfathers of the four grandfathers, they all have atrial fibrillation. We have a different way of treating each one of them based on what's happened. One of them is on anticoagulation and fine. One of those off anticoagulation because we couldn't tolerate and didn't think he was going to benefit.
One who was, we were afraid of stopping anticoagulation, wasn't having a true problem, but lives on an island, had hit his head a couple of times mechanically, and he has a watchman. It's been two and a half years. He's doing fine. He may have been done fine with placebo. Right. Because. These are that rear, you know, now, and then one who's just going, you know, that we're just watching because he's, you know, start going back and forth and parents, right.
So we say, okay, , it's going to be difficult to judge and how you're going to approach it. You know, one of the questions as I was preparing for tonight is. Other than the patient who you've electrically isolated the appendage, is there any other patient you would say, yeah, that's the guy that I think it's not unreasonable for because, you know, patients want this and we have to sort of inform them.
And I, I think that it's, it doesn't, it doesn't say the service sometimes because we, we tend to lead towards things that, , our partners are doing, you know, I used to do them , we like to, , we like to do procedures. We like to help people. We like to have errors of commission rather than errors of omission.
John: , I've got a thought, but I want to respond to one of the things you said about patients want them, patients want them because there's marketing and there's momentum and there's, it's sort of what I call the therapeutic fashion.
And , I'm not convinced that, from any of the data, that there's, that there's an obvious net benefit. Now, , I will say, as far as the perfect patient, I will tell you who's the least likely to benefit. And the least likely to benefit is the most common patient that we're doing in, in the U.
S. The most common patients that we're doing, and I've actually published a study with Andrew on this, um, is that when you look at the patients that are being done, when you look at the NCDR database, we're doing. older patients with comorbidity. And the problem with that, Tom, is that patients that we're doing have the most competing risks for a stroke.
, they're the least likely to benefit. They have a shorter longevity, more comorbidity, more competing risk. And, , paradoxically the patient who's most likely to benefit is a patient , who has the most stroke risk from the appendage, from the, Uh, from the atrial fibrillation and the least amount of comorbidities.
And the problems that I've had, and I've argued with this, is that we really should have the guts to do a clinical trial where a patient who, , has, , challenges to taking anticoagulation is randomized to no treatment. And we don't have, we don't have a no treatment arm. We don't have, if we don't expose that patient to a 5 percent risk, we Uh, of, of the procedural complications or even a 3 percent risk or 2 percent risk if say it's a trial with the highest level operators.
If we don't expose them to that upfront risk, and we just give no treatment, then we would, have a better comparator. The problem is that we have anti platelets and, , we don't have , the guts to do that. And if we did, and it was better, I would consider it. But the, I would say that the patients that we're doing now most commonly are the least.
Likely to benefit, and I, furthermore, the last thing to say is that when I see a patient who really cannot take an anticoagulant because they have an AVM or they, they truly have a bleed every time they take an anticoagulant, these are the scariest patients to do because if they develop a device related thrombus, they are really between a rock and a hard place because They're going to have 4x, 5x, the risk of thrombosis with that, uh, device related thrombus, , and , they actually can't take anticoagulants.
And so I'm really scared about patients who have an absolute contraindication to anticoagulants.
Tom: , and , your point, which is a little bit subtle there is that, you know, there are other things that we're treating with the anticoagulants rather than just the appendage thrombosis. , and , you look at the other antithrombotic pieces of this, Whether it's peripheral arterial disease, whether it's coronary disease, et cetera, you're getting other benefit here beyond.
And so, you know, you can stop aspirin on patients who have coronary disease when they're on a DOAC, , you there's benefit in coronary disease with warfarin, right? You know, we're doing other things beyond just Uh, stopping the thrombosis in the appendage. , you can't extrapolate, , just because we think we know the etiology.
, like you said, if we could do a trial, the trial that I wanted 15 years ago, which was to put it in, in patients who we could not put them on warfarin. Or needed to stop the warfarin or wanted to stop the warfarin and that really hasn't been done, which is too bad, it's in nobody's economic interest.
Um, but maybe it will be, , one of the questions, , was trying to figure out other ways of identifying risk. , , as you said, , patients who have clinical atrial fibrillation for more than 24 hours, um, those patients are clearly going to be at higher risk.
Are there other things that we should be doing in order to determine their risk? In coronary disease, we're starting to finally look at plaque morphology using CTAs, maybe using a fat attenuation index as a reasonable surrogate. But is there other things that you are looking at, say inflammation in the atrium, uh, there's some information that the more fibrosis, the more inflammation there, maybe they'll run higher risk. Have you been looking at any of that data
John: , so I have competing forces in my brain. These two, these two, uh, uh, friends in my brain. One friend is One friend is saying, John, don't be skeptical because this may be one role for machine learning and artificial intelligence where, can't we imagine a future where someone can look at echo clinical parameters, left atrial strain, left atrial size, maybe, LV fibrosis, , and mix in that with clinical data and come up with this risk score and say, okay, okay.
Here we have all of this clinical data. Uh, we've calculated a risk score with machine learning, and it's going to tell us which patients that we need to, uh, be more aggressive with. And and so that's 1 voice. Um, and because, of course. Tom, you're my age or you're close to it. And so you know that the things that we do now weren't even conceived when we were training.
So we have to, have that, humility about the future. But on the other hand, the other, voice in my brain says that. You know, we have been at this for a long time since Chad's VASC, which is a silly little interger score, fast and furious, um, uh, kind of thing like Gigerenzer talks about.
And, , we haven't shown anything better than that. And so, fast and frugal may, may be very difficult to beat. , an analogy is coronary calcium scores , we have. Polygenic risk scores, coordinated calcium, but is it really that much better than putting your basic parameters in a pooled cohort equation and coming up with a 10 year risk?
And so that's our challenge is, will all of this fancy, , amalgam of imaging and clinical data be any better than just basically the CHADS-VASC, which is equivalent to a pooled cohort equation? I don't know. I'm like 51 49 on it. I don't, I don't know.
Tom: the, well, the, one of the problems with the equations is the patient in front of you is one person. When you're looking at the equations, it's a, it's a, it's a whole group of people, you know, you're, and you have to make that decision. And, , a number of people have said this, but of all the decisions that we make, we only have data on about 10 to 20 percent of them.
Everything else, we're, we're making it up or we're, we're extrapolating from the data. We look at the person in front of us and say, yeah, this is the way we should go. , when we were looking at, uh, some of the original PCI versus cabbage. Data. We saw that.
Yes, cabbage was better, but when you looked at the patients who were removed from the trials and they remained in the registries, the decisions that the clinician made were actually better than the outcomes from the trial. And so there, there, there are certain things that we're going to get right because we look at the patient, but , we need more help and more data .
You can't just say. Yeah, they have a 7 percent risk. I'm going to treat them with statins. . It will be interesting. Although your coronary calcium comment.
I just was talking with Matt Budoff and I think that you might have started a fight with him on that.
John: Oh yeah. Yeah. We, we've had plenty of debates.
Tom: But, but, but I like his data, , and I'm pleased that they're utilizing this as ways of getting into clinical trials as well, right?
Because if you include that as part of your entrance criteria, then you can use that data to extrapolate out from there. If you go back and you then look at it. And you say, Oh yeah, it seemed to work. And here's the groups that it worked in. That's a different question. , but a lot of the new trials that we're doing, especially for coronary disease, but that's a coronary disease question.
And we're, I'm trying to keep this on atrial fibrillation. Um,
John: I know, I know, I know. I hear you just, just to, to, to inject a little humility into the, into the whole Chad's fast thing. , there's a paper I like to cite. It doesn't get much press, but it's, and the first author is Gene Quinn. And it was from Danny Singer's group. And. They published in 2016 in Circulation, and they basically looked at different cohorts of AFib.
Uh, and it could be observational, it could be randomized trials, and they were interested in what's the untreated stroke risk? And what they found is huge variations in stroke risk. So with our little chadsvasc calculator that we get from a CC, we say Chadsvasc, uh, two is 2.2% risk. And that is just, that is just like a fantasy that we think it's 2.2% because what they've shown is that it could be 0.9%, it could be 5%.
And so, uh, depending on which cohort you looked at. So I think. I, I just think we have to be humble about what, what we can predict about the future.
Tom: I was, uh, I was talking to Sharon Hayes and she was, um, talking about the, uh, the AI that they're doing up at Mayo and they, I'm sure you've seen that
John: Been there, been there. Yeah. I gave rounds there and they showed it to me.
Tom: Totally cool, right? You look at that EKG and it tells you what the percentage of chance that this patient has heart failure, what percent of the chance they have, and You gotta love that.
John: Yeah. I mean, I know we have to move on, but I feel the same way about the, the AI ECG stuff at Mayo as I do about the coronary calcium score. If they, if they spent equal amount of time, um, doing what Gert Van Gursing, Gert Jan Gursing did in Netherlands. If they spend an equal amount of time just randomizing patients into a colonoscopy trial where you have a 20, 000 patients who you offer it to and 20, 000 patients you don't and you, in five years, you look at outcomes.
I think that's the ultimate thing that would convince me. And so we'll see if that's going to be done. I don't know if it will, but, , that's why I love FRAIL AF so much.
Tom: So last quick question, , because I, I want to be, cognizant of your time and I, and I thank you again for spending so much time, um, is the Apple watch wearables, and atrial fibrillation. Every day in my clinic, I have at least four patients who come in and say, and by the way, this is what I've seen on my Apple watched over the last week.
And they'll come and they'll be really concerned, you know, my heart rate went down to 40 last night. What does that mean? , I think I had atrial fibrillation, , and , , what's interesting is now we're using Apple watches for a number of the HFib trials. Uh, I think for a couple of the ablation trials are using them, they're, , they're utilizing them for looking at end points.
Uh, which I think is going to be important because then we'll be able to again, extrapolate out the database ,
John: , Let me start with a positive because have some negative things to say too, but let me start with a positive. , here's where I think, , I've enjoyed having the Apple watch. Let's say I have a patient who has paroxysmal AFib and their symptoms have been shown to correlate with their, with their, uh, notifications and the patient's recording a rhythm and they're sending it to me.
And yes, I can confirm. the diagnosis of AFib, so that's good, and I can show that there's pretty good correlation with their symptoms and their episodes and their notifications. And then I do AFib ablation. And then, uh, six months to one year later, they're asking me, can I get off anticoagulants? And let's say they're not, let's say they're a medium risk, CHA2DS2 VASc1 or 2, and so they're, it's really on the fence.
And so they're like, doc, I'm not having a fib. You're not recording a fib. And my Apple watch gives me zero notifications. Whereas before the ablation, it was giving me notifications. And so I'm just speculating. I don't know the negative predictive value of that, of that situation, but I feel pretty good that if the patient's not having symptoms, I'm not recording a fib.
And they're not getting notifications, then I feel like, okay, that would be a really, uh, that would be a patient to put into a trial where we would randomize to continuing at that coagulation or stopping at a coagulation, or at least I'm leaning towards stopping at a coagulation that patient. So, I think it has some clinical.
Utility and I could say the same thing for , the cardio mobile, the problem though, is the apple doesn't want to market it. Uh, and smartwatch makers don't want to market for a specific, nice, clinical, clean indication like that.
They want to market for this digital health to all kinds of people. And we've already talked earlier in the podcast about how in patients with established heart disease, enough to have a cardiac device. Uh, uh, pacemaker or defibrillator, uh, in these patients who have established heart disease, who have a little bit of AFib, that's probably not enough to treat.
And now we're having 40 and 50 year olds who are getting diagnosed with these paroxysmal atrial arrhythmias. So we, we don't know, plus, and as you mentioned, all the over diagnosis and medicalization that goes into that. And when they meet somebody like you, uh, and you. You see somebody with a heart rate of 40 or 38, they're healthy, you just tell them, don't worry about it.
Other people might do a monitor, they might do an echocardiogram, and that might lead to, you know, a TEE, and that might lead to whatever. I do really worry about the medicalization and the overuse and overdiagnosis. Though, I will say that I think that's getting a little better with time as we learn , where these new normals are.
The next thing is the, the blood pressure monitors without a cuff. And, you know, so those are going to be fraught, right? Because blood pressure is going to be even more variable and noisy. Then, then rhythm management. So , , I'm mostly negative on them. I don't have one.
I don't recommend them, but if a patient like the one I described has one, I won't ignore the data that comes from it.
Tom: , I completely agree with you 100%. I do think that there's a little bit of, there's going to be some interesting things when we talk about athletes. And I know that you're a cyclist. I, I like to describe myself as a reformed triathlete. And I take care of a lot of athletes and, , I'm going to be having a discussion with Aaron baggage, uh, coming up about ischemic coronary artery disease and, endurance athletics.
And part of the part of that is actually going to be talking about utilizing these washes and , what the data was really helping us with. And the vast, vast, vast, vast, vast majority of the time, it's not helping us at all. I took off my watch years ago.
I said, okay, enough, , I was doing a lot of training. I said, this is not changing anything. We love looking at the data. We love analyzing it and scrutinizing it and pulling the, , the fly shit out of the pepper but it's not probably not making any difference. And it's probably not helping us as much, , and for all the people that it's.
I wouldn't say harming, , but, making them worried for days and days and days, uh, until they get into see us you know, that's, that's not nothing, right.
John: No, , and you downplay it because , if one out of 50 people get into a cascade that causes a complication or, or excess cost or whatever, I think it's a, it's a negative., they used to think it was all minus, but I, I do now see some occasional positives.
Tom: so last, or there actually are two more questions. One is what else are we missing? What did we, so we've had a lot of discussion about atrial fibrillation. I wanted to try to sort of focus it on that. Right. But, um, is there something that you would say, okay, , , this is something that people are missing.
, what should we really be talking about?
John: , I'm glad you asked. , one of the biggest, deficits in knowledge that we have in EP and AFib is that we have never had an AFib ablation trial, with a proper sham control. In other words, where. People go in and , they don't do the procedure.
We, we've had that for PCI. Um, we've had that for knee surgery. We've had that for so many things, but we've never had it for AFib. And I've been barking up this tree for years that somebody needs to have the guts. And two days ago Two days ago, uh, Malcolm Finlay, Finlay's group at BARTs in London have published a proof of concept study , they call it Orbita AFib, where they actually have randomized, 20 patients, to, AFib ablation PVI, uh, versus cardioversion and sham PVI, and they, they don't have any results, they just showed that, Patient blinding was upheld to 1 year.
So, their conclusion was that it can be done and there's also a sham control trial going on in Leipzig and I think there's another one in the UK. So, I think that what we'll be able to have in the next couple of years. is we'll have some idea of what the placebo resistant quality of life effect is with AFib.
Because right now, , AFib ablation is done primarily for, quality of life. And patients feel better after they have AFib ablation, but we always randomize patients to, tablets, which make them worse, or ablation. And so one group gets a big procedure, one group gets nothing. And so we've not had a fair comparison.
So, , the paper, the Orbita AF papers published the American heart journal just, December 17th. So, uh, it's feasibility, no data, but they've shown that they can do it and they're going to proceed. So I'm very excited about this.
Tom: Do you think that the sham, uh, control for that should be a true sham , because, you know, there's more to that, the procedure than just isolating, right? You're killing cells there.
John: , I agree. There's going to be a debate. . What they did is they put two sheets in the femoral vein and then they just cardioverted a patient under sedation and, um, and then versus PBI with, I think they did cryo balloon mostly. But we, we absolutely would, the proper way to do it would be to do some sort of ablation, uh, either CTI ablation, like a flutter ablation or possibly like a partial.
But , the issue, of course, is the ethical, the ethical boundaries of, of, of what you can do. And so, uh, any sham. Is going to be better than nothing and we'll we'll have some idea. I think my personal opinion is that it's when we ablate a fib and we're successful.
It's more than placebo, but there's a strong placebo effect. So it's going to be 1 of those things like CRT where there's a percentage. Um, I think Daryl Francis's group has shown that the percentage of placebo resistant benefit with CRT is in the range of like 16 to 20 percent. Most of it is placebo effect, but there is a real effect.
And I think with AFib ablation, there's going to be a real effect, but it's going to be a large placebo effect as well.
Tom: So , the final question, is, who do you want to give a shout out to? Who's, whose research are you, are you reading? It doesn't have to be on AFib, but you just say, yeah, I really like that. I think that's, that's really cool reading. I'm really excited about what they're producing.
Um, go take a look at this because , that's the next thing.
John: Yeah, I think the person that I would, the person that I would give a shout out to, and of course I'm biased because he's a friend, but I think the work that Andrew Foy is doing at, in Penn State University in Hershey. If your readers or listeners go and they just go on PubMed and put in Andrew Foy, F O Y.
I mean, what he's doing is a ton of work on multi morbidity. And looking at how multimorbidity affects our, uh, translation of trials. , of course the big trialists, they always, they do a great job and , they get shout outs from everybody, the, the DCRI and McMasters. And, of course, Richard Whitlock has done amazing work in the LAOS3 study.
But I think Andrew's work with multimorbidity really interests me because I see so many patients that are unlike trial patients. But just for instance, all of the heart failure data comes mostly in ambulatory outpatient heart failure patients. But we're basically , pushed to put, , inpatients, frail, older patients on four drug therapy.
But all of the data comes from, uh, really much more stable patients.
Tom: Well, once again, Dr. Mandrola, thank you for joining us today. I I've completely gone over the timeframe and, uh, so I owe you one, uh,
John: we just did an hour. It's not bad.
Tom: But thanks so much for joining us today. And, , I look forward to having another conversation in a year or two and saying, okay, , how far have we gotten , and I'm going to bet that, uh, what do they say?
, the arc of history leans towards justice. , I'm sure I'm destroying that one, , but, I believe that the arc of science is going to lean towards a more conservative approach . So thank you.
John: You're welcome. Thank you for having me.