Feb 21 2025 This Week in Cardiology

You're listening to This Week in Cardiology from TheHeart.org, Medscape Cardiology. This podcast is intended for healthcare professionals only. Any views expressed are the presenter's own and do not necessarily reflect the views of WebMD or Medscape. Hi, everyone. This is John Mandrola from TheHeart.org, Medscape Cardiology, and this is This Week in Cardiology for February 21, 2025. This week, we're talking blanking period after ablation, periprocedural MI after PCI and non-STEMing, predicting AFib after ischemic stroke, and proper standards for mitral valve repair in primary mitral regurgitation. Before starting, I'll apologize a bit for my voice, as this is day eight of the influenza A and my utter failure to respond to the mandated flu shot that I received. I and other members of the Mandrola compound have been decimated by influenza A. Our hospital's been overflowing, and it has been a tough time for influenza here in this community. But I have improved and survived, as have all the Mandrolas. And let's start talking about these cardiology topics. And the first is the blanking period after AFib ablation. And I have numerous studies to speak about and some opinions about this blanking period business. As many of you know, EP doctors have blessed themselves with something called the blanking period after AFib ablation. That is that AF that occurs in the 12 weeks or eight weeks or sometime after the procedure is uncounted against procedural success. It's one of these dogmas that got established that has never been questioned. The idea being that thermal energy from the ablation could actually cause irritability of the atrium, and then it might cause AFib that then settles down. That's the idea. Now, before I tell you about my friend Jason Andrade's study that's published in CERC, Arrhythmia and Electrophysiology, I'll remind you about the blanking period after PCI, or the blanking period after TAVI, or after left atrial appendage closure. Wait, you ask, what blanking period? And you would be correct. There are no blanking periods after these other procedures. In fact, I know of no other cardiac procedure, which has gifted this period of time after the procedure where failures don't count. Now, the research letter from Dr. Andrade's group in Vancouver centered on how AFib ablation blanking period impacts AF recurrence and AF burden. They used data from their 2019 CERCADOS randomized control trial. This was a three-arm RCT comparing RF energy, a four-minute cryo-balloon freeze, and two-minute cryo-balloon freeze. All patients in that trial had loop recorders. Now, the main CERCADOS trial showed that all three modalities provided similarly poor efficacy of the primary endpoint of time to first AF. It was about 53% in all three groups. And I say poor because, remember, AF ablation results are poor and have not improved in 20 years. But recently, the European guidelines have recommended shortening the blanking period from 12 weeks to eight weeks. Now, having ILR data on these 346 patients in the CERCADOS trial allowed the authors to describe differences in outcomes based on the timing of recurrence after the procedure. And they had really four main findings with the eight-week rather than 12-week blanking period. First, the one-year freedom from any tachyarrhythmia was significantly lower with eight weeks. It was 50% versus 55% with 12 weeks. Now, that 5% drop in efficacy was statistically significant. The absolute number of patients who recurred early was only 14 of the 346 or 4%. Now, the good news was that AF burden, which can be measured with the loop recorder, was not statistically significant. It was very low in both groups, 0.18 versus 0.23, that is whether it's eight weeks or 12 weeks. And then they were able to progressively shorten the blanking period from six to four to two weeks. And this was associated with progressively lower freedom from AF, which was the primary endpoint. But none of these appreciably changed the AF burden. And the most interesting finding of their paper was that patients who recurred early versus only late had a 20 times higher burden than those that failed later. Now, this latter finding caused the authors to conclude that shortening the blanking period would enable earlier detection, but at the expense of potentially overtreating a minority of patients of about 4% who recurred but have a very low burden going forward. So, there's a second study. The Basel EP group has also published a sub-analysis of a trial looking at the impact of this eight-week blanking period. They published their paper in HeartRhythm two months ago in December, and their trial was a head-to-head comparison of the Boston Scientific CryoBalloon versus the Medtronic CryoBalloon at their center, and also all patients had loop recorders. And they did a similar thing to the Andrade group. They took the more than 5,200 episodes of early recurrence of atrial tachyarrhythmias. They called it ERAT, or early recurrence of atrial tachyarrhythmias, and studied the relationship of the timing of that and late failures. And they found similar things. First, early episodes of atrial tachyarrhythmia were common. 58% of patients had something in the early period. Now, pause there. And we have a humbling job that 58% of our patients have a recurrence in the first two to three months. The second finding was that it mattered. If there was an early recurrence and there was a much higher failure rate later. Success occurred in 82% versus 37% in patients without versus with early recurrences of atrial tachyarrhythmia. Third finding, it was better to have the early recurrence of ATAC early. When the atrial tachyarrhythmia recurrence occurred in the first, second, or third months of the blanking period, the freedom from late recurrence was 60%, 56%, and 11%. Pause there again. Recurrence in the third month translates to an 89% chance of failure. And the fourth finding was relatively good news. Only seven of their 200 patients in this trial, or 3.4%, had AF recurrence in the third month. So the new eight-week blanking period would reclassify only a small number of patient successes to patient failures. And now I have yet another paper on blanking period after now PFA, pulse field ablation. This is from Mohanty et al in Heart Rhythm. And they published on a consecutive series of 337 patients who had PFA for AFib. They found 53 patients, or 16%, had early recurrence of atrial tachyarrhythmia, 10 in the first month, 12 in the second month, and 31 in the third month. This group noted a huge difference in the risk of late recurrence or failures. Of the 10 patients with very early recurrence in the first month, 70% remained in sinus rhythm with no late recurrences. However, but at one year, all patients, all patients with recurrences in the second, third month after PFA ablation, ended up having late recurrences, most of whom underwent repeat procedures. And they concluded that the blanking period with PFA could be reduced to one month because recurrences in months two or three were so predictive of true failure. Now I have one more unpublished data point, quasi-unpublished. My friend, Stefano Mordignon, an EP in Frankfurt, Germany, who may have done the most PFA ablation in the world, sent me an email about data he presented at last year's ERA Congress, and this was from the Europea Registry. And it showed similar findings, namely that early recurrences after PFA were highly predictive of late recurrences. He and Dr. Vivek Reddy communicated that all the data with PFA, however, is from symptomatic recurrences, not with ILR data. So that is a limitation. I found no PFA studies with lubricators. My friend Stefano made the point that is worth sharing, that he said that in his opinion, we should really question the blanking period after PFA. Quote, I think that after PFA, a recurrence is a recurrence, even if it happens during the blanking period. Here's his thinking. Because, he writes, the rate of durable PVI in experienced fans is so high with PFA that even during the blanking period should be considered a sign that the patient is a PVI non-responder. Some of these patients would therefore profit from a redo ablation even during the blanking period. But the big problem, he adds, is that PFA is so fast and has such a favorable safety profile that in Europe, we started to ablate patients that we would have never ablated a few years ago. Therefore, some of the recurrence in the blanking period maybe are not a technology failure, but a failure in the indication process and patient selection. All right. That's a lot of studies, a lot of comments, but I think I can make this whole thing simpler. Mohamed Roussier, Andrew Foy, and I have written a brief editorial in the American Heart Journal, open, A-H-J, open, and our title of our short piece is called, Patients' Lives Don't Pause for a Blanking Period. We argue that there should be no blanking period, none. Foy told me this week that when he admitted patients at Penn State for AFib early after ablation, the patient who gets admitted, put in a hospital bed with an IV, and then has to have cardioversion or drugs, does not feel like the episode doesn't count. Of course it counts. I would add that some of these early recurrence are for left atrial flutter, which are worse than AFib because the slower atrial rate means a faster ventricular rate. And that's actually, that comment is actually a really important property of cardiac electrophysiology that we can talk about later. Well, thankfully though, we're seeing less pure left atrial flutters because we stopped carpet bombing so much on the first procedure. If you just do PVI, you're less likely to get left atrial flutters. Our other argument for vanquishing the blanking period is that it complicates trial interpretation. As I've just shown you, counting the early recurrences reduces the success rate to what is probably closer to truth. And now with PFA, it looks more and more like there should be no blanking period. But our strongest argument for starting a success or failure count from the AFiblation procedure is that no other procedure in cardiology has such a thing. Consider the issue of counting MIs in the ischemia trial of invasive versus conservative strategy in patients with ischemia and coronary artery disease. In the main analysis of ischemia, there was no difference in the composite MACE nor in MI whether you were randomized to delayed or conservative. However, if ischemia had a blanking period wherein post-PCI MIs were not counted, then the trial would have shown a statistically significant 33% reduction in MI in the invasive arm versus the conservative arm. And if you don't believe me, go to page 92 on the supplement of the ischemia trial. But it is not fair to exclude times around the procedure. I was hugely critical of the option trial of left atrial appendage closure plus AFIblation versus DOACs after AFIblation because the authors chose the primary safety event to exclude bleeding during or right after the procedure. You can't do that. Patience accounts for patience. I would argue that in the same way that these two examples are wrong, so has it been wrong to even have a blanking period at all after AFIblation. Now, the ESC has decreased the blanking period to eight weeks, but I would like to try and persuade you that we should reduce it to zero. Then the playing field is level for everyone. And perhaps if that's true, we don't look so foolish to our non-EP colleagues with this blanking period business. All right, second topic. Speaking of periprocedural events, I found a great paper, a circulation paper. It's a nice study from an Italian group, first author Matteo Armolada, looking at future events after a periprocedural myocardial injury with or without type 4 MI after PCI, which was being done for non-STEMI. I like this study for many reasons, but mainly it clearly illustrates why it's important to consider events around the time of an invasive procedure. Here it is, PCI for patients who come in with non-STEMI. So they looked at consecutive patients from an observational registry that was done in Bologna, Italy region. Now, let me just tell you about some definitions. Two definitions. To have periprocedural myocardial injury, or PMI, required a 20% increase in troponin after and an absolute post-procedural value of greater than five times the 99% upper reference limit. Now, to have a type 4 AMI, this had to be diagnosed in the presence of myocardial injury if there were A, new ECG changes, B, new QAVC imaging evidence of loss of viable myocardium or regional wall motion abnormality that was consistent with the ischemic origin of the non-STEMI, or D, angiographic findings consistent with a procedural flow-limiting complication such as a dissection, loss of a side branch, slow flow, thrombus, or distal embolization. Now, the primary endpoint they were looking out forward was all-cause death, but they also looked at secondary endpoints, which were a compositive MACE-type thing. They had about 1,400 patients, and their main findings were that a substantial portion of non-STEMI patients in the series, 37%, met the criteria for periprocedural injury. About 17% also had criteria for a type 4 AMI. These patients were about 70 to 73 years old, one-third female. Now, if you had periprocedural myocardial injury that increased the risk of MACE, the adjusted hazard ratio for future death was 2.7, injury versus no injury. If there was injury with a 4 AMI versus injury with no AMI, then the hazard ratio was 2.94 for death. So, substantially higher rates of death if you had injury and or AMI. For MACE, the trends were similar, though the hazard ratio was a little less robust, 1.4 hazard ratio for periprocedural AMI versus no microinjury, and 2.2 for AMI versus no AMI. Now, they also found that a post-PCI troponin change of greater than 20%, but less than 40%, was associated with one-year outcomes similar to those observed with post-procedural troponins less than 20%. But, conversely, a post-procedural troponin rise greater than 40%, combined with an absolute post-procedural value of greater than 5 times the 99th percentile for upper reference limit, was identified as the best threshold for diagnosing a prognostically relevant periprocedural myocardial injury. And patients exceeding this threshold had a significant four-fold and three-fold risk of one-year death and or MACEs, respectively. Now, some contextual things discussed in the paper. The threshold they estimated of more than 40% in troponin rise is higher than that's noted in PCI of chronic coronary syndromes. Of course, the authors say that those previous thresholds were grounded only in expert opinion. But this data suggests that they may be overly sensitive and it may take more of a rise of troponin to have prognostic utility. That said, they write, and I agree, that none of these predictions or prognosis things speak to any therapeutic changes because, of course, patients are going to have maximized secondary prevention therapy whether the troponin rises 20%, 40%, or 50%. My point in showing you this very nice data is that events around the procedure, even a surrogate marker such as troponin, should be counted in outcomes. There's no blanking period after PCI because these events are prognostic, as are early recurrences of atrial tachyarrhythmia after ablation. And I think this is a really nice paper and it's worth you take a look at it. All right, third topic is a little speculative. This is using artificial intelligence to detect AFib after ischemic stroke. The journal, eClinical Medicine, which I didn't hear of but is part of the Lancet journals, published an interesting study using AI to brain MRI scans in patients who've had stroke in an attempt to detect patients who will have AF. This is an important concept because we monitor the heck out of people after ischemic stroke and we monitor everybody and our yield is low so maybe AI can look at these brain scans and tell us which patients we should monitor. That's the idea. It's extremely clinically relevant. I mean, embolic ischemic stroke without obvious evidence of AFib always brings up the question of which antithrombotic to use, right? So you have this embolic stroke on MRI, you think it's probably due to AFib, but you don't know. But we know that there are two negative trials of so-called ESIS, or embolic stroke of unknown source, two trials of DOACs, RESPECT-ESIS, NAVIGATE-ESIS, both of which were negative. So you just don't use DOACs. The answer, of course, is to monitor for AFib, but it's costly and it takes time. So a team from China developed a convolutional neural network to sort out patterns from brain MI to try and identify patients who had had AFib. And again, I'm out on a limb here because I'm no expert in AI or machine learning, so it's hard for me to comment on the guts of this paper. I did ask around to a few friends, but the idea of this study is super important because it's about prediction and it's something that doctors struggle with. So AI models generally follow these steps, generally. First, they had 510 stroke patients that they looked at retrospectively to develop the algorithm. Then they used 73 patients who prospectively enrolled patients to test the algorithm. But they also had an external cohort from two other hospitals, which retrospectively used for validating the model. Now, the endpoints that they looked at were the area under the curve, AUC, sensitivity, specificity, accuracy, PPV, and NPV. And they had a combined classifier leveraging predefined radiomics features, and they had de novo features extracted by this convolutional neural network. And that was what was proposed to identify AFib in these acute ischemic stroke patients. Before I tell you the results, I want you to first notice the small number of patients. Two experts and friends in machine learning I reached out to said training models with neural networks usually take many more patients. I can't comment yay or nay on this, but it did seem like a really small number of patients. So in the three groups, the internal retrospective, the external retrospective, and the prospective test, AFib occurred in about a third of the stroke patients, which is kind of what you'd expect. In the end, if the combined radionics features and the convolutional neural network had an AUC of about 0.87, a sensitivity of 87%, specificity 79%, positive predictive value only 68%, and negative predictive value 94%. This is all for predicting AFib. The authors concluded that there was a potential association between brain ischemic lesion pattern on MR and underlying AFib. And they wrote, furthermore, that with additional validation, the AI model we developed may serve as a rapid screening tool for AFib in clinical practice in stroke units. My comments, again, I feel like we're at a point similar to practicing medicine without internet or smartphones. AI feels like that. It feels transformative. But let me say also that there are some post-stroke cases which are highly suggestive with the use of normal human intelligence, such as a patient who has multiple cortical stroke areas. They have a large LA. They might have had a history of AFib 10 years ago. And you don't find any atherosclerosis. You don't need AI to suspect AFib here is causal. You monitor these patients aggressively. But there are many, many more cases of these ischemic stroke patients that are gray areas. And that is where you hope an AI model looking at brain MRIs could help. But sadly, I don't think these numbers are good enough for decision making. So sensitivity-wise, 87 percent sounds good. But that model misses 13 percent of patients with AF. That's a lot. The specificity of 80 percent means that one in five is going to be a false positive. That's not good either. Another weakness in this case is that AF monitoring was only done for three days. That's not enough. The validation cohort was only 73 patients. That doesn't seem like enough. I think we have to see how these models work out with larger cohorts of patients. I cover this paper not because their model is actionable, but I feel like it's the beginning of something really big. And I wonder whether in the months and years to come, I'll be telling you about bigger, stronger models that are more predictive and that have higher sensitivity and specificity. Okay, final topic today is mitral valve repair. Now, I don't know about your zip code, but in my zip code, transcatheter edge-to-edge repair is super hot. I like to think of the field of dreams saying, field of dreams is said, build it and they will come. Well, the mitral device is built and it's being used. The thing is, and we mustn't forget that surgical mitral valve repair is the standard, at least for degenerative disease. And Jack has a descriptive paper from the team at San Rafael Center in Milan, the home of Dr. Alfieri of the Alfieri Stitch. And they report on 3,300 patients who had surgical mitral valve repair for degenerative MR between 2008 and 2017. These were young people, 57 years old, two-thirds men. The most used technique was posterior leaflet resection followed by edge-to-edge repair. They write that in patients in whom one repair technique was used, the second most common procedure was the central Alfieri Stitch. Almost all of their patients, 98%, received a concomitant mitral aneuplasty, which of course cannot be done with a transcatheter device. About a quarter of these patients had something else done, most often tricuspid repair. They have remarkable results. Overall, in-hospital mortality was 0.5% for all cases and 0.2% for isolated mitral valve repair. That's five of 2,400 patients who had in-hospital mortality. 10-year survival was 89%, which is also amazing. And the biggest result was that freedom from re-operation was 97% at 10 years. 10-year freedom from re-hospitalization for heart failure was 93%. And they noted that age, prior hospitalization for heart failure, pre-op symptoms, and anterior leaflet disease were identified as independent risk factors for a re-hospitalization for heart failure. Okay, why do I report on this? My comments. This paper exposes an uncomfortable truth in medicine, and this has been present since I started cardiology. Mitral valve repair requires surgical centers of excellence, including surgeons of excellence. Now, to have surgeons of excellence requires them to have excellent training and lots of practice. That latter part, lots of practice, could be increasingly challenging in the coming era of non-surgical interventions on valves. Now, I call this an uncomfortable truth because policy people like to think of doctors as interchangeable. And for some things, increasingly AF ablation, for instance, that is true, but not for valvular heart surgery. Surely not. The results from this work, from this center, is extraordinary. It's the kind of heart center you'd like to work in or close to. Now, you might say, come on, John, this is their own data and they're reporting their own data. It's going to be great. But I would say, okay, that is a limitation. But the authors did use an independent organization to match up patient outcomes and to do impartial assessments of outcomes. So, good on them. Now, right now, there are three ongoing trials that are going to compare transcatheter repair to surgical mitral valve repair. But I'm telling you that transcatheter repair is going to have to pass a high bar and we're going to have to look carefully at endpoints, length of follow-up, and that sort of thing. I know it's tempting for patients to choose non-surgical repair because who wouldn't, right? You get a catheter in the femoral vein. You get to go home either the same day or next day. Recovery is easy. It's certainly a lot easier than having your chest open. But at the age of these patients, right, 57 years, not having another re-op on your mitral valve is a pretty important metric. I think before we cardiologists go crazy with these mitral valve procedures, we ought to know the results of our surgical colleagues. Patients should know that at centers like the one in Milan, they will have a less than three in 1,000 chance of dying from the procedure and a 98% chance of not having another intervention over the next decade. And I wonder if our transcatheter friends are going to be able to replicate that. We shall see what the trials show, but centers like San Rafael and Milan are going to be hard to beat. So that's it for this week in cardiology. As always, I'm grateful that you listened. Thank you. And remember, if you liked this podcast, take the time, please, friends, take the time to give us a rating, write us a review. These help others find us. Also, if you don't agree with something I said, let me know. We'll discuss it. Until next week, this is John Mandrola from TheHeart.org, Medscape Cardiology. You're listening to This Week in Cardiology from TheHeart.org, Medscape Cardiology. This podcast is intended for healthcare professionals only. Any views expressed are the presenter's own and do not necessarily reflect the views of WebMD or Medscape.