Subscribe
Share
Share
Embed
Welcome to EP Edge Journal Watch — where cardiac electrophysiology meets evidence, precision, and perspective.
Hosted by Dr. Niraj Sharma, this bi-weekly podcast distills high-impact cardiovascular and EP research into clear, clinically meaningful insights. Each episode goes beyond headlines and abstracts to uncover what new studies actually mean for patient care, decision-making, and the future of electrophysiology.
What EP Edge Journal Watch stands for:
Evidence-based practice
Precision electrophysiology
A forward-thinking, edge-driven approach to how we interpret and apply data in real-world clinical settings.
Whether you’re an electrophysiologist, cardiologist, researcher, trainee, or allied health professional, EP Edge Journal Watch brings you the signal — not the noise. Expect sharp summaries, thoughtful commentary, and practical takeaways designed for the busy clinician who wants to stay ahead of the curve
This
Disclaimer:program is for educational purposes only and reflects independent editorial commentary. It is not medical advice and should not replace clinical judgment or review of primary sources and guidelines. The views expressed are those of the host and contributors.
Niraj Sharma:Welcome back to EP Edge Journal Watch. I'm Doctor. Sharma and I'm very grateful that you're here with me for this breaking news special edition. Thank you for listening. Thank you for your time and thank you for being part of this growing EP Edge community.
Niraj Sharma:We are stepping outside our usual release schedule today because CHAMPION AF is a truly landmark atrial fibrillation trial. When data of this magnitude arrive, it is worth pausing, recalibrating, and discussing what it really means for practice right now. In this episode, we are looking at a very important late March left atrial appendage closure data cluster: CHAMPION AF, CLOSURE AF, and the longer term perspective from PROGUE-seventeen. And then we will briefly discuss why the COBRA trial, even though it was not an atrial fibrillation trial, still matters when we interpret what medical therapy really means in 2026. The central question is straightforward: do these new data justify a major shift away from contemporary anticoagulation and toward routine first line percutaneous left atrial appendage closure in patients who are eligible for direct oral anticoagulants?
Niraj Sharma:EP Edge read is no. These trials do not support a broad migration away from modern anticoagulation. What they support is a much narrower and more careful conclusion. Left atrial appendage closure remains an important selective tool for the right patient, particularly when anticoagulation intolerance, recurrent bleeding, or other major treatment constraints are present, but it does not dethrone contemporary anticoagulation as the default evidence based standard. Before we go trial by trial, one key point: these studies are not interchangeable.
Niraj Sharma:They enroll different populations, use different devices, different endpoint structures, different bleeding definitions, and different medical comparators. So any simplified headline suggesting that left atrial appendage closure simply matches anticoagulation misses the design asymmetry. Let's start with CHAMPION because this is the trial most likely to trigger enthusiasm for earlier expansion of device therapy. CHAMPION AF was designed to test a bold and clinically relevant question: Could WATCHMAN FLEX function not just as a bailout strategy for patients unable to take anticoagulation, but as a genuine first line alternative to a direct oral anticoagulant based strategy in patients who were actually suitable for anticoagulation? This was a large prospective international randomized trial conducted across 141 sites in 16 countries.
Niraj Sharma:Just under three thousand patients were randomized with fourteen ninety nine assigned to left atrial appendage closure and fifteen oh one assigned to direct oral anticoagulant therapy. The mean age was about 71.7 years, about thirty two percent were women. The mean CHA2DS2 VASc score was 3.5, and the mean HasbLED score was only 1.3. That last number matters. This was not the classic extreme bleeding risk left atrial appendage closure population.
Niraj Sharma:This was a relatively lower bleeding risk, more procedure eligible group. Most patients had paroxysmal atrial fibrillation and nearly half had undergone prior atrial fibrillation ablation. So this was a modern, fairly active atrial fibrillation cohort, not simply a registry of patients already boxed into device therapy because drugs had failed them. The primary efficacy endpoint was cardiovascular death, stroke, or systemic embolism at three years, tested for non inferiority with an absolute margin of 4.8 percentage points. The primary safety endpoint was also very important to understand.
Niraj Sharma:It was non procedure related bleeding. Specifically ISTH major bleeding plus clinically relevant non major bleeding. That endpoint structure already tells you what the trial was trying to capture: the long tailed bleeding burden of chronic anticoagulation rather than the front loaded procedural liability of device implantation. Also remember that the device arm was not immediately drug free. Most patients were discharged on a direct oral anticoagulant early after implantation, and a smaller proportion were discharged on dual antiplatelet therapy.
Niraj Sharma:So this was not device only versus drug only from day one. There was an overlap period which is clinically relevant when we later talk about bleeding and periprocedural hazard. Now the results. The primary efficacy endpoint occurred in five point seven percent of the device arm and four point eight percent of the direct oral anticoagulant arm. The hazard ratio was 1.2, and the trial met its pre specified non inferiority criterion.
Niraj Sharma:So by protocol, this was a statistically positive non inferiority trial. Now, before we accept CHAMPION AF as a practice changing trial, we need to slow down and explain what the statistics actually mean, because this is exactly where the limitations become important. The trial was called non inferior, but non inferior does not mean equivalent, and it certainly does not mean superior. It simply means the device strategy was not worse than the drug strategy by more than a pre specified amount. In CHAMPION AF, that allowed margin was 4.8 percentage points for the primary efficacy endpoint.
Niraj Sharma:In practical terms, the trial is set ahead of time that if the device arm was not worse than the anticoagulation arm by more than 4.8 percentage points, they would declare success. Now here is the key statistical issue. That margin was based on an assumption that the primary endpoint rate would be about twelve percent in each group and that a relative risk below 1.4 would be clinically acceptable, but the actual event rates were much lower than expected, five point seven percent in the device group and four point eight percent in the anticoagulation group, so the observed absolute difference was 0.9 percentage points against the device. The 95% confidence interval ranged from -0.8 to plus 2.6 percentage points. What does that mean in plain language?
Niraj Sharma:It means the true difference could have modestly favored the device or it could have meaningfully favored anticoagulation. Because the upper end of that interval 2.6 was still below the preset 4.8 margin, the statistical test for non inferiority was technically met. So yes, the trial was positive by protocol. But because the real event rate was less than half of what had been expected, the study becomes harder to read with confidence. That raises concern about power, which is the trial's ability to detect a real difference when one truly exists.
Niraj Sharma:And when event rates are much lower than anticipated, one major risk is a false negative result, meaning the study may reassure us too easily that there is no important difference when in fact there might be one. That concern becomes more important when you look at the ischemic outcomes. Ischemic stroke or systemic embolism occurred in three point two percent of the device group and two point two percent of the anticoagulation group, so the most clinically feared thromboembolic outcomes numerically went in the wrong direction. By contrast, hemorrhagic stroke was not different between groups, so the argument here is not that the device clearly failed, but that the most important ischemic outcomes did not provide the kind of reassurance one would want before replacing a well established medical standard. The second major limitation is the bleeding story, because the headline safety result depends heavily on which bleeding events were counted.
Niraj Sharma:The primary safety endpoint was non procedure related bleeding over three years, and by that definition the device strategy looked better, but once procedure related bleeding was included, major bleeding was not meaningfully different between the two groups. That distinction matters enormously. If you exclude the front loaded bleeding cost of an invasive procedure, the device naturally looks safer. If you include the total bleeding burden more comprehensively, the advantage becomes much less convincing. There is another practical point here: patients in the device arm were not instantly free from antithrombotic therapy.
Niraj Sharma:Antiplatelet therapy after closure was still recommended, and early after implantation many patients also remained on anticoagulation. So this was never a simple procedure versus no drugs comparison. And that matters because the editorial makes the important point that apixaban in particular has a very favorable bleeding profile, in some settings approaching aspirin level bleeding risk and perhaps even lower intracranial bleeding risk than aspirin. So if the comparator arm included a substantial proportion of apixaban use, then the claim that device closure should obviously reduce bleeding becomes much less self evident, and CHAMPION AF does not clearly report the proportions of the specific direct oral anticoagulants used. That is a real limitation because not all direct oral anticoagulants have the same bleeding phenotype.
Niraj Sharma:The editorial also raises a trial context limitation. These results now have to be interpreted alongside CLOSURE AF, which did not show convincing non inferiority for left atrial appendage closure. Yes, the trials differ. CLOSURE AF allowed multiple devices whereas CHAMPION AF used a single platform. But CLOSURE AF was also not industry sponsored, whereas in CHAMPION AF, the sponsor manufactured the device, collected the data, and conducted the analyses.
Niraj Sharma:That does not invalidate CHAMPION AF, but it does raise the evidentiary bar and makes independent confirmation more important. And finally, there are the unresolved clinical questions beyond the primary endpoint tables. Five year follow-up still needs if one wants to argue that very long term anticoagulant exposure carries harms that a device might avoid. Patient preference clearly matters, but if indications expand, shared decision making needs to be genuinely balanced and not distorted by operator incentives. There are also biologic and device specific unknowns.
Niraj Sharma:Could occluding the appendage affect hemodynamic or endocrine function, especially in athletic patients? Do any cognitive or dementia related benefits associated with anticoagulation anticoagulation translate when only the appendage is excluded, and device associated thrombus seen in four point eight percent of image patients, along with post implant peridevice leak, still requires better understanding. So the take home message for the audience is this: CHAMPION AF is best understood as a statistically positive but clinically incomplete trial. It may justify broader case by case discussion in selected patients, but it does not yet justify declaring left atrial appendage closure as broadly as effective as conventional direct oral anticoagulant therapy for the great majority of patients with atrial fibrillation. Now let's move to closure AF and in many ways this is the harder trial to ignore.
Niraj Sharma:If CHAMPION asked whether device therapy could move earlier into a lower risk anticoagulation eligible population, CLOSURE AF asked the opposite question: What happens in older atrial fibrillation patients who are simultaneously at high risk of stroke and high risk of bleeding, the very population in whom left atrial appendage closure is so often pitched as the solution? This was a prospective multicenter randomized probe trial conducted in Germany. Nine twelve patients were randomized with the primary analysis including four forty six assigned to left atrial appendage closure and four forty two assigned to physician directed best medical care. The mean age was 77.9, nearly thirty nine percent were women, the mean CHA2DS2 VASc score was 5.2, and the mean has bled score was three point zero. So compared with CHAMPION AF, this was an older, frailer, more vulnerable cohort.
Niraj Sharma:Medical therapy was individualized. It could include direct oral anticoagulants when appropriate, vitamin K antagonists when indicated, and no antithrombotic therapy when anticoagulation was felt to be contraindicated. That makes this trial clinically messy in the way real world decision making is messy. It was not a pristine head to head single drug efficacy contest. It was a test of device therapy against physician directed care in a genuinely difficult population.
Niraj Sharma:The primary endpoint was time to first stroke, systemic embolism, major bleeding, or cardiovascular or unexplained death tested for non inferiority with a hazard ratio margin of 1.3. So unlike CHAMPION AF, where the headline result came from a protocol defined win on non inferiority, here the statistical structure was already less forgiving and the clinical population was substantially less forgiving as well. The result was clear: Left atrial appendage closure failed non inferiority. A first primary event occurred in 155 device patients versus one hundred and twenty seven medical therapy patients. The incidence rates were sixteen point eight versus thirteen point three per one 100 patient years.
Niraj Sharma:The adjusted hazard ratio was one point two eight, with the confidence interval leaning against device therapy. And the deeper you go into closure AF, the more sobering the picture becomes. Stroke counts were identical in absolute terms, 27 in each arm, so there was no ischemic advantage for device therapy. Major bleeding was not reduced with left atrial appendage closure, and cardiovascular or unexplained death was numerically higher with the device strategy. Early procedural harm remained very real.
Niraj Sharma:There were major bleeding events around the procedure, tamponade, periprocedural deaths, and even a device embolization requiring surgery. That matters because the entire conceptual appeal of left atrial appendage closure in this kind of patient is that it should, in theory, mitigate long term bleeding liability. But if early procedural vulnerability is high enough, that theoretical advantage can be neutralized or even reversed. This is why closure AF is such an important reality check. If there is a patient group in whom the bleeding avoidance thesis of left atrial appendage closure should have looked strongest, it is this one: older, higher stroke risk, higher bleeding risk, and clinically fragile, and yet the trial did not deliver a meaningful win for device therapy.
Niraj Sharma:It is also a reminder that the phrase high bleeding risk is not itself a therapeutic argument for a procedure. The relevant question is net clinical benefit after you include the upfront procedural penalty, the heterogeneity of background medical care, and the reality that some patients may still derive acceptable outcomes with carefully tailored medical management. The EP Edge Take is straightforward: Closure AF materially raises the evidentiary bar for any broad enthusiasm around left atrial appendage closure. In elderly, high risk atrial fibrillation patients, optimized physician directed medical therapy remains a very credible default strategy. Device therapy cannot be assumed to provide an automatic net clinical advantage simply because bleeding risk is high on paper.
Niraj Sharma:So if CHAMPION AF was the headline generating trial, closure AF is the counterweight that should keep the field grounded. And now that brings us to PROG-seventeen, which remains the most supportive randomized evidence for selective left atrial appendage closure, but still requires careful interpretation. PRAG-seventeen still matters because it is the most important randomized counterbalance to an overly negative reading of left atrial appendage closure. This was an investigator initiated multicenter randomized non inferiority trial of four zero two high risk atrial fibrillation patients assigned one to one to left atrial appendage closure or direct oral anticoagulant therapy. The mean age was 73.3, the mean CHADS VASc score was 4.7, and the mean HasbLED score was about three point zero.
Niraj Sharma:Entry required a genuinely high risk profile, such as prior bleeding requiring intervention or hospitalization, prior cardioembolism despite anticoagulation, or a combined moderate to high stroke and bleeding risk phenotype. So unlike CHAMPION AF, this was much closer to the kind of patient in whom a device first conversation may naturally arise. Now, one of the most important methodological features of PRAG-seventeen is the comparator. The direct oral anticoagulant arm was overwhelmingly apixaban based. Ninety five point five percent of patients in the drug arm received apixaban, with only a very small number receiving dabigatran or rivaroxaban.
Niraj Sharma:That makes the trial highly relevant, because it is not comparing device therapy against an abstract anticoagulant class it is, in practical terms, comparing against an apixaban dominant strategy, which is exactly the real world benchmark many clinicians would consider the most difficult medical standard to beat on bleeding. The primary endpoint was deliberately broad and balanced. It included stroke or transient ischemic attack, systemic embolism, cardiovascular death, clinically significant bleeding, and significant procedure or device related complications. That broad endpoint is useful because it captures both the benefits and the costs of device therapy instead of artificially isolating only one side of the trade off. The initial annual primary endpoint rates were ten point nine nine percent with left atrial appendage closure and thirteen point four two percent with direct oral anticoagulants.
Niraj Sharma:The trial met non inferiority. All stroke or transient ischemic attack was essentially identical between groups. Clinically significant bleeding numerically favored left atrial appendage closure, though it was not significantly different in the initial report. What became more interesting in PROG-seventeen was the longer term follow-up. At four years, non inferiority persisted, and non procedural clinically relevant bleeding became significantly lower with left atrial appendage closure.
Niraj Sharma:That is the strongest randomized signal supporting selective device use in high risk patients where long term bleeding reduction matters, but this trial also should not be oversold. The ischemic protection was not superior. The advantage was largely a late non procedural bleeding signal, not better stroke prevention. The device arm was heterogeneous, including amulet, Watchman and Watchman FLX. So this was not a single platform contemporary device test in the way CHAMPION AF was.
Niraj Sharma:Major left atrial appendage closure related complications still occurred and remind us that the early procedural cost is real. That combination of strengths and limitations is exactly why PRAG-seventeen deserves a nuanced reading. It is the best randomized evidence supporting selective use, but it is not evidence for superiority, and it is not evidence for universal substitution. It supports a narrower claim, that in carefully chosen high risk patients, especially when non procedural bleeding reduction over time is clinically meaningful, device therapy can be a reasonable alternative to an apixaban dominant anticoagulation strategy. So the EP Edge take for PROG-seventeen is that it supports carefully selected use, not automatic substitution.
Niraj Sharma:It tells us that in the right high risk patient, especially when long term non procedural bleeding is a dominant concern, left atrial appendage closure can be a reasonable alternative, but it does not justify a default move away from anticoagulation for the broader atrial fibrillation population. Now where does COBRA fit into this discussion? Importantly, COBRA was not an atrial fibrillation trial. It randomized patients with acute venous thromboembolism to apixaban or rivaroxaban, so it cannot be used to recalculate atrial fibrillation specific trade offs or to prove that one left atrial appendage closure comparator arm was unfair. But it does teach a very important lesson: direct oral anticoagulants are not a monolith.
Niraj Sharma:In COBRA, clinically relevant bleeding at three months was lower with apixaban than with rivaroxaban, while recurrent symptomatic venous thromboembolism was similar. That matters here because any device strategy claiming a bleeding advantage over direct oral anticoagulants should be judged against the actual anticoagulant mix used in the trial. And CHAMPION AF does not provide drug specific bleeding or ischemic outcomes by anticoagulant agent in the main publication or source material. That omission is not trivial. If a comparator arm is apixaban heavy, the practical bleeding hurdle for device therapy may be lower than a generic direct oral anticoagulant label implies.
Niraj Sharma:If it is more rivaroxaban leaning, then a class level summary may overstate what left atrial appendage closure is really being compared against in day to day practice. So better comparator transparency is not an academic detail. It directly affects how we interpret the claim bleeding advantage, And this is where the EP Edge concern becomes especially important. My concern is not simply how CHAMPION AF is read by trialists or statisticians, but how it may be translated into practice by the broader field. The real risk is indication creep.
Niraj Sharma:In other words, a statistically positive but clinically nuanced non inferiority trial may be interpreted too quickly, too broadly, and too favorably, leading to more devices being implanted before the results, methodology, and limitations have been fully understood. That would be a mistake, because CHAMPION AF did not show superior ischemic protection. It showed a protocol defined non inferiority result for the primary composite endpoint, while ischemic events numerically moved in the wrong direction. And its main safety advantage was heavily influenced by endpoint construction, especially the focus on non procedure related bleeding rather than the total bleeding burden including the procedure itself. So if this trial is reduced to a simple message that left atrial appendage closure is now just as good as anticoagulation, that would be a substantial over reading of the data.
Niraj Sharma:The practical concern is that once a landmark trial enters the clinical bloodstream, nuance is often the first thing to disappear. Busy clinicians may hear non inferior and interpret it as interchangeable. Patients may hear device and assume freedom from long term risk. Operators may feel greater momentum to offer closure earlier in the disease pathway. But that is exactly why methodology matters.
Niraj Sharma:We still have to ask who was studied, how the endpoints were defined, what was counted, what was excluded, what drug strategy the device was compared against, and whether the trial truly supports broad first line substitution in a modern apixaban era practice environment. So how should this change practice right now? Routine first line left atrial appendage closure for broadly direct oral anticoagulant eligible atrial fibrillation should not become the new default based on CHAMPION AF. For older, frailer, higher bleeding risk patients, closure. AF reinforces that optimized physician directed medical therapy remains not only viable but often preferable.
Niraj Sharma:Left atrial appendage closure remains reasonable for carefully selected patients with recurrent major bleeding, true anticoagulation intolerance, inability to sustain long term oral anticoagulation, or other compelling individualized constraints, but selective use is very different from indiscriminate expansion. So the EP Edge bottom line is this: the danger now is not under reacting to CHAMPION AF, the danger is over reacting to it. If this trial drives broader implantation before the field has carefully absorbed the statistical meaning, the endpoint structure, the comparator ambiguity, and the unresolved clinical questions, then indication creep may outpace evidence. And that is exactly what we should avoid. Very briefly, here is the recap: CHAMPION AF was statistically positive but clinically mixed, with ischemic events numerically higher in the device arm and the main bleeding advantage driven by non procedural bleeding definitions.
Niraj Sharma:AF is the harder signal to ignore because in older high risk patients, left atrial appendage closure did not reduce stroke, did not reduce major bleeding, and did not beat optimized medical therapy. PROG-seventeen remains the best support for selective use in carefully chosen high risk patients, especially when long term non procedural bleeding reduction matters. And COBRA reminds us that the choice of direct oral anticoagulant matters when we interpret what medical therapy really means. All references and graphics are available on the LinkedIn newsletter, EP Edge Journal Watch, as well as on Substack at eph.substack.com. Thank you again for listening, thank you for your time and thank you for being part of EP Edge.
Niraj Sharma:Bye for now, take care and I'll see you till the next episode.