Jun 19 2026 This Week in Cardiology

First up, let's do some listener feedback on my story of declining V T rates last week. And I have to say that this podcast has very sharp listeners. Doctor Yannick Kirst or Kirste. I'm sorry about the pronunciation from Eslegen, Germany writes to me regarding my coverage of the fifteen year declining rate of VT in patients with ICDs in two centers in Europe. I was struck by the substantial decline over time of VT therapies while VF therapy remained low and unchanged.

And I liken this to the heart failure trial this paper in the New England Journal of Medicine showing a temporal decline of sudden death in heart failure trials. But doctor Kirst writes that another potential cause of declining VT intervention in ICD patients may be the gradual change in more permissive programming of after the 2012 Made It RIT study, which stunningly showed that permissive programming resulted in both a reduction in shocks, inappropriate shocks, and mortality.

Now if you're not totally familiar with Made It RIT, it's worth a look as it was one of uh seminal trials and it was ten out of ten on the bold scale. Patients were with ICDs were randomized to three different strategies of programming, a conventional out of the box type, a high rate group, and a delayed type.

The conventional arm had two detection zones, for heart rate one seventy to one ninety nine. There was a two point five second delay and the second zone of greater than two hundred beats per minute with the second delay before ATP. Patients in the high rate group received a device programmed to monitor only zone between one hundred seventy and one hundred ninety-nine per minute and to a therapy zone beginning at two hundred beats per minute after a two point five second monitoring therapy.

And patients in the delayed therapy group received a device program to three detection zones, one at 170 to 199, uh with rhythm detection on and a sixty second delay before initiation of therapy. There was a second tachyrhythmie detection zone beginning at two hundred. with rhythm detection on and a twelve second delay before therapy, and a third zone at two fifty beats per minute, which was VF or higher with a two point five second delay before initiation of therapy.

And during an average follow up of one point four years, high rate therapy and delayed therapy were associated with a seventy nine percent and seventy six percent reduction in the first occurrence of inappropriate therapy respectively.

and there were reductions in all cause mortality of fifty five percent for high rate versus conventional and forty four percent reduction for delayed versus conventional. I will tell you that at the time this was a truly shocking result, because few of us thought that being this permissive with V T detection would work. We thought collectively that VT was terrible for patients and the earlier we intervened was better and made it RIT completely exploded that convention or therapeutic fashion.

And I think the trial taught us that A groupthink should always be tested in R C T's and B A lot of V T goes away on its own and C, high rates are often non-VT rhythms like AFib or sinus tachycardia. So doctor Yannick Kurst is correct some of the decline in V T noted in the Basel led study may have been due to permissive programming.

And good on him for bringing this up as it is always good to revisit seminal trials and boy was made it RIT worth a re looked. This New England Journal of Medicine twenty twelve.

Alright, now the vast majority of this podcast is going to be on the long-term data on Taver versus Saver that just came out in Jack. So the conundrum faced by patients with severe isolated AS without a bicuspid valve continues with this newest long term data and Jack has published ten year data on the intermediate risk partner two trial, that's balloon expandable trial, as well as an observational comparison between the newest balloon expandable valve called Sapien 3.

versus the balloon expandable valve that was used in the trial partner to which is no longer used called the Sapien XT. And this is complicated stuff. Let me first tell you the results, then rehash the recent six to seven year data on the self expander valve that was published this winter, and then some comments on the totality of the evidence and the difficulty of the decision.

But before any details, let's set out that there is a tension between having long-term data and having new iterations of devices. A trial starts out with the valve or stent or pacing device or whatever that's available at that time. The trial takes a couple years. uh during or after the trial, the biomechanical engineers d are not sitting on their butts doing nothing.

they're iterating and new iterations of the valve or stent or device come out. And that iteration is probably or clearly better in some way. And then how do we translate evidence generated from old devices to the new devices. And I'm asking, I don't know. So now to the data. Partner two randomized nineteen hundred patients with severe AS and intermediate risk between 2011 and In 2013, and the trial was planned for five years. So to obtain tenure data, patients had to reconsent at five years.

And the authors did something called the vital status sweep, and this was implemented to improve data completeness for all-cause mortality. So it was about a thousand patients per group and three quarters received transfemoral TAVR and twenty four percent had transapical. The mean age of patients was old at eighty two years and almost half were women The S T S surgical risk was about five point eight percent and the results at ten years

They had some data on about ninety percent of both groups. All cost ten year mortality with the vital status sweeps was eighty six point one percent TAVER, eighty two point eight percent after surgery, That three point three percent increase in absolute terms had a hazard ratio of one point one three, which was statistically significant. When stratified by excess root, however, rates of all cause mortality for tavern surgery in a transphemoral group were similar.

83.9 versus 82.1, that's not significant, whereas mortality was higher with Taver in the transacal group, 93.2 versus 85.1, and the P for interaction in that subgroup was positive. The f the key finding, I think, though, was this. Cumulative incident rates of aortic valve reintervention at 10 years were 6.3% for Tavra and 1.6% for surgery, and that's highly significant.

Now my comments when you look at the overall data, Tavi looks worse, higher mortality and about four X higher rate of valve intervention interventions. However There are two big caveats, and I mean big ones. One is that most of the harm in the trial came in a transapical root. And when you look at transfemoral mortality is still higher, but it's so close as to not be significant.

Now the problem with hanging the excess mortality on the transavical route is not only that it is an underpowered subgroup susceptible to noise, But the transapical root requires a decision, and many things go into that decision. So these patients could be different, say there's more atherosclerosis, and these differences may drive the outcomes rather than just the uh valve and uh surgical approach.

But the biggest caveat I think is that partner two used the sapien XT valve which has been replaced. now with the Sapien three valve. Now I don't have enough content expertise to explain the exact differences, but suffice to say, implanters obviously thought it a bet it's better because sapien XT is no longer taken off the shelf and these valves are done with uh sapien three.

Now another paper in Jack, which I've linked to, is a reasonable effort to assess sapien three performance, the new valve. A large group of authors, some from Partner 2, took about 900 patients who had the new sapien 3 valve for severe AS and intermediate risk. in the partner two sapiens three intermediate risk registry and compared them to a matched group in the actual trial who had surgical aortic valve replacement.

So they used propensity matching to approximate baseline characteristics in these two non randomized groups, one from the registry, the other from the trial. So this is a observational comparison study.

The registry, like the trial, was designed as five year studies, so again they had to reconsent patients and then use vital status sweeps to get mortality data. Notably, and I'll just cite the paper Here, quote, the trial sponsor, Edwards Life Sciences, owned and maintained study data, performed statistical analysis in collaboration with the authors, and provided editorial support under the direction of the authors, end quote, no comment, just I read that.

Results after propensity score at matching there were seven hundred three eighty-three patients in each group and in the matched cohort 367 Taver and 378 surgical patients completed. The five-year follow-up, and of these, 40% and 59% respectively did not reconsent for extended follow-up. So at ten year follow up there were fifty nine TAVER and fifty-four surgical patients who were alive and on study.

And four hundred and nine Taver and three hundred and thirty-six surgical patients who died, resulting in sixty percent of the Taver group and fifty-four percent of the surgical patients. with a known vital status for all-cause mortality analysis. Implementation of the vital status sweep improved the completeness of mortality data to 88% and 90% for the two for the TAVER and surgery groups at 10 years.

So in sum at 10 years, all cause mortality with this vital status sweep was 83.4% for Tavern, 82.3% for surgery. And that hazard ratio is one point zero one and obviously not statistically different. Interesting was that death with the vital status sweep was similar between TAVR and surgery for the transfemoral and transapical roots in this non randomized comparison.

Uh reintervention in this study was also lower and no different, two point zero versus one point nine percent in the Saver in the Taver versus Savour arm. Echo data at ten years was available in in only sixty patients and it was uh valve function seemed similar in terms of gradient and effective orifice area, though aortic insufficiency was uh two times more in the tavern arm. And I wouldn't make much of the echo data since it's only available in sixty patients.

The authors were very fair about the limitations of this analysis, which were numerous, right? It's non-randomized. And of course, there could be unmeasured confounders. It of course only uh uses balloon expandable valves. Uh reconsenting uh resulted in asymmetrically and potentially non-random loss of follow-up, which was greater in the surgical patients. And the vital status sweep, which is better than nothing, cannot address missingness of data for other endpoints.

and there was no central adjudication of clinical events after five years. But on the other hand, The registry was started within a year of the completion of the trial. The patients were similar in characteristics and taking a Bayesian view. you'd expect a new valve to be better in many ways, including the fact that in the registry and uh including the fact that in the registry and now it is normal to use uh CTA to guide valve choice, which was not typical in the main trial.

Now doctors David Cappadano and Camille Bujack have an excellent accompanying editorial. Uh they have a figure. Figure one is a table of five studies that have gone on for more than six to seven years. Other than these two that I just mentioned, let me just mention the other three.

Uh Notion, which is an R C T in Denmark. Small n small numbers of patients, two hundred and eighty, randomized one to one, with either a self expanding core valve or SAVR These patients were seventy nine years old but had a lower STS score of three percent versus five point three percent in partner two, and the main results of Notion

was that after ten years the risk of the composite outcome, composite of all cause mortality stroke or MI, was sixty five point five percent in both arms. So obviously no difference. Now there's the Partner three trial of low surgical risk repa patients, has reported outcomes uh at seven years last October in the New England Journal of Medicine. There were slightly more than 900 patients randomized.

Uh recall that partner three was the first Haver trial to mysteriously add rehospitalization in the primary endpoint as all the other trials had stroke and death. At seven years the incidence of the primary endpoint

outcome of death stroke rehospitalization was thirty-four point six percent Tav, thirty-seven point two percent surgery, which was not significantly different. However, for death It was higher with Tavern, nineteen point five versus sixteen point eight percent, though not statistically significant. Stroke was about the same. Uh reinterventions was about the same. And slightly more than half of the interventions were surgical explants.

The fifth trial that was mentioned in the editorial that has reported long-term results is the Evolute Low Risk. And here there's six to seven year results. This is the self-expanding valve. There were about fifteen hundred patients. Death again was higher in the Taver arm, but not significant. It was twenty-three point three versus twenty point two.

Reintervention though was statistically higher in the ta taver arm, 9.8 versus 6%. And most reinterventions in evolute low risk were surgical explants. The editorialists note different patterns of valve failure after five years. In evolution risk, aortic regurge predominated, whereas in partner 2A, stenosis was slightly more common, especially beyond five years. And of course these patterns mirror observations in surgical valves with re stenosis more typical of bovine.

pericardial valves and regurgitation more frequent in porcine valves. The modes of intervention were also different. In Partner two A nearly all patients who had reintervention uh had Taver in Taver. Whereas surgical reintervention predominated in the evolute low risk group, and that likely reflects differences in patient age, the risk of coronary occlusion relative to prosthesis type and mechanism of failure. The last paragraph of their editorial is good and I'll just cite it.

Uh quote the ten year partner two A results and the accompanying observational analysis, despite the challenges of long term follow up, represent an important contribution to the evidence informing the lifetime management of aortic stenosis and at the same time make the forthcoming longer term outcomes of partner three and evolute low risk even more awaited.

They do not diminish the transformative impact of Taver, but serve as a timely reminder that durability remains a central consideration as the field enters its second decade. End quote.

Let me add one more commentary that was published. Victor Diane, Sanjay Call and Benoit Shaw have written an editorial in European Heart Journal this past week. It's entitled Taver versus Saver Age versus Lifetime Perspective. And they push back strongly on the recent ESC guidelines which Which have lowered the default age threshold for TAVI from age 75 to 70. This is now a class one.

Level of evidence A, the strongest type, so obviously the guideline writers have confidence in Tavi as first line therapy for younger lower risk patients. Now, the argument of the editorialists Diane, Call, and Shaw, which is well argued, I'd say, make the case that for Saver is the best first-line choice. First, they cite five lines of evidence. One, partner three at seven years.

Death or disabling stroke was thirty percent thirty one percent higher in a tavern arm. Two inevolute low risk, non inferiority may have. been met, but a non significantly higher rate in death or disabling stroke favored Savour. This was observed and the reintervention rate was sixty eight percent higher in a Taverarm. Evidence three, in the Notion two trial, which included younger patients at low risk, mean age seventy one,

This trial failed to establish non inferiority of TAVI relative to SAVR for death, stroke, or rehospitalization. Evidence four. UK TAVI trial, which enrolled older patients mean age eighty one, at Low to Intermediate Risk reported that TAVI was not inferior to surgery regarding all cause mortality at one year. However, unpublished five year follow-up uncovered a seventy eight percent higher risk of stroke with TAVI.

And f evidence five, a Bayesian hierarchical meta-analysis studies with at least five-year follow-up showed TAVI was associated with a higher risk of death and stroke in low to intermediate risk patients. And so robust RCT data for low risk patients only extend to seven years. Seventy years old has a much greater life expectancy and using life tables up to fifteen years.

So the art the editorialists argue there's not enough data. And while some of the data they were say is reassuring, there are signs too of a catch up where death and disabling stroke and reintervention seem to be going the wrong way with Tavi. And so they make the argument that SAVR first, four young people with greater than ten years lic life expectancy, and they give four big reasons. One is it preserves coronary access. Two

Saver optimizes hemodynamics for future interventions. For instance, surgical aric valve allows for annular or root enlargement, enabling the placement of larger prostheses, which then decreases the risk of prosthesis from the patient mismatch and higher gradients during future TAVI and SAVIR interventions. Three third reason, there's flexibility in reinterventions. Patients starting with TAVI may be limited to TAVI and TAVI.

and if unf unfeasible TAVI explant becomes necessary and that's a procedure with very high operative mortality, up to seventeen percent in some studies. But even Tavi and Tavi is not nothing as the authors cite a sapien three registry with a one-year mortality of approximately twenty percent.

And the fourth reason why they would choose SAVR first is procedure related complications. While atrial fibrillation or bleeding after surgery are transient complications, Higher pacemaker rates and perivalvular AI from Taver are often persistent and have cumulatively negative effects.

Okay, so my final comments are that first it's exemplary that we have this tenure data and that we're having this discussion. We we don't have this uh uh kind of data with many other cardiac interventions. While vital status sweeps are not the same as staying in a trial, it's pretty good. And as I have said, for the older patient who has high surgical risk, Taver is amazing. It's been a great invention. The question is younger people who will likely outlive their first valve replacement.

Now, the main tension for us here on this evidence-based medicine podcast is the is the tension between long term evidence and iteration. Now partner two trial looks worse for Taver, but it was with a sapien XT valve. Now we implant sapien 3 and it's a better valve. So does the trial data apply to the new valve? I don't think it does. David Cohn uh wrote to me on X that his favorite aphorism for this issue is, quote, you can either have long-term outcomes data or new technology, but not both.

To me though, altogether, Savour does look slightly better over the long term, though it's close. Doctors Diane Call and Shaw make a compelling case for Saver First in Young Patients, but honestly. Patients at age 70 face a tough decision. Taver is clearly going to be easier on them. Yes, it may not last as long, but patients may say, Who knows what happens to me? I could die of cancer in five years, and then I've avoided six months of that five years during having a hard recovery from surgery.

And I think patients feel differently. I feel differently. I'm having dental surgery this afternoon and I've chosen to do an easier procedure knowing that two to five to ten years down the road I may need to do another surgery. But I may not be here then, so I'm going with the easier surgery now, rather than the more complete surgery. And some patients with aortic stenosis may want SAVR, knowing that longer term there'd be better results and surely more options for reintervention.

Specifically, they may want to avoid a TV Tavi explant, which is terrible. But Tavi now is an understandable choice. I also suspect that anatomy matters too. Now everybody gets a CTA. Maybe the presence of ideal anatomy for TAVI, and the odds for a great result are high. Now trials give us average effects, but with modern imaging and modern valve techniques and an utterly biased multidisciplinary team,

may recommend the Tavi over SAVR in a young person based on anatomy. The problem though is that I'm not sure there are many unbiased multidisciplinary teams in the US. I'm not even sure we have MDT teams in the US. I could be wrong, but what I saw in the UK and in Denmark were doctors that make the same amount of money regardless of the number of procedures they do. And to me that reduces the risk of overuse of Taver.

In our system in the US, and perhaps in Germany as well, there may be outside pressures to do Tavi when SABR would be better. Plus we know that cardiologists control the patients, so of course there is that bias. Now we will know more when the low risk trials uh get out here to ten years.

but I predict a clear best choice in all patients is going to remain uncertain and largely dependent on getting wise advice from an unbiased multidisciplinary team and strongly, strongly uh gonna be preference sensitive decisions from a patient point of view.

All right, next topic is the preview valve study. Now did you ever wonder what the prevalence of valvular heart disease was in the population? Well, a group led by doctor David Cohn at Columbia University did, and their study was unique and novel and worth a mention. The authors chose a random sample of sixty five to eighty five year olds who had filled a prescription at C V S or Walgreens, and they invited them to participate in the study via email or text or direct main mail.

The sampling strategy was based on zip code and demographics and was used to identify potential candidates. From among the pool of eligible CVS and Walgreens pharmacy customers. Oh, individuals who were interested in participating were referred to an online portal and then they had to go through and complete an electronic screening that verified their eligibility.

and uh solicited basic demographic data. Then, an independent research organization contacted eligible individuals to obtain informed consent and scheduled a study procedure. Patients were compensated with a twenty-five dollar gift card after they completed the in-home study procedures.

But get this, all study procedures to detect the valvular heart disease were performed during a single study at the participants' home, where they had a comprehensive two D transthoracic echo uh by a certified cardiac sonographer. The primary endpoint of the study was the prevalence of moderate or greater valvular heart disease, which was defined by more than moderate heart disease of any valve.

And so they sent invitations to more than six hundred and fifty thousand individuals and offered the opportunity to participate in this study. Of these, about 8,200 or 1.3% visited the study portal. Ultimately thirty-three hundred uh completed the informed consent process and agreed to enroll, and three thousand of them had completed the study procedures. Three thousand patients had in-home echo.

That's pretty wild. And the sample included individuals from all fifty states and was racially and ethnically diverse. The median age was seventy-one years and fifty-seven percent were female. Overall prevalence of valvular heart disease, guess. Guess what it was? 8.2%. And of course it had an age gradient, right? So 5.1% for 65 to 69 year olds, and then gradually getting more up to nearly 15% for 80 to 85 year olds.

The weighted prev prevalence of greater than moderate valvular heart disease by valve type was zero point seven percent for aortic insufficiency. Three point one percent aortic stenosis, two point zero percent MR and three point seven percent T R. And so the authors applied this prevalence data to the population of the US and estimated that four point seven million U.S. adults have more than moderate valvular heart disease.

Also notable and I actually didn't know this was a lower incidence of valvular heart disease among black individuals, especially AI and AS, which Of course, helps explain why some studies show that black patients are less likely to have Taver or Savour. So my comments, this is a neat study, right? It avoids the bias of studying patients who come to a clinic. Now at first I thought that it might be a biased assessment because patients who agree to have an echo and fill out all this paperwork

uh might have worrisome symptoms and thus these estimates might be an overestimate. But doctor Cohn messaged me that uh they doesn't think that's the case because their K C C Q was off the chart uh for good numbers. So these were happy healthy patients. I still wonder though about sampling. Now another lesson of course from this is that accurate population level prevalence estimates can actually be obtained

pretty easily with these new age tools like databases and modern telecommunications with text messages and stuff. I think that's pretty nifty. Recall that the Dan Kavas screening trial in Denmark enrolled with invitations via national email. That said, I do worry that studies such as these could create disease recognition campaigns. which are a modern day scourge. I despise disease creation because ultimately it is to drive profits. Now the authors don't do this. It's not them.

So I don't wanna implicate them. It's just that I can see a company that makes valve intervention tools creating advertisements describing the high prevalence of TR or MR in the community. If this would happen, our profession should resist it strongly. But finally, finally, none of this should be necessary, right? Because if patients come in for health checks,

and patients have health checks. I mean everybody has a health check. You might go to the th an urgent care center for an ear infection, for instance. A clinician who does a basic exam should be able to hear moderate to severe valvular heart disease. You would think that if someone had moderate to severe MR and they had a clinician listen to their heart, it would be detected.

But sadly, I'm sorry to say that A, fewer patients even get their heart examined, and B, those that do could easily have their moderate to severe AI or MR missed. But I just wanna say still to this day in twenty twenty six, I still listen to the every to the heart of every patient I see and I actually do listen for stuff. So I still recommend that when you listen to the heart, that you actually listen in a quiet room

and just listen for valvular heart disease. Sometimes you'll pick it up. Alright, next topic is that CMS in the US has proposed new coverage for Taver. This proposed decision memory was released and is open for comment until july fifteenth, and the big changes are that it will expand coverage for Taver to Asymptomatic severe AS, of course, with coverage with evidence development, and I'll talk about that. And it will also allow for Taver to be done with one Taver operator.

It also revises coverage criteria for pre-procedural patient assessment and comments on hospital procedure volume requirements. Now the coverage with evidence I like. This is a good thing. And it includes seventeen requirements for the studies to be conducted, which makes it look like they're being really strict about what kind of studies can be done. But sadly not one of these seventeen requirements is that they d

do an RCT, which is of course the best way to study things. The American Association of Thoracic Surgeons are opposed to this new coverage and so am I. The AATS writes to its members that, quote, we believe the current framework represents a deeply problematic outcome that eliminates mandatory in-person surgeon evaluation of patients. Undergoing consideration of TAVR or intraprocedural surgical expertise. This potentially puts patients at risk.

The AATS maintains that the following positions which were not included in the proposed framework are of the utmost importance for patient safety. One, dual specialty operator standards. The perfor the proposed framework eliminates joint operator requirement, and we ask that the coverage decision should explicitly note that dual specialty operators consisting of both a cardiac surgeon and an interventional cardiologist are optimal for patient care.

Number two, requirements for new programs The AATS asks for the first year of operation any new Taversite must be required to utilize a dual specialty operator composition

consisting of both a cardiac surgeon and an interventional cardiologist. And the third uh position is a coverage with evidence development. We asked that Tabber be covered Under coverage with evidence development for the following indications or conditions and there's multiple ones that they want at coverage with evidence, and that is asymptomatic AS, bicuspid anatomy. Moderate aortic stenosis, valve and valve taver, aortic insufficiency,

And low risk aortic stenosis patients. They also want coverage with evidence for new TAVER programs for hospitals without Taver experience. And they would like coverage with evidence for newly FDA approved Taver valves.

And finally they write, we ask that for the above C E D indications and conditions, the heart team and hospital must participate In a prospective national audited registry that consistently enrolls Tabber patients, follows them for at least one year, and reports on the relevant clinical outcome. But I I actually would go farther than this. I think this is great. It's obviously what we should do, because this is new and it has extreme bearing on low risk patients and

We have a chance to mandate collection of data and we should do it. But I would even go farther and I would add that the evidence base for allowing aortic valve replacement with Taver in asymptomatic AS is terrible. It's terrible. I was asked this winter by the British Cardiac Interventional Society to present there on the three worst trials in interventional cardiology that affect practice still, and I chose Early Taver to lead the list. Oh early Taver, which is the underpinning uh R C T

This was an RCT comparing early TAVR to clinical surveillance in patients with severe asymptomatic AS. And of course, that trial included an endpoint of death, stroke, and rehospitalization for cardiac causes. Which which sounds kinda normal on the surface if you just let it buzz over you.

and death and stroke and a trial were similar, but the entire positive end point was driven by rehospitalization, the majority of which was for conversion to TAVR, which is ridiculous because just months before these patients passed a stress test to show that they were asymptomatic. This was subtraction anxiety run amok, meaning patients had severe AS, one group got fixed and the other group didn't get fixed, and the group that didn't get fixed just said, okay, I want to get fixed.

So it was a trial designed by industry to be positive and sell more valves, which it is likely to do based on this new CMS proposal. So I'm against it. I'm against even having uh it as a uh uh indication asymptomatic AS. So that's it for this week in cardiology. I know I went a little long this week because of all this data on Taver and Saver. Um

And so if you have comments, please let me know. And remember, if you like this podcast, take the time, give us a rating, write us a review, and if you don't agree with something I say, uh write me a note and we'll do some more listener feedback and I'm really learn a lot from listener feedback. So thank you very much. And until next week, this is John Mandrola from the Heart.org Medscape Cardiology.