Mar 13 2026 This Week in Cardiology

You're listening to This Week in Cardiology from org Medscape Cardiology. Any views expressed are the presenter's own.

Hi everyone, this is John Mandrola from the Heart.org Medscape Cardiology, and this is This Week in Cardiology for march thirteenth, twenty twenty six. This week, some listener feedback, beta blocker use in non-obstructive hypertrophic cardiomyopathy. A head-to-head Apixaban versus River Oxaban RCT? A observational study of deltazam versus metoprool when combined with direct acting oral anticoagulants, and accuracy of smart watches for atrial fibrillation.

First some feedback. Dr. Daniel McIntyre writes to me from Sydney, Australia about my coverage of the emphatic trial. which tested guideline directed medical therapy in patients who had heart failure and advanced cancer. Not surprising was that medical therapy had no effect on a primary hierarchical endpoint that I actually loved and I want to repeat.

This was a combination of one, days alive and able to wash oneself, two, ability to walk four meters, and three, a patient reported global assessment of just well being. I highlighted the study because you would not believe how many very ill patients, not only with cancer, but other conditions like dementia, COPD, end stage renal disease, frailty. Who, instead of prescriptions for cheeseburgers and ice cream, get for drug therapy for heart fire?

doctor McIntyre read the study carefully and noted that most patients in the study with heart failure actually had HEF PEF, and only one of the selected medicines has the backing of an RCT and that's epiglyflosin. But I would note that Sacubitral Valsartan came close in Hefpeh. But McIntyre writes that this is important because in this population his approach has been to prioritize diuresis for symptom relief over ras inhibition.

I was a shame to see, he said, no patients randomized to MRA or even ferrosamide. He notes that only forty percent of patients in both arms were on these medications, and I agree with this concern about MRAs. I think MRAs are highly effective. And I leaned towards the sub-analysis of the Topcat trial, which excluded Russia and Georgia and showed a large benefit for sperolactone in patients with HAPPES. I call MRAs or sperno lectone my secret weapon for these patients.

But in a patient with advanced cancer or other severe diseases, simply preventing congestion with lupe diuretics might be the least disruptive and most effective thing.

All right, first topic is beta blockers and non-obstructive hypertrophic cardiomyopathy. Now when I discussed this study of beta blocker versus calcium channel blockers in patients with non-obstructive HCM, with a number of my friends in academia, they said, John, we've known for years that beta blockers impair exercise capacity.

Which I thought was kind of funny because almost every patient I see who has any degree of hypertrophic myocardium of any sort is on a beta blocker. It's like the sky is blue and patients with HCM take beta blockers. And guidelines still give beta blockers or calcium channel blockers a class two A recommendation in patients without obstruction.

Yet there are no RCTs in non obstructive HCM. Beta blockers and calcium blockers were established in patients who had obstructive HCM, and this is likely a very different condition. Well, Jack has published a very nice crossover double-blind RCT of bisoprolol versus verapamil in patients who have non-obstructive HCM. A group in Ahuos, Denmark, conducted the mainly exercise capacity trial. Now, whenever I mention Ohu's Denmark,

I fondly remember I visit there a couple of years ago when a group of very young people took me on an eighty kilometer bike ride in the hills and along the sea. And it turns out that you just can't sit on wheels in Denmark. These This is a very hilly country, at least in the north. Now, the goal of this trial was to assess the peak oxygen consumption, PVO2, as well as exercise response, symptoms, biomarkers, and structural changes on the two drugs.

Thirty two patients. Now that sounds small but Thirty-two patients were randomized and they underwent a crossover, so each patient was actually their own control. There were one-month sessions on bisoprolol, and then they had cardiopulmonary exercise stress tests. Then one month washout, then one month on placebo and testing, then one month on verapamil, and then the stress test, and each patient served as its own control.

Patients were blinded, investigators were blinded. This is actually a three stage trial and this report is stage one, the exercise part. Stage two and three will look at myocardial remodeling and arrhythmias respectively.

These were young patients, 54 years old, 34% female, and half had a pathogenic mutation. About a quarter had an ICD. Okay, the main finding was that the PVO2 was twenty five point seven milliliters per kilogram per minute with basoprolol, twenty eight point two with verapamil, and twenty eight point seven with placebo.

So when comparing the differences of the combination, basoprolol versus verapamil, ver basoprolol was minus one point eight milliliters per kilogram per minute, which was statistically significant. The soprolol versus placebo was even worse, minus 2.5, and here the p-value was more significant. Virapamil versus placebo was not significantly different.

Peak heart rate was lower with basoprolol. Neither basoprolol nor verapamil changed the oxygen consumption at anaerobic threshold or the minute ventilation over CO2 production slope. Basoperol worsened the KCCQ questionnaire while verapamil did not. Basopolol increased the BNP while verapamil decreased the BNP. My comment.

I see this as an elegant demonstration that beta blocker clearly worsens exercise capacity and has negative effects on other outcomes relative to varapamil and placebo, such as quality of life and BNP. Now you could go on and on about why beta blockers worsen exercise capacity, like things like uh blunting of the heart rate, or maybe it's the negative anotropy, but I don't think the why question matters.

Here, you have evidence that is clearly worse, and there's no other data to support beta blocker use in nonobstructive hypertrophic cardiomyopathy. So, in the absence of reasons to use beta blocker, like atrial fib with a high ventricular rate in these patients with non-obstructive HCM. I think this study strongly suggests that we should avoid beta blocks.

Now some may say we already knew this and this study was not necessary, and I don't know, I don't think so. I see tons of patients with HCM who take beta blocker for no obvious reason. It's like one of those dogmas that persist, uh like the aspirin for prevention dogma. And I think it needs widespread reversal.

As for verapamil, it was better than beta blocker for sure, but it was no different from placebo. So again, in the absence of s symptoms, I don't see any reason to use beta blocker or calcium blocker. But if there were symptoms, in these patients, say palpitations or maybe dyspion exertion, verapamil would clearly be better than beta block. So I say good on the OH team. Uh this is an excellent study and I look forward to stages two and three.

Alright, next topic is a pixaban versus riveroxaban for bleeding risk. Now, you like me have probably wondered about the superiority of these two drugs. In our neighborhood here in Louisville, the two drugs are basically used interchangeably, almost at random. River Roxaban is once a day, which is a huge advantage regarding compliance. An opixaban has the strongest data against warfarin in the Doak v. Warfarin trials.

And it's BID dosing with lower doses gives an impression of maybe smoother levels without the spike in um anticoagulant effect. This week New England Journal of Medicine has published a head-to-head comparison of the two drugs. Not in patients with atrial fibrillation, but in patients with VTE and PE, which is a bit different. Now Canadian, Australian, and Irish investigators conducted this open label pragmatic RCT. There was no role of industry.

VTE and PE treatment is different than AFib in that both drugs are loaded, they have a loading phase. and actually this loading phase may be important for trial interpretation, so stay tuned. More than twenty seven hundred patients were randomized in one to one fa fashion, A pixaban dosing was ten milligrams twice daily for a week, followed by five milligrams twice daily. Riveroxaban was fifteen milligrams twice daily, not for a week, but for three weeks, followed by twenty milligrams daily.

And I think that's an important distinction. One week versus three weeks of loading. The primary endpoint of this trial was clinically relevant bleeding, which was a composite of A major bleeding or B clinically relevant non-major bleeding. Notably, this was only a 90-day trial. So a primary outcome event occurred in three point three percent in the Epixaban group and seven point one percent in the River Oxevan group. That relative risk is zero point four six, so a fifty four percent reduction

in major bleeding, and that is a highly statistically significant result. The absolute risk reduction was three point eight percent, corresponding to a number needed to treat of only thirty six, or number needed to harm, actually, if you're going the other way of thirty six. If you use the lower bound of zero point three three, then the number needed to harm could be as low as twenty one.

Now, the death from any cause occurred in one patient in the Pixagand group and four in the riveroxigran group, and that was too few events to make any uh conclusion. Serious adverse events related to unrelated to bleeding, such as venous thrombosis occurred in two point seven versus two point two percent in the uh epixaban and riveroxagran groups, respectively, and that was not different.

When you look at the Kaplan-Meyer curves had most of the separation by day 21, then the curves were separating only by a small degree. Medication adherence was actually a bit better in the riveroxiband arm as you'd expect. So my comments. The first thing to say is kudos to the Amer to the Canadian, Australian, and Irish investigators. This is an excellent trial, and we need more of these kinds of studies.

Observational data would struggle with this question because if clinicians, not randomization, chose the drug, then you would never know if there was selection bias. This data also aligns well with previous trials, though none have directly compared the two DOAC drugs for VTE. In the trial called Einstein, riveroxaban was compared against VKA, and it was eight point one percent riveroxaban, eight point one percent uh VKA, so no different. Not worse, though.

In Amplify trial, it was 4.3% of the patients in the Apixagran group had bleeding versus 9.7% in the VKA group. So when you look at these two previous uh trials, Apixaban hugely outperforms VKA, but Riveroxaban just had the same bleeding rate. But again, cross trial comparisons are tempting, but you should resist the urge. That's why this head to head comparison is so valuable.

When I first looked at the Kaplan Meyer Curse, when it came across on social media, I thought, gosh, this is huge and this is a crushing defeat for River Oxebay. But then I looked a little deeper, now I wonder if the answer isn't simply a too aggressive loading phase because most of the bleeding was early. It's hard for me to tell how these different loading regimens came to be. A quick look at the Einstein trial reveals that the authors seemed to make an estimate of what it should be.

Now, I did not find a dose ranging study or anything like that. On the other hand, I did find a dose ranging study for a Pixaban. I could be wrong here in the origins of these two regimens, but they're clearly different. then that means that there's an asymmetric exposure to higher doses of anticoagulation for three times longer uh in the riveroxaban arm versus the epixaban arm.

Loading dose seems like it may play strongly as a role because the difference in the primary endpoints came early. Not all of it was early, but most of the difference comes within 21 days. So as always, there are two messages from this superb trial. Specific messages. For these patients with VTE and PE, with these loading regimens, a pixeban is the superior drug because efficacy is similar and bleeding is substantially reduced by half.

And the absolute risk reduction is also clinically substantial, with a number needed to harm from riveroxaban over a pixaban at thirty six. The larger message, though, is that trials like this can be done. It's well worth the government's money, in my opinion, to sort out these differences. As when you multiply this by millions of people treated, the impact can be huge on a societal level.

The only caveat here in Gosh there are always caveats, aren't there? Is that what if the river oxaban loading dose was the was only a week like a pixaban? Then would the two have been equal? Would efficacy have been similar? Now, I know there's a temptation to extrapolate this to patients with AFib. Millions more people are treated with these two doacs for AFib than VTE.

Should we also prefer a Pixaban for patients with AFib as well? And I don't think so. The problem is that we don't have different loading programs in AFib. We don't load the drugs at all. What we do have in AFib, however, that's different than PE, is we have much longer treatment time. So if there were a small signal of higher bleeding risk with riveroxaban over a pixeband at this maintenance dosing for aphet,

then it might be borne out in a three or five year study. And I think that study should be done, especially with this data. Now it won't be done by the company, so it'll have to be done with government funding, but I think it's well worth the money and effort.

Okay, next topic. There was actually another DOAC study to mention, and this one is really interesting: deltazam versus metoperol, when combined with the doac. The Annals of Internal Medicine has this very nice observational study from a Vanderbilt team Looking at the risk of bleeding in patients with AFib from concumbent use of either epixaban or riveroxaban with dotazam versus metoprolol.

So the scenario is this. Your patient is put on a direct acting anticoagulant these days. It's either a pixaban or riveroxaban. then they need rate control and the choice is either metoprolol or deltazam. And this is an everyday issue because we either use those two drugs for rate control. But here's the issue though. Diltiazam is a potent inhibitor of cytochrome P four fifty three eighty four and it's a P glycoprotein substrate.

Both of these features can affect the metabolic pathways of factor 10A inhibitors, and this could predispose patients. with atrial fibrillation to higher levels of doac and serious bleeding complications. Metoprolol doesn't have these effects. So the authors wanted to compare bleeding incidence in patients who use these doax and take either uh diciltiasam or metoprolol. Now obviously uh Rct would be best, but there is no RCT.

So this group did a retrospective cohort review and they used a large administrative health care database. They had a lot of patients, 23,000 patients on one of these acticoagulants who took deltazam and about 85,000 who took metoprolol. So these are not random comparisons, these were clinicians' choice. So the next step is they balance the two groups with propensity matching.

And they end up using twenty-three thousand patients who on paper had similar characteristics, and then the patients were followed for a year. Their primary endpoint of interest was serious bleeding that required hospital admission admission, and of course that's available in the database. Now the authors did some extra analyses and one in particular that was important was a comparison of bleeding events on low versus high doses of deltazam.

They also compared the uh Pixaban versus River Oxaband, and I'll tell you more about that in a minute. So the main result. Patients who took Diltiazam. had an 18% higher rate of bleeding than those on metoprolol. The 95% conference intervals went from 1.02 to 1.37, so this is considered statistically significant. However, the absolute differences are going to be small. There were 34 versus 29 bleeding events per 1,000 patient year.

Another way of thinking of that is 0.34% versus 0.29%, and that would translate to a number needed to harm of approximately 300 patients using deltazam over metoprolol. They also found a dose response and that the risk for bleeding increased with high versus low dose of deltizam, which was broken down and dichotomized at greater than one hundred twenty. versus less than one hundred twenty milligrams per day.

The absolute bleeding incidence was 0.9% versus 0.3% high versus low. They also did a river oxaban versus a pixeband comparison with Giltiazim, and it favored a pixeband. The rates of bleeding were thirty eight point five per thousand patient years with the riveroxaban deltiazam combination versus thirty per one thousand patient years with a epixaban deltazam combination.

This difference did reach statistical significance, but it was not the primary endpoint, nor did I see any testing for multiple comparisons. Now the rates of stroke and systemic embolism were not statistically different. I think this is one of those rare, non-com random comparison studies that we can actually make causal inference from, namely that there probably is more bleeding when using deltazam over metoprolos.

Why do I say this about this non randomized comparison? For a number of reasons. First De Kaplanmeyer curves do not separate immediately. It takes time or exposure to the separate levels. This suggests causation as longer exposure to higher energy. Anticoagulant levels in the deltazan group should gradually increase bleeding rates, and that was what that is what was seen. Another suggestion of causation comes from the analysis of deltazam dose.

you'd expect a higher dose of deltazam to have more blockade of the metabolic features and higher anticoagulant levels, and that should lead to higher rates of bleeding and indeed The absolute rates of increase of bleeding events was greater in the high dose deltazim group. And of course, dose response is one of those Bradford Hill criteria for causation.

Another clue comes in the differential bleeding rates with riveroxaban versus epixaban. It turns out that riveroxaban depends more on CYP3A4 mediated interactions than does epixaban. When Diltiazan blocks SIP three hundred A four, riveroxaband levels are likely pushed up more than a pixaban level. And again, bleeding rates were slightly higher with the riveroxyband deltazam combination than the epixiban deltazam combination.

Now this was a secondary comparison and conference intervals barely excluded no effect, so caution is warranted, but it's directionally consistent with the drug drug interaction effect. But I think the strongest evidence for causation is that the choice of Metobroll versus Diltaizam is often random. Now the drugs both slow the rate of conduction of AFib through the AV node. Now there's a tendency to use deltaism in patients with lung disease.

and maybe metoprololol in patients with coronary disease, but in the vast majority of cases, whether a person ends up on metoprolol or deltaism is largely random. So what should we do about this study tomorrow in clinic? My first point is that drug-drug interactions are real and they're underappreciated. We should be more mindful of drug-drug interactions.

Of course, they come in two varieties, right? One type of interaction is pharmacokinetic, and that's what this is, wherein one drug alters the metabolism of another. The other variety of drug-drug interactions is pharmacodynamic, where the two drugs have additive effects, say digoxin and mitoprolol. They interact because both slow heart rate. The second take home is that a deltazam doac combination probably does increase bleeding versus a metoprolol doac combination.

The difference is tiny. It's less than 0.5% in absolute terms. The number needed to the harm is 300. Most people on the combinations would have no effect. No difference. The absolute risk increase, however, may be significant in a population of hundreds of thousands, but on an individual level, it's very small.

Then and so there are patients who feel worse on beta blockers. We see this every day. For these patients, deltazam can work well. And so I don't see a problem with using the drug combination in the average patient. The one caution here I would say is that this was only a one-year study, so it might be an accentuated effect over three to five years or five to ten years. I might also be concerned m um with deltazam in patients who had an inherently higher bleeding risk.

Let's say they were on antiplatelet drugs or they had low platelet count. In these patients, this tiny risk could increase and you might be more apt to use metoprolol for rate control. And I think it's really nice for me to cover an observational study that gets very close to making causal conclusions because it's very rare.

All right, next topic is on the actual use, clinical use of smartwatch. EuropAce has published an excellent study from the Circa Dose trial investigators led by Canadian electrophysiologist and hockey star Jason Andrade. Their goal was to assess how three different AFib watch algorithms would stack up to the gold standard using implantable loop recording. The study was doable only because of the novel characteristics of the original circadose trial.

Circados was an RCT comparing three ablation strategies in patients with PAF paroxysmal AFib. It was contact force RF ablation versus a four-minute versus two-minute cryoballoon ablation. The follow up was twelve months and it was one of the first trials to use implantable loop recorder monitors that allow calculation of AF burden. The other neat feature of Circados was that

The more you monitor, the less successful AF ablation looks. Implantable loop recorders give a much more modest success rate than, say, intermittent Holter monitoring. So the main trial results published in circulation in twenty nineteen found no difference between the three strategies. They all looked similar. All three reduced AF burden, though I have to say AF burden was quite low in these patients to begin with.

The study question for this Europace paper centered on how useful can consumer data be in making clinical decisions? So all patients in the circados trial had lube recorders, so the research team used these. To simulate smartwatch algorithms. Now they did not have patients actually wear smart watches, instead they used the continuous ILR rhythm data as a ground truth recording, then mathematically applied The sampling rules rules of each smartwatch algorithm on top of it.

So basically if a smart watch was checking the same rhythm at its programmed intervals, what would it have detected? And these three algorithms that they tested were the Apple Watch AF Burton algorithm. W that was simulated by sampling the ILR data in one minute windows every 15 minutes between 8 a.m. and 10 p.m. Then they would just calculate what fractions of those checks showed a fifth.

The second algorithm was the Apple Watch irregular notification algorithm. This checked every two hours, and if AF was found, it switched to every 15 minutes, triggering an alert if three of five consecutive checks were positive. And they also checked the Fitbit irregular heart rate uh algorithm. This required eleven consecutive five-minute checks every 2.5 minutes to sh to all show AF before generating an alert.

Now, while this simulation likely overestimates real world smartwatch performance, because often the PPG signal is less than ideal, and also patient compliance was assumed ideal. The research team argued these were still reasonable estimates and I agree.

Overall, the watch algorithms did very well. Very well. For AF recurrence detection, The seph sensitivities of these algorithms versus the loop recorder were eighty-two percent for the A Apple Watch AF burden, seventy-one percent for the Apple Watch irregular notification. and 64% for the Fitbit irregular heart rate determination. All three wearables, however, outperformed or equated conventional intermittent monitoring with up to three 14-day ECG patches.

How did they do for AF burden estimation? The correlation for all three algorithms was superb. 0.97 to 0.99. And AF burden in missed detections was clinically net negligible, so less than zero point zero two percent. The effect of wear time. Sensitivity of these things improves substantially with longer daily use. And it rose from forty-two to fifty percent if they were at one hour per day, up to eighty-two to ninety percent for fifteen to twenty-four hours per day.

But get this, even one hour day of watch use outperformed three repeated seven-day halter monitors. Patient activity did modestly reduce sensitivity. From like eighty-two to sixty-four percent with no activity to seventy-one to fifty percent with high activity. And finally, nearly all missed recurrences had extremely low AF burden. Virtually no patient. with greater than zero point one percent AF burden were missed by the Apple Watch AF burden algorithm at standard wear time.

So my comments again, I want to emphasize again that this is a best case scenario because it was a simulation study of the algorithm using very clean data from the loop recorder. The data from the wrist is not so clean. But this study shows A smartwatch algorithms outperform conventional non invasive monitoring that we do in the clinic. B

The Apple Watch AF burden algorithm seemed modestly better in terms of sensitivity than the other ones. C. AF burden estimates are strongly correlated with the AF burden estimates from the loop recorder. D. missed detections are likely clinically insignificant, and e. longer wear times improve performance. And overall I think this is a really important paper because many, many more patients wear these devices.

And we now have two trials, Ocean and Alone AF, which suggest interruption of anticoagulation may be reasonable in patients who do not have AFib after ablation. So when these smart watches came to be, I was highly critical and I was worried about false positives. and overdiagnosis and I still am, especially when they're deployed in young, healthy people who do not have AFib and have a low incidence of AFib. Because remember, if you deploy a screening device

in a low incidence condition you're going to have mostly false positives. But this study, and also my clinical experience in AF Clinic, reveal a possible role of these consumer techs in actual decision making. For instance, You ablate a person who had symptomatic AFib. Maybe even they have a Chad's vask of three, four, or even five. These patients had before the ablation reporting many episodes of AFib. They felt their AFib. And then they had watch notifications pre-ablation.

Now, this patient is six months after ablation and they report no symptoms. Doctor, I'm fixed. Every ECG you do in the clinic shows sinus rhythm. And then they tell you their watch shows no AFib. Whereas before it was showing AFib. Week in and week out, the notch watch notifications are negative. So in this case, your pretest suspicion is that their AFib is gone. The data from the watch strengthens this prior.

And if the patient strongly wants to be off anticoagulation, I think it's hard to argue that they should remain on anticoagulation. Yes, I'd like to see a study like this one take the next step and use the actual watch derived data, not just simulation from the loop recorder. But this is strongly suggestive that watch data can be carefully used in the AFib clinic.

Okay, so one preview for next week is that Gyma Cardiology has published two RCTs. of conduction system pacing versus biventricular pacing. The trials have discordant findings and I will delve deeper and cover this next week, though I would say don't worry too much because both of these trials were very small and I think the C S P versus Bivacing story is not even close to being told yet.

So that's it for this week in cardiology. As always, I'm grateful that you listened. Thank you. And remember, if you like this podcast, please take the time, give us a rating, write us a review. Ah, this helps others find us. Until next week, this is John Mandrola from the Heart.org Medscape Cardiology. You're listening to This Week in Cardiology from dot org Medscape Cardiology Any views expressed are the presented.