Aug 01 2025 This Week in Cardiology

0:00 [Music] You're listening to this week in cardiology from the heart.org Medscape Cardiology. This podcast is intended for health care 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 Mandola from the heart.org Medscape Cardiology and this is this week in cardiology for August 1st, 2025. This week, exercise and cardiovascular outcomes, eldsterone modulation, AI for ECG reading, GLP1 comparisons, end of life decisions, and yet another well-meaning policy that caused substantial harm, this time in US veterans. First, listener feedback. My former

0:48 partner Anthony Pearson, an echo cardiographer, super smart person and blogger known as a skeptical cardiologist, wrote to me about my comments last week on exercise. He wrote, quote, I was shocked to hear you quote major guidelines and numerous observational studies to support the concept that physical activity lowers cardiovascular death, etc. I was even more shocked to hear you discuss the 10,000step study. Again, it's purely observational data as if the higher step counts were reducing your CV death rate, etc. John, have you looked at the RCTs in this area? Why do you have such intense blinders on when it comes to critical analysis of the data supporting exercise and physical activity? You don't have to

1:34 answer. I know why. The logic is that exercise is good. It's good for me. People should exercise more. We can't possibly critique the epidemiologic data that establishes the CB health benefits because that people might exercise less. At the very least, you should throw in a limitation statement when quoting these types of study. Non sedentarily yours, he signs. I love this comment. Indeed, the empirical evidence for exercise is weak to non-existent. It would be nearly impossible to study lifelong exercise and cardiovascular outcomes. Problem A, of course, is the timeline. Problem B is the ethics of randomizing people to no exercise. And problem C is maintaining treatment adherence. You can no more

2:20 study exercise empirically than you can study smoking sessation. Yet, I believe exercise is one of those interventions that does not require randomized data. But I also believe the epidemiologic data is utterly confounded. Not least because a people who exercise regularly do other things that promote health and b there's reverse causation. Right? It's surely present. That is people who are healthier likely exercise more than people who do not because of their good health. And of course C is that scan few people can accurately estimate their exercise amounts. So my answer to the skeptical cardiologist who is worth following and reading, he has an excellent blog, is that he is correct on the technicalities of the empirical

3:06 evidence, which is terrible. But I'm also correct to promote exercise in the same way we could promote clean air and being kind to others. You don't need data to think kindness is beneficial. And I don't need RCTs to believe in the benefits of exercise. All right, next topic is aldsterone modulation in cardio kidney disease. Jack, one of our top journals has a state-of-the-art review on aldsterone modulation in cardioidney disease. I mention it because this is a classic scenario, isn't it? Right after the approval of a new expensive medication, you can bet there will be disease review articles. FDA approved the non-steroidal MRA mineraloid receptor antagonist

3:52 finerone on July 9th. Now of course we have a major review article discussing the two conditions for which the drug will be used CKD and heft. Now, this will be a brief topic because I don't want to be a broken record, but it boggles my mind that finerone was not studied against spironolactone. Yes, finerone beat placebo mostly for non-fatal outcomes, but we don't know if it would beat the $4 per month generic spiritactone. Later on in this podcast, I will discuss a comparison drug study for the GOPs, which is the sort of thing we need more of these days. My points here are three. One is to highlight the pattern of seeing review articles that follow expensive drug approvals. Second point, I will be starting with

4:37 spironolactone in these cases. And my third point is that MRAS are still the most underused drugs in all of cardiology. Go back and take a look at the RALS trial of spiralactone versus placebo in patients with half ref is one of the largest effect size in cardiac therapeutics. 9% absolute risk reduction in death and N&T to prevent death 11. If I were the ZAR of heart failure therapeutics, I would push for more MRA focus. I see it nearly every week. A patient comes in with heart failure. They get sacubetrol, valartan, mtopralol XL and daboglyphloin. Two of these threes drugs will crush the patient's budget. Far better I think for

5:23 many patients who have normal American drug coverage which isn't good would be lysinopril at the highest dose possible spironolactone and carvdalol carvdalol given last and only when stable on ACE inhibitor and MAS I like to shoot for minimally disruptive care whenever possible and minimally disruptive care is often finding the least expensive drugs for patients. All right, third topic today is an AI versus Dr. ECG reading uh paper. The American Journal of Emergency Medicine has a neat study out this week comparing the accuracy of Kathlab activation for STEMI equivalent and STEMI mimic ECGs. I like this study and it's an important problem as the identification of STEMI

6:10 in the first minutes of a person's presentation is crucial. The aim of this study was to measure the doctor accuracy versus the machine learning based AI algorithm from Queen of Hearts which is a deep neural network model. You can use it on your phone. It determines the presence or absence of an occlusion MI. The PM cardio STEMI AI ECG model received FDA breakthrough device designation in March of this year but is yet to be cleared by FDA for marketing in the US. Americans need to wait for FDA approval, but there is an opportunity to get early access to the PM cardio AI bot through a beta signup. It's available on Android and Apple app stores in the European Union and UK. Over 200 hospitals are on the waiting

6:56 list and it's currently being tested in pilot programs at over 60 global centers. For this study done in San Antonio in a hospital system that has two community hospitals and four standalone ERS, the authors chose 18 tough ECGs. Now I know this because they actually put the ECGs in the supplement and I had to really study them. These included four STEMI equivalent types which require immediate reprofusion therapy. They added an ECG with Wellins T-waves and AVR STEMI. They also included transient esti and right bundle branch block left anterior vicular block OMI and a number of ECGs that would represent STEMI mimics. These were included to test false positives uh Kathlab activations.

7:42 So again maybe you have the initial reaction that I had to this study that these could be highly selected ECGs and they might be selected to accentuate the MDAI differences. you know, maybe they were, but then when I looked at them in the supplement, these are real ECGs and they're the types of ECGs that cause a lot of stress reading them. The kind of ECGs you see in peer reviews for miss STEMIS. Now, an important exclusion were ECGs seen in paricarditis, takubus, and prince metal since there are limited published criteria differentiating them from OMI. So in some there were about 12 ECG types that warranted immediate angography and about six ECGs that were mimics and warranted no Kathlab activation and the ECGs were shown to 53 EM docs, emergency

8:30 medicine docs, 42 cardiologists and then of course the AI algorithm and the reference standard was angography or was there an occlusion MI or not and the outcome was a binary outcome a Cathlab activation or not and the results The interpretation accuracies were similar between emergency medicine docs and cardiologists, but both were about 66%. And both of these were hugely lower than the AI model, which accurately called cath activation in 89%. Doctors most frequently mclassified the dwinter pattern, transient stemmies, and hyperaccute T-wave occlusion MIS, as well as bundle branch ECGs. The Queen of Hearts AI algorithm mclassified only two ECG types. The left

9:16 bundle branch block OMI, the Sarosa positive and left ventricular aneurysms. At these DCG types were also challenged by physicians with only 14% and 58% of physicians correctly interpret them respectively. Finally, emergency medicine docs missed 41% of true OMIs and overall 32% of nonomis whereas Queen of Hearts missed only 11% and overcalled 11%. So the summary is that physician accuracy was quite low at only 2/3. This is consistent with prior studies. There was very little difference between emergency medicine and cardiologists. The EC ECG types most frequently

10:02 misinterpreted including left bundle branch block transient STEMI and hyperacute T- waves as well as dwinter T- waves and finally the Queen of Hearts algorithm was more accurate than physicians 89 versus 66%. And it correctly classified all ECG types except LV aneurysm and left bundle branch block. So my comments I find this a remarkable study. The AHI algorithm is clearly better for these ECGs and the ECGs were very hard but they're real and I've seen them reviewed in peerreview meetings where we review miss ST stemmies and no one misses the 3 to 4 millimeter tombstone ST elevations. It's these subtle STEMI equivalents and STEMI mimics that are tough. If you're a

10:48 patient with an oluded LED but not a conclusive ECG, you hope either for good luck or a master ECG reader or perhaps a really good AI algorithm like this one. Now scientifically, I wonder if the best solution is smart doctors who have seen the patient and so they have this basian priors based on history and general appearance. For instance, patients with MIS look like patients with MI from the door plus AI versus just AI. Now, I this is a false comparison because I don't think that a study will ever be done with just AI as it's hard for me to envision an ER without a doctor, but I could not have imagined a medicine without smartphones before smartphones. So, you never know. Nonetheless, I don't

11:36 know why FDA would not approve such a device for use. looks like an important adjunct for getting to the proper diagnosis. I see it as similar to using point of care ultrasound for central venus access. Sure, you could get into a central vein without ultrasound, but why would you? So, in the case of ECGs and cathab activation, sure you can do it without AI, but why would you? These STEMI equivalents and STEMI mimics aren't rare, and the Queen of Hearts looks quite good, and technology is amazing. All right, next topic is GLP news via press release came this week. Eli Liy, the maker of Tzepatide, which is a a GIP GLP1 dual agonist, announced results of

12:21 the surpass cardiovascular outcomes trial, surpass CVOT trial and this was comparing trespathide to dual glutide, which is truly in patients with diabetes and established cardiovascular disease. The trial began in 2020. It enrolled about 13,000 patients and the company reported the topline results this week. Now, dualaglutide was shown to reduce cardiovascular outcomes in patients with type 2 diabetes and established cardiovascular disease or high risk for CV disease in a trial called rewind. And this was published in Lancet 2019. and rerun results were close on the primary endpoint of MI stroke and CV death was 12% in the dual glutide group versus

13:07 13.4% in placebo that hazard ratio and rewind was 0.88 confidence interval 0.79 to 0.99 and a p value of 0.026. Now in the surpass cardiovascular outcomes trial, Lily says the risk of cardiovascular death, heart attack or stroke was 8% relative risk lower tepatide versus duoglutide that hazard ratio was 0.92 confidence interval 0.83 to 1.01 p value 0.086 086 and you think, "Oh, that didn't meet the criteria, but it did meet the prespecified criteria for non-inferiority."

13:55 Now, showed consistent results across all three components of the mace. The rate of all cause mortality in the press release was noted to be 16% lower. Uh, this hazard ratio was 0.84. Confidence interval 0.75 to 0.94. There are also positive results in secondary endpoints. There was a slower slope of EGFR decline. There was more reduction of A1C uh and a 12 - 12% versus - 5% body reduction with tzepide versus dual glutide. A key opinion leader Mutu Vatu Ganathan wrote on X that quote the game has changed to surpass CVBOT meets its primary and secondary endpoints and first

14:41 head-to-head CV outcomes trial. He emphasized the 16% lower risk of all cause mortality. Yet the always reasonable Dr. Sanjay call on Twitter noted that surpass was powered for a 15% relative risk reduction in mace and the primary efficacy endpoint comes out only 8% lower with 95% conference going from 0.83 to 1.01 and that barely contains the hazard ratio 0.85 for which the trial was powered. call also notes the superiority of the comparator here dual glutide has not been established in this patient population. Now you might question that what I told you the rewind trial of

15:27 doulide versus placebo was positive and yes it was but call notes that the subgroup of patients with established athoscorotic vascular disease about a third of the patients in rewind the hazard ratio of doublutide versus placebo was 0.87 87, but the conference intervals went from 0.74 to 1.02. Call also asked what to make of the 16% reduction in all cause mortality. And I think this is a good question because you only have an 8% reduction in mace. And Lily tells us that reduced A1C, body weight, and slowed chronic kidney disease, but no significant difference in CV events. My two cents here are that all cause mortality is likely to be a noise issue.

16:14 The p value was not adjusted for multiple testing. But more important is that if a drug is a cardiac disease modifier, then CV death and CV outcome should drive the reduction in death and we need to see the full trial results. Another issue is that was titrated to max dose and duoglutide was fixed at one dose. And finally, I have a real problem with the non-inferiority design here. Non-inferiority designs are to be used for interventions that offer something less invasive, less costly, or less risky. And I don't see any of that true with tratide. In these early trial results, my takehome is that tritide failed to show superiority of over dual glutide. The

17:00 hazard ratio was only 8% relative risk reduction and the conference intervals went above one with a p value well above our established rest of 0.05. We'll have to wait for details of the full trial presentation which I understand will be presented in September at the European Association of Diabetes meeting. Our next topic is end of life decisions. The BMJ ethics journal published a survey of physicians preferences for their own end of life. The survey included docs from Belgium, Italy, Canada, US and Australia. More than,00 responses were analyzed and physicians rarely rarely considered life sustaining practices a very good option in cancer and Alzheimer's respectively. CPR

17:47 resuscitation 0.5.2% mechanical ventilation 0.8 8 and 0.3%. Tube feeding a little higher at 3.5 and 3.8%. About half of physicians considered euthanasia a very good option, respectively 54% and 51%. Physicians practicing in a jurisdiction with a legal option for both euthanasia and physician assisted suicide were more likely to consider euthanasia a very good option for both cancer odds ratio 3.1 and Alzheimer's odds ratio 1.9. Now I cover this short paper because I continue to be struck by the severity of illness in hospitalized patients.

18:33 Nothing nothing has changed from when I started 29 years ago. I used to remember coming home and telling my wife Stacy how much we were torturing these older people in ICUs and that was in the 1990s. Well, it's the same. I see consults nearly every day at our place and many of the people we are asked to see because of VT or Aphib or brada cardia are literally weeks or months from dying. not of the arhythmia and not of one disease but rather a multitude of diseases resulting in severe frailty. So you read the survey of doctors and you get the impression that since doctors know better they would not be stuck in this loop of hospitalization and ICU stays. But whenever one of these

19:18 surveys on doctor's preferences comes out, I go back to Dan Matlock's paper in 2016 and it's titled how US doctors die, a cohort study of healthcare use at the end of life. They actually found when looking at Medicare data of US docs, doctors spent the same number of days in the hospital and ICU in the last six months of life as did non-d doctors. Doctors in this study did spend a few more days in hospice than non-MDs, but the take-home was that while doctors may express a desire not to have fetal care at the end of life, in reality, they suffer as much as non-d doctors. Now, no idea that I've had gets stronger than this one. The idea is the challenge

20:05 of modern cardiologist is not having something to do for people, but whether we should do it. With every new advance of perccutaneous valve procedures, PFA for aphib ablation, CTO, PCI procedures, the question of using this these procedures in older, sicker patients gets harder and harder. I mean, we can do tavies and open up valves. We can put in pacemakers and fix brada cardia. We can put in CRT devices and fix left bundle wrench block. And now with PFA, we can ablate about anything in the left atrium. But in many of the inpatient consults I see, none of what we can do in EP or in cardiology can fix the dying process of old age. These are super hard decisions

20:51 and I don't have an answer for all the suffering we inflict in the last months of years of life. One specific example, take VT ablation for instance. It's one of the sexiest newest movements in EP on X. You see tons of it. Gorgeous pictures of these diastolic buffets of electrograms. colorful 3D maps. But I will tell you that in reality, many of these patients have VT because they have endstage cardiammyopathy. Now, you want to of course have the skills to ablate VT because a minority of patients have an isolated scar that can be ablated and that patient can then years live years of good life. But gosh, many of these patients have VT because they've successfully survived an MI for

21:36 and heart failure for 20 years. They've had a great run. And I don't mean to be preachy in this topic. In reality, I often don't know when to stop, but I do know that stopping is often the right choice. I would remind listeners that all of us humans have end. And the job of a modern physician more and more is to help people have a good life and good death. We are much better at the former than the latter. All right. Next topic, final topic is well-meaning policies that make sense. I want to close today with another chapter on these well-meaning policies that make good sense. It's one of the most dangerous concepts in healthcare. A few years ago here in the US, there

22:23 was an uproar about access to care in VA hospitals. Veterans often live far from a facility and there were also substantial weight times. So, Congress passed the mission act which stands for maintaining internal systems and strengthening integrated outside networks. This mission act allowed veterans who lived longer than an hour drive to get care outside the VA closer to home because that makes sense. Well, Gemma has published a very interesting observational study of cardiac outcomes from the mission act. The authors led by a team in Philadelphia did a retrospective difference in difference cohort study of veterans who had PCI cabbage or AVR between 2016 and 2022 in

23:12 nonVA hospitals covered under the mission act or in VA hospitals. The two outcomes were mace MI stroke or hospitalization for CV cause or death within 30 days of the procedure and travel times. This was a huge database study looking at the three procedures. Tens of thousands of patients in each group. The two main groups were patients who lived far and patients who live near. Now the first finding was that mission act implementation for PCI cabbage and AVR. There were much larger percentages of far rather than near patients who received these procedures in nonVA hospitals. So the mission act did accomplish that veterans who lived far from a VA hospital did get more

23:58 procedures outside of the VA system. The second finding and hint is that FAR patients who received procedures at nonVA hospitals were more likely to receive care at non-eing smaller rural and for-profit hospitals than near patients receiving nonVA care. The third finding was to look at outcomes before mission act October 1st 2016 to June of 2019. The difference in travel times the probability of choosing the VA and 30-day MAC showed no statistically significant differences in differences between the two groups. That's important because it is provides support for the pre-intervention parallel trends assumption critical to the validity of

24:43 difference and difference analyses. So basically in other words before the mission act there were no significant differences but after the mission act which was in implemented in 2019 travel times increased a tiny bit in near patients but decreased by a lot in far patients. I think the travel time increased a bit in near patients because it was not just distance but also weight times that could allow veterans to go to other hospitals and non and nonVA hospitals may be farther away than the VA. Indeed, PCI cabbage and AVR volume in VA hospitals decreased quite a bit after mission implementation. But here's the key results. FAR patients, FAR veterans undergoing

25:30 PCI had a 2.3 percentage point adjusted mean increase in 30-day MACE rates compared with a 0.5 percentage point adjusted mean decrease in mace rates among near patients. So the difference in difference was 2.8 percentage points highly statistically significant. FAR patients undergoing cabbage had a 1.6 6 percentage point adjusted mean increase in 30-day mace rates compared with a 6.5 percentage point decrease among near patients and that difference in difference was huge at 8.1 percentage points and near and far patients undergoing AVR had similar adjusted mean increases so there was no difference in

26:15 difference for AVRs and the authors concluded that quote mission act implementation was associated iated with substantial decreases in travel time among veterans who became geographically eligible for nonVA care for these patients undergoing PCI or cabbage mission implementation was also associated with worsened 30-day MACE rates my comments I remember thinking this was going to be the likely result yes it's nice to get care closer to home I often see rich endurance athletes who travel to see me if they should have a procedure I tell them to get it close the home. Well, that's because AF ablation is a wellpracticed procedure that can be done in all major cities. But PCI, cabbage, and AVR are procedures

27:03 that not only require a skilled doctor, but also a skilled team in a system. And while VA hospitals may not have the best food or the greatest decorations, they often have great processes and dedicated staff. In fact, in the introduction of this paper, the authors site three observational studies finding that VA cath labs have better mortality rates than nonVA kath labs. And I don't find this a surprising finding at all. So, the mission act focuses on improving access to care as if that's a good thing. And it does. Veterans have shorter drive times to get care, but increasing care outside the VA results in worse results, at least for PCI and cabbage. Now, I should add that this is observational and there may be

27:50 confounding. While baseline characteristics in the two groups near and far were similar, those who live farther from the VA may be sicker. But I doubt this is a confounder here because if there's one thing that US hospitals are good at, it's making patients look sicker on paper. So I find these results highly likely to be right. Care in the US has lots of variability. VA care on the other hand is pretty standardized. I see a similarity to say Canadian health care. When I visit Canada, I'm struck by how cardiac procedures are done in in small numbers of hospitals. So large volume, small numbers of hospitals. This means that Canadian patients having procedures have doctors and teams who do a lot of the procedure. Now they may

28:36 have to travel, they may have to wait, but when they have the procedure, it is done by experts. Now in the periphery of major cities in the US, it's the wild west. For instance, in Louisville, there's about eight or nine centers doing AF ablation. Now, you may get a skilled doctor in the US who has done has tons of experience, but you may not. This paper suggests the policy of allowing veterans to seek faster and closer care resulted in worse outcomes. And the lessons, as they always are, are both specific and general. Specifically, it was a bad idea to think that in the US, more convenient healthcare was a positive. And generally it would have been far better to implement this policy

29:22 in an RCT pilot form first. Then instead of looking back and seeing all this harm that it caused, policymakers could have adjusted mid-stream and maybe mitigated the harm. I don't know why we feel that trials are needed for new drugs and devices but not policies. In fact, policies may affect more people than drugs and procedures. And I think it's even more important to study these in RCT form. So that's it for this week in cardiology. As always, I'm grateful that you listen. Remember, if you liked this podcast, please give us a rating, write us a review, and by all means, if you don't agree with something I said, send me an email, write a comment, and we'll make it into listener feedback. Until next week, this is John Mandola from

30:08 theart.org Medscape Cardiology. You're listening to this week in cardiology from theheart.org Medscape Cardiology. This podcast is intended for health care professionals only. Any views expressed are the presenter's own and do not necessarily reflect the views of WebMD or Medscape.