Nov 18, 2022 This Week in Cardiology Podcast

John M. Mandrola, MD


November 18, 2022

Please note that the text below is not a full transcript and has not been copyedited. For more insight and commentary on these stories, subscribe to the This Week in Cardiology podcast, download the Medscape app or subscribe on Apple Podcasts, Spotify, or your preferred podcast provider. This podcast is intended for healthcare professionals only.

In This Week’s Podcast

For the week ending November 18, 2022, John Mandrola, MD comments on the following news and features stories.

Three Brief Announcements

First a hearty thanks to our listeners. This Week in Cardiology podcast is up to 543 ratings on Apple Podcasts. There are oodles of podcasts, so I am grateful for your support. Remember to give us a rating or review on whichever app you use.

Second, I had some great feedback this week. On Twitter, Dr Andre Zimerman, a research fellow in the TIMI Study Group, pushed back on my comment last week about triglycerides (TG) being a marker of risk but not a causal factor. He cited genetic data supporting TG-rich lipoproteins as causal and noted that lowering TG is not enough. He argued that pemafibrate may not have reduced cardiovascular (CV) events because it did not reduce both TG and apoB. In fact, in the active arm of PROMINENT, apoB increased slightly.

Thank you, Andre, and please friends, keep it up. If you disagree, let me know.

Third, next week is Thanksgiving and #TWICPodcast takes a break and returns on Friday December 2. Thanksgiving for the Mandrola’s means lots of family time, crushing the trainer bike on Zwift and reading. I’m finishing the Gentleman in Moscow and will plan the Managerial Revolution by John Burnham next.

Now, a lot more from the 2022 American Heart Association (AHA) meeting to review.


We start with the IRONMAN trial, not because it has to be the best trial acronym I’ve seen in years, but because it has both elements of direct clinical concern, and because it provides a major lesson is trial interpretation.

Clinically, nearly half of patients with heart failure with reduced ejection fraction (HFrEF) have low iron levels. Students learn very early on that humans require iron to make hemoglobin which transports oxygen to the tissues. So common sense holds that if iron is low, replacing it will help.

IRONMAN addressed that exact question. This was a randomized controlled trial (RCT) conducted in the National Health System of the United Kingdom. It used a novel iron replacement compound called ferric derisomaltose. Iron deficient patients HFrEF with an left ventricular ejection fraction of less than 45% were randomly assigned to intravenous iron infusions or usual care.

These were mostly outpatients. Trialists aimed to repeat infusions with the goal being long-term maintenance of adequate iron levels. The primary endpoint (PEP) was CV death or recurrent HF admissions.

  • Slightly more than 1100 patients were enrolled over 5 years.

  • Over 2.7 years, the primary outcome was reduced by 18% in the iron group compared with the usual care group. This nearly made statistical significance (rate ratio [RR], 0.82; 95% confidence interval [CI], 0.66 - 1.02; = .070).

  • Hospitalizations for heart failure drove the reduction. Deaths from CV causes, all-cause mortality, and total hospitalizations did not differ significantly.

  • By 20 months, quality of life scores did not differ significantly between the two treatment groups—remarkable because this was an open-label trial.

  • Serious adverse events also did not differ.

The authors did a sensitivity analysis that excluded the pandemic time. The good news was that only 9% of patients were affected by lockdowns. This analysis, of fewer patients, found a 24% reduction in the primary endpoint that met statistical significance (RR, 0.76), albeit with a wide CI, 0.58 - 1.00; P = .047). This was again driven by lower hospitalizations for HF, with no significant reductions in CV death or all-cause death.

Three Main Comments. With an 18% reduction in the primary and P of 0.07, was IRONMAN positive or negative trial?

  • Technically, the upper bound of the 95% CI went to 1.02, and anything one or greater means no effect, and the P-value, a measure of how surprising the data were, given the made-up hypothesis of no difference in the treatment arms, did not meet our accepted, and probably lenient, standard of 0.05.

  • So, no, IRONMAN must be considered a trial with a non-significant primary endpoint. Notice I did not say negative. I think we should jettison that binary distinction. Consider that in IRONMAN, the lower bound of the 95% CI also included the possibility of a 34% reduction in the primary endpoint.

  • And if you consider the totality of the CI of the primary endpoint, ranging from 0.66 to 1.02, the probability of some benefit seems more likely than not.

But the next question is, if there was some benefit, was it clinically meaningful? As an aside: That’s always what we do, right: was a result signal or noise? That requires looking at the CI and some measure of probability, like a P-value or posterior probability in Bayesian analyses. But the next issue is always, was the effect size clinically meaningful.

  • In IRONMAN, the modestly lower rate of the PEP was driven by HF hospitalizations. More important endpoints, like CVD, all-cause death, and all-cause hospitalizations, or even QOL, were not different.

  • HFrEF is not a benign condition. You’d like a therapy — especially one that requires patients to come to an infusion center, give up hours of their day — to move something more than one type of hospitalizations.

  • Again, the matter of HF hospitalizations. Who cares about a reduction in HF hospitalizations if it does not reduce total hospital admissions?

Which brings me to the third point: Are iron infusions worth the trouble, given these results? Patients with HF incur a significant burden with taking bunches of pills and making doctors’ appointments and paying attention to diet and daily weights. So, adding another really significant burden of iron infusions is no small thing.

  • As clinicians, we often hear about all the great therapies for HF, and there surely are, but we must never forget that there is more to life than being a patient. We must let patients have time to live life.

  • What’s more, IRONMAN enrolled relatively young 73-year-old patients who were well enough to be in trials. Many patients with HF have significant co-morbid conditions —arthritis, balance issues, dementia, etc — and many patients don’t have the family support required to get to transfusions centers.

  • Therefore, before adding and paying for outpatient visits for infusion therapy, you’d want a more meaningful effect size and more statistically robust result.

The weakness of IRONMAN data also informs payers. Again, in a system where cost is no issue, sure, go ahead, offer iron infusions. But this includes exactly zero systems. Iron infusion is a challenging area. Let me know what you think.


You don’t have to be in cardiology long to realize how bad cardiogenic shock (CS) is. This may sound a bit goofy, but CS makes you realize the importance of that pump in our chests. The heart is so reliable — 100,000 beats per day, for years on years — we just forget about it. Then when it fails to send enough blood, things go south quickly. Patients turn gray, organs shut down, the brain stops working, and death is imminent.

You hope upon hope that for patients who have a correctable cause of CS, aggressive support can see the patient through that time of reduced cardiac flow. Support should work in some patients. And the only way to sort out if things work in medicine, is with an RCT.

Few forms of support are more aggressive than venoarterial (VA) extracorporeal membrane oxygenation or ECMO. I did not know that VA ECMO was increasing in use, but apparently it is. My partner tells me hospital re-imbursement is generous. I am sure that has nothing to do with anything.

At AHA, thanks to investigators from the Czech Republic, we have results of an RCT comparing immediate VA ECMO to an initially conservative strategy in patients with CS.

A brief note on VA ECMO, which will be an oversimplification: VA ECMO is when blood is removed from the veins (preload reduction), sent to an oxygenator outside the body, and pumped back into the arterial system, usually via the femoral artery. That’s right, oxygenated blood is pumped backwards or retrograde back up the aorta and mixes with blood coming out of the heart at some point.

I spoke with our ECMO person, and read up on this topic, and what I have learned is that VA ECMO is akin to a super-complex symphony. You have to monitor many parameters and adjust flows accordingly. This requires serious human expertise. Recall that blood pumped back into the aorta is an increase in afterload, for a failing pump. There is an excellent 2018 review of VA ECMO by Rao and colleagues.

  • The ECMO CS trial randomly assigned 122 patients at four centers. Most patients had CS as a complication of MI, about one-fifth had decompensation of chronic HF.

  • The primary endpoint was a composite of death, resuscitated circulatory arrest, and implant of another mechanical circulatory support at 30 days.

  • ECMO-CS was not an ECMO vs no-ECMO trial. Instead, it was immediate ECMO vs initial conservative therapy with ECMO if needed. In fact, 23 patients, nearly 40% of the control arm, eventually received downstream ECMO.

  • The PEP occurred in 64% of patients assigned to immediate ECMO and 71% of those in the conservative strategy group, for a hazard ratio (HR) of 0.72. That 28% reduction, however, had 95% CI, 0.46 - 1.12: and a P = .21. That is a strongly non-significant endpoint, though you could argue that the CI was wide, allowing for a more than 50% lower rate and 12% higher rate.

  • But the two main components — death and cardiac arrest — showed no major differences: all-cause death, 50% vs 47%; resuscitated cardiac arrest 10.3% vs 14%.

  • Serious adverse events such as, sepsis, pneumonia, stroke, leg ischemia, and bleeding were not different.

In summary, ECMO-CS yields no signal of benefit to immediate ECMO, on average. This places the trial firmly in line with other CS support trials. A smaller VA ECMO study also failed to show benefit over medical therapy. The large IABP-SHOCK-II trial found no benefit to intra-aortic balloon pump (IABO) vs conservative care.

There are four ongoing trials in this space. ECLS-SHOCK and EURO-SHOCK trials compare VA ECMO vs conservative therapy, the ANCHOR trial compares VA ECMO plus IABP vs conservative therapy, and DanGer Shock trial compares Impella vs conservative therapy.

I want to say something I normally don’t say regarding trials.

What ECMO-CS and the other trials tell us is that on-average, there is not a benefit from support. But another interventional cardiology partner told me something that stuck with me: I know what the average results are, but I can tell you there are occasional patients who we would not get out of the lab without support.

I believe him and what that tells me is that it while it would be unwise to routinely use these devices, it might also be unwise to not have support available for any patient.

This is where I usually add the caveat here that the onus is on proponents to find a population in whom the therapy works. This might be very hard to do in a disease like CS.

Sadly, and I wish I need not add this, but I wonder whether generous re-imbursement for some forms of CS support may influence decision making.


Buckle up for this one.

A few facts before I tell you about the CTS-AMI trial, which studied an intervention to reduce hard outcomes after an ST-segment elevation myocardial infarction (STEMI) and was presented as a late-breaking clinical trial at one of the worlds’ biggest cardiology conferences, the AHA. Patients with STEMI, especially those who present late, still face high in-hospital mortality, myocardial no-flow, and reperfusion injury. There are no drugs available that have proven beneficial in this space.

I want to lead with the results of CTS-AMI:

  • 30-day major adverse cardiac and cerebrovascular events (MACCE) was reduced by highly significant 36% (3.4% vs 5.2%).

  • Cardiac death was reduced 30%.

  • Reinfarction was reduced by 65% (0 vs 9 events).

  • The benefit of this intervention was similar across subgroups.

  • At one year, the MACCE results held up. PEP, CV death, reinfarction, and stroke all were significantly lower.

Now let me tell you that CTS-AMI randomly assigned nearly 3800 patients from 124 hospitals. An important feature of these patients was that many had delayed presentation – symptom to balloon time was 5.7 hours.

Pause there and ponder the results. Huge reductions in MACCE, CV death, and reinfarction in a large multicenter trial, presented in a marquis congress. What’s the intervention? Surely it is set for FDA approval.

It’s a Chinese medicine compound made of insects and herbs called Tongxinluo (tong-zin-lo). Really it is. Traditionally tongxinluo has been thought to have clinical benefits for patients with angina, including the reduction of the occurrence of acute MI, and complications of some types of heart surgery. A 2006 Cochrane review found some evidence that it helps patients with unstable angina, though all the trials had severe methodological limitations.

  • Trial discussant Dr. Ken Mahaffey was surprised. He said something like we need more details about the analysis, trial conduct, and the potential mechanisms.

  • There were calls to see this trial repeated in a Western population.

  • And there were calls to see the angiograms, echos, and biomarkers. In other words, people want to see source data.

While I agree that it’s always good to see confirmatory trials, and the published paper, the disbelief kind of sort of got me thinking about standard science judging.

Again, this was not a poster from a trainee. This was a major clinical trial with hard-endpoint reduction presented at a marquis session at AHA. We don’t understand Chinese medicine, or this compound so, boom, we are skeptical. That’s good. Skepticism is a core tenet of #TWICPodcast. Carl Sagan said skepticism is science.

But. Calls to see source data make me smile. Ok, if this is your position, let’s sit down with an espresso and go through the last few years of New England Journal of Medicine (NEJM) papers and look at the data-sharing statements attached to every trial.

Most of them, especially those from industry will say, nope, you don’t get to see the source data; trust us. Few are skeptical of these data.

I don’t know if CTS-AMI replicates, but boy if it does, we all learn a bunch. There are a lot of late presentations MIs and I think we ought to be randomizing them to tongzinluo or placebo quite soon.


Jason Andrade presented results of a study called PROGRESSIVE-AF. NEJM published it but did not allow the acronym, presumably because this was 3-year follow-up data from the EARLY AF study, published in January 2021.

  • EARLY AF randomly assigned 300 patients with symptomatic paroxysmal atrial fibrillation (AF) who had not yet received rhythm control to either cryoablation (CB) ablation or antiarrhythmic drugs (AAD). All patients had implanted loop recorders (ILR), and the PEP was first documented AF.

  • Patients had to have had AF in the previous 2 years; drug therapy was standardized, and fair doses were used. These were young patients at age 57 years. EARLY AF was a well-conducted trial.

  • At one-year, EARLY AF found that CB ablation resulted in a 52% reduction in first recurrence of AF, though the recurrence rate was very high in both groups — 43% in the CB arm, and 68% with AADs.

  • The ILR allowed the authors to calculate AF burden and here the median AF burden in the CB arm was 0 but it was only 0.13% in the AAD arm. This was statistically significant, but the point was that these patients had low-burden AF.

In the PROGRESSIVE-AF 36-month follow-up, the authors looked at a different endpoint. Here it was the conversion of paroxysmal AF to persistent AF and CB compared with AAD reduced the presence of persistent AF by 75%.

  • Again, the absolute numbers were small: three patients in the CB arm had persistent AF vs 11 in the AAD arm.

  • The AF burden results were similar to the EARLY AF one-year results: very low AF burden with CB and very low AF burden with AAD, as these were low-burden young patients with AF.

  • At 3 years, eight patients (5.2%) in the ablation group and 25 (16.8%) in the antiarrhythmic drug group had been hospitalized (relative risk, 0.31; 95% CI, 0.14 to 0.66).

  • Complication rates were low in both arms. There was a cardiac tamponade in a patient who received ablation and three patients with transient phrenic nerve injury that resolved in 3 months.

Comments. I respect Jason Andrade and many of the Canadian investigators in this trial. I love it that so many RCTs in our field come from this group. And I have no doubt that ablation suppresses AF-episodes-recorded-on-a-monitor a bit better than drugs. I have no qualms with the internal validity or trial conduct.

What I have trouble with here is that ablation vs drug trials are not advancing the field.

The authors write that, “Catheter ablation is a tailored procedure that is designed to modify the pathogenic mechanism of AF initiation and perpetuation by means of a combination of PVI, neural modulation, or substrate modification.”

  • They offer no references. Why? Because it’s speculation. In fact, all of AF pathogenesis is speculation. And that is the problem. When we ablate an atrioventricular pathway AP to cure supraventricular tachycardia (SVT), we understand the mechanism.

  • When we ablate a channel of slow conduction in a post-MI scar to vanquish VT, we understand the mechanism.

  • With AF, there is no understanding of causation. Patients ask how do you know where to burn? Answer: We don’t.

  • Studies show that while inadequate pulmonary vein isolation (PVI) increases the odds of AF recurrence, it is well known that many patients with incomplete PVI do not recur.

  • Another issue with ablation vs drug trials, especially in young people with low-burden AF, is a lack of a proper control arm.

  • Soren Diedrichsen and fellow LOOP authors have elegantly shown that patients with low burden AF often have spontaneous resolution of AF. I had AF for a year, and it stopped. Had I had an ablation it would be counted as a success.

I want you all to think about the choice of AF-episodes-recorded-on-a-monitor as an endpoint. It is surely not meaningless, but it is still a surrogate. Future trials need a non-ablation, non-drug arm, perhaps with a multi-disciplinary team managing them. A team that would say to a patient, “Yes, you had a few hours of AF. It’s ok. Keep taking your oral anticoagulant, lose weight, reduce alcohol intake, use your CPAP device, and live life. If things worsen, we have plenty of choices of therapy.”

These trials also need a proper sham control arm. Put catheters in. Don’t ablate. Let’s see what happens. Medicine is replete with examples that shocked us with sham or placebo-controls. If percutaneous coronary interventions of proximal left anterior descending artery lesions don’t reduce angina over a sham, we surely need to study PVI.

Another problem I have with studies like this is 100% not the authors’ issue but is on us as a field. Electrophysiology (EP) doctors love to do procedures. We get paid to do procedures. Studies like this foster that urge. Medtronic sells more CB, hospitals make tons of money on ablation. In the United States, EP doctors are facing substantial cuts in re-imbursement for AF ablation.

What do you think this will do to the urge to ablate? Do you think ablation docs will say, “No worries, I am fine with 30% less pay,” or do you think some will find more patients to ablate?

My approach to patients with AF remains the same: Remove fear. Educate. Treat risk factors. Give peace a chance to heal, sometimes use AAD, always for the short-term to allow time to reverse risk factors, and use ablation when patients remain symptomatic, and are well-versed in the potential risks.

Next week is Thanksgiving. I’ll have more from AHA and other stories in 2 weeks.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.