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 September 8, 2023, John Mandrola, MD comments on the following news and features stories.
ESC Part 2
In part 2 of my European Society of Cardiology (ESC) meeting review, most of the trials for which I make comments, I do as a neutral Martian. That is, as an outside observer, focusing mostly on data. Last week’s ESC review involved a lot of trials in my area of electrophysiology. This week’s trials are different. If those with content expertise think I missed some nuance, please do comment.
We had a few comments and ratings from last week, so keep it up. Ratings and reviews really help others find this podcast.
I choose important trials based on two big criteria: one is how they answer a specific clinical question and the other is what they teach about medical science. ECLS-SHOCK fulfills both these criteria, and even has a bonus — bravery.
Some brief background points:
Cardiogenic shock related to acute MI is bad. Death rates are high despite all that modern cardiology can deliver.
The problem with CS is poor oxygen delivery caused by pump failure. Opening closed coronaries will help, but recovery of pump function takes time.
This core problem leads doctors to look to mechanical circulatory support, to replace or assist the heart’s function as a pump during the time of injury.
ECLS, also called veno-arterial ECMO (extracorporeal membrane oxygenation) is a way of providing mechanical support during the time of pump failure. It’s use has risen 10 times over the past 10 years.
ECLS is seriously invasive. It requires large bore access to central veins and the femoral artery. Recall that the patient with acute MI just had percutaneous coronary intervention (PCI) and is on high dose antiplatelets. Blood is routed to an extracorporeal circuit where it is oxygenated. It is then sent back retrograde in the aorta. These flow adjustments require lots of human expertise.
Doctors choose mechanical support in the setting of MI-related CS because it is therapy directed at the problem — lack of pump function. But there is scant data to support this practice, even though it makes sense. In fact, the current data is downright depressing. The intra-aortic balloon pump IABP has not been found effective. A smaller trial called ECMO-CS, published last year, found no benefit.
So the ECLS-SHOCK authors, first author Holger Thiele from Leipzig, set out to conduct a more definitive trial.
Slightly more than 420 patients were recruited from multiple centers. All had MI-related CS.
Half got venoarterial (VA) ecmo, the other half standard of care.
These were 63 year-old patients, mostly men.
Primary endpoint was death at 30 days.
There was no significant difference In outcomes; 48% in the VA ECMO group vs 49% in the standard of care group (hazard ratio [HR] 0.98).
Safety outcomes were way worse in the VA ECMO group.
Moderate or severe bleeding: 23% VA ECMO vs 10% standard of care. That’s more than double the risk of a bad outcome.
Peripheral vascular complications: 11% vs 4%; nearly 3 times higher in the VA ECMO group
Comments. First, the bonus. This is a brave study because it requires taking people with pump failure and not putting them on a life supporting therapy. It sets an example.
Second is the specific message: VA ECMO does not work in MI-related CS. Period. There was no signal of death reduction and 2 to 3 times more serious complications. There was talk about patient selection and trial conduct, but you already had one null trial last year, and now this more definitive trial.
The authors spent many words discussing why VA ECMO did not work. This was not necessary and probably was meant to appease peer reviewers and editors. This could have been the shortest discussion section ever published. In fact, you barely need a discussion given the clear results.
Specifically, don’t do VA ECMO in this setting. I also think this puts a serious damper on mechanical support of any sort in CS, though we await the IMPELLA trials before making any judgements based on observational data.
Third is the general lesson about medical science: I’ve said this many times.
Always be cautious of doing things that make sense.
Yes, on paper, VA ECMO supports the low cardiac output. It’s plausible. But that is way too low a bar in medicine.
You have to show something works to reduce an important outcome in real patients.
If Professor Thiele had not been curious and brave, we would be doing this invasive procedure for years—we’d be harming patients and not knowing it.
ECLS-SHOCK is a shining example of why we need a culture of randomization in this profession. Many if not most things can be tested in randomized controlled trials (RCTs).
The question in the FIRE trial centered on the strategy of treating older patients with acute MI who have multivessel disease. There are two choices in the cath lab — fix the culprit occlusion and treat the others medically or, after fixing the culprit, do the other significant lesions.
Culprit-only or complete revascularization? There is a general consensus that in younger patients, complete revascularization is preferred. (I am not sure how strong the evidence is for this, as the largest trial, COMPLETE, found no difference in cardiovascular (CV) death, nor all-cause death.) But, for the sake of argument, let’s accept the consensus.
We know from last week’s most important trial, FRAIL AF, what works in younger patients may not work in older patients.
In the FIRE trial, Italian investigators (first author Simone Biscaglia) enrolled older patients (mean age 80 years) who presented with either ST-elevation MI (STEMI) or non-STEMI (NSTEMI) to receive complete revascularization guided by physiology assessment or culprit-only PCI.
The primary endpoint was a composite of major adverse cardiac events (MACE) including death, MI, stroke, or ischemia-driven revascularization.
This is important — patients were randomized after successful culprit PCI was done.
About 720 patients were in each arm. (For reference, the seminal COMPLETE trial enrolled nearly 3 times more — about 2000 patients per arm)
The composite MACE endpoint occurred in 21% of patients in the culprit only vs 15.7% in the complete arm. (HR, 0.3; confidence interval [CI] 0.57-0.91, P = 0.01)
Death from any cause: 12.8 vs 9.2%; HR 0.70 (0.51-0.96)
MI: 7% vs 4.4%; HR 0.62 (0.4-0.97)
Stroke: numbers were too small
Revascularization: 6.8 vs 4.3%; HR 0.63 (0.40-0.98)
Safety outcomes — acute kidney injury, stroke, bleeding – no difference.
Comments. The results are clear. The lower rates of the primary outcome are statistically and clinically significant. Three of the four components of the primary outcome are positive for complete revascularization. Safety was not different. I am surprised.
First, it’s unusual that older patients have stronger effects than younger patients. That’s typically because of competing risks.
Second, after you fix a culprit, you are, essentially, treating stable disease.
And there is no signal that revascularization improves outcomes over medical therapy.
But the results are the results. Although there were far fewer patients in FIRE as opposed to COMPLETE, FIRE had a higher event rate.
One caveat the authors point out that may have been important is that non-culprit interventions were guided by physiology, which reduces the number the of PCIs.
The authors note that 483 non-culprit vessels (50.9%) were not treated with PCI on the basis of physiologic measurements that did not indicate the need for revascularization at the time of functional testing.
Based on the totally biased FAME-2 trial of fractional flow reserve, I am not sold on physiology-based treatment, but I have no trouble believing that anything that reduces non-culprit PCI is a good thing.
The main problems with translating this evidence are that it is open label and randomization was done after culprit PCI.
First, patients in the medical arm know that they have other bad blockages that were not treated. That may not influence death, but it surely influences revascularization.
I calculated a Fragility index of 9. Meaning, if 9 patients having events in the control arm were converted from not having a primary outcome to HAVING a primary outcome, the study would lose significance at P > 0.05. That’s not a lot of revascularization decisions in a trial of older pts with acute MI.
Second, randomizing after culprit PCI affects translation. Meaning, interventional cardiology docs are unlikely to randomize patients with complex non-culprit PCI. So, to me, a non-PCI doctor, means that I would not translate this evidence to people with complicated non-culprit lesions such as chronic total occlusions (CTO). For instance, more than one in three non-culprit PCIs were in the circumflex or ramus arteries.
Caveats aside, this is strong evidence. It surprises you. And that fact adds heft to my contention in the first segment that randomization is one of the most important innovations of modern medicine. For, if we were guessing, we would have guessed that older patients would have benefited less, not more, from COMPLETE revascularization.
Interventional cardiologists, if you feel differently, please weigh in.
The FIRE trial, taken together with previous data from COMPLETE, argues for complete vs culprit-only PCI when patients have acute MI and multivessel disease.
Now the question is how best to accomplish complete revascularization. The two choices: Do it all at once during the index procedure right after stopping the STEMI. Or do it in a staged second procedure in the days to weeks later.
Lead author Barbara Stahli and colleagues in the MULTISTARS AMI trial set out to answer that question by randomly assigning more than 800 patients who had STEMI and multivessel coronary artery disease (CAD) to immediate multivessel PCI or PCI of the culprit followed by staged multivessel PCI within 15 to 45 days.
Their primary endpoint was MACE — death, MI, stroke, ischemia-driven revascularization or hospitalization for heart failure (HHF).
Mean age of patients 64 years, mostly men, mostly white.
They designed the comparison as a noninferiority (NI) trial with immediate as the new treatment and staged as the standard.
The results were crystal clear.
A primary end-point event occurred in 8.5% in the immediate group vs 16.3% in the staged group (risk ratio, 0.52; 95% CI, 0.38 to 0.72; P < 0.001 for noninferiority and P < 0.001 for superiority)
MI and revascularization were the drivers of the benefit. Death, stroke and HHF were not much different.
Serious adverse events (AEs) favored the immediate strategy: 104 vs 145 had AEs in the immediate vs staged strategies.
Comments on Limitations. This data does not apply to CTO, left main, vein grafts, or CS, as they were excluded.
We also have the same translational issue as FIRE, namely, that patients were randomly assigned after successful culprit PCI. So, interventional cardiologists are choosing which patients to put in the trial based on knowledge of anatomy.
One question I have is, why NI design? It’s not clear to me that doing all the PCI in the first procedure offers that much of a safety or cost advantage. Maybe it’s a slight convenience advantage. To me a straight superiority comparison would have been better.
Another issue is that the drivers of the MACE were not the strongest or hardest endpoints. Ischemia-driven revascularization is a difficult endpoint because it is so decisional (by doctors and patients) and thus, susceptible to non-nefarious bias.
MI could have favored the immediate group because any periprocedural MI will be hidden by the initial STEMI whereas any periprocedural MI at a month, would be obvious.
CV death, HHF, stroke were all the same.
So, yes, immediate multivessel PCI seems reasonable to do, but I could see a scenario wherein a clinician might decide staging is better. I’d hate to see guidelines or quality measures incentivize doctors to algorithmically choose immediate multivessel PCI based solely on this trial.
In patients with HF, iron deficiency is common and contributes to symptoms. When important things are low in the body, doctors look towards replacing them. Thus, the IV-iron replacement strategies were born. A handful of small trials hinted at benefit. But none were statistically robust.
The HEART FID trial was designed to be the trial to answer the question.
8000+ patients were screened to enroll about 3000 patients randomly assigned 1:1 to iron infusions or placebo.
The primary outcome was a hierarchical composite of first death, then HHF, then change in 6-minute walk test. The significance cut off was set at P = 0.01.
This significance level fulfilled a special protocol agreement with the Food and Drug Administration that permitted the assessment of efficacy on the basis of a single randomized trial.
The patients were about 69 years old, mostly male, white, and about half had class 2 and the other half had class 3 HF. Mean ejection fraction was 31%. Mean heomglobin was 12.6 gm/dL, mean serum ferritn was 56 mcg/L.
The results were similar to previous iron infusion Trials — almost but not quite there.
Deaths were 8.6% vs 10.3% in the iron arm vs placebo.
HHF were 297 vs 332, iron arm vs placebo.
Delta in the 6-min walk test no real difference.
Overall win ratio – 1.10 (CI 0.99-1.23) with a P-value of 0.02. Higher than prespecified 0.01.
Secondary outcomes – CV death or HHF – HR 0.93 (CI 0.81-1.06) so, not significant.
The authors wrote:
“We found no apparent difference between the ferric carboxymaltose group and the placebo group with respect to the primary outcome — a hierarchical composite that included death, hospitalizations for heart failure, or change in 6-minute walk distance. Results of prespecified sensitivity and supportive analyses appeared to be consistent with results of the primary analysis.”
In a patient level meta-analysis of three trials of iron replacement (CONFIRM-HF, AFFIRM-HF, HEART-FID, presented at ESC and published in European Heart Journal, there was a trend towards reduction of the co-primary endpoint of CV death and total HHF with an HR of 0.87 but a CI of 0.75-1.01 and P-value of .08.
Comments. When I look at this data taken together, it is hard to get excited. Iron infusions are burdensome for patients. It takes time, a trip to a center. And it’s not free.
When we assess whether a therapy is worth recommending, we consider the effect size and the statistical robustness of the findings. I am a neutral observer. I have never prescribed iron replacement. I have no conflicts.
To me, the effect size seems not very clinically significant. Nor is it statistically robust. I don’t believe this therapy, on average, in these trials, made a big difference.
Could there be people with super low iron levels, super low transferrin saturation levels, that might gain more? Maybe, but on average, I am not seeing it.
As with the two previous trials, I assess here as an evidence-person who has everyday cardiology knowledge but no specific content expertise. So let me know if I am off base.
Sick sinus syndrome or now, the new name, sinus node dysfunction (SND), is one of the most common bradycardias. Its main cause is birthdays, so as patients live longer, we will continue to see more SND.
When the bradycardia causes enough symptoms, we place a dual chamber pacer, so as to pace in the atrium.
Many of these patients have or develop atrial fibrillation (AF). My entire career, experts have wrestled over how best to prevent the AF seen with SND.
Some have advocated for higher rates of pacing and pacing overdrive algorithms. The idea is that higher rates of pacing will smooth out refractory periods and act as anti-fibrillatory therapy. BOOM. Another idea that makes sense. I bought into it too.
Then over the years, I started to see that it doesn’t really work, and patients may feel palpitations. Longer term trials have not born it out.
Now the pendulum has flipped to, “let’s minimize atrial pacing.” Let the atrium do its thing.
A Danish team, led by Mads Brix Kronberg, from Aarhus, designed the elegant DANPACE II trial to answer the question of pacing in this scenario.
540 patients got a pacer for SND.
One group gets a base rate of 60 and rate responsive pacing — DDDR pacing.
The other group gets a base rate of 40 with no atrial rate responsive pacing — DDD pacing.
The DDDR-60 group paces in the atrium about half the time.
The DDD-40 group paces about 1% of thetime; a major difference in the amount of atrial pacing.
The primary endpoint of first AF occurs in 46% of both groups; HR 0.97 (CI 0.76-1.25) P = 0.8
There were no between-group differences in AF exceeding 6 or 24 hours, persistent AF, or cardioversions for AF.
The incidence of syncope or presyncope was higher in patients assigned to DDD-40 (HR 1.71, 95% CI 1.13–2.59, P = .01).
Comments. We have two themes. One is the specific question of programming patients who have pacing for SND. This trial shows no AF prevention from lower rates, and a higher rate of syncope. So we can program it in the normal way at a base rate of 60 and rate responsive pacing. Good to know that avoidance of atrial pacing does not affect AF. Question answered.
The second theme gets back to epistemology or ways of knowing. The team in Aarhus did not go about answering the question with charismatic key opinion leaders wearing Italian suits. Or retrospective observational studies. They designed and conducted an elegant RCT. They showed it can be done, and it answers questions.Good on them. Thank you.
© 2023 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Sep 08, 2023 This Week in Cardiology Podcast - Medscape - Sep 08, 2023.