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 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 October 28, 2022, John Mandrola, MD, comments on the following news and features stories.
First, I want to thank the Royal College of Edinburgh for inviting me to speak. Especially Drs Anne Scott and Andrew Flapan. Thanks also to the many doctors who spent time talking with my Dad, who went to Scotland with me — especially Marc Dweck.
My assigned talk title was “Evidence Translation, Lies, Damn Lies and Statistics.” It was a fun lecture, one that I have not given previously.
I want to give some follow-up from my coverage of the ERASE-AF trial, which was a randomized controlled trial (RCT) of pulmonary vein isolation (PVI) vs PVI plus low voltage area (LVA) ablation in patients with persistent AF. The surprising result was that LVA ablation led to a statistically significant 38% reduction of first recurrence of AF vs PVI alone.
In my comments last week, I praised the fact that the authors tested their idea of identifying and targeting low voltage areas in a randomized trial. I pointed out that the prior evidence of additional ablation beyond PVI was extremely pessimistic; adding ablation beyond basic PVI offered no benefit. That’s why ERASE AF was so notable — it’s the first trial showing that ablation beyond PVI definitively reduced first AF recurrence.
Another point I made was that only one-third of the active arm actually had LVA to ablate. Yet ERASE-AF was positive. I said that this was akin to a drug trial being positive despite the fact that only one-third of patients got the drug.
The pessimistic Bayesian priors plus the fact that only a third of active arm patients got the intervention led me to raise the possibility of performance bias – operators did better PVI in patients randomized to LVA arm.
Senior author Christopher Piorkowski and I communicated by email. He made some important points; points that I hope #TWICPodcast listeners will be interested to hear.
First is that ERASE AF is the first study that documented that two-thirds of patients with persistent AF were LVA negative vs one-third who were LVA positive. That the majority of patients with persistent AF had normal electrical substrate helps explain why so many previous trials have failed to show any benefits from additional ablation.
Piorkowski emphasized and I think he is right, that ERASE AF is the first study that acknowledges and really exploits the different phenotypes in patients with persistent AF.
While we have generally thought that patients with persistent AF represent more advanced AF, the documentation that most of these patients do not have demonstrable electrical disease is potentially important.
So the benefit of ERASE-AF can be explained by ablating the electrical substrate as well as NOT ablating patients with normal atria.
Piorkowski also noted that the ablation of LVA in ERASE-AF included great care in ablating the LVA andcreating a block in the lines that transected the LVA.
I want to re-iterate the notion that good trials often lead to new areas of research, and I believe ERASE-AF has done that. The field should now work to confirm and build on these results. I want to know:
Whether these results hold up longer than one year? They may not.
Would this reduction of AF lead to better quality of life and perhaps lower stroke risk?
Can other centers duplicate the results?
Recreational Drugs Associated With AF
The UC San Francisco group led by Dr. Greg Marcus (first author Anthony Lin) have published an observational study in the European Heart Journal regarding the association of recreational drugs and AF. This is the first large longitudinal cohort study examining use of cocaine, methamphetamine, opioids, and cannabis with the risk of incident AF. All four of the drugs are associated with higher incidences of AF.
It’s a nice study of adults who received care in a California emergency department (ED); they use a bunch of statistical methods to adjust for confounders.
Cannabis use had the weakest association of the four but was still significant. But it may be the most clinically significant of the four because legalization of cannabis will surely lead to more of its use.
The thing is, if a person is using these drugs, having AF seems quite low on the diseases to worry about.
Even if the associations with AF were weak or null, my advice to people is to avoid cocaine, meth, and opiates. Period.
If you take meth, for instance; relative to your teeth falling out, I am not sure higher AF incidence is high on the worry scale.
As for cannabis, we definitely need more studies on its efficacy and harms for medical uses. It could be that the conditions cannabis is being used for are the causal agents in its use.
There may be a new drug for the treatment of supraventricular tachycardia (SVT). Etripamil is a self-administered nasal spray that can be used to terminate SVT. The drug is a calcium-channel blocker — similar to verapamil — and reaches peak effect at 10 minutes. The antiarrhythmic effects decrease after 30 minutes.
It is not yet approved by the US Food and Drug Administration (FDA) but the drug maker, Milestone Pharmaceuticals, announced in a press release that the drug met its primary endpoint in the RAPID trial, which randomly assigned patients to etripamil vs placebo.
The primary endpoint of SVT termination was 2.6 times more likely in the active arm; 64.3% vs 31.2%.
The company also said they pooled this analysis with a previous trial, called NODE-301 trial, which technically did not make its primary endpoint.
The company also reported acceptable safety and tolerability.
They plan to present detailed results as a late-breaking session at the American Heart Association (AHA) meeting coming up in 2 weeks in Chicago. They also plan to apply for a new drug application next year.
I think the drug could have a niche use. Here is why: SVT is one of the most preference-sensitive decisions we make in electrophysiology. SVT is rarely life-threatening. I offer patients 3 options — no treatment other than education on how to terminate episodes; daily drugs like beta blockers; or ablation.
Many patients, after hearing the discussion, choose to live with SVT and choose to learn to terminate the arrhythmia. In my younger days I was enthusiastic about ablation. I still love ablation but realize many fully educated patients do fine without an invasive procedure and with as needed measures.
Normally this entails resting, then Valsalva maneuvers. It doesn’t always work, however, and when that occurs the patient has to go to an ED, which is expensive and time consuming and often means getting adenosine, one of my least favorite drugs because it is causes so much distress. Patients hate adenosine. I call it the heart stopper. Anyway, a nasal spray that could help terminate SVT and keep patients with infrequent episodes out of the ED could be a potentially handy drug.
The other thing with SVT is that it often goes dormant or stops, so as-needed treatment for SVT with a nasal spray may allow for natural history to eliminate the problem.
We will see the actual data at AHA, and then if approved, cost will be a big factor, as it is with any new and novel drug. I am afraid it will be cost-prohibitive for many.
Misuse of Science
Theheart.org | Medscape Cardiology has a longform features article on the important topic of misuse of scientific data. The author leads the story with an egregious case of violent misuse of a social science paper. But then she delves into the more common forms of misappropriation of science. Say, when a controversial politician Tweets out a paper on myocarditis related to a vaccine.
The longform article makes you think about the bigger picture of science interpretation. While COVID brought more intense politicization of some topics in science, the abuse and misunderstanding of science long pre-dates the pandemic.
In one part of the article, the author considers the role of scientists. Quoting an expert, she writes: “The way science hits the world is part of science.”
“Modern researchers should think about the social context of their work” says another.
Having been embroiled in this mess, in the ‘before’ times, when I foolishly believed critical appraisal could be applied to COVID science as if it were statins, stents or ablation, I have thought a lot about the presentation of evidence in the digital landscape. In fact, the grand idea of this podcast is not just to help with specific issues in cardiology, but more to create a culture in which critical appraisal is the norm.
There will always be bad actors who can twist a study’s findings; supplement and vitamin sellers, or politicians, for instance.
But here is my idea: imagine a populace, a society, that has a different view of how science worked — how much uncertainty there is, how most things we test don’t work. How science is a process of knowing, a process that does not finish.
If this process of knowing was the normative view, there would be less market for dubious actors to work in. If doctors held such norms, there’d be a lot fewer left atrial appendage closure (LAAO) devices placed.
My conclusion, and I know many will disagree, please do, is that society’s struggle with medical science is largely on us.
I’ve co-authored a paper documenting the shocking amount of spin in the cardiac literature. Spin is when authors use language that distracts from a non-significant endpoint. Every single time an author spins a scientific paper, cynicism and distrust is advanced. It may seem like nothing to emphasize that silly surrogate endpoint or subgroup, but it shreds confidence in our profession.
This month we had a decidedly non-significant trial looking at embolic protection devices after transcatheter aortic valve implantation. The primary endpoint of stroke did not get close to reaching statistical significance. Yet, at the meeting, the authors of the study, many of whom are proponents of the device, diverted attention to a secondary endpoint with small numbers of events. And then in the conclusion of the paper, they added a phrase suggesting the confidence interval did not exclude the possibility of benefit.
At the Heart Rhythm Society meeting last May, prominent researchers presented a study comparing two types of implantable cardioverter defibrillator (ICD). The Atlas trial had a primary endpoint with five components, four of which could only occur in one arm. When the newer, fancier ICD proved superior, and of course it would, the next day, Boston Scientific was publishing ads touting the “positive results.”
These are just two of many examples. Our literature is replete with stuff like this.
Here is another thing we take as normative that average people would find shocking. I’ve heard Joe Rogan rant about his discovery that scientists who publish papers don’t have to show their data. Regular people hear that and think, wait...the only people who see the data are the authors?
Then there is the process of codifying standards of care. In many cases, not only do proponents of a therapy conduct the study, these same authors also write the guidelines, which establish what doctors do and payers pay for.
Then come the medical reversals, which can overturn a therapeutic fashion. Look no further than cardiac cooling after cardiac arrest. This began with weak data. Then trials showed no benefit. Still there was no change in practice. Not until a much larger trial was done has practice begun to change.
Some degree of medical reversals ought to be normal as knowledge changes. But you hardly ever see leaders get up to the microphone or pen editorials about how they were wrong, and what they learned from the process.
The model of science that we now embrace borders on anti-science.
We have proponents of interventions doing studies largely designed to show a therapy works.
Much of a bias in a trial comes in its design. Say the choice of uneven run-in periods, weak comparators, dubious endpoints.
Don’t even get me going on selective reporting of endpoints that look unfavorable.
So disordered is the use of medical evidence that devices that fail to reach a lax non-inferiority margin become a standard therapy. Yes, a pivotal regulatory trial of LAAO, PREVAIL, did not reach non-inferiority in its first co-primary endpoint of stroke, systemic embolism, and CV death. Missing a lax non-inferiority margin should tell an important story. But now LAAO is normative and advertised to patients by doctors and hospitals alike.
This model of science sews cynicism and distrust. Now, even when a new therapy actually works, it’s less believable.
My arguably naïve solutions would be that:
Content experts and proponents be less involved with the design, conduct and especially interpretation of trials.
Methodologists who have less skin in the game ought to lead the trials. Big industry ought to fund more basic science, and phase 3 and 4 science funding should come from government.
Guidelines should be narrative reviews written by neutral parties.
I would do away with those boxes that summarize recommendations as if patients fit neatly into an algorithm. The narrative and evidence review part of guideline documents are often wonderful ways to learn the evidence base. We don’t need the boxes. Doctors are smart.
Mostly, I would want public facing scientists to always lead with uncertainty. Viz, “We don’t (exactly) know but here is what we think and why we think it.” And since we know so little, there would be more equipoise, and this might create a culture of randomization.
In Scotland, I heard a colleague say, “The only way we could get that device is if it were in a trial.” That’s what we need more of. Science would actually advance faster if properly studying things before adoption were the norm. Medical science, therefore, needs to heal itself.
Diabetes as a Risk Factor
The Journal of the American Medical Association has published an interesting research letter from Canadian authors who used administrative healthcare data from Ontario to study the changing association of diabetes (DM) and CV events over the past 3 decades.
Their findings were provocative, because diabetes has always been one of the scariest risk factors for heart disease. Here is what they found:
DM prevalence increased substantially from 3.1% in 1994 to 9% in 2014. (I bet it’s way higher now.)
In 1994, the presence of DM increased the risk of a CV event by 2.1 times.
In 2014, the presence of DM increased the risk of a CV event by 1.6 times.
As a comparison, they also looked at the presence of CV disease as a predictor of a CV event. The relative risk (RR) of a person having a CV event if they had prior CV disease was also about double, but it did not change over time.
The message here could be that the DM of new may not be as strong a risk factor as the DM of old. It’s still a risk factor, but in this one study, it was no longer equivalent to having prior CV disease. The authors attribute this to advances in the treatment of diabetes. That’s possible although there was not much SGLT2 inhibitor or GLP use in 2014. Another factor — especially when using administrative data — could be the lower thresholds for the diagnosis of DM.
I highlight this research letter not to make light of DM as a risk factor, but to highlight the changing environment in which new therapies are tested. One of the topics I spoke about in my lecture in Edinburgh was that trials might need an expiration date. Take ICD therapy for nonischemic cardiomyopathy: 20 years ago, ICDs reduced mortality. The most recent DANISH trial in 2016 found no benefit to having an ICD, which is likely because of decreasing sudden cardiac death event rates along with better background therapy.
Beta-blockers after myocardial infarction may not be useful in the post-reperfusion era. And now we have DM becoming less of a risk factor. Times change.
The final reason to highlight this paper is that it is a good use of observational data. Instead of trying to match two non-randomized groups and making causal inference, this group is reporting on changing trends of association over time.
© 2022 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Oct 28, 2022 This Week in Cardiology Podcast - Medscape - Oct 28, 2022.