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 July 19, 2024, John Mandrola, MD, comments on the following news and features stories: Vulnerable plaque, scientific method, and prominent journals; industry payments to trainees; tirzepatide or semaglutide; trial interpretation, intention to treat (ITT) vs on-treatment; and pulsed field ablation (PFA).
Non-invasive Imaging for Vulnerable Plaque
More than 30 years ago, I was taught, first at Hartford hospital and then at Indiana University, that most myocardial infarctions (Mis) occur because of young, immature plaques that did not cause severe stenosis. For whatever reason, possibly, no reason, a non-stenotic plaque would rupture and cause clot and an MI.
This mostly ignored lesson is why negative stress tests don’t exclude the chance of having an MI. It’s why coronary artery bypass grafts (CABG) and percutaneous coronary intervention (PCI) of stable stenotic lesions don’t reduce MI or death.
Here’s the problem: Over the years and decades since this core knowledge has been known, many investigators have tried to reach the twin pinnacles of modern cardiology — identification of the vulnerable plaque and effective treatment of that plaque.
The Journal of the American College of Cardiology (JACC) has published a study from the famous Edinburgh group led by Professor David Newby that has led to headlines in the news such as “PET Imaging Finds Vulnerable Plaques That Cause MI.”
Gosh, if that were true, that would be amazing. We could all buy more PET imaging machines, image patients, and shred the incidence of heart disease.
But.
If the identification and treatment of vulnerable plaques is a marathon, the study presented in JACC might be the first kilometer. Or, you might be running on the wrong course.
Before I give you the Neutral Martian/non-PET-imaging-expert take on this evidence, let me first set out the most likely possibility of heart attacks and most human disease. Wait for it, sit down, slow the bike.
The most likely cause of most terrible events is chance. Many of you may know the term stochastic, but I have come to believe that stochastic is the explanation of many of the conditions we treat.
My understanding of stochastic is that events are ultimately random, although there may be a probability effect. For instance, smokers are more likely to have an MI, but when or if an MI occurs is random. Or radiation exposure from cardiac procedures increases the risk of cancer, but whether a cancer occurs is random. If you studied 1000 cardiologists and 1000 internists, more cardiologists may have a cancer, but whether a cardiologist gets cancer is still random.
I believe that is the story with vulnerable plaques. Stochastic could be translated as “stuff happens.”
Now to the JACC study. This was a secondary analysis of a prospective cohort study of F-flouride-PET imaging in more than 2600 patients who had an MI and multivessel coronary artery disease (CAD) first published in JAMA-Cardiology in 2023.
The JACC paper, first author Kang-Ling Wang, aimed to test whether vessel-level coronary plaque activity is associated with vessel-level MI. It’s a great concept right? Do a PET scan. If there is increased uptake in a vessel, you will know that the vessel is hot. And perhaps be treated. (Treated with what, exactly is a question, but forget that, because we aren’t close to answering that yet). Let’s just see if we can use PET imaging to detect an at-risk artery.
Here is the top-line.
Over 2 years of follow-up, MI occurred in 4% of vessels with increased uptake and 2% of vessels without increased uptake. The hazard ratio (HR) was 2.08 and the confidence intervals (CI) were significant at a P of 0.01.
The authors then tested the association in vessels that were treated with revascularization vs those not treated with revascularization and, lo and behold, those treated with revascularization had no association but untreated vessels had a positive association.
The authors also noted that increased coronary plaque activity in multiple vessels was associated with increased patient-level risk of cardiac death and MI (HR 2.4).
Comments. When I read this study and the supportive editorial, senior-authored by Dr. Fuster himself, I thought, gosh, this is amazing. The Edinburgh team is on to something. Vulnerable plaques detected non-invasively is something.
Well, there are problems.
First, this vessel-level analysis is a secondary analysis of an initially negative cohort study. It’s not even a randomized controlled trial (RCT).
Journalist Sue Hughes quoted senior author Dr. David Newby saying, “this is the first multicenter randomized trial to show [PET] can be used to ID vulnerable plaques….”
But I don’t think that is right, because the original analysis, in JAMA-Cardiology, first author, Alastair Moss, is a cohort study of patients post-MI who had multivessel disease.
That cohort study had a primary endpoint of cardiovascular (CV) death, MI, and unplanned coronary revascularization. Of the 2600 patients screened, 995 (less than half) were eligible and 712 had imaging.
The rate of the primary endpoint did not differ in those with low or high coronary plaque activity as measured by Flouride PET. The primary endpoint was non-significant.
Here is the first line of the conclusions in the original analysis:
In this cohort study of patients with recent myocardial infarction, coronary atherosclerotic plaque activity was not associated with the primary composite end point.
Did the authors say, crap, I guess this did not work. No. They did not.
Instead, they looked at a secondary endpoint. The composite of CV death or MI was indeed higher (HR, 1.82; CI 1.07 to 3.1) in patients with high plaque activity. The authors put a lot of emphasis on that endpoint because CV death and MI was supposed to be the primary endpoint, but was changed to CV death, MI, and unplanned revascularization because of lower-than-expected events.
I am just a Neutral Martian here, but if you don’t like your more specific endpoint because events are low, and you fear your study will have no difference, and then you change your endpoint to capture more events, and then that endpoint, with more events, is negative, you can’t go back and emphasize your first endpoint after you know the results.
Second problem: If your original study has a nonsignificant primary endpoint, it’s highly dubious to go back into the data and publish a positive secondary analysis.
Third problem: Even if you ignore the issue of going back into a nonsignificant original analysis to find a positive secondary finding, the difference in the vessel specific PET study was 4% vs 2%. That means that 96% of vessels with high uptake did not cause a future event.
Fourth problem: Even if you ignore the first three problems, and you believe PET identifies vulnerable vessels, or vulnerable plaques, or vulnerable patients, how else do you treat these patients over and above lifestyle changes and statins?
I am sorry to say that the holy grail of cardiology, the identification and treatment of a vulnerable plaque remains totally, 100% elusive.
But you would not realize that from reading our most important cardiology journal, and that is a problem.
Industry Payments to Cardiology Fellows
The team of Sanket Dhruva and Rita Redberg have a new research letter out in JAMA-Internal Medicine looking at Industry payments to cardiology fellows.
Before I tell you about the study, I recall my training at Indiana University many years ago. Industry was everywhere. Thursday echo conference, Friday cath conference, were always sponsored by a drug or device company. There were trips to pacing company offices for “education.” And in clinic, there was often food laid out by industry. I am not sure we were outliers, but when I came to private practice, having industry people around was normal.
The authors write in their second sentence that “Payments to trainees merit scrutiny because these payments may be formative on future practice.” Of course, these payments are formative. That is the point.
The study attempted to quantify industry payments to fellows and then measure the association of these payments to receiving future payments. These were non-research-related payments.
The study sample was slightly more than 5000 fellows. Nearly half were procedural; 21% female.
During the year before graduation, 80% of procedural fellows and 67% of non-procedural fellows received payments from industry. The median amount was $1800 in the procedural group vs $200 in the non-procedural field.
Three years after fellowship, 96% of the cardiologists in the procedural fields and 81% in the non-procedural fields had received industry payments.
Physicians who received general payments in the year before fellowship graduation were more likely to receive payments after graduation (HR, 2.12; 95% CI, 1.89 to 2.37) .
Docs in procedural fields were more likely than docs in nonprocedural fields to receive any payments (HR 2.2).
If the amount in fellowship was more than $5000, then the HR for receiving payments in practice was 3.3.
Comments. I commend the authors for doing this work. As I have said many times in this podcast, I am not against innovation, nor am I against a profit-motive for industry innovation. We have better tools, devices, and drugs today because of profit motive.
But there should be strong barriers between doctors and industry. When a device company spends money taking fellows out to dinner, or sending them to meetings, etc, that money has a specific goal. Industry may say it is for education, but it is actually 95% for marketing.
If this is done on a grand scale, industry can help establish therapeutic fashions.
Imagine a world where a specific company — say one who makes a left atrial occlusion device, or a left ventricular support device — takes care of fellows in many programs. Now, when the fellows get out, nearly everyone has had experience with the company and its device. Boom. Use of those devices is normal.
Marketing and fear of missing out is not the only way therapeutic fashion occurs, but it is a big contributor.
A specific word here about my field. It saddens me greatly that so much of our education is done via industry. At the Heart Rhythm Society meeting, I went to a lunch symposium sponsored by industry to hear a lecture from a friend. I stayed for the other three lectures, and it was shocking how biased the presentations were regarding the company’s product. I did not Tweet or write about it, but perhaps I should have. Some of the other lectures were from key opinion leaders, and hearing such biased presentations decreased my respect for them.
I am sorry, but industry payments to doctors is often unseemly. And unprofessional. It should be possible to learn new techniques without taking money from industry. With the advent of social media, I learned conduction system pacing (His bundle and then left bundle area pacing without taking money from industry.
When you learn from industry courses, you are being marketed to. Period.
Tirzepatide vs Semaglutide
The SELECT trial of semaglutide in patients with obesity and established atherosclerotic CV disease (ASCVD) found that semaglutide reduced hard CV outcomes by 20%, notably, in addition to baseline therapy with statins. SELECT makes semaglutide a disease-modifying drug. That is big.
The FLOW trial of semaglutide in patients with chronic kidney disease (CKD) has helped confirm the disease-modifying properties of semaglutide. FLOW studied 3500 patients with type 2 diabetes and CKD and had as a primary endpoint a composite of kidney outcomes, one of which was CV death.
And, semaglutide over placebo reduced the composite outcome by 24%. This included a statistically significant 29% lower rate of CV death. Again, 80% of these patients were on lipid-lowering agents.
How semaglutide improves outcomes is a major question. Surely a large driver of benefit is weight loss. There have been arguments that some of the benefit is independent of weight loss.
The mechanism question is important because tirzepatide, which has a dual GLP-1a and gastric inhibitory polypeptide action, also induces weight loss, but has no CV outcomes data, as yet.
The notion of class effect here is huge. Should we assume that tirzepatide has the same effects on CV outcomes as semaglutide?
JAMA-Internal Medicine has published a comparative observational study of tirzepatide and semaglutide, looking at the degree of weight loss.
This was a cohort study of 18,000 propensity matched patients who initiated either tirzepatide or semaglutide. First author, Patricia Rodriguez, PhD, and colleagues used electronic health record (EHR) data from multiple US healthcare systems,
The total sample size was 41,000; about 25% on tirzepatide and 75% on semaglutide. Before matching, patients taking tirzepatide were slightly younger, more often female, and a more often had a college education.
The authors then matched the patients using a propensity score.
In short, patients on tirzepatide had more weight loss.
For ≥5% weight loss, the HR] was 1.76, 95% CI, 1.68, 1.8.
For ≥10%, the HR was 2.54; 95% CI, 2.37, 2.73.
For ≥15%, HR was 3.24; 95% CI, 2.91, 3.61.
Gastrointestinal side effects were similar.
Comments. The authors strengthened this observational study by confirming these results with sensitivity analyses. For instance, using inverse probability weighting, the results were similar.
These results also mirror the findings from clinical trials, although the two drugs have not been compared head-to-head in a single trial. In the SURMOUNT studies of tirzepatide vs placebo for weight loss, there was greater weight loss in tirzepatide vs placebo group than in the STEP trials of semaglutide vs placebo.
The authors tell us that there is a SURMOUNT 5 trial comparing tirzepatide vs semaglutide but only looking at weight loss.
Adding this observational data to the trials, I suspect that tirzepatide does induce more weight loss than semaglutide.
But does that mean it has the same CV outcomes data as semaglutide ? I looked but have not found a CV outcomes trial for tirzepatide. There is the SURPASS-CVOT of tirzepatide vs liraglutide looking at major adverse cardiac events (MACE) but it is not yet completed. Otherwise, we have no specific CV outcomes data with tirzepatide.
Tirzepatide has a GLP-1a mechanism of action and so it is likely that it also induces the same CV outcomes as semaglutide. But we don’t know. Gosh it would be nice to know if it was a class effect.
We have CV outcomes trial for multiple beta-blockers, RAS inhibitors, MRA drugs, and SGLT2 drugs. We ought to have them for GLP-1 drugs in obesity.
Colchicine and Trial Interpretation in the Lancet
Here is the idea: inflammation plays a role in causing vascular events. Non-cardiometabolic stroke or vascular stroke is a common manifestation of CV disease. Would an inflammation blocker, like colchicine reduce the risk of recurrent vascular stroke?
The first reason to discuss the CONVINCE trial is to emphasize to a cardiology audience that many, if not most strokes, are not cardioembolic or related to atrial fibrillation (AF). Many strokes are caused by small vessel atherosclerosis.
The second reason to discuss the CONVINCE trial of colchicine vs placebo in patients with a transient ischemic attack (TIA) or vascular stroke is to discuss trial interpretation. David Brown alerted me to this issue via Twitter (X).
More than 3000 patients were randomized after a high-risk TIA or vascular stroke to either colchicine or placebo. The primary endpoint was a composite of first fatal or non-fatal recurrent ischemic stroke, MI, cardiac arrest or hospital admission for unstable angina. Basically, a broad MACE endpoint.
Here were the results. A primary endpoint occurred in 9.8% of patients on colchicine vs 11.7% on placebo. This corresponded to a 3.32 vs 3.92 event per 100 patient years, resulting in an HR of 0.84; 95% CI, 0.68 to 1.05, P=0.12).
That’s nonsignificant right?
Well, here was the conclusion:
Although no statistically significant benefit was observed on the primary intention-to-treat analysis, the findings provide new evidence supporting the rationale for anti-inflammatory therapy in further randomized trials.
Comment. This is strange. I thought we had a P-threshold of 0.05. This is way over that.
The way the authors justified it, I think, is that they used the on-treatment analysis. Here the results were 8.0% vs 11.7% and HR was 0.796 (0.63 to 0.9992)
That is funny, isn’t it. They carry the HR out to 3 decimals and the upper bound of the CI to 4 decimals.
The thing is that they planned for colchicine discontinuations of 15% but had an observed discontinuation rate of 20.5%.
The authors also emphasized the lower CRP levels in the colchicine arm.
The issue here is that you cannot use on-treatment analyses. You have to use ITT to preserve randomization. People who stop the drug or don’t take the drug are likely to be different from those who took the drug. On-treatment analyses are fine to do, they teach you things for further study, but I don’t think this particular on-treatment analysis with an upper bound of 0.999 lends much confidence in the anti-inflammatory theory of treating vascular stroke.
Another reason to highlight this paper is that I find it very interesting that the trialists are spinning results of a non-significant trial, even though the trial is not sponsored by industry and colchicine is inexpensive. It speaks to an intellectual not financial bias. Scientists, it seems, like to find positive results.
I wish it were not that way. I wish scientists simply like to do experiments, and interpret them more neutrally, and more skeptically.
Speaking of Skeptical, Let’s Talk PFA
I wrote a column on the outsized enthusiasm for PFA for AF among US electrophysiologists. It surprises me how much marketing and the fear of missing out affects the brains of US electrophysiology (EP) doctors. Our enthusiasm for first generation PFA systems is remarkable and inconsistent with the empirical evidence.
Empirical evidence finds no difference in efficacy. No difference in safety. Manyfold higher costs. And despite the rumors, empirical evidence does not confirm that PFA is that much faster than thermal ablation.
Yet you cannot be a Top Person in EP unless you are adopting first generation PFA systems. I will link to the short piece. I hope you read it.
© 2024 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Jul 19, 2024 This Week in Cardiology Podcast - Medscape - Jul 19, 2024.
Comments