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 March 21, 2025, John Mandrola, MD, comments on the following topics: A large trial in cardiac pacing finally published, PVCs and cardiomyopathy, cannabis, CV risk and the danger of observational studies, and the tale of two disparate statin trials.
BioPace Trial
EuroPace journal has published the results of the BioPace trial, first presented at the 2014 ESC meeting in Barcelona. The trial compared biventricular (BiV) pacing to standard right ventricular (RV) pacing in patients with bradycardia who will require lots of pacing.
It was — basically — a preventive trial. The idea being: standard RV pacing fixes the bradycardia, but RV paced beats create, much like left bundle branch block (LBBB), and induces pacing-induced left ventricular (LV) dysfunction in a certain (highly debated) number of patients.
But BiV pacing is harder; it requires a third CS/LV lead. The question BioPace tested is whether the preventive more synchronous BiV pacing would lead to better outcomes than the old standby — the single RV pacing lead.
It was a multicenter RCT — conducted more than 20 years ago — of 1800 patients randomized to either of the two pacing strategies. The primary endpoints were two: (1) death or heart failure hospitalization (HHF), and (2) survival time.
These patients were in their 70s and had about 90% RV pacing at one month. The mean ejection fraction (EF) was normal at 55%. But nearly 10% had EF < 35% and 23% had EF between 35% and 50%. Listen later for the EF subgroup analysis. (Recall here that the positive-for-biV-pacing Block HF trial included patients with EF < 50%).
Results
Over a mean follow-up of nearly 6 years, a primary outcome of death or HHF occurred in 38.4% in the BiV arm vs 40% of the RV pacing arm. HR was 0.78 but CI went from 0.76-1.02 and the P value was 0.09, so nonsignificant. The second co-primary of survival was also not different, with both groups having a death rate of 33.8%.
Trials like this have crossovers, and the authors provide an as-treated analysis which also showed no significant differences.
You might wonder if the patients with lower EFs did better with BiV pacing. Nope. Subgroup analyses showed no heterogeneity of treatment effect for any endpoint.
For secondary endpoints, there was no difference in 6 min-walk, and quality-of-life (QOL) questionnaires.
There were, however, more adverse events (AE) in the BiV pacing arm: 115 vs 72 procedure related. And 87 vs 54 lead-related AE noted in the BiV vs RV group.
The authors concluded, rightly I think, that "in patients with preserved EF who require pacing and will pace in the RV, BiV pacing did not prove superior and was associated with numerically more AE." The authors add that RV pacing may be less harmful for this type of patient than previously thought.
Comments
What a nice trial this is! It directly challenges my beliefs — an enthusiastic believer in synchronous RV pacing.
While it is older data with BiV pacing, the publication come at a crucial time for the electrophysiology (EP) field. Why do I say that? Because the hottest topic in all of EP is conduction system pacing (CSP), wherein we can achieve ventricular pacing with synchronous RV and LV pacing by placing the ventricular pacing lead at either the His bundle or the left bundle (LB) in the septum.
Essentially, CSP purports to do cardiac resynchronization therapy (CRT) with one lead instead of two required for BiV pacing. That is the RV and LV lead.
I, and huge numbers of my colleagues, feel that pacing the conduction system will prevent pacing-induced cardiomyopathy. I, and many of colleagues, also do CSP routinely as first line.
But… But… BioPace should cause us proponents to reconsider two prevailing ideas: one is that RV pacing is so terrible. Second, that more complicated pacing is better than simple RV pacing.
On the matter of pacing-induced cardiomyopathy, observational studies put this incidence in the range of 10%–20%. And we know from the early implantable cardioverter-defibrillator (ICD) trials, that preexisting heart failure (HF) makes RV pacing really bad. But BioPace enrolled the more common patient — one with bradycardia but minimal to no previous heart muscle disease. They found very little harm from RV pacing. HF events leading to death or hospitalization were not different and by my inspection of the Kaplan-Meier curves, HF occurred in less than 10% of either group over the first three years.
A neutral observer — one not affected by the beauty of the narrow-paced QRS from CSP or the satisfaction of finding the perfect CS lead — would look at BioPace and strongly argue for sticking with standard RV pacing.
They would say the trial had more than 700 primary outcome events, long follow-up, great generalizability, and yet no benefit. This was not an underpowered trial.
And I would have trouble refuting that.
Except there are some limitations. First the trial recruited patients in 2003. I was young then. Many of TWIC listeners may have been children then. It's important because BiV pacing has iterated quite a bit. LV leads and delivery systems — as well as knowledge has improved — a bunch. In BioPace, LV lead implants failed in 13%, and I suspect it is lower now.
Another criticism I saw online was that the death should not have been an endpoint because these patients were too healthy. There is some merit to this criticism. The more typical HF endpoint of CVD or HHF would have been better. Or even total hospitalizations, because this would capture the AE. But on the other hand, death is the least biased of all endpoints.
I would also argue that CSP is likely better than BiV pacing in both safety and efficacy. Safety is likely better because we use one lead in the RV septum vs cannulation of the CS, venograms and CS lead placement. Smaller studies and observational studies suggest — albeit weakly — that the hemodynamics of CSP are better than BiV. But I want to emphasize that empirical support for this contention is weak.
BioPace should infuse us CSP evangelists — like me — with a big bolus of humility. The standard RV pacing is not dead yet. And surely we should have more data with CSP.
Now a couple comments on medical evidence. I worry that the 11-year delay in publishing this study had something to do with the negative results. BioPace was industry supported.
At the time, BiV pacing was hot. It was sexy; the devices were more profitable. You might think this is cynical thinking. Well, let me ask you about the counter-factual: had BioPace been a huge win for BiV pacing, do you think the funder of the trial would have let it sit unpublished for 11 years?
No, no, double no. I really worry about this delay in publication.
Another Belief Challenged in EP This Week — PVCs and Cardiomyopathy
In my opinion, one of the most misunderstood and overtreated conditions in cardiology today are premature ventricular complexes (PVCs). And overtreatment of PVCs can get into the hazardous area quite quickly because ablation of PVCs can bring doctors to scary areas of the heart — left ventricular outflow tract (LVOT), thin-walled RV and para-Hisian areas, for instance.
A recent paper from the University of Colorado group provided strong findings about another strong belief in EP — that is, if patients reach a threshold PVC-burden, then they will develop cardiomyopathy (CM) and HF. The number in our town is 20% PVCs. But in other places it can be as low 10%
Observational studies vary considerably in the predictors of cardiomyopathy. JACC featured a state-of-the-art review paper last year, I will link to it. In one paragraph on predictors of PVC CM they had over 40 references. Things like PVC burden, coupling interval, variability of coupling interval, QRS width, epicardial origin and many more.
Another issue: I do not trust echo reading in the presence of PVCs. In the era of digital echo, if an apical four chamber includes a PVC, I swear the EF is underestimated.
The Colorado group adds to the melee of observational studies: They analyzed more than 30,000 ambulatory monitors on patients between 2017-2023. They included those who had PVC burden > 5% and more than 24 hours of monitor. That latter point, about duration of monitoring, is important, as just one day of monitoring can misrepresent (often underestimate) the degree of PVC burden.
They also searched for people who had a transthoracic echocardiogram (TTE) done within 3 months of the monitor. They defined CM as an LVEF < 50%.
Results:
They found a total of about 1400 patients who fulfilled criteria: Average age 68, 40% female. The average PVC burden was 12%.
Of the 746 patients who had a TTE, the mean EF was normal at 55%, with 22% (1 in 5 patients) having an EF of < 50%.
Here is the key finding:
In both unadjusted and adjusted analyses, they found no significant association between percentage of PVCs and LVEF (P = 0.78), nor with PVC burden and depressed left ventricular function (P = 0.13).
The conventional 20% cutoff of PVC burden showed no predictive value, and no trends were seen linking increasing PVC burden or prolonged PVC exposure to reduced function.
They concluded that in this sample, that PVC burden alone was not an independent predictor of CM.
Comments
I include this single-center observational study, even though there have been many similar studies in this space, because it is a) recent, and b) challenges a prevailing therapeutic fashion that PVC burden more than 20% should trigger treatment, such as ablation.
It's an important paper because treatment decisions for PVCs is one of the hardest areas of EP. It can never be algorithmic.
I have seen oodles of patients with more than 20% PVC burden who have no symptoms, no HF and no LV dysfunction. Yet I have also seen these same asymptomatic PVC patients made sick by their doctor's words. Once happy and living life, the patient with asymptomatic PVCs comes to the EP consult fearful and addicted to their Kardia device—because a doctor scared the patient.
I have also seen patients with lower burden PVCs who are terribly symptomatic, and counseling reassurance do not work.
So please, friends, be careful with your words when counseling these patients. Be cautious in the use of PVC burden cut-offs.
I do recommend this paper, as well as the excellent State of the Art review.
Perhaps the most important sentence from the review, first author, Pouri Shoureshi,
in a prospective cohort study from 2012 to 2017, 100 asymptomatic or minimally symptomatic patients with frequent idiopathic PVCs (≥5% in 24 hours, normal LVEF, and no identifiable cause), 44% experienced a decrease to <1% PVCs without ablation or anti-arrhythmic drug therapy, highlighting high rate of spontaneous PVC resolution.
Yes, as it is with atrial fibrillation, with PVCs I also say, give peace a chance. There is very often spontaneous resolution. Nothing in Medicine can be more elegant than resisting the urge to intervene when nature can cure the problem.
While I do PVC ablation, I try first to make sure the PVCs will not disappear on their own, I also try big doses of reassurance.
Ablating PVCs can become harrowing. Be careful out there. Do not be influenced by EP Twitter. Because on Twitter, no one puts their 6-hour cases that fail or cause tamponade or heart block or stroke or aortic damage.
Cannabis and CV Risk
JACC Advances has published a brief report that hit the social media jackpot this week. The question was to assess the long-term CV risk of cannabis.
The authors used a data source called TriNetX, which is drawn from 53 healthcare organizations.
The study population included adults younger than 50. The time course of study was 2010-2018.
The comparison group by ICD codes were simply cannabis user vs non-cannabis user. Incldued patients were free of co-morbid conditions such as prior MI or heart disease, DM, HTN, or tobacco use. The relatively good health of these patients was planned and important.
This was a huge data base study of more than 4.6 million young adults. About 2% (93,000) were cannabis users and 4.5 million were not users. So, you are comparing 2% of the sample to 98% of the sample.
Of course, the authors attempt matching here; they use propensity scoring.
Key findings
Over the 5 years of follow-up, myocardial infarction (MI) absolute risk was 0.55% vs 0.09%. That is a relative risk of 6x higher in cannabis users. But the absolute risk difference is 0.45%. A number needed to harm is in the range of 220.
There were similar findings for ischemic stroke. Big relative risk difference of 5x higher but tiny absolute risk difference of 0.3%
Same with major adverse cardiovascular events (MACE). For AF the relative risk increase was only 2x higher and the absolute risk increase was in the range of 0.43%
For all-cause mortality, the relative risk was 1.5x higher with an absolute risk difference of 0.45%.
The authors conclude that this data links cannabis use to adverse CV events. Notably this was in a population without comorbid conditions.
The last tine of the paper uses causal words:
In conclusion, the findings underscore significant cardiovascular risks associated with cannabis use, reinforcing the need for public health initiatives and heightened clinician awareness.
Note the words "underscore" and "reinforcing the need" for health initiatives.
Comments
I don't love cannabis, but I also don't love the conclusions of this paper. It's fine to make associations in epidemiologic studies. But recall how Bradford Hill established the causative effect of smoking in heart disease and cancer.
Epidemiologic studies such as this one were one component of causation. But Hill also had biologic mechanism, which are clear for tobacco (carcinogens, endothelial dysfunction). And natural experiments wherein countries and regions that started smoking bans saw less heart attack. And Hill had consistency and specificity, meaning that the smoking link holds up in different populations and different time periods. There was also a clear dose–response curve with smoking.
This week's study is but one of the components of showing a causal connection, and that is the epidemiologic association.
These types of studies are perhaps the weakest component of causal inference. Let's review the three main issues with these nonrandom, look-back comparisons.
First is selection bias: do cannabis users have higher rates of MACE because of the cannabis, or because of other factors, such as the ones that induce cannabis use. Things like insomnia, mental stress, etc. The authors might say the corrected for baseline factors, but adjustment can only be made by those factors listed in a database.
Second is the matter of Dena Zeraatkar's work on specification curve analysis, where the McMaster University team shows that observation studies like this have many thousand different analytic choices when looking at data from 5 million patients. How one chooses analytic method can affect outcomes. A good example, I remember the 2018 HRS meeting where a Colorado group showed in MI patients, cannabis users vs non-users had a lower in-hospital mortality. The study is published in PLOS-1.
Third, this is a chart review using codes. There is no data on dose-response, for instance, and you would really want to see dose–response if you plan to infer causation.
In sum, please be sure. I think the likelihood that cannabis use is protective agent is super low. In fact, I suspect that recreational use of cannabis likely leads to bad health. While studies like this garner tons of attention, they are nowhere near robust enough to infer causation. The most useful aspect is in teaching critical appraisal.
Final comment: While I don't think cannabis will likely pan out as a useful therapeutic, the fact that it has been illegal has precluded proper randomized controlled trials (RCTs). I would not be surprised if a proper RCT in a selected cohort found that cannabis could have some short-term benefit in say chemo-induced nausea or anorexia or for pain relief.
Cardio-Oncology
JACC CardioOncology has published a prespecified subanalysis of the STOP-CA trial of Statins to Prevent Cardiotoxicity from Anthracyclines.
It's an important topic because up to 20% of patients who get A develop HFrEF at 5 years.
Recall that JAMA published STOP-CA in 2023. The highly positive trial was controversial — at least to me. One reason for curiosity, before even looking at the trial procedures etc., was that a Wake Forest-led RCT called PREVENT of statins for prevention of LV dysfunction failed to show a benefit. This was in NEJM Evidence in 2022.
STOP-CA had 300 patients with lymphoma who had anthracycline-based chemo received either 40 atorvastatin or placebo for 12 months.
The primary outcome of the proportion of patients experiencing a decline in left ventricular ejection fraction (LVEF) of ≥10% from baseline to a final value of <55% over 12 months was significantly lower in the atorvastatin group (9%) compared to placebo (22%), P = .002.
In subgroups, atorvastatin showed greater benefit in patients who were older than median age (52 years), those treated with higher cumulative anthracycline doses (≥ 250 mg/m²), obese patients (BMI ≥ 30), and females.
I remember there was great celebration about this, but there were problems: missing data where not all subjects had MRI at baseline and follow-up. And of course, change in EF is very much a surrogate outcome — which I previously mentioned is highly imprecise. What's more, how exactly does a statin exert such a great effect in 12 months? It would seem longer time frames would be needed.
The substudy in JACC Oncology took a subset of about half the patients in STOP-CA and looked at before and after MRI scans.
They chose another surrogate marker — a >3% increase in extracellular volume (ECV) on MRI. I am not an imager, but they chose this because it is a marker of interstitial fibrosis, a purported mechanism of anthracycline-induced toxicity. The cutoff value was chosen based on its association with heart failure and cardiovascular death in cancer patients treated with anthracyclines.
Indeed, the subanalysis was quite positive. Atorvastatin significantly reduced the likelihood of experiencing elevated ECV.
Specifically, the odds of elevated ECV were 80% lower in the atorvastatin group compared with the placebo group (OR, 0.20; P = .002). Although 29% of patients in the placebo group showed elevated ECV, only 8% of those receiving atorvastatin exhibited this increase. Furthermore, an ≥3% increase in ECV was linked to an absolute decline of 8 percentage points in LVEF.
There was also a trend — albeit with small numbers — to suggest that ECV expansion induced HF events with 24% vs 8% in those without 3% ECV expansion.
The authors suggest that they found a mechanistic explanation for the benefits in statin therapy in the main trial.
Comments
I like the editorial from Dr. Giselle Melendez and Angelica Riojas.
There is no doubt that further studies are needed to address unresolved questions before statins can be widely recommended for the prevention of cancer therapy–related cardiovascular toxicity.
It remains to be determined whether the expansion of ECV can predict the long-term development of heart failure and, more importantly, whether statin-mediated prevention of cardiac fibrosis can halt the progression of LV dysfunction.
Good on the editorialists.
To me, an outside observer in this important field, I would say you have one negative trial and one positive trial. Just do a third trial. with more patients and longer follow-up. NIH should fund such a trial.
Use a clinical endpoint and have EF as a secondary endpoint. I worry about surrogate measures and mechanistic studies. They are fine for forming hypotheses, but the ultimate thing practicing doctors need are trials.
Next weekend I will see my colleagues in Vienna for European Heart Rhythm Association (EHRA) meeting. And I will have a preview of the big ACC meeting. Gosh, we have to figure out a way to separate these two important meetings.
© 2025 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Mar 21, 2025 This Week in Cardiology Podcast - Medscape - Mar 21, 2025.
Comments