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 May 17, 2024, John Mandrola, MD, comments on the following news and features stories: An Impella update, another trial of transcatheter aortic valve implantation (TAVI) vs surgical aortic valve replacement (SAVR), two studies on angina and percutaneous coronary intervention (PCI), another null substudy from REVIVED-BCIS, and semaglutide.
Update on Impella
When I wrote about and discussed DanGer-Shock I made note that this was a highly select patient population and the positive results for the Impella should not be translated to other indications, such as use in high-risk PCI.
Dr Mark Petrie, the senior author of REVIVED-BCIS, emailed me to tell me that his colleague Dr Divaka Perera is leading a trial called CHIP BCIS3 that will randomly assign patients with high-risk PCI to unloading with Impella vs not. The primary outcome will be a hierarchical win-ratio composite of death, stroke, myocardial infarction (MI), cardiovascular (CV) hospitalization, or periprocedural MI. This is great news.
TAVI vs SAVR in Low-risk Patients but With a Super Interesting Subgroup
The low-risk trial space is crowded. We have PARTNER 3, EVOLUT Low-risk, NOTION-1, and DEDICATE-DZHK6. Roughly speaking the results look really similar for the two procedures out to 4 to 5 years.
The NOTION-2 trial is the latest head-to-head comparison to be published. And it was presented at EuroPCR.
The trial looked at TAVI vs SAVR in especially low-risk patients. And, slightly more than one-quarter of the 370 patients had bicuspid aortic stenosis (AS).
The primary endpoint was death, stroke, rehospitalization at 12 months.
Average age 71 years, which is a few years younger than patients in the other trials. The average Society of Thoracic Surgeons (STS) score was super-low at 1.1.
The analysis plan was weird; it was non-inferiority (NI) for TAVI vs SAVR. I dunno. I think we have moved well past the time when TAVI is a new therapy.
Also curious was that the authors expected a 10% rate of events in TAVI arm and 15% in the SAVR arm. That, I think, is too high. I tell you about the estimates because it will inform the results.
In DEDICATE-DZHK6, it was 5.4% in the TAVI group and 10.0% in the SAVR group. And these patients were 4 years older patients and had higher STS scores.
The NI margin therefore was the difference between 10 and 15 or 5% in absolute terms. If the upper bound of the confidence interval (CI) was greater than that, then NI would not have been achieved. I am also curious why they did NI if they thought TAVI was superior.
To convert the NI margins to relative risk (RR), with SAVR as the standard intervention as it was here, then 5% margin translates into a RR of 15+5/15 = 1.33.
Here was the primary endpoint — focus on the confidence intervals.
At 1 the primary endpoint occurred in 10.2% in the TAVI group vs 7.1% in the SAVR group (absolute risk difference 3.1%; 95% CI, −2.7% to 8.8%; hazard ratio [HR] 1.4, 95% CI: 0.7 to 2.9; P=0.3)
They also presented an important subgroup –the primary endpoint incidence with regular AS vs bicuspid AS.
With standard trileaflet AS, the HR was 1.0 (CI 0.5-2.3) vs bicuspid AS where TAVI was nearly 4 times worse; HR 3.5 (CI, 0.8-18.5).
Comments. First of all, the wide CIs are sad. I say sad because this was an experiment on nearly 400 patients and with a RR going from TAVI is 30% better to 2.9 times worse, we simply cannot know.
The upper bound of the absolute risk difference (ARD) also went from -3% to + 9%, which is greater than the NI margin of 5%. So, we have to say that TAVI was not non-inferior to SAVR, but we can’t say anything else. Basically, the trial was inconclusive in the overall results.
Why the authors thought they would a) see this high of an event rate and b) TAVI would be this much better with nearly one-third bicuspids is quite strange to me. Had they estimated a smaller effect size or a lower event rate, then they would have had to recruit a lot more patients.
That said, the subgroup analysis of tricuspid vs bicuspid AS is interesting. TAVI looked much worse than SAVR for bicuspids. That goes against a fair amount of observational data, which has suggested TAVI does ok in well-selected patients with bicuspid AS.
Proponents of TAVI will argue that the lower bound of the 95% CI for bicuspid allows for TAVI to be 20% better. I’d counter that it also allows for it be 18 times worse.
Indeed, this is the problem with underpowered trials. We want to conclude: let’s not do TAVI in bicuspid given this subgroup analysis. But, since the CI are so wide it’s hard to say.
What I, an outside observer, would say is that before embracing TAVI for bicuspid valves, we should have a properly powered trial. Though, as a patient, I am not sure I would sign up for such a trial because SAVR looks so much better.
Angina and PCI
Two neat studies on angina from the group of Rasha Al-Lamee at Imperial College London Have been published in the Journal of the American College of Cardiology (JACC).
The first was a substudy of the ORBITA-2 study looking at two things: the relationship between presenting symptoms and disease severity and the ability of symptoms to predict the placebo-controlled efficacy of PCI; first author is Dr. Florentina Simader
Recall that ORBITA -2 was a randomized controlled trial (RCT) looking at the sham-controlled effect of PCI (without meds) to relieve angina. Patients with coronary artery disease (CAD) were stabilized on meds. Then meds were stopped. Then they had either PCI or sham. PCI reduced the primary end point, which was an angina symptom score that patients recorded on a novel smartphone app.
Briefly, ORBITA-2 differed from ORBITA-1. ORBITA-1 studied the placebo-controlled effect of PCI in patients who were maximally medically managed. It was negative. ORBITA-2 found that PCI alone, without meds, had a placebo-resistant effect.
The goal of this new sub-analysis was to further study symptom character to disease as well as PCI effect. It’s an important distinction because these days more and more people end up diagnosed in the cath lab not because of life-altering symptoms of angina, but diagnostic tests, like stress tests and coronary CT scans.
Symptom nature was separated into categories. One was the Rose angina questionnaire. To be honest I did not know about Rose angina. After I read the authors description, Rose angina is as typical as it gets. Exertional. Relieved with rest. Mid sternal. Etc. Everything.
They also used other angina descriptors. Things like the short form McGill pain questionnaire and Medical Research Council (MRC) dyspnea scale.
The results were interesting: There was little relationship between symptom severity and nature and anatomic or ischemic markers of disease severity.
Second finding: Symptom severity did matter. Patients with more severe angina were more likely to achieve (odds ratio [OR] = 4.3) a better placebo-controlled “health state.”
Patients who met criteria for Rose angina (vs those who did not) were more likely (OR 1.9) to have a placebo-controlled benefit from PCI on the angina symptom score and the number of angina episodes.
Patients with Rose angina were also more likely to have a placebo-controlled benefit of PCI on exercise time, CCS class, and domains of symptom and quality of life scores.
Weird though, Rose angina-patients did not have a placebo-controlled benefit of PCI on stress echo scores.
Patients who reported guideline-based angina (about two-thirds) also had a strong evidence of a placebo-controlled effect of PCI on angina symptom score and fewer angina episodes.
The other angina scores (McGill and MRC dyspnea score) did not have an association with placebo-controlled effects.
Comments. The authors wrote that they were surprised that symptom severity did not correlate with anatomic or ischemic severity but noted that this lack of correlation was not because presenting symptoms were not meaningful. Rather, it was the nature of the symptoms.
I don’t completely agree with their first statement, but their finding that the type of symptoms – typical angina — matters, is neat. And it is important. Here is why: at least in my zip code, oodles of patients get referred for angiography for all manner of symptoms. shortness of air, fatigue, atypical stuff, nonsustained ventricular tachycardia (VT).
And, once angiography finds disease, a stent goes in.
What the Imperial college team has shown, or suggested at least, is that these symptoms are unlikely to be relieved by PCI. if the patient gets better, it is likely placebo effect.
Back to my minor disagreement with their comments about *severity* of symptoms not correlating with disease severity. I can’t cite studies, but I recall, vividly, being taught back at Indiana from older cardiologists that you could not exclude left main disease based on mild symptoms. I remember cath lab conferences (many of them) where the fellow showed a runner or cyclist who could still put out big watts with a tight left anterior descending artery or left main. So, I am not surprised about the lack of severity correlation.
I do believe it’s immensely useful that they showed that typical angina (which, in my practice, is less common than atypical symptoms) is the only symptom that resists the strong placebo-effect from PCI.
These observations — or correlations if you will — likely explain the ORBITA-2 observation that nearly 60% of patients had continued symptoms in ORBITA-2 despite PCI totally relieving ischemia. Why? Because they had cath, angiography, diagnosis of CAD, PCI and, originally, their symptoms were not due to the CAD. If this happened in a careful RCT at Imperial College, imagine how often it happens in everyday US cardiology wherein doctors get paid to find and treat CAD, regardless of the nature of the symptoms.
This is a very nice paper and great use of data from a randomized controlled trial.
N of 1 Trial – the ORBITA -STAR Trial
The second study, ORBITA-STAR, was an N of 1 trial of 51 patients with CAD. It was led by Dr Christopher Rajkumar, also from Imperial College London.
The question was to determine whether independent symptom verification using a placebo (or sham) ischemic stimulus could distinguish which patients achieve greatest symptom relief from PCI.
Patients had single-vessel disease. In the cath lab, each patient had four episodes of 60 seconds of low-pressure balloon occlusion across the stenosis. These were randomly paired with four episodes of placebo or sham inflation.
Following each episode (real or placebo) patients reported the similarity of their symptoms compared with their normal symptoms.
To maintain blinding, the operator inflated and deflated a second “decoy” balloon positioned outside of the patient’s body to recreate the sound of a real balloon occlusion episode.
The symptom score had a range from -10 (placebo inflation replicated the symptoms more than balloon occlusion) to +10 balloon occlusion replicated the symptoms.
The primary endpoint was the ability of the similarity score to predict symptom relief later after the PCI.
Before I tell you the results, we should pause and consider how clever this is. No one does work like this. It’s impressive.
The results:
The median similarity score was 3 (interquartile range [IQR] 0.87-5.25). Pause there. I will come back to that. Remember that 0 is no correlation.
The similarity score was a strong predictor of symptom improvement following PCI: a patient with an upper quartile similarity score of 5.25 was significantly more likely to have lower angina frequency at follow-up after PCI (OR: 8.01; 95% credible interval: 2.39-15.86) than a patient with a lower quartile similarity score of 0.875 (OR: 1.31; 95% credible interval: 0.71-1.99).
Statistician Dr Frank Harrell from Vanderbilt was a co-author, and the study has beautiful Bayesian statistics. These reveal a 99.9% probability of a difference in how symptom score predicts angina relief after PCI.
I mention the stats thing, because, some may think, it’s only 51 patients. This would be a mistake because each patient had multiple entries in the symptom score. The primary endpoint was not binary but a symptom score marked by daily episodes of angina. So, Dr. Harrell constructed an analysis that took advantage of all these measurements to maximize the power.
Comments. These observations complement the secondary analysis of ORBITA-2. That is, if balloon occlusion replicates our patient’s day-to-day symptoms, he or she is likely to gain relief from the PCI. The more similar the symptoms the greater the probability of relief.
This, too, explains why so many patients in studies (and in practice) continue to have symptoms after PCI. Because the CAD is incidental. And the fault is ours for not diagnosing it properly.
These two studies provide the cardiology community with a wake-up call. We need to go back to basics. Does the patient have classic angina? If they do, and pills aren’t working, angiography and PCI will improve symptoms. Because the problem is like a tourniquet.
But for the vast majority with atypical symptoms and positive stress tests, mildly positive troponins, the CAD is likely incidental and best treated with disease modifying medicines and lifestyle changes, not metal in the coronary artery.
The late American writer Upton Sinclair pinpointed the challenge of translating these elegant studies when he famously said: “it is difficult to get a person to understand something when his salary depends on his not understanding it.”
Seriously though, I believe the next major breakthrough in cardiology will be when we figure out a way to explain chronic CAD to new students and patients alike.
The funny thing is that the late Bernard Lown was absolutely right about nearly all things related to CAD.
More Insights from REVIVED-BCIS
Another sub analysis of REVIVED BCIS has produced sobering results.
Recall that REVIVED BCIS randomly assigned the best possible patients for PCI. Those with left ventricular (LV) dysfunction, CAD amenable to PCI, and viable myocardium to PCI or simple pills. The main trial found no difference in the primary outcome of death or heart failure (HF) hospitalizations. There was also no improvement of LV ejection fraction (EF). No subgroup seemed to benefit. And no substudy has yet to reveal any way to find benefit from PCI over pills in ischemic LV dysfunction.
This most recent substudy looked at the impact of complete revascularization in these patients. Surely this would make a difference?
The authors measured completeness of revascularization with anatomic and viability-guided means.
All but 30 of the total of 700 patients had available data to assess the completeness of revascularization. In the main trial complete anatomic revascularization was recommended.
About half of the PCI group had complete revascularization whether assessed by anatomy or viability. So, the groups were at least similar in size. Chronic total occlusions often figured in the completeness of revascularizations. That is, the operator decided it was not doable so that patient had incomplete revascularization.
Compared with optimal medical therapy (OMT) alone, complete anatomic revascularization did not reduce the primary outcome (HR 0.90 with CI 0.62-1.32). As in the main trial, there was no association of complete revascularization with improvement of LVEF.
The authors also assessed complete revascularization with viability studies. Complete viability guided revascularization by PCI was not associated with a reduction in the primary outcome, HR 0.95, or any other secondary outcome.
A secondary analysis using late gadolinium enhancement transmurality cut off of 25% to define viability also found no interaction with the primary outcome in the group who had complete revascularization.
Comments. The authors state — and I did not know this — that no RCTs of complete vs incomplete revascularization have been done in chronic CAD. Previous studies, like COMPLETE and CULPRIT-SHOCK were in patients with acute coronary syndrome (ACS).
They also remind us that complete vs incomplete revascularization was not randomized in this study. It was up to a clinician. So, patients with incomplete revascularization were sicker and had more extensive disease.
Still. The healthier complete revascularization group did not do better. Gosh that is surprising to me. Complete revascularization did not improve outcomes over incomplete despite the fact that these were healthier patients at baseline.
One twist of this analysis was the viability guided revascularization. Here is why it was nifty: Many people think functional-guided revascularization is better than anatomy-guided revascularization. Proponents like the strategy because they think it’s more accurate. I like the strategy because it usually results in fewer stents (think FAME 1trial).
Anyway, the authors have this novel measure of viability-guided revascularization that assesses the completeness of revascularization relative to the extent of viable myocardium supplied downstream from the stenosis.
In this study, the degree of functional revascularization was high at about 85% and, as I told you, no difference in the primary outcome vs pills.
Final comments.
This is an observational study. It corroborates and strengthens the findings of the main trial.
My friends, I am shocked at how surprising and disruptive REVIVED BCIS is. These were perfect patients for PCI. Yet the authors have not been able to find any way for PCI to look like it improves outcomes over medical therapy.
How in the world does fixing proximal severe CAD not help these patients — even when revascularization is complete? I am asking. I don’t know.
But I do know this: The professional societies, the KOLs, the thinkers in cardiology ought to be out there educating people that heart failure is a medical—not ischemic disease.
This search for CAD in every new patient with LV dysfunction ought to be reversed.
But two years on from the main publication of REVIVED in 2022, my zip code and probably yours too, patients who have an abnormal LVEF on echo get a hot line to the cath lab.
Semaglutide
Some brief comments on two substudies of the secondary prevention SELECT trial of semaglutide vs placebo in patients with a) established heart disease and b) no diabetes. Both were presented at the European Congress on Obesity in Venice.
Recall that SELECT randomly assigned 17,000 patients to semaglutide or placebo and semaglutide reduced the primary endpoint of CV death, MI, and stroke by a statistically robust 20%; HR 0.80 (CI 0.72-0.90). The average length of follow-up was 34 months.
The first substudy, also published in Nature Medicine, looked at weight loss based on body weight, anthropometric outcomes, and safety and tolerability by body mass index (BMI).
The follow-up is now longer at 4 years. The main findings here were that weight loss was sustained, about 10% over placebo, more in females, which the authors thought might be due to higher drug exposure.
The report also noted meaningful weight loss in all races, body sizes and regions. Notable was that over 40% of those in the semaglutide arm no longer had obesity at 2 years. Recall that these are patients with established heart disease.
As for adverse events (AE), in the main trial, an AE leading to permanent discontinuation of the drug was 16% vs 8% with placebo.
In this study, the authors saw a graded increase in the proportion discontinuing semaglutide, but not placebo. For lower BMI classes, discontinuation rates are higher in the semaglutide group but not the placebo group.
This paper adds knowledge. Semaglutide works in mostly everyone. Weight loss is sustained when patients remain on it. Note that in SELECT there was not aggressive up-titration as it was a CV outcomes trial not a weight loss trial.
The second study was more interesting, and well covered on theheart.org | Medscape Cardiology by journalist Becky McCall. Dr John Deanfield reported results of SELECT based on baseline weight as well as weight change after 20 weeks of treatment.
First, they found no relationship between CV outcomes benefit based on baseline measures of adiposity.
Second, and most interesting, was semaglutide reduced outcomes even when it did not reduce weight by that much.
For this analysis, the authors split the semaglutide group into two groups — those with more than 5% body weight loss and those with < 5% body weight loss.
The HR for CV reductions was lower in both, albeit the group with more than 5% loss had a 33% reduction in outcomes and those with less than 5% loss had a 15% relative risk reduction (RRR) that barely made significance.
Comments. As with the REVIVED substudies, these bolster the main findings. How you feel about semaglutide depends on what kind of hat you are wearing.
If you are a Neutral Martian clinician, you look at SELECT and see that semaglutide is a disease-modifying drug, much like statins. The risk reduction in major adverse cardiac events is similar. Statins are mandated as secondary prevention agents and SELECT supports the addition of semaglutide, although a skeptical Neutral Martian would want a confirmatory trial.
So, I guess my thoughts on semaglutide as a statin-like disease modifier turn on the assumption that SELECT would be confirmed in a second trial.
But if you wear the hat of a pragmatic person who considers things beyond trials, you have a more complex view of GLP-1 drugs. You think, wait, at these prices, and with this much heart disease and obesity, we cannot easily afford these drugs. You also think, wait, if we achieved these same weight loss gains with lifestyle, could we not approach this level of MACE reduction.
These are serious tensions. And I am not yet sure how I feel about 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: May 17, 2024 This Week in Cardiology Podcast - Medscape - May 17, 2024.
Comments