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.
In This Week’s Podcast
For the week ending November 19, 2021, John Mandrola, MD comments on the following news and features stories.
Next week is Thanksgiving break. This Week in Cardiology will return December 3.
Before I get to the American Heart Association (AHA) 2021 meeting review, I want to say thank you to all who have sent me comments. I talk into this microphone each week and wonder how it is received. Both positive and negative feedback is appreciated. I also want to thank my colleagues and editors at theheart.org | Medscape Cardiology for their support and mentorship in helping me learn critical appraisal.
Now to the AHA review. I won’t be able to get to all the studies in one podcast so look for AHA Part 2 the first Friday of December.
AVATAR was a randomized controlled trial (RCT) of early surgery vs watchful waiting in patients with severe aortic stenosis (AS). Patients had to pass an exercise test where they achieved 85% maximum heart rate, which ensured that they were actually asymptomatic. Circulation published the paper.
The study enrolled 157 patients, mean age 67 years. The cause of AS was degenerative in more than 80% and bicuspid in only 14%.
Over a median follow-up of 32 months, the primary endpoint (PEP) of all-cause death, myocardial infarction, stroke, or heart failure hospitalization occurred in 16.6% patients in the early surgery arm vs 32.9% in the conservative arm; hazard ratio (HR) of 0.46 (95% CI, 0.23 - 0.90; P =.02).
The Kaplan-Meier curves for the primary endpoint separated at about a year.
There were some peculiar observations regarding the components of the composite:
The PEP was driven by all-cause death; 16 deaths in the conservative arm vs nine in the early surgery group.
But cardiovascular (CV) death rates were not different (approximately 9% in both arms).
This is peculiar because if early surgery helped young patients with AS, you’d expect it do so by reducing cardiac death.
My translation of this important trial is that you have to go deeper than the topline results. On the surface you have a significant 54% reduction of the PEP, and results in AVATAR are similar to the South Korean early surgery vs watchful waiting RECOVERY trial. Sounds like a no-brainer, right? I don’t think so. First, the results are fragile. If only two patients in the active arm converted from not having the PEP to having the PEP, the study would lose statistical significance. Look at the differences in death. Why, if early surgery was so beneficial, wasn’t CV death lower? It turns out, the authors tell us in the supplement, that three patients in the watchful waiting group vs none in the early surgery group died of viral pneumonia. That surely seems due to the play of chance.
The other issue translating AVATAR and RECOVERY to practice is generalizability or external validity. These were young mobile patients with AS. Most patients I see with tri-leaflet degenerative severe AS are older and have more co-morbidity. I think we should be very careful applying fragile results to the typical older patient with AS. Why? Because older patients are going to have a much tougher time with surgery.
Preventive medicine always means doing something to people who have no complaints. Since all interventions have a risk of harm, before we do intervene, we should have a high bar that benefits are greater than harm. When the preventive intervention involves sternotomy and cardiopulmonary bypass, that bar should be much higher than these two small trials. In fact, the AVATAR authors called for this in their paper, writing that these “findings will require further confirmation in a larger study."
I know what your rebuttal may be, because one of my structural heart partners reminded me that transcatheter aortic valve replacement (TAVR) is different from surgical AVR. TAVR is less invasive and yes, maybe you don’t want to subject an older patient to surgical AVR, but TAVR represents a less invasive option. This makes sense, but I would argue it is an entirely testable hypothesis. Eighty-year-old patients with severe asymptomatic AS: randomize to TAVR or watchful waiting and measure outcomes.
Another big trial at AHA also studied a heart-surgery question, one that I have always wondered about. When patients go for mitral valve surgery for degenerative mitral regurgitation, should the surgeon also do a tricuspid repair? I wonder about this because electrophysiology is consulted for pacemakers after heart surgery.
NEJM published the trial from the CTSN investigators. The study enrolled 401 patients. Patients had to have moderate tricuspid valve regurgitation (TR) or, if less-than-moderate regurgitation, annular dilatation of 40 mm or more. The active arm was mitral surgery plus tricuspid annuloplasty (TA) and the control arm was mitral surgery alone.
The key feature of this trial was the PEP, a composite of death, reoperation for TR, or progression of TR by two grades from baseline or the presence of severe TR. Thus, you have two hard binary endpoints (alive or dead; reop or not) and an ultrasound estimation of a TR jet.
The primary endpoint occurred in 10.2% of patients who underwent mitral-valve surgery vs 3.9% who underwent concomitant tricuspid annuloplasty (relative risk, 0.37; 95% CI, 0.16 - 0.86; P = .02).
But the endpoint was driven exclusively by less TR progression in the repair group; just 0.6% of the annuloplasty group had TR progression compared with 6.1 % of the surgery-alone group.
The other two components of the PEP: zero re-ops in either group and no significant difference in deaths.
Here is the kicker: The rate of permanent pacemaker implantations was 2.5% with surgery alone vs 14.1% with concomitant tricuspid annuloplasty—a greater than 5 times higher rate. That is because sewing in an annular ring involves placing a stich in close proximity to the conduction system.
I wrote a column on this trial. Proponents would say it’s a positive trial. Tricuspid valve repair reduced TR and that is a good thing because TR is bad, and reoperations for TR are especially high risk. Better to prevent TR.
I disagree. This is an underpowered trial for hard endpoints. TR is a terrible endpoint because it is susceptible to measure error, dependent on pre- and after-load, and no patient ever has complained to me that his or her TR jet was abnormal.
For the price of a better ultrasound image, there was a 5 times greater rate of pacemakers in a group of patients younger than 70 years old. As a pacemaker implanter, I say this is a bad tradeoff. You want to hold off on pacers for as long as possible.
Here is the thing: the trial could have been larger. The same CTSN recruited more than 2000 to study rate vs rhythm control for post-op AF. Whitlock and his group randomized more than 2300 for left atrial appendage closure at the time of surgery. I have heard Vanderbilt stats professor Frank Harrell say that underpowered trials that cannot answer a clinical question are problematic. Why? Because a clinical trial is an experiment on humans. If you are going to do that, you ought to strive to get a clear answer.
If you are going to increase the odds of needing a pacer because of AV block, we need a trial with more patients who are followed longer. Until then, in the average patient with moderate TR or a dilated annulus, we ought to favor the conservative, less is more approach.
Post-cardiac surgery atrial fibrillation (AF) is one of the toughest problems in all of cardiology. In two decades of practice, everything I have seen tried has failed to prevent AF after heart surgery. I have come to believe the only way to avoid AF after heart surgery is not to have heart surgery! The PALACS RCT randomly assigned cardiac surgery patients to a posterior pericardiotomy or no intervention. Lancet published the paper. Investigators screened 3600 patients but randomly assigned only 420. They excluded those undergoing mitral or tricuspid valve surgery or with a history of previous atrial arrhythmias. These were mostly CABG and AVRs.
Posterior left pericardiotomy is a simple surgical procedure that connects the pericardial sac with the left pleural space and drains fluids and thrombi from the pericardial cavity in the postoperative period. The idea is that even a small amount of fluid in the pericardium is causal in AF.
AF occurred in 17% in the posterior pericardiotomy group vs 32% in the control arm; The odds ratio (OR) was 0.44 with a confidence interval (CI ) from 0.27 to 0.70.
The incidence of postop pericardial effusion was lower in the posterior left pericardiotomy group than in the no intervention group; 12% vs 21%.
Post op adverse events were not different.
No complications related to the pericardiotomy were seen and there were only 4 extra minutes of pump time.
Sounds amazing, right? Well, this was a single-center proof of concept study. One surgeon performed the procedures in nearly 80% of all the patients in the trial. The investigators are already planning another bigger trial. Good on them.
I love this study for many reasons. One is that the trial received no funding. The surgeons at Weill Cornell Medical College were curious, and instead of just doing things willy-nilly, they studied it with proper controls. We in electrophysiology could learn from this. Take the ablation approach to persistent AF. Some do just pulmonary vein isolation (PVI), other perform PVI plus posterior wall, others PVI plus triggers, others PVI plus lines. Whenever there is an expert panel on the ablation approach to persistent AF, there are almost as many approaches as panelists.
We’ve been ablating AF for decades now and all it would take is a couple of systematic trials. There have been a handful of trials, notably, STAR-AF-2 from Dr. Atul Verma, but not many others. The point is that if these heart surgeons had simply done posterior pericardiotomies and then noticed less AF, then went back and did an observational chart review, you’d never know the answer, no matter how many statistical adjustments and propensity matching maneuvers you did. By doing an RCT, you increase the chance of detecting true differences b/c randomization balances the known and unknown patient characteristics.
Left Main CAD
In a featured science presentation at AHA, Marc Sabatine from the TIMI group at BGH presented results of an independent analysis of the left main PCI vs CABG trials. The Lancet published the paper. The reason for the meta-analysis stems from controversy surrounding the EXCEL trial—the largest of the four left main PCI vs CABG trials.
The three EXCEL controversies:
The signal of increased death in the PCI arm, which was not seen in NOBLE, the other big left-main RCT. EXCEL authors posited that the higher death rate was due to chance because they found no difference in CV death.
The second debate around EXCEL stemmed from the authors’ biomarker-dependent definition of myocardial infarction (MI), which favored PCI because the number of periprocedural MIs was higher with CABG.
The third and perhaps biggest issue was that the EXCEL trial protocol called for reporting MI rates by universal definitions of MI (UDMI), but they did not include these in either paper. This was discovered because the trial data was leaked to the BBC News. Had there not been an insider who leaked the data, it’s possible no one would have known. MI rates defined by UDMI favored the CABG arm.
So, the way to resolve this issue: get an independent review. The meta-analysis took patient-level data from the four trials—Syntax, PRECOMBAT, NOBLE and EXCEL. Follow-up was 5 years. They chose all-cause death as the PEP and reported it in both standard frequentist terms with P-values and in Bayesian terms.
Death was 11.2% in the PCI arm vs 10.2% in the CABG arm. The HR was 1.10 and 95% CI ranged from 0.91 to 1.32; P = 0.33.
In their Bayesian analysis, the probability of any mortality benefit of CABG over PCI was 85.7% and the probability that the mortality advantage of CABG was at least 1% in absolute terms was 49%.
Spontaneous MI was 2.3 times more likely in the PCI group. Stroke was not significantly different at 5 years.
Repeat procedures for revascularization were 78% higher after PCI
And for the broader composite outcome of death, MI, stroke, or revascularization was 31% higher for PCI vs CABG.
I see three main takeaways:
The first is that this a breakthrough in medical-evidence communication. TIMI investigators, the group of Eugene Braunwald, from Harvard, reported an important finding using a Bayesian analysis. This is HUGE, because as clinicians, we love the Bayesian point of view; we want to know the probability of benefit given the data. The traditional approach with P-values tells you how surprising the data are given the made-up hypothesis of no difference between treatments.
Clinicians don’t make up null hypotheses and calculate the probability of the data given no difference. We look at event rates in two arms and want the probability of benefit given the results. In this case, there is an 85% probability that mortality is lower with CABG and a 50-50 chance that CABG produces a 1% absolute risk reduction in death.
I encourage you all to look at the PDF in Figure 5 in the supplement. It shows how they get these percentages. This can be done easily with most trials, and I think the coolest aspect is you can decide what you believe the minimum clinically important benefit is. It might a 1% reduction, it might be a 5% reduction. Whatever it is, the PDF of the Bayesian approach can tell you what the probability of that is.
The second takeaway is that given the average patient who was eligible for either PCI or surgery and randomized in these trials, CABG is the better strategy--on paper. At 5 years, CABG delivers an 85% prob of lower mortality, less MI, less repeat procedures, and no difference in stroke.
Patients who have left main disease ought to come off the cath lab table for a discussion about options. Not “Sir, you have blockages, and I can fix them.”
Which leads to my third takeaway, which is that even though the odds favor CABG, there is a clear difference in tradeoffs. PCI is way easier on people in the short run. Some patients may decide that surgical outcomes aren’t that much better. They will take the band-aid on the wrist and discharge in one day. Some patients will have subtle anatomical or clinical features that may favor one therapy. A neutral heart team approach would be ideal, but this is easier said than done.
The Friday after Thanksgiving I will talk more about key AHA sessions: The AMAZE Trial – epicardial left atrial appendage (LAA) closure with LARIAT added no benefit to AF ablation; Greg Marcus and colleagues presented two studies on AF triggers—including, you guessed it, caffeine; a study looking at the effect of preventive aspirin on dementia.
I’ll discuss the failure of electronic health record alerts about heart failure prognosis to change outcomes or clinician practice—a nonsignificant trial that teaches us a bunch.
There were also two studies on SGLT2 inhibitors that were both positive but had huge caveats. The teaser is that this class of drugs can deliver big value in some patients, but I am afraid, in other classes of patients, the value is likely low. It should come as no surprise that the same treatment can be on both the steep and flat part of the value-utility curve.
© 2021 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Nov 19, 2021 This Week in Cardiology Podcast - Medscape - Nov 19, 2021.