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 April 15, 2022, John Mandrola, MD comments on the following news and features stories.
VT Ablation
VT Ablation After First ICD Shock Boosts Survival: PARTITA
Two small recently published randomized controlled trials (RCT) got the electrophysiology (EP) community excited about “early” ventricular tachycardia (VT) ablation. Both trials had topline “positive” results but, while I love RCTs, and wish we did more of them, these two studies should not change our cautious approach to VT ablation. There are just far too many limitations.
As background, I will discuss VT ablation in patients with substantial structural heart disease, usually ischemic. This differs substantially from ablating VT in normal hearts, which is less risky and curative—like supraventricular tachycardia (SVT) is. For the non-cardiologists out there, scar in the heart provides the three components of reentry tachycardia: multiple pathways, unidirectional block in one pathway, slowed conduction in another. Think of a scar with many channels of surviving myocardium.
Ablating VT in patients with structural heart disease confers added risk because patients usually have systolic dysfunction. They may not tolerate VT; prolonged procedures with saline-irrigated catheters can lead to heart failure (HF), and stroke is also possible in dilated ventricles.
That said, implanted cardioverter-defibrillator (ICD) shocks are bad. They hurt and can cause PTSD. Shocks are associated with worse prognosis. Preventing a patient with an ICD from getting shocks is a good goal.
PARTITA: At the American College of Cardiology (ACC) meeting, Professor Paolo Della Bella presented results of the PARTITA trial; Circulation simultaneously published the paper. PARTITA was billed as a two-phased trial, but it is really a very small trial comparing VT ablation in patients who received a shock vs continuation of standard care.
Phase A, as they called it, involved about 500 patients who had an ICD placed. Of these, 11% or 56 patients received an appropriate ICD shock for VT; 47 of these of these patients agreed to be randomly assigned to ablation vs standard care. That is right: 23 in the ablation arm, 24 in the standard arm. Eighteen of the 23 got ablated.
The primary endpoint had nothing to do with VT. It was instead a composite of death from any cause or hospitalization for worsening HF. Right off the bat you should be thinking, wait, VT ablation is a therapy designed to reduce VT or stop ICD shocks. Would we not want to measure that?
After 2 years, the primary end point occurred in one of 23 (4%) patients in the ablation group and 10 of 24 (42%) patients in the control group
The hazard ratio (HR) was a whopping 0.11 or 89% reduction. Here’s the problem: the 95% confidence interval (CI) ranged from a 99% benefit to 15% benefit.
The P-value squeaked under the threshold at 0.034.
There were no deaths in the ablation arm and eight deaths in the standard care arm. This drove the composite endpoint.
HF hospitalizations (HHF) were lower in the ablation arm but it was not significant.
Here’s the thing: cardiac death, recurrent VT, VT with anti-tachycardia pacing (ATP), or VT with shocks did not differ. (Editor's Note: The occurrence of VT with shocks was lower in the ablation arm: 2 vs 10; P = .039.) You might wonder how a procedure that does not reduce VT or cardiac death drives overall mortality. I did too.
So, I went to the supplement where they listed the causes of the eight deaths: Deaths in the standard arm included one for skin cancer, one for sarcoma, two “non-cardiac,” and one for sepsis. That is, five of the eight mortality primary outcome events driving the difference had absolutely no relation to the ablation.
You don’t have to be a stats expert to know that taking five events out of the standard care arm would have easily made this a nonsignificant finding. Also, it seems strange that there are no deaths in the ablation arm. In a population of ICD patients with a mean ejection fraction (EF) of ~ 32% and a mean age of 72 (Table 1) with or without VT and with or without ablation, you would NOT expect 100% of patients to survive for 24 months. This isn’t nefarious, it simply highlights the problem with small numbers.
Finally, the authors report zero complications in the ablation arm. Which differs substantially from the next study. The authors write that “PARTITA provides evidence that catheter ablation should be considered after the first ICD shock in patients with cardiomyopathy.” I don’t think so, but ....
SURVIVE VT: The Journal of the ACC (JACC) published the SURVIVE VT trial from multiple centers in Spain. The trial compared VT ablation using substrate-based ablation or anti-arrhythmic drugs (AAD) in patients with ischemic cardiomyopathy who had an ICD shock.
A word on substrate ablation: this means we identify the scar tissue and ablate all late or fractionated potentials in that area. There is not an attempt to induce or map one or more specific tachycardias. Some have called it homogenization of the scar. The idea is that by ablating the entire scar you ablate all possible channels for reentry.
SURVIVE VT had 144 patients, aged 70 years, mean EF 35% and about 40% of patients had EF’s less than 30%. Again, we get a curious endpoint. Not VT or just recurrent ICD shocks. Instead, the authors chose a primary outcome of a composite of cardiovascular (CV) death, appropriate ICD shock, unplanned hospitalization for worsening heart failure, or severe treatment-related complications.
It’s the latter component, severe treatment related complications, where the issue comes to play. Here are the results: A primary outcome occurred in 28% of patients in the ablation group and in 46% of those in the AAD group (HR, 0.52; 95% CI: 0.30-0.90; P = 0.021).
So boom, it’s positive. Wait a minute, let’s look at the four components of the primary outcome. Let’s see what drove the results.
CV death: Three in each arm.
App ICD shocks:12 v 13 in each arm, so no difference.
HHF: 8 vs 13 in each arm, no significant difference.
The positive primary endpoint was driven by “severe treatment-related complications;” 7 vs 21 (HR 0.30).
What are these treatment-related complications? In the AAD, examples of “severe” issues: sinus bradycardia in two patients, dyspnea in another, low thyroid in another. Many of the other severe events were slower VTs that required hospital admission and subsequent ablation.
Also of note: there were 10 major complications in the ablation arm. And 10/71 is a 14% rate of major complications. Keep in mind that this is the rate in the confines of a carefully controlled clinical trial with sites selected for their skilled operators. Some of the secondary outcomes are also interesting: Of note, total mortality, appropriate ICD therapies, and any documented VT were all similar.
The authors concluded that “in ICD patients with ischemic cardiomyopathy and symptomatic VT, catheter ablation reduced the composite endpoint of CV death, appropriate ICD shock, hospitalization due to HF, or severe treatment-related complications compared to AAD.”
And they write this in the discussion: “Based on these results, catheter ablation should be strongly considered before AAD as a first-line strategy for patients with ischemic cardiomyopathy and ventricular arrhythmias.”
I don’t think so. If VT ablation works, it is supposed to stop VT, which reduces ICD shocks and CV death. That is not what happened. Appropriate ICD shocks, CV deaths, total deaths, and any VT were all similar.
Note that a much stronger VT ablation trial, called VANISH, enrolled 260 patients who had VT despite AAD. Here the choice was more AAD vs ablation and ablation barely proved beneficial, with a P-value of 0.04.
In the SURVIVE supplement, the authors report results of their trial using the endpoint of VANISH, which was death, ≥ 3 episodes of VT within 24 hours, or appropriate ICD shock. Using this endpoint, SURVIVE was null.
Technically, SURVIVE was a positive trial, but it had a 14% rate of major procedural complications, and was driven by a dubious endpoint of serious treatment related complications. I say dubious, because:
This is not a blinded trial, and it’s done at ablation centers. There is likely performance bias.
Sinus brady, dyspnea, and low thyroid in an amiodarone patient are hardly severe complications.
Of course, you’d expect more incidents of slow VTs in AAD arm. That is not a severe complication. This is what AADs do. If the patient has a slow VT and is symptomatic, it can then be ablated. Many patients don’t get slow VTs on AADs, especially after one shock.
What’s more, and this worries me a lot, SURVIVE authors don’t tell us how many patients they screened over 7 years in these nine centers to get these 145 patients. It makes you wonder how representative these patients are to the average patient who gets an ICD shock in real-world practice.
Back to PARTITA, we have roughly 20 patients in each arm and the results are driven by non-cardiac deaths.
In sum, VT ablation has a role in the treatment of patients with significant scar-related heart disease. I do VT ablation. It can be quite rewarding. But it is the most hazardous procedure we EP docs do. It deserves huge doses of caution. Neither PARTITA nor SURVIVE suggest we be more aggressive with this complicated and risky procedure.
Alerts for HF Care
Alerts Improve Prescribing of HF Quadruple Therapy: PROMPT-HF
One of the biggest issues with the care of patients with HF is getting patients on the four guideline-directed medical therapy (GDMT) drugs and getting the doses titrated to the best possible dose. This is job for the out-patient clinic, and everyone who has practiced in an outpatient clinic knows it is a busy place. There is inertia to mess with meds. Messing with meds creates extra work. Extra work means running behind.
Clinicians want to do the best by their patients, but clinicians are also human. We don’t have an endless supply of energy. Well, an energetic, seemingly tireless research group at Yale is interested in the implementation process. They presented their latest work at ACC and JACC published the PROMPT-HF trial. PROMPT-HF was a pragmatic, cluster-randomized trial of 100 clinicians. Recall that cluster RCTs randomly assign clinicians not patients. Here, one group got electronic health record (EHR) alerts and the other did not. The alert notified clinicians of GDMT recommendations.
The primary endpoint was simple: an increase in the number of GDMT classes of drug prescribed at 30 days. And it seemed to work.
The primary outcome occurred in 26% participants in the alert arm vs 19% in the usual care arm, increasing GDMT class prescription by > 40% after alert exposure [adjusted RR: 1.41 (1.03, 1.93); P=0.03].
Nearly 80% of alerted providers agreed that the alert was effective at enabling improved prescription of medical therapy for HF.
Comments: PROMPT-HF is an outlier. Most efforts at improving quality have not worked. The authors cite at least four earlier studies that have failed to show that alerts or quality improvement programs improve outcomes, including one from their group. Perhaps this EHR alert worked because it was simple and the only alert the clinicians got. Perhaps it was because it was designed with clinician imput. Perhaps it worked because it was in the out-patient system.
Yes, there are limitations.
The trial was done in one system of academic centers.
The Yale authors seemed like serious champions of EHR alerts.
EHR alerts without said champions might not have such an effect.
But this is a start. I really like this work. Of course, it is important to try to improve systems of implementation. But what really excites me about the Yale group is that they are studying these efforts empirically. This cannot be under-stated.
Empirical study of policy may be more important than drugs or devices. So much of the quality movement has simply been foisted on clinicians because it made sense — at least to policy wonks
This week, my friend Lisa Rosenbaum, the national correspondent for NEJM, wrote a brilliant piece titled “Reassessing Quality Assessment — The Flawed System for Fixing a Flawed System.” It is a must-read.
Triple Vessel Disease Care
FAME 3 Subanalysis Adds Twist to Negative Primary Results
The FAME-3 investigators presented a substudy of the main trial at ACC, and Circulation simultaneously published the results. It’s a head-shaker and I am going to try to remain measured.
Recall that FAME-3 tested noninferiority of fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) vs coronary artery bypass graft (CABG) in patients with three-vessel disease.
The results on the primary endpoint (PEP) of a major adverse cardiac or cerebrovascular event (MACCE; defined as death, myocardial infarction (MI), stroke, or repeat revascularization) were decisive.
The 1-year incidence of a PEP was 10.6% in FFR-guided PCI and 6.9% with CABG (HR, 1.5; 95% CI, 1.1 to 2.2),
Note, the upper bound of the CI was 2.2, meaning that FFR-PCI could be up to 2 times worse than CABG.
That did not even come close to meeting non-inferiority.
FFR-guided PCI was clearly not noninferior to CABG.
Now the authors are back with a sub-analysis looking at quality of life (QOL) as measured by a European QOL–5 Dimensions (EQ-5D) questionnaire. The results at 1 year were nearly identical. No difference.
The authors then tell us, “The trajectory of QOL differed. PCI QOL was better at 1 month.” But this is like saying you will get wet riding a bike in the rain. Of course, CABG has a tougher short-term recovery. Everyone knows this.
The other secondary finding they note is that the proportion of patients with class 2 or greater angina at 12 months was 6.2% in PCI vs 3.1% CABG. The odds ratio (OR) is 2.5. So more than 2 times greater chance of significant angina.
However, the authors conclude that, “The rate of significant angina was low in both groups and not significantly different.” That is because the CI of that OR of 2.5 worse for PCI went from 0.96 to 6.8.
The take-home from the investigators was that this data — primarily the faster recovery from PCI — might be important for shared decision making.”
Here is my take-home: If you choose FFR-PCI over CABG, you will have a higher rate of MACCE, no difference in QOL at 12 months, and a 2.5 higher odds of class 2 angina.
Why can’t we accept that CABG is the better strategy for three-vessel disease? FAME-3 confirmed the results of SYNTAX done a decade before. That this sort of post-hoc study gets into Circulation and is spun as a new twist is disheartening.
PCI and its iterations are awesome. It has utterly transformed the care of acute coronary syndrome and MI. Why is this not good enough?
Treating Drug Side Effects with More Drugs
DIAMOND: Adding Patiromer Helps Optimize HF Meds, Foils Hyperkalemia
Speaking of disheartening, I have to tell you about the story of the DIAMOND trial of patiromer. This was a late-breaker at ACC. There is not yet a paper.
Patiromer is a novel potassium (K+) binder that is used to lower K+ and maintain normal K+ levels. The DIAMOND trial was designed to assess whether patiromer in patients with HF with reduced EF, with current or a history of hyperkalemia due to renin-angiotensin-system inhibiting (RASi) therapy, can simultaneously control K+ levels, enable optimal RASi therapy, and improve clinical outcomes. A 30-day supply of the drug retails for about a $1000.
DIAMOND enrolled symptomatic patients with HFrEF who had hyperkalemia or history of hyperkalemia while on RASi, which led to a dose reduction or discontinuation. Since the trial was treating patients with serious illness and the drug is designed to allow patients to stay on life-saving drugs, the trialists set out to show patiromer reduced CV death or hospitalization, which is an excellent outcome. Alas, the investigators changed the endpoint to changes in serum K+ levels.
After randomly assigning about 400 patients per group, the primary endpoint was significantly better in the patiromer arm. The P-value was < 0.001. Your next question should be, Okay Mandrola, what was the effect size?
Answer: it was 0.10. Yes, I said that right. The $1000 per month drug, which is used to treat a side effect from another drug, reduced the K+ levels by 0.1. CV deaths were now secondary endpoints and showed absolutely no difference.
This led to the final bullet point: “These data suggest the potential for improved outcomes with patiromer enabled optimization GDMT in patients who are unable to tolerate optimal RASi treatment due to hyperkalemia.
The principal investigator of the trial told Steve Stiles “It stands to reason that optimization of RASi therapy without a concomitant risk of hyperkalemia may, in the long run, lead to better outcomes for these patients,"
Comments:
This reads like something out of the book, House of God. A drug creates an adverse effect, and we give another drug to treat that effect.
Then the pharma-sponsored trial changes a real endpoint to a Mickey Mouse endpoint guaranteed to be positive.
Then the P-value is super positive but he effect size is literally miniscule.
Finally, I am not saying a K+ sequestering agent is useless, but come on, patients with HF already have to take tons of pills and if you want me to prescribe a drug to treat another drug side effect, you have to show me outcomes data.
© 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: Apr 15, 2022 This Week in Cardiology Podcast - Medscape - Apr 15, 2022.
Comments