Aug 18, 2023 This Week in Cardiology Podcast

John M. Mandrola, MD


August 18, 2023

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 August 18, 2023, John Mandrola, MD comments on the following news and features stories.

First a note of thanks. #TWICpodcast received a number of comments on the site, and a few reviews on the Apple Podcast app. Thanks for taking the time to write. It’s nice, and it helps us reach more people. Another note — I am trying to name authors. I am often bad at pronunciation. I am sorry.

ICD Generator Change

We start today with one of the most common and important decisions I make in the clinic: The ICD generator change. I am not an outlier. Nearly one in four ICDs implanted in the United States are implanted during generator changes.

This may seem like a niche topic, but it is not. This may be the most important 5-minute topic I’ve done all year. The ICD gen change is a massive blind spot in cardiology. A very nice paper stimulated my interest, but first some background.

  • The ICD, or implantable cardioverter defibrillator, has been a great innovation in cardiology. Seminal trials done 20 years ago showed robust mortality reductions vs medical therapy in patients with heart failure (HF), though most of the benefit occurred in patients with ischemic rather than non-ischemic causes of HF.

  • Patients in these trials were relatively young and free of co-morbidity. They were ambulatory outpatients with HF.

  • What I will say now seems almost too obvious to state. Patients with ICDs age. As humans accumulate birthdays, they also accumulate co-morbidities. Battery life is about 8 to 10 years after a primary prevention ICD is implanted.

  • And 8 to 10 years after the initial ICD implant, at the time of battery depletion, the health situation is often changed. For instance, the risk of ventricular tachycardia (VT) and ventricular fibrillation (VF), the risk of the procedure, quality of life, and co-morbid conditions all may differ from when the primary implant was done.

Here is just one example: The left ventricular ejection fraction (LVEF) may have improved since the initial ICD implant so that patients would no longer meet eligibility requirements for a new ICD. Also, what if the patient has had zero therapies? What if it’s there third generator change and still no VT therapies?

Yet, what happens way too often, is patients are told their ICD has a low battery and they “need” a new one. They are often told it is an easy operation. And it sort of is, except the downside risk is far greater during a generator change. Guess why.

The reason is that if an infection is found during the generator change, the 10-year-old leads have to be extracted. That is a big procedure, requiring docs and labs and staff with a lot of experience. I like to say it is a low-risk event with asymmetrically terrible consequences.

What’s crazy is that there is not a shred of evidence to support ICD generator surgeries. Exactly zero randomized controlled trials (RCTs) have been done. I propose that this may be one of cardiology’s biggest blind spots.

Maybe you don’t believe me about the equipoise. Let me cite some evidence for the equipoise:

  • First consider ICDs in nonischemic cardiomyopathy (NICM). Most of the data supporting this is weak. MADIT2 was all ischemic CM. 

  • The DEFINITE trial of ICD vs medical therapy in NICM did not reach significance.

  • The SCD-HeFT included about 50% NI pts but that subgroup did not meet statistical significance.

  • The strongest evidence against an average benefit was DANISH in 2016. They enrolled 550 pts with NICM and found absolutely no mortality difference of ICD vs conventional HF therapy.

And I have one more, a reference that I did not recall.

  • Jeanne Poole, in the Journal of the American College of Cardiology (JACC) 2020, reported 11-year follow-up from SCD-HeFT, just when you’d be considering a generator change for a primary prevention ICD. While the mortality reduction held up, on average, there was a decrease in effect size after 6 years.

  • “When the treatment benefit was examined post hoc, as a function of time since randomization, we observed an attenuation of the ICD benefit after 6 years (≤ 6 years hazard ratio [HR]: 0.75 [95% confidence interval (CI): 0.64 to 0.88] and > 6 years HR: 1.09 [95% CI: 0.90 to 1.32]; P value for the interaction = 0.0015).”

You might still argue with me that we can’t study this because patients are attached to their ICDs. You can’t just not replace it. It would go against their wishes.

Ah, but here I have a new study to tell you about. First author, Sarah Montembeau, from Emory University, published in Circulation Quality and Outcomes.

Their question was how patients make ICD generator change decisions.

  • They recruited patients from 2013 to 2021 to complete in-depth interviews about their ideas on generator changes.

  • The interview included both open and closed ended question. This focused on patients’ subjective understanding of the generator change, the factors driving their decision, and the patients willingness to forgo a generator change based on risk of sudden cardiac death (SCD) and also their clinicians’ recommendation to not have the procedure.

  • To investigate benefit thresholds for pursuing generator exchanges, patients were presented standard gamble-type hypothetical scenarios where their 5-year risk of SCD at that time varied (10%, 5%, and 1%).

  • The survey included 50 patients. Most were older than 65 years; 17 of the 50 had improved their EFs.

  • As sudden cardiac death risk decreased from 10% to 5% to 1%, the number of participants willing to undergo a generator exchange decreased from 48 to 42 to 33, respectively.

  • When a clinician recommended against the generator change, only 17 of 50 sill wanted it.

  • Of the 33 who said they get a generator change even at a very low SCD risk of 1% over 5 years, only 16 or half would still want the generator change if the clinician recommended against it. Wow.

Other qualitative factors that played into the patient-decision. These included the probability data, but also prior experiences with the device or implant.

The research team also picked up a lot of misconceptions about “needing” the device for survival. Therapeutic inertia, such as just wanting to continue the same, was also common.

Comments. What got me pointed to this study was the excellent editorial from Drs Stacey Howell and Eric Stecker titled, “Just a Generator Change?” It is excellent. It reminded me of this major blind spot in our field.

  • SCD risk is decreasing in patients with HF. DANISH and the long-term follow-up of SCD-HeFT strongly argue that even a de novo primary ICD might not add much to the excellent HF therapies that we have now.

  • But a generator change is even more complicated because of the downside risk of infectious complications and increasing co-morbidities of patients 8 to 10 years after the initial implant.

  • We 100% need RCTs for ICD generator changes. Especially in primary prevention. Especially when there have been no therapies for VT. Especially when they’re in nonishcemics.

Can one be certain about equipoise? Yes. I am 99.9% sure there is enough equipoise here to conduct a trial. Of course, due to the lack of industry confluence of interest, that is, they don’t fund studies that decrease the number of units sold, this would have to be a government-funded study. Young researchers should start writing grants for this study — this weekend. I can help.

The other major message of this paper is that patients’ attitudes are not a barrier. When educated properly, many, if not most, may change their views of having surgery. I have witnessed this first-hand.

This paper clearly shows that our field of electrophysiology (EP) and really all of cardiology should intensify efforts to inform patients before a generator change. I start these discussions many years before battery depletion, so it is not such a big shock. As in, we don’t necessarily have to replace this when the battery depletes. Because. Things change.

At a minimum, all patients at battery depletion should have an outpatient office visit in regular clothes, before generator change surgery is scheduled.

VT Management

This topic melds into general cardiology and dogma. Here is another common scenario in the hospital setting. A patient with a previous history of structural heart disease presents with sustained monomorphic VT. They were fine until they had VT, but in VT they feel awful, with chest pain, breathing difficulties, syncope, etc.

What happens in your hospital? In ours, most of these patients go to coronary angiography. Why? Because of the idea that ischemia causes the VT.

While it is true that acute ischemia from plaque rupture may cause VT, that’s usually polymorphic VT or VF due to R on T stuff.

Monomorphic sustained VT is almost always from reentry within scar — often an old scar. Why old, I don’t know, but often the VT comes from very remote myocardial infarction (MI).

The EP group at Cleveland Clinic has published a nice observational study, first author, Feras Alkhalaileh, supporting what most EP docs believe; that is, in patients with monomorphic VT, just get the EP doc involved and avoid the delay and risk of coronary angiography and a percutaneous coronary intervention (PCI).

  • They looked at nearly a 100 consecutive patients with VT without acute coronary syndrome who went to VT ablation.

  • These were patients with sick LVs — mean EF 30%, two-thirds with ischemic CM and HF.

  • Nearly half (45%) of the cohort underwent coronary evaluation;  angiography 10%, nuclear scans 26%; and both 9%.

  • The yield of these tests was super low.

  • There were zero acute coronary occlusions

  • They found no associations between ischemic evaluations and acute ablation outcomes or mortality.

  • Four of the 97 total patients had revascularization with PCI. In all 4 patients, VT reoccurred after revascularization and patients ultimately underwent ablation.

  • The authors also added a second cohort of 91 patients admitted to the critical care unit for VT during the same time period. These patients were not necessarily destined for VT ablation. Many were medically treated.

  • In this cohort, 46 patients had coronary evaluations. Five of these had revascularization for chronic stable disease. Get this: four of the five had recurrent VT.

Comments. First, this is a good use of observational data. They aren’t making non-random observations. They are describing a large case series.

I think you have to combine their findings with baseline knowledge that monomorphic VT is almost always a scar-related problem, not an acute or subacute ischemic issue. If that is your baseline, then these observations only strengthen that belief.

  • They found zero acute occlusions.

  • They found no association with ischemia workups and outcomes and nearly every patient who had a revascularization had recurrent VT.

  • Why? Because they found incidental coronary disease (CAD) that wasn’t the cause.

  • I am not sure we would want to proscribe ischemia evaluation in all patients with monomorphic VT, but we should stop and think about the knee-jerk idea that VT means evaluate coronaries.

By the way, I also see angiography used in cases of clear cut normal-ventricle VT in young people. That too is a practice that should be re-thought.

The first thought one should have about monomorphic VT — after resuscitation — ought to be getting an EP doc involved right quick. I thought that before this series, but this data bolsters that view.

Critics might push back. Come on Mandrola, this is a small observational series. We surely need better data. And I also believe we could study this question in pragmatic RCT form.

Here is the protocol. Get a bunch of VT centers to participate. One group of monomorphic VT gets usual care. One group gets an immediate call to the EP team. You could look at acute outcomes and coronary angiography rates. The problem many of us EP docs have is that we don’t see the patient until after they have had angiography or PCI.

Co-morbidity in HF

I did not know of Hull in the United Kingdom until I read a nice paper from a group at University of Hull.  The study was on hospital admissions in the last year of life of patients with HF, first author  Alexandra Abel,  published in the European Heart Journal Quality and Outcomes

Their idea was to look at the frequency, causes, and patterns of hospitalizations for patients with chronic HF in the year before death.

This will be another good use of observational data because it describes patterns, here, patterns that I am afraid HF treatment enthusiasts don’t often put up on Twitter.

  • Their series included nearly 5000 patients referred to their HF clinic over 20 years.

  • They separated the group into patients with HF with reduced EF (HFrEF), HF with preserved EF (HFpEF) with lowish BNP, and HFpEF with highish BNP.

  • This is sort of a follow-back type study, from death backwards.

  • They then classified hospital admissions in the last year of life as due to HF, other cardiovascular (CV) causes or non-CV causes.

In this series, the median number of hospitalizations in the last year of life was two.

Before we get into the causes, let me remind of you of my palliative care wife Staci’s Four Horsemen of the Apocalypse – Weight loss, immobility, falls, and repeat hospital admissions. When you see any or all of these it is time to think of palliative care. Who are pals for people with serious disease.

  • In this series, 83% of patients had at least one hospitalization but only 20% had at least one HF hospitalization.

  • Let’s repeat for the HF community; 83% of patients had one or more hospitalizations but only 20% had one or more HF hospitalization.

  • 24% had one or more CV hospitalization; 70% had one or more non-CV hospitalization.

  • HF hospitalizations were most common in patients with HFrEF, but in all groups, at least two-thirds of admissions were for non-CV causes.

  • There were 788 (16%) deaths due to progressive HF, of which 74% occurred in hospital.

They have an amazing bar graph listing causes of death. The bar depicting non-CV causes dominates all types of HF, but especially patients with HFpEF.


For patients with chronic HF in the last year of life, most hospitalizations were for nonCV causes regardless of HF phenotype. Most patients had no HF hospitalizations in their last year of life. Most deaths were from causes other than progressive HF.

Comments. This is important work. The authors have a nice paragraph explaining why.

Most discussions about the care of patients with HF are focused on the treatment of the syndrome itself with emphasis placed on the need to maximize medical and device therapy for HF.
However, most hospital admissions in the last year of life are for non-CV causes, suggesting that the needs of patients with HF extend far beyond the optimization of treatment for HF.
Our findings point to a need for care that is multidisciplinary, patient-centered and coordinated between the hospital and community.

I don’t know about you, but I see more patients like these 75 to 87 year old’s than I do robust ambulatory HF patients who attend HF clinic and take all four drug classes.

  • I agree that, when possible, we should maximize all possible therapies in patients with HF who are like those in trials. But. Many and, I would argue, most patients with HF are like these described from the University of Hull.

  • For these patients, we should emphasize different things other than SGLT2 inhibitors or sacubitril/valsartan. We should emphasize ways to improve their quality of life, and one sure way is to minimize the burden of being a patient.

  • I’ve seen elders who fall and are losing weight who are told to take their blood pressure (BP) twice daily. I’ve seen elders who have two or three of the four horsemen on 80 mg of atorvastatin. I’ve seen elders like these have nuclear perfusion studies. This is a failure of Medicine 101.

It is crazy. Most of my HF patients are unlike like those in trials. We can still help them, a lot, by caring, by minimizing their work of being patient, and by maximizing the things that give them comfort and joy in their remaining time.

This is a great paper from the University of Hull. In the United Kingdom. Near the sea.

SGLT2 Inhibitors and AF

About 15 years ago, Prash Sanders and his team in Adelaide, Australia, changed the world of EP by showing the importance of managing cardio-metabolic risk factors. Their work helped bring primary care back to EP. Things like weight loss, management of BP and glucose, and sleep apnea and exercise all helped to reduce atrial fibrillation (AF), as well as increase the chance of success of AF ablation.

Because it is hard to accomplish these goals, and because it is easier and more lucrative to ablate AF, the effort plateaued a bit. Yes, we still talk about it, but ablation technology steals most of the oxygen in the halls of science.

Yet, new drugs seem to be reviving the interest in treating AF as a symptom of upstream cardiometabolic disorders. First there were SGLT2 inhibitors, which improve CV outcomes in patients with diabetes, HFrEF, and reduce renal outcomes in patients with chronic kidney disease. And now there are the new weight shredders, the GLP-1 agonists.

JACC EP has published an observational study from a group led by docs at the Lahey Clinic in Massachusetts, first author Mohammed Rasheed Abu-Qauod.

They studied the effect of SGLT2 inhibitor use on reducing AF after ablation. Notice the construction of that sentence. After talking about Bobby Yeh’s and Miguel Hernan’s papers last week, wherein they suggest not trying to hide causal inference intentions of observational research, I chose to use causal intent in the description.

The authors sort of did this too. They wrote their objective was to assess the impact of SGLT2i use on the recurrence of AF among patients with DM after catheter ablation.

My friends, if you haven’t listened to last week’s podcast, please, do. I think we should avoid shying away from causal intent in these non-random comparisons. Our job henceforth will be to assess how well the authors do it. How close it is to a trial emulation.

  • It was a simple study. They used a big database called TriNetX, a US-based multicenter federated health research network aggregating anonymized data from electronic health records (EHR) of more than 250 million patients and more than 120 US health care organizations.

  • They took patients with diabetes who had AF ablation between 2014 to 2021 and made two groups: Those on SGLT2 inhibitors and those not on SGLT2 inhibitors.

  • Their main outcome was something you can get from an EHR database—cardioversion, antiarrhythmic drugs, or repeat AF ablation.

  • The two groups were not that far off in baseline characteristics, which Bobby Yeh says is a first good sign.

  • They then propensity matched them (the robustness of the matching is hard to assess for me, at least).

The main results were:

  • SGLT2 inhibitors use in patients with type 2 diabetes undergoing AF ablation was associated with a significantly lower risk of cardioversion, new antiarrhythmic drug (AAD) therapy, and re-do AF ablation (adjusted odds ratio [OR]: 0.68; 95% CI: 0.602-0.776; P < 0.0001).

  • At 12 months, patients on SGLT2 inhibitors had a higher probability of event-free survival (HR: 0.85, 95% CI: 0.77-0.95; log-rank test χ2 = 8.7; P = 0.003).

  • All secondary outcomes were lower in the SGLT2 inhibitor group; however, the ischemic stroke did not differ between groups. One possible clue was the mortality signal. All-cause mortality was lower among patients on SGLT2 inhibitors (adjusted OR: 0.62; 95% CI: 0.41-0.93; P = 0.019)

Comments. You know where I always start: The main problem is the risk of selection bias wherein “healthier” patients get the SGLT2 inhibitors and that is why there were fewer episodes of CV, AAD, or repeat AF ablation.

The authors list it as their first limitation. Evidence for confounding is the mortality signal. These are young 65-year-old patients in whom you would not expect SGLT2 inhibitors to have an effect on mortality.

  • In a non-random observational study, whenever the OR for death is lower than odds for the primary non-fatal outcome, here repeat procedures or AAD for AF, then you worry about bias.

  • On the other hand, the two groups were pretty well matched. They had a time zero, index ablation, they did some matching, and this data comports with RCT-level data.

  • Here I cite a large meta-analysis of RCTs of SGLT2 inhibitors looking at AF from the group at McMasters. It’s published in Journal of the American Heart Association, 2021, first author, Arjun Pandey. That paper pooled 31 trials of SGLT2 inhibitors that had reported on AF and found a statistically significant 25% risk reduction in AF in the SGLT2 inhibitor arms.

My take-home here is that SGLT2 inhibitors have proven CV benefit in diabetes, CKD, and HFrEF. These types of patients often have AF. When we ablate AF in patients with these conditions it seems wise to use these drugs, as there may be an added boost in AF outcomes as well.

ESC Preview

I leave Wednesday for Amsterdam for the European Society of Cardiology (ESC) Congress. This week I will have a written preview.

I will also write a number of stories. The team from the|Medscape Cardiology will have full news coverage.

At first glance, there are going to be many interesting studies coming from ESC. If you are in Amsterdam, please say hello.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.