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 22, 2025, John Mandrola, MD, comments on the following topics: Conduction system pacing, withdrawing HF meds when AF is corrected and patient selection for left atrial appendage occlusion (LAAO).
The CSPACE Trial
JACC has published the CSPACE trial, from a group in Melbourne, comparing conduction system (CS) pacing to right ventricular (RV) pacing for patients with a pacing indication. As I have said many times on this podcast, I believe strongly in conduction system pacing. I am in my ninth year of doing it, starting initially with his bundle pacing and now using left bundle pacing.
By placing the lead in the conduction system, either his or left bundle branch, you activate the heart naturally, without the dyssynchrony inherent in RV pacing. Why is this? Let me just say something basic here: a standard lead placed in RV muscle activates the muscle around the lead; the wave of depolarization spreads first through the RV, then to the LV. This results in a dyssynchronous contraction—and in 10%-30% of the time, can cause cardiomyopathy. Make no mistake, dyssynchronous contraction from RV leads is better than no contraction, so pacing was transformative technology.
But the idea with CS pacing is activating the his bundle or left bundle branch, allows for simultaneous activation and contraction. The problem is that CS pacing is harder than RV pacing. It requires skill in evaluating electrograms. It may cause perforation of the septum, and may require more lead revisions. His bundle pacing was marred by high pacing thresholds, and less battery longevity.
Skeptics of CS pacing stand on solid ground; they are correct to want evidence of benefit. Because beautifully narrow paced QRS complexes are not enough. EP has been late to get data on CSP because industry is not interested in funding trials, as CSP would a) reduce the need for more costly CRT devices, and b) may be able to do with two leads (A, and LB) what can be done with three leads (A, RV, CS).
The CSPACE study, therefore, is a welcome addition. 100 patients were randomly assigned to CS pacing, 100 patients to RV pacing. Investigators encouraged implanters to use RV septal lead positions over the RV apex.
The primary outcome was a composite of pacemaker induced cardiomyopathy (PICM), upgrade to biventricular CRT, heart failure hospitalization (HFH), and all-cause mortality. PICM was defined as a left ventricular ejection fraction (LVEF) reduction of > 10% resulting in an EF< 50%
Patients were excluded if they had an indication for CRT, such as an EF < 35%, and QRS width of 120 msec. They also excluded patients with end stage renal disease, dementia and limited life expectancy.
The first result was that CS pacing was successful in 88% vs 99% for RV pacing. CS pacing required longer procedures, more fluoroscopy time, and more lead revisions (8 vs 1). After 2 years of follow up, CSP was associated with lower composite endpoint (7.2 vs 20.7 events per 100 person-years; HR: 0.35; 95% CI: 0.19-0.64; P < 0.001)
This was primarily driven by lower PICM (CSP 4.6 vs RVsP 14.7 events per 100-person-years; HR: 0.31; 95% CI: 0.15-0.67, and need for CRT upgrade. (0 vs 2 events per 100 person years.
There were no differences in HFH or all-cause mortality. And when LVEF was averaged, there were no significant differences.
The authors conclude that
In patients with AV block, CSP resulted in a lower incidence of PICM and need for biventricular CRT upgrade, at the expense of higher requirement for lead revision, compared with RVsP. There was no difference in HFH or mortality. This RCT supports the broadening of the indication of CSP as an upfront pacing technique for pacemaker implantation in patients with AV block.
My comments
This is exactly what should be done with new procedures. We should randomize vs standard pacing. Hundreds of thousands of pacemakers are placed worldwide, and randomizing only a fraction of them would provide answers. Retrospective look back studies trying to adjust for baseline differences is not the way to answer the question of CS pacing vs standard pacing. Because, obviously, healthier patients will have CSP and sicker patients will get RV pacing.
At minimum, this trial shows that CS pacing is achievable though harder than standard pacing. There was more radiation required and more lead revisions. This is not nothing. Hard outcomes like HFH or death were not statistically different (though 7 vs 1 HF hospitalizations in the RV vs CSP group is a strong trend), but the reduction of pacing induced cardiomyopathy is an important endpoint, especially when doing younger patients.
Applied to a larger scale, a 70% reduction in pacing induced cardiomyopathy is a large benefit.
Yet we should not be over interpret this study: this was a very small study with small numbers of events. The main driver was a LVEF measure, which is always fraught with imprecision. As I have said, LVEF differences are a bit like judges who give harsher sentences before vs after lunch.
The follow-up too was short. People with pacemakers live decades after. And a two year window is not nothing, but longer term lead performance is important. The 8 vs 1 lead revision rate is a huge headwind for this technique, because if you apply such a ratio to a 100 hundred thousand cases, that is a lot of harm—as lead revisions have much higher rates of device infection than de-novo implants.
Another factor is that patients with an EF > 35 and less than 50% were included. I cannot tell how many, but these are patients in whom the BLOCK HF trial found benefit from CRT over standard pacing. Randomizing these patients to RV pacing alone may have created a bias in favor of CSP
The CSPACE trial is very preliminary data. It does not satisfy the medically conservative clinicians who want evidence for net benefit of this new approach to pacing.
Let’s not forget the recently published BioPace trial of CRT vs RV pacing in patients with a pacing indication but narrow QRS and normal LVEF. BioPace did not find a clinical benefit to CRT over standard pacing. Arguably, CSP is less technical and less complex than CRT, but, still, BioPace found that less was more in standard pacing.
CSPACE does have enough signal of benefit to support funding of a very large trial of hard outcomes—which will have to come from government. I see this as a huge unmet need in the evidentiary landscape of cardiac pacing.
I believe CS pacing will win out in a proper trial, and if I were to require a pacer, I would want a skilled EP to place a LB area lead, but beliefs are not evidence.
WITHDRAW AF RCT
Yet another Melbourne Australia group has published an RCT asking an important clinical question: can guideline directed medical therapy (GDMT) be stopped in patients who had HFrEF from AF after correction of AF.
The scenario comes up often: a patient who has not had HF develops AF with rapid rates. This causes shortness of air, BNP rises and pulmonary edema—or clinical HF. The echo done during rapidly conducting AF shows LV dysfunction.
Too many times, the HF docs don’t ask electrophysiologists to help, and these patients get put on GDMT, and AF treatment is delayed. Rod Tung has been trying to promote the concept of sinus rhythm therapy for HF.
The idea is if you phone a friend in EP, we can help correct the AF, and if successful, the EF may normalize, and voila, the patient no longer has HF.
Then the question becomes can we stop the GDMT (beta blockers, ACE inhibitors, mineralocorticoid receptor agonist (MRA) and SGLT2 inhibitors). The TRED HF trial found that withdrawal of GDMT in patients with nonischemic cardiomyopathy (NICM) who recovered to normal EF was harmful. This and other studies have led guidelines to rec indefinite GDMT, regardless of cause of HF.
But, clearly, AF cardiomyopathy is different. Resolution of AF, and maintenance of SR eliminates the cause of the LV dysfunction.
The Withdraw AF study had a neat design. 60 patients who had recovered EF and were maintaining SR were randomized to early withdrawal of meds (Group A) vs delayed withdrawal (Group B).
Then were crossed over with reintroduction of medical therapy in the early withdrawal group and subsequent withdrawal of the delayed withdrawal arm. The crossover design was selected to ensure: (i) similar follow-up between groups; (ii) each subject served as their own control to mitigate confounding and for secondary analyses; and (iii) to examine whether the timing of treatment withdrawal influenced the primary outcome. CMR was done at crucial periods—at initiation of the trial; after 6 months of withdrawal, then again after re-starting, In the delayed arm CMR was done at study onset, then after being on GDMT for the first 6 months, then after being off for the second 6 months.
A note on the patients. Trialists were careful to make sure these patient did not have evidence of other causes of cardiomyopathy, such as scar burden on screening CMR. And they had to have documented SR. AF burden was assessed with twice daily ECG tracings in 90% of patients. Implantable loop recorder (ILR) would have been better, but daily ECGs are not bad. 97% of the patients had SR maintained via AF ablation
Supervised withdrawal comprised a stepwise reduction in HF pharmacotherapy every 2 weeks up to a maximum of 12 weeks, adapted from the TRED-HF protocol. However, unlike the original TRED-HF medication weaning protocol, in the present study, beta-blockers were weaned first to allow earlier detection of atrial arrhythmias in this AFCM population, where beta-blockers might otherwise suppress early signs of relapse due to arrhythmia recurrence.
Medication withdrawal was extremely careful and stepwise with lots of surveillance. This will be important in external validity.
The primary endpoint was maintenance of CMR LVEF ≥50% at 6 months assessed by blinded endpoint evaluation and compared between randomized groups (Group A vs Group B)There were many secondary outcomes.
Results and Caveats
During the first 6 months 27 of 30 participants in Group A maintained LVEF ≥50% (90%), compared with 30 of 30 (100%) in Group B. Not statistically different. CMR derived LVEF were nearly identical.
At 6 months, participants crossed over to the alternate treatment arm. Following medication withdrawal in Group B during the second 6-month period, 28 of 30 participants (93%) maintained LVEF ≥50% following medication withdrawal.
A reduction in LVEF to <50% was observed in five participants following short-term medication withdrawal—three from Group A during the initial randomized phase and two from Group B during the crossover withdrawal phase, detected on CMR at 6 months post-medication withdrawal (median LVEF reduction: −7%).
The reduction in LVEF was not apparent on TTE performed 4 weeks after medication withdrawal. LVEF reduction was accompanied by an increase in LV end-diastolic volume and elevated NT-proBNP levels but occurred without clinical HF or MACE. One of these five had AF recurrence, leading to antiarrhythmic drug therapy. All 5 had normalization of LVEF with restarting GDMT.
On serial CMR imaging, four of the five individuals who relapsed developed new LV LGE. Two additional individuals without LVEF decline developed new LV late gadolinium enhancement (LGE) at 12-month follow-up (one with focal sub-epicardial/inferolateral LGE and one with fine septal linear mid-wall LGE). AF recurred in 26 of 60 participants overall (43%), with similarly low AF burden on and off HF therapy (1.4% vs 0.6%, P = .50). Rates of cardioversion (n = 6) and repeat ablation (n = 7) were similarly low between groups and across treatment periods with no associated LVEF decline.
For secondary outcomes: No differences were observed in NT-proBNP, cardiac imaging (CMR/TTE), NYHA class, symptoms and quality of life scores. At 12 months, 55 of 60 participants had maintained LVEF ≥50% off HF pharmacotherapy. A handful restarted meds for HTN.
During an additional 12-month observational follow-up, all 50 participants remained clinically stable with sustained LVEF normalization and no adverse clinical sequelae.
My Comments:
This is really nice work. It’s a good example of a small sample that provides important data. The crossover design was also informative. Coming off GDMT is a huge deal for patients. Not just in cost, but also in the burden of taking this many pills—forever. I take from this study the following…if you are extremely careful, and I mean really careful, patients with clear AF related CM can be taken off GDMT. Careful means being sure they do not have AF recurrence, that you go slowly, and that you vigorously watch LV function. Unlike in TRED HF where 44% of patients taken off HF meds relapsed, in this study it was only 8%.
Surely this reflects that fact that AF CM is a different entity. And that AF CM is reversible with sinus rhythm therapy.
But I also take from this that a small number of patients may have enough underlying CM that they see reductions in EF and should be put back on GDMT. Five patients out of 60 is not a small a number.
This is a nice study but a study we must be extremely careful translating to the clinic. These were highly highly selected patients who had documented reversal of AF, and great care was taken to exclude other causes of heart failure. Plus, there was no other reason to take meds, such as htn or diabetes. CMRs in everyone is perhaps something that is difficult to achieve in real world practice.
We must remember that many patients have persistant AF because of cardiomyopathy. Not all patients with AF and CM have AF-CM. Some have CM that then causes AF. The causal arrow is the other way. These patients should not have meds withdrawn. They were excluded from this study. And 5 of 60 patients in this study were thought to have AFCM but actually had CM exposed by medication withdrawal. We must be vigilant about this people.
That said, I do believe it is worth considering withdrawal of GDMT in these types of patients. If you have fixed the cause of HF, and there is no other pathology, you have done a great service to the patient by removing the need to take four classes of medicine. The other take home from this trial—perhaps the most important one—is that patients with AF and HF benefit from collaboration between HF and EP specialists. Rod Tung is correct; SR is indeed a beneficial HF therapy.
Patient Selection in Percutaneous LAAO
The Journal of the American Heart Association has published a retrospective review of long-term outcomes after LAAO from the NCDR registry. I mention this paper, because there is one data point that is very reliable and very important.
The authors looked at all patients getting Watchman from 2016-2019. They then matched outcomes using linking to CMS data basses. The sample was nearly 35,000. The mean age was 77 years; 42% female. Mean CHADS VASC 4.7.
Here is the datapoint to discuss:
Death occurred in 11% at one year, and 44% at 5 years. These high death rates way outpaced the stroke rates which were only 1.7% at one year and 5.7% at five years.
Comments
This observation shows how badly we are choosing patients for LAAO. The point of LAAO is to prevent stroke with less bleeding—over years and decades, not months or one year. In this study one in ten patients, those who died in a year, garnered no benefit from the device. Nearly half of patients having this procedure were dead at 5 years. This highlights the crucial concept of competing causes of bad outcomes. Death rates in this cohort are nearly 10x the rate of stroke.
And…this was from 6 years ago. Multiple analyses have shown that we continue to implant LAAO in older patients with co-morbidities. My friends: I know that many of you disagree with me. You think there are patients that benefit. I can’t find evidence for this, but let’s say you are correct, and there are a fraction of patients who might benefit. Surely you can agree that if your patient is sick enough to die within a a year or so, they cannot benefit from LAAO because it is a long-term not short term strategy. During my first ever debate on Watchman, Professor Horst Sievert said it in the Q and A: that patients with frailty, older age and co-morbid conditions, are the least likely to benefit.
This makes perfect sense, right, because both stroke and bleeding are relatively rare causes of death. Look up the population attributable risk of stroke and bleeding. Stroke causes about 1 in 10 deaths, bleeding is around 1%-3%.
If your patient has 15 life threatening medical problems, including, immobility, falls, poor nutrition, frailty, then you will almost certainly harm them with LAAO.
ESC Next Week
I leave for Madrid next week. I will go as a medical journalist and plan to cover the major studies. And there are going to be some great studies to discuss. My preview column will come out early in the week. I think there will be some great discussions.
I also have a post on my likes and dislikes of the new ACC/AHA blood pressure guidelines out on the website. Take a look.