12 Comments

I disagree with Dr. Mandrola’s praise of this study, but not for the reasons he might expect.

He argues frequently and persuasively for clinical trials that challenge existing dogma. I agree wholeheartedly. I think these are most important in two circumstances.

The first circumstance is when the dogma was never based on quality evidence to begin with. Such is the case with most “medical reversals.” Examples include antiarrhythmic drugs for PVC suppression (CAST trial, 1989), or avoidance of peanuts to prevent allergies in children (LEAP study, 2015).

The second circumstance is when the science is stagnant and outdated. The best recent example is beta blocker discontinuation in patients with preserved ejection fraction post-MI (ABYSS, 2024). This practice was collecting cobwebs in our collective consciousness.

However, primary prevention ICDs do not fit neatly into either category. Although primary prevention ICD trials were conducted before the advent of modern medical therapies, they met a rigorous standard that few other treatments in contemporary medicine achieve—demonstrating improved overall survival in more than one randomized trial. Far from being stagnant, the allocation of primary prevention ICDs is a constant source of debate and deliberation among electrophysiologists. The field is ripe for innovation, not repetition.

Ejection fraction remains a crude measure, as does the ischemic/non-ischemic classification. Despite the neutral results of DANISH, critical questions persist. What should be done for patients with TTN or LMNA mutations, mitral valve prolapse with significant enhancement on MRI, cardiac amyloidosis with normal ejection fraction but frequent NSVT, or young adults with significant scarring post-myocarditis? Emerging therapies challenge traditional assumptions: for instance, do myosin inhibitors reduce sudden death risk in hypertrophic cardiomyopathy, thus diminishing the need for ICDs? In ischemic cardiomyopathy, the identification of low-risk patients through MRI, continuous monitoring, and the application of artificial intelligence to electrocardiograms are intriguing possibilities.

In fairness, some of the above classification schemes (MRI, genetics) are included as optional sub-studies in PROFID EHRA according to the website. The trial is not designed to properly test any of them. At best, PROFID will call into question a well-supported practice and need to be duplicated (again). At worst, it will take a decade to land us right back where we started.

ICD therapy in patients with ischemic cardiomyopathy and low ejection fraction should not be re-evaluated now. Not because the practice is beyond reproach – but because doing so is a perplexing and unimaginative allocation of resources. I urge more senior cardiologists to test more innovative ICD algorithms with randomization so we can really move the field forward.

Expand full comment

That's a beautiful and essential comment on the trial, sir. It's important for us to realise that when we talk about EBM, the E stands for evidence and not expertise. Which essentially translates to conducting clinically relevant trials which can even challenge generations of practice. Good one!

Expand full comment

I would quibble a bit with characterizing this study as “knowledge transfer”. I would instead suggest that it represents actual knowledge generation.

We don’t currently know if an ICD in 2025 actually adds all-cause survival benefit to med Rx in 2025 for ischemic CM. This study will hopefully provide new knowledge and allow us to update from the prior foundational trials.

And hopefully, if it demonstrates “no ICD” to be non-inferior, then the relevant guidelines will also update to reflect this new information, and allow that knowledge to be better transferred to clinician end-users.

Sadly, I have found in recent years that guideline writers are not necessarily particularly faithful to the foundational studies in their quest to “transfer” knowledge. But that is a gripe for another day.

Expand full comment

Very interesting trial and discussion. I would wonder why they would not test first in non-ischemic cardiomyopathy patients who are not CRT candidates where there would be more equipoise? When PROFID EHRA is completed, will have to interpret results carefully, as likely that more marginal ICD candidates (ie, LVEF 34%, little CHF, lots of comorbidities) are more likely to get enrolled than excellent candidates (ie, LVEF 15% with class 2-3 CHF, no other comorbidities).

Expand full comment

I wonder if they considered Danish to have answered the question for contemporary non-ischemics.

I agree, will need to see mean EF or interquartile EF range of study participants. If the “study arm” is found to be non-inferior, that would affect my decision-making for those with similar EF, but may not change my approach for those with very low EF (unless such a cohort is well represented in the study).

Expand full comment

Dr.Mandrola. I’m nervous again. It is as if you’re every word was pulled from my brain. Like you I think this study courageous and important. Participated in.SCD-HEFT and MADIT with the subsequent thought that the results were marginal. Also participated in. AVID and again the ICD benefit was relatively small. I’m concerned that the roadblocks to completing this study are formidable.

One other comment – as you know an iCD discharge does not necessarily mean a life saved(perhaps the save is implied in your comment)

John Mcanulty

Expand full comment

Could this trial have been blinded?

Expand full comment

Is there a database, country-based, in which a regression discontinuity-designed trial could be undertaken? Look at patients a smidge above and below an EF cut point and examine outcomes in who did and didnt get a device? That would be a quick and dirty way to get a signal of efficacy in the modern era.

Expand full comment

A general comment on absolute risk reduction: if 5-7% of patients with a given condition gain benefit from an intervention then 93-95% gain no benefit. I suspect that if clinicians discussed this concept with patients few would opt in to an intervention

Expand full comment

The number to treat to save one life seems a bit lower than the standard applied to other interventions. For flu vaccines, I read that NNT is over 200 to save 1 hospitalization. (Full disclosure; haven't vetted those data.) Is it really 1 in 14, or am I reading it wrong?

Expand full comment

I agree and the number electing to have a procedure or medication would decrease to almost the vanishing point if they had to pay for it. This is why the whole medicopharmatologic industry love a system where the true costs are hidden and everything is perceived as "free". Another form of deception is the frequent (and sometimes exclusive) use of relative risk reduction figures.

Expand full comment

Isn’t there a mortality benefit in non-beta specific beta blockers? Not surprised that enrollment is low, and am not convinced that the study can be effectively randomized. Physician bias WRT recruitment is a possibility. Furthermore, I could see how the overall health and life expectancy of those self-selecting for inclusion in the study could differ. Are there IRBs in the EU?

Expand full comment