Should We Do 'Whatever It Takes' or 'Whatever Is Best' to Prevent Cardiac Arrest in High-risk Patients?

Sami Viskin; Raphael Rosso


Eur Heart J. 2019;40(35):2962-2963. 

Cardiac arrest survivors—and the physicians treating them—are likely to embrace the 'second chance' offered to them and do 'whatever it takes' to prevent a recurrent event. When the cardiac arrest is due to a 'treatable cause' [such as the ischaemic ventricular fibrillation (VF) taking place during acute myocardial infarction], treatment may safely be limited to treating the underlying cause (urgent revascularization in the example given). Some concerns, however, will remain, particularly during the early months, due to, amongst other reasons, the non-trivial nature of defining what is a 'reversible cause'[1] of cardiac arrest or defining 'ischaemic VF'.[2] For all other cases of non-fatal cardiac arrest, the implantable cardioverter-defibrillator (ICD) will rightfully be the mainstay of therapy, together with disease-specific prophylactic medications.[3] Arrhythmogenic channelopathies [such as the long QT syndrome, Brugada syndrome, or catecholaminergic polymorphic ventricular tachycardia (CPVT)], being inborn diseases, fall into the last category. Thus, ICD implantation (or implantation of 'the box', as is often referred to) is a fairly 'nominal' procedure after cardiac arrest. Thinking out of the box, however, a few investigators have pioneered the concept that well-defined categories of cardiac arrest survivors may be treated without ICD implantation, arguing that specific drug therapy is so effective that the added value of ICD implantation will be too limited to justify its inherent risks: a 'whatever is best' instead of a 'whatever it takes' approach. Indeed, in 2009, Michael G. Vincent presented retrospective data showing that the effectiveness of beta-blockers in long QT syndrome type 1 (LQT1) is so remarkable, that even cardiac arrest survivors with LQT1 could safely be treated with beta-blockers alone, without ICD implantation, if their original event occurred in the absence of therapy.[4] A similar concept for cardiac arrest survivors with idiopathic VF[5,6] or Brugada syndrome,[7,8] using quinidine as the sole therapy, was pioneered by Bernard Belhassen. Guideline documents eventually endorsed these observations, recommending medical therapy (beta-blockers for LQT1 and quinidine for Brugada syndrome) in circumstances where ICD implantation is problematic (such as in children).[9]

Now, in this issue of the European Heart Journal, van der Werf et al. present, for the International CPVT Registry (led by Ackerman and Wilde), the largest set of follow-up data on CPVT patients who survived a cardiac arrest event.[10] The obvious 'whatever it takes' approach for secondary prevention of cardiac arrest dictates protective ICD implantation in addition to preventive drug therapy. However, unique characteristics of CPVT may call for a 'whatever is best' policy instead: not only is drug therapy (with flecainide and beta-blockers and/or left cardiac sympathetic dendervaion) highly effective for preventing CPVT-related arrhythmias,[11,12] but the ICD is also particularly problematic and potentially dangerous in this disease. This is because the VT of CPVT is a triggered arrhythmia. In contrast to the scar-related VT of patients with organic heart disease, which is a classic reentry arrhythmia, triggered arrhythmias do not respond to anti-tachycardia pacing or shock therapy. If anything, the shocks delivered by the ICD are likely to be pro-arrhythmic by causing pain with a consequent adrenergic surge, a catastrophic consequence for adrenergic-mediated arrhythmias. For ICD shocks to be effective in CPVT, sufficiently long detection times ought to be programmed so that shocks are delivered only after the polymorphic VT degenerated to VF, a true reentry arrhythmia. On the other hand, excessive delay prior to shock delivery may render ICD shocks ineffective, as delaying shock delivery to as little as 10 s after VF initiation leads to a 50% rise in defibrillation threshold even with modern biphasic shocks.[13] Again, 'thinking out of the box', this study reasons that ICD implantation is not warranted for well-characterized cardiac arrest survivors.

This retrospective study describes 136 CPVT patients who experienced a cardiac arrest event in the absence of preventive drug therapy (generally before they were correctly diagnosed), survived the event, and then received appropriate secondary prevention therapy. Of these cardiac arrest survivors, 79 (58%) patients underwent ICD implantation and drug therapy, whereas 57 (42%) patients received only drug therapy.[10] Since this is not a randomized study, it is important to question whether patients treated with and without the ICD were comparable. Indeed, patients treated without an ICD were more likely to be European, were younger (mean age of 11 vs. 16 years), and were more likely to be prescribed nadolol/propranolol (as opposed to metoprolol/bisoprolol) than the ICD group. The first characteristic is reflective of physicians'—rather than patients'—characteristics (namely, that European physicians had more confidence in the drugs they prescribed). The second characteristic actually places the non-ICD group at a disadvantage because being of a younger age at the time of diagnoses correlates with a higher risk of arrhythmic events.[14] On the other hand, nadolol, used more often in the non-ICD group, appears to be the most effective beta-blocker (albeit based on retrospective studies).[14] Importantly, cardiac and non-cardiac comorbidities, often important confounders in non-randomized studies of cardiac arrest survivors with heart disease, are a non-issue here because the CPVT patients were young and otherwise healthy. Finally, the fact that both groups consisted exclusively of cardiac arrest survivors frankly overshadows the group differences mentioned, by categorizing them all as very high-risk patients.[14] This is why the results of the study are compelling: ICD implantation was not associated with a mortality benefit or with a reduction in sudden cardiac death among CPVT patients who had survived a cardiac arrest prior to the onset of appropriate medical therapy and were subsequently medicated properly. In fact, during a mean follow-up period of 4 years (with ≥10 years of follow-up for nearly half of all patients), only three patients died, and all three had an implanted ICD. Post-mortem interrogation was available for two of these fatalities and, in both of them, ICD shocks appeared to have played a pro-arrhythmic role in the fatal event. The composite endpoint of sudden death, non-fatal cardiac arrest, or appropriate ICD shock also occurred more frequently in patients with implanted ICD (47% vs. 16%), clearly demonstrating that in CPVT, ICD shocks, even when considered appropriately delivered for ventricular arrhythmias, cannot be equated with 'lives saved.'

Before one rushes to embrace the 'no-ICD therapy' for CPVT cardiac arrest survivors, one should note two caveats: (i) the upper limit of the 95% confidence interval, for a study showing zero mortality risk among patients treated without ICD, reaches 5% because there were only 57 patients in that group;[15] and (ii) nine (15%) of patients treated only with medications experienced a second non-fatal cardiac arrest. Practically all these patients were treated with monotherapy (beta-blockers without flecainide), received inappropriate low doses, and/or were non-compliant with the drugs prescribed. Fortunately, none of them developed anoxic brain damage. Whether this outcome simply represents good luck or reflects the fact that some degree of cardiac output continues during clinical cardiac arrest caused by polymorphic VT is an unanswered question. Nevertheless, physicians and patients willing to consider 'conservative' therapy based exclusively on medications after a cardiac arrest should be committed to continue combined therapy with beta-blockers and flecainide at the maximal tolerated dosages and should confirm that such doses effectively prevent VT during periodically repeated maximal exercise tests. Defining tolerability of a drug is obviously a function of the drug's indication: patients may need to come to terms with some degree of fatigue from beta-blocker therapy (or undergo left cardiac sympathetic denervation) when the drug is prescribed for the prevention of arrhythmic death. The same caveat holds true, however, for patients undergoing ICD implantation. Patients and physicians should realize that compromising on maximal drug therapy for CPVT because of the perceived safety provided by the ICD is a terrible compromise. Where defensive medicine is the rule, physicians may still feel compelled to implant an ICD in patients similar to those described here, simply because sudden deaths occurring despite an implanted defibrillator being perceived as 'inevitable', whereas similar events, in appropriately treated patients receiving 'only medication' are considered 'preventable'. The present study is a good reminder that more is not always better.