Prediction of ventricular arrhythmia and sudden death in arrhythmogenic right ventricular cardiomyopathy

This editorial refers to ‘A new prediction model for ventricular arrhythmias in arrhythmogenic right ventricular cardiomyopathy’†, by J. Cadrin-Tourigny et al., on page 1850. The recognition of arrhythmogenic right ventricular cardiomyopathy (ARVC), a cardiomyopathy with arrhythmia as the cardinal manifestation, dates back almost 40 years.1 , 2 ARVC’s distinguishing clinical feature of ventricular arrhythmia, predominantly from the right ventricle, contrasts with the more common finding in adults of left ventricular arrhythmias dominated by coronary artery, valvular, and hypertensive heart disease. The fact that ARVC often goes unrecognized during life and even at post-mortem has contributed to the historical focus by international experts on diagnostic criteria.3 , 4 Analogous international efforts to define arrhythmia management and sudden death prevention have been more recent.5 , 6 Arrhythmia treatment and sudden cardiac death (SCD) prevention strategies have evolved over time, with early reliance on electrophysiology (EP)-guided pharmacological management and ablation being superseded by use of the implantable cardioverter defibrillator (ICD).7–11 Although highly effective for prevention of SCD, unwanted device complications of the ICD have mandated efforts to define SCD risk accurately. This effort is particularly poignant given that ARVC patients are young (age at diagnosis 38 years in the Cadrin-Tourigny study), with potential lifelong issues related to the ICD.12 The identification of the high-risk patient and decision-making regarding ICD continues to be challenging, contributed to by the absence of an established risk algorithm and variability in both physician perception and patient acceptance of risk. The risk model presented by the international collaboration of 18 centres from Europe and North America in this issue of the European Heart Journal is potentially a major advance in defining risk of ventricular arrhythmia in individual patients.12 Relative youth, male gender, ECG and imaging features reflecting more extensive right ventricular disease, and the severity of ventricular arrhythmia were the most accurate identifiers of the high-risk cohort in this study. This profile is consistent with previous attempts to define the high-risk patient.5 , 6 The model also provides 1–5 year ventricular arrhythmia event-free survival rates for the predicted high-risk group and the potential to determine 5-year risk in the individual patient. An analogous scenario is the experience of defining arrhythmic risk in patients with hypertrophic cardiomyopathy (HCM), where over several decades appreciation of the combined power of individual risk factors (e.g. syncope, non-sustained ventricular tachycardia, and family history of sudden death) eventually led to the creation of a 5-year individual patient sudden death risk model.13 The HCM online risk calculator (http://doc2do.com/hcm/webHCM.html) which is populated with readily obtained clinical data provides the basis for informed discussion between patient and physician. Accordingly, the data set acquired by this ARVC collaboration is the beginning of individual patient absolute risk assessment. As the authors note in the Discussion, their intention is to create an online calculator to establish absolute individual 5-year risk.12 The current publication of this risk score, and its eventual accessibility via a risk calculator will be a laudable achievement for the consortium that brought together collective experience of proband and family member treatment from multiple international centres. Importantly, the authors emphasize that the model does not aim to ‘prescribe’ treatment for a given patient, but rather to provide the data for ‘well informed shared clinical decision making’. This emerging shared-care concept is only achievable when based on clinically achievable data-based risk estimation to inform discussion between provider and patient. Thus, the published model and the anticipated online risk calculator represent critical steps forward in ARVC disease management. Can the model to identify ARVC patients without prior sustained ventricular arrhythmia who are at risk of sudden death be improved? Inclusion of inducibility of ventricular arrhythmia on programmed ventricular stimulation might be considered. Data from various registries do not, however, provide definitive support for this, suggesting rather that syncope, the burden of non-sustained ventricular arrhythmia, and left ventricular dysfunction are more predictive of appropriate ICD interventions than inducibility during programmed ventricular stimulation (Figure 1).9 , 10 , 12 Further, the impact on survival of ablation procedures beyond their immediate success, particularly with endo/epicardial approaches, remains uncertain.14 Is successful ablation a ‘short-term gain for long-term pain’? Arrhythmia risk model variables. Those included in the model are depicted in blue. Additional variables that have been associated with disease arrhythmia severity/progression and which might add to the model are depicted in orange. Another unresolved issue is the importance, and challenge, of incorporating left ventricular involvement as a risk predictor in a disease that is traditionally conceived of as specific to the right ventricle. Previous studies have reflected left ventricular involvement by progression of T wave changes across anterior chest leads, and by measures of left ventricular function, usually ejection fraction. Though not included as a diagnostic feature, left ventricular late gadolinium enhancement on cardiac magnetic resonance imaging is likely to be a more sensitive marker of electrical vulnerability and arrhythmia risk than ejection fraction, and efforts to standardize and include the evaluation of late gadolinium enhancement across large ARVC data sets should be considered. Ultimately including left ventricular assessment in risk projections may well improve model fit. This is particularly relevant given recognition of phenotypes of ARVC that predominantly affect the left ventricle or that have biventricular expression. That the model predicts ventricular arrhythmia in ARVC is a step towards prediction of sudden death. However, the distinction between fatal and non-fatal ventricular arrhythmias is problematic. This confounds interpretation of the data, particularly as ICD discharges were the most common endpoint in this study.12 To what extent is the conventional interpretation in outcome studies that an appropriate ICD discharge is potentially a life-threatening event correct in the context of ARVC, and to what extent has this influenced the clinical features incorporated into the model? The number of sudden death/cardiac arrest endpoints is often a limitation of outcome studies. During a median of 4.8 years, 9 of 528 patients (1.7%) experienced an unequivocal life-threatening event (sudden death/cardiac arrest). An additional 102 patients (19%), however, had an appropriate ICD discharge, of which 44 had rapid sustained monomorphic ventricular tachycardia with cycle length <240 ms (8% of the total population). The inclusion of an appropriate ICD discharge enabled an endpoint in 27.7% of the 528 patients, ∼6% per year. Were the majority of these events legitimate sudden death surrogates? Their inclusion is an accepted and often necessary compromise; however, the data suggest that most appropriate ICD discharges in ARVC if untreated would not have been fatal events. In HCM patients who have not had documented sustained ventricular arrhythmia, appropriate ICD discharge rates from multiple registries are in the range of 2–4% per year, while sudden death rates are in the range of 0.5–1% per year, with a ratio of ICD discharge to death of ∼4:1.15 , 16 These findings are consistent with the interpretation that a significant minority but not all appropriate discharges terminated what would have been a fatal arrhythmia. In ARVC, robust data indicate an appropriate ICD discharge rate of ∼10–15% per year in patients with, and ∼5% in patients without documented sustained ventricular arrhythmia; mortality data are less clear, but mortality is estimated to be <1% per year. These data suggest that the majority of appropriate ICD discharges would not have been fatal events. This is a difficult to resolve limitation of this model and any outcome study that relies on ICD discharges as a major outcome endpoint. It is of interest that a study in the early ICD era (1990s) of 81 patients with ARVC who presented with ventricular arrhythmias from an expert German centre which involved EP-guided pharmacological treatment implanted only one ICD and had only one sudden death during 3 year follow-up.7 Despite these considerations, the development of a risk model with absolute individualized predictive data represents a major advance in managing patients and asymptomatic family members in relation to arrhythmic risk. Early vascular risk profiles started with 5- to 10-year event data; the bigger question is lifelong risk and, similar to the HCM risk calculator, this is a valuable start to a longer journey. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.