This editorial refers to ‘Contemporary Predictors of Death and Sustained Ventricular Tachycardia in Patients With Repaired Tetralogy of Fallot Enrolled in the INDICATOR Cohort’ by Valente et al.1

Challenge of risk stratification

Tetralogy of Fallot (ToF) is the most common cyanotic heart defect at birth. Major advances in cardiac surgery and paediatric cardiology now enable the vast majority of patients born with this condition to survive to adulthood. However, surgical repair cannot cure these patients and life expectancy continues to be compromised. Haemodynamically relevant pulmonary regurgitation (PR) is common and appears to have a major impact on long-term morbidity and mortality. It causes RV volume overload eventually associated with RV dysfunction and is—in itself—associated with life-threatening arrhythmias and sudden cardiac death. While timely pulmonary valve replacement may help to avoid some of the long-term complications associated with PR,2 it appears that merely replacing the pulmonary valve is not sufficient to avoid sudden death in this young population.3 As a consequence, numerous researchers have been struggling to identify reliable risk predictors in ToF patients over the last decades. This is not an easy undertaking: the heterogeneity of the condition combined with the relatively low annual mortality rate requires large patient cohorts and considerable follow-up periods to ensure sufficient statistical power. In addition, not all patients die suddenly and information on the cause of death is not always available. Not surprisingly, the number of risk predictors remains therefore limited (for an overview of studies, see table 1): the most prominent and widely used one is QRS prolongation on surface ECG. This risk factor is based on the seminal publication by Gatzoulis et al4 1.5 decades ago, suggesting a strong association between a QRS duration ≥180 ms and poor prognosis. While this association has been confirmed by other studies, it appears that improving surgical techniques and trends towards earlier repair may compromise the prognostic value of QRS duration or, at least, will shift the 180 ms cut-off to lower values in future. In addition, a broad QRS complex—in isolation—is insufficient to identify individual patients with annual event rates high enough to justify the primary prophylactic implantation of an implantable cardioverter-defibrillator (ICD). Subsequently, Khairy et al5 have proposed a risk score based on clinical history, QRS duration, results of ventricular stimulation and invasive assessment of LV end-diastolic pressures. Uptake of this risk score has been mainly hampered by the fact that invasive pressure measurements and electrophysiological studies are not performed routinely for risk assessment in ToF patients. More recently, the prognostic value of LV dysfunction has been highlighted. This is prima facie surprising as ToF mainly represents a right-sided heart condition. Nevertheless, ToF patients are affected by myocardial ischaemia before corrective surgery that may in part account for late LV dysfunction.6 Furthermore, myocardial fibres are shared between the ventricles and ventricular–ventricular interaction7 may also be responsible for LV dysfunction in patients with advanced RV volume overload and dysfunction. However, systolic LV dysfunction assessed by measuring EF8 lacks sensitivity in detecting early LV dysfunction and is not well suited to identify ToF patients at risk of death or those likely to benefit from ICD implantation. Recently, modern myocardial deformation parameters (such as strain on speckle tracking analysis) have been increasingly employed and preliminary studies investigating the utility of these methods for risk stratification in ToF patients are promising.9 Although further confirmation is required, these parameters seem to be more sensitive compared with EF in detecting myocardial damage and their prognostic values appear to be superior to that of conventional measures of ventricular size and function.

Role of CMR

Regarding risk stratification in ToF, until now, only small, single-centre cardiac MRI (CMR) studies, employing composite end points (including symptomatic deterioration) have been available, not providing convincing results. Despite its prominent role in the evaluation of patients before pulmonary valve replacement,10 CMR has therefore not gained a major role for risk stratification in ambulatory ToF patients. This may change in the future. Valente et al present the results of a large multicentre effort combining the CMR expertise of four large adult congenital heart disease centres in North America and Europe. Using a well-defined study protocol,11 the authors assembled CMR data on 873 ToF patients and investigated the association between CMR parameters and outcome. This is an unusually large number of ToF patients using a clinically relevant, end point of death or sustained ventricular tachycardia (VT). The authors identified LVEF and RVEF, RV mass-to-volume ratio ≥0.3 g/mL and history of atrial tachyarrhythmia as outcome predictors on multivariate analysis. Interestingly, VT was rare and confirmed sudden cardiac death accounted for only approx. 14% of deaths. In contrast, over 50% of patients died due to unknown or non-cardiac reasons. This limitation should be considered when applying the results of the current study directly for risk stratification of sudden cardiac death in ToF patients. The study by Valente et al also highlights the limitations of using EF as a marker of ventricular dysfunction. Unlike in patients with acquired heart failure, even mildly reduced EF was associated with poor outcome in this study. Therefore, it appears that EF may be suboptimal in this setting and more sensitive markers of subclinical ventricular dysfunction are desired and should be increasingly investigated.

RV pressure overload-not a benign condition?

One interesting and novel aspect of the study is the reported association between elevated RV pressure or RV hypertrophy and worse outcome. The authors report that patients who died or experienced an episode of sustained VT had on average a RV outflow tract (RVOT) gradient of 64 mm Hg and this gradient was significantly higher compared with the remaining patients investigated with echocardiography. While the detrimental impact of PR on RV mechanics and risk of sudden death12 has been highlighted, RV pressure load has received much less attention. As discussed by the authors, RV pressure and RV hypertrophy were related to each other, but provided independent prognostic information. Therefore, one could speculate that beyond the obvious effect of RV pressure overload on RV hypertrophy, alternative factors such as inadequate RV adaptation in the setting of PR may account for this finding. This interpretation is also supported by the fact that the correlation between RV pressures and mass was only modest overall.

Although current guidelines suggest that interventions for RVOT obstruction should be considered even in asymptomatic patients, the level of evidence is low (C) and the proposed RVOT pressure gradient necessitating intervention is relatively high.13 Given the results presented by Valente et al, this recommendation may need to be reconsidered. This is especially true in patients with a RVOT conduit who may be good candidates for a percutaneous valve implantation. In addition, the results of the current study raise the question whether accepting a certain degree of RVOT obstruction to avoid PR is indeed a desirable option, both at initial surgical repair and during catheter inventions later in life.

Overall, the current paper by Valente et al will not solve the problem of accurate risk stratification in ToF patients but brings us one step closer to develop a comprehensive risk score, now incorporating CMR data. In addition, the authors have to be commended, as this study also serves as a model for how multicentre efforts can help to overcome the natural limitations of efforts aiming to risk stratify adult congenital heart disease patients.