Clinical characterisation and risk stratification of patients with arrhythmogenic right ventricular dysplasia/cardiomyopathy ≥50 years of age

The study population was recruited from the University Medical Center Utrecht ARVD/C registry [4]. We identified 29 patients enrolled in the registry who (1) presented for first clinical evaluation at ≥50 years, and (2) were diagnosed with definite ARVD/C in accordance with the 2010 Task Force Criteria (TFC) by last follow-up [5]. Ascertainment of the study population is shown in Fig. 1.

Patients were evaluated as described previously [6]. Mutation analysis of the desmosomal genes encoding plakophilin-2 (PKP2), desmoplakin (DSP), desmoglein-2 (DSG2), desmocollin-2 (DSC2), and plakoglobin (JUP), and the non-desmosomal gene phospholamban (PLN) was performed in all patients, as reported previously [6, 7]. Criteria for pathogenicity of variants were determined as done previously [4]. Twelve-lead electrocardiograms (ECGs) (recorded at rest, 10 mm/mV at paper speed 25 mm/s) were evaluated for repolarisation (precordial T‑wave inversion in V1–2 or beyond) and/or depolarisation (epsilon waves or terminal activation duration ≥55 ms) criteria for ARVD/C. All patients were off drugs that influence ventricular depolarisation or repolarisation at the time of ECG recording. Twenty-four hour Holter monitoring was used to determine the frequency of premature ventricular complexes (PVC). Exercise stress testing was performed to detect ventricular tachycardia (VT) and ischaemia. Since signal-averaged ECG was not available in this cohort, we used invasive electrophysiological study as a measure of late potentials. Electrophysiological studies were performed in 15 subjects, the majority of whom (11/15, 73 %) had prior VT, while the remainder (4/15, 27 %) were symptomatic and had frequent ectopy on Holter monitoring. Late potentials on electrophysiological study were defined as local electrogram signals after termination of the QRS complex on the surface ECG during sinus rhythm. Echocardiography, cardiac magnetic resonance (CMR) imaging, and right ventricular angiography were reviewed to determine cardiac structural abnormalities. Per study design, all study participants fulfilled diagnostic criteria for definite ARVD/C according to the revised 2010 TFC [5].

The primary outcome measure was the occurrence of a major ventricular arrhythmia, which was a composite measure of the occurrence of sudden cardiac death (SCD), resuscitated sudden cardiac arrest (SCA), or spontaneous sustained VT (≥100/min, lasting ≥30 s or shorter with electrical or pharmacological interruption), as done previously [8]. In subjects with multiple endpoints, the first event was considered to be the censoring event.

Continuous data are displayed as mean ± standard deviation or median (interquartile range [IQR]) and as proportions for categorical variables. Differences in continuous data were calculated using the independent samples t-test or Mann-Whitney U test as appropriate; for categorical data the Chi-square test or Fisher’s exact test was used. Kaplan-Meier curves were calculated for time to an event of a major ventricular arrhythmia. P-values <0.05 were considered significant. Data were analysed using SPSS software (version 22.0 for Mac).

We included 29 patients who presented at ≥50 years of age and were diagnosed with ARVD/C by last follow-up. Their phenotypical characteristics are summarised in Table 1. Overall, 16 (55 %) study participants were male, with a mean age of 59.0 ± 5.8 years at the time of presentation. As shown in Fig. 2, the majority (n = 15, 52 %) of patients presented between the age of 50 and 60 years. ARVD/C-associated pathogenic mutations were found in 21 (72 %) patients; mutations were most frequently observed in the plakophilin-2 (n = 14, 48 %) gene, while the remainder harboured a phospholamban mutation (n = 7, 24 %).

The majority (n = 23, 79 %) of patients fulfilled the diagnostic TFC at the time of first presentation; the remainder (n = 6, 21 %) were diagnosed on average 3.6 ± 2.2 years after presentation. Reasons for clinical presentation were sustained monomorphic VT in 12 (41 %) patients, family screening in 10 (34 %) patients, and syncope/presyncope or palpitations in 7 (24 %) patients. No study subjects presented with an SCD or resuscitated SCA.

As shown in Table 1, T‑wave inversions were observed in 20 (69 %) elderly patients, most commonly (n = 16, 55 %) in leads V1–3. Depolarisation abnormalities were observed in 21 (72 %) patients, of which terminal activation duration ≥55 ms was the most frequent (n = 17, 59 %). Twenty-four hour Holter monitoring showed >500 PVCs in 16/21 subjects (76 %). The majority of patients (n = 24, 83 %) had structural abnormalities on imaging studies. The median TFC score was 6 (IQR 5–8) at last follow-up.

Among the overall population, 15 (52 %) patients experienced a major ventricular arrhythmia, either at first presentation (n = 12, 80 %) or during follow-up (n = 3, 20 %). All first major ventricular arrhythmias presented with spontaneous sustained VT. Clinical characteristics of these patients are shown in Table 2. Compared with subjects without arrhythmia, patients who experienced an arrhythmic event were more likely to be male (n = 11 [73 %] vs. n = 5 [36 %]; p = 0.042) and proband (n = 13 [87 %] vs. n = 3 [21 %]; p < 0.001). There were no other differences in demographics or in any domain of the TFC between subjects with and without arrhythmia.

In elderly patients, an obvious arrhythmic substrate may occur in the setting of coronary artery disease (CAD). Electrocardiographic evidence of previous myocardial infarction was absent in all patients. In addition, we evaluated the presence or absence of CAD using coronary angiography and exercise stress testing in 15 patients with ventricular arrhythmia. Two (13 %) of these patients had evidence of CAD, defined as >50 % stenosis on coronary angiography. The first subject presented at age 77 with left bundle branch block (LBBB) inferior axis VT, and was found to have a 70 % stenosis in the circumflex artery on coronary angiography. CMR revealed right ventricular abnormalities suggestive of ARVD/C, and further evaluation revealed T‑wave inversions in V1-4 and a radical mutation in the plakophilin-2 gene (deletion exon 1–14). The second subject presented at 76 years of age with VT of LBBB superior axis morphology and a normal left ventricular ejection fraction of 54 %. However, a 70 % stenosis in margo obtusus 1 was found and subsequently treated with a bare metal stent. The LBBB superior axis VT recurred, after which another percutaneous coronary intervention was performed. After several years of recurrent VT, a comprehensive follow-up revealed ARVD/C based on late potentials, major structural CMR abnormalities, and LBBB superior axis VT. All other patients had no evidence of CAD on coronary angiography (n = 11) and/or exercise stress testing (n = 14).

Fig. 3 shows cumulative survival free from major ventricular arrhythmia in 17 patients who did not present with an arrhythmia. Over a mean follow-up of 5.4 ± 3.2 years, 3/17 (18 %) patients experienced a sustained monomorphic VT. Median cycle length of the VT was 240 ms (range 231–429). VT episodes showed LBBB morphology: two with superior, one with inferior axis. As shown in Fig. 3, median time to first VT was 4 months after first cardiological evaluation. Cumulative survival free from major ventricular arrhythmia after 6, 12 and 24 months was 88 % (95 % CI 72.3–100.0), 88 % (95 % CI 72.3–100.0), and 80 % (95 % CI 60.4–99.6), respectively.

The characteristics of the three patients who experienced an arrhythmic event during follow-up are shown in Table 3. The first event was experienced at a median age of 59 (range 55–69) years. Interestingly, all patients with arrhythmia were male and proband, and one of them was a carrier of an ARVD/C-associated pathogenic mutation (plakophilin-2). Prior to the arrhythmia, all three patients had reported symptoms including exercise-induced syncope (n = 2), palpitations (n = 2), and/or presyncope (n = 1). At the time of the arrhythmia, all patients fulfilled the TFC for ARVD/C: repolarisation criteria were fulfilled in 1 patient (major), depolarisation criteria in 2 patients (1 major, 1 minor), and arrhythmia criteria in 3 patients (2 major, 1 minor) (Table 3). In addition, cardiac imaging showed that all 3 subjects had major structural abnormalities. Distribution of age at diagnosis (p = 0.885), sex (p = 0.516) and presence of a pathogenic mutation (p = 0.242) was similar in patients with arrhythmic events at presentation and during follow-up.

Prior studies have extensively described phenotypical characteristics in ARVD/C patients [3, 7, 9]. As such, we know that depolarisation and repolarisation changes on 12-lead ECG and frequent ectopy on Holter monitoring are often observed in ARVD/C patients [3, 7, 8]. However, it is important to note that elderly patients were underrepresented in prior ARVD/C cohorts, which typically had a mean age between 20–40 years [3, 7, 9]. To the best of our knowledge, the present study is the first to specifically focus on those ARVD/C patients who only come to clinical attention after the age of 50 years. As in younger ARVD/C subjects, ECG abnormalities, frequent ventricular ectopy, and structural alterations are commonly observed in this elderly cohort.

It is important to recognise that ventricular arrhythmias in the elderly population have a broad differential diagnosis and CAD, in particular, is much more prevalent than ARVD/C. Even in the 76-year-old patient in our ARVD/C cohort, monomorphic VT was initially thought to be related to the 70 % stenosis in a coronary artery. However, monomorphic VT in CAD in the absence of a previous myocardial infarction is very rare. Although our study was not designed to compare the clinical characteristics of elderly subjects with and without ARVD/C, these results provide a cautionary note to cardiologists, who may wrongly disregard the diagnosis of ARVD/C based on older age at presentation. On the other hand, wall motion abnormalities of the right ventricle are common and cardiologists should be aware that non-pathological right ventricular wall motion disorders can easily be mistaken for a pathological regional wall motion contraction, particularly in patients with ARVD/C [10]. Future studies should address specific treatment recommendations with regard to coronary artery disease and its prevention among ARVD/C patients.

The yield of molecular genetic testing in ARVD/C typically lies between 50–63 % [7]. Our yield of 72 % is higher than in previously published reports. This may be due to a founder effect of the phospholamban p.Arg14del mutation [11], which was observed in a remarkably large subset of our population (24 %). Prior reports have shown that ARVD/C patients who harbour a phospholamban mutation present at an older age, frequently with left ventricular involvement [12, 13]. In this regard, it is interesting that only 62 % of our population had T‑wave inversion in the right precordial leads, in contrast to >85 % in the ARVD/C literature [3, 7, 8]. This may reflect a left predominant phenotype dominated by phospholamban mutations in our population.

Prevention of SCD is the most important therapeutic goal in ARVD/C [14]. Therefore, it is an important finding of our study that none of the elderly patients in our cohort experienced SCD/SCA either at presentation or during follow-up. This is in contrast to prior studies with young ARVD/C patients which reported a 3–22 % rate of SCD [3, 4, 7]. This finding is particularly interesting in the context of recent basic science studies showing that desmosomal mutations lead to reduced sodium current which promotes re-entry susceptibility at an early disease stage [15‐17]. In contrast, heterogeneous fibrofatty replacement of the myocardium leads to monomorphic sustained VT that is haemodynamically stable and can be adequately treated. In our study population, 82 % already had structural abnormalities on cardiac imaging. As such, we believe that our study subjects were more likely to experience haemodynamically well-tolerated monomorphic VT, rather than ventricular fibrillation. Future studies are necessary to determine whether a true difference in arrhythmic substrate exists in younger vs. elderly ARVD/C subjects.

Despite a low risk of SCD, malignant ventricular arrhythmias are not uncommon among elderly ARVD/C patients. Our study shows that more than half of ARVD/C patients who present beyond 50 years of age experience a ventricular arrhythmia during more than five years of follow-up. Almost all of these events occurred at presentation. As such, the risk of ventricular arrhythmias seems highest for as yet unrecognised cases. Ventricular arrhythmias were significantly associated with male gender and proband status, and arrhythmic events during follow-up occurred exclusively among male probands. These data suggest that further screening and invasive risk stratification in elderly ARVD/C subjects may be most beneficial among male probands.

This study was limited by its small sample size. However, this is consistent with a low prevalence of ARVD/C, possibly due to under-recognition. Since this is a registry-based study, not all subjects underwent all clinically available tests such as Holter monitoring. It is possible that Holter monitoring was mainly employed in those with mild disease, which may explain a relatively modest PVC count in this cohort. Data on exercise participation were not available. Given recent studies confirming the role of strenuous exercise in phenotypic development of ARVD/C [18, 19], it would be interesting to further investigate exercise participation in elderly ARVD/C subjects. Detailed analysis of the relationship between structural phenotype and age were beyond the scope of this study. Since a large proportion of elderly patients presented with ventricular arrhythmias, only cross-sectional analyses could be performed to determine predictors of arrhythmic events. While we are aware of the broad differential diagnosis of arrhythmias in an elderly population (particularly ischaemic heart disease), we were unable to obtain coronary angiography in all patients. Of note, all patients who did not undergo coronary angiography had no electrocardiographic (exercise stress testing or resting ECG) and CMR evidence of previous myocardial infarction.