Clinical and Genetic Characterization of Families With Arrhythmogenic Right Ventricular Dysplasia/Cardiomyopathy Provides Novel Insights Into Patterns of Disease Expression
Background— According to clinical-pathological correlation studies, the natural history of arrhythmogenic right ventricular dysplasia/cardiomyopathy is purported to progress from localized to global right ventricular dysfunction, followed by left ventricular (LV) involvement and biventricular pump failure. The inevitable focus on sudden death victims and transplant recipients may, however, have created a skewed perspective of a genetic disease. We hypothesized that unbiased representation of the spectrum of disease expression in arrhythmogenic right ventricular dysplasia/cardiomyopathy would require in vivo assessment of families in a genetically heterogeneous population.
Methods and Results— A cohort of 200 probands and relatives satisfying task force or modified diagnostic criteria for arrhythmogenic right ventricular dysplasia/cardiomyopathy underwent comprehensive clinical evaluation. Desmosomal mutations were identified in 39 individuals from 20 different families. Indices of structural severity correlated with advancing age and were increased in long-term endurance athletes. Fulfillment of modified criteria indicated phenotypically mild disease, whereas asymptomatic status did not. In >80%, ECG, rhythm monitoring, and/or gadolinium-enhanced cardiovascular magnetic resonance were suggestive of LV involvement, the extent of which often was marked among individuals with chain-termination mutations and/or desmoplakin disease. Three patterns of disease expression were identified: (1) classic, with isolated right ventricular disease or LV involvement in association with significant right ventricular impairment; (2) left dominant, with early and prominent LV manifestations and relatively mild right-sided disease; and (3) biventricular, characterized by parallel involvement of both ventricles.
Conclusions— LV involvement in arrhythmogenic right ventricular dysplasia/cardiomyopathy may precede the onset of significant right ventricular dysfunction. Recognition of disease variants with early and/or predominant LV involvement supports adoption of the broader term arrhythmogenic cardiomyopathy.
Received August 25, 2006; accepted January 12, 2007.
The natural history of inherited disorders is best defined by serial, lifelong follow-up of large, unselected cohorts comprising both probands and relatives. Attaining this optimum is difficult in newly described and underrecognized diseases. Although accounts of sudden cardiac death in young people are as old as recorded history, modern awareness of the genetic myocardial diseases that underlie most of these cases dates back a brief half-century. The first contemporary description of arrhythmogenic right ventricular (RV) dysplasia/cardiomyopathy (ARVD/C) appeared in 1961, but its name was not coined until 1978; a preliminary outline of its clinical profile appeared 3 years later.1,2
Current understanding of the disease evolution in ARVD/C stems from 2 sources: clinical–pathological correlation and medium-term clinical follow-up studies.2–6 The former has the incontrovertible benefit of allowing exploration of the disease process at a cellular level through histological assessment but is disadvantaged by its obligatory focus on ex vivo hearts from probands with lethal disease forms.3,4 Early clinical follow-up studies were based primarily on probands presenting with symptomatic arrhythmia of RV origin, with the strong familial preponderance of the disease becoming apparent in 1988.2,5–7
Clinical Perspective p 1720
In the past 6 years, the identification of causative mutations in plakoglobin, desmoplakin, plakophilin-2, desmoglein-2, and desmocollin-2 has fostered the view that ARVD/C is a disorder of the desmosome and provided insight into its pathogenesis.8–11 Early studies of genotype–phenotype associations have challenged prevailing conceptions of the natural history of ARVD/C, although most have hitherto been dedicated to specific disease-causing genes.12–16 We hypothesized that unbiased representation of the spectrum of disease expression in ARVD/C would require in vivo assessment of families in a genetically heterogeneous population.
The cohort comprised 200 individuals attending the ARVD/C clinics who fulfilled the following criteria: (1) age ≥12 years; (2) referral with arrhythmia of RV origin, proven ARVD/C in the family, and/or features suggestive of ARVD/C on prior investigations; and (3) clinical diagnosis of ARVD/C in accordance with either task force (TF) guidelines1 or modified diagnostic criteria for familial ARVD/C17 on the basis of standard noninvasive evaluation. The workup included history, pedigree compilation, 12-lead ECG, signal-averaged ECG with 40-Hz filter, 2-dimensional echocardiography, exercise testing, ambulatory ECG monitoring, and cardiovascular magnetic resonance (CMR) as previously described.18
Mutation screening of 5 desmosomal genes (plakoglobin, desmoplakin, plakophilin-2, desmoglein-2, and desmocollin-2) was performed by direct sequencing.11,13 Informed consent was obtained before CMR assessment and genetic testing.
Notable ventricular arrhythmia was arbitrarily defined as any of the following: >200 premature ventricular contractions (PVCs) per 24 hours on ambulatory ECG monitoring, ≥4 PVCs on exercise testing, nonsustained ventricular tachycardia (VT) (≥3 consecutive beats at >120 bpm), or sustained VT (>30 seconds’ duration). PVCs of right bundle-branch block (RBBB) and left bundle-branch block (LBBB) configuration were presumed of left ventricular (LV) and RV origin, respectively. Biventricular arrhythmia referred to the presence of PVCs of both RBBB and LBBB morphology.
CMR-derived end-diastolic volumes (EDVs) were expressed as a percentage of predicted throughout to adjust for normal variations related to age, sex, and body surface area.
The extent of myocardial fatty replacement and/or late enhancement (LE) in the LV (LVlesion) was visually graded as an ordinal score: 0=nil; 1=minor, affecting ≤2 segments; 2=moderate, affecting 3 segments; 3=prominent, affecting 4 segments; 4=extensive, affecting ≥5 segments.18
The extent of fatty replacement and/or LE in the RV was visually graded by intensity (0 to 2) at each site affected (RV outflow tract, subtricuspid region, free wall, and apex) to give a total score (RVlesion) ranging from 0 to 8.18
LV involvement was considered present in subjects with ≥1 of the following on CMR: LVEDV above the upper limit of normal, LV ejection fraction (EF) <55%, LV intramyocardial fat, LVLE in a nonischemic pattern, and/or LV wall motion abnormalities (WMAs) excluding isolated septal hypokinesia.
Statistical calculations were performed using Web-based computation facilities at http://departments.vassar.edu/lowry/VassarStats.html. The relationship between continuous and dichotomous variables was assessed through the use of point-biserial correlation coefficient (rpb). The Fisher exact probability test was used to analyze dichotomous frequency data. Continuous variables were compared between 2 subgroups with the Student unpaired t test and between multiple subgroups with 1-way ANOVA. Ordinal variables were compared between 2 subgroups with the Mann-Whitney test and between multiple subgroups with the Kruskal–Wallis test. The relationship between 2 continuous variables was assessed with Pearson’s correlation coefficient; Spearman’s rank correlation was used when either or both variables were ordinal.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Demographics, ECG, and Arrhythmic Findings
Of 200 patients with a clinical diagnosis of ARVD/C, 116 (58%) fulfilled TF guidelines, whereas 84 (42%) were relatives satisfying modified criteria for familial ARVD/C. The cohort included 43 families with ≥2 affected individuals (total, n=145); 25 of these families had ≥3 members fulfilling TF or modified criteria. Symptoms or signs of right-sided heart failure were absent. Although 9 (4.5%) reported episodic exertional dyspnea, none had physical signs of LV failure. Baseline characteristics are summarized in Table 1.
Pedigree and Mutation Analysis
Pedigrees were consistent with autosomal dominant transmission in the 43 families with ≥2 affected individuals, although other modes of inheritance could not be definitively excluded.
In the 42 months after the commencement of the study in January 2003, mutations in desmosomal genes (desmoplakin, plakophilin-2, desmoglein-2, and desmocollin-2) were identified in 39 individuals from 20 different families (Figure I in the online Data Supplement). No mutations were isolated in plakoglobin. A further 49 families, making up a total of 117 individuals, had no identifiable mutations in the 5 desmosomal genes screened, resulting in an overall pickup rate of ∼30%.
Of the 39 genotyped individuals, 38 (97%) had proven familial disease. The exception was a 36-year-old man of Pakistani origin with a homozygous missense mutation in desmoglein-2 (3175T→A) considered pathogenic because it alters a conserved amino acid (S1059T). His relatives were not available for evaluation, however, and a racial polymorphism cannot be definitively excluded at present.
In 1 family, the proband had suffered sudden cardiac death at 16 years of age with postmortem diagnosis of ARVD/C. Both his parents fulfilled TF criteria for ARVD/C. Subsequent genetic analysis revealed a nonsense mutation in exon 5 of plakophilin-2 in the mother, whereas the father carried a missense mutation in exon 15 of desmoglein-2. The proband was therefore presumed a double heterozygote, although DNA extracted from retained blocks was of insufficient quality to provide molecular confirmation.
CMR volume analysis is presented graphically in Figure II of the online Data Supplement. Mitral valve prolapse was identified in 8 (4%); a further 23 had mild bowing of the anterior leaflet (11.5% prevalence). RV aneurysms were present in 70 (35%). RV fatty replacement was discernible in 141 (70.5%).
RVWMA and tissue characterization findings are outlined in online Data Supplement Table I. RVLE was present in 137 (68.5%), commonly affecting the free wall, subtricuspid region, and outflow tract. The RVlesion score showed significant positive correlation with RVEDV (Spearman r=0.45, P<0.0001) and weak negative correlation with RVEF (Spearman r=−0.28, P<0.0001).
Among 200 affected patients, 168 (84%) had CMR evidence of LV involvement. Within this subgroup, the relevant LV findings were LE in 155 (92%), myocardial fat in 72 (43%), LV dilation in 68 (40%), and reduced LVEF in 31 (18%). Sixty-eight patients had LVWMA, of which septal hypokinesia was the most common manifestation (n=42) but not considered evidence of LV involvement per se. LV free WMAs were present in 55 (33%), with the apex the most frequently affected site (n=13); 11 of these patients (7%) had apical aneurysms.
Among 155 patients with LVLE, favored locations were as follows: inferolateral wall, 131 (85%) (Figure 1A through 1C); inferior wall-septal junction, 85 (55%) (Figure 1B); inferior wall, 45 (29%); septum, 30 (19%); anterolateral wall, 28 (18%); and anterior wall, 11 (7%). Dominant patterns of LE were subepicardial in 71 (46%), midmyocardial in 27 (18%), and both in 57 (36%).
One 49-year-old man had subendocardial LE anteriorly, whereas the septum, as well as inferior and inferolateral walls, showed the typical subepicardial and midmyocardial distributions. He had originally presented with sustained VT of LBBB morphology accompanied by profound hypotension, requiring emergent direct current cardioversion. Multiorgan failure resolved during extended intensive care. Angiography revealed unobstructed coronary arteries, suggesting that prolonged hemodynamic compromise may have resulted in anterior infarction.
LVLE frequently was observed without concomitant WMA in the same segment. LE was detected at the inferior wall or inferior wall-septal junction in 115 patients, but only 6 (5%) had inferior WMA; of 131 patients with inferolateral LE, only 4 (3%) had inferolateral WMA. Conversely, localized hypokinesia or dyskinesia in the LV free wall generally coincided with discernible LE and/or fat: of 8 patients with inferior WMA, 7 received gadolinium-DTPA and had inferior LE. Similarly, the 4 patients with inferolateral WMAs had evidence of coincident LE, as did the 3 patients with anterolateral hypokinesia. The main exception to this observation was asynchronous septal motion, which could be related to fibrofatty substrate in only 8 of 42 cases (19%).
Global LV size (LVEDV 112±19% of predicted; range, 72% to 183%) and function (LVEF 61±7%; range, 30% to 80%) were relatively well preserved in the cohort. The LVlesion score showed weak positive correlation with LVEDV (Spearman r=0.22, P=0.0032) and weaker negative correlation with LVEF (Spearman r=−0.19, P=0.0099) but did not correlate significantly with LV mass (Spearman r=0.11, P=NS).
Age and Sex
The following parameters showed positive correlation with age: RVEDV (Pearson r=0.32, P<0.0001), RVlesion (Spearman r=0.30, P<0.0001), LVlesion (Spearman r=0.27, P=0.0003), and more weakly, LVEDV (Pearson r=0.18, P=0.01). Notable ventricular arrhythmia became more common with advancing age (rpb=0.33, P<0.0001), as did nonsustained VT (rpb=0.28, P<0.0001).
Compared with women, men had higher RVEDV (mean, 119±24% versus 131±30%; P=0.004), lower RVEF (mean, 55±6% versus 50±8%, P=<0.0001), and lower LVEF (mean, 62±7% versus 60±8%; P=0.0249), although the latter attained only marginal significance. Sex differences in LVEDV, RVlesion, and LVlesion did not reach statistical significance. Neither was there any significant difference in the frequency of notable ventricular arrhythmia, nonsustained VT, or LV involvement between the sexes.
Symptomatic Versus Asymptomatic
The 116 patients fulfilling TF criteria were subdivided into those with (n=79) and without (n=37) cardiac symptoms (online Data Supplement Table II). The asymptomatic subgroup comprised 35 relatives undergoing screening and 2 probands with an incidental finding of frequent PVCs during routine health examinations. The main differences between the 2 subgroups were in the ratio of male to female (3:1 in asymptomatic individuals versus 0.65:1 among symptomatic subjects) and the prevalence of notable ventricular arrhythmia and nonsustained VT, which were more common in the symptomatic group (P<0.01). Aneurysms were more frequent among symptomatic individuals, but the difference reached only marginal significance. No significant differences existed in volumetric indices, RVlesion, or LVlesion.
TF Versus Modified Criteria
Key points of distinction between individuals satisfying TF guidelines and relatives fulfilling modified criteria only are outlined in Table 2.
Impact of Endurance Training
A history of participation in athletic activity was sought to identify subjects who had engaged in long-term endurance training (equivalent to >1 marathon a year for >10 years). The cohort included 11 such individuals, although detailed characterization of endurance levels was outside the scope of the clinical workup. Prior to comparative analysis, we excluded significant differences in age between athletes and the rest of the cohort. Compared with other subjects, athletes had increased RVEDV (mean, 174±30% versus 122±25%; P<0.0001), lower RVEF (mean, 44±7% versus 53±7%; P=0.0001), increased LVEDV (mean, 134±27% versus 110±18%; P<0.0001), and higher RVlesion scores (P=0.0004). No significant differences existed in LVlesion scores, the prevalence of notable ventricular arrhythmia, or nonsustained VT.
Depolarization Abnormalities on 12-Lead ECG
The presence of localized QRS prolongation in V1 through V3 (n=36) was associated with significantly higher RVEDV (mean, 145±35% versus 120±24%; P<0.0001), RV-to-LV volume ratio (P=0.0003), and RVlesion scores (P=0.0003) and lower RVEF (P=0.0010).
Patients with late potentials (n=56) showed significantly increased RVEDV (mean, 138±32% versus 116±20%; P<0.0001), reduced RVEF (mean, 49±7% versus 54±6%; P<0.0001), and higher RVlesion scores (P<0.0001) and were more likely to have RV free wall LE (P<0.001) than those with negative signal-averaged ECGs. Epsilon waves in the lateral leads were recognized in 2 patients, both of whom had LVLE at the inferolateral wall.
The 16 patients with T-wave inversion inferiorly had evidence of LVLE in the inferior wall (P=0.0002). Altered repolarization in ≥1 LV leads (V4 through V6, I, aVL) was observed in 37 (18.5%); 34 of these patients received gadolinium-DTPA, revealing LE with or without myocardial fat in the inferolateral, anterolateral, and/or apical regions of the LV. Lateral T-wave inversion/attenuation was associated with significantly higher LVlesion scores (P<0.0001).
Individuals with ventricular arrhythmia of LBBB morphology (n=65) had significantly increased RVEDV (mean, 136±24% versus 119±23%; P<0.0001), lower RVEF (mean, 50±8% versus 53±7%; P=0.0036), and higher RVlesion scores (P<0.0001).
Individuals with ventricular arrhythmia of RBBB morphology (n=34) had significantly increased LVEDV (mean, 125±22% versus 109±17%; P<0.0001), lower LVEF (mean, 58±8% versus 62±7%; P=0.0056), and higher LVlesion scores (P<0.0001).
The QRS axis of the PVCs was determinable in 48 patients. Among patients with ventricular arrhythmia of LBBB configuration, the most frequent pattern was inferior axis and transition at V3 to V4, compatible with RV outflow tract origin (n=34). Thirty-three of these patients had evidence of disease in the RV outflow tract in the form of localized dilation, WMA, fatty replacement, and/or LE. The 11 patients with ventricular arrhythmia of LBBB morphology and superior axis had structural abnormalities of the subtricuspid region (Figure 1D) and/or RV apex. PVCs of RBBB configuration and superior axis were observed in 21 patients, commonly in addition to a primary arrhythmic focus in the RV (n=13); 20 of these patients received gadolinium-DTPA, revealing LVLE at the inferior or inferolateral wall, and 4 had LV apical aneurysms.
LV involvement was weakly associated with advancing age (rpb=0.21, P=0.0032), increasing RV dilation (rpb=0.2, P=0.0037), and higher prevalence of notable ventricular arrhythmia (P=0.0055). There was, however, no significant difference in RVEF between patients with and without LV involvement. Furthermore, of 168 patients with LV involvement, 58 (35%) had normal RV volumes and preserved RV systolic function.
Correlation of ECG, arrhythmic, and CMR findings in the study sample revealed 3 patterns of disease expression: (1) classic, defined as isolated RV disease or LV involvement in the presence of notable RV enlargement and/or dysfunction; (2) left dominant, with prominent LV manifestations in the setting of relatively mild right-sided disease; and (3) biventricular, characterized by equal bilateral involvement with no apparent predilection for either ventricle.
Of 200 study patients, 10 (5%) had left-dominant disease, and 112 (56%) showed a biventricular pattern. Classic disease expression was observed in 78 patients (39%), 46 of whom (59%) had LV involvement. Probands who presented with symptomatic ventricular arrhythmia of LBBB morphology (n=35) were more likely to show the classic pattern of disease expression (P<0.0001) than those with other presentations. Among 43 families, 14 demonstrated phenotypic concordance, with all members belonging to the same disease subgroup (biventricular in 11 families, classic in 3). In the remaining 29 families, relatives showed distinct patterns of disease expression.
Comparison of Disease Patterns
The classic, left-dominant, and biventricular patterns of disease expression are compared in Figure 2. Within the classic subgroup, disease appeared significantly more advanced among subjects with LV involvement (Table 3). Key differences between the subset of the classic group with LV involvement and those with left-dominant and biventricular disease expression are outlined in Figure 3.
Disease Progression Within Subgroups
Significant positive correlation existed between age and ratio of RV to LV volume in the classic group (Pearson r=0.41, P=0.0002), weak negative correlation in the left-dominant group that did not attain statistical significance (Pearson r=- 0.24), and essentially no correlation among subjects with the biventricular disease pattern (Pearson r=−0.07, P=NS).
The 6 individuals with the highest ratios of RV to LV volume in the cohort (mean, 1.83±0.21; range, 1.70 to 2.24) were men 42 to 63 years of age demonstrating classic disease expression with LV involvement; all had presented with symptomatic VT of LBBB morphology. In contrast, the 6 individuals in the biventricular group with the largest percentage increase in ventricular volumes included 4 women. Ventricular arrhythmia of both RBBB and LBBB morphology was present in 5 of 6 patients, and all maintained RV-to-LV volume ratios close to 1 (mean, 1.06±0.09; range, 0.94 to 1.16).
Phenotypic Features of Desmosomal Disease
Patients with defects in desmoplakin, plakophilin-2, desmoglein-2, and desmocollin-2 (n=39) were compared with the 117 individuals who had screened negative for changes in the 5 desmosomal genes (online Data Supplement Table III). The only significant differences were higher LVlesion scores and an increased prevalence of left-dominant disease and notable ventricular arrhythmia (but not nonsustained VT) in the genotyped subset.
Individuals with and without mutations in each of the 3 main desmosomal genes were subsequently compared (online Data Supplement Table IV). Results were unremarkable for the plakophilin-2 and desmoglein-2 subgroups. However, carriers of desmoplakin defects had increased LVEDV, reduced LVEF, higher LVlesion scores, and greater prevalence of notable ventricular arrhythmia, nonsustained VT, lateral T-wave inversion, ventricular arrhythmia of RBBB morphology, and left-dominant disease expression than those who did not. A cross-comparison of individuals with mutations in desmoplakin, plakophilin-2, and desmoglein-2 is presented in Table 4. Finally, individuals carrying nonsense/frameshift mutations resulting in truncated protein products were compared with the rest of the genotyped sample (Table 5).
Of the 9 genotyped families with ≥2 affected members, 3 showed concordant patterns of disease expression, whereas the remaining 6 included relatives belonging to different subgroups. We sought to assess whether unrelated families harboring identical mutations might have similar phenotypes. The previously reported 419C→T mutation in plakophilin-2 was isolated in 2 families.8,13 The probands in both families had suffered sudden cardiac death, with ARVD/C diagnosed on postmortem. Relatives in both families demonstrated concordant biventricular disease expression. Nevertheless, more detailed examination of their clinical profiles revealed key differences. The first family had 2 genetically affected members, the father (54) and sister (33) of the proband. Neither had documented ventricular arrhythmia or ECG abnormalities; both satisfied modified criteria only on the basis of RVWMA. In contrast, the 3 gene carriers in the second family fulfilled TF criteria. The eldest, a 36-year-old woman, had inverted T waves in V2 through V6 and inferiorly. Her daughter developed right precordial T-wave inversion and late potentials by 16 years of age, and her 18-year-old son had frequent PVCs of both RBBB and LBBB configuration. RV aneurysms were evident in all 3 on CMR. An additional first-degree relative in this family had otherwise unexplained RVWMA and nonsustained VT but was negative for the 419C→T mutation.
Elucidation of the natural history of ARVD/C has hitherto relied primarily on correlation of clinical and pathological findings, an approach that inevitably focuses on sudden cardiac death victims and transplant recipients.3,4 Four phases of disease evolution have been documented. In the early “concealed” phase, individuals are often asymptomatic but may nonetheless be at risk of sudden cardiac death, notably during extreme exertion. Structural changes are absent or subtle and frequently are confined to 1 region of the triangle of dysplasia: the inflow, outflow, and apical portions of the RV. The second phase marks the onset of the overt electrical disorder, with symptomatic arrhythmia of RV origin and more prominent RV morphological abnormalities readily discernible by conventional imaging. In the third phase, diffuse RV disease results in isolated right heart failure. LV involvement with biventricular pump failure occurs in the final stage, leading to a phenotype that may resemble dilated cardiomyopathy.1
The few existing studies of familial ARVD/C in genetically heterogeneous populations have used echocardiography as the sole imaging modality.17,19 We conducted comprehensive clinical and CMR assessment of 200 probands and relatives satisfying TF or modified criteria for ARVD/C. Our cohort was distinguished by inclusion of 43 families with ≥2 affected members; 89% of participating individuals had proven familial disease. The study design was unique in using CMR as a surrogate for anatomic examination. Morphological assessment was based on high-temporal-resolution steady-state free-precession sequences, whereas T1-weighted turbo–spin-echo sequences and delayed-enhancement imaging allowed in vivo tissue characterization. Delayed enhancement is a well-established means of demarcating scarring after myocardial infarction; more recently, correlation with fibrofatty replacement in ARVD/C has been demonstrated.20,21 Furthermore, the pattern of LE may be an aid to determining the cause of heart failure; subendocardial involvement is characteristic of ischemic damage, whereas midwall enhancement is observed in dilated cardiomyopathy.20,22
Corrado and coworkers3 presented a clinical–pathological overview of ARVD/C in a multicenter study of 42 patients who had suffered arrhythmia- or heart failure–related death or had undergone cardiac transplantation. CMR findings in our study showed close agreement with the pathological series in several key aspects. First, we observed regional involvement and aneurysmal segments in the RV, preferentially affecting the triangle of dysplasia and the free wall. In the LV, the inferior wall-septal junction and inferolateral wall were favored sites for LE, which occurred in a subepicardial or midwall pattern in concordance with the distribution of fibrofatty replacement described by Corrado and colleagues.
Results of clinical–anatomic correlation also were in keeping with previous investigations. Inverted T waves in the inferolateral leads were an ECG marker of LV involvement, whereas localized QRS prolongation in V1 through V3 was related to increasing RV dilation and fibrofatty replacement.3,23 Consistent with the findings of Turrini and colleagues,24 late potentials were associated with enlarged RVEDV, reduced RVEF, and RVLE, which was considered a marker of fibrosis. Arrhythmia originating in either ventricle or a specific region thereof was frequently associated with identifiable substrate in the form of LE, myocardial fat, or WMA.
Because individuals with early “concealed” ARVD/C are thought to be asymptomatic, symptom-free subjects were expected at the outset to have morphologically mild disease.3 This, however, was not the case. Symptomatic and asymptomatic individuals fulfilling TF criteria did not differ in the majority of structural indices assessed, implying that symptoms may be a poor guide to disease severity. Significant differences were observed in sex ratio and the prevalence of ventricular arrhythmia. Women were more likely than men to report cardiac symptoms, and symptomatic patients were more likely to have ventricular arrhythmia. Conversely, no sex difference existed in the prevalence of ventricular arrhythmia, suggesting that women are more inclined to perceive certain sensations and report them as abnormal. Furthermore, the apparent association between symptoms and ventricular arrhythmia may be spurious because clinician requests for event recorders in symptomatic patients enhance the likelihood of documenting arrhythmia in this population.
Lacking evidence that asymptomatic status equates to mild disease, we sought alternative clinical indicators of the so-called concealed phase. Compared with patients satisfying TF guidelines, those fulfilling modified criteria17 were younger and less likely to have documented ventricular arrhythmia. The 2 diagnostic subsets showed significant differences in all structural indices assessed, affirming a potential role for the modified criteria in identifying individuals with early disease.
Most of the study patients had CMR evidence of LV involvement, of which LE was the most sensitive indicator. That LV involvement is common in ARVD/C is an emerging concept; Lindstrom and colleagues25 reported LV abnormalities in 93% of ARVD/C patients undergoing thallium-201 single-photon emission tomography. In our cohort, localized fat and LE frequently occurred without concomitant WMA or volume expansion, accounting perhaps for the lower estimation of LV involvement by echocardiography. The exception was the frequent occurrence of septal hypokinesia without discernible fibrofatty substrate, for which the phenomenon of ventricular interdependence offers a possible explanation.25
The main departure from the purported natural history of ARVD/C was in the onset of LV disease, with 40% of the cohort demonstrating LV abnormalities in the context of preserved RV systolic function. Furthermore, clinical heart failure, whether right or left sided, was absent from the cohort, even among individuals with significant systolic impairment. These findings challenge the conventional tenet that LV involvement occurs solely as an end-stage complication, preceded by isolated right-sided heart failure and accompanied by biventricular pump failure. Variations in the apparent timing and extent of LV involvement allowed recognition of 3 distinct patterns of disease expression: classic, left dominant, and biventricular. The 3 categories showed similar age distributions, lending support to the premise that the patterns are independent and coexist in the ARVD/C population.
Classic disease expression was characterized by an RV-to-LV volume ratio exceeding 1 (average, 1.2), which increased with advancing age. Individuals with LV involvement had an average RV/LV volume ratio of 1.4, with global RV dilation and various degrees of RV systolic impairment. The inferred pattern of disease progression in this subset was (1) regional RV disease, (2) isolated, global RV dysfunction, (3) localized LV involvement, and (4) biventricular dilation and systolic impairment. This pattern corresponds to the classic sequence of events, although none had signs or symptoms of heart failure. Within the classic subgroup, advanced disease was easily distinguishable from dilated cardiomyopathy owing to the increased RV-to-LV volume ratio; the RV was consistently more severely affected than the LV. Classic disease expression was particularly common among probands presenting with symptomatic arrhythmia of RV origin. Studies recruiting patients on this basis are therefore likely to emphasize this pattern. Interestingly, the classic subgroup showed male predominance, as described in earlier studies,2 whereas the cohort as a whole demonstrated an equal sex ratio, as might be expected in a genetic disease.
Left-dominant disease expression was associated with inverted T-waves confined to the (infero) lateral leads, extensive LVLE with septal involvement, arrhythmia of LV or biventricular origin, and/or isolated LV dilation and impairment. The RV-to-LV volume ratio was typically <1 and showed negative correlation with age, although this did not attain statistical significance. Variants of ARVD/C with a predilection for the LV are increasingly recognized both on postmortem and in vivo.14,26–28 The relative rarity of the left-dominant pattern in our sample is a probable consequence of restrictive inclusion criteria; probands with arrhythmia of exclusively LV origin would not have been eligible. All individuals in the left-dominant group had a family history of ARVD/C, which was the basis of initial enrollment, plus RVWMA and/or aneurysms within the RV, enabling fulfillment of TF or modified criteria. The TF guidelines are likely to exclude many patients with left-dominant disease because several criteria stipulate mild or absent LV involvement.1
Biventricular disease expression was defined by early and parallel involvement of both ventricles. Milder cases typically demonstrated segmental RV dilation and/or WMA and localized LVLE. At least 15% had ventricular arrhythmia of both RBBB and LBBB morphology, underscoring the presence of significant substrate in both ventricles. Advanced disease was characterized by biventricular dilation and/or systolic impairment, with the RV-to-LV volume index remaining close to 1. The phenotype often was a composite of classic features (as outlined in the TF criteria) and left-sided findings such as arrhythmia of LV origin and isolated (infero)lateral T-wave inversion. Both the biventricular and left-dominant patterns may resemble dilated cardiomyopathy with disease progression, although the latter typically presents with heart failure, whereas symptoms of arrhythmia predominate among ARVD/C patients. Correct diagnosis in late presenters may depend on distinguishing features such as prominent WMA, aneurysms, and myocardial fat; in their absence, familial evaluation may prove revealing.
At the molecular level, ARVD/C is considered a disease of the desmosome, the protein complex that anchors intermediate filaments to the cytoplasmic membrane in adjoining cells, thereby imparting mechanical strength to cardiac and epithelial tissue.10 Because desmosomal genes are expressed in both ventricles, no inherent reason exists why a mutation should preferentially affect 1 side of the heart. We therefore postulate that the most common variant of the disease may be biventricular from its early stages. A disease-causing mutation in any component of the desmosome may compromise both intercellular adhesion and intermediate filament function, leading to myocyte detachment and death, followed by fibrofatty repair. Because wall thickness and shear stress are heterogeneous, certain regions of the heart will be more vulnerable than others: the subtricuspid portion, outflow tract, apex, and mid–free wall of the RV, as well as the inferior wall–septal junction and inferolateral wall of the LV.
Early myocyte loss and fibrous replacement in the LV are seldom accompanied by localized deformation and dysfunction, perhaps owing to its thicker walls. In the thin-walled RV, however, fibrofatty atrophy manifests as segmental dilation, WMA, and aneurysm formation. Progressive myocyte loss and fibrotic repair cause gradual biventricular dilation. A proportion of individuals, however, will develop classic RV or predominant left-sided disease. Desmosomal mutations that primarily affect cell adhesion may selectively affect the distensible, thin-walled RV; conversely, myocytes in the high-pressure LV may be more susceptible to defects that impair intermediate filament function.
Mutation screening of 5 desmosomal genes allowed successful genotyping of 39 of 156 probands and relatives. The 2 other genes hitherto implicated in ARVD/C, transforming growth factor-β3 and cardiac ryanodine receptor (RyR2), have not yet been screened in this cohort but are not expected to substantially augment the detection rate.10 Although a significant proportion of the remainder may have mutations in other desmosomal components, this is unlikely to account for the full discrepancy. Phenotypic differences between the subgroups with and without desmosomal mutations were largely absent; in particular, none of the subjects demonstrated the stress-induced polymorphic VT characteristic of RyR2 disease.10 The left-dominant disease pattern was more common among desmosomal mutation carriers, and the extent of LVLE was greater; however, the sparsity of subjects with left-dominant disease in the cohort obscures the clinical significance of these findings.
All 3 disease patterns were observed among the genotyped subset, enabling preliminary investigation into their molecular basis. Overall, structural and ECG indicators of LV involvement were more prominent among individuals with mutations in desmoplakin (online Data Supplement Table IV), as were notable ventricular arrhythmia and nonsustained VT, although differences in the prevalence of arrhythmia became insignificant when subjects with desmoplakin disease were compared with other desmosomal mutation carriers (Table 4). Cross-comparison of phenotypes associated with the 3 main desmosomal genes (desmoplakin, plakophilin-2, and desmoglein-2) was otherwise unremarkable.
Although the predominance of left-sided disease features and ventricular arrhythmia among carriers of desmoplakin mutations may be due to a gene-specific effect, we hypothesized that the high frequency of chain-termination mutations in this subgroup might be another key factor (online Data Supplement Figure I). Individuals with frameshift and nonsense mutations resulting in truncated protein products were found to have a higher prevalence of ventricular arrhythmia and nonsustained VT than the rest of the genotyped subset. Chain-termination mutations also were associated with an increase in the severity of LV involvement, manifested as significantly lower LVEF, more extensive LVLE, and higher prevalence of lateral T-wave inversion and arrhythmia of RBBB morphology. This association is consistent with the premise that defects causing truncation of the C terminus of desmosomal proteins may result in dominant and/or severe LV disease through disruption of intermediate filament binding, a mechanism that may take on particular importance in desmoplakin mutants owing to its direct interaction with desmin.10,12,14,29
Several lines of evidence suggest that the genetics of ARVD/C may be more complex than hitherto appreciated. First, although LV involvement appeared more severe among subjects with chain-termination mutations and/or desmoplakin disease, this association was by no means exclusive. Carriers of missense mutations also showed early and extensive LV involvement, albeit less frequently. Second, the genotyped sample mirrored the cohort as a whole in demonstrating discordant patterns of disease expression within families. Furthermore, because the classic, left-dominant, and biventricular categories each encompass a wide spectrum of disease expression, concordance in this respect does not preclude phenotypic heterogeneity. Thus, 2 unrelated families with identical missense mutations in plakophilin-2 showed concordant biventricular disease expression, although the relatives in 1 family were distinguished by early onset of ECG abnormalities and ventricular arrhythmia. Environmental influences may contribute to the differences. In our study, long-term participation in endurance training resulted in structurally severe disease, a probable corollary of intense and prolonged mechanical stress on the heart. It is more difficult, however, to ascribe distinct patterns of disease expression to variations in physical activity. Genetic factors are likely to play a significant role.
Experience from other inherited cardiovascular diseases, particularly hypertrophic cardiomyopathy and long-QT syndrome, suggests that a small but important minority of families have >1 mutation, although this has yet to be demonstrated in ARVD/C. Double heterozygosity and homozygosity were inferred in only 2 probands in our cohort. Digenic inheritance with a gene-dose effect would, however, account for many hitherto unexplained observations, including relatives with mild but unequivocal disease expression who are “negative” for the family mutation, marked and early LV involvement in individuals with missense mutations, and variable penetrance and expressivity in families with apparently identical genotypes.30 Furthermore, the 30% prevalence of desmosomal mutations in our cohort contrasts with the findings of Gerull et al8 and subsequent studies15,16 that reported a similar (or higher) frequency of mutations in plakophilin-2 alone. Founder effects in certain populations16 and differences in patient selection may be invoked31; our cohort had a predominance of familial disease, often in the context of a deceased proband, whereas the original plakophilin-2 study recruited only live unrelated index cases.8 Nevertheless, among populations of similar ethnicity, selection factors are insufficient to wholly account for the discrepancy, which raises the possibility that some of the numerous missense mutations may be pathogenic only when accompanied by additional genetic changes.8 Although the presence of 2 independently pathogenic mutations may be rare, coexistence or homozygosity of otherwise nonpenetrant “polymorphisms” could result in the development of clinical disease. Modifier genes may therefore play a major role in producing the ultimate phenotype in ARVD/C. Forthcoming multicenter studies of genotype–phenotype associations will shed light on the undoubtedly complex interplay of genetic factors in ARVD/C.
The use of modified diagnostic criteria ensured recognition of early disease forms but also may have resulted in inclusion of false positives. Because specific criteria are currently lacking, a wide variety of features were considered markers of LV involvement on CMR, with consequent potential for overdiagnosis. The small sample size allowed only preliminary analysis of genotype-phenotype associations, which should be interpreted with caution pending large-scale validation in a genotyped population.
In view of the relatively common occurrence of early LV involvement in ARVD/C, we support adoption of the previously proposed term “arrhythmogenic cardiomyopathy,”26,28 with appropriate subclassifications.
We are grateful to Ricardo Wage, DCR(R), and Gillian C. Smith, MSc, for technical support and to Sripurna Das, PhD, for her helpful comments on the manuscript. This work is dedicated to the memory of Asifa Quraishi, MBBS, MRCP.
Sources of Funding
This work was funded by the British Heart Foundation (Dr Sen-Chowdhry, Dr Syrris, Dr Ward, and Dr McKenna) and the European Commission Fifth Framework Program (ARVC/D project QLG1-CT-2000–01091).
Marcus FI, Fontaine G, Guiraudon G, Frank R, Laurenceau JL, Malergue C, Grosgogeat Y. Right ventricular dysplasia: a report of 24 adult cases. Circulation. 1982; 65: 384–398.
Corrado D, Basso C, Thiene G, McKenna WJ, Davies MJ, Fontaliran F, Nava A, Silvestri F, Blomstrom-Lundqvist C, Wlodarska EK, Fontaine G, Camerini F. Spectrum of clinicopathologic manifestations of arrhythmogenic right ventricular cardiomyopathy/dysplasia: a multicenter study. J Am Coll Cardiol. 1997; 30: 1512–1520.
Basso C, Thiene G, Corrado D, Angelini A, Nava A, Valente M. Arrhythmogenic right ventricular cardiomyopathy: dysplasia, dystrophy, or myocarditis? Circulation. 1996; 94: 983–991.
Blomstrom-Lundqvist C, Sabel KG, Olsson SB. A long-term follow-up of 15 patients with arrhythmogenic right ventricular dysplasia. Br Heart J. 1987; 58: 477–488.
Gerull B, Heuser A, Wichter T, Paul M, Basson CT, McDermott DA, Lerman BB, Markowitz SM, Ellinor PT, MacRae CA, Peters S, Grossmann KS, Drenckhahn J, Michely B, Sasse-Klaassen S, Birchmeier W, Dietz R, Breithardt G, Schulze-Bahr E, Thierfelder L. Mutations in the desmosomal protein plakophilin-2 are common in arrhythmogenic right ventricular cardiomyopathy. Nat Genet. 2004; 36: 1162–1164.
Pilichou K, Nava A, Basso C, Beffagna G, Bauce B, Lorenzon A, Frigo G, Vettori A, Valente M, Towbin J, Thiene G, Danieli GA, Rampazzo A. Mutations in desmoglein-2 gene are associated with arrhythmogenic right ventricular cardiomyopathy. Circulation. 2006; 113: 1171–1179.
Bauce B, Basso C, Rampazzo A, Beffagna G, Daliento L, Frigo G, Malacrida S, Settimo L, Danieli G, Thiene G, Nava A. Clinical profile of four families with arrhythmogenic right ventricular cardiomyopathy caused by dominant desmoplakin mutations. Eur Heart J. 2005; 26: 1666–1675.
Syrris P, Ward D, Asimaki A, Sen-Chowdhry S, Ebrahim HY, Evans A, Hitomi N, Norman M, Pantazis A, Shaw AL, Elliott PM, McKenna WJ. Clinical expression of plakophilin-2 mutations in familial arrhythmogenic right ventricular cardiomyopathy. Circulation. 2006; 113: 356–364.
Norman M, Simpson M, Mogensen J, Shaw A, Hughes S, Syrris P, Sen-Chowdhry S, Rowland E, Crosby A, McKenna WJ. Novel mutation in desmoplakin causes arrhythmogenic left ventricular cardiomyopathy. Circulation. 2005; 112: 636–642.
Dalal D, Molin LH, Piccini J, Tichnell C, James C, Bomma C, Prakasa K, Towbin JA, Marcus FI, Spevak PJ, Bluemke DA, Abraham T, Russell SD, Calkins H, Judge DP. Clinical features of arrhythmogenic right ventricular dysplasia/cardiomyopathy associated with mutations in plakophilin-2. Circulation. 2006; 113: 1641–1649.
van Tintelen JP, Entius MM, Bhuiyan ZA, Jongbloed R, Wiesfeld AC, Wilde AA, van der Smagt J, Boven LG, Mannens MM, van Langen IM, Hofstra RM, Otterspoor LC, Doevendans PA, Rodriguez LM, van Gelder IC, Hauer RN. Plakophilin-2 mutations are the major determinant of familial arrhythmogenic right ventricular dysplasia/cardiomyopathy. Circulation. 2006; 113: 1650–1658.
Hamid MS, Norman M, Quraishi A, Firoozi S, Thaman R, Gimeno JR, Sachdev B, Rowland E, Elliott PM, McKenna WJ. Prospective evaluation of relatives for familial arrhythmogenic right ventricular cardiomyopathy/dysplasia reveals a need to broaden diagnostic criteria. J Am Coll Cardiol. 2002; 40: 1445–1450.
Sen-Chowdhry S, Prasad SK, Syrris P, Wage R, Ward D, Merrifield R, Smith GC, Firmin DN, Pennell DJ, McKenna WJ. Cardiovascular magnetic resonance in arrhythmogenic right ventricular cardiomyopathy revisited: comparison with task force criteria and genotype. J Am Coll Cardiol. 2006; 48: 2132–2140.
Tandri H, Saranathan M, Rodriguez ER, Martinez C, Bomma C, Nasir K, Rosen B, Lima JA, Calkins H, Bluemke DA. Noninvasive detection of myocardial fibrosis in arrhythmogenic right ventricular cardiomyopathy using delayed-enhancement magnetic resonance imaging. J Am Coll Cardiol. 2005; 45: 98–103.
McCrohon JA, Moon JC, Prasad SK, McKenna WJ, Lorenz CH, Coats AJ, Pennell DJ. Differentiation of heart failure related to dilated cardiomyopathy and coronary artery disease using gadolinium-enhanced cardiovascular magnetic resonance. Circulation. 2003; 108: 54–59.
De Pasquale CG, Heddle WF. Left sided arrhythmogenic ventricular dysplasia in siblings. Heart. 2001; 86: 128–130.
Sen-Chowdhry S, Syrris P, McKenna WJ. Desmoplakin disease in arrhythmogenic right ventricular cardiomyopathy: early genotype-phenotype studies. Eur Heart J. 2005; 26: 1582–1584.
Corrado D, Thiene G. Arrhythmogenic right ventricular cardiomyopathy/dysplasia: clinical impact of molecular genetic studies. Circulation. 2006; 113: 1634–1637.
Arrhythmogenic right ventricular dysplasia/cardiomyopathy is an inherited myocardial disorder that is a recognized cause of sudden cardiac death. Salient pathological findings are myocyte loss and fibrofatty replacement with an early predilection for the right ventricle. Typical clinical features include arrhythmia of right ventricular origin, right precordial T-wave inversion, and regional followed by global right ventricular dysfunction. The clinical–pathological correlation studies underpinning contemporary understanding of disease evolution have inevitably focused on sudden cardiac death victims and transplant recipients with advanced disease. The aim of the present study was in vivo characterization of both probands and relatives in a genetically heterogeneous sample, using gadolinium-enhanced cardiovascular magnetic resonance imaging as a surrogate for anatomic examination. Our findings challenge the traditional tenet that left ventricular (LV) involvement and biventricular pump failure occur solely as end-stage complications, preceded by isolated right-sided heart failure. Forty percent of the cohort had LV involvement in the presence of preserved right ventricular function. Three patterns of disease expression were apparent: the well-described classic form; a biventricular form characterized by early and parallel involvement of both ventricles, initially localized but progressing to biventricular dilation and/or dysfunction, with arrhythmia of both right ventricular and LV origin; and a left-dominant variant with prominent (but not exclusive) LV disease, manifested as lateral T-wave inversion, arrhythmia of LV origin, and/or LV late enhancement. The spectrum of disease expression in arrhythmogenic right ventricular dysplasia/cardiomyopathy appears far broader than hitherto appreciated, commonly encompassing arrhythmic, ECG, and morphological abnormalities in both ventricles. This finding supports adoption of the broader term arrhythmogenic cardiomyopathy with appropriate subclassifications.
The online-only Data Supplement, consisting of tables and figures, is available with this article at http://circ.ahajournals.org/cgi/content/ full/CIRCULATIONAHA.106.660241/DC1.