Family Screening in Relatives at Risk for Plakophilin-2–Associated Arrhythmogenic Right Ventricular Cardiomyopathy

Abstract

BACKGROUND:

Penetrance and risk of ventricular arrhythmias (VAs) in arrhythmogenic right ventricular cardiomyopathy (ARVC) are increasingly recognized as being genotype specific. Therefore, genotype-informed family screening protocols may lead to safer and more personalized recommendations than the current one-size-fits-all screening recommendations. We aimed to develop a safe, evidence-based plakophilin-2 (PKP2)–specific longitudinal screening algorithm.

METHODS:

We included 295 relatives (41% male; age 30.9 years [18.0–47.7 years]) with a pathogenic or likely pathogenic PKP2 variant from 145 families. Phenotype was ascertained with ECG, Holter monitoring, and cardiac imaging and classified by the 2010 Task Force Criteria. VA was defined as a composite of sudden cardiac arrest or death, spontaneous sustained ventricular tachycardia, ventricular fibrillation, or appropriate implantable cardioverter defibrillator intervention. We performed Cox regression to determine predictors of ARVC development and multistate modeling to assess the probability of ARVC development and occurrence of VA.

RESULTS:

At baseline, 110 relatives (37%) had definite ARVC. During 8.5 years (4.2–12.9 years) of follow-up, 62 of 185 relatives (34%) without definite ARVC at baseline progressed to definite ARVC diagnosis, and 35 of 295 of all relatives (12%) had VA. VAs occurred only in relatives who previously fulfilled definite ARVC diagnosis. Relatives with borderline ARVC (fulfillment of one minor criterion plus the major family history criterion) progressed 5 times faster in the multistate model to definite ARVC diagnosis and compared with genotype-positive/phenotype-negative (G+/P−) relatives (ie, major family history criterion alone). Relatives 20 to 40 years of age had increased risk for developing definite ARVC (hazard ratio, 2.23; P=0.012) compared with those ≥40 years of age. New Task Force Criteria fulfillment most commonly occurred first on ECGs, followed by Holter monitoring and cardiac imaging. Consequently, 3 risk profiles were identified, and appropriate screening protocols were derived: relatives with borderline ARVC (annual ECG and Holter monitoring; complete evaluation [ie, ECGs, Holter monitoring, and imaging] every 2 years), younger (<40 years of age) or symptomatic G+/P− relatives (every 2 years an ECG and Holter monitoring; complete evaluation every 4 years), and older (≥40 years of age) and asymptomatic G+/P− relatives (complete evaluation every 5 years).

CONCLUSIONS:

An evidence-based longitudinal screening algorithm that integrates age, symptoms, and baseline clinical phenotype may improve patient care and improve efficiency of clinical resource allocation.

Clinical Perspective.

What Is New?

• Relatives with a pathogenic or likely pathogenic variant plakophilin-2 (PKP2) variant with borderline arrhythmogenic right ventricular cardiomyopathy (ARVC) progress ≈5 times faster to definite ARVC diagnosis compared with relatives who are genotype positive/phenotype negative (G+/P−).

• Relatives 20 to 40 years of age or who are symptomatic have an ≈2 times higher risk of progressing to definite ARVC diagnosis.

• Relatives with a pathogenic or likely pathogenic PKP2 variant without definite ARVC diagnosis at baseline rarely develop ventricular arrhythmias during medium-term follow-up and only after fulfilling definite ARVC diagnostic criteria.

What Are the Clinical Implications?

• We identified 3 risk profiles for at-risk relatives: (1) borderline ARVC; (2) symptomatic or <40 years of age and G+/P−; and (3) asymptomatic ≥40 years of age and G+/P−.

• Using these risk profiles, we developed longitudinal screening algorithms that ensured diagnosis >5 years before ventricular arrhythmias occur: (1) for borderline ARVC, annual screening with an ECG and Holter monitoring and a complete evaluation every 2 years; (2) for symptomatic or <40 years of age and G+/P−, screening every 2 years with an ECG and Holter monitoring and a complete evaluation every 4 years; and (3) for asymptomatic ≥40 years of age and G+/P−, screening every 5 years with a complete evaluation.

Arrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited disease that is most commonly caused by a pathogenic or likely pathogenic (P/LP) variant in plakophilin-2 (PKP2).¹ The hallmark features of the disease are sustained ventricular arrhythmias (VAs), including sudden cardiac death.^1,2 Because a familial predisposition to PKP2-associated ARVC can be identified through genetic testing, longitudinal screening of relatives of patients with ARVC is strongly recommended.^3–6 Indeed, the current guidelines recommend initiation of screening at 10 to 12 years of age and re-evaluation with an ECG, Holter monitoring, and an imaging modality every 1 to 3 years.^3–6

Unfortunately, the currently recommended screening algorithms do not account for genotype-specific differences in penetrance, risk of VA, or their associated risk factors.^1,7–20 Therefore, this one-size-fits-all family screening algorithm may result in overscreening of patients with a low likelihood of developing penetrant ARVC and underscreening of those who are at the highest risk, leading to inefficient use of clinical resources and potential missed windows of opportunity for primary prevention. Although longitudinal family screening studies have been performed previously, these studies enrolled relatives of patients with ARVC with different underlying genotypes, as well as gene-elusive families.^21–23 Thus, to date, there is limited evidence to support genotype-specific longitudinal screening algorithms.²⁴

To address this gap, we evaluated outcomes of a large cohort of relatives of patients with ARVC with P/LP PKP2 variants to develop such an algorithm. The purpose of this study was to develop a safe, evidence-based longitudinal screening algorithm for relatives with a P/LP PKP2 variant.

METHODS

Study Population

The study population was recruited from the Netherlands ACM registry (Arrhythmogenic Cardiomyopathy²⁵; www.acmregistry.nl; UCC-UNRAVEL No. 12-387²⁶) and the Johns Hopkins ARVC registry. From both registries, we identified all families in which the proband fulfilled definite ARVC diagnosis as per 2010 Task Force Criteria (TFC)²⁷ and who carried a P/LP PKP2 variant. All PKP2 variant classifications were adjudicated by an experienced cardiac genetic counselor (B.M.) according to the standards and guidelines for the interpretation of sequence variants set forth by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology.²⁸ Among the families identified, all relatives with a P/LP PKP2 variant who underwent a baseline screening evaluation in which definite ARVC diagnosis could be ascertained were included, as described later. We excluded relatives who had a sustained VA before their first clinical evaluation because they were no longer at risk for a first VA. The study population partially consisted of relatives who had been included in previous ARVC studies.^{22,23,29–33} The data used in this study are available on reasonable request. This study followed the code of conduct and the use of data in health research and was approved by local ethics or institutional review boards.

Clinical Evaluation

Participants were evaluated as described previously.^22,23,34 Medical history was obtained by review of medical records or clinical evaluation. Detailed clinical information on demographics, presentation, symptom onset, and noninvasive testing was obtained. Pedigree analysis was performed by genetic counselors with expertise in ARVC. Relatives were categorized according to their relationship to the proband as first-, second-, or third-degree relatives; first-degree relatives were subdivided into parents, siblings, and children of the proband.

All relatives underwent guideline-recommended baseline screening evaluation in which definite ARVC diagnosis could be ascertained, defined as a 12-lead ECG, Holter monitor of at least 24 hours, and an imaging evaluation of cardiac structure and function (echocardiogram or cardiac magnetic resonance imaging [CMR]).^3–6 Testing results from other modalities (eg, electrophysiology studies) were also collected. For follow-up evaluations, we included all available clinical testing performed at the discretion of the treating cardiologist, including 12-lead ECG, Holter monitoring of ≥24 hours, and repeat imaging (echocardiogram or CMR).

ARVC Diagnosis

Diagnostic evaluation was based on the 2010 TFC,²⁷ with definite ARVC diagnosis defined by fulfillment of 2 major, one major plus 2 minor, or 4 minor criteria. By study design, all subjects fulfilled one major criterion in the family history category given their genetic predisposition (ie, harboring a P/LP PKP2 variant). Thus, relatives were stratified by their baseline clinical phenotype: (1) genotype positive/phenotype negative (G+/P−; ie, only the major family history criterion without any other TFC fulfillment on diagnostic tests), (2) borderline ARVC (ie, fulfillment of one minor criterion plus the major family history criterion), or (3) definite ARVC (ie, fulfillment of 2 minor or one major plus the major family history criterion).

Study Outcomes

The primary outcomes of this study were the development of definite ARVC during follow-up as per 2010 TFC and the occurrence of sustained VA during follow-up. For the purpose of this study, sustained VA was defined as a composite of sudden cardiac death, sudden cardiac arrest, spontaneous sustained ventricular tachycardia (ventricular tachycardia lasting ≥30 seconds at ≥100 bpm or with hemodynamic compromise requiring cardioversion), ventricular fibrillation/flutter, or appropriate implantable cardioverter defibrillator intervention, as done previously.²⁹

The secondary outcome of this study was the occurrence of a new TFC during follow-up that was absent at the first evaluation. This was evaluated for each diagnostic test (ie, ECG, Holter monitor, echocardiogram, and CMR) separately.

Statistical Analysis

Nominal variables were expressed as number (percent) and continuous variables as mean±SD or median (interquartile range [IQR]) as appropriate. Comparisons for binary variables were performed by the χ² or Fisher exact test. For continuous variables, independent t tests or Mann-Whitney U tests were used. To visualize progression to definite ARVC diagnosis, new TFC, and VA, Kaplan-Meier curves were plotted, and groups were compared with the log-rank test. The event rates for VA, including the 95% CI, were calculated with the Kaplan-Meier method. Because progression to definite ARVC diagnosis and fulfillment of new TFC are dependent on diagnostic test results, our analysis could potentially be biased because of the unavailability of tests. To assess this potential bias, we repeated the analyses for progression to definite ARVC diagnosis and progression to a new TFC including only relatives who had a complete repeat evaluation (ie, those who had an ECG, Holter monitoring, echocardiography, and CMR during follow-up available) as a sensitivity analysis. No missing data were imputed. A value of P<0.05 was considered statistically significant. Data were analyzed with R version 4.1.2 (Boston, MA), including the msm package for the multistate model.³⁵

Overcoming Interval Censoring

Because this was a retrospective study without predefined follow-up dates, the date of diagnosis in our study population was dependent on the timing of outpatient clinic visits; that is, our analyses are sensitive to interval censoring. To overcome that limitation, we created a multistate model (continuous-time Markov chain). In this model, we determined G+/P−, borderline ARVC, definite ARVC, and VA as separate clinical states through which an individual can progress unidirectionally to a more advanced disease stage (Figure 1A shows a visualization of how an individual could progress). Transitioning to a less severe disease state (eg, from borderline ARVC to G+/P−) was deemed impossible. The multistate model considers transitions from one disease state to another to occur at an unknown time point between 2 outpatient clinic evaluations at which the current disease state could be assessed. The model therefore estimates the probability of transitioning to a more severe disease state between 2 time points. VA, in contrast to the other disease states, was modeled as a time-stamped event. An in-depth description of the type of Markov chain and the assumptions and restrictions of the model is provided in the Supplemental Methods.

Figure 1.

Visualization of the multistate model. A, Multistate model of progression of disease using the diagnostic 2010 Task Force Criteria (TFC) consisted of 4 states: genotype positive/phenotype negative (G+/P−), borderline arrhythmogenic right ventricular cardiomyopathy (ARVC), definite ARVC, and ventricular arrhythmia (VA). A relative could only stay in the same state or progress to a more severe state during follow-up, which is shown by the blue arrows. Numbers represent the estimated transition intensities per year. B, Multistate models of progression to new TFC by diagnostic test. Numbers represent the estimated transition intensities per diagnostic test per year.

Determining When to Screen

We used the fitted risks from the multistate model as mentioned in 3 complementary ways to determine clinically acceptable risks for progression to definite ARVC. First, because the current guidelines^3–6 recommend a 1- to 3-year screening interval, we evaluated the 1- to 3-year risk for definite ARVC and VA for G+/P− relatives and relatives with borderline ARVC. Second, we determined which screening intervals for G+/P− relatives and relatives with borderline ARVC had comparable risks for definite ARVC and VA. Third, we determined which screening intervals for G+/P− relatives and relatives with borderline ARVC would hold acceptable risks in daily clinical practice. To do so, we deemed a clinically acceptable risk for definite ARVC development to be between 6% and 16%, as previously reported for risk for definite ARVC development in the overall at-risk ARVC population.²² For VA, a clinically acceptable risk was deemed to be <0.5% in the setting of family screening.

To determine how (ie, by which modalities) relatives should be screened, we created multistate models for every diagnostic test separately in which no progression and progression were the available states (Figure 1B shows a visualization of how an individual could progress). We performed these analyses only in those who did not fulfill definite ARVC diagnosis at baseline evaluation.

Establishing Predictors of Progression

To determine which individuals were at a higher risk of progression (ie, who would benefit from more frequent follow-up visits), we additionally examined individual risk predictors for progression of development of ARVC using predefined variables known to be associated with progression to definite ARVC.^22,29 Predictors for the definite ARVC diagnosis were tested by a frailty Cox proportional hazard regression to adjust for a potential effect of family membership, and hazard ratios were reported with their 95% CIs. The Cox proportional hazard assumptions were tested with the Schoenfeld residuals.³⁶

Longitudinal Screening Algorithm Development

To develop a safe, evidence-based longitudinal screening algorithm, we combined the natural rate of progression (ie, fitted risk from all multistate models) and the risk factors to identify different risk profiles. Using these different risk profiles, we subsequently proposed a longitudinal screening algorithm accounting for the natural progression of disease (ie, when to screen) and new TFC per diagnostic test (ie, how to screen) for relatives in each of these risk groups.

RESULTS

Study Population

Our study cohort consisted of 295 relatives of probands from 145 families (relatives per family: median, 1 [IQR, 1–3]; range, 1–9) who harbored a familial P/LP PKP2 variant (Table S1 gives P/LP PKP2 variants) and had follow-up evaluations available (Figure 2). Median age at first evaluation was 30.9 years (IQR, 18.0–47.7 years); 122 (41%) were male; and 205 (69%) were first-degree relatives (Table 1). Most relatives were asymptomatic (n=159/236; 67%), whereas among the symptomatic relatives (n=77; 33%), palpitations were the most prevalent symptom (n=41/77; 53%). During the follow-up period (median, 8.5 years [IQR, 4.2–12.9 years]); 7 of 295 relatives (2.4%) died. All causes of death were likely not ARVC related, and no sudden cardiac deaths were observed. Table S2 summarizes individual causes of death.

Figure 2.

Flowchart of the study population. *Complete baseline evaluation defined as at least 12-lead ECG, Holter monitoring, and imaging (cardiac magnetic resonance imaging [CMR] or echocardiography). Boxes around the study population visualize specific analyses as described in the text.

Table 1.

Baseline Characteristics

Progression of Disease

Progression to Definite ARVC Diagnosis

Sixty-two of the 185 relatives (34%) without definite ARVC diagnosis at first clinical evaluation (dotted box in Figure 2; median follow-up time, 8.2 years [IQR, 4.0–12.1 years]) progressed to definite ARVC diagnosis at a median time after initial evaluation of 5.3 years (IQR, 2.5–8.7 years; Figure 3A). Relatives with borderline ARVC (4.1 years [IQR, 1.6–7.0 years]) progressed significantly faster to definite ARVC diagnosis compared with G+/P− relatives (7.7 years [IQR, 4.6–9.5 years]; P<0.001; Figure 3B).

Figure 3.

Progression of disease. Survival curve of definite arrhythmogenic right ventricular cardiomyopathy (ARVC) diagnosis in the overall, nondefinite population (A) and stratified by baseline clinical phenotype (B). Relatives who are genotype positive/phenotype negative (G+/P−) and who have borderline ARVC are depicted as yellow and blue lines, respectively. Shaded areas indicate 95% CI. Censoring is indicated by a vertical bar.

Occurrence of VA

Of all 295 relatives, including those with definite ARVC (solid box, Figure 2; median follow-up, 8.5 years [IQR, 4.2–12.9 years]), 35 (12%) had a VA during follow-up, with a corresponding annual rate of 1.4% (95% CI, 0.9%–1.8%). In those 35 relatives, median time to VA was 6.0 years (IQR, 2.3–9.3 years; Figure 4A). The youngest relative experiencing VA was 14 years of age. Of note, no VAs occurred before a definite ARVC diagnosis. Consequently, of the 172 relatives with a definite ARVC diagnosis by last follow-up (dashed box in Figure 2; Table S3), the 35 VAs corresponded to an annual rate of VA of 2.7% (95% CI, 1.7%–3.5%) while fulfilling definite ARVC diagnosis (Figure 4B).

Figure 4.

Occurrence of VA. Survival curve of ventricular arrhythmia (VA) after baseline evaluation (A), arrhythmogenic right ventricular cardiomyopathy (ARVC) diagnosis (B), baseline evaluation stratified by baseline clinical phenotype (C), and ARVC diagnosis stratified by baseline clinical phenotype (D). Relatives with and without definite ARVC diagnosis at baseline are depicted as red and turquois lines, respectively. Because only 3 relatives (1.6%) without definite ARVC had a VA during follow-up, we combined the genotype-positive/phenotype-negative (G+/P−) relatives and relatives with borderline ARVC into a “no ARVC at baseline” group for visual purposes in C and D. Shaded areas indicate 95% CI. Censoring is indicated by a vertical bar.

In total, 3 of 185 relatives (1.6%) without a definite ARVC diagnosis at baseline had a VA during follow-up. These events all occurred after definite ARVC diagnosis (time from definite ARVC diagnosis to VA, 6.8, 9.6, and 14.5 years). It is notable that these 3 relatives without a definite ARVC diagnosis at baseline progressed significantly more slowly and less often to VA from baseline evaluation (time from baseline evaluation to VA, 7.8, 14.9, and 18.3 years) compared with those with definite ARVC diagnosis at baseline (median time to VA, 5.5 years [IQR, 2.1–8.4 years]; P<0.001; Figure 4C and 4D). Phenotypic expression at time of diagnosis was less severe in those who developed diagnosis during follow-up compared with those who were diagnosed at baseline (Table S3).

Yield of Diagnostic Testing in Relatives Without ARVC

We analyzed the development of new TFC fulfillment per diagnostic test in the 185 relatives without definite ARVC diagnosis at the baseline evaluation. Disease trajectory for every relative is visualized in Figure S1.

Overall, 84 of 185 relatives (45%) developed a new TFC during follow-up. Median time to a new TFC was 4.5 years (IQR, 2.1–7.1 years; Figure S2A). New TFC were most commonly first observed on ECG (n=45/84, 54%), followed by Holter monitoring (n=25/84; 30%) and imaging (n=14/84; 17%; Figure S2B). New imaging criteria were most commonly observed first on CMR (n=11/14, 79%), followed by echocardiography (n=2/14; 14%) and both modalities at one time point (n=1/14; 7%; Figure S2C).

G+/P− relatives and relatives with borderline ARVC progressed at rates similar to those for new TFC (P=0.380; Figure S3A). In addition, no significant difference in rate of progression to a new TFC per diagnostic test was found between G+/P− relatives and those with borderline ARVC (all P>0.05; Figure S3B through S3F).

Sensitivity Analyses

We next repeated these analyses in patients with complete repeat evaluations (ie, all had a repeat ECG, Holter monitoring, echocardiography, and CMR during follow-up available). Of note, relatives with a complete repeat evaluation were significantly younger (P=0.002) and more often had a prolonged terminal activation duration (P=0.019) compared with those without a full repeat evaluation (Table S4). Nonetheless, these sensitivity analyses yielded similar results for both progression to definite ARVC diagnosis (Figure S4) and progression to new TFC criterion per diagnostic test (Figure S5), including when the full repeat evaluation cohort was stratified by baseline clinical phenotype (Figures S6 and S7).

Multistate Modeling of Progression of Disease to Overcome Interval Censoring

Probabilities of Definite ARVC and VA

The probabilities of progression to definite ARVC diagnosis, VA, and new TFC based on multistate modeling are presented in Figure 5 (Figures S8 and S9 show the continuous-time Markov model survival curves). Patients with borderline ARVC had a 5-fold higher rate of progressing to definite ARVC and occurrence of VA compared with G+/P− relatives, which was consistent throughout the follow-up period (Figure 5A and Figure S10 for smaller fitted risks for VA; P<0.001 for definite ARVC and P<0.05 for VA). Calibration plots are presented in Figures S11 and S12.

Figure 5.

Progression to developing definite ARVC, VA, and new TFC. A, Yield of screening stratified by baseline clinical phenotype for definite arrhythmogenic right ventricular cardiomyopathy (ARVC) and ventricular arrhythmia (VA; y axis) over different screening intervals (x axis). Every screening interval consists of 3 bars, which depict the genotype-positive/phenotype-negative (G+/P−; yellow), borderline ARVC (blue), and definite ARVC (gray) groups. The bars are divided into solid and striped bars, which differentiate between the probability of definite ARVC and VA, respectively. Of note, because relatives with definite ARVC have a 100% - “probability of VA” probability of definite ARVC, we visualize only the probability of VA in that population. B, Yield of screening new Task Force Criteria (TFC) and ECG, Holter monitoring, and imaging TFC over different screening intervals. C, Yield of screening for new cardiac magnetic resonance imaging (CMR) and echocardiography TFC over different screening intervals. ECG, Holter monitoring, imaging, CMR, and echocardiography are indicated by red, green, blue, dark blue, and light blue bars, respectively. Dotted black line indicates the clinically acceptable risk for definite ARVC (ie, 6%–16%). Error bars indicate 95% CI.

The 1- to 3-year risk for definite ARVC and VA in relatives with borderline ARVC was as follows: risk for definite ARVC at 1 year, 11% (95% CI, 9%–14%) and at 3 years, 29% (95% CI, 23%–35%); risk for VA at 1 year, 0.05% (95% CI, 0.02%–0.1%) and at 3 years 0.5% (95% CI, 0.2%–1%). These risks were higher compared with G+/P− relatives (risk for definite ARVC at 1 year, 0.5% [95% CI, 0.3%–0.7%] and at 3 years, 4% [95% CI, 3%–5%]; risk for VA at 1 year, <0.01% [95% CI, <0.01%–<0.01%] and at 3 years, 0.04% [95% CI, 0.02%–0.1%]). The 1- to 3-year risk for definite ARVC or VA in relatives with borderline ARVC was similar to the 4- to 7-year risk for definite ARVC or VA in G+/P− relatives (Figure 5A). Both the 1-year screening interval in relatives with borderline ARVC and the 4- to 7-year screening interval in G+/P− relatives fell within the clinically acceptable thresholds (ie, 6% to 16% and <0.5% screening risk for definite ARVC and VA, respectively; Figure 5A).

Probabilities of New TFC by Diagnostic Test

Patients without definite ARVC were more likely to progress to new TFC based on electrocardiographic findings (1-year risk of new TFC, 5.3% [95% CI, 4.2%–6.9%]; 3-year risk of new TFC, 15.2% [95% CI, 11.4%–20.5%]), followed by Holter monitoring (1-year risk of new TFC, 3.3% [95% CI, 2.3%–4.5%]; 3-year risk of new TFC, 9.5% [95% CI, 6.5%–14.1%]) and then imaging criteria (1-year risk of new TFC, 2.0% [95% CI, 1.4%–3.0%]; 3-year risk of new TFC, 6.0% [95% CI, 3.9%–9.1%]; P_ECG-imaging<0.05, P_{Holter-imaging} and P_ECG-Holter>0.05; Figure 5B). A nonsignificant trend was observed showing that the progression to new imaging criteria was driven mostly by new CMR TFC, which was followed at a later stage by echocardiography TFC (P>0.05; Figure 5C).

Predictors of Progression

Relatives 20 to 40 years of age showed a significantly higher hazard for definite ARVC (hazard ratio, 2.33; P=0.012) compared with relatives ≥40 years of age (Table 2). Female sex was associated with higher risk of definite ARVC diagnosis at follow-up (hazard ratio, 2.38; P=0.002). In addition, symptomatic patients had a trend toward development of definite ARVC (hazard ratio, 1.79; P=0.058).

Table 2.

Frailty Cox Proportional Hazard Regression for Definite ARVC Diagnosis

Risk Profiles and Longitudinal Screening Algorithm

We defined 3 distinct risk profiles for relatives of patients with ARVC based on multistate modeling and the predictors: (1) relatives with borderline ARVC (n=68/185; 37%); (2) G+/P− relatives who were <40 years of age or symptomatic (n=94/185; 51%), and (3) asymptomatic G+/P− relatives who were ≥40 years of age (n=23/185; 12%). The proposed longitudinal screening algorithm derived from the 3 risk profiles is shown in Figure 6 and described in detail here.

Figure 6.

Longitudinal screening algorithm for at-risk pathogenic or likely pathogenic PKP2 carriers. The longitudinal screening algorithm. The baseline evaluation should start between 8 and 10 years of age, and the serial evaluation depends on the baseline clinical phenotype, presence of symptoms, and age: (1) genotype-positive/phenotype-negative (G+/P−) relatives who are ≥40 years of age and are asymptomatic should be evaluated every 5 years with an ECG, Holter monitoring, and imaging test; (2) G+/P− relatives who are <40 years of age or are symptomatic should be evaluated every 2 years with an ECG and Holter monitoring and every 4 years with an ECG, Holter monitoring, and imaging test; and (3) relatives with borderline arrhythmogenic right ventricular cardiomyopathy (ARVC) should be evaluated every year with an ECG and Holter monitoring and every 2 years with an ECG, Holter monitoring, and imaging test. If a relative remains in the same clinical phenotype (G+/P− or borderline ARVC) or progresses from G+/P− to borderline ARVC, screening intervals and management should be adjusted accordingly as shown by the dotted lines. If a relative progresses to definite ARVC, management and risk stratification should follow the most current guideline recommendations. Following this screening algorithm ensures that all relatives will be screened at a similar risk for progressing to definite ARVC (G+/P−, 9%; borderline ARVC, 11%) and ventricular arrhythmia (VA; both G+/P− and borderline ARVC, <0.5%). CMR indicates cardiac magnetic resonance imaging.

Borderline ARVC

Relatives with borderline ARVC had an 11% risk for development of definite ARVC and a 0.06% risk for VA at the 1-year follow-up (Figure 5A). Of note, the majority of new diagnoses were based on new ECG- and Holter monitoring–based TFC, with only 2% of new diagnoses explained by new imaging TFC (Figure 5B). Consequently, the number needed to image at a 1-year screening interval was 50. Therefore, proposing to annually screen relatives with borderline ARVC with only an ECG and Holter monitoring and a complete evaluation (ie, ECG, Holter monitoring, and imaging) every 2 years to use resources efficiently while still upholding a safe screening interval concerning the risk of definite ARVC and VA is clinically acceptable. Regardless, if a new TFC on ECG or Holter monitoring is established and a definite ARVC diagnosis is made, an imaging test is warranted for VA risk stratification.

Younger (<40 Years of Age) or Symptomatic G+/P− Relatives

Because a 4-year screening interval is well within the clinically acceptable risk for definite ARVC (6.4% [95% CI, 4.5%–8.9%]) and VA (0.09% [95% CI, 0.04%–0.25%]) in younger or symptomatic G+/P− relatives, a complete evaluation (ie, ECG, Holter monitoring, and imaging) at a 4-year interval is adequate (Figure 5A). However, to avoid missing any relevant disease progression but also to use clinical resources efficiently, we propose an additional evaluation every 2 years with an ECG and Holter monitoring. The combined yield for new TFC on ECG and Holter monitoring at a 2-year interval was 16.8% (95% CI, 12.4%–23.2%), and the added yield of an imaging modality for new TFC was 4.0% (95% CI, 2.7%–6.2%; Figure 5B). Of note, if a new TFC on ECG or Holter monitoring is established, an additional imaging test is warranted to adjudicate a potential definite ARVC diagnosis.

Older (≥40 Years of Age), Asymptomatic G+/P− Relatives

A 5-year interval (9.3% [95% CI, 6.5%–13.0%] for definite ARVC; 0.2% [95% CI, 0.07%–0.45%] for VA; Figure 5A) for older and asymptomatic G+/P− relatives is well within the acceptable risk. Therefore, a 5-year screening interval with an ECG, Holter monitor, and an imaging test in older (≥40 years of age), asymptomatic G+/P− relatives is acceptable in daily clinical practice.

DISCUSSION

This study represents the first systematic investigation of genotype-specific progression to definite ARVC and VA in relatives who harbor a familial P/LP PKP2 variant. These results provide direct evidence to support a genotype-specific longitudinal family screening approach. From our findings, we propose distinct intervals of longitudinal screening for: (1) relatives with borderline ARVC (every year with an ECG and Holter monitoring and every 2 years with a complete evaluation [ie, ECG, Holter monitoring, and imaging]); (2) G+/P− relatives who were <40 years of age or symptomatic (every 2 years with an ECG and Holter monitoring and every 4 years with a complete evaluation), and (3) asymptomatic G+/P− relatives ≥40 years of age (every 5 years with a complete evaluation; Figure 6).

Relationship Among Family Screening, Early Diagnosis, and VA

Contemporary clinical guidelines recommending genetic testing in patients with ARVC and families have brought an increasing number of G+ relatives at risk for ARVC to clinical attention. Since the inception of family screening recommendations,³⁷ a number of studies have established the rate of progression to definite ARVC in G+ relatives.^21–23 a recent study found that the rate of structural progression does not differ between relatives and probands, regardless of phenotypic expression.³⁸ Our results enforce and extend these findings by showing that the rates of both structural and electrical progression are similar between clinical phenotypes. These findings are important to recognize because they underline that PKP2-associated ARVC follows a specific mode of progression, which is in contrast to other types of arrhythmogenic cardiomyopathy such as desmoplakin (DSP) cardiomyopathy.^7,8,12

Prior studies have shown that up to one-third of relatives have definite ARVC at baseline.³⁹ Indeed, more recent studies have confirmed this yield of baseline evaluation and extended it by showing that, after a median follow-up of 4 years, an additional one-third of relatives without definite ARVC at baseline progressed to definite ARVC diagnosis,^21–23 thus demonstrating the benefit of longitudinal family screening in at-risk relatives for ARVC. However, these studies did not model the impact of genotype on the development of ARVC.^21–23 The present study extends previous findings by including solely relatives harboring the familial P/LP PKP2 variant and establishing a safe, evidence-driven, genotype-specific longitudinal screening algorithm, including risk of VA.

Impact of Family Screening on Relatives Not Fulfilling Definite ARVC Diagnosis

Similar to previous family screening studies,^21–23 we found that the baseline yield of screening is approximately one-third and that an additional one-third can be diagnosed after 4 to 5 years of follow-up. We extended these findings by showing that relatives with the familial P/LP PKP2 variant who progress to definite ARVC diagnosis during follow-up had less severe phenotypic expression at the time of diagnosis compared with those with definite ARVC at baseline. This is important because these relatives with PKP2-associated ARVC are diagnosed in a phase when lifestyle modification (ie, limiting exercise) may have a great influence on slowing ARVC progression and first VA occurrence.^40–43 Although we are fully aware that G+/P− relatives are recommended and should adhere to the exercise restrictions to slow ARVC progression,⁴² a recent survey⁴⁴ showed that a nonnegligible portion of patients with ARVC will start to limit their exercise intensity at the time of definite ARVC diagnosis. This shows that early diagnosis, besides risk stratification for VA, has a practical implication in daily clinical practice to initiate lifestyle modification in PKP2-associated ARVC, which is in contrast to relatives with phospholamban (PLN)-associated ARVC in whom lifestyle modification does not influence progression of disease.¹⁷

Furthermore, we established that the rate of VA is very low and VA occurred only in those who fulfilled 2010 TFC during follow-up. In contrast to other studies,^1,45 we did not report any VAs in the “concealed ARVC” phase (ie, P/LP PKP2 carriers not fulfilling definite ARVC diagnosis). This is reassuring for relatives harboring a P/LP PKP2 variant because it shows that, when following our proposed longitudinal screening algorithm, all relatives are diagnosed with definite ARVC well before their first VA. These PKP2-specific findings are notably different from the much higher risks of G+ relatives with other genotypes, for example, the risks seen in relatives with the ARVC-associated transmembrane protein 43 (TMEM43) P.S358L variant.¹³

We also established clinical characteristics that are associated with definite ARVC in PKP2 relatives. We showed that age-related progression peaked in early adulthood with a 2-fold higher hazard for definite ARVC diagnosis in relatives 20 to 40 years of age compared with those ≥40 years of age and shows that age-specific screening intervals are justified. Our results showed that female sex is associated with an ≈2-fold higher hazard of developing definite ARVC. However, previous studies have shown that female sex is associated with a lower hazard of VA. The well-established high risk of VA in young male patients with ARVC partially explains why men are more likely to be the proband of a family,^29,46 leading to a female predominance in this and other ARVC family screening studies.

Diagnostic Tests Used in Family Screening

We also investigated how relatives progressed to new TFC to ultimately determine which diagnostic tests should be prioritized in the outpatient clinic. In line with our previous studies,^21,22 we showed that G+/P− relatives and relatives with borderline ARVC progress to new TFC at a similar rate. We extended this finding by showing that the modality of progression (ie, ECG, Holter monitoring, or imaging) in these relatives is also similar. In addition, a trend was observed in relatives without definite ARVC diagnosis who develop imaging TFC, indicating that CMR TFC precede echocardiography TFC. This may reflect the expertise required to perform precise right ventricular structural and functional quantification on echocardiography.

Longitudinal Screening Algorithm in At-Risk Relatives

When a longitudinal screening algorithm is proposed, a major consideration is given to when to initiate family screening. Consistent with prior studies,^1,23 we found that the penetrance of definite ARVC diagnosis at <14 years of age is rare (n=1/110; 0.9%) and that VAs can occur in adolescence (youngest individual who experienced VA in our study population was 14 years of age). Therefore, initiating screening at 8 to 10 years of age is justifiable to diagnose severe cases before VA occurrence. After a baseline evaluation, 2 scenarios are possible: (1) if a definite ARVC diagnosis is made, decisions on the management (including VA risk stratification^29–31,47) should be based on guideline-directed recommendations^3–6; and (2) if no definite ARVC diagnosis is made, the evidence-driven PKP2-specific and personalized algorithm can be used for longitudinal screening (Figure 6).

Although our longitudinal screening algorithm warrants external validation, we believe that the findings of this study are reassuring for clinicians and families, considering the large sample size of relatives not fulfilling 2010 TFC at baseline with long-term follow-up, yielding a very low VA rate that occurred only after being diagnosed with definite ARVC for ≥5 years when our longitudinal screening algorithm was followed.

Study Limitations

Although our cohort of comprehensively evaluated relatives with PKP2 P/LP variants is a relatively large cohort of relatives of an individual with ARVC, we were underpowered to perform extensive multivariable analyses to ascertain independent predictors, particularly in the occurrence of VA. In addition, we did not include medication use (eg, beta-blockers) as a factor for disease progression, but they were initiated in some patients. In addition, because of the unavailability of exercise data for many relatives, we did not include exercise as a factor for disease progression despite its well-documented influence on development of ARVC and VA in PKP2-associated ARVC.^40,42,43 Certainly, screening intervals should be shortened in family members who choose to participate in competitive or frequent high-intensity exercise. Last, our multistate model was built mostly on G+/P− relatives who were younger or symptomatic, which could have led to a slight overestimation of progression of older asymptomatic G+/P− relatives.

Conclusions

In this study, we showed that although the development of ARVC is slow in relatives with P/LP PKP2 variants, younger relatives are at higher risk of disease progression. In addition, disease-related abnormalities were typically appreciated first on ECG and Holter monitoring. We propose a longitudinal screening algorithm that differentiates between relatives with: (1) borderline ARVC (annual screening with ECG and Holter monitoring and a complete evaluation every 2 years), (2) G+/P− relatives who are younger (<40 years of age) or symptomatic (an evaluation with ECG and Holter monitoring every 2 years with complete evaluation every 4 years), and (3) older (≥40 years of age), asymptomatic G+/P− relatives (complete evaluation every 5 years). When the proposed algorithm was evaluated in the present cohort, all potentially at-risk relatives were diagnosed with definite ARVC diagnosis at least 5 years before their VA.

ARTICLE INFORMATION

Acknowledgments

The authors thank the ARVC relatives who have made this work possible.

Sources of Funding

Dr Muller is supported by the Netherlands Heart Foundation, grant CVON 2018-30 Predict2 Young Talent Program. Dr Asatryan is supported by the 2022 research fellowship for aspiring electrophysiologists from the Swiss Heart Rhythm Foundation and a postdoctoral research fellowship grant from the Gottfried und Julia Bangerter-Rhyner-Stiftung (Switzerland). Dr Carrick is funded by a National Institutes of Health T32 grant (T32HL007227) and the National Institutes of Health loan repayment program (L30HL165535). Dr te Riele is supported by the ZonMW clinical fellows grant 2024 and ERC HORIZON IMPACT (#101115536). The Johns Hopkins ARVC Program is supported by the Leonie-Wild Foundation; the Leyla Erkan Family Fund for ARVD Research; The Hugh Calkins, Marvin H. Weiner, and Jacqueline J. Bernstein Cardiac Arrhythmia Center; the Dr Francis P. Chiramonte Private Foundation; the Dr Satish, Rupal, and Robin Shah ARVD Fund at Johns Hopkins; the Bogle Foundation; the Campanella family; the Patrick J. Harrison Family; the Peter French Memorial Foundation; and the Wilmerding Endowments. The Netherlands ACM Registry is supported by the Netherlands Heart Institute (project 06901).

Disclosures

Dr James has received research support from Stride Bio Inc, Lexeo Therapeutics, and ARVADA Therapeutics. C. Tichnell has received salary support on these grants. Dr te Riele is a consultant for Tenaya, Rocket Pharmaceuticals, and BioMarin for unrelated work (gene therapy trials). Drs James and Gasperetti have served as compensated consultants for LEXEO Therapeutics for unrelated work. Dr Yap has received honoraria (speaker or consultancy fees) from Boston Scientific, Medtronic, Biotronik, and Acutus Medical. In addition, he has received research grants from Medtronic, Biotronik, and Boston Scientific. The other authors report no conflicts.

Supplemental Material

Methods

Tables S1–S4

Figures S1–S13