ORIGINAL ARTICLE

Performance of the 2015 International Task

Force Consensus Statement Risk Stratification
Algorithm for Implantable Cardioverter-
Defibrillator Placement in Arrhythmogenic
Right Ventricular Dysplasia/Cardiomyopathy

See Editorial by Indik Gabriela M. Orgeron, MD

Anneline te Riele, MD,
BACKGROUND: Ventricular arrhythmias are a feared complication of PhD
arrhythmogenic right ventricular dysplasia/cardiomyopathy. In 2015, Crystal Tichnell, MGC
an International Task Force Consensus Statement proposed a risk Weijia Wang, MD, MPH
Brittney Murray, MS
stratification algorithm for implantable cardioverter-defibrillator placement
Aditya Bhonsale, MD,
in arrhythmogenic right ventricular dysplasia/cardiomyopathy.
MHS
METHODS AND RESULTS: To evaluate performance of the algorithm, Daniel P. Judge, MD
365 arrhythmogenic right ventricular dysplasia/cardiomyopathy patients Ihab R. Kamel, MD, PhD
Stephan L. Zimmerman,
were classified as having a Class I, IIa, IIb, or III indication per the
MD
algorithm at baseline. Survival free from sustained ventricular arrhythmia
Harikrishna Tandri, MD
(VT/VF) in follow-up was the primary outcome. Incidence of ventricular Hugh Calkins, MD
Downloaded from http://ahajournals.org by on March 17, 2021

fibrillation/flutter cycle length <240 ms was also assessed. Two hundred Cynthia A. James, ScM,
twenty-four (61%) patients had a Class I implantable cardioverter- PhD
defibrillator indication; 80 (22%), Class IIa; 54 (15%), Class IIb; and 7
(2%), Class III. During a median 4.2 (interquartile range, 1.7–8.4)-year
follow-up, 190 (52%) patients had VT/VF and 60 (16%) had ventricular
fibrillation/flutter. Although the algorithm appropriately differentiated
risk of VT/VF, incidence of VT/VF was underestimated (observed versus
expected: 29.6 [95% confidence interval, 25.2–34.0] versus >10%/year
Class I; 15.5 [confidence interval 11.1–21.6] versus 1% to 10%/year Class
IIa). In addition, the algorithm did not differentiate survival free from
ventricular fibrillation/flutter between Class I and IIa patients (P=0.97) or
for VT/VF in Class I and IIa primary prevention patients (P=0.22). Adding
Holter results (<1000 premature ventricular contractions/24 hours) to
International Task Force Consensus classification differentiated risks. Correspondence to: Cynthia A.
James, ScM, PhD, Johns Hopkins
CONCLUSIONS: While the algorithm differentiates arrhythmic risk well ARVD/C Program, Carnegie 568D,
overall, it did not distinguish ventricular fibrillation/flutter risks of patients 600 N Wolfe St, Baltimore, MD
with Class I and IIa implantable cardioverter-defibrillator indications. 21287. E-mail: cjames7@jhmi.edu
Limited differentiation was seen for primary prevention cases. As these are Key Words: arrhythmia
vital uncertainties in clinical decision-making, refinements to the algorithm ◼ arrhythmogenic right ventricular
dysplasia/cardiomyopathy
are suggested prior to implementation. ◼ implantable cardioverter-
defibrillator ◼ ventricular
fibrillation ◼ ventricular
tachycardia
© 2018 American Heart
Association, Inc.

Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593 February 2018 1

 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
WHAT IS KNOWN?
• Ventricular tachycardia and sudden cardiac death
are common in arrhythmogenic right ventricular
dysplasia/cardiomyopathy, so following diagnosis a
critical decision is whether to implant an implant-
able cardioverter-defibrillator.
• A recently published international consensus state-
ment for the treatment of arrhythmogenic right
ventricular dysplasia/cardiomyopathy proposes
Downloaded from http://ahajournals.org by on March 17, 2021
a risk stratification algorithm with Class I, IIa, IIb,
and III indications for implantable cardioverter-
defibrillator placement based on expert opinion
and review of observational cohort studies.
WHAT THE STUDY ADDS?
• We assess the performance of the proposed algo-
rithm for the first time in a cohort of patients
and assess whether integration of Holter monitor
results (absent from the algorithm) improve differ-
entiation of arrhythmic risk.
• Although the algorithm differentiated risk of
sustained ventricular tachycardia well overall, it
underestimated incidence of ventricular tachy-
cardia and failed to distinguish risk of ventricular
fibrillation/flutter in Class I and IIa patients and had
limited differentiation of risk in primary prevention
patients.
• Adding premature ventricular complex count to
algorithm classification significantly improved dif-
ferentiation of risks.
A
rrhythmogenic right ventricular dysplasia/cardio-
myopathy (ARVD/C) is an inherited cardiomy-
opathy characterized by fibrofatty myocardial
replacement and intercalated disk remodeling.1 These
changes predispose patients to ventricular arrhythmias
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
and sudden cardiac death (SCD). Thus, once the di-
agnosis of ARVD/C is established, a critical decision is
whether to implant an implantable cardioverter-defibril-
lator (ICD). Several studies have validated the efficacy
of ICD therapy in patients with ARVD/C.2–4 However, a
long course with an ICD is associated with a substan-
tial risk of device-related and psychosocial complications
and inappropriate shocks.5,6 Accurate identification of
patients at sufficiently high risk of SCD to warrant ICD
placement is therefore critical not only to save lives but
also to protect low-risk patients. Unfortunately, the clini-
cal course of ARVD/C is highly variable, and arrhythmic
risk stratification can be difficult, particularly for patients
without a history of sustained ventricular arrhythmias.
To address this challenge, an International Task
Force Consensus (ITFC) statement for the treatment
of ARVD/C7 was recently published. These recom-
mendations included Class I, IIa, IIb and III indications
for ICD placement that were based on review of
observational cohort studies and estimation of yearly
incidence of appropriate ICD therapy. In this article,
we take advantage of a large cohort of ARVD/C
patients enrolled in the Johns Hopkins ARVD/C regis-
try to assess the performance of the proposed algo-
rithm. Specifically, we (1) assess the extent to which
the ITFC algorithm predicts sustained ventricular
arrhythmia (ventricular tachyarrhythmia/ventricular
tachycardia [VT/VF]) and also ventricular fibrillation/
flutter (VF/VFL; cycle length [CL] ≤240 ms), (2) assess
the algorithm’s performance in patients implanted
for primary prevention, and (3) explore opportuni-
ties to improve the algorithm, with particular atten-
tion to the results of 24-hour Holter monitoring, an
inexpensive readily available test previously shown3
to effectively predict sustained ventricular arrhyth-
mias in ARVD/C patients but not included in the ITFC
algorithm.
February 2018 2
 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
MATERIALS AND METHODS
Patient Population and Follow-Up
The Johns Hopkins ARVD/C Registry was established in 1999
to prospectively follow patients and at-risk family members.
For the current study, we included all patients enrolled in the
registry with (1) a definite diagnosis of ARVD/C by the 2010
Task Force Criteria8 and (2) a follow-up period of at least 30
days. The study was approved by the Johns Hopkins School
of Medicine Institutional Review Board. All participants pro-
vided written informed consent. To maintain patient confi-
dentiality, the data, analytic methods, and study materials
will not be made available to other researchers for purposes
of reproducing the results or replicating the procedure. A
limited data set may be made available on request.
Patient Classification by the ITFC
Algorithm
Each patient was classified by whether they met Class I,
IIa, IIb, or III indications per the ITFC algorithm7 at baseline.
To do so, the clinical history, demographics, family history,
and cardiac testing results at the time of ICD implantation
(in patients with an ICD) or diagnosis by 2010 criteria8 (in
patients without an ICD) were reviewed. Medical record
review included 12-lead ECG, exercise testing, and 24-hour
Holter monitoring when available. Echocardiographic and
magnetic resonance imaging studies were reviewed to
determine the severity and extent of right (RV) and left
ventricular (LV) dysfunction. Electrophysiological study was
deemed inducible if a sustained VT or VF that lasted >30
Downloaded from http://ahajournals.org by on March 17, 2021
seconds or required termination because of hemodynamic
compromise was induced. History of syncope, spontaneous
sustained VT and VF, and spontaneous nonsustained VT
(NSVT) was obtained for each patient. Syncope was defined
as a transient loss of consciousness and postural tone with
spontaneous recovery. NSVT was defined as 3 consecutive
ventricular premature beats with a rate >100 beats per min-
ute, lasting less than 30 seconds. Probands were affected
individuals ascertained independently of family history.
As in the ITFC algorithm,7 patients meeting Class I indi-
cations (ICD indicated) had history of sustained VT or VF
before ICD implantation, severe RV dysfunction (fractional
area change ≤17% or RV ejection fraction ≤35%), and severe
LV dysfunction (LV ejection fraction ≤35%). Individuals with
Class IIa indications (ICD should be considered) had ≥1 major
risk factors: syncope, NSVT, and moderate dysfunction of RV
(RV fractional area change between 24% and 17% or RV
ejection fraction 36% to 40%), LV (LV ejection fraction 36%
to 45%), or both ventricles. Class IIb indications (ICD may be
considered) included ≥1 minor risk factors.7 We report the
most prevalent in our population: male sex, proband, induc-
ibility at electrophysiological study, and ≥3 precordial T-wave
inversions on ECG. Patients with Class III indications (ICD not
indicated) had no ITFC algorithm risk factors.
Arrhythmia Outcomes
The primary study outcome was first sustained ventricular
arrhythmia (VT/VF) in follow-up. First VF/VFL was a sec-
ondary outcome. VT/VF was a composite measure of the
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
occurrence of spontaneous sustained VT, aborted SCD, or
appropriate ICD intervention for a ventricular arrhythmia. In
patients without an ICD, VT/VF outcome was adjudicated
based on reviewing ECGs and medical records; in patients
with an ICD, the device-stored ECGs were reviewed for
appropriateness of ICD therapy according to following
definitions.9 VF or VFL was defined as an irregular or regu-
lar tachycardia with a mean CL ≤240 ms. VT was defined
as a regular tachycardia arising from the ventricle with a
CL >240 and <600 ms. Decisions regarding ICD program-
ming were made by the managing cardiovascular special-
ist. When complete ICD interrogation information or ECG
tracings were not available, interpretation by the referring
electrophysiologist was used to classify arrhythmic events.
Statistical Analyses
Continuous variables are summarized as either mean±SD
or median (interquartile range) and categorical variables as
N (%). Incidence rates were calculated in person-years and
compared via 2-sided Fisher exact tests. The cumulative prob-
ability of survival free from first sustained ventricular arrhyth-
mia (VT/VF) and from first VF/VFL was determined by the
Kaplan–Meier method and differences in survival between
groups evaluated with the log-rank test. Log-rank tests were
also used for pairwise comparisons. ICD implantation was
the start of follow-up in survival analysis in patients with an
ICD. For patients without an ICD, diagnosis by 2010 criteria
was the start of follow-up. In patients without an outcome,
follow-up was to most recent evaluation, transplantation, or
date of death, whichever came first. Univariate Cox regres-
sion was used to assess the association between each risk
factor and survival from first VT/VF or first VF/VFL. Cox pro-
portional hazard models were used to assess the association
of premature ventricular contraction (PVC) count and survival
from first VT/VF or first VF/VFL independent of ITFC class.
The proportionality of hazards assumption was checked on
the basis of Schoenfeld residuals. All statistical analyses were
performed using STATA (version 14.2 College Station, TX). A
2-sided P value ≤0.05 was considered significant.
RESULTS
Study Sample
The study sample consisted of 365 patients with definite
ARVD/C. Table  1 summarizes their clinical and demo-
graphic characteristics. Half were male (181, 50%) and
the majority probands (264, 72%). Most (312; 85%) had
an ICD. Mean age at ICD implantation was 36.5±13.5
years. The majority (177, 55%) had their device implant-
ed for secondary prevention, with 158 (51%) experienc-
ing VT and 19 (6%) VF/VFL before implantation. Clinical
characteristics of the cohort without an ICD are sum-
marized in Table I in the Data Supplement.
Classification by ITFC Algorithm
The majority of patients (224, 61%) had a Class I ICD
indication per the ITFC algorithm, 80 (22%) had a Class
February 2018 3
 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance

Table 1.  Clinical Characteristics of the Study Sample Table 2.  Association of Survival Free From Sustained
Ventricular Tachycardia With ITFC Risk Factors
Clinical Variables Overall (N=365)
Demographics No. of
Indications Patients Patients
 Male sex 181 (50) for ICD With Risk With VT/ Hazard Ratio P
 Caucasians 348 (95) Implantation* Factor VF (95% CI) Value

 Age when follow up started, years* 36.6±14 Class I indication

Family history/genetics  Prior VT/VF 183 127 (69) 2.66 (1.97–3.61) <0.001
 Proband 264 (72)  Severe RV
67 37 (55) 1.68 (1.15–2.46) 0.008
dysfunction
 Mutation carrier 224 (63)
Clinical characteristics at baseline  Severe LV
12 8 (67) 1.48 (0.70–3.11) 0.303
dysfunction
 History of VT/VF 183(50)
Class IIa indication
 Syncope 99 (27)
 Syncope 99 61 (62) 1.53 (1.13–2.00) 0.006
 NSVT 125 (35)
 NSVT 125 68 (54) 1.24 (0.92–1.67) 0.164
 Inducibility at EPS 167/231 (72)
 Moderate RV
 Inverted T waves in ≥3 precordial leads 247/320 (77) 36 19 (53) 1.77(1.06–2.94) 0.029
dysfunction
RV dysfunction†
 Moderate LV
23 14 (61) 1.40 (0.84–2.34) 0.192
 Severe RV dysfunction 67/254 (26) dysfunction
 Moderate RV dysfunction 36/254(14) Class IIb indication
LV dysfunction‡  Male 181 117 (65) 1.87 (1.39–2.51) <0.001
 Severe LV dysfunction 12/308(4)  Proband 264 173 (66) 6.26 (3.86–10.16) <0.001
 Moderate LV dysfunction 23/308 (8)  Inducibility at
167 120 (72) 3.45 (2.19–5.41) <0.001
 PVCs on Holter monitor; median [IQR] 1903 [366–5211] EPS

 PVCs ≥1000/24h on Holter monitor 128/206 (62)  Inverted T waves

in ≥3 precordial 247 127 (51) 1.72 (1.13–2.61) 0.011
 ICD implanted 312 (85)
leads
Values n (%) or n/N (%). EF indicates ejection fraction; EPS, Other
electrophysiology study; FAC, fractional area change; ICD, implantable
Downloaded from http://ahajournals.org by on March 17, 2021

cardioverter-defibrillator; ITFC, International Task Force Consensus; LV, left
ventricular; NSVT, nonsustained ventricular tachycardia; PVC, premature
ventricular contraction; RV, right ventricular; VF, ventricular fibrillation; and VT,
ventricular tachycardia.
*Age at ICD implantation for patients with defibrillators or age when
meeting ITFC criteria for those without defibrillators.
†RV dysfunction per ITFC: severe, FAC ≤17% or RV EF ≤35%; moderate,
RV FAC 24% to 17% or RV EF between 36% and 40%.
‡LV dysfunction per ITFC: severe, LV EF ≤35%; moderate, RV FAC 7% to
24% or RV EF between 36% and 40%.
IIa indication, and 54 (15%) had a Class IIb indication.
Seven (2%) patients had a Class III indication, with
no ITFC risk factors at baseline, despite meeting TFC.8
Among those with a Class I indication, 183 (82%) had
a history of VT/VF, 67 (44%) severe had RV dysfunction,
and 12 (6%) severe had LV dysfunction per the ITFC algo-
rithm.10 The most common major risk factor in patients
with Class IIa indication was NSVT (61, 77%) followed by
syncope (34, 43%). The most common minor risk factors
in patients with Class IIb indications were ≥3 precordial
T-wave inversions (42/53, 79%), male sex (20, 37%),
inducibility on electrophysiological study (10/25, 40%),
and proband status (9, 17%). Table 2 shows the preva-
lence of each ITFC risk factor in the overall study sample.
Ventricular Arrhythmias During Follow-Up
During a median follow-up of 4.2 years (interquartile
range, 1.7–8.4), 190 (52%) patients experienced VT/
 PVCs ≥1000/24
h on Holter 128 65(51) 3.48 (2.00–6.03) <0.001
monitoring
 Mutation 224 119 (53) 0.99 (0.74–1.33) 0.962
EPS indicates electrophysiology study; ICD, implantable cardioverter-
defibrillator; ITFC, International Task Force Consensus guidelines; LV, left
ventricular; NSVT, nonsustained ventricular tachycardia; PVC, premature
ventricular contraction; RF, risk factor; RV, right ventricular; VF, ventricular
fibrillation; VFL, ventricular flutter; and VT, ventricular tachycardia.
*For each variable, the referent group is patients without the ITFC risk
factor.
VF (mean CL 278±60 ms). VF/VFL was observed in 60
(16%) patients (mean CL 220±15 ms). Table 2 shows
the association of each risk factor in the ITFC algorithm
with survival free from first VT/VF. Table II in the Data
Supplement shows association of risk factors with sur-
vival free from a rapid ventricular arrhythmia (VF/VFL).
Performance of the ITFC Algorithm
Figure 1 shows outcomes of study participants stratified
by ITFC algorithm class indication. Patients with a Class
I indication were most likely to meet the primary and
secondary study outcomes, with 148 (66%) patients
experiencing VT/VF and 43 (19%) experiencing VF/VFL
at last follow-up. As shown in Figure 1, the incidence
rate of VT/VF ranged from 29.6 per 100 person-years
(95% confidence interval [CI], 25.5–34.8) for patients

Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593 February 2018 4

 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance

Figure 1. Flow chart of risk

stratification, indications for ICD
implantation, number of events,
and annual incidence rate per ICD
class indication.
ARVD/C indicates arrhythmogenic
right ventricular dysplasia/cardio-
myopathy; EPS, electrophysiological
study; ICD, implantable cardioverter-
defibrillator; LV, left ventricle; NSVT,
nonsustained ventricular tachycardia;
SCD, sudden cardiac death; RV, right
ventricle; VF, ventricular fibrillation;
VFL, ventricular flutter; and VT, ven-
tricular tachycardia.

with a Class I indication, 15.5 per 100 person-years Figure 3 shows the performance of the ITFC algo-
(95% CI, 11.1–21.6) for Class IIa patients, and 2.4 rithm in this cohort. Forty-one (22%) primary pre-
per 100 person-years (95% CI, 1.1–5.0) for Class IIb vention patients were classified as having a Class I
Downloaded from http://ahajournals.org by on March 17, 2021

patients. The 7 patients with Class III indications had
no events. In contrast, mean incidence rate of VF/VFL
was highest among patients classified as IIa (4.2 per
100 person-years [95% CI, 2.5–7.0]). Incidence rates of
VT/VF and VF/VFL among patients with each ITFC class
indication are summarized in Table 3.
Figure 2 and Figure I in the Data Supplement show
survival from first VT/VF and VF/VFL, respectively, strati-
fied by ITFC class indication. The ITFC algorithm accu-
rately differentiated survival from any sustained VT/VF
among the 4 risk groups (P<0.001; Figure 2). Patients
ICD indication, 80 (44%) as Class IIa, 54 (30%) as
Class IIb, and 7 (4%) as Class III. As can be appreci-
ated, incremental risk of VT/VF was again observed
for Class III and IIb ICD indications. However, survival
free from first VT/VF (P=0.22; Figure  3) and annual
incidence rate of VT/VF (Table 3) were similar for pri-
Table 3.  Incidence Rates of VT/VF and VF/VFL
Stratified by ITFC Class Designation
ITFC All Primary Prevention
Classification (N=365) %/y (95% CI) (N=182) %/y (95% CI)

with a Class I indication had significantly worse survival Class I

from VT/VF than patients with a Class IIa (P=0.002) or  VT/VF 29.6 (25.2–34.8) 25.7 (16.7–39.4)
Class IIb (P<0.001) indication. In contrast (Figure I in the  VF/VFL 3.8 (2.8–5.1) 9.8 (5.6–17.2)
Data Supplement), the ITFC algorithm did not differen- Class IIa
tiate survival free from VF/VFL between patients with a
 VT/VF 15.5 (11.1–21.6) 15.5 (11.1–21.6)
Class I and IIa indication (P=0.97).
 VF/VFL 4.2 (2.5–7.8) 4.2 (2.5–7.0)
Class IIb
Performance of the ITFC Algorithm in  VT/VF 2.4 (1.1–5.0) 2.4 (1.1–5.0)
Primary Prevention Patients  VF/VFL 0.6 (0.2–2.6) 0.6 (0.2–2.6)
We subsequently focused our analyses on the popu- Class III
lation without prior history of sustained ventricular  VT/VF 0 0
arrhythmias (N=182) because the clinical question of  VF/VFL 0 0
whether to implant an ICD is most challenging in this
CI indicates confidence interval; ITFC, International Task Force Consensus
group. Annual incidence rate of VT/VF in these patients guidelines; VF, ventricular fibrillation; VFL, ventricular flutter; and VT,
was substantial at 10.0% (95% CI, 7.8–12.8) per year. ventricular tachycardia.

Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593 February 2018 5

 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
Figure 2. Cumulative survival free from VT/VF stratified
by International Task Force Consensus class indication.
P values are for pairwise comparison with Class I indica-
tion. VF indicates ventricular fibrillation; and VT, sustained
ventricular tachycardia
mary prevention cases with Class I (25.7%/year [95%
CI, 16.7–39.4]) and Class IIa (15.5%/year [95% CI,
Downloaded from http://ahajournals.org by on March 17, 2021
11.1–21.6]) indications.
Figure 3. Survival free from VT/VF in follow-up in
patients without history of sustained ventricular tachy-
cardia or ventricular fibrillation stratified by Interna-
tional Task Force Consensus class.
P values are for pairwise comparison with Class I indica-
tion. VF indicates ventricular fibrillation; and VT, sustained
ventricular tachycardia.
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
VF/VFL was also somewhat common with an inci-
dence rate of 3.5%/year (95% CI, 2.5–5.1). Of particular
importance, as shown in Table 3, the small portion of pri-
mary prevention cases with a Class I indication had a high
incidence of VF/VFL (9.8%/year [95% CI, 5.6–17.2]).
Refining the Consensus Guideline via
Inclusion of Holter Results
We next investigated opportunities to improve differ-
entiation of risk by the ITFC. We began by exploring
whether regrouping risk factors based on hazard ratios
calculated in univariate analyses could lead to better dif-
ferentiation. As shown in the analyses presented in the
Data Supplement, this approach led to no improvement.
We next assessed whether Holter monitor results
could be used to refine risk stratification. Median PVC
count of our population was 1903 per 24 hours (inter-
quartile range, 366–5211), with 62% (128/206) hav-
ing >1000 PVCs in 24 hours. Patients with a high PVC
burden (≥1000 PVCs/24 hours) had poorer survival from
both VT/VF (hazard ratio, 3.48; 95% CI, 2.00–6.03;
P<0.001; Table 2) and VF/VFL (hazard ratio, 6.69; 95% CI,
2.03–22.12.6; P=0.002; Table II in the Data Supplement).
As PVC count is not part of the ITFC algorithm,
we explored whether integrating the results of Holter
monitoring with the ITFC algorithm could improve cali-
bration of arrhythmic risks. First, we assessed whether
Holter monitor results are an independent predictor to
class status for survival from VT/VF and VF/VFL. Tables 4
and 5 presents Cox proportional hazard models includ-
ing PVC count and ITFC class status as predictors of
survival free from VT/VF and VF/VFL, respectively. As
shown, for both outcomes having >1000 PVCs/day is
an independent predictor of arrhythmias. For rapid,
likely life-threatening events (Table  5), Holter monitor
result remained the only significant predictor. Frequent
PVCs are also independent predictors of VT/VF (hazard
ratio, 2.94; 95% CI, 1.28–6.77; P=0.01) in the primary
prevention cohort (Table III in the Data Supplement).
Next we assessed whether there was practical pre-
dictive utility of adding Holter monitor result to class
Table 4.  Prognostic Value of PVCs on Holter Monitor
for VT/VF in ARVD/C
Hazard Ratio (95%
Number of Patients = 206 Confidence Interval) P Value
PVCs≥1000/24 h on Holter 2.21 (1.24–3.93) 0.007
 Class IIa* 0.7 (0.42–1.16) 0.163
 Class IIb* 0.23(0.09–0.60) 0.003
 Class III* 0 (0) NA
ARVD/C indicates arrhythmogenic right ventricular dysplasia/
cardiomyopathy; ICD, implantable cardioverter-defibrillator; NA, not
applicable; PVC, premature ventricular complex; VF, ventricular fibrillation; and
VT, ventricular tachycardia.
*Class I was used as the reference. There was no event in Class III.
February 2018 6
 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
status. As shown in Figure 4 and presented in detail in
Table IV in the Data Supplement, for each ITFC class
indication, incidence rate of VT/VF was lower among
the subgroup with fewer PVCs. In the Class IIa group,
this difference was significant (P<0.004), with Class IIa
patients with >1000 PVCs having risks indistinguish-
able from Class I patients and those with <1000 PVCs
similar to Class IIb risks. Regardless of class indication,
incidence of VF/VFL was low among patients with
<1000 PVCs (Figure II and Table IV in the Data Supple-
ment).
DISCUSSION
Main Findings
This study was designed to retrospectively assess the
performance of the proposed ITFC risk stratification
algorithm7 in predicting sustained ventricular arrhyth-
mias in definite ARVD/C patients. It has 4 main find-
ings. First, the ITFC algorithm accurately differentiated
survival from any sustained VT/VF among 4 proposed
ICD class indications. The incidence rate was somewhat
higher than estimated in the guideline for those with
Class I and Class IIa indications: observed versus expect-
ed incidence 30% versus >10%/year for Class I and
15% versus 1% to 10%/year for Class IIa. The observed
incidence for those with Class IIb and III ICD indications
was as predicted (observed versus expected 2.4% ver-
Downloaded from http://ahajournals.org by on March 17, 2021
sus 1% to 10% for Class IIb and 0% versus <1% for
Class III). Second, the ITFC algorithm did not differen-
tiate survival from VF/VFL well, particularly between
patients with Class I and IIa indications. Third, when
we focused our analyses on the population without a
previous history of sustained ventricular arrhythmias,
the ITFC again performed well, although there was lim-
ited differentiation of VT/VF risk between patients with
Class I and IIa indications. Finally, a high PVC burden
(≥1000 PVCs/24 hours) was an independent predictor
of arrhythmic events and may be particularly useful to
further differentiate risks of rapid ventricular arrhyth-
mias within ITFC risk groups.
Table 5.  Prognostic Value of PVCs on Holter Monitor
for VF/VFL in ARVD/C
Hazard Ratio (95%
Number of Patients = 206 Confidence Interval)
PVCs≥1000/24 h on Holter 5.24 (1.51–18.18)
 Class IIa* 1.05 (0.48–2.27)
 Class IIb* 0.52 (0.11–2.38)
 Class III* 0 (0)
ARVD/C indicates arrhythmogenic right ventricular dysplasia/
cardiomyopathy; ICD, implantable cardioverter-defibrillator; NA, not
applicable; PVC, premature ventricular complex; VF, ventricular fibrillation; and
VFL, ventricular flutter.
*Class I was used as the reference. There was no event in Class III.
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
Figure 4. Incidence rates of VT/VF in arrhythmogenic
right ventricular cardiomyopathy/dysplasia patients by
International Task Force Consensus classification and
PVC counts on Holter.
Two-sided P values for difference in incidence rates within
each class by PVC >1000/day or not were reported. Line
lengths represent 95% confidence intervals. PVC indicates
premature ventricular complex; VF, ventricular fibrillation;
and VT, sustained ventricular tachycardia.
Prior Studies
ARVD/C is characterized by frequent ventricular arrhyth-
mias and a high risk of SCD. Thus, a key element of
treatment is reducing the risk of SCD by properly risk-
stratifying patients who would benefit from an ICD. Over
the past 2 decades, numerous efforts have been made
to improve arrhythmic risk stratification. Key studies
described the clinical course of ARVD/C, established the
efficacy of ICDs, and explored clinical risk factors.2–4,10,11
While laying critical groundwork, most of these studies
had important limitations including small sample size,2–
4,10,12
use of the 1994 Task Force Criteria,2,4,11 and inclu-
sion of patients not meeting diagnostic criteria.3
The quest to define factors associated with life-threat-
ening arrhythmias accelerated when the discovery of
genes associated with ARVD/C13 allowed at-risk family
members to be identified. The opportunity to prevent SCD
in these genetically at-risk relatives spurred development
and evaluation of both traditional and novel noninvasive
methods to refine assessment of arrhythmic risk.10–12,14
The ITFC algorithm7 is an important attempt to inte-
grate this literature into a risk stratification algorithm
P Value for ICD placement in ARVD/C patients. This was a dif-
0.009 ficult task as the authors were required to take into
0.908
account results of heterogenous studies, generally with
small sample sizes, a variety of predictor variables (vari-
0.398
ably defined), and different clinical outcomes (eg, any
NA
ICD therapy, VF/VFL). Here, for the first time, we test
the ITFC algorithm in a group of patients—providing
an opportunity to assess its overall performance, define
possible weaknesses, and consider opportunities for
improvement.
February 2018 7
 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
Overall ITFC Algorithm Performance
The ITFC7 algorithm groups patients into Class I, IIa, IIb,
and III indications for ICD implantation based on pre-
dicted incidence of ICD therapy combined with “general
health, socioeconomic factors, the psychological impact,
and the adverse effects of the device.” On the whole,
this approach seems to have been effective. Incidence
rate of VT/VF was highest in patients with a Class I indi-
cation, followed by patients with Class IIa, IIb, and III indi-
cations, respectively. It is worth noting that our patients
with Class I and IIa ICD indications had a somewhat
higher than predicted incidence rate of VT/VF than esti-
mated in the ITFC algorithm. It is uncertain whether this
represents a higher risk of events in these groups in the
overall population of ARVD/C patients or is specific to
our population in which 85% had ICDs. The algorithm
also performed well for stratifying risk of primary preven-
tion patients, although in this limited analysis Class I and
IIa risks for VT/VF were not well differentiated.
Limitations of the ITFC Algorithm
While the ITFC algorithm differentiated incidence of
future ventricular arrhythmias reasonably well overall,
our study uncovered an important weakness. It did not
distinguish incidence of VF/VFL between those implant-
ed for a Class I and IIa indication. Given the high likeli-
hood of VF/VFL being fatal in the absence of an ICD, it
Downloaded from http://ahajournals.org by on March 17, 2021
is concerning that the algorithm does not differentiate
these risks. This limitation is most likely to affect young
ARVD/C patients as they disproportionately experience
VF/VFL early in their disease course.15,16
It is possible that the algorithm had weaknesses
in differentiating incidence of VF/VFL in part because
relatively few prior studies have reported the incidence
and risk factors specifically for developing VF/VFL. This
limited literature impacts both the amount and quality
of data used to develop the ITFC algorithm and also
leaves physicians limited evidence for which to guide
their decisions beyond the algorithm. Studies that have
addressed this topic include work by Corrado et al2,4
who described syncope as a predictor for VF/VFL as well
as younger age and a prior history of sustained ventric-
ular arrhythmias. Link et al11 and Mazzanti et al12 agree
that younger patients are at increased risk of develop-
ing life-threatening arrhythmias; however, age was not
included in the ITFC algorithm.
The ITFC algorithm was also limited in distinguish-
ing between outcomes of patients implanted for pri-
mary prevention with a Class I versus IIa indication. The
implantation of ICDs for primary prevention in ARVD/C
patients continues to be one of the most challenging
questions in clinical practice. Unfortunately, data are
limited. The largest study of ARVD/C patients with pri-
mary prevention ICDs (N=106) was by Corrado et al4
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
who identified syncope as a significant predictor for ICD
therapy for both VT and VF/VFL. Bhonsale et al3 found
NSVT and inducibility on electrophysiological study to
be independent predictors of appropriate ICD therapy.
Refining the Consensus Guideline via
Inclusion of Holter Results
In our analysis we show that by adding Holter monitor-
ing results to the ITFC risk classification, we were able
to further distinguish arrhythmic risks. We found that
Holter monitor result was an independent predictor to
ITFC Class status in predicting survival from an arrhyth-
mia in follow-up. For predicting VF/VFL, Holter result
remained the only significant predictor. Holter result
remained a significant predictor when only primary
prevention cases were entered into regression analysis,
highlighting the utility of integrating these results into
risk stratification of this difficult population. We also
found that within risk groups, incidence rates for VF/
VFL are low and indistinguishable among patients with
Class I, IIa, and IIb indications who have <1000 PVCs,
while rates are high in Class I and IIa patients with fre-
quent PVCs. Therefore, our results suggest that strati-
fying patients by Holter result after ITFC classification
helps further differentiate risks particularly in address-
ing situations imperfectly addressed by the ITFC: (1) Dif-
ferentiating risk specifically of rapid, likely life-threat-
ening arrhythmias and (2) differentiating risk among
primary prevention cases. This finding supports the role
of Holter monitoring as a widely available and inexpen-
sive, noninvasive way to obtain valuable information for
the risk stratification of patients with ARVD/C.
Study Limitations
This study has several limitations. First, while the ITFC
risk stratification algorithm was developed by consensus
of clinical experts, it is likely prior publications describ-
ing risk factors and outcomes of Johns Hopkins ARVD/C
registry patients influenced its design. Assessment of its
performance in other ARVD/C cohorts is, thus, particu-
larly important. Second, in this retrospective analysis,
we include individuals with devices implanted many
years ago. While this enhances follow-up duration, it
also includes patients with historical device program-
ming and, thus, possibly a higher rate of therapy for
VT, but not VF/VFL, than a contemporary sample might
experience. Finally, while our data strongly suggest
integrating the results of Holter monitoring as a second
step of risk stratification after classifying patients by the
ITFC algorithm, not all our patients had Holter monitor
results at study entry available, which may limit gener-
alizability of our findings. Finally, while our sample size
is large relative to other published ARVD/C populations,
we had limited power in stratified analyses.
February 2018 8
 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
Conclusions/Clinical Implications
This study assessed the performance of the risk stratifi-
cation algorithm for ICD placement in ARVD/C patients
published as part of the recent consensus statement for
ARVD/C treatment. We found that while the algorithm
performed well overall, there were weaknesses in distin-
guishing incidence of VF/VFL. Adding the results of Holt-
er monitoring improved differentiation of risks. While
these results are promising, they should not be viewed
as a definitive answer for utilization and improvement of
risk stratification using the algorithm. Validation of our
findings in other ARVD/C cohorts would be valuable.
More critically, given the limited evidence base on which
the ITFC algorithm was built—particularly with regards
to primary prevention cases and VF/VFL events—the
time may be right for development and validation of a
clinical risk prediction model for SCD that can be used
to generate individualized risk estimates for SCD as has
recently been done for hypertrophic cardiomyopathy
patients.17 Such an effort would require a multi-center,
statistically rigorous approach measuring the perfor-
mance of prespecified predictors. Although undeniably
a significant undertaking, it would have the potential to
at last provide a rigorous, evidence base for improving
the targeting of ICD therapy in ARVD/C patients.
ACKNOWLEDGMENTS
Downloaded from http://ahajournals.org by on March 17, 2021
We are grateful to the arrhythmogenic right ventricular dys-
plasia/cardiomyopathy (ARVD/C) patients and families who
make this work possible.
SOURCES OF FUNDING
We acknowledge funding from the Dr Francis P. Chiaramonte
Private Foundation, the St Jude Medical Foundation, and Bos-
ton Scientific Corp. The Johns Hopkins ARVD/C Program is
supported by the Leyla Erkan Family Fund for ARVD Research,
the Dr Satish, Rupal, and Robin Shah ARVD Fund at Johns
Hopkins, the Bogle Foundation, the Healing Hearts Founda-
tion, the Campanella family, the Patrick J. Harrison Family,
the Peter French Memorial Foundation, and the Wilmerding
Endowments. Dr Te Riele is supported by the Dutch Heart
Foundation (grant 2015T058), the University Medical Center
Utrecht Clinical Research Talent Fellowship, and the Cardio-
vascular Research the Netherlands (CVON) PREDICT Young
Talent Program. Dr Calkins received support from the Leducq
foundation—RHYTHM Network. Dr James was supported by
Netherlands Organization for Scientific Research (NWO, visi-
tor’s travel grant).
DISCLOSURES
Dr Calkins is a consultant for Medtronic Inc and St Jude Medi-
cal. Dr Calkins receives research support from the St Jude
Medical Foundation and Boston Scientific Corp. C. Tichnell
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
and Dr James receive salary support from these grants. Dr
Tandri receives research support from Abbott. The authors
have no conflicts of interest.
AFFILIATIONS
From the Department of Medicine, Division of Cardiology
(G.M.O., C.T., W.W., B.M., A.B., D.P.J., H.T., H.C., C.A.J.) and
Department of Radiology (I.R.K., S.L.Z.), Johns Hopkins Hos-
pital, Baltimore, MD; and Department of Heart and Lungs,
Division of Cardiology, University Medical Center Utrecht, and
Netherlands Heart Institute (A.t.R.).
FOOTNOTES
Received June 30, 2017; accepted December 15, 2017.
The Data Supplement is available at http://circep.ahajour-
nals.org/lookup/suppl/doi:10.1161/CIRCEP.117.005593/-/
DC1.
Circ Arrhythm Electrophysiol is available at http://circep.
ahajournals.org.
REFERENCES
1. Delmar M. ARVC/D and the dyad: a long distance relationship? Heart
Rhythm. 2016;13:742. doi: 10.1016/j.hrthm.2015.11.036.
2. Corrado D, Leoni L, Link MS, Della Bella P, Gaita F, Curnis A, Salerno JU,
Igidbashian D, Raviele A, Disertori M, Zanotto G, Verlato R, Vergara G,
Delise P, Turrini P, Basso C, Naccarella F, Maddalena F, Estes NA III, Buja
G, Thiene G. Implantable cardioverter-defibrillator therapy for prevention
of sudden death in patients with arrhythmogenic right ventricular cardio-
myopathy/dysplasia. Circulation. 2003;108:3084–3091. doi: 10.1161/01.
CIR.0000103130.33451.D2.
3. Bhonsale A, James CA, Tichnell C, Murray B, Gagarin D, Philips B, Dalal
D, Tedford R, Russell SD, Abraham T, Tandri H, Judge DP, Calkins H. Inci-
dence and predictors of implantable cardioverter-defibrillator therapy in
patients with arrhythmogenic right ventricular dysplasia/cardiomyopathy
undergoing implantable cardioverter-defibrillator implantation for pri-
mary prevention. J Am Coll Cardiol. 2011;58:1485–1496. doi: 10.1016/j.
jacc.2011.06.043.
4. Corrado D, Calkins H, Link MS, Leoni L, Favale S, Bevilacqua M, Basso C,
Ward D, Boriani G, Ricci R, Piccini JP, Dalal D, Santini M, Buja G, Iliceto S,
Estes NA III, Wichter T, McKenna WJ, Thiene G, Marcus FI. Prophylactic
implantable defibrillator in patients with arrhythmogenic right ventricular
cardiomyopathy/dysplasia and no prior ventricular fibrillation or sustained
ventricular tachycardia. Circulation. 2010;122:1144–1152. doi: 10.1161/
CIRCULATIONAHA.109.913871.
5. Olde Nordkamp LR, Postema PG, Knops RE, van Dijk N, Limpens J, Wilde
AA, de Groot JR. Implantable cardioverter-defibrillator harm in young
patients with inherited arrhythmia syndromes: a systematic review and
meta-analysis of inappropriate shocks and complications. Heart Rhythm.
2016;13:443–454. doi: 10.1016/j.hrthm.2015.09.010.
6. James CA, Tichnell C, Murray B, Daly A, Sears SF, Calkins H. General and
disease-specific psychosocial adjustment in patients with arrhythmogenic
right ventricular dysplasia/cardiomyopathy with implantable cardioverter
defibrillators: a large cohort study. Circ Cardiovasc Genet. 2012;5:18–24.
doi: 10.1161/CIRCGENETICS.111.960898.
7. Corrado D, Wichter T, Link MS, Hauer R, Marchlinski F, Anastasakis A, Bauce
B, Basso C, Brunckhorst C, Tsatsopoulou A, Tandri H, Paul M, Schmied C,
Pelliccia A, Duru F, Protonotarios N, Estes NA III, McKenna WJ, Thiene G,
Marcus FI, Calkins H. Treatment of arrhythmogenic right ventricular cardio-
myopathy/dysplasia: an international task force consensus statement. Eur
Heart J. 2015;36:3227–3237. doi: 10.1093/eurheartj/ehv162.
8. Marcus FI, McKenna WJ, Sherrill D, Basso C, Bauce B, Bluemke DA, Calkins
H, Corrado D, Cox MG, Daubert JP, Fontaine G, Gear K, Hauer R, Nava A,
Picard MH, Protonotarios N, Saffitz JE, Sanborn DM, Steinberg JS, Tandri
H, Thiene G, Towbin JA, Tsatsopoulou A, Wichter T, Zareba W. Diagnosis
February 2018 9
 Orgeron et al; ARVD/C Risk Stratification Algorithm Performance
of arrhythmogenic right ventricular cardiomyopathy/dysplasia: proposed
modification of the Task Force Criteria. Eur Heart J. 2010;31:806–814.
doi: 10.1093/eurheartj/ehq025.
9. Maron BJ, Shen WK, Link MS, Epstein AE, Almquist AK, Daubert JP, Bardy
GH, Favale S, Rea RF, Boriani G, Estes NA III, Spirito P. Efficacy of im-
plantable cardioverter-defibrillators for the prevention of sudden death in
patients with hypertrophic cardiomyopathy. N Engl J Med. 2000;342:365–
373. doi: 10.1056/NEJM200002103420601.
10. Schuler PK, Haegeli LM, Saguner AM, Wolber T, Tanner FC, Jenni R, Corti
N, Lüscher TF, Brunckhorst C, Duru F. Predictors of appropriate ICD thera-
py in patients with arrhythmogenic right ventricular cardiomyopathy: long
term experience of a tertiary care center. PLoS One. 2012;7:e39584. doi:
10.1371/journal.pone.0039584.
11. Link MS, Laidlaw D, Polonsky B, Zareba W, McNitt S, Gear K, Marcus F, Es-
tes NA III. Ventricular arrhythmias in the North American multidisciplinary
study of ARVC: predictors, characteristics, and treatment. J Am Coll Car-
diol. 2014;64:119–125. doi: 10.1016/j.jacc.2014.04.035.
12. Mazzanti A, Ng K, Faragli A, Maragna R, Chiodaroli E, Orphanou N, Mon-
teforte N, Memmi M, Gambelli P, Novelli V, Bloise R, Catalano O, Moro
G, Tibollo V, Morini M, Bellazzi R, Napolitano C, Bagnardi V, Priori SG.
Arrhythmogenic right ventricular cardiomyopathy: clinical course and pre-
dictors of arrhythmic risk. J Am Coll Cardiol. 2016;68:2540–2550. doi:
10.1016/j.jacc.2016.09.951.
13. Groeneweg JA, Bhonsale A, James CA, te Riele AS, Dooijes D, Tichnell
C, Murray B, Wiesfeld AC, Sawant AC, Kassamali B, Atsma DE, Volders
PG, de Groot NM, de Boer K, Zimmerman SL, Kamel IR, van der Heijden
JF, Russell SD, Jan Cramer M, Tedford RJ, Doevendans PA, van Veen TA,
Tandri H, Wilde AA, Judge DP, van Tintelen JP, Hauer RN, Calkins H. Clini-
Downloaded from http://ahajournals.org by on March 17, 2021
Circ Arrhythm Electrophysiol. 2018;11:e005593. DOI: 10.1161/CIRCEP.117.005593
cal presentation, long-term follow-up, and outcomes of 1001 arrhyth-
mogenic right ventricular dysplasia/cardiomyopathy patients and family
members. Circ Cardiovasc Genet. 2015;8:437–446. doi: 10.1161/CIRC-
GENETICS.114.001003.
14. Saguner AM, Vecchiati A, Baldinger SH, Rüeger S, Medeiros-Domingo A,
Mueller-Burri AS, Haegeli LM, Biaggi P, Manka R, Lüscher TF, Fontaine G,
Delacrétaz E, Jenni R, Held L, Brunckhorst C, Duru F, Tanner FC. Different
prognostic value of functional right ventricular parameters in arrhythmo-
genic right ventricular cardiomyopathy/dysplasia. Circ Cardiovasc Imaging.
2014;7:230–239. doi: 10.1161/CIRCIMAGING.113.000210.
15. Bhonsale A, Groeneweg JA, James CA, Dooijes D, Tichnell C, Jongbloed
JD, Murray B, te Riele AS, van den Berg MP, Bikker H, Atsma DE, de Groot
NM, Houweling AC, van der Heijden JF, Russell SD, Doevendans PA, van
Veen TA, Tandri H, Wilde AA, Judge DP, van Tintelen JP, Calkins H, Hauer
RN. Impact of genotype on clinical course in arrhythmogenic right ven-
tricular dysplasia/cardiomyopathy-associated mutation carriers. Eur Heart
J. 2015;36:847–855. doi: 10.1093/eurheartj/ehu509.
16. te Riele ASJM, James CA, Sawant AC, Bhonsale A, Groeneweg JA, Mast
TP, Murray B, Tichnell C, Dooijes D, van Tintelen JP, Judge DP, van der Hei-
jden JF, Crosson J, Hauer RNW, Calkins H, Tandri H. Arrhythmogenic right
ventricular dysplasia/cardiomyopathy in the pediatric population: clinical
characterization and comparison with adult-onset disease. JACC Clin Elec-
trophysiol. 2015;1:551–560.
17. O’Mahony C, Jichi F, Pavlou M, Monserrat L, Anastasakis A, Rapezzi C, Bi-
agini E, Gimeno JR, Limongelli G, McKenna WJ, Omar RZ, Elliott PM; Hyper-
trophic Cardiomyopathy Outcomes Investigators. A novel clinical risk predic-
tion model for sudden cardiac death in hypertrophic cardiomyopathy (HCM
risk-SCD). Eur Heart J. 2014;35:2010–2020. doi: 10.1093/eurheartj/eht439.
February 2018 10