Elsevier

Heart Rhythm

Volume 13, Issue 1, January 2016, Pages 208-216
Heart Rhythm

High interobserver variability in the assessment of epsilon waves: Implications for diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia

https://doi.org/10.1016/j.hrthm.2015.08.031Get rights and content

Background

Revision of the Task Force diagnostic criteria for arrhythmogenic right ventricular cardiomyopathy/dysplasia (ARVC/D) has increased their sensitivity for the diagnosis of early and familial forms of the disease. The epsilon wave is a major diagnostic criterion in the context of ARVC/D, which, however, remains not quantifiable and therefore may leave room for substantial subjective interpretation.

Objective

The purpose of this study was to assess interobserver agreement in epsilon wave definition and epsilon wave importance for ARVC/D diagnosis.

Methods

Electrocardiographic (ECG) tracings depicting leads V1, V2, and V3 collected from individuals evaluated for ARVC/D (n = 30) were given to panel members who were asked to respond to the question whether ECG patterns meet epsilon wave definition outlined by the Task Force diagnostic criteria. The prevalence and importance of epsilon waves for ARVC/D diagnosis were assessed in a pooled data set of patients with definite ARVC/D from European and American registries (n = 815).

Results

The number of ECG patterns identified as epsilon waves varied from 5 to 18 per reviewer (median 13 per reviewer). A unanimous agreement was reached for only 10 cases (33%), 2 of which qualified as epsilon waves and 8 as non–epsilon waves by all panel members. From a pooled data set, 106 patients reportedly had epsilon waves (13%). In 105 of 106 patients with epsilon waves (99%), exclusion of epsilon waves from the diagnostic score would not affect the “definite” diagnostic category.

Conclusion

Interobserver variability in the assessment of epsilon waves is high; however, the impact of epsilon waves on ARVC/D diagnosis is negligibly low. The results urge to exercise caution in the assessment of epsilon waves, especially in patients who would not otherwise meet diagnostic criteria.

Introduction

The introduction of the revised diagnostic criteria for arrhythmogenic right ventricular cardiomyopathy/dysplasia (ARVC/D) in 20101 has reflected the molecular genetic and clinical progress made after the original criteria were introduced in 1994.2 In addition, to minimize subjectivity in imaging and tissue characterization, quantification of right ventricular imaging findings and morphometric analysis of histology preparations have improved the strictness of some of the major and minor diagnostic criteria that compose the 2010 revised Task Force criteria (TFC). These improvements increase sensitivity for the diagnosis of early and familial forms of the disease.

Characteristic abnormalities of ventricular depolarization and repolarization are important components of the diagnostic criteria for ARVC/D. The establishment of the diagnosis of ARVC/D requires 2 major criteria, 1 major and 2 minor criteria, or 4 minor criteria. Because of this, for example, a patient undergoing downstream testing for ARVC/D after the diagnosis of ARVC/D in a first-degree relative of a patient with ARVC/D (a major criterion by family history) will meet diagnostic criteria for ARVC/D if 1 additional major criterion is observed on the subsequent diagnostic test. One of these major criteria is the presence of an epsilon wave. It is therefore important to be sure that epsilon waves, as assessed on a 12-lead electrocardiogram (ECG), can be diagnosed accurately. However, it is notable that the epsilon wave remains one of the few 2010 TFC that is not quantifiable and therefore may leave room for subjective interpretation.

Using the 2010 TFC, an epsilon wave is defined as a “reproducible low-amplitude signal between end of QRS complex to onset of the T wave in the right precordial leads (V1 to V3).”1 This definition evolved from the one originally described in 1977 by Fontaine’s observation of tiny signals that consistently occurred after the end of each QRS complex on the surface electrocardiogram3 or as “a slur at the end of right precordial QRS complexes.”4 Epsilon waves have also been described as low-amplitude electrical potentials that occur “at the end”5 of or “immediately after”5, 6, 7, 8, 9 the QRS complex in the right precordial leads, “at the beginning of the ST segment,”10, 11 as “notches buried in the end of the QRS complex,”12 “constant or inconstant small afterdepolarizations in the transition of right precordial QRS complex and ST segment,”13 or “terminal deflection within or at the end of the QRS complexes.”14 Understandably, a number of publications do not provide any study-specific definition of an epsilon wave, but rather refer to the Task Force documents.15, 16 Nevertheless, available publications suggest that there is a variation in epsilon wave definition with regard to the appearance and exact location of the epsilon wave.

An increasing appreciation of the importance of ARVC/D as an important sudden death risk factor has led to an explosive growth of ARVC/D-related research worldwide and highlighted the necessity of assessment of whether epsilon waves are defined uniformly across different cohorts. The aim of the present epsilon wave initiative by the International Society for Holter and Noninvasive Electrophysiology undertaken by the writing group was to assess (1) interobserver agreement in the identification of epsilon waves and (2) the importance of epsilon waves for ARVC/D diagnosis in the national and international ARVC/D registries represented by the writing group members.

Section snippets

Construction of data set for ECG readings

Thirty ECG tracings were collected from patients with ARVC/D and family members screened for ARVC/D in the Nordic ARVC Registry17 (n = 21) and high-resolution ECG examples (n = 9) available from the literature.8, 14, 15, 18 Seven panel members (H.C., W.Z., D.C., T.W., R.N.H., J.H.S., E.K.B.) were asked to express their opinion on each of the ECG examples and answer “Yes” or “No” to the question whether the tracing would be considered as an epsilon wave on the basis of the 2010 TFC. Only QRST

Consensus (all panel members in agreement)

Of the 30 ECG examples, a consensus between panel members was reached for 10 cases, 2 of which were unanimously identified as epsilon waves and 8 as non–epsilon waves. The number of ECG patterns identified as epsilon waves by individual reviewers varied from 5 to 18 per reviewer (median 13 per reviewer). The results of the ECG assessment are graphically illustrated in Figure 1.

Agreement (≤6 panel members in agreement)

By applying less strict criteria for the assessment of interobserver agreement (≥6 panel members in agreement on ECG

Main findings

To our knowledge, this is the first study to report interobserver variability in the interpretation of epsilon wave definition performed by principal investigators of large ARVC/D registries in Europe and North America. Our main finding is the surprisingly low agreement in the interpretation of epsilon waves on the basis of the review of ECG patterns depicting QRS complex abnormalities in leads V1, V2, and V3, which had not improved after case-by-case ECG review and an attempt to refine the

Conclusion

Interobserver variability in the assessment of epsilon waves performed by principal investigators of several large international ARVC/D registries is high. However, the relative impact of the epsilon wave on ARVC/D diagnosis in patients from the participating registries is negligibly low, as patients are unlikely to express an epsilon wave as an isolated finding and in the vast majority of patients it is accompanied by other clinical manifestations that are sufficient for a definite ARVC/D

References (23)

  • K. Nasir et al.

    Electrocardiographic features of arrhythmogenic right ventricular dysplasia/cardiomyopathy according to disease severity: a need to broaden diagnostic criteria

    Circulation

    (2004)
  • Cited by (74)

    View all citing articles on Scopus

    The study was supported by the Swedish Heart-Lung Foundation grant number #20110875. Dr. te Riele participated in this work during tenure as the Mark Josephson and Hein Wellens Research Fellow of the Heart Rhythm Society.

    View full text