|Home | About | Journals | Submit | Contact Us | Français|
We analyzed the shape and distribution of epsilon waves by 3 various methods of electrocardiographic recording in patients with arrhythmogenic right ventricular cardiomyopathy.
Thirty-two patients who met recognized diagnostic criteria for arrhythmogenic right ventricular cardiomyopathy were included in this study (24 men and 8 women; mean age, 42.3 ± 12.9 yr). Epsilon waves were detected by standard 12-lead electrocardiography (S-ECG), right-sided precordial lead electrocardiography (R-ECG), and Fontaine bipolar precordial lead electrocardiography (F-ECG). We found 3 types of epsilon waves: wiggle waves, small spike waves, and smooth potential waves that formed an atypical prolonged R' wave. The most common configuration was small spiked waves. In some circumstances, epsilon waves were evident in some leads (especially in leads V1 through V3), but notches were recorded in the other leads during the corresponding phase. These waves could be detected only by S-ECG in 1 patient, R-ECG in 3 patients, and F-ECG in 5 patients; the rates of epsilon-wave detection by these 3 methods were 38% (12/32), 38% (12/32), and 50% (16/32), respectively. However, the detection rate using combined methods was significantly higher than that by S-ECG alone (SF-ECG 56% vs S-ECG 38%, P = 0.0312; and SRF-ECG 66% vs S-ECG 38%, P = 0.0039). In addition, the rate of widespread T-wave inversion (exceeding V3) was significantly higher in patients with epsilon waves than in those without (48% vs 9%, P = 0.029), as was ventricular tachycardia (95% vs 64%, P = 0.019).
These 3 electrocardiographic recording methods should be used in combination to improve the detection rate of epsilon waves.
Arrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited heart-muscle disease that predominantly affects the right ventricle (RV), especially the triangle of dysplasia. It is characterized pathologically by RV myocardial atrophy with fibrofatty replacement and clinically by ventricular electric instability with ventricular tachycardia (VT) or ventricular fibrillation that may lead to sudden death, primarily in young people and athletes.1,2 Epsilon waves are an important diagnostic clue for ARVC. Peters and colleagues3 found the prevalence of epsilon waves to be 23% in standard 12-lead electrocardiography (S-ECG) and 75% in Fontaine bipolar precordial lead ECG (F-ECG). In the present study, we applied jointly the S-ECG, right-sided precordial lead ECG (R-ECG), and F-ECG to analyze preliminarily the shape and distribution of epsilon waves, and to compare the impact of the various methods of ECG recording on the detection rate of epsilon waves.
Study Population. From 1 January 2000 through 15 March 2008, 32 consecutive patients (mean age, 42.3 ± 12.9 yr; 24 men) who met International Society & Federation of Cardiology/European Society of Cardiology (ISFC/ESC) diagnostic criteria for ARVC2 were recruited for this prospective study. All participants provided written informed consent to participate. The clinical characteristics and diagnostic scores, according to the ISFC/ESC criteria of ARVC, are summarized in Table I. None of the patients was receiving antiarrhythmic drugs known to affect the QRS complex before being tested for ECG tracings.
Electrocardiographic Analysis. The ECG was obtained from multilead electrophysiologic recording instruments (MAC® 5000 Resting ECG System, GE Healthcare; Chalfont-on-Giles, UK) in order to analyze the value of different ECG criteria. All ARVC patients underwent S-ECG, R-ECG (V3R, V4R, V5R), and F-ECG (FI, FII, and FIII) recordings during sinus rhythm with a paper speed of 25 mm/sec, filter setting of 0.05–30 Hz, and amplitude of 10 mV/mm. In addition to surface ECG tracing, each patient underwent comprehensive testing including signal-averaged ECG, Holter, stress testing, and 2-dimensional echocardiography. Some patients also underwent magnetic resonance imaging, 3-dimensional mapping, and RV angiography. In recording the S-ECG, the V1 electrode was placed on the 4th intercostal space at the right sternal edge, the V2 electrode was placed on the 4th intercostal space at the left sternal edge, the V3R electrode was placed halfway between V1 and V4, the V4R electrode was placed on the 5th intercostal space at the right mid-clavicular line, and the V5R electrode was placed on the 5th intercostal space at the right anterior axillary line.
The Fontaine bipolar precordial leads were placed at the manubrium of sternum, xiphoid, and V4 positions using the right arm connection, left arm connection, and left foot connection, respectively (Fig. 1).4 To determine the ECG values, all recorded ECGs were scanned with a high-resolution digital scanner and digital calipers capable of measuring to within 1 msec (horizontal axis) and 0.01 mV (vertical axis). Three independent clinicians who had no access to the clinical data measured the following ECG values: epsilon waves, QRS duration of 110 msec or more in V1 through V3, T-wave inversion, QRS ratio of 1.2 or higher in (V1 + V2 + V3)/(V4 + V5 + V6),5 and prolonged terminal activation of 55 msec or higher.6 Results were considered to be reliable when at least 2 of the 3 observers agreed. Epsilon waves were defined as any potential after the depolarization between the end of the QRS complex and the beginning of the ST segment.4
Statistical analyses were conducted with Stata® 7.0 software (StataCorp LP; College Station, Texas). Data were expressed as the mean ± SD and as percentages. Comparisons of proportions for dichotomous data were performed by an exact McNemar significance probability analysis. A probability value of P ≤0.05 was considered statistically significant.
Patterns of Epsilon Waves. We found 3 patterns of epsilon waves: wiggle waves, small spike waves, and smooth potential waves with the QRS duration in V1 exceeding the QRS duration in V3 by at least 25 msec. The small spike waves were divided into 2 subtypes, one upward and the other downward (Fig. 2).
S-ECG. Epsilon waves detected by S-ECG in 12 patients (38%) were found chiefly in the right-sided precordial leads (V1 through V3) and to a lesser extent in leads II, III, and aVF (Table II). The mean duration of those waves was 0.039 ± 0.011 sec (range, 0.02–0.06 sec), and the mean amplitude was 0.086 ± 0.04 mV (range, 0.05–0.2 mV). Smooth potential waves were recorded in 1 patient, and small spike waves were observed in the remaining patients. The distribution of epsilon waves varied from patient to patient, but in 2 patients epsilon waves were observed in almost all 12 leads. Epsilon waves occurred in leads V1 through V3 in 6 patients, while notches were observed on the upstroke of the S wave or the downstroke of the R wave in the other leads during the corresponding phase (Fig. 3A). In another patient, epsilon waves were recorded in the limb leads while notches were recorded in the precordial leads.
R-ECG. Epsilon waves were detected by R-ECG in 12 patients (38%) (Table II). The mean duration of those waves was 0.039 ± 0.008 sec (range, 0.03–0.06 sec), and the mean amplitude was 0.08 ± 0.046 mV (range, 0.05–0.2 mV). Epsilon waves detected by R-ECG were configured as small spike waves. In 2 patients, small spike waves were recorded in lead V5R while notches were observed on the upstroke of the S waves in leads V3R and V4R during the corresponding phase (Fig. 3B).
F-ECG. Epsilon waves were detected by F-ECG in 16 patients (50%) (Table II). The mean duration of those waves was 0.054 ± 0.019 sec (range, 0.038–0.08 sec), and the mean amplitude was 0.127 ± 0.072 mV (range, 0.08–0.3 mV). Wiggle waves were recorded in 4 patients, and small spike waves were recorded in 12 patients. In 2 patients, small spike waves were recorded in lead FI, while notches were observed on the upstroke of the S waves in leads FII or FIII (or both) during the corresponding phase (Fig. 3C).
The detection rates of epsilon waves by use of these 3 ECG recording methods (S, R, and F) were 38% (12/32), 38% (12/32), and 50% (16/32), respectively. There was no significant difference in detection rates between individual methods (all comparisons were P > 0.05). Epsilon waves could be detected only by S-ECG in 1 patient (Fig. 4A), only by R-ECG in 3 patients (Fig. 4B), and only by F-ECG in 5 patients (Fig. 4C). The detection rate of epsilon waves was 50% (16/32) by using S-ECG combined with R-ECG (SR-ECG), 56% (18/32) by using S-ECG combined with F-ECG (SF-ECG), and as high as 66% (21/32) when all 3 methods were combined (SRF-ECG). The detection rate of epsilon waves when using some of the methods of ECG recording in combination was significantly higher than that when using S-ECG alone (SF-ECG, 56% vs S-ECG, 38%, P = 0.0312; and SRF-ECG, 66% vs S-ECG, 38%, P = 0.0039) (Fig. 5).
The rate of QRS duration ≥110 msec in leads V1 through V3 was 71% (15/21) in patients with epsilon waves and 73% (8/11) in patients without epsilon waves. In 10 of the 21 (48%) patients with epsilon waves and 1 of the 11 (9%) patients without epsilon waves, we saw T-wave inversion exceeding V3 (P=0.029). In 7 of the 21 (33%) patients with epsilon waves and 2 of the 11 (18%) patients without epsilon waves, we saw diffuse RV dilation. In 20 of the 21 (95%) patients with epsilon waves and 7 of the 11 (64%) patients without epsilon waves, we saw VT (P=0.019). The rates of QRS terminal activation lasting ≥110 msec and QRS ratio ≥1.2 were 62% (13/21) and 52% (11/21), respectively, in patients with epsilon waves. The rates of QRS terminal activation lasting ≥110 msec and QRS ratio ≥1.2 were 27% (3/11) and 36% (4/11) in patients without epsilon waves. See Figure 6 for a comparative summary of these findings on diagnostic markers.
One patient, when first diagnosed for ARVC, had no epsilon waves upon S-ECG (Fig. 7A). The 2nd day, she reported palpitations and another S-ECG was recorded, revealing a wide complex tachycardia with a left bundle branch block morphology. Then the VT was terminated with 70 μg of propafenone administered intravenously. The S-ECG was recorded for a 3rd time, after the patient returned to sinus rhythm (Fig. 7B), and epsilon waves were seen in V1 through V2 (small spike waves with a duration of 0.04 sec and an amplitude of 0.08 mV). However, those waves disappeared on the following day (Fig. 7C).
In 1977, Guy Fontaine first described epsilon waves, which had been recorded at the time of epicardial surgery on patients who manifested sustained VT.7 Since then, ARVC has been noted to be due to genetic defects in cell–cell adhesion that can cause not only myocyte injury—eventually leading to cell death and fibrofatty tissue replacement characterized by islands of viable cardiomyocytes surrounded by fat and fibrosis8—but can also cause sudden death by remodeling gap junctions and altering electrical conduction.9 This characteristic pathogenesis of ARVC explains its ECG features. This kind of cardiomyopathy is regarded as an important cause of sudden death in young people and especially in athletes10; but confirmation of an ARVC diagnosis can be challenging. More and more new ECG features have been proposed, such as a QRS ratio ≥1.25 and a terminal activation lasting ≥55 msec.6 Both measurements of activation delay were developed from RV fibers. However, according to ISFC/ESC diagnostic criteria published in 1994 for ARVC,2 epsilon waves are considered to be one of the major diagnostic criteria. Moreover, they are more specific for diagnosis of ARVC than are the other ECG features.
In our study, we found 3 configurations of epsilon waves, the most common pattern being the small spike wave. These results are similar to those reported by Fontaine.4 The duration and amplitude of epsilon waves detected by F-ECG were longer and larger than those detected by the other 2 ECG recording methods. In some circumstances, epsilon waves were evident in some leads—especially in leads V1 through V3—but presented as notches in the other leads during the corresponding phase. We assume that those notches are the manifestation of other epsilon waves, but this will need confirmation.
Epsilon waves are relatively low in sensitivity, manifesting themselves during S-ECG in only 20% to 25% of ARVC patients1; and those waves are usually seen in leads V1 through V3. However, the percentage can be doubled or tripled by using F-ECG.4 In the present study, we found epsilon waves in 38% of all patients by using S-ECG and increased that to 50% by using F-ECG.
There was no significant difference in detection rates between individual ECG recording methods (all comparisons were P >0.05). The detection rate of epsilon waves using combined methods of ECG recording was significantly higher than that of S-ECG alone (SF-ECG 56% vs S-ECG 38%, P=0.0312; and SRF-ECG 65.63% vs S-ECG 37.5%, P=0.0039). This may be due to the discrepancy in spatial distribution of pathologic lesions11 from patient to patient. The placement of R-ECG leads is closer to the RV than is lead placement for routine surface ECG, while F-ECG is bipolar chest-lead ECG used specifically to record the potentials developed in the RV from the infundibulum to the area of the diaphragm. Therefore, using a combination of ECG recording methods improves the detection rate of epsilon waves. Obviously, even by combining the 3 various methods of ECG recording, the rate of epsilon-wave detection cannot be increased to 100%. This is an inherent defect of the surface ECG, which is a 2-dimensional projection of cardiac electrical activity that cannot truly and totally indicate the prevalence of epsilon waves. Studies on catheter ablation treatment of ventricular arrhythmias in ARVC have shown that delayed potentials can be recorded clearly by local-contact mapping in scar regions while epsilon-wave detection is negative upon surface ECG recording.12
Arrhythmogenic RV cardiomyopathy is characterized by a high incidence of VT, which often occurs early in the disease; left ventricular and biventricular failure may occur at the end stage. In our study, 10 patients had widespread T-wave inversion, but none of these had significant left ventricular contractile dysfunction. Kaplan and colleagues13 found that gap-junction remodeling develops early, before overt structural and functional deterioration has occurred; therefore, patients who display gap-junction remodeling need long-term follow-up. In addition, we found that the rates of VT and widespread T-wave inversion in ARVC patients with epsilon waves were significantly higher than those in patients without epsilon waves and that 70% of patients with widespread T-wave inversion had diffuse RV involvement. However, we cannot conclude that the more extensive the involvement of T-wave inversion, the greater the possibility of epsilon-wave detection. Electrocardiographic detection of epsilon waves relates not only to the distribution of the diseased regions but to the amplitude of delayed potentials and to the extent of delay in activation. Wu and associates14 found that epsilon waves correlate with diffuse RV involvement, but we found no such relationship. It is very difficult to identify patients who are likely to have epsilon waves. Because of our small sample size, we were unable to perform logistic regression to analyze the impact factors of epsilon waves.
In 1 hospitalized patient who had received propafenone intravenously, we noticed epsilon waves in leads V1 through V2, although they had been absent in the original ECG. This observation can be best explained by the impact of propafenone on conduction of normal and viable myocytes that are surrounded by fibrofatty tissue. This special case also shows that the detection of epsilon waves can be affected by the extent of delay in sluggish activation. In addition, epsilon waves might display dynamic changes in some patients during follow-up, which could be the result of disease progression.
In regard to the cost-effectiveness of the procedures, we should bear in mind that S-ECG has a blind region. Therefore, it is worth adding R-ECG and F-ECG to improve the detection rate. Moreover, the 3 ECG recording methods are very inexpensive; in China, each patient pays the equivalent of only US $5 for each examination.
Some of our findings, such as the sharp increase in detection of epsilon waves by F-ECG, might have been affected by our comparatively small population sample or by the racial disparity between our study population and that (for example) of Peters and colleagues,3 who detected epsilon waves in 23% of their study sample by S-ECG and in 75% by F-ECG. Our efforts to detect epsilon waves by R-ECG were successful in only 3 patients, and that might be because the amplitude of the delayed potentials detected by R-ECG is so small that these waves could not be detected. In addition, we did not analyze the correlation between epsilon waves and the distribution of lesions, which certainly warrants further observation.
Because various ECG recording methods can complement each other in detecting epsilon waves, we suggest using the 3 various ECG recording methods in combination to improve the detection rate of epsilon waves.
We are grateful to the patients with arrhythmogenic right ventricular cardiomyopathy who have made this work possible.
Address for reprints: Minglong Chen, MD, Department of Cardiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing 210029, PRC
Dr. Wang and Dr. Yang contributed equally to this work.
This work was supported by the National Basic Research Program of China (973 Program, 2008CB517303).