The value of the 12-lead electrocardiogram in localizing the scar in non-ischaemic cardiomyopathy

Introduction

Based on magnetic resonance imaging (MRI) or using the electroanatomical mapping (EAM) unipolar voltage distribution, non-ischaemic cardiomyopathy (NICM) patients referred for ventricular tachycardia (VT) ablation have been categorized into two discrete scar patterns, anteroseptal (AS) and inferolateral (IL).^1,2 The AS pattern is associated with more extensive endocardial scarring and left ventricular (LV) dysfunction, and an arrhythmic substrate located within the mid-wall of the interventricular septum, leading to higher VT recurrence rates after ablation.³ On the other hand, a high proportion of patients with IL scar had a clinically suspected myocarditis with a predominant epicardial substrate.⁴

Therefore, important relationships between the scar patterns and aetiology, clinical presentation, and the VT ablation procedural approach and outcomes have been established by previous studies.^1,2 However, few centres perform MRI routinely in patients with implantable defibrillators whereas the EAM implies an invasive procedure.

Although in NICM, the electrocardiogram (ECG) during sinus rhythm is usually abnormal, few data have been described to localize the scar. We therefore analysed the 12-lead ECG of the baseline rhythm in patients with NICM admitted for catheter ablation of sustained ventricular arrhythmias dividing them into those with preserved ejection fraction (EF) vs. dilated cardiomyopathy. The aim of this study was to correlate the ECG findings with the endo-epicardial voltage maps performed for the purpose of substrate ablation and identify the ECG characteristics that may help to distinguish the scar as AS or IL. These data may provide adjunctive information to schedule an epicardial approach if an IL scar is suspected by the 12-lead ECG.

Methods

This is a retrospective single-centre study. Patients with structural heart disease and drug-refractory VT that underwent catheter mapping and ablation at the San Raffaele University Hospital, Milan, between January 2010 and December 2013 were screened for inclusion. Patients with coronary artery disease or abnormal loading conditions (congenital, hypertensive, or valvular) were excluded. Patients without EAM scar, defined as at least one segment of unipolar low voltage area (LVA) either in endocardium or in epicardium were excluded. Of 207 patients with a primary cardiomyopathy, a further 57 were excluded due to arrhythmogenic right ventricular cardiomyopathy, restrictive cardiomyopathy, sarcoidosis, hypertrophic cardiomyopathy, and LV non-compaction.⁵ Of the remaining 150 patients, 108 underwent a Carto^® procedure (Biosense Webster, Diamond Bar, CA, USA) with the availability of both contact bipolar and unipolar voltage data and were included in the study. Within this group, 36 NICM patients showed a defined scar at the EAM and normal or mildly impaired LV function and dimensions; the remaining 72 patients fulfilled conventional diagnostic criteria for non-ischaemic dilated cardiomyopathy (NIDCM) (EF < 45% and moderate or severe LV dilatation).^6,7 All patients provided written informed consent.

Electroanatomical mapping

Procedures were performed under general anaesthesia. Electroanatomical mapping was performed with a 3.5 mm open-irrigated catheter including contact force assessment since available (Thermocool^® and Thermocool^® SmartTouchTM, Biosense Webster). The CARTO system filter set for unipolar recording was 2/240 Hz and for bipolar recording 16/500 Hz for high and low pass, respectively. Patients underwent high-density EAM (defined as a 5 mm EAM interpolation fill threshold in areas with reduced unipolar voltage and 10 mm elsewhere). The endocardial LV EAM was segmented using the standardized 17-segment model with the aorta, mitral valve annulus, and apex as references.⁸ Segments not fulfilling fill threshold criteria were excluded. For the epicardium, we have used the transparency tool in CARTO system to analyse the 12 segments. A figure showing an example of endo- and epicardial segmentation is included as Supplementary material online. For each segment, the predominant bipolar and unipolar voltage was evaluated. Unipolar scar and LVA were defined by unipolar voltage of <5 and <8 mV, respectively.⁹ Patients were categorized, as previously published, having either a predominant AS scar sub-type or a IL scar sub-type based on the majority percentage of endocardial segments (eight segments each) displaying a unipolar voltage <8 mV, excluding the apex.² The AS group of segments included the basal anterior, basal AS, basal inferoseptal, mid-anterior, mid-AS, mid-inferoseptal, apical anterior, and apical septal. The IL group of segments are the basal inferior, basal IL, basal anterolateral, mid-inferior, mid-IL, mid-anterolateral, apical inferior, and apical lateral. In patients with a normal endocardial unipolar map (n = 7) or only epicardial map (n = 4), the same system was applied epicardially, using three AS and eight IL segments. In cases of similar percentages of unipolar LVA in the eight AS or IL group of segments (n = 5), we have classified the scar pattern based on bipolar scar distribution, that was only present either anteroseptally or inferolaterally in those patients, and concordant with the site of the ablation lesions.

Electrocardiogram recording and analysis

Before starting the procedure, we programmed the pacing mode to VVI 40 to allow and maximize the analysis of the spontaneous rhythm. Baseline 12-lead ECGs (frequency range: 0.05–100 Hz, 25 mm/s, 10 mm/mV; Prucka CardioLab GE Medical Systems) were registered. The intervals were measured using digitized tracings with calipers allowing 1 ms and 0.01 mV resolution at a screen velocity of 100 mm/s. Low voltage in the extremity leads was considered if <0.6 mV in all the limb leads.¹⁰

The definition of a fragmented QRS included various triphasic QRS patterns such as notching of the R or S wave or the presence of more than two R waves crossing the baseline, in the absence of a wide QRS (>120 ms).¹¹ Two observers blinded for the EAM data made all measurements. In cases of discordance of >20 ms or >0.2 mV, a third observer was used. A mean of the two measurements was used to minimize the error and in cases of three observers, an average of the two closer values was used deleting the outlying value.

Statistical analysis

Continuous variables are presented as means (±SD) or medians (25th and 75th percentiles) and categorical variables as numbers and percentages. Comparisons between groups were undertaken using the t-test or non-parametric tests for continuous variables and Fisher's exact or χ² tests for proportions. The differences were considered statistically significant at the two-sided P< 0.05 level. Interobserver variability of ECG measurements was determined using intra-class correlation (ICC). We sought to create an algorithm that could consistently identify the AS vs. IL scar pattern. Only those significant variables were selected for the sensitivity and specificity analysis, including those that provided a high specificity ≥90% with a sensitivity of at least 20%. For continuous variables (PR and QRS intervals), receiver operating characteristic curves selected a cut-off value. The correlation between areas and ECG features was assessed using Spearman ranks correlation coefficient (r_s). All analyses were undertaken using IBM^® SPSS^® Statistics Version 21 (IBM Corporation).

Results

Patient population

Of the 72 patients meeting DCM diagnostic criteria, 49 were categorized as predominant AS (AS-NIDCM) and 23 IL (IL-NIDCM). Of the remaining 36, 10 had predominant AS scar (AS-NICM) and 26 IL (IL-NICM). The baseline characteristics and EAM data are presented in Table 1.

The IL-NICM vs. AS-NICM group had a higher prevalence of previous clinically suspected myocarditis (65 vs. 0%, P < 0.001) and younger age (52 ± 16 vs. 64 ± 16 years, P = 0.046). Prior myocarditis was also more suspected with IL-NIDCM (17%) vs. AS-NIDCM (2%), P = 0.041. On the other hand, prior cardiac resynchronization therapy with defibrillator implantation was more frequent in AS-NIDCM when compared with IL-NIDCM (65 vs. 13%, P < 0.001).

The overall epicardial procedure rate was 67%. A total of 1364 of 1530 endocardial and 688 of 792 epicardial segments fulfilled threshold criteria and were analysed.

Electrocardiographic characteristics anteroseptal-non-ischaemic cardiomyopathy vs. inferolateral-non-ischaemic cardiomyopathy

The AS-NICM group displayed a longer mean PR interval compared with the IL-NICM (208 ± 39 vs. 162 ± 28 ms, P = 0.001) associated with a longer median HV interval [55 ms (48–64) vs. 40 ms (36–50), P = 0.005] without differences in the QRS duration.

In the IL-NICM group, 13 patients (50%) exhibited a low QRS voltage in the limb leads vs. 1 patient (10%) in the AS group, P = 0.05. A fragmented QRS in the inferior leads and the presence of a q wave in the lateral leads were specifically found in this group (39 and 35% vs. 0% in AS-NICM, P = 0.035 and P = 0.038, respectively). Finally, the median R wave in V6 was reduced in IL-NICM compared with AS-NICM [0.72 mV (0.44–0.80) vs. 1.02 mV (0.79–1.19), P = 0.031] (Figures 1 and 2). The electrocardiographic data are given in Table 2.

Table 2

Open in new tab

Electrocardiographic characteristics

.	AS-NICM (n = 10) .	IL-NICM (n = 26) .	P .	AS-NIDCM (n = 49) .	IL-NIDCM (n = 23) .	P .
Paced/AVN ablation	0 (0%)	0 (0%)	1.00	0 (0%)	2 (9%)	0.03
Paced/no rhythm	0 (0%)	0 (0%)	1.00	11 (22%)	1 (5%)	0.09
Atrial fibrillation	1 (10%)	0 (0%)	0.27	3 (8%)	1 (5%)	1.00
PR interval	9	26		35	19
PR duration (ms)	208 ± 39	162 ± 28	0.001	212 (183–265)	182 (171–195)	0.008
PR >200 ms	4 (44%)	4 (15%)	0.16	21/38 (60%)	4/19 (21%)	0.006
Any AV block	4 (44%)	4 (15%)	0.16	32/49 (65%)	5/20 (25%)	0.001
QRS duration	122 ± 33	109 ± 22	0.18	171 (134–191)	113 (102–159)	<0.001
QRS axis frontal plane
From −30 to +90	8 (80%)	21 (81%)	0.71	22 (58%)	10 (50%)	0.39
QRS <120 ms
QRS <120 ms	5 (50%)	21 (81%)	0.10	6 (16%)	12 (60%)	0.001
Normal ECGa	3 (30%)	10 (39%)	0.71	0 (0%)	2 (10%)	0.11
QRS ≥120 ms	5	5		32	8
IVCD	1 (10%)	2 (7%)	1.00	7 (18%)	1 (5%)	0.24
RBBB	2 (20%)	2 (7%)	0.30	7 (18%)	2 (10%)	0.47
LBBB	2 (20%)	1 (4%)	0.18	18 (47%)	5 (25%)	0.09
QRS voltage
Amplitude limb leads	0.59 ± 0.15	0.50 ± 0.15	0.11	0.52 (0.43–0.68)	0.45 (0.35–0.58)	0.038
Amplitude inferior	0.68 ± 0.26	0.53 ± 0.21	0.08	0.64 (0.44–0.86)	0.42 (0.31–0.67)	0.037
Limb leads <0.6 mV	1 (10%)	13 (50%)	0.05	13 (34%)	14 (70%)	0.009
Inferior leads <0.6 mV	1 (10%)	14 (54%)	0.024	14 (37%)	15 (75%)	0.006
Inferior leads
Q wave	0 (0%)	2 (7%)	1.00	3 (8%)	5 (25%)	0.11
q wave	1 (10%)	10 (39%)	0.12	5 (13%)	7 (35%)	0.08
Frag. inferior QRS	0 (0%)	10 (39%)	0.035	1 (3%)	6 (30%)	0.005
Lateral leads
Q wave	0 (0%)	2 (7%)	1.00	0 (0%)	0 (0%)	1.00
q wave	0 (0%)	9 (35%)	0.039	1 (3%)	9 (45%)	<0.001
Frag. lateral QRS	0 (0%)	2 (7%)	1.00	0 (0%)	3 (15%)	0.037
Precordial leads
Data excluding BBB	6	23		13	11
R in V1	0.25 (0.0–0.94)	0.20 (0.14–0.22)	0.38	0.14 ± 0.11	0.21 ± 0.11	0.13
R in V2	0.26 ± 0.25	0.55 ± 0.36	0.08	0.34 (0.19–0.64)	0.40 (0.32–0.63)	0.27
R in V3	0.75 ± 0.53	0.77 ± 0.38	0.92	0.35 (0.23–0.51)	0.72 (0.38–1.12)	0.018
R in V4	0.77 ± 0.19	0.93 ± 0.50	0.47	0.58 (0.41–0.91)	0.76 (0.56–1.41)	0.27
R in V5	0.99 (0.87–1.18)	0.78 (0.60–1.08)	0.19	0.74 ± 0.35	0.86 ± 0.53	0.51
R in V6	1.02 (0.79–1.19)	0.72 (0.44–0.80)	0.031	0.67 ± 0.27	0.64 ± 0.30	0.76

.	AS-NICM (n = 10) .	IL-NICM (n = 26) .	P .	AS-NIDCM (n = 49) .	IL-NIDCM (n = 23) .	P .
Paced/AVN ablation	0 (0%)	0 (0%)	1.00	0 (0%)	2 (9%)	0.03
Paced/no rhythm	0 (0%)	0 (0%)	1.00	11 (22%)	1 (5%)	0.09
Atrial fibrillation	1 (10%)	0 (0%)	0.27	3 (8%)	1 (5%)	1.00
PR interval	9	26		35	19
PR duration (ms)	208 ± 39	162 ± 28	0.001	212 (183–265)	182 (171–195)	0.008
PR >200 ms	4 (44%)	4 (15%)	0.16	21/38 (60%)	4/19 (21%)	0.006
Any AV block	4 (44%)	4 (15%)	0.16	32/49 (65%)	5/20 (25%)	0.001
QRS duration	122 ± 33	109 ± 22	0.18	171 (134–191)	113 (102–159)	<0.001
QRS axis frontal plane
From −30 to +90	8 (80%)	21 (81%)	0.71	22 (58%)	10 (50%)	0.39
QRS <120 ms
QRS <120 ms	5 (50%)	21 (81%)	0.10	6 (16%)	12 (60%)	0.001
Normal ECGa	3 (30%)	10 (39%)	0.71	0 (0%)	2 (10%)	0.11
QRS ≥120 ms	5	5		32	8
IVCD	1 (10%)	2 (7%)	1.00	7 (18%)	1 (5%)	0.24
RBBB	2 (20%)	2 (7%)	0.30	7 (18%)	2 (10%)	0.47
LBBB	2 (20%)	1 (4%)	0.18	18 (47%)	5 (25%)	0.09
QRS voltage
Amplitude limb leads	0.59 ± 0.15	0.50 ± 0.15	0.11	0.52 (0.43–0.68)	0.45 (0.35–0.58)	0.038
Amplitude inferior	0.68 ± 0.26	0.53 ± 0.21	0.08	0.64 (0.44–0.86)	0.42 (0.31–0.67)	0.037
Limb leads <0.6 mV	1 (10%)	13 (50%)	0.05	13 (34%)	14 (70%)	0.009
Inferior leads <0.6 mV	1 (10%)	14 (54%)	0.024	14 (37%)	15 (75%)	0.006
Inferior leads
Q wave	0 (0%)	2 (7%)	1.00	3 (8%)	5 (25%)	0.11
q wave	1 (10%)	10 (39%)	0.12	5 (13%)	7 (35%)	0.08
Frag. inferior QRS	0 (0%)	10 (39%)	0.035	1 (3%)	6 (30%)	0.005
Lateral leads
Q wave	0 (0%)	2 (7%)	1.00	0 (0%)	0 (0%)	1.00
q wave	0 (0%)	9 (35%)	0.039	1 (3%)	9 (45%)	<0.001
Frag. lateral QRS	0 (0%)	2 (7%)	1.00	0 (0%)	3 (15%)	0.037
Precordial leads
Data excluding BBB	6	23		13	11
R in V1	0.25 (0.0–0.94)	0.20 (0.14–0.22)	0.38	0.14 ± 0.11	0.21 ± 0.11	0.13
R in V2	0.26 ± 0.25	0.55 ± 0.36	0.08	0.34 (0.19–0.64)	0.40 (0.32–0.63)	0.27
R in V3	0.75 ± 0.53	0.77 ± 0.38	0.92	0.35 (0.23–0.51)	0.72 (0.38–1.12)	0.018
R in V4	0.77 ± 0.19	0.93 ± 0.50	0.47	0.58 (0.41–0.91)	0.76 (0.56–1.41)	0.27
R in V5	0.99 (0.87–1.18)	0.78 (0.60–1.08)	0.19	0.74 ± 0.35	0.86 ± 0.53	0.51
R in V6	1.02 (0.79–1.19)	0.72 (0.44–0.80)	0.031	0.67 ± 0.27	0.64 ± 0.30	0.76

IVCD, intraventricular conduction delay; Frag., fragmented; BBB, bundle brunch block.

^aIt includes SR, normal PR, QRS < 120 ms, no fragmented QRS, normal depolarization.

Table 2

Open in new tab

Electrocardiographic characteristics

.	AS-NICM (n = 10) .	IL-NICM (n = 26) .	P .	AS-NIDCM (n = 49) .	IL-NIDCM (n = 23) .	P .
Paced/AVN ablation	0 (0%)	0 (0%)	1.00	0 (0%)	2 (9%)	0.03
Paced/no rhythm	0 (0%)	0 (0%)	1.00	11 (22%)	1 (5%)	0.09
Atrial fibrillation	1 (10%)	0 (0%)	0.27	3 (8%)	1 (5%)	1.00
PR interval	9	26		35	19
PR duration (ms)	208 ± 39	162 ± 28	0.001	212 (183–265)	182 (171–195)	0.008
PR >200 ms	4 (44%)	4 (15%)	0.16	21/38 (60%)	4/19 (21%)	0.006
Any AV block	4 (44%)	4 (15%)	0.16	32/49 (65%)	5/20 (25%)	0.001
QRS duration	122 ± 33	109 ± 22	0.18	171 (134–191)	113 (102–159)	<0.001
QRS axis frontal plane
From −30 to +90	8 (80%)	21 (81%)	0.71	22 (58%)	10 (50%)	0.39
QRS <120 ms
QRS <120 ms	5 (50%)	21 (81%)	0.10	6 (16%)	12 (60%)	0.001
Normal ECGa	3 (30%)	10 (39%)	0.71	0 (0%)	2 (10%)	0.11
QRS ≥120 ms	5	5		32	8
IVCD	1 (10%)	2 (7%)	1.00	7 (18%)	1 (5%)	0.24
RBBB	2 (20%)	2 (7%)	0.30	7 (18%)	2 (10%)	0.47
LBBB	2 (20%)	1 (4%)	0.18	18 (47%)	5 (25%)	0.09
QRS voltage
Amplitude limb leads	0.59 ± 0.15	0.50 ± 0.15	0.11	0.52 (0.43–0.68)	0.45 (0.35–0.58)	0.038
Amplitude inferior	0.68 ± 0.26	0.53 ± 0.21	0.08	0.64 (0.44–0.86)	0.42 (0.31–0.67)	0.037
Limb leads <0.6 mV	1 (10%)	13 (50%)	0.05	13 (34%)	14 (70%)	0.009
Inferior leads <0.6 mV	1 (10%)	14 (54%)	0.024	14 (37%)	15 (75%)	0.006
Inferior leads
Q wave	0 (0%)	2 (7%)	1.00	3 (8%)	5 (25%)	0.11
q wave	1 (10%)	10 (39%)	0.12	5 (13%)	7 (35%)	0.08
Frag. inferior QRS	0 (0%)	10 (39%)	0.035	1 (3%)	6 (30%)	0.005
Lateral leads
Q wave	0 (0%)	2 (7%)	1.00	0 (0%)	0 (0%)	1.00
q wave	0 (0%)	9 (35%)	0.039	1 (3%)	9 (45%)	<0.001
Frag. lateral QRS	0 (0%)	2 (7%)	1.00	0 (0%)	3 (15%)	0.037
Precordial leads
Data excluding BBB	6	23		13	11
R in V1	0.25 (0.0–0.94)	0.20 (0.14–0.22)	0.38	0.14 ± 0.11	0.21 ± 0.11	0.13
R in V2	0.26 ± 0.25	0.55 ± 0.36	0.08	0.34 (0.19–0.64)	0.40 (0.32–0.63)	0.27
R in V3	0.75 ± 0.53	0.77 ± 0.38	0.92	0.35 (0.23–0.51)	0.72 (0.38–1.12)	0.018
R in V4	0.77 ± 0.19	0.93 ± 0.50	0.47	0.58 (0.41–0.91)	0.76 (0.56–1.41)	0.27
R in V5	0.99 (0.87–1.18)	0.78 (0.60–1.08)	0.19	0.74 ± 0.35	0.86 ± 0.53	0.51
R in V6	1.02 (0.79–1.19)	0.72 (0.44–0.80)	0.031	0.67 ± 0.27	0.64 ± 0.30	0.76

.	AS-NICM (n = 10) .	IL-NICM (n = 26) .	P .	AS-NIDCM (n = 49) .	IL-NIDCM (n = 23) .	P .
Paced/AVN ablation	0 (0%)	0 (0%)	1.00	0 (0%)	2 (9%)	0.03
Paced/no rhythm	0 (0%)	0 (0%)	1.00	11 (22%)	1 (5%)	0.09
Atrial fibrillation	1 (10%)	0 (0%)	0.27	3 (8%)	1 (5%)	1.00
PR interval	9	26		35	19
PR duration (ms)	208 ± 39	162 ± 28	0.001	212 (183–265)	182 (171–195)	0.008
PR >200 ms	4 (44%)	4 (15%)	0.16	21/38 (60%)	4/19 (21%)	0.006
Any AV block	4 (44%)	4 (15%)	0.16	32/49 (65%)	5/20 (25%)	0.001
QRS duration	122 ± 33	109 ± 22	0.18	171 (134–191)	113 (102–159)	<0.001
QRS axis frontal plane
From −30 to +90	8 (80%)	21 (81%)	0.71	22 (58%)	10 (50%)	0.39
QRS <120 ms
QRS <120 ms	5 (50%)	21 (81%)	0.10	6 (16%)	12 (60%)	0.001
Normal ECGa	3 (30%)	10 (39%)	0.71	0 (0%)	2 (10%)	0.11
QRS ≥120 ms	5	5		32	8
IVCD	1 (10%)	2 (7%)	1.00	7 (18%)	1 (5%)	0.24
RBBB	2 (20%)	2 (7%)	0.30	7 (18%)	2 (10%)	0.47
LBBB	2 (20%)	1 (4%)	0.18	18 (47%)	5 (25%)	0.09
QRS voltage
Amplitude limb leads	0.59 ± 0.15	0.50 ± 0.15	0.11	0.52 (0.43–0.68)	0.45 (0.35–0.58)	0.038
Amplitude inferior	0.68 ± 0.26	0.53 ± 0.21	0.08	0.64 (0.44–0.86)	0.42 (0.31–0.67)	0.037
Limb leads <0.6 mV	1 (10%)	13 (50%)	0.05	13 (34%)	14 (70%)	0.009
Inferior leads <0.6 mV	1 (10%)	14 (54%)	0.024	14 (37%)	15 (75%)	0.006
Inferior leads
Q wave	0 (0%)	2 (7%)	1.00	3 (8%)	5 (25%)	0.11
q wave	1 (10%)	10 (39%)	0.12	5 (13%)	7 (35%)	0.08
Frag. inferior QRS	0 (0%)	10 (39%)	0.035	1 (3%)	6 (30%)	0.005
Lateral leads
Q wave	0 (0%)	2 (7%)	1.00	0 (0%)	0 (0%)	1.00
q wave	0 (0%)	9 (35%)	0.039	1 (3%)	9 (45%)	<0.001
Frag. lateral QRS	0 (0%)	2 (7%)	1.00	0 (0%)	3 (15%)	0.037
Precordial leads
Data excluding BBB	6	23		13	11
R in V1	0.25 (0.0–0.94)	0.20 (0.14–0.22)	0.38	0.14 ± 0.11	0.21 ± 0.11	0.13
R in V2	0.26 ± 0.25	0.55 ± 0.36	0.08	0.34 (0.19–0.64)	0.40 (0.32–0.63)	0.27
R in V3	0.75 ± 0.53	0.77 ± 0.38	0.92	0.35 (0.23–0.51)	0.72 (0.38–1.12)	0.018
R in V4	0.77 ± 0.19	0.93 ± 0.50	0.47	0.58 (0.41–0.91)	0.76 (0.56–1.41)	0.27
R in V5	0.99 (0.87–1.18)	0.78 (0.60–1.08)	0.19	0.74 ± 0.35	0.86 ± 0.53	0.51
R in V6	1.02 (0.79–1.19)	0.72 (0.44–0.80)	0.031	0.67 ± 0.27	0.64 ± 0.30	0.76

IVCD, intraventricular conduction delay; Frag., fragmented; BBB, bundle brunch block.

^aIt includes SR, normal PR, QRS < 120 ms, no fragmented QRS, normal depolarization.

Figure 1

Examples of 12-lead ECG findings in patients with NICM and an AS (left panel) or IL scar pattern (right panel).

Open in new tabDownload slide

Figure 2

The figure illustrates a patient with NICM and preserved EF and inferolateral scar. The epicardial unipolar voltage map (left lateral view) and a 12-lead ECG showing a low QRS voltage in the extremity leads.

Open in new tabDownload slide

Electrocardiographic characteristics anteroseptal-non-ischaemic dilated cardiomyopathy vs. inferolateral-non-ischaemic dilated cardiomyopathy

Within the AS-NIDCM group, 32 patients (65%) presented any degree of AV block vs. 5 patients (25%) in the IL-NIDCM group, P = 0.001. Concordantly, 34 patients with AS-NIDCM (69%) had an HV interval longer than 55 ms and only 6 patients in the IL-NIDCM group (33%), P = 0.008. The median PR [212 ms (183–265) vs. 182 ms (171–195), P = 0.008) and QRS duration (171 ms (134–191) vs. 113 ms (102–159), P < 0.001] were significantly longer in AS-NIDCM vs. IL-NIDCM (Figures 3 and 4). The mean r voltage in V3 was reduced in the AS-NIDCM group [0.35 mV (0.23–0.51) vs. 0.72 mV (0.38–1.12), P = 0.018] (Figures 3 and 4).

Figure 3

Examples of 12-lead ECG findings in patients with a dilated cardiomyopathy and an AS (left panel) or IL scar pattern (right panel).

Open in new tabDownload slide

Figure 4

A patient with dilated cardiomyopathy is represented. The endocardial unipolar voltage map (right anterior oblique view) showed an AS scar and the 12-lead ECG a widened QRS.

Open in new tabDownload slide

Within the IL-NIDCM group, a low QRS voltage in the limb leads was commonly found (70 vs. 34% in the AS-NIDCM, P = 0.009). A fragmented QRS in the inferior leads (30 vs. 3%, P = 0.005) or in the lateral leads (15 vs. 0%, P = 0.037) and the presence of a q wave in the inferior leads (45 vs. 3%, P < 0.001) were also more frequent in the IL group.

Anteroseptal and inferolateral scar pattern

In the NICM patients, with preserved LV function, the scar was highly localized in the AS or IL area. However, particularly in the dilated forms, the extent of the scar is greater, sometimes beyond the corresponding AS or IL area. Twenty-six per cent of patients in the IL group had a prolonged PR, or a prolonged QRS duration and exhibited significantly greater areas of total endocardial unipolar LVA [85 (33–104) cm² vs. 18 (3–32) cm², P = 0.002] and specifically greater areas of unipolar LVA in the eight AS segments [13 (0–30) cm² vs. 0 (0–4) cm², P = 0.003] compared with the IL-NIDCM with normal PR and QRS duration. In the AS group, the 29% of patients with low voltage in the limb leads had greater areas of unipolar LVA in the eight IL segments [35 (12–66) cm² vs. 13 (0–27) cm², P = 0.041].

Correlation between electroanatomical mapping data and electrocardiogram findings

A significant negative correlation was found between the extension of the endocardial unipolar LVA and LVEF (r_s = −0.719, P < 0.001) whereas no correlation was observed with the epicardial unipolar LVA (r_s = −0.029, P = 0.825). The unipolar LVA within the eight AS segments correlated with both QRS duration (r_s= 0.680, P < 0.001) and PR interval (r_s = 0.583, P < 0.001) (Figure 5). A significant negative correlation was found between the percentage of unipolar LVA in the eight IL segments in the epicardium and the mean voltage of the limb leads (r_s = −0.639, P < 0.001).

Figure 5

The scatterplot on the left shows a positive correlation between the percentage of the unipolar low voltage in the endocardial AS segments (x-axis) and the QRS duration (y-axis). The scatterplot on the right shows the negative correlation between the percentage of the unipolar low voltage in the epicardial IL segments (x-axis) and the mean voltage of the limb leads (y-axis). Anteroseptal patients are plotted in turquoise and IL patients in red.

Open in new tabDownload slide

Bipolar and imaging scar related to unipolar voltage mapping

No bipolar scar was identified in the AS-NICM group. In the IL-NICM group, bipolar scar was detected in 9 patients (16 segments, 81% epicardially) with the basal IL and basal inferior segments most affected (31%) and 94% of segments with bipolar scar were included within the IL group of segments. In the AS-NIDCM group, bipolar scar was detected in 13 patients (26 segments, 85% endocardially, 35% basal AS segments) and 92% of segments were included within the AS group of segments. In the IL-NIDCM group, 11 patients showed bipolar scar (30 segments, 73% epicardially, 37% basal IL segment) and 87% of segments were included within the IL group of segments. Overall, 65 patients out of 94 (69%) did not exhibit any segment with bipolar dense scar (<0.5 mV). In the 64 segments with bipolar dense scar, there was also underlying unipolar LVA (97% unipolar scar and 3% unipolar LVA).

Forty-three patients underwent delayed-enhanced pre-procedural imaging (20 MRI, 23 CT). The agreement for scar location with EAM was high (Κ = 0.769, P< 0.001), entirely concordant with MRI, whereas in three patients an AS scar and in one patient an IL scar were missed by CT, readily detectable with EAM.

Twelve-lead electrocardiogram algorithms

The three-step algorithm for NICM included a PR interval of <170 ms [the area under the curve (AUC) 0.863, sensitivity 73%, specificity 100%] or a QRS voltage in the inferior leads of <0.6 mV (sensitivity 50%, specificity 90%) or the presence of a q wave in the lateral wall (sensitivity 35%, specificity 100%). It showed a high sensitivity (92%, 24/26 IL-NICM) and specificity (90%, only 1 AS-NICM fulfilled the criteria) for identifying the IL scar pattern.

The four-step algorithm for predicting AS scar pattern in NIDCM included a paced ventricular rhythm (sensitivity 22%, specificity 95%) or a QRS duration >170 ms (AUC 0.829, sensitivity 50%, specificity 95%) or a PR interval >230 ms (AUC 0.724, sensitivity 42%, specificity 95%) or an r wave in V3 ≤0.3 mV in the absence of bundle brunch block (BBB) (sensitivity 25%, specificity 100%). Forty-four of the 49 AS-DCM patients were correctly identified and 4 of 20 IL-DCM were misdiagnosed (92% sensitivity, 81% specificity). As Figure 6 shows, with a single ECG criterion, the algorithm is fulfilled due to the highly specific criteria utilized, achieving a good sensitivity.

Figure 6

The flowchart represents the initial evaluation of non-ischaemic patients with VT and the algorithms, for NICM and for NIDCM to localize the scar by using the baseline 12-lead ECG.

Open in new tabDownload slide

Interobserver variability/intra-class correlation coefficient

A third measurement was required in 22 of 87 PR intervals, 17 of 94 QRS durations, and 38 of 1128 voltage measurements (113 if R and S single measurements were included). Good interobserver correlations were noted for PR interval and QRS duration (both ICC 0.98), QRS voltage in the limb leads (ICC 0.97–0.98), QRS voltage in precordial leads (0.98–0.99), and ‘r’ wave in V3 (ICC 0.94).

Discussion

Suspecting the presence of a scar and distinguishing its location in patients with NICM has clinical implications that can improve their management. In this population, the extent of unipolar scar in the endocardial AS or epicardial IL area was significantly correlated with the following ECG parameters: PR interval, QRS duration, and the mean voltage in the limb leads. Simple ECG analysis therefore allows scar location to be predicted in most of these patients.

Unipolar voltage in non-ischaemic cardiomyopathy

Unipolar recordings often contain potentials that are not generated by local activation, but instead, originate at a distance from the region where the electrode is placed. However, despite its limitations, they are commonly used for the study of ventricular arrhythmias in NICM.^2,4,12,13 In fact, bipolar abnormalities are scarcely represented in this population, and in this cohort only around 30% of patients showed some segment with bipolar dense scar. The information that can be provided by imaging is also restricted as MRI is not usually performed in defibrillator carriers. In our study, the bipolar findings, when present, and the imaging studies, when available, were concordant with the unipolar mapping data and demonstrated the same AS or IL predominance in scar distribution for NICM patients. The unipolar scar detected by EAM has been related to fibrosis in NICM and the progression of unipolar scar detected by two EAMs separated in time was associated with a decrease in LVEF.^12,13 Also in this cohort, the extension of low unipolar voltage in the endocardium was well correlated with the LVEF. In the AS-NICM group, we observed a long PR interval with a non-widened QRS whereas in the AS-DCM group the most important difference was found in the duration of the QRS. Both parameters were correlated to the amount of unipolar scar in the AS area. Therefore, the specific AS scar area has an impact on the 12-lead ECG by prolonging the QRS duration and PR interval. This finding probably leads to a higher need for CRT and pacing-dependency in the AS-DCM group.

Low voltage in the limb leads in non-ischaemic cardiomyopathy

In this study, a low QRS voltage in all limb leads was present in 70% of the patients with IL-DCM and in 50% of the patients with IL-NICM. Of interest is the observation by Chinitz et al.¹⁴ in 100 patients with a low QRS voltage in the limb leads that no demonstrable traditional aetiology of low voltage was found in half of them whereas the other half had DCM concluding that low limb lead amplitudes in the presence of even normal precordial voltages may be a useful ECG marker for DCM. This is in agreement with an earlier publication by Goldberger¹⁵ who found a decrease in the bipolar limb leads voltage when intracardiac volume increases, in contrast to the voltage of the unipolar precordial leads. Interestingly, the IL-NICM patients with normal or only mildly altered LV function frequently showed low voltage in the limb leads. A significant negative correlation was found between the voltage in the limb leads and the size of the epicardial voltage abnormalities in the IL area, often associated with a fragmented QRS, both in IL NICM and in NIDCM, suggesting that the low voltage in the limb leads may be caused by myocardial tissue loss and substitution by fibrosis distributed in a patchy, non-compact transmural manner, typical of NICM.¹⁶

Clinical implications

Prior reports have described the epicardial predilection of the arrhythmogenic substrate for NICM although the indication for an epicardial approach is still not well defined.¹⁷ Up to 60% of the NICM patients with ventricular arrhythmias do not need an epicardial approach guided by contrast-enhanced MRI VT ablation according to a recent publication by Andreu et al.¹⁸ Our study showed similar results with radiofrequency delivered epicardially in only 65% (47 of 72) of the total epicardial maps performed, significantly 87% corresponded to the IL scar sub-type. A first-line endo-epicardial approach will be more helpful in cases with an IL scar pattern because most of the scar is located epicardially in IL-NICM and endo-epicardially in IL-NIDCM. In the subgroup of patients with preserved EF, an IL scar pattern is more common (53% of IL-NICM showed preserved EF whereas only 17% in case of AS-NICM), particularly in the presence of a low limb QRS voltage, normal PR or fragmented ECG in the inferior or lateral wall. These patients usually presented with an arrhythmic debut (94%) as syncope, fast VT, or cardiac arrest usually without an ICD in place, allowing a cardiac MRI to be performed, and clinical suspicion for myocarditis as the aetiology. On the other hand, the NICM group with DCM usually presented many years earlier with heart failure and patients were already defibrillator carriers (CRT in case of the AS group) presenting ventricular arrhythmias late in their clinical course.

It is unknown whether different scar locations in NICM are due to disease-specific sites or to areas more vulnerable to different types of insults, namely excess of mechanical stress and the cardiomyopathic process. Laminopathies, especially mutations in lamin A/C, are preferentially localized in basal and midventricular septal sites and often present with conduction abnormalities.¹⁹ On the other hand, previous studies showed that myocarditis mainly involves the basal free wall of the LV.^4,20 The typical IL ECG features are also frequently present in patients with no structural abnormalities. This observation may generate further investigations to screen asymptomatic patients with preserved EF that are carriers of this pattern before an episode of syncope or cardiac arrest occurred in a quarter of patients in this NICM subgroup.

Limitations

The study population constituted a selected cohort with ventricular arrhythmias, questioning its extrapolation to all NICM patients. Biopsy-proven diagnoses or imaging studies are not provided for the entire group. The dichotomous classification may be a limiting factor given the heterogeneity of patients. As voltage amplitude and signal recording shape can be changed by the wavefront of electrical activation, particularly working with unipolar signals, there may be some differences for unipolar data in paced patients. In relation to the discussion on low QRS voltage, we do not have accurate measurements of the total right ventricular volume to make a definite distinction between patients in whom low voltage was caused by myocardial tissue loss, increased intracardiac volume, or both.¹⁰

Conclusion

A small r in V3 and the presence of substantial conduction delay with lengthening of the PR interval, or QRS duration or a paced rhythm with a CRT device are all criteria that suggest an AS scar pattern in NICM. A PR interval of <170 ms and a low voltage, q wave or fragmented QRS in the limb leads is more frequently observed in cases with an IL scar pattern, both in patients with dilated cardiomyopathy and in patients with preserved EF. Further developments are required to improve our knowledge about the aetiology, location, and characteristics of the scar.

Supplementary material

Supplementary material is available at Europace online.

Conflict of interest: P.D.B. is consultant for St Jude Medical and has received honoraria for lectures from Biosense Webster, St Jude Medical, and Biotronik. T.O. and J.S. are Advanced European Heart Rhythm Association Fellow with a grant funded by Biosense Webster.

References