Abstract
Background
To develop a new algorithm to differentiate ventricular tachycardia (VT) from preexcited tachycardia (pre‐ET) according to left bundle branch block (LBBB) and right bundle branch block (RBBB) patterns.
Methods
This study included 67 electrocardiograms (ECGs) with VT and 63 ECGs with pre‐ET, collected from our hospital and through PubMed. Of those, 64 were allocated to the derivation cohort and the rest to the validation cohort. The diagnoses of the ECGs were confirmed using an electrophysiological study. Parameters and classifiers from prior algorithms along with the propagation speeds in the early portion of the QRS complex (initial deflection index) in leads V1, V6, aVR, II, and III were manually measured. The performance of the new algorithm was compared with that of prior algorithms.
Results
The initial deflection index in lead III was the strongest predictor of pre‐ET in LBBB‐pattern wide‐QRS tachycardia (p = 0.003, AUC 0.805). The initial deflection index in lead V1 was the most powerful predictor of pre‐ET in RBBB‐pattern wide‐QRS tachycardia (p = 0.001, AUC 0.848). Compared to earlier algorithms, those using the initial deflection indexes: lead III in LBBB patterns (cutoff value >0.3) and lead V1 in RBBB patterns (cutoff value ≤0.48), demonstrated superior performance in screening VT, with AUC values of 0.828. The initial deflection indexes proved effective as discriminators between VT and pre‐ET in the validation cohort.
Conclusions
In LBBB‐pattern wide‐QRS tachycardia, the early propagation speed of pre‐ET was faster than that in VT. Conversely, in RBBB‐pattern wide‐QRS tachycardia, it was slower.
Keywords: diagnosis, differential; electrocardiography; electrophysiology; preexcitation syndromes; tachycardia, ventricular
Previous algorithms for differentiating between SVT and VT were unable to distinguish wide‐QRS preexcited tachycardia from VT; however, our algorithm utilizing the propagation speeds in the early portion of the QRS complex (initial deflection index) proved useful for this differentiation.
1. INTRODUCTION
There have been many studies comparing ventricular tachycardia (VT) and supraventricular tachycardia (SVT) on distinguishing wide‐QRS tachycardia by electrocardiogram (ECG). Discriminating between VT and SVT is crucial as the prognosis and treatment vary based on the diagnosis. Similarly, distinguishing wide‐QRS preexcited tachycardia (pre‐ET) from VT is vital, but studies comparing VT and pre‐ET are scarce. Only three studies related to discrimination between VT and pre‐ET were found in the literature. One of the three studies compared only specificity and did not included control group (Jastrzębski et al., 2018), while the others by Steurer and Verechei, reported sensitivity and specificity (75% and 100%; 73% and 74.6%), respectively (Steurer et al., 1994; Vereckei et al., 2023). The importance of distinguishing pre‐ET from VT is underscored by the limited data on pre‐ET; the incidence of pre‐ET with a wide QRS complex stands at a mere 2% (Brady et al., 2017). In a report of 228 subjects with Wolff–Parkinson–White (WPW) syndrome, the overall incidence of tachycardia was 1% per patient‐year (Fitzsimmons et al., 2001), and in another report, the incidence of antidromic atrioventricular reentrant tachycardia (AVRT) in WPW syndrome patients was 5% (Ali et al., 2018). Among 807 patients with preexcitation, 63 (8%) had inducible antidromic AVRT during electrophysiological studies (Brembilla‐Perrot, Pauriah, et al., 2013).
Accessory pathway (AP) can exhibit diverse circuits and can be closely or remotely connected to the His‐Purkinje system. The rate of early propagation can differ, since atypical bypass tracts can span from the atrium to the bundle branch, the AV node to the bundle branch, the AV node to the ventricular tissue, and from the bundle branch to the ventricular tissue (Anderson et al., 1975). The morphology and duration of QRS components due to the varying circuits of pre‐ET are often indistinguishable from those of VT.
This study aimed to devise a new algorithm that can distinguish between VT and pre‐ET by comparing the propagation speed in the early portion of the QRS complex according to ECG patterns (left bundle branch block [LBBB] or right bundle branch block [RBBB] pattern). Additionally, it sought to identify the variables that can differentiate between pre‐ET and VT and to compare the algorithms' performance in previous studies.
2. METHODS
2.1. Study design and setting
This retrospective observational cohort study utilized ECG data from a tertiary teaching hospital and a PubMed search of case reports and original studies (Data S1). Patients whose ECGs were diagnosed by electrophysiological study as VT and pre‐ET with a wide QRS tachycardia were included. A total of 130 patients were included after excluding ECGs with obscure or misaligned electrographic lines, those where the measurement unit of ECG duration was not recorded, those with a heart rate below 100 beats/min, and those indicating an orthodromic reentrant circuit through APs. From this cohort, 67 ECGs with VT and 63 ECGs with pre‐ET were identified, comprising 65 RBBB patterns and 65 LBBB patterns. These were randomly subdivided into two groups: 32 ECGs with VT and 32 ECGs with pre‐ET formed the derivation cohort. The remaining 31 ECGs with pre‐ET and 35 ECGs with VT were allocated to the validation cohort. Pre‐ET encompassed not only antidromic AV reentrant tachycardia but also atrial flutter with 2:1 or 1:1 AV conduction and atrial tachycardia with preexcitation. A qualified specialist in both emergency and internal medicine reviewed all ECG data twice. This study was approved by the institutional review board of Dong‐A University Hospital on January 18, 2023 (DAUHIRB‐23‐010). ECG data were anonymized and analyzed in accordance with the approved protocol.
2.2. Data acquisition and definition
The 12‐lead ECGs were analyzed by the author (J. H.), who was not blinded to the final diagnosis. The included ECGs were mainly measured at a sweep speed of 25 mm/s. Any ECGs from PubMed articles measured at different speeds were converted to the standard speed of 25 mm/s. All ECGs were scanned into Adobe Photoshop of version 23 (San Jose, CA), were scaled and aligned if they were askew. All measurements were made using Photoshop's guide shortcut and ruler tool.
Most parameters in previous studies (Ding & Mahida, 2021; Kashou et al., 2021) that assessed the accuracy of ECG algorithms were considered and collected as potential discriminators between VT and pre‐ET. Additionally, to measure the conduction speed at the farthest point from the left ventricle, the propagation components in lead III (as well as those in leads V1, V6, aVR, and II, as included in previous studies) were further investigated in this study.
Baseline ECG characteristics, including heart rate, axis, AV dissociation, positive and negative concordance in precordial lead, late QRS transition in precordial (V5, V6) and limb (II, III, aVF) lead, and importantly parameters related to the propagation speed, were explored. Concordance was defined as having uniformly monophasic polarity in V1 to V6. Late QRS transition in precordial leads denoted observing V4 and V5 where a dominant of S wave in lead V4 and a dominant of R wave in lead V5 were noted. In the context of limb leads, late QRS transition implied a polarity change in leads II, III, and aVF, as per the Cabrera system (Lam et al., 2015). To observe the speed of propagation in the early and late parts of the QRS complex in VT and pre‐ET, durations (ms) up to the initial R or S peak (initial deflection), the largest amplitude deflection, the last deflection, S or Q nadir and R peak were investigated (Figure 1). These values were divided by the QRS duration (ms) and were expressed as the initial deflection index, maximal deflection index (Letsas et al., 2014), last deflection index, RS/QRS ratio (Kim et al., 2021), and Rpeak/QRS ratio (Figure 1). The QRS duration was determined based on the lead with the clearest bidirectional deflection or the lead showing the transition where the margin is observed most clearly (Figures 2 and 3, Red box). Additionally, the boundaries of the QRS complex referred to extension lines of QRS deflections in each lead, notably lead II in the bottom line (Figures 2 and 3). The initial R or S was defined as the initial protruding inflection point [initial r or R in rR(s), Rr(s), rs(r)R, and Rsr pattern of RBBB‐pattern wide‐QRS complex; initial s or S in sS(r), Ss(r), sr(s)S, and Srs pattern of LBBB‐pattern wide‐QRS complex] because of the varied morphologies of the QRS complexes, such as the fragmented QRS complex (Figure 1). Regardless of whether the QRS complex is fragmented, the maximum and last deflection times were measured as the durations up to the S or Q nadir and R peak showing the maximum amplitude and the last deflection, respectively. The RS interval was defined as the interval from the onset of the QRS complex to the nadir of the S wave, and for a QS or Qr pattern, the RS interval was the duration to the nadir of the Q wave (Brugada et al., 1991; Kim et al., 2021). The R peak time was defined as the interval from the onset of the QRS complex to the R peak or S nadir, or to the second peak regardless of the relative height of the R and R' waves in a fragmented QRS complex (Perez‐Riera et al., 2016). The concept of ventricular activation time (De Pooter et al., 2018; Sun et al., 2022) and intrinsicoid deflection (Briceno et al., 2020; Del‐Carpio Munoz et al., 2013; Ye et al., 2022) have various interpretations.
FIGURE 1.
Readings of fragmented QRS complexes: interpretation of initial, maximum, last deflection, RS duration, and Rpeak related to propagation time. III: QRS duration (black to red line); initial defection (black to blue line); maximum deflection (black to blue line); last deflection (black to green line); RS duration (black to green line); Rpeak (black to yellow line). V6: QRS duration (black to red line); initial defection (black to blue line); maximum deflection (black to green line); last deflection (black to green line); RS duration (black to yellow line); Rpeak (black to green line).
FIGURE 2.
Differentiating pre‐ET and VT in LBBB‐pattern wide‐QRS tachycardia cases. QRS duration and deflection times were measured using extension lines (sky blue lines) in the leads exhibiting bidirectional deflection (red box) to accurately identify deflection points, specifically in leads II and V4. (a) LBBB‐pattern pre‐ET (QRS duration = 135.3 ms, initial deflection index was 0.201 [<0.3]). (b) LBBB‐pattern VT (QRS duration = 192.6 ms, initial deflection index was 0.671 [≥0.3]).
FIGURE 3.
Differentiating pre‐ET and VT in RBBB‐pattern wide‐QRS tachycardia cases. QRS duration and deflection times were measured using extension lines (sky blue lines) in the leads exhibiting bidirectional deflection (red box) to accurately identify deflection points, specifically in leads aVR and V1. (a) RBBB‐pattern pre‐ET (QRS duration = 189.1 ms, initial deflection index was 0.632 [>0.48]). (b) RBBB‐pattern VT (QRS duration = 149.1 ms, initial deflection index was 0.46 [≤0.48]).
To evaluate the diagnostic accuracy of prior studies, several algorithms were applied to our electrographic data, including: Steurer, Wellens, Kindwall for the LBBB pattern, Akhtar, Brugada, Vereckei, Pava, Kim for the RBBB pattern, and modified Vereckei (Ding & Mahida, 2021; Kashou et al., 2021; Vereckei et al., 2023). In the present analysis, one of the morphologic criteria from the Wellens algorithm, which stipulates that the R amplitude in wide QRS tachycardia should be larger than the R amplitude in V1, and the criterion from the Aktar algorithm regarding dissimilar QRS morphology during tachycardia compared to baseline preexisting bundle branch block, were excluded and subsequently applied because previous ECGs were not acquired. The two algorithms for comparing VT and pre‐ET, i.e., the Steurer and modified Vereckei algorithms, were directly compared with the newly developed algorithm reflecting propagation speed in the early portion of the QRS complex.
2.3. Statistical analysis
Continuous variables were presented as medians (interquartile ranges [IQRs]) and analyzed by the Mann–Whitney U test. Fisher's exact test was used to compare categorical variables. All electrocardiographic parameters were analyzed for ECGs classified as LBBB and RBBB‐pattern wide‐QRS tachycardia. Receiver operating characteristic analysis was used to determine the optimal cutoff values and the performance of the continuous variables related to propagation speed, including the initial, maximum, and last deflection index, as well as the RS/QRS ratio, and Rpeak/QRS ratio, and to compare the performance of various algorithms, including Steurer, Wellens, Kindwall, Akhtar, Brugada, Vereckei, Pava, Kim, and modified Vereckei. A p value of <0.05 was considered statistically significant.
3. RESULTS
3.1. Baseline characteristics
Patients with pre‐ET had a higher heart rate, shorter QRS duration, less AV dissociation, more positive concordance, and late transition in precordial leads and limb leads as the Cabrera format of the 12‐lead ECG (Table 1). Especially in pre‐ET, heart rate, late transition in precordial leads, and AV dissociation were markedly distinct (p < 0.001). Patients with LBBB‐pattern wide‐QRS pre‐ETs tended to show more left axis deviation, shorter QRS duration, no right axis deviation, and more late transition in precordial and limb leads. Patients with RBBB‐pattern wide‐QRS pre‐ETs tended to present faster heart rate, less left axis deviations, more positive concordance, less late transition in precordial leads.
TABLE 1.
Baseline electrocardiographic characteristics.
| Pre‐ET | VT | p | |
|---|---|---|---|
| Total (N = 63), RBBB/LBBB | Total (N = 67), RBBB/LBBB | ||
| Heart rate, beats/min | 193.4 (163.2, 218.7), 200**/191.8 | 164.9 (139.4, 196.6), 155.5/177.4 | .002 |
| Axis | 168.2 (79.2, 303.2), 189/178.7 | 214.5 (100.6, 290.3), 276.4*/101.9 | .73 |
| Left axis deviation, n (%) | 23 (36.5), 8/15* | 23 (34.3), 17/6 | .855 |
| Right axis deviation, n (%) | 15 (23.8), 15/0 | 20 (29.9), 9/11*** | .553 |
| Northwest axis, n (%) | 7 (11.1), 7/0 | 8 (11.9), 6/2 | 1 |
| QRS duration, ms | 145.9 (128.6, 168), 153.5/144 | 159.2 (140.7, 181.9), 170.1/156.8** | .009 |
| AV dissociation, n (%) | 0 (0), 0/0 | 15 (22.4), 8**/7* | <.001 |
| Positive concordance, n (%) | 21 (33.3), 21**/0 | 9 (13.4), 9/0 | .012 |
| Negative concordance, n (%) | 4 (6.3), 0/4 | 5 (7.5), 0/5 | 1 |
| QRS transition in V5, V6, n (%) | 30 (63.8), 16***/14* | 15 (25), 10/5 | <.001 |
| QRS transition in II, III, aVF, n (%) | 28 (44.4), 5/23*** | 17 (25.4), 9/8 | .027 |
*p < .05, **p < .01, ***p < .001.
3.2. ECG characteristics related to the early propagation speed between pre‐ET and VT
For distinguishing LBBB‐pattern wide‐QRS tachycardia, the initial deflection index in lead V1 and III was a powerful predictor (p = 0.008, AUC 0.773; p = 0.003, AUC 0.805), as shown in Table 2. Conversely, for RBBB‐pattern wide‐QRS tachycardia, the index in lead V1 was most effective (p = 0.001, AUC 0.848), as shown in Table 3. Compared to the maximum, last index, RS/QRS ratio, and Rpeak/QRS ratio, the initial deflection index was more effective in screening VT. Except for leads V1 and III, the maximum deflection index in lead V6, while not statistically significant, was a meaningful predictor for differentiating LBBB‐pattern wide‐QRS tachycardia (p = 0.07, AUC 0.688) and for distinguishing RBBB‐pattern wide‐QRS tachycardia (p = 0.05, AUC 0.703). The RS interval and R peak time were not substantially good predictors.
TABLE 2.
Discrimination of LBBB‐pattern wide‐QRS tachycardia in the derivation cohort.
| Pre‐ET | VT | p | AUC | |
|---|---|---|---|---|
| Initial deflection index in V1 | 0.26 (0.13, 0.44) | 0.48 (0.4, 0.55) | .008 | 0.773 |
| Maximum deflection index in V1 | 0.42 (0.39, 0.54) | 0.5 (0.45, 0.63) | .076 | 0.684 |
| last deflection index in V1 | 0.42 (0.39, 0.54) | 0.53 (0.45, 0.67) | .046 | 0.707 |
| RS/QRS ratio in V1 | 0.42 (0.39, 0.54) | 0.5 (0.45, 0.63) | .076 | 0.684 |
| Rpeak/QRS ratio in V1 | 0.44 (0.4, 0.55) | 0.53 (0.47, 0.67) | .076 | 0.684 |
| Initial deflection index in V6 | 0.44 (0.39, 0.49) | 0.5 (0.28, 0.63) | .366 | 0.594 |
| Maximum deflection index in V6 | 0.45 (0.39, 0.53) | 0.54 (0.46, 0.65) | .07 | 0.688 |
| last deflection index in V6 | 0.57 (0.42, 0.73) | 0.62 (0.51, 0.67) | .598 | 0.555 |
| RS/QRS ratio in V6 | 0.45 (0.39, 0.56) | 0.54 (0.5, 0.69) | .132 | 0.656 |
| Rpeak/QRS ratio in V6 | 0.58 (0.44, 0.71) | 0.62 (0.51, 0.67) | .763 | 0.531 |
| Initial deflection index in aVR | 0.39 (0.34, 0.45) | 0.45 (0.35, 0.54) | .163 | 0.645 |
| Maximum deflection index in aVR | 0.4 (0.35, 0.51) | 0.49 (0.39, 0.57) | .142 | 0.652 |
| last deflection index in aVR | 0.65 (0.47, 0.78) | 0.54 (0.42, 0.69) | .291 | 0.391 |
| RS/QRS ratio in aVR | 0.4 (0.35, 0.51) | 0.52 (0.39, 0.6) | .105 | 0.668 |
| Rpeak/QRS ratio in aVR | 0.61 (0.43, 0.7) | 0.54 (0.42, 0.69) | .763 | 0.469 |
| Initial deflection index in II | 0.32 (0.26, 0.42) | 0.4 (0.29, 0.58) | .083 | 0.68 |
| Maximum deflection index in II | 0.55 (0.31, 0.69) | 0.53 (0.37, 0.66) | .91 | 0.512 |
| last deflection index in II | 0.65 (0.54, 0.7) | 0.63 (0.51, 0.74) | .821 | 0.477 |
| RS/QRS ratio in II | 0.65 (0.34, 0.7) | 0.57 (0.38, 0.7) | .474 | 0.426 |
| Rpeak/QRS ratio in II | 0.61 (0.47, 0.7) | 0.58 (0.4, 0.71) | .91 | 0.488 |
| Initial deflection index in III | 0.29 (0.24, 0.38) | 0.49 (0.34, 0.61) | .003 | 0.805 |
| Maximum deflection index in III | 0.65 (0.35, 0.7) | 0.57 (0.39, 0.65) | .327 | 0.398 |
| last deflection index in III | 0.69 (0.64, 0.74) | 0.6 (0.52, 0.67) | .032 | 0.277 |
| RS/QRS ratio in III | 0.69 (0.6, 0.74) | 0.58 (0.47, 0.66) | .055 | 0.301 |
| Rpeak/QRS ratio in III | 0.65 (0.56, 0.7) | 0.58 (0.47, 0.66) | .175 | 0.359 |
Note: Variable values were presented as medians (interquartile ranges).
TABLE 3.
Discrimination of RBBB‐pattern wide‐QRS tachycardia in the derivation cohort.
| PreET | VT | p | AUC | |
|---|---|---|---|---|
| Initial deflection index in V1 | 0.54 (0.51, 0.66) | 0.34 (0.24, 0.47) | .001 | 0.848 |
| Maximum deflection index in V1 | 0.6 (0.52, 0.7) | 0.6 (0.49, 0.68) | .572 | 0.559 |
| last deflection index in V1 | 0.7 (0.66, 0.76) | 0.7 (0.62, 0.81) | 1 | 0.5 |
| RS/QRS ratio in V1 | 0.6 (0.52, 0.7) | 0.65 (0.52, 0.78) | .678 | 0.457 |
| Rpeak/QRS ratio in V1 | 0.72 (0.67, 0.76) | 0.66 (0.59, 0.75) | .2 | 0.633 |
| Initial deflection index in V6 | 0.34 (0.26, 0.46) | 0.34 (0.21, 0.42) | .366 | 0.594 |
| Maximum deflection index in V6 | 0.57 (0.45, 0.62) | 0.46 (0.39, 0.58) | .05 | 0.703 |
| last deflection index in V6 | 0.69 (0.56, 0.75) | 0.67 (0.58, 0.76) | .792 | 0.527 |
| RS/QRS ratio in V6 | 0.62 (0.55, 0.7) | 0.59 (0.46, 0.71) | .407 | 0.586 |
| Rpeak/QRS ratio in V6 | 0.58 (0.45, 0.73) | 0.49 (0.39, 0.62) | .142 | 0.652 |
| Initial deflection index in aVR | 0.35 (0.25, 0.5) | 0.35 (0.28, 0.4) | .598 | 0.555 |
| Maximum deflection index in aVR | 0.52 (0.35, 0.68) | 0.43 (0.35, 0.54) | .498 | 0.57 |
| last deflection index in aVR | 0.73 (0.64, 0.82) | 0.7 (0.56, 0.76) | .386 | 0.59 |
| RS/QRS ratio in aVR | 0.52 (0.35, 0.68) | 0.43 (0.35, 0.54) | .498 | 0.57 |
| Rpeak/QRS ratio in aVR | 0.52 (0.39, 0.74) | 0.56 (0.39, 0.7) | .792 | 0.473 |
| Initial deflection index in II | 0.34 (0.18, 0.43) | 0.34 (0.29, 0.42) | .678 | 0.457 |
| Maximum deflection index in II | 0.55 (0.39, 0.72) | 0.46 (0.41, 0.66) | .346 | 0.598 |
| last deflection index in II | 0.75 (0.54, 0.83) | 0.66 (0.5, 0.78) | .327 | 0.602 |
| RS/QRS ratio in II | 0.58 (0.39, 0.8) | 0.46 (0.41, 0.66) | .243 | 0.621 |
| Rpeak/QRS ratio in II | 0.72 (0.54, 0.83) | 0.69 (0.51, 0.78) | .429 | 0.582 |
| Initial deflection index in III | 0.27 (0.17, 0.42) | 0.24 (0.2, 0.4) | .651 | 0.547 |
| Maximum deflection index in III | 0.56 (0.39, 0.59) | 0.44 (0.42, 0.52) | .214 | 0.629 |
| Last deflection/QRS in III | 0.69 (0.57, 0.82) | 0.67 (0.46, 0.78) | .624 | 0.551 |
| RS/QRS ratio in III | 0.56 (0.39, 0.59) | 0.44 (0.42, 0.56) | .346 | 0.598 |
| Rpeak/QRS ratio in III | 0.71 (0.44, 0.83) | 0.65 (0.43, 0.79) | .97 | 0.504 |
Note: Variable values were presented as medians (interquartile ranges).
3.3. Differential diagnosis using previously published algorithms and a newly developed algorithm
Different algorithms were tested on our data to differentiate VT from SVT. However, those performed poorly in distinguishing between pre‐ET and VT as the AUCs were <0.65 (Figure 4a). The lead V1‐III index and V1‐V1 algorithm outperformed both the Steurer and modified Verechei algorithms in entire ECGs showing wide QRS tachycardia (Figure 4b) and in ECGs showing each LBBB and RBBB pattern wide QRS tachycardia (Figure 4c,d). Although the Steurer and modified Verechei algorithms demonstrated better performance (AUC 0.734 and 0.672) than other previously published algorithms, those efficacies dropped in LBBB‐pattern wide‐QRS tachycardia (AUC 0.625 and 0.625), as shown in Figure 4. For LBBB‐pattern wide‐QRS tachycardia, only the Akhtar, Pava, and Kindwall algorithms were decent predictors in distinguishing pre‐ET from VT (AUC 0.688 in all). The most robust algorithm was the lead V1‐III index with an AUC of 0.828. To distinguish LBBB‐pattern wide‐QRS tachycardia, this algorithm employed an initial deflection index in lead III. A cutoff value >0.3 indicated VT with an AUC of 0.781, signifying initial slow conduction after ventricular activity (Figure 4c). To differentiate RBBB‐pattern wide‐QRS tachycardia, the algorithm utilized an initial deflection index in lead V1. A cutoff value of ≤0.48 indicated VT with an AUC of 0.869, denoting initial fast conduction after ventricular activity (Figure 4d). Additionally, the lead V1‐V1 index algorithm emerged as promising for VT screening (AUC 0.813), with cutoff values of >0.27 for LBBB‐pattern and ≤0.48 for RBBB‐pattern. Enhanced performance of these algorithms was further demonstrated in the validation cohort, with AUCs of 0.844 and 0.816 (Figure 5).
FIGURE 4.
Performance of algorithms using ROC curves to distinguish between SVT versus VT and pre‐ET versus VT in the derivation cohort, with separate analyses for patterns of both LBBB and RBBB, LBBB alone, and RBBB alone. (a) Previous algorithm performance in wide QRS tachycardia for SVT. (b) Performance of the initial deflection index, Steurer, and modified Vereckei algorithm for pre‐ET. (c) Algorithm performance in LBBB‐pattern wide‐QRS tachycardia. (d) Algorithm performance in RBBB‐pattern wide‐QRS tachycardia.
FIGURE 5.
Performance using ROC curves in differentiating between pre‐ET and VT in the validation cohort.
4. DISCUSSION
Compared with previous algorithms, the algorithms that exhibited the best performance in distinguishing between VT and pre‐ET were the lead V1‐III and V1‐V1 index algorithms. The early propagation speed of pre‐ET was typically faster than that of VT in LBBB‐pattern wide‐QRS tachycardia and slower in RBBB‐pattern wide‐QRS tachycardia.
Previous algorithms for distinguishing between SVT and VT incorporated criteria related to slow propagation in the early portion of the QRS complex. Examples include: R ≥ 30 ms and S nadir >60 or 70 ms in lead V1 or V2 as proposed by the Wellens, Kindwall, and Brugada algorithms; RS interval ≥ 100 ms in precordial leads as per the Brugada algorithm; q > 40 ms, Vi/Vt ≤1 in lead aVR according to the Vereckei algorithm; R wave peak time ≥40 or 50 ms in lead II following the Pava algorithm; and RS/QRS ratio in lead V6 as presented in the Kim algorithm (Ding & Mahida, 2021). Overall, the initial deflection index demonstrated superior performance in discriminating VT from pre‐ET compared to the maximum deflection index, RS/QRS ratio, and R peak/QRS ratio. This discrepancy in criteria, other than the initial deflection index, might arise from QRS fragmentation observed in VT or pre‐ET. QRS complex fragmentation may occur due to myocardial dysfunction (MacAlpin, 2010), and such fragmentation can alter the duration of deflections in QRS complex. The early propagation speed, represented by initial deflection index, might have a limited effect on the fragmented QRS complex. Another issue in previous algorithms for distinguishing between SVT and VT is the portion of pre‐ET cases included in the SVT population. If the portion of pre‐ET cases in an SVT population is relatively high, the sensitivity and specificity of previous algorithms for diagnosing VT may be reduced as indicated by the accuracy (68.8%–77.5%) determined in a study including 23 pre‐ETs cases among 101 SVT cases (22.8%) (Jastrzebski et al., 2012). The early propagation speed of pre‐ET is difficult to be distinguished from that of VT; however, it is relatively easy to distinguish that of SVT excluding pre‐ET, which may be explained through studies on previous algorithms; consequently, pre‐ET should be analyzed separately from SVT with aberration.
It is a complex matter as to why the LBBB pattern VT displayed slower propagation during the early portion of the QRS complex than the LBBB pattern pre‐ET, and conversely, why the RBBB pattern VT had a quicker early propagation speed than the RBBB pattern pre‐ET. In cases of SVT, the difference of propagation speed in the early portion of the QRS complex between SVT and VT was presumed due to the absence of organic heart disease other than differences in distance from the His‐Purkinje system (Jastrzębski et al., 2018). Some VT‐specific features (qR in leads V2–V6, S‐wave notch in lead V1/aVR, monophasic R in lead aVR, no RS in leads V1–V6) may be exhibited in VT because patients with VT tend to have myocardial scar due to organic heart disease (Jastrzębski et al., 2018). Nevertheless, initial slow propagation due to a distance difference outside of the His‐Purkinje system was demonstrated. The time interval from a far‐field RV lead to a near‐field RV lead using the classic conduction pathway (AV node, His, Purkinje) was short in patients with SVT, but prolonged in patients with VT (Kapoor et al., 2020). Similar to SVT, the changes in early propagation speed in pre‐ET and VT might be affected not only by the distance difference from the His‐Purkinje system but also by myocardial dysfunction or scar formation near the pathway tract.
The conduction mechanism of VT and pre‐ET, as well as the sites where VT and pre‐ET commonly occur in the heart, should be considered to understand the propagation speed in the early portion of the QRS complex. The AP was commonly located in the left free wall, especially the left lateral wall (Brembilla‐Perrot, Moejezi, et al., 2013) although approximately 10% of all patients had multiple Aps (Gallagher et al., 1979). Septal APs, especially those involving the posteroseptal wall, were the next most common, and APs of the right free wall and other locations occurred with less frequency. Some patients with WPW syndrome have accessory pathways in which conduction slows at faster rates of stimulation (decremental conduction); however, this ultimately leads to very fast ventricular rates because APs characteristically do not demonstrate decremental conduction. Rarely, APs that show decremental conduction in the antegrade direction give rise to wide‐QRS tachycardia, and this phenomenon usually occurs in the posteroseptal wall (de Chillou et al., 1992). Most antidromic AVRTs function via a left‐sided accessory pathway (RBBB pattern) as the antegrade route for conduction (Cain et al., 1992). Fascicular VT (RBBB pattern) is the most common form of idiopathic VT arising from the left fascicular system (Tanawuttiwat et al., 2016). Left‐sided focal VT with structural heart disease originates from the basal lateral wall of the left ventricle (50%), posteromedial papillary muscle (33.3%), and anterolateral papillary muscle (16.7%) (Anderson et al., 2020). Right ventricular outflow tract (RVOT) VT (LBBB pattern) accounts for 70%–80% of idiopathic VT, and less commonly outflow tract VTs originate from the left ventricular outflow tract (LVOT), aortic cusps, and aortic mitral continuity (Tanawuttiwat et al., 2016). Most VT patients have structural heart disease (90%), and the site and corresponding frequency of right‐sided focal VT with myocardial scar are as follows: moderator band (33.3%); RVOT anteroseptal wall (25%); RVOT free wall (16.7%); RVOT septal wall (8.3%); lateral tricuspid annulus (8.3%); and apical septal right ventricle (8.3%) (Anderson et al., 2020). On the other hand, among 158 patients with antegrade AP conduction and right‐sided pre‐ET, the distribution was as follows: right posterior (29%), lateral (19.4%), posteroseptal (19.4%), midseptal (12.9%), anterior (11.3%), and anterospetal (8.1%) APs (Liu et al., 2017). Overall, LBBB‐pattern VT originates from the anterior or septal aspect of the RVOT or moderator band, while LBBB‐pattern pre‐ET inserts on the tricuspid annulus. The reason for the longer early propagation speed in pre‐ET than in VT in RBBB‐pattern wide‐QRS tachycardia might be related to differences in the propagation speed via the common pathways of pre‐ET (left‐sided accessory pathway in RBBB pattern) and VT (either the basal lateral wall of the left ventricle or the left fascicular system in RBBB pattern). Similarly, the reason for the shorter early propagation speed in pre‐ET compared to VT in LBBB‐pattern wide‐QRS tachycardia might be related to differences in the propagation speed via the common pathways of pre‐ET (the tricuspid annulus) and VT (either RVOT or moderator band). The common conduction locations and conduction pathways may affect the early propagation speed.
Discriminating between VT and preET using an ECG is crucial for guiding diagnostic and therapeutic strategies. While an ECG can be used for screening, a definitive diagnosis ultimately requires an electrophysiological study. It is important to remember that initially distinguishing VT from preET with an ECG always provides insufficient information.
This study had several limitations. First, this retrospective study was conducted at a single center using a small sample. Implementing a prospective multicenter study might be challenging due to the low incidence of pre‐ET and difficulties in collecting data on both LBBB and RBBB pattern pre‐ET. Second, the intraobserver and interobserver reliabilities of the measures were not evaluated. Although we provided predefined criteria and definitions for ECG interpretation, the reproducibility of algorithmic results likely depends on the examiner's experience in reading ECGs. Third, the results of two prior algorithms might be inaccurate because the criteria requiring data from a previous ECG in the Wellens and Akhtar algorithms were omitted. Finally, ECGs sourced from PubMed may not represent typical cases, as those presented as case reports could be outliers.
5. CONCLUSION
The lead V1‐III index algorithm for discriminating between VT and pre‐ET exhibited better performance than previous algorithms. In LBBB‐pattern wide‐QRS tachycardia, the propagation speed in the early portion of the QRS complex for pre‐ET was generally faster than that for VT, while in RBBB‐pattern wide‐QRS tachycardia, it was generally slower. VT should be distinguished with pre‐ET separately from SVT with aberration in discriminative algorithms.
AUTHOR CONTRIBUTIONS
Lee JH conceived, drafted, planned, investigated, and managed the manuscript concerning all aspects of the study. Lee JH contributed to data acquisition, data analysis, interpretation, statistical analysis, revisions. The author has read and approved the final version of the manuscript.
CONFLICT OF INTEREST STATEMENT
None.
ETHICS STATEMENT
The study design and plan, including the informed consent form, were approved by the institutional review board of each hospital (DAUHIRB‐23‐010). In accordance with national requirements and the principles of the Declaration of Helsinki, the current study was performed. Written informed consent was not required.
Supporting information
Data S1.
ACKNOWLEDGMENTS
This work was supported by the Dong‐A University Research Fund.
Lee, J. H. (2024). Discrimination between ventricular tachycardia and wide‐QRS preexcited tachycardia. Annals of Noninvasive Electrocardiology, 29, e13112. 10.1111/anec.13112
DATA AVAILABILITY STATEMENT
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
REFERENCES
- Ali, R. , Tahir, A. , Nadeem, M. , Shakhatreh, M. I. , & Faulknier, B. (2018). Antidromic atrioventricular reentry tachycardia with Wolff Parkinson white syndrome: A rare beast. Cureus, 10(5), e2642. 10.7759/cureus.2642 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Anderson, R. D. , Lee, G. , Trivic, I. , Campbell, T. , Pham, T. , Nalliah, C. , Kizana, E. , Thomas, S. P. , Trivedi, S. J. , Watts, T. , Kalman, J. , & Kumar, S. (2020). Focal ventricular Tachycardias in structural heart disease: Prevalence, characteristics, and clinical outcomes after catheter ablation. JACC. Clinical Electrophysiology, 6(1), 56–69. 10.1016/j.jacep.2019.09.013 [DOI] [PubMed] [Google Scholar]
- Anderson, R. H. , Becker, A. E. , Brechenmacher, C. , Davies, M. J. , & Rossi, L. (1975). Ventricular preexcitation. A proposed nomenclature for its substrates. European Journal of Cardiology, 3(1), 27–36. [PubMed] [Google Scholar]
- Brady, W. J. , Mattu, A. , Tabas, J. , & Ferguson, J. D. (2017). The differential diagnosis of wide QRS complex tachycardia. The American Journal of Emergency Medicine, 35(10), 1525–1529. 10.1016/j.ajem.2017.07.056 [DOI] [PubMed] [Google Scholar]
- Brembilla‐Perrot, B. , Moejezi, R. V. , Zinzius, P. Y. , Jarmouni, S. , Schwartz, J. , Beurrier, D. , Sellal, J. M. , Nossier, I. , Muresan, L. , Andronache, M. , Moisei, R. , Selton, O. , Louis, P. , & de la Chaise, A. T. (2013). Missing diagnosis of preexcitation syndrome on ECG: Clinical and electrophysiological significance. International Journal of Cardiology, 163(3), 288–293. 10.1016/j.ijcard.2011.06.051 [DOI] [PubMed] [Google Scholar]
- Brembilla‐Perrot, B. , Pauriah, M. , Sellal, J. M. , Zinzius, P. Y. , Schwartz, J. , de Chillou, C. , Cismaru, G. , Beurrier, D. , Voilliot, D. , Selton, O. , Louis, P. , Andronache, M. , Nosu, R. , & de la Chaise, A. T. (2013). Incidence and prognostic significance of spontaneous and inducible antidromic tachycardia. Europace, 15(6), 871–876. 10.1093/europace/eus354 [DOI] [PubMed] [Google Scholar]
- Briceno, D. F. , Santangeli, P. , Frankel, D. S. , Liang, J. J. , Shirai, Y. , Markman, T. , Enriquez, A. , Walsh, K. , Riley, M. P. , Nazarian, S. , Lin, D. , Kumareswaran, R. , Arkles, J. S. , Hyman, M. C. , Deo, R. , Supple, G. E. , Garcia, F. C. , Dixit, S. , Epstein, A. E. , … Schaller, R. D. (2020). QRS morphology in lead V(1) for the rapid localization of idiopathic ventricular arrhythmias originating from the left ventricular papillary muscles: A novel electrocardiographic criterion. Heart Rhythm, 17(10), 1711–1718. 10.1016/j.hrthm.2020.05.021 [DOI] [PubMed] [Google Scholar]
- Brugada, P. , Brugada, J. , Mont, L. , Smeets, J. , & Andries, E. W. (1991). A new approach to the differential diagnosis of a regular tachycardia with a wide QRS complex. Circulation, 83(5), 1649–1659. 10.1161/01.cir.83.5.1649 [DOI] [PubMed] [Google Scholar]
- Cain, M. E. , Luke, R. A. , & Lindsay, B. D. (1992). Diagnosis and localization of accessory pathways. Pacing and Clinical Electrophysiology, 15(5), 801–824. 10.1111/j.1540-8159.1992.tb06847.x [DOI] [PubMed] [Google Scholar]
- de Chillou, C. , Rodriguez, L. M. , Schläpfer, J. , Kappos, K. G. , Katsivas, A. , Baiyan, X. , Smeets, J. L. , & Wellens, H. J. (1992). Clinical characteristics and electrophysiologic properties of atrioventricular accessory pathways: Importance of the accessory pathway location. Journal of the American College of Cardiology, 20(3), 666–671. 10.1016/0735-1097(92)90022-f [DOI] [PubMed] [Google Scholar]
- De Pooter, J. , El Haddad, M. , Kamoen, V. , Kallupurackal, T. T. , Stroobandt, R. , De Buyzere, M. , & Timmermans, F. (2018). Relation between electrical and mechanical dyssynchrony in patients with left bundle branch block: An electro‐ and vectorcardiographic study. Annals of Noninvasive Electrocardiology, 23(4), e12525. 10.1111/anec.12525 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Del‐Carpio Munoz, F. , Powell, B. D. , Cha, Y. M. , Wiste, H. J. , Redfield, M. M. , Friedman, P. A. , & Asirvatham, S. J. (2013). Delayed intrinsicoid deflection onset in surface ECG lateral leads predicts left ventricular reverse remodeling after cardiac resynchronization therapy. Heart Rhythm, 10(7), 979–987. 10.1016/j.hrthm.2013.03.045 [DOI] [PubMed] [Google Scholar]
- Ding, W. Y. , & Mahida, S. (2021). Wide complex tachycardia: Differentiating ventricular tachycardia from supraventricular tachycardia. Heart, 107(24), 1995–2003. 10.1136/heartjnl-2020-316874 [DOI] [PubMed] [Google Scholar]
- Fitzsimmons, P. J. , McWhirter, P. D. , Peterson, D. W. , & Kruyer, W. B. (2001). The natural history of Wolff‐Parkinson‐white syndrome in 228 military aviators: A long‐term follow‐up of 22 years. American Heart Journal, 142(3), 530–536. 10.1067/mhj.2001.117779 [DOI] [PubMed] [Google Scholar]
- Gallagher, J. J. , Sealy, W. C. , & Kasell, J. (1979). Intraoperative mapping studies in the Wolff‐Parkinson‐white syndrome. Pacing and Clinical Electrophysiology, 2(5), 523–537. 10.1111/j.1540-8159.1979.tb04261.x [DOI] [PubMed] [Google Scholar]
- Jastrzebski, M. , Kukla, P. , Czarnecka, D. , & Kawecka‐Jaszcz, K. (2012). Comparison of five electrocardiographic methods for differentiation of wide QRS‐complex tachycardias. Europace, 14(8), 1165–1171. 10.1093/europace/eus015 [DOI] [PubMed] [Google Scholar]
- Jastrzębski, M. , Moskal, P. , Kukla, P. , Fijorek, K. , Kisiel, R. , & Czarnecka, D. (2018). Specificity of wide QRS complex tachycardia criteria and algorithms in patients with ventricular preexcitation. Annals of Noninvasive Electrocardiology, 23(2), e12493. 10.1111/anec.12493 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kapoor, R. , Tyagi, S. , Dohmen, C. , Oujiri, J. , Roth, J. , Rubenstein, J. C. , & Berger, M. (2020). Tachyarrhythmia discriminator for implantable cardioverter‐defibrillators in bundle branch block. Heart Rhythm, 17(9), 1561–1565. 10.1016/j.hrthm.2020.04.031 [DOI] [PubMed] [Google Scholar]
- Kashou, A. H. , Evenson, C. M. , Noseworthy, P. A. , Muralidharan, T. R. , DeSimone, C. V. , Deshmukh, A. J. , Asirvatham, S. J. , & May, A. M. (2021). Differentiating wide complex tachycardias: A historical perspective. Indian Heart Journal, 73(1), 7–13. 10.1016/j.ihj.2020.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim, M. , Kwon, C. H. , Lee, J. H. , Hwang, K. W. , Choi, H. O. , Kim, Y. G. , Lee, K. N. , Ahn, J. , Park, H. S. , & Nam, G. B. (2021). Right bundle branch block‐type wide QRS complex tachycardia with a reversed R/S complex in lead V(6): Development and validation of electrocardiographic differentiation criteria. Heart Rhythm, 18(2), 181–188. 10.1016/j.hrthm.2020.08.023 [DOI] [PubMed] [Google Scholar]
- Lam, A. , Wagner, G. S. , & Pahlm, O. (2015). The classical versus the Cabrera presentation system for resting electrocardiography: Impact on recognition and understanding of clinically important electrocardiographic changes. Journal of Electrocardiology, 48(4), 476–482. 10.1016/j.jelectrocard.2015.05.011 [DOI] [PubMed] [Google Scholar]
- Letsas, K. P. , Asvestas, D. , Vlachos, K. , Karlis, D. , Korantzopoulos, P. , Efremidis, M. , & Sideris, A. (2014). Electrocardiographic characteristics of premature ventricular contractions in subjects with type 1 pattern of Brugada syndrome. Journal of Electrocardiology, 47(3), 351–355. 10.1016/j.jelectrocard.2014.02.012 [DOI] [PubMed] [Google Scholar]
- Liu, R. , Chen, Q. I. , Chen, Y. , Zhang, Y. , Xu, Z. , & Wang, G. (2017). Effects of Antegrade accessory pathway conduction on QRS terminal vector in patients with Preexcitation syndrome. Pacing and Clinical Electrophysiology, 40(3), 264–270. 10.1111/pace.13025 [DOI] [PubMed] [Google Scholar]
- MacAlpin, R. N. (2010). The fragmented QRS: Does it really indicate a ventricular abnormality? Journal of Cardiovascular Medicine (Hagerstown, Md.), 11(11), 801–809. 10.2459/JCM.0b013e32833b9816 [DOI] [PubMed] [Google Scholar]
- Perez‐Riera, A. R. , de Abreu, L. C. , Barbosa‐Barros, R. , Nikus, K. C. , & Baranchuk, A. (2016). R‐peak time: An electrocardiographic parameter with multiple clinical applications. Annals of Noninvasive Electrocardiology, 21(1), 10–19. 10.1111/anec.12323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Steurer, G. , Gursoy, S. , Frey, B. , Simonis, F. , Andries, E. , Kuck, K. , & Brugada, P. (1994). The differential diagnosis on the electrocardiogram between ventricular tachycardia and preexcited tachycardia. Clinical Cardiology, 17(6), 306–308. 10.1002/clc.4960170606 [DOI] [PubMed] [Google Scholar]
- Sun, W. , Upadhyay, G. A. , & Tung, R. (2022). Influence of capture selectivity and left Intrahisian block on QRS characteristics during left bundle branch pacing. JACC. Clinical Electrophysiology, 8(5), 635–647. 10.1016/j.jacep.2022.01.012 [DOI] [PubMed] [Google Scholar]
- Tanawuttiwat, T. , Nazarian, S. , & Calkins, H. (2016). The role of catheter ablation in the management of ventricular tachycardia. European Heart Journal, 37(7), 594–609. 10.1093/eurheartj/ehv421 [DOI] [PubMed] [Google Scholar]
- Vereckei, A. , Simon, A. , Szenasi, G. , & Miller, J. M. (2023). The application of a new, modified algorithm for the differentiation of regular ventricular and pre‐excited Tachycardias. Heart, Lung & Circulation, 32(6), 719–725. 10.1016/j.hlc.2023.03.016 [DOI] [PubMed] [Google Scholar]
- Ye, Z. , Xu, Z. , Gong, K. , Chen, X. , Wang, W. , Chen, J. , Chen, L. , & Zhang, F. (2022). Electrocardiogram parameters that affect the success rate of radiofrequency ablation in patients with outflow tract ventricular premature complexes. Cardiovascular Therapeutics, 2022, 8160144. 10.1155/2022/8160144 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1.
Data Availability Statement
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.